How Sound Transit Cheated During Planning to promote Light Rail

i
How Sound Transit Abused the Planning Process
to Promote Light Rail
and also
CETA Comments on DSEIS for Sound Transit’s
“Regional Transit Long-Range Plan”
by
R.C. Harkness, PhD Urban Systems Planning, member CETA Technical Committee
January, 2005
This report in its entirety is submitted as one of the Coalition for Effective
Transportation Alternatives (CETA) comments on Sound Transit’s DSEIS for their
Regional Transit Long-Range Plan as solicited in Sound Transit’s news release
dated Dec. 2, 2004 and on the web as:
http://www.soundtransit.org/newsroom/releases/pr_20041202_1.asp
Abstract
Areas of weakness and apparent bias in the planning process behind Seattle's planned
“Link” light rail system are identified. Problems with Sound Transit’s one and only
alternatives analysis – a 1993 FEIS comparing rapid rail and bus alternatives-- are
documented. It is concluded that there is no valid alternatives analysis behind Link light
rail. Link’s performance is summarized using objective data and qualitative comments.
Key advantages of BRT are described. Particular attention is given to the issue of rail
versus bus capacity, and how it was mishandled in the 1993 FEIS in order to favor rail.
Raw cost-effectiveness data in the 1993 FEIS is reformatted and presented in a way that
better communicates the relative merits of the rail and bus alternatives. The bus
alternative is modified to overcome alleged capacity deficiencies, and then re-compared
with the rail alternative. It is found that a modified bus alternative would have had more
than enough capacity to meet long-term needs and could have achieved the same
ridership as rapid rail, at far lower cost. Results are brought forward in order to estimate
the costs of the full regional light rail system envisioned by Sound Transit versus a BRT
alternative. It is concluded that switching to the BRT alternative would save the Puget
Sound region about $900 million dollars a year over a 30 year period. It is argued that
Link is the failed result of a faulty planning process. The effects on public trust in
government are documented with newspaper quotes, as are requests made of Sound
Transit to consider alternatives to light rail. Detailed recommendations are offered on
how to improve the process and conduct the proper alternatives analysis that is still
needed.
ii
Funding and Sponsors
This report was written by the author as a community service. There has been no outside
sponsorship or funding.
Author’s Background
BSEE Duke University; LtJg US Navy; MSE Civil Engineering/Urban Trans. Planning
Univ. of Washington; summer work at Trans. Dept. of Puget Sound Regional
Conference; PhD Urban Systems Planning, Univ. of Washington; market researcher for
Boeing’s Morgantown PRT project; Policy Analyst and project manager at Stanford
Research Institute for large National Science Foundation study on Telecommunications
Substitutes for Transportation; various planning and management positions in strategy
and new business/product planning in telecom industry and Boeing. Life Member Sierra
Club. Now retired but active in CETA and Eastside Transportation Association.
Author’s Email: harknessr@w-link.net
Web Sites for additional information on Link light rail
http://www.effectivetransportation.org/ (CETA site)
http://www.globaltelematics.com/pitf/index.htm
Copies of this report
Copies of this report can be obtained from Sound Transit since Sound Transit is required
to publish comments received on the DSEIS. Copies may also be obtained from CETA.
See CETA web site for contact information.
iii
Table of Contents Page
Specific requests for Sound Transit’s reply to these DSEIS comments ……...…….. iv
Executive Summary …………………………………………………………………. v
Extended Summary …………………………………………………….……………vii
Main Report
Part 1: Sound Transit has never done a proper alternatives analysis …………….1
Part 2: Sound Transit morphed rapid rail into light rail ………………………....19
Part 3: The shortcomings of Link Light Rail …………………………………… 23
Part 4: Bus rapid transit (BRT) and other alternatives ………………………….. 39
Part 5: The Capacity Issue ………………………………………………………. 49
Part 6: The Cost Issue …………………………………………………………… 81
Part 7: Evaluation and Recommendations …………………………………..……97
Part 8: Guidelines for a proper alternatives analysis …………………………….123
Part 9: Conclusions …………………...………………………………………….133
Brief Introduction to Least Cost Planning ……………………………………………..135
Tech Notes (includes the cost calculations behind the charts) ………………………………………..139
References ……………………………………………………………………………. 145
Key Abbreviations:
ST = Sound Transit, the agency charged with detailed planning, implementation of high capacity
transit in the Puget Sound area, known officially as the Regional Transit Authority or RTA.
RTP = Regional Transit Project, the name of the planning project taken over by RTA and whose main
product was the 1993 FEIS for rapid rail (the author uses the terms RTA and RTP somewhat
interchangeably)
PSRC = Puget Sound Regional Council, the agency responsible for strategic transportation planning,
and the regions official MPO
EIS = Environmental Impact Statement, FEIS = Final EIS, as opposed to draft EIS
DSEIS= Draft Supplemental EIS, in this case for Sound Transits “Regional Transit Long-Range Plan”
iv
Specific requests to Sound Transit concerning it’s response
to these CETA comments on the DSEIS
As Sound Transit knows, CETA members have almost certainly spent more effort
over the last several years scrutinizing Sound Transit’s light rail plans than any other
organization.
Due to the importance of Sound Transit’s Draft Long-Range Plan DSEIS a large
amount of effort has gone into preparation of this document and CETA would
appreciate a commensurately detailed and through response.
Individual and specific responses are requested to the following:
* Parts 1.5 thru 1.15 pertaining to specific deficiencies in the 1993 FEIS
* Parts 3.1 through 3.16
* Each of the eight points in Part 5.4.9
* Each of the reports three key recommendations at the beginning of Part 7.4
* Parts 8.2.2 and 8.3
* Each of the report’s main conclusions as listed in the Executive Summary and
Part 9.
In responding we would urge Sound Transit to consider not only what existing
regulations may or may not require the agency to do, but also Sound Transit’s moral
obligation to provide citizens and public officials with the sufficient and objective
information they need to make intelligent decisions about mass transit.
v
Executive Summary
Sound Transit’s Board is increasingly committing this region to use light rail, as opposed
to bus rapid transit, for the region’s mass transit backbone. In 1996 voters approved
spending $1.8 billion for a 21-mile light rail system. In 2001 Sound Transit admitted their
initial cost estimates were wrong and shortened the line to 14-mile miles. However ST
still hopes to eventually build over 125 miles of light rail and is taking administrative steps
toward doing so.
Sound Transit justified its choice of light rail technology on an alternatives analysis done
in 1993 by Sound Transit’s predecessor the RTP. That particular study compared a 125-
mile rapid rail system costing $11.5 billion against an express bus alternative costing $4.7
billion.
During the course of that study RTP predicted year 2020 ridership for both alternatives,
and then evaluated their capacity to handle the predicted ridership. In looking at bus
system capacity through downtown Seattle the RTP assumed the bus tunnel could only
carry 100 buses per hour in each direction, although six previous studies had concluded its
capacity was significantly higher than that.
At this point the RTP had a choice. Either it could verify that 100 was the correct value,
and if so, apply one of the remedies that staff had already identified. Or it could penalize
the bus alternative. It chose the latter, and proceeded to reduce predicted ridership for the
bus alternative and claim it didn’t have enough capacity to meet the region’s needs. In
addition, because the bus alternative now had lower ridership the RTA also down-rated it
on all other ridership related benefits, such as its ability to improve mobility and support
land use goals.
In short, the RTP compared a robust rapid rail alternative against a deliberately hobbled
bus alternative and used the results to rule-out bus technology for the region’s main transit
spine along I-5. In its recently released Draft Long-Range Plan Sound Transit still relies
on that corrupted and now obsolete study to justify proceeding with light rail.
Fortunately, it is possible to estimate what would have happened if RTP had elected to
remedy the alleged capacity problem rather than penalize the bus alternative. However, to
compare apples-to-apples it was necessary to have two alternatives that are either equal in
benefit or equal in cost. Therefore the author elected to modify the bus alternative so it
would attract the same ridership as the rail alternative, then compare costs. The first step
was to remove the alleged capacity bottleneck using –to be conservative-- the most
expensive remedy identified by staff, namely building a second parallel bus tunnel costing
$600 million. This allowed the bus alternative to carry its originally predicted ridership,
which was 93 % of what the rail system was predicted to carry. To get that last 7% the
author used an RTP estimate for the cost of attracting extra riders.
The result is that a modified bus alternative would be $400 million per year less expensive
(in 1991$) than the rapid rail system chosen by RTP. This is the picture that the RTP
vi
could have produced using information available at that time. However, RTP chose not to
do so because officials wanted rail to win. What can that study tell us today when we are
concerned with light rail, not rapid rail?
Today there is every indication Sound Transit’s Board wants to build at least 125 miles of
light rail. There has never been an apples-to-apples comparison between bus rapid transit
(BRT) and any of the different size light rail networks that Sound Transit is
contemplating, much less a 125-mile system. However; it is possible to make an
approximation.
First, it was assumed that 125 miles of Link light rail would attract as many riders as 125
miles of rapid rail. Clearly, it wouldn’t because Link’s slower, but this is the conservative
approach. The remaining task was to estimate the cost of a 125-mile version of Link and
compare that with an all-bus or BRT alternative. The 1993 cost comparison can be
reused, but only after adjusting it for the facts that Sound Transit’s early rail cost estimates
were 44% too low and that many of the HOV lanes needed for the bus alternative have
now been completed. The results show that a 125-mile light rail system would cost about
$900 million (02$) per year more than a comparable BRT system. This cost differential
would continue over the 30-year period needed to repay the construction bonds. It’s also
likely that BRT could replace the 14- mile Initial Segment or the 21-mile Central Link
system for less than half their costs.
This information about a potential $900 million per year savings opportunity is new, and it
needs to be published widely so taxpayers can decide whether it makes better sense to
abandon Sound Transit’s light rail strategy and switch to an equally effective BRT
alternative.
It’s unfortunate that knowledge of this opportunity has been suppressed. The best
explanation may be the major disconnect that exists between what most citizens of this
region want (reduced traffic congestion at the lowest possible cost) versus what the
members of Sound Transit’s Board want (light rail regardless the cost and despite the fact
it won’t reduce congestion).
To paper-over the gap, Sound Transit has systematically and continually resorted to
disseminating biased, misleading, and even false information about the merits of Link
light rail in order to bolster public support and justify Federal funding. One result is
deterioration in the public’s trust in government to spend scarce tax dollars wisely. It’s
also evidence that the current transportation planning process in Puget Sound is broken.
The process is not providing the through and objective information officials and voters
need to make multi-billion dollar decisions.
This report finds that the very foundation of Sound Transit’s Draft Long-Range Plan is
invalid because it’s based on one corrupted and obsolete study done in 1993. It
recommends that Link be placed on hold until and unless a proper, honest alternatives
analysis demonstrates it’s superior to BRT and other alternatives. This report further
recommends that federal and local officials take steps to fix the process.
vii
Extended Summary
The main purpose of this report is to trigger a reconsideration of Sound Transit’s light rail
plans, and the FTA’s willingness to fund them by challenging the fundamental basis for
Sound Transit’s entire rail-centric strategy including their recently released Draft Long-
Range Plan. A secondary purpose is to trigger improvements to the planning process that
will: 1) help ensure taxpayers get the most “bang for the buck” from their investments in
transportation, and 2) eliminate deceptive and manipulative practices on the part of
agencies such as Sound Transit. Still a third objective is to give planning students and
citizens in other cities a case study example of ways in which a transit agency has abused
the planning process in order to promote a favored outcome.
There are three key reasons Sound Transits light rail strategy should be reconsidered:
1) Link light rail costs too much for what little it accomplishes. At the same time
there appear to be better alternatives such as Bus Rapid Transit (BRT).
2) Sound Transit has never proven in any logical or business-like manner that Link is
superior to these other alternatives.
3) Decisions to approve Link have been based on incomplete, misleading, biased,
and false information which Sound Transit disseminated in order to garner support for
its light rail plan.
These observations are not new. Sound Transit has heard, and ignored, them many times
before. What this report adds to the record is a detailed and carefully footnoted analysis
of how Sound Transit has abused the planning process to promote light rail, and, for the
first time, a dollar estimate of just what pursuit of Sound Transit’s light rail strategy
would probably cost this region in relation to a bus rapid transit (BRT) alternative.
In short, the reader of this report will be privy to information that has not been available
to date. It will show that the emperor (Sound Transit) has no clothes as regards its
rationale for proceeding with light rail.
The full story of Sound Transit’s abuse of the planning process is beyond the scope of
this report. This report focuses on the alternatives analysis and environmental impact
study (EIS) that supposedly justified the choice of rail technology over bus technology
for this region’s mass transit backbone. The alternatives analysis and closely related EIS
are equivalent to a “business case” in the transportation-planning arena. They provide
just about the only cost and performance information available to those trying to decide
whether or not to fund projects like Link.
Background and context—
The Puget Sound Region is becoming committed to a light rail-centric strategy. Sound
Transit recently began construction on Links 14-mile “Initial Segment”. Sound Transit’s
viii
Board is already committed to extending it to Northgate, although they lack the money.
In addition, Sound Transit has just published a Draft Long-Range plan which envisions
extending Link into a regional system over 125 miles long. Over two billion dollars are
already committed to the first 14-mile miles of light rail, and it would take many billions
more to construct the full system.
Unfortunately even at this late date there is no solid analysis showing that light rail makes
sense. The main failing is that there has never been a proper apples-to-apples alternatives
analysis that compares the merits of light rail against an all-bus alternative based on bus
rapid transit or BRT technology.
Therefore, even as the region increasing commits to Sound Transit’s multi-billion dollar
rail strategy, neither public officials nor taxpayers know whether spending X billions on
light rail would yield more benefit than spending the same amount on buses, or on other
alternatives such as car and van pools, demand management, and so forth. The relevant
data needed to make intelligent decisions simply does not exist.
The region is being led unwittingly into a light rail plan that will influence the quality of
life in this region for decades and be the largest public works project in local history,
without having done the same due-diligence homework that MBA schools teach
businesses to use on far smaller investments. The public’s trust in government to spend
transportation dollars wisely is among the casualties.
This issue is highly relevant at present because Sound Transit is beginning to plan for a
Phase 2 that would seek more federal and local money to expand Link’s Initial Segment,
and because it is still not too late to stop light rail and switch to a to a different strategy if
new information –such as in this report-- shows that would make more sense.
The initial stimulus for this report was the fact that Sound Transit’s current light rail plan
didn’t reduce congestion and didn’t seem cost-effectiveness as an alternative to driving.
Sound Transit’s stubborn resistance to all criticism, its misrepresentations, and its failure
to resubmit its much altered plan to a public vote of confidence –making all of us feel
manipulated--have also been motivators. Perhaps most fundamental, was the knowledge
that Sound Transit had abused the planning process by never having done a proper
alternatives analysis. This greatly offended this planner’s sense of what’s right.
Link didn’t just happen by accident; it emerged as the end result of a planning process. A
key part of that process is the alternatives analysis.
Unfortunately, the alternatives analysis had been rigged to justify rail. In short, the
process had been abused. Unfortunately that’s not obvious to the casual reader of Sound
Transit’s 1993 FEIS and alternatives analysis report. It’s an impressive and seeming
well-written document. Most would assume it was competent and objective. Reluctantly
the author came to conclude that the 1993 FEIS was essentially a sham. Something that
appeared objective, but wasn’t. Something intended more to sell, than to inform.
ix
Part 1: Sound Transit has never done a proper alternatives analysis
In the transportation arena the alternatives analysis is critical. It’s equivalent to a
business case and is supposed to identify and evaluate the most promising alternatives
available to solve a given transportation problem. It’s about all that elected officials and
voters have to rely on, if they wish to make rational decisions.
Sound Transit maintains that the region’s one and only rail/bus alternatives analysis
provides adequate rationale for selecting rail rather than express bus for the backbone of
the regions transit system. However, for a wide variety of reasons this study –
documented in a 1993 Final Environmental Impact Statement (FEIS) prepared by Sound
Transit’s predecessor the RTA-- is inadequate and misleading. Nothing more up to date
or more relevant has been done since, as Sound Transit recently made clear in its DSEIS
for the Regional Transit Long-Range Plan.
The 1993 FEIS compared rail and bus technologies for the region’s main transit
corridors. Four alternatives were studied. One was the baseline or do nothing scenario
called “No Build”. Next was an extensive system of express buses operating on HOV
lanes. This $4.7 billion scenario was called “TSM”. A second all-bus alternative – called
“Transitway/TSM” and costing $5.5 billion-- was similar to TSM but used exclusive
busways in lieu of HOV lanes. Finally, there was a hybrid rail/bus alternative called
“Rail/TSM”. This $11.5 billion scenario called for 125-miles of rapid rail on a 100%
grade separated right-of-way. It also included many of the TSM improvements, except
those that would compete with rail.
There are two broad reasons why decisions about Link should not be based on this 1993
FEIS. First, the 1993 FEIS was neither adequate nor honest at the time it was completed.
Second, what we are planning today is not what the 1993 alternatives analysis studied. In
addition, circumstances have changed in the intervening 12 years.
A fundamental problem with the 1993 FEIS is that it compared alternatives that differed
in both cost and benefit. In theory this can be dealt with by careful focus on costeffectiveness,
but RTA botched the job. Essentially, RTA concluded that an $11.5 billion
rail alternative would perform better than a $4.7 bus alternative. This was simpleminded,
and is essentially like comparing proposals to build a brick wall costing $20,000
with a concrete wall costing $10,000; then deciding bricks are a better technology since
the $20,00 wall would be higher than the $10,000 wall.
The 1993 FEIS contained very little on cost effectiveness but what it did contain was
presented in a biased fashion. Essentially, it masked the high cost of using rail as a
means to increase transit ridership, and thus made the cost of the rail alternative appear
more competitive with the bus alternative than it really was. Also it failed to show that
the high marginal cost of rail riders was reasonable, or to put them in the context of costs
in other cities or of other alternatives. In short, it didn’t provide the kinds of information
needed to make wise decisions. This topic is expanded in the explanation of Part 6
below.
x
Another serious shortcoming of the 1993 FEIS is that it didn’t quantify or emphasize the
impact of the alternatives on traffic congestion. This is critical because polls have
repeatedly shown that the public’s main transportation concern is reducing traffic
congestion.
However, the main fault of the 1993 FEIS is that the bus alternatives were deliberately
designed to fail. In particular, the RTA unfairly alleged they lacked sufficient capacity
through downtown Seattle and then did nothing to rectify the problem. The RTA used
“inadequate capacity” as their primary reason for dismissing the bus alternatives, even
though they were more cost-effective. This topic is further pursued in the summary of
Part 5 below. The sum, the 1993 FEIS was neither adequate nor fair at the time it was
completed.
The 1993 FEIS is even less a valid alternatives analysis for Link light rail.
The 1993 FEIS was an alternatives analysis for rapid rail not light rail. This fact was
stated explicitly and repeatedly in the final report, and is not just a matter of semantics.
Certainly there were similarities between the rapid rail in the 1993 FEIS and Link today
in that the main rail corridors and station locations were similar and the train sizes were
identical. However, the train speeds, capacities, rights-of-way, and network sizes were
not the same. The maximum speed of the rapid rail in the 1993 FEIS was 70 mph,
whereas Link’s maximum speed is 55 mph. Rapid rail had an average speed of 36 mph
whereas Link’s would be 26 mph. The rapid rail system in the 1993 FEIS was asserted to
have a maximum capacity of 22,000 persons per hour (pph) whereas Link’s maximum is
16,400. The ridership forecasts in the 1993 FEIS were based on a system that was 100%
grade separated, whereas Link runs down the middle of the street in the Rainier valley
and may have additional at-grade street crossings on the Eastside. These differences
obviously affect ridership, reliability, and safety. In all these respects the Link system is
different from, and inferior to, the rapid rail system studied in the 1993 FEIS.
In addition, the 1993 FEIS evaluated only one particular 125-mile long rapid rail
network. This does not provide a comparison between the 14-mile system Sound Transit
is presently constructing, or the 21-mile system approved by voters in 1996, versus BRT
alternatives. Nor does it provide a comparison with any of the various networks that may
emerge from Phase 2 planning: like Central Link extended to Everett and Tacoma but not
to the eastside, or Central Link extended to the Eastside but not Everett and Tacoma, or to
all those but not to Issaquah and Totem Lake. The 1993 FEIS did not even compare the
ultimate light rail system envisioned by Sound Transit in their recent Draft Long-Range
Plan. That Plan includes a curious, never before contemplated loop following the
monorail from downtown Seattle to Ballard and then east to the University District. It
also contains light rail along I-405 in spite of prior decisions that BRT be used in that
corridor.
Since it’s not clear how much light rail it makes sense to build, how much is affordable,
or how much voters would ever approve, it is not adequate that the only rail/bus
alternatives analysis this region has to rely on studied just one particular 125-mile long
xi
configuration. In fact, rational decision-making requires that a range of possible light rail
networks –any of which might represent the optimum or final configuration—be
compared against BRT alternatives.
An important circumstance that has changed dramatically since the 1993 FEIS is the
assumed cost of constructing light rail. In the material handed to voters at the time of the
1996 ballot Sound Transit said: “Sound Move is based on extremely conservative cost
and ridership assumptions and methodologies reviewed by an independent expert review
panel appointed by the governor, the state Legislature and the state Transportation
Department.” However, this became front-page scandal in 2000 when Sound Transit
was forced to admit this was all-wrong, and had to increase the project budget by over
one billion dollars, or roughly 44%. For this reason any rail cost assumptions made prior
to 2000 were probably grossly underestimated. This fact alone makes the 1993 FEIS, as
it stands, obsolete and misleading.
Still another reason the 1993 FEIS is obsolete is that over half the HOV network --whose
costs were included in the TSM or bus alternative—has since been completed. Those are
sunk costs. Thus the all-bus alternative is now less expensive than it was in 1993.
The net effect of the underestimated rail costs and partial completion of the HOV
network is to make the cost difference between the rail and all-bus alternatives even
greater today that it appeared to be in 1993. To be even roughly relevant today, the 1993
FEIS cost estimates would need to be updated. They haven’t been. In fact Sound Transit
hasn’t even published an updated cost estimate for the 21-mile system approved by the
voters in 1996.
Sound Transit’s recent release of a Draft SEIS states quite clearly that its Long-Range
Plan is based on the 1993 FEIS. However, since the 1993 FEIS was not a valid exercise
in the first place, and is now obsolete, the entire foundation for Sound Transit’s Long-
Range Plan –not to mention the part already under construction-- is faulty. The FTA
should recognize this and force Sound Transit to conduct the proper alternatives analysis
that is long overdue and which is recommended so often throughout this report.
Part 2: Sound Transit morphed rapid rail into light rail
Sound Transit has not wanted to do an alternatives analysis for Link, so they have
maintained that the 1993 FEIS for rapid rail was really about light rail. In one recent
presentation the agency falsely states that the 1993 FEIS compared light rail and BRT. In
fact the 1993 FEIS never mentioned BRT, but it did address light rail (LRT) in a cursory
fashion along with monorail and other alternatives the RTA wasn’t interested in. This is
what the FEIS said about light rail.
"Surface LRT options were analyzed to the point that it became clear that these options did not
adequately serve the goals and objectives of the Regional Transit Project. Because of the superior
performance of the grade-separated RTP system in terms of consistency with land-use objectives, level
of service, and ridership, it was recommended as the rail technology in the recommended draft
Systems Plan." (Ref 1: page 2-61)
xii
In spite of their one and only alternatives analysis having recommended against it, Sound
Transit plans to put some “surface LRT” into the very backbone of the light rail system
they are building. How then is it possible for Sound Transit to claim that the 1993 FEIS
supports their decision to make light rail the technology of choice?
Part 3: The shortcomings of Link Light Rail
Light rail has many appeals to the superficial observer. It promises fast effortless trips
bypassing congestion. Many hope it will lure others off the road thus leaving more room
for them. Some believe it’s a way to control sprawl, clean the air and reduce energy
consumption. As visitors we’ve all benefited from riding rail systems in other cities.
Civic boosters think Seattle can’t be world class without rail. Unfortunately, at least in
the Puget Sound setting, light rail is one of those things where the less you know about it;
the better you probably like it.
Link’s fundamental problem is that it costs too much and does too little.
Polls consistently show that traffic congestion is one of the public’s highest concerns.
Link was sold to voters by implying it was a solution to traffic congestion. Yet Sound
Transit’s own studies prove Link would have almost no effect. The FEIS for the 21-mile
Central Link shows that the $2.6 billion (in 95$) system would only reduce road traffic
about 1/3 of one percent. This is equivalent to taking two and a half cars off a lane on
520 that is packed bumper to bumper with cars from Montlake to the east shore of Lake
Washington. It’s hardly noticeable, as the bar chart below illustrates.
0
5000
10000
15000
20000
Regional VMT in PM Peak in 2020
Without
Central
Link
With
Central
Link
Source: Central
Link FEIS,
Table 3.1-4
VMT in
000’s/day
16996 16933
xiii
The $1.5 billion (95$) Initial Segment now under construction would accomplish even
less. It would only reduce traffic 1/10th of one percent.
Even auto traffic into the Seattle CBD is minimally impacted.
It turns out the miniscule traffic reduction effects of Link’s 14-mile Initial Segment
would be wiped out within 29 days of its opening by the normal growth trend in regional
traffic as the following chart shows. The 21-mile Central Link has greater ridership, but
its traffic reduction effect would be wiped out in about three months.
Sound Transit has repeatedly implied that Link would reduce traffic congestion even
while knowing it wouldn’t. This was done with pictures, innuendo, and carefully chosen
wording so there was never any direct statement that could be challenged legally. It was
done extensively during Sound Transits pre-vote marketing campaign and continues
today. The statements that Link will not reduce traffic congestion are brief and don’t
stand out in the EIS’s, which very few voters read anyway. Sound Transit never
acknowledged publicly that light rail wouldn’t reduce congestion until, in December
2000, the author used Sound Transit’s own data to publicize that knowledge in a Seattle
Times OP ED. Unfortunately Sound Transit’s admission came years after the 1996 vote.
Light rail: There will never be a better time Richard Harkness got one thing right in a guest
column (The Times, Dec 22): Light rail will not ease traffic congestion. Yes, that’s a fact. (Dave
Earling, OpEd, Seattle Times, Dec 26, 2000)
The effects light rail could have on environmental issues like air pollution or energy use
are proportional to its impact on traffic, and thus similarly miniscule.
LRT effect on daily regional VMT
growth between 1998 and 2020
Reduction due
to Link IS = 81
Growth = 22,384
Values are daily
VMT in 000's
14 mile Link IS = 81
21 mile Central Link = 271
Link IS' savings
are wiped out in
29 days by
normal traffic
growth
9
xiv
As a way to get cars off the road Link is notably costly. If that’s the primary aim of the
project then it’s fair to divide the project’s cost by the number of cars it removes to get a
sanity check on Link’s cost-effectiveness. The result is that it would cost taxpayers
$100,000 per year for each car which Link IS removes from peak period traffic. This
cost would continue every year until the bonds were paid off in about 30 years.
The chart below shows how much this particular cost metric increased from the time
voters approved Link. The left bar is based on Central Link costs at the time of the vote.
The center bar is based on Central Link costs after Sound Transit admitted having
underestimated them and raised Link’s cost by $1 billion in Jan 2001. The right bar is
based on Link IS costs as reported in the Feb. 2002 Environmental Assessment.
In stark contrast, a recent Seattle Times article reported on a company that had
significantly increased car pool use by giving employees just $75 a month per person to
carpool. And, for $100,000 per year, it may be cheaper to simply pay people to quit their
day jobs and stay home.
Once the congestion reduction myth was publicly debunked, rail advocates –such as King
County Executive and Sound Transit Board Member Ron Sims-- switched to claiming
that light rail offers an alternative to driving. They said it provides “choice”. Indeed it
does offer choice to a favored few, but again that gift costs society roughly $80,000 per
year for each individual who –according to Sound Transit’s ridership estimates for 2020--
would find Link IS attractive enough to stop using a car. In this case the cost of choice is
huge. To paraphrase Winston Churchill: Never in the region’s history will so many,
have paid so much, to benefit so few.
Annual cost per vehicle removed
from peak period traffic
$ 30,000
$ 43,000
Operating cost (02$) $ 100,000
Capital cost (02$) + 233%
from vote
Vote in 1996 Jan 01 EA in 2002
xv
Link light rail has other deficiencies beside cost-ineffectiveness. It is highly inequitable
in that it provides service to a relatively narrow corridor while the cost is borne across a
wide region. Capacity on the south line –due to the decision to run on the surface along
Rainier Avenue—is only one third the capacity of the north line, and thus forever
shortchanges the entire south Puget Sound area.
The Initial Segment serves only one of the region’s 21 designated urban growth centers.
Even a full 100+ mile light rail system –which is not guaranteed and which the region
may never be able to afford-- would bypass many important commercial and employment
centers such as South center, Renton, Bothell, Tukwila, West Seattle, south Lake Union,
Bell town, Seattle Center, Ballard, Magnolia, south Seattle below the stadiums, West
Seattle, key Boeing sites, the Sammamish plateau office park, the emerging biotech area
on Elliott Bay, and so forth.
Link’s central control system and extensive tunneling make it vulnerable to power
outages and terrorism.
If Link is not extended into a full regional system its preemption of the Downtown
Seattle Transit Tunnel would compromise the remaining express bus system and dim
prospects for a regional BRT system.
Some of the region’s most precious resources are its preexisting rights-of-way (ROW),
and those should be used efficiently. Having the center lanes on I-90 devoted solely to
the occasional light rail train is not a good use of ROW since it would reduce the total
people moving capacity of the bridge relative to having a mixture of BRT buses, car and
vanpools, emergency vehicles, etc. use those lanes. Adding light rail would actually
reduce capacity in this corridor.
Currently over ten times more daily trips are made by car pool than by mass transit in this
region. (283,000 by transit, versus 3,554,000 by car/van pool) Arguably it is far more
important to maintain and expand car/van pooling than to expand an already excellent
mass transit system. Displaced them into less protected lanes than they have today is not
progress.
In short, upon close inspection of the facts, Link light rail doesn’t seem to be a good idea,
or to be worth the money. This would be true even if there were no obvious alternatives.
In economic terms, one must examine the “opportunity costs” of going with light rail
when there appear to be less costly ways to achieve much the same benefit. What else
could the region do with the money that could be saved? Alternately, if taxpayers are
willing to spend the same amount, how much further might we get toward reducing
congestion and improving the environment if some more cost effective technology were
employed? What if we could have 200 route miles of BRT for what – according to the
GAO study-- 14 miles of light rail is costing? What if some of the light rail money could
be diverted to accelerate the Alaska Way viaduct and 520 bridge replacements?
xvi
Part 4— BRT and Other Alternatives to Light Rail
If the objective is to reduce congestion or travel delay, the author has identified about 50
alternatives to light rail. They include things like widening roads, increasing car and
vanpool usage, telecommuting, and clearing accidents and breakdowns faster. However,
if only various forms of mass transit are of interest, then bus rapid transit or BRT is the
most likely contender to light rail. If light rail is “Plan A”, then BRT is “Plan B”.
BRT is actually a systems solution comprised of several elements. The core would be
express buses operating on HOV lanes, essentially what we have today, but more and
better. Dedicated busways are possible but probably not necessary. Service frequencies
would match that of rail. A BRT alternative would also involve bus priority lanes on
certain arterials, bus priority signalization, off-bus fare collection, direct access ramps,
and a range of other things all intended to make bus travel more rapid and attractive.
BRT hasn’t gotten much publicity here in Puget Sound. However, BRT has been
implemented elsewhere, particularly overseas, with great success. The U.S. General
Accounting Office published a report comparing it to light rail and encouraging cities
planning light rail to give BRT serious consideration as a less expensive alternative. It’s
proven technology and the advent of hybrid buses makes it even more attractive.
There is no question that BRT has enough capacity to handle the regions mass transit
needs. As shown in the chart below, BRT routes operating in other countries already
carry far more people than Link could carry, or that Sound Transit estimates Link would
need to carry.
Figure 5.12a
Pittsburgh 3700 *
Brisbane 9000 *
Ottawa 10000 *
Curitiba 11000 *
Bogata 25000 *
NYC 44000 **
Ultimate 16400 ***
Supplied 5480
Used 5400 ****
Link Light Rail BRT
Plan for 2020:
Persons/hr one way
at peak load point
North line
*Actual current use
** Based on current bus
volumes, all seated
*** Includes standees
Capacity: BRT vs Link north Ultimate 50000+
Harkness 3/03
**** Assumes LRT goes to Northgate
40,000
xvii
In terms of ability to meet future needs it is worthwhile to note that there is no way to
increase the capacity of Link light rail since train lengths would be limited by stations
already in place and headways can’t be reduced. For these reasons even a second rail
tunnel through downtown in some distant year wouldn’t help. In other words the limits
on rail system capacity are systemic and not subject to local remedies. In contrast bus
capacity can be increased by building short parallel paths around local bottlenecks, and
then only when needed.
Based on ST’s 1993 FEIS the highest volume that a BRT system would need to carry by
2020 is 12,000 persons per hour between downtown Seattle and the U District. This
volume could be handled by about 110 articulated buses per hour. This volume of buses
would use about 10% of the capacity of a single express lane along that stretch.
Elsewhere BRT would need much less than 10% of the capacity of an HOV lane.
BRT should also be faster. BRT buses can operate in express non-stop mode once loaded
whereas light rail must stop at every station.
BRT would probably require fewer transfers. Buses can circulate in neighborhoods
picking up passenger before entering the HOV lanes for non-stop travel to major
destinations.
Still, the most compelling reason for BRT is that much of the “guideway” needed for
BRT already exists in the form of the region’s 200 miles of HOV lanes. And every
investment made in expanding or improving these guideways for BRT has the double
benefit of encouraging more car and vanpooling.
Car and Van pooling—This report focuses on BRT versus light rail. However, car/van
pooling is probably even more cost effective than BRT, and should be among the
alternatives considered most carefully as the region charts its transportation strategy.
There are currently about 250,000 people car or van pooling to work. Car and vanpools
are already far more effective in getting people out of single occupancy vehicles than is
mass transit, as the following chart based on PSRC data makes clear.
0%
20%
40%
60%
80%
100%
All
Trips
Work
Trips
3
35
62
Transit
Car/van pool
Single
occupancy
vehicle
7
21
72
Mode Share for Trips in Puget Sound Region in 1998
Source: Ref 17:
p.xvii)
xviii
If the number of people car or van pooling to work could be increased by just 5% it
would take the same number of cars off the road as would building the 21-mile Central
Link light rail system.
At last estimate Central Link would cost $2.6 billion in 02$. Simple calculations show it
would cost $43,000 per year to take a car off the road using Central Link. Many people
who don’t car pool today could probably be induced to do so for considerably less than
$43,000 per year. If so, car and van pooling would be a far more cost-effective way to
relieve traffic congestion than building Link light rail.
Part 5—The capacity issue
Part 5 investigates the capacity issue in depth because back in 1993 the RTP alleged that
buses had insufficient capacity to meet the regions needs and dismissed bus alternatives
largely for that reason. Ever since, buses or BRT have remained off the table as far as
the core portion of the regional transit network is concerned.
RTP’s treatment of the capacity issue in the 1993 FEIS appears to have been deliberately
manipulated to favor rail. Basically, RTP postulated an all-bus alternative called TSM
and compared it against a hybrid rail/bus alternative called Rail/TSM. During their
evaluation RTP estimated year 2020 ridership for both alternatives and concluded that the
Downtown Seattle Transit Tunnel (DSTT) could not handle the number of riders the bus
alternative would attract. RTP’s response was to reduce the predicted ridership of
518,000 daily riders to what they said the tunnel could handle, namely 474,000. This
significantly degraded the cost-effectiveness of the bus alternative as well as reduced
other ridership dependent benefits, such as impacts on air pollution.
What the RTP did not do –after recognizing that a capacity problem might exist—was to
verify the assumptions they had made about tunnel capacity and/or seek ways to modify
the bus alternative so as to eliminate the alleged capacity problem. In short, RTP put a
knowingly and deliberately hobbled bus alternative into competition with the rail
alternative.
The key assumption that RTP did not verify concerned tunnel capacity. There had been
six prior paper studies that each reached different conclusions, ranging from 125 to 192
buses per hour in each direction. The RTP chose to assume 100, a value based on
operating buses inefficiently. About 135 buses per hour were needed to carry the
predicted demand. Why --when capacity was so critical-- didn’t RTP assume the tunnel
would be managed efficiently? Why --with the whole multi-billion dollar rail vs. bus
decision hanging in the balance-- didn’t RTP take the trouble to verify the tunnels true
capacity with real world trials? This could not have been oversight or incompetence, it
must have been deliberate.
If the RTP had properly verified tunnel capacity and still found it below 135 buses per
hour they could, and should, have found other remedies so capacity problems in this 1.5-
mile segment of a 125-mile network didn’t become the tail that wagged the dog.
xix
In prior studies, a range of fixes had been identified. They ranged from increasing bus
capacity on downtown streets to building a second parallel bus tunnel, which RTP staff
had estimated would cost $600 million. As a worst case, RTP could have added this
second tunnel to the bus alternative thus completely eliminating the downtown bus
capacity problem. Failure to do so was apparent bias or manipulation. It seems
especially egregious since the RTP elected to provide 20 miles of tunnel for the $11.5
billion rail alternative yet was unwilling to provide even a mile or two for the $4.7 billion
bus alternative.
In short, had RTA resolved the alleged bus capacity constraint –either by finding it didn’t
exist, or fixing it-- the RTA could not have claimed the bus alternative was unable to
handle its predicted ridership, and –able to carry its full ridership-- the TSM alternative
would have been much more competitive in its comparison against rapid rail.
Part 5 also addresses ridership forecasts that appear to conflict. In 1993 RTP estimated
that by 2020 the peak load on the rail system would be 15,000 persons per hour at the
peak load point just north of the DSTT. Sound Transit still asserts the long term demand
for rail transit would create a peak load point demand of 15,000 persons per hour on the
north line, and that we need a system able to handle it. However, the much more recent
forecast for Central Link predicts a peak load of only 5415 persons per hour. This
discrepancy is something the large difference between the 125 and 21-mile systems does
not appear to explain.
This is an important issue in that if 15,000 is the correct number it appears that Link
would run out of capacity soon after 2020 and is therefore not a long range solution for
the region’s capacity needs. Indeed one of the reasons RTA gave for choosing rapid rail
in 1993 was that its assumed capacity of 22,000 gave it headroom for growth well beyond
2020. In addition, if 15,000 is correct, Link may not be enough capacity on its south line
to even meet demand in 2020.
On the other hand if 5415 is correct, it is possible that system ridership is simply lower
than originally thought.
It seems that no matter which forecast is correct, Sound Transit faces an embarrassing
situation. If the higher forecast is correct, light rail is inadequate. If the lower forecast is
correct, it completely destroys Sound Transit’s claim that buses lack sufficient capacity
through downtown, even if the tunnel could only handle 100 buses per hour.
Sound Transit’s claim that light rail “would provide the same people moving capacity as
a 12 lane highway” is simply false. Just one freeway lane pair full of buses could carry
far more people than Link. Rather than compare what these systems could carry, it is
more meaningful to compare what highways actually carry on a daily basis versus what
light rail is actually expected to carry. Sound Transit’s ridership forecasts for Central
Link –at the ship canal where it is heavily loaded-- show it would carry only slightly
more people in 2020 than a single lane pair on I-5 carries today. Along the bulk of its
xx
route Central Link would carry only a fraction of what one lane pair on I-5 is now
carrying.
Part 6: The cost issue
Part 6 has two broad objectives. One is to demonstrate how the 1993 FEIS was
inadequate and biased in the way it addressed cost-effectiveness. The second objective is
to use the raw data in the 1993 FEIS and related documents to produce an estimate of
what Sound Transit’s current rail-centric strategy might end up costing this region in
relation to the all-bus alternative they rejected.
The 1993 FEIS gave very little attention to either cost or cost-effectiveness. Almost
everything that large report had to say was contained in a single small table listing cost
and ridership totals for each of the four alternatives. (Table 4 in the 1993 FEIS)
This form of presentation was deceptive. The difference between the rail and bus
alternatives, shows in the numbers, but it doesn’t appear dramatic. This author, probably
like many others, saw that table and felt that while the bus alternative was more cost
effective, it was not dramatically so. This was because the cost per rider figures RTP
chose to present were averaged across all riders and thus did not allow the costeffectiveness
of the rapid rail element to be separated out and examined for
reasonableness. It could be claimed this information was deliberately hidden.
To better illustrate what could have been done, and what should have been done, the data
in that FEIS Table 4 has been reformatted into Figure 6.1 below. Remember, this is the
picture as it would have appeared in 1993, with all costs in 91$ and before the rail cost
estimates were found to have been underestimated.
This chart conveys a very different message. Not only would it cost much more than we
are paying today to increase transit ridership by implementing TSM, but it would cost
dramatically more still for those additional riders that rapid rail could add above and
beyond what TSM could attract on its own.
xxi
As to the detail, transit would attract 389,000 daily riders by 2020 if we did nothing but
grow the existing bus system. RTP called this the “No Build” alternative. The chart
shows that society is now paying about $3.30 for each one-way bus ride, less the 80 or so
cents recovered at the fare box. (The transit industry calls a one-way ride a “rider”.)
Next, it shows that implementing just the TSM alternative would increase daily transit
ridership by about 80,000, and that these extra 80,000 rides would cost society about
$15.50 each. Finally, it shows that if we build the 125-mile rapid rail system it would
attract an additional 85,000 riders above and beyond what the all-bus TSM alternative
could achieve. However, these extra rides would cost $25 each.
If the RTA had elected to fix the alleged capacity bottleneck by building a second bus
tunnel this cost picture would have changed to that shown in Figure 6.2c.
0
5
10
15
20
25
30
0 200 400 600
R iders 000's/da y
Cost in 91$/rider
$15.50 each for those
added by TSM
$3.30 each with No
Build
$24.70 each for those
added by Rail
389 474 560
Figure 6.1
xxii
Figure 6.2c emphasizes the point that society would need to spend far more to attract
these new riders than it has been willing to spend in the past. The upsweep of the bars
illustrates very dramatically the law of diminishing returns. Once those with relatively
little choice have gotten aboard, it becomes increasing expensive to improve transit
service enough to attract others. This raises the questions of affordability and
reasonableness. There is some point where the cost of making transit more attractive so it
will attract more riders begins to exceed the benefits. To make rail worthwhile the
benefits of each rider it adds would need to exceed $47.50 per one-way ride. Again, that
0
5
10
15
20
25
30
35
40
45
50
0 200 400 600
Riders 000's/day
Cost in 91$/rider
$47.50 for
Rail/TSM
$11.50 for TSM
with 2nd bus
tunnel
$ 3.30 for
No Build
390 560
Figure 6.2c
518
xxiii
is what would have been visible in 1993. After accounting for ST’s underestimated rail
costs, and inflating to current dollars, that $47.50 becomes $93. The 1993 FEIS did not
get into any of this, perhaps because they felt it would have worked against their desire to
promote rail.
Figure 6.2a is another chart –based on Table 4 in the FEIS--that would have been helpful.
The cost ($960 million/yr) and ridership (172,000/day) of the Rail/TSM alternative is
represented by the dot at the end of the Rail/TSM line. The same is true of the TSM line.
This chart shows that while the Rail/TSM alternative costs more and does more, the TSM
alternative is actually more cost effective since its “trend line” is not as steep.
0
200
400
600
800
1000
1200
0 50 100 150 200
Riders above NB 000's/day
Cost above NB millions of 91$/yr
Rail/TSM
TSM (constrained)
Why not just do
more TSM ?
85 172
Figure 6.2a
xxiv
NOTE: If the RTA’s goal had been to increase ridership by just 85,000 above the
NoBuild baseline, rather than 172,000, this chart shows that the original TSM
alternative would have had adequate capacity, as well as being less expensive.
Alternately if the goal had been to reach 300,000 riders the rapid rail alternative
itself would have failed for lack of capacity. Thus, where the goal is set can
sometimes determine which alternative wins. Planners can manipulate this to get
the answer they want. In the case of the 1993 FEIS the goal was set high enough to
(allegedly) break the bus alternative but not high enough to break the rail
alternative. This is one reason Part 8 calls for a range of different size rail networks
to be compared with bus alternatives.
Presenting the data with a chart like Figure 6.2a is useful because it graphically suggests
the following: Why not just intensify or extend the TSM alternative until it achieves the
same ridership as rail, because it looks as though the savings would be worthwhile? In
other words, whatever we were doing in the TSM alternative, just do more of the same.
In practice this would have meant more frequent bus service, more routes, and perhaps
more HOV lanes, direct access ramps, and park & ride lots. The arrows suggest this idea.
RTP staff was aware of this option and had already estimated its cost.
However, the TSM alternative can’t be extended until the alleged bus capacity constraint
in downtown Seattle is dealt with. Maybe the RTAs assumption about tunnel capacity
was wrong and there really isn’t any bottleneck. However, to be conservative the author
assumed that the bottleneck was real, and that it takes the most costly of the available
remedies –namely a second bus tunnel—to fix it. This was an option that RTA staffers
had already identified and estimated would cost $600 million. The RTA had also
estimated the cost of extending TSM services so as to achieve greater ridership. It short
the RTA had shown how the TSM alternative could be extended to achieve more
ridership, and they had estimated the costs of doing so. However, all this information
was hidden in backup technical reports and none of it was used to fix the bus alternative’s
alleged capacity problem. Instead the 1993 FEIS presented a hobbled bus alternative
unable to match rails ridership or achieve its other benefits.
The chart below shows what would happened if the RTP had un-hobbled the bus
alternative. The first part of the TSM trend line shows the result of including $600
million for a second tunnel, which allows TSM to achieve the “unconstrained” ridership
forecast by the RTP. The second part of the line is based on RTP’s cost estimate for
extending it so as to attract and handle as many riders as the Rail/TSM alternative.
xxv
At this point it is apparent that fixing the alleged TSM capacity constraint and extending
TSM is a good idea. It would save taxpayers about $400 million (in $91$) per year. The
RTP could have done this analysis and included it in the 1993 FEIS. Was their failure to
do so a matter of incompetence, or of deliberate bias? Again, this is the picture that RTP
should have presented in 1993. But what does it mean for Link light rail today?
The above analysis can be updated and used to show the implications of proceeding with
Sound Transit’s light rail-centric strategy, which calls for building as much as 125 miles
of light rail along with supporting TSM elements such as bus feeders, HOV
improvements, and park and ride lots. In other words the RTP created and evaluated a
0
200
400
600
800
1000
1200
0 50 100 150 200
Riders above NB 000's/day
Cost above NB millions of 91$/yr
TSM (unconstrained)
Rail/TSM
TSM extension
$ 404 million/yr
difference
Figure 6.2d
xxvi
“Rapid Rail/TSM” alternative with 125 miles of rapid rail. What the author does in this
report is create a “Light Rail/TSM” alternative with 125 miles of light rail, and then
compare it against an all-bus or TSM alternative using BRT. Figure 6.4 shows the result.
It now appears that Sound Transit never intends to make this comparison, but it’s
essential information because the region’s at a fork in the road. It can either proceed to
implement Sound Transit’s light rail-centric strategy, working out along the Rail/TSM
line in Figure 6.4 to build as much light rail as possible. Or it can switch to an equally
effective all-bus or BRT strategy while there’s still time. The public needs to understand
the dramatic difference in cost.
0
200
400
600
800
1000
1200
1400
1600
1800
0 1 2 3 4
Reduction in VMT % under NB
Cost millions of 02$/yr over NB
Rail/TSM
TSM, extended
Initial Segment
(14-miles)
Central Link
(21-miles)
Figure 6.4
$ 916 million/year
xxvii
Thus Figure 6.4 compares the cost of Sound Transit’s light rail-centric strategy with the
cost of an all-bus or BRT alternative able to reduce regional Vehicle Miles of Travel or
VMT by the same amount. (Reduction in VMT is a better metric for comparing benefit
or effectiveness of the alternatives than is transit ridership. Travel delay would have been
a better metric yet, but the RTP did not provide that data.)
Figure 6.4 is based on conservative assumptions. For instance the author assumed that a
125-mile light rail system would attract the same ridership as the 125-mile rapid rail
system studied by RTP. Actually, light rail would probably attract fewer riders since it’s
slower. On the cost side, Figure 6.4 is based on light rail costing an average of $120
million/mile whereas Link IS is actually costing $138 million/mile and Central Link is
expected to cost $158 million/mile. (All these are in 02$) Also, it was assumed that the
all-bus alternative would require a second bus tunnel costing $600 million (91$) through
downtown Seattle, although the existing bus tunnel may well suffice, and if not there are
probably less expensive remedies than a second tunnel. In other words, the Tail/TSM
line in Figure 6.4 is probably steeper than shown, the TSM line is probably flatter, and
the difference between them is probably greater than $900 million/yr.
As to details, the RTP originally estimated the 125-mile rapid rail system would cost $7.9
billion in 91$. The construction bids Sound Transit received a few years ago showed that
actual costs were 44% higher than originally estimated. Figure 6.4 assumes the 125-mile
light rail system would cost 1.44 times that $7.9 billion, or $11.4 billion in 91$. In 02$
the 125-mile light rail system would cost $15 billion. (This is probably the best estimate
publicly available today for what the light rail system in ST’s Long-Range Plan might
cost) The original TSM costs were updated in two ways. The cost component allocated
for building HOV lanes was cut in half since half the HOV network has now been
completed. Second, $600 million was added for the second bus tunnel. All capital costs
were annualized by assuming a 30-year, 6% bond. Annual O&M costs from the RTP
table were included without modification. Finally, cost totals were inflated from 91$ to
02$ to make them more timely.
The dots at the end of the Rail/TSM and TSM lines show the annual cost of these two
alternatives. For instance it would cost about $1.5 billion per year to fully implement
Sound Transits light rail-centric strategy by building 125 miles of light rail. The lines
from the origin to the dots give some rough indication of the costs of smaller systems.
For instance, a system with 63 miles of light rail rather than 125 might cost about $750
million per year. By the same token a BRT system able to achieve the same VMT
reduction might cost about $300 million per year. The actual cost and performance of
Link IS and Central Link are shown for reference.
Figure 6.4 is probably the most important chart in the entire report. What does it tell us?
It shows that to achieve the same level of transit ridership estimated for a 125-mile rail
system, switching from light rail to an all bus strategy would save taxpayers about $900
million per year. This would continue over the 30-years life of the bonds.
xxviii
The huge cost difference between Link light rail and BRT simply reflects the fact that the
best technology for one region is not necessary the best for another. Light rail doesn’t
have an easy fit in this region because we lack the abandoned railroad rights-of-way or
flat terrain that makes constructing light rail relatively easy in other cities. On the other
hand our excellent bus system and extensive HOV network makes BRT particularly
attractive.
The diagram below attempts to clarify any remaining confusion as to how the RTA
concluded buses were unsuitable while the author reached the opposite conclusion.
The flowchart reads as follows. The RTP started by designing an all-bus alternative
called TSM and a 125-mile rapid rail plus feeder bus alternative called Rail/TSM.
Consultants then estimated the cost and ridership for each alternative. At that point it
became apparent that predicted bus ridership would exceed RTP’s assumption for tunnel
capacity. RTP elected not to fix the capacity problem using any of the remedies that
staff had previously identified. This resulted in a crippled bus alternative being compared
against the rail alternative, and thence to RTP’s conclusion that buses couldn’t meet the
regions needs, whereas rapid rail could.
In this report the author starts with RTP’s bus and rail system designs, and with RTP’s
cost and ridership forecasts, but reaches a different conclusion. He did that by adding a
second bus tunnel to “fix” the alleged bus capacity bottleneck downtown, then adding
more bus service so it would attract the same ridership as the rail alternative. The cost of
these modifications was added in. He then concluded that the modified bus alternative
could equal rail in ridership, have sufficient capacity for long-term growth, and be much
less expensive.
Cost and
ridership
estimates
Fix the bus
capacity
problem?
Crippled bus
alternative
Competitive
bus alternative
FEIS’ conclusion that
bus alternative was
Bus alternative unsuitable
(TSM)
Rail alternative
(Rail/TSM)
NO
YES
Authors’ conclusion
that bus alternative is
probably superior
xxix
Part 7: Evaluation and recommendations
Part 7 tries to explain why things have gone astray, and some of the broader implications
that Sound Transit’s preoccupation with light rail is having on public trust and on
competing uses for scarce tax dollars. This report presents new information that will
hopefully trigger action. Part 7 recommends several specific actions that seem
appropriate based on this new information.
Why things went wrong- There is a huge disconnect between the problem (traffic
congestion) and the proposed solution (light rail).
This disconnect shows that what drives officials and decisions in this arena was not an
honest attempt to find congestion remedies based on solid analysis. Instead, what has
happened is best explained by some complex mixture of myth, fact, wishful thinking,
uninformed opinion, altruistic and not so altruistic motives, hard-ball politics, ego,
psychology, bureaucratic maneuvering, and most of all, money. When all this enters the
mix it is not surprising that actually spending taxpayer money wisely so as to make the
most progress against traffic congestion fell by the wayside.
As to motives, one must acknowledge the superficial appeal of rail transit. Emotionally
rail seems like a simple “silver bullet” solution for a complex and intractable problem.
Promoting it appears to be “doing something”.
But elected officials have another set of reasons to favor rail. They revolve around power
and money. If it proceeds, Link light rail will be among the largest public works projects
in Puget Sound history. Billions of dollars will be spent. Officials can feel important
making decisions about how billions are spent. They can proudly leave a legacy of
concrete and steel. They can please certain powerful parties on the receiving end of those
billions. They can be seen as providing jobs and stimulating the economy. Perhaps
most of all they can be seen as bringing in “free” money from Washington D.C.
Fundamentally Link is a “pork barrel” project.
In terms of who pays and who benefits, Link is a clever way to transfer money from the
pockets of many to the pockets of a few. Not enough is taken from the pockets of the
average taxpayer (in the form of sales tax and auto registration tax) to cause him or her to
‘fight city hall’. On the other hand the relative few directly benefiting from Links design,
financing, and construction maintain a behind-the-scenes pressure to keep Link going.
Another reason Sound Transit’s a clever mechanism for wealth transfer is that taxpayers
have practically no way to stop it. The State Legislature created Sound Transit, but
forgot to make Sound Transit’s Board members directly elected, and they forgot to give
voters in the Sound Transit taxing district any practical control over Sound Transit via
initiative. As a result the only way local voters can control Sound Transit is by mounting
a costly statewide initiative. This self-confidence in its own untouchability gave Sound
xxx
Transit’s attorney the hubris to say in open court that Sound Transit recognized no limits
on how much it could spend or how long it could take building light rail.
Key officials have long wanted a rail transit system in Puget Sound, period. In the early
1990’s consultants were hired to prepare the necessary paperwork in order to get federal
funding and meet state law. Although alternatives analyses are supposed to be objective
studies, consultants are generally fairly astute at figuring out what the client really wants,
and repeat business means delivering it. The client wanted rail. As a result the 1993
FEIS or alternatives analysis became a pro forma exercise meant to justify this
preordained conclusion.
In short, the 1993 FEIS is a sham; a document meant to satisfy legal requirements, but
almost totally useless in providing objective information that would help officials or the
public make a wise decisions about spending billions of dollars.
The FTA colludes with Sound Transit in these abuses of the planning process. FTA
overlooks faulty work such as the 1993 FEIS, and their requirements for funding projects
like Link appear lax. FTA is probably among those Federal agencies that have been
accused of being in bed with the organizations they are supposed to regulate.
It is also clear that Sound Transit’s board has little concern for spending tax dollars
efficiently. In other words achieving the “most bang for the taxpayer buck” is not a high
priority with that group. If the Sound Transit Board really wanted cost effective solutions
they would have scrutinized the 1993 FEIS in the way the author has done and probably
chosen BRT, since it appears BRT would save billions. After finding that Link’s costs
had been underestimated and needed to be increased by 44%, they would certainly have
revisited their choice of rail, rather than simply looked for additional money. They would
have published and agonized over, rather than hidden, the high cost-per-rider data for
Sounder and Link. They would have calculated the approximate cost of fully
implementing their current rail centric strategy and compared that against BRT. These
are the minimum things a board really concerned with spending tax dollars wisely would
have done.
The Puget Sound Regional Council (PSRC) must also take considerable blame. If that
organization were sincerely concerned with getting the most bang for the taxpayer buck it
would have eagerly embraced and practiced Least Cost Planning –as State Law requires--
rather than trying so hard to avoid it. It would have analyzed Sound Transit’s light rail
plans and ensured they were cost effective relative to other alternatives rather than simply
downloading them intact into the Metropolitan Transportation Plan (MTP). It would
consider van and car-pooling as a major stand-alone alternative to investing billions in
mass transit and highways, rather than just as window dressing attached to the main
transit and highway alternatives.
In fact, it is not exaggerating to state that if the PSRC had obeyed the law as regards
Least Cost Planning (LCP), Sound Transit’ light rail plan would probably never gotten
this far.
xxxi
The importance of trust-- It has become increasingly clear that voters don’t trust Sound
Transit and Sound Transit doesn’t trust voters. In fact, Sound Transit has become the
poster child for why voters distrust government to spend their transportation tax dollars
wisely. There are good reasons for this mistrust.
Voters overwhelmingly voted to rescind Sound Transit’s tax on vehicle registration, but
ST still fights to overturn that expression of public will in the courts. ST changes the
project greatly from what voters approved in 1996 but vigorously fights a lawsuit that
would have put the much altered plan back on the ballot for a vote of confidence.
Sound Transit wants more money. However, while claiming it has a public mandate and
support for light rail, ST is deathly afraid to ask voters to increase the existing light rail
tax. Thus ST supporters used political strong-arm tactics to force light rail into a broad
package of regional transportation projects where voters couldn’t reject it without
rejecting everything else as well.
On this November’s advisory ballot voters were asked if they supported that package,
which the ballot explicitly said was intended to reduce congestion and improve safety.
Since Sound Transit had already admitted Link wouldn’t reduce congestion, putting it
into that package was a remarkable example of cynical manipulation and deception.
The problem is magnified since that ballot –which still implies Link would help reduce
congestion-- was seen by millions, while few have seen the truth.
When one party has the funds to broadcast misleading statements through a megaphone
while their critics have only the occasional Op-Ed or letter to the editor to whisper a
rebuttal, there is simply no way the public will get a balanced story. Critics, such as
CETA, have found the facts don’t matter when there’s no money to get them
disseminated. This has been a fundamental structural problem throughout the entire light
rail debate here in Puget Sound.
The role that trust, or lack thereof, plays in reaching any solution for Puget Sound’s
transportation problems is hard to overstate. Part 7 contains a long list of newspaper
quotes like the following:
Sound Transit: a matter of trust The 10-year plan for increasing transportation system capacity in
the Central Puget Sound area was dubbed ”Sound Move” by its creators at the regional transportation
authority, Sound Transit. …
….Half of those ten years are now past, and the Sound Move plan has fallen well short of its billing.
The most visible culprit of course is the light rail project, which is $1 billion over budget, three years
behind schedule and the subject of a federal audit.
…It’s the sort of indecision that makes one wonder, despite the agency’s official denial, just how
much of the 10-year plan was completed in a vacuum, without input from the very people is meant to
serve. It’s the kind of day to day waffling and mismanagement that wastes time, overruns budgets
and over time, causes people to lose trust. Some of us in the Legislature have noticed the loss of trust
in Sound Transit.
xxxii
… Why should taxpayers support long term financial commitment to transportation when Sound
Transit provides such a convenient example of a commitment gone sour?
…The second offers revote on the grounds that the Sound Transit board’s actions have significantly
altered the proposition citizens approve in 1996.
…Because of Sound Transit’s lackluster performance so far, any taxes directed toward transit and
transportation projects from here on out need to be the best spent money in state government.
….At the heart of the public trust , President Abraham Lincoln once wrote, is trusting the public.
Government should not be – and cannot afford to be- afraid of letting the people judge how well their
money is being spent. Let’s Vote. (Op Ed by 15 members of the State Legislature, Seattle Times,
Feb. 20, 2001)
Requests that Sound Transit consider alternatives- There have been numerous
requests that Sound Transit reconsider its light rail plans. Again this story is told via
articles quoted from local newspapers. Two examples are:
Sound Transit Board: It’s time to do your job After the overwhelmingly negative wave of recent
events, you would think a board that calls itself “Sound” would stop shelling out our money to move
forward on such a monumental undertaking as a $4.2 billion light rail plan. You would think they
would call a timeout not just to patch up the holes that have been revealed, but a timeout to actually
rethink whether this flawed vessel, light rail, is going to get us to the vital goal of reducing traffic
congestion in central Puget Sound. Yet in response to the damaging report issued last week by the
US Inspector General’s office… all we get here at home is more patch-up.
… My response is this: Is there any event or combination of facts, any misgivings about cost,
funding, ridership, or concerns over the Inspector General’s criticisms, anything at all that would
finally cause the Sound Transit Board to ask, “Is proceeding with light rail still a good idea?” It
seems no issue exists that’s significant enough to prompt the board’s serious review of alternative
solutions.
…The board’s irresponsibility is found in the decision to remain silent despite their growing
awareness of that misinformation, including possession of significant evidence that light rail may not
be a cost-effective transit alternative.
Civic groups, critics and other elected officials are calling ever more loudly for a complete review of
the project including available alternatives… (Booth Gardner, former Governor Washington State,
Op Ed Seattle Times, April 11, 2001)
________________________
Full speed ahead for light rail In a hasty attempt to secure $500 million in federal funding before
the Clinton administration steps down, Sound Transit will move forward with light rail despite
mounting objections about costs, the agency’s leaders say.
…But Sound Transit Executive Director Bob White and board Chairman Dave Earling say they
already know what the board will decide on that pivotal day: The agency will not explore
alternatives to light rail… (Chris McGann, Seattle P-I, Jan. 5, 2001)
xxxiii
Recommendations-- This report recommends the following actions:
1) The FTA should reject the DSEIS for Sound Transit’s Long-Range Plan
and should withhold any additional money for Link until a proper alternatives
analysis has been completed, its results fully communicated to the public, and a
public vote of confidence confirms voter support for continuation of Sound
Transit’s light rail strategy. Part 8 of this report outlines key requirements for
conducting a proper alternatives analysis.
2) The Sound Transit Board of Directors should voluntarily undertake the
above actions with or without the FTA requiring same. They should do this to
confirm they are on the right path and to restore public trust.
3) Congress should investigate the manner in which mass transit grants are
approved to ensure that FTA controlled planning processes are not abused in
the ways chronicled throughout this report. They should investigate Sound
Transit as one case example. Congress should insure that the process is
redesigned to obtain the most “bang for the taxpayer buck”. That is: greatest
improvement in transportation at least cost to Federal and local taxpayers.
The Sound Transit Board has no good reason to balk at taking these actions. If Board
members are confident they are on the right path, a proper alternatives analysis can do
nothing but confirm it, silence the critics, and help restore public trust. There is no
reason why Link can’t be put on hold since its completion wouldn’t have much beneficial
effect.
Vetting-- Authorities should immediately cause this report to be reviewed by an
objective team of experts to confirm or refute the logic, calculations and conclusions
herein. This could be done in about six weeks. If those are upheld by the team of experts
then there would exist reasonably credible, but still not conclusive, evidence that a an allbus
strategy could achieve much the same benefits as light rail, and do so at a far lower
cost. Such vetting would justify temporally halting construction on Link until a fullfledged
alternatives analysis is complete in 12 to 18 months. Finally, Link could be
restarted if the full-fledged analysis is favorable, or terminated if it is not.
Does it matter if the region spends billions more on mass transit than it needs to?--
This report concludes that pursuit of Sound Transit’s light rail strategy could end up
costing the region roughly a billion dollars a year more than switching to an all-bus
alternative.
So what? Does anyone care?
The answers are not obvious. It will be interesting to see if anyone gets concerned
enough to act. Perhaps the best way to make the cost of Sound Transit’s rail plans
meaningful is to list some of the other good projects that are being shortchanged because
they are, in the ultimate analysis, competing with light rail for limited tax dollars.
xxxiv
School-renovation fund $11 million short A Seattle School District review shows that its schoolrenovation
programs are running deficits that could mean some projects will be delayed, trimmed or
eliminated. (Sanjay Bhatt, Seattle Times, Aug. 4, 2004)
_______________________
$878 million more sought by Bergeson for schools Terry Bergeson, state superintendent of public
instruction, yesterday asked for an additional $878 million for public schools over the next two years,
an amount she says is essential to reach the goals of the state’s decade old education reform law.
(Linda Shaw, Seattle Times, Sept. 23, 2004)
_______________________
Legislators Brace for Extra-Hungry Interest Groups …There will be pressure to expand colleges
and universities and pay for multibillion-dollar transportation projects such as replacing the Alaskan
Way viaduct.
Much of this year’s problem is pent up demand. Colleges, for instance, have not kept pace with
population growth and many state workers have gone years without a pay increase. But it comes at a
time when the state projects an budget deficit of around $1.8 billion, after already struggling through
several years of huge shortfalls. (Seattle Times, Jan. 10, 2005)
Sound Transit’s 2004 Financial Plan says that $2.437 billion will be spent on Link’s
Initial Segment between 1997 and 2009. In 2004 alone Sound Transit will collect $271
million in taxes.
The amount of money going to even the Initial Segment of Link light rail would make a
big dent in the funding needed to reconstruct the Alaska Way viaduct and rebuild the 520
bridge. But the Initial Segment is just the first step in Sound Transit’s ambitions for light
rail. Clearly the $900 million per year difference between a 125-mile version of Link
versus an all-bus alternative would be more than enough to pay for both projects.
In short, the money that might be wasted on Sound Transits rail-centric strategy is not an
abstraction; it comes at the expense of opportunities foregone, and of other good ways to
spend tax dollars.
Nor is the impact on the Federal budget, of questionable projects all across the country
like Link, something to ignore.
Congress lift debt ceiling New borrowing to avert default
Congress last night sent President Bush an $800 billion boost in the federal borrowing limit,
spotlighting how the budget has lurched out of control in recent years and how difficult it will be to
afford future initiatives. … “I want someone to explain to me how it can be moral for a father to stick
his kids with his bills,” said Rep. Gene Taylor, D-Miss. (Seattle Times, Nov. 19, 2004)
xxxv
Part 8: Guidelines for a proper alternatives analysis
Sound Transit is beginning to plan for Phase 2 of Link light rail. The DSEIS Long-
Range Plan shows this will include a limited version of an alternatives analysis, but the
only alternatives being looked at are the alternative ways to expand HCT above and
beyond the 24-mile Central Link line which Sound Transit is taken as a given.
Limiting Phase 2 to just consideration of ways to expand Central Link does not provide
voters and officials a full view of the alternatives before this region. Since Link
construction has hardly begun, one option is to halt its construction and shift to a more
cost effective alternative like BRT. Another is to build Link IS but truncate it south of
the bus tunnel so it wouldn’t interfere with a regional BRT system. Still another option
concerns the full 100+ mile version of Link called for in Sound Transit’s draft Long-
Range Plan, and in PSRC’s Metropolitan Transportation Plan. This 100+ mile version of
Link needs to be compared apples-to-apples with a BRT alternative that is either equal in
cost, or equal in benefit.
Part 8 describes these and other options in more detail along with technical guidelines to
ensure the rail/bus comparisons are done objectively.
Early indicators for Phase 2-- Part 8 concludes with a litmus test that will give voters
an early indication of whether or not Sound Transit plans to conduct an adequate and
objective alternatives analysis as part of their Phase 2 planning.
Part 9: Main conclusions of this report:
1) The existing planning process is not producing the kind of information needed by
officials and the public to make intelligent decisions about major mass transit
projects. Important information is missing or obscured. Promising alternatives are
ignored. Reports seem intended more to sell than to inform.
2) Sound Transit and its predecessor agency the RTA have abused the planning process
in order to promote light rail. They biased key studies by making inappropriate
assumptions and masking key information. They compared a robust rail alternative
against a deliberately hobbled bus alternative. They disseminated misleading
information to the public.
3) ST justifies its choice of light rail on the one and only rail vs. bus alternatives analysis
conducted here since the 1980s. However, that study was deliberately biased to favor
rail. When that bias is removed the underlying data shows that an all-bus solution
could probably achieve the same level of benefit at far lower cost.
4) ST and RTP dismissed bus alternatives largely on false claims that buses lacked
adequate capacity. Their analysis was deliberately manipulated to support these
claims. BRT has more than adequate capacity to meet the region’s long-term needs.
Light rail has less capacity than BRT and is therefore less strategic.
xxxvi
5) As construction begins on Link there is still no study which compares the benefits of
spending $X billions on light rail plan versus spending the same amount on bus rapid
transit.
6) If the money now intended for light rail were instead redirected toward other projects
such as BRT, car and vanpool enhancement, and other transportation projects the
region could probably make considerably more progress in solving our transportation
problems, because these other alternatives are more cost-effective.
7) By objective measures Link does not seem like something worth pursuing. Among
other faults it would have almost no effect on traffic congestion and is not costeffective
as an alternative to driving.
8) BRT is a viable alternative to light rail in the Puget Sound Region. It could achieve
the same benefits at a much lower cost and has more than adequate capacity to handle
long term growth.
9) Link is the failed result of a faulty planning process. The process can and has been
manipulated to favor preordained outcomes. It is not objective. It does not produce
the type of information needed to make intelligent decisions. It fosters distrust. It is a
process that needs to be fixed. This would take local and Federal action.
10) The region is embarking on a rail-centric mass transit strategy, which could result in
over 125 miles of light rail. If fully implemented, that strategy will probably cost the
region about a billion dollars per year more than an all-bus (BRT) strategy having the
same level of transit ridership and related benefits. Meanwhile Link IS and Central
Link are probably costing over twice what comparable all-bus alternatives would
cost.
11) Link light rail is an example of the “waste, and abuse” that is driving up the Federal
budget deficit, because it was sold on the basis of misleading information and because
there are more cost-effective alternatives.
12) There has been insufficient public discussion about the merits or consequences of
committing the region to this multi-billion dollar rail-centric strategy, and there is no
solid analysis demonstrating it’s the best strategy.
13) Transportation planning in the Puget Sound region has not placed a high priority on
spending taxpayer money efficiently or in finding the lowest cost solutions.
14) Link should be put on hold and further Federal funding withheld until and unless a
new and honest alternatives analysis is completed, and that analysis demonstrates that
light rail is superior to BRT and other options. On the basis of available evidence,
such a conclusion seems unlikely.
xxxvii
15) The FTA should review, and if appropriate withdraw, its prior acceptance of the 1993
FEIS as meeting FTA requirements for a proper alternatives analysis, since that
particular analysis is faulty in so many respects and contains no apples-to-apples
comparison between Sound Transit’s light rail plans and all-bus alternatives. By the
same token FTA should not allow Sound Transit to proceed with any Phase 2
planning until a new and proper alternatives analysis has been completed.
16) Sound Transit’s planning process for Phase 2 will not provide the information needed
for this region to make intelligent decisions about massive investments in mass transit
unless the recommendations listed in Part 8 of this report are adopted.
17) The PSRC should be forced to obey the State Law requiring Least Cost Planning
because imposition of that planning technique is the single most important thing that
can be done to help ensure that scare transportation tax dollars are spent wisely.
18) Sound Transit’s latest Long-Range Plan should not be approved because it’s very
foundation, the 1993 FEIS, is obsolete and corrupt.
--end--
1
Part 1: Sound Transit has never done a proper alternatives
analysis.
1.1 Overview
Part 1 will document the fact that Sound Transit has never conducted a proper
alternatives analysis for Link light rail". In particular Sound Transit has never done an
apples to apples comparison of the light rail system they intend to build against other
alternatives such as bus rapid transit. As a result there still exists no factual basis for
asserting that light rail is a better solution for this region than other alternatives might be.
In fact, Parts 3,4,5 and 6 will provide strong indications to the contrary.
To be even more specific there exists no apples to apples comparison between:
a) The 14-mile Link “Initial Segment” (Ref 2) now under construction versus an
all-bus alternative
b) The 21-mile Central Link (Ref 18) voters approved versus an all-bus alternative
c) The 24-mile Central Link to Northgate version of Link (Ref 2) versus and allbus
alternative
d) The 100+ mile, fully built-out version of Link described in Sound Transit’s
Long-Range Vision (Ref 49), Draft Long-Range Plan (Ref 65), and in the PSRC’s
Metropolitan Transportation Plan (Ref 17: p. 73) versus an all-bus alternative
At the time of this writing Sound Transit has just released diagrams of the alternatives it
plans to evaluate for Phase 2. One alternative is a light rail network that includes all of
Central Link plus extensions to Northgate, Totem Lake, Redmond, and to the east of
Issaquah. It appears to involve between 50 and 70 route miles of light rail. There will be
two rail/bus alternatives, one using busways, the other using HOV lanes.
Sound Transit intends to conduct an alternatives analysis of the rail extensions per se, but
not the entire rail network. In other words they plan to take the core light rail network
from SeaTac to Northgate as a given in all three of the alternatives they plan to examine
for Phase 2. Thus when this Phase 2 planning exercise completes there will also be no
apples to apples comparison between:
e) The 50 to 70 mile (Phase 1 + Phase 2) version of Link versus an all-bus
alternative.
Because none of these alternatives analyses have been done, Sound Transit has no
scientific or analytic basis for its choice of light rail. Neither elected officials nor the
public has the data needed to see whether or not Sound Transit’s preoccupation with light
rail makes sense.
2
In lieu of a proper alternatives analysis, what Sound Transit uses to justify light rail is an
outdated, irrelevant and biased FEIS/alternatives analysis completed in 1993. (Ref:1)
That study compared a 125-mile, fully grade separated rapid or heavy rail system against
two bus alternatives. It is referred to frequently throughout this report as the “1993
FEIS”. See the attached Maps for the rapid rail and bus alternatives studied in the 1993
FEIS.
There is apparently no other, or more recent, alternatives analysis. The PSRC’s
"Summary of Prior All-Bus and Integrated Rail/Bus Alternatives Analysis" dated April 5,
2001 states: "The most recent and comprehensive system-wide analysis that evaluated
the relative costs and benefits of a rail/bus alternative and an all-bus alternative was
conducted as part of the Regional Transit System Plan (RTSP), completed in 1993." (Ref
28, p.2)
Nor are any other alternatives analyses mentioned in "Documentation of Major
Investment Study", dated March 12, 1996 (Ref: 3), or in a short report called "Sound
Transit Link light rail project, Transit Technology Review" dated 2/2/99 (Ref: 5), or in a
comprehensive listing of all high capacity transit studies conducted from 1967 to June
2001. (Ref 36, Appendix P) Moreover a very recent Sound Transit presentation called
"SEIS/Long-Range Plan Scoping Summary Report and Definition of Phase 2
Alternatives", Office of Policy and Planning, June 24, 2004 points back to the 1990-1993
alternatives studies but mentions nothing more recent. (Ref: 8) Finally, the just released
DSEIS for ST’s Long-Range Plan states definitively that it’s based on the 1993 FEIS.
In short, the 1993 FEIS is the one and only rail versus bus alternatives analysis this
region, and Sound Transit, has to rely on.
Unfortunately the rapid rail system studied in the 1993 FEIS is not like the light rail
system Sound Transit plans to build. The sizes of the networks are different, the speeds
of the trains are different, the capacity and levels of service are different, and the cost
estimates have proven wrong. In addition, the bus alternatives considered in 1993 were
not well designed; in fact they were designed to fail.
In short, the 1993 study does not provide adequate rationale as to why light rail should be
Sound Transit's technology of choice, nor does it demonstrate that Sound Transit's plan to
spend about $2 billion on a 14-mile Initial Segment of light rail is more beneficial than
spending an equivalent amount on some other alternative such as bus rapid transit. The
public needs to know. Public officials need to know. It is the essential and fundamental
question that the millions spent on planning should have answered. Yet it remains
unanswered as Sound Transit begins construction on Link’s “Initial Segment”.
Sound Transit is now beginning "Phase 2" planning that would extend that 14-mile initial
segment, and it appears the agency plans to rely again on that same irrelevant, obsolete
1993 study. That must not be allowed.
3
Before beginning any Phase 2 planning, and indeed before spending any more on Phase
1, our region needs an honest apples to apples comparison wherein Sound Transit's plan
to spend $X billion on light rail is compared with spending the same amount on BRT
and/or other alternatives. (Or alternately the costs of achieving a given level of benefit
using different plan alternatives are compared.) Equal cost or equal benefit is what is
meant by apples-to-apples.
The author believes the way Sound Transit has abused good planning practice is not
unique and has happened elsewhere. If so, what's needed at the national level is a
change in FTA regulations or enforcement thereof that forces transit agencies to conduct
honest and adequate alternatives analyses whenever new rail systems or extensions are
contemplated. For information on abuse in other cities, see References 12, 13, 21, 24 and
50.
Having a proper alternatives analysis is not the only requirement for a good planning
process, but it's an essential one. If there is one part of the process that should be singled
out and fixed, this may be the one with the highest leverage. That’s why it's the focus of
this report.
1.2 Alternatives analyses are required
According to the FTA's Office of Planning and Program Management:
"FTA's FY 2003 Strategic Business Plan identifies a commitment to delivering the highest value for
Americas investment in public transportation…" (Ref 31, p. 3)
"The name "alternatives analysis" has as its basis the New Starts planning provisions contained in
federal legislation…" (Ref 31, p. 10)
"Regardless of what the study is called, its intent is the same: to identify and compare the costs,
benefits and impacts of a range of transportation alternatives as a means of providing local decision
makers with the information necessary to implement the most appropriate transportation solutions in
priority corridors." (Ref 31, p. 10)
Additionally, common sense, taxpayer interest, the principles of business analysis, and
urban planning techniques taught in graduate schools all require proper alternative
analyses. Decision makers and taxpayers need to know what options are available for
solving their transportation problems. And, they need some way to compare the costs
and benefits of the different options so as to spend money wisely.
Much of this report is devoted to detailing the deficiencies in the 1993 FEIS which is
Sound Transits one and only alternatives analysis and apparently the one with Sound
Transit submitted to the FTA in support of their full funding grant agreement for Link.
The FTA apparently accepted this flawed document because their formal acceptance of
the grant states:
4
WHEREAS, the Government has determined that the Project is based on the results of an alternatives
analysis…. (Ref 60)
This author is not familiar with the details of FTA requirements as to how alternatives
analyses should be conducted. He does not know if they are so loose that practically
anything that a transit agency says is an alternatives analysis will be accepted by the
FTA; but considering what Sound Transit has been able to get away with that must be the
case. I leave it to others to determine whether or not Sound Transit has met the "letter of
the law” in terms of FTA’s formal requirements. It's clear they have not met the sprit of
the law, or the needs of those wanting objective information.
1.3 Author's definition of a minimal alternatives analysis
In this letter I define an alternatives analysis as a study wherein one fairly detailed transit
system plan is compared against other detailed transit or non-transit system plans such
that the costs and benefits, including environmental impacts, of each alternative are
quantified. The systems plans must be detailed enough to show routes, stations, and
frequency of service. The analysis must include running mode split and other models so
as to predict transit ridership, impacts on road usage, and other relevant impacts and
benefits. These are what I consider the minimal requirements for something to be called
an alternatives analysis.
A qualitative description of the pros and cons of different alternative technologies does
not comprise an alternatives analysis. Sound Transit and the RTA have produced a
number of these. Most are very short. See for instance the Appendix in Ref 5.
Using my definition there has been only one alternatives analysis relevant to major
regional mass transit planning in the Puget Sound region since 1990; namely the Final
Environmental Impact Statement Regional Transit System Plan, dated March 1993. (Ref:
1) It was produced by Sound Transit's predecessor, the Regional Transit Authority or
RTA.
1.4 Author's definition of a proper alternatives analysis
To be a proper alternatives analysis other common sense requirements apply. For
instance the systems alternatives compared should include the one actually planned to be
built not some obsolete or irrelevant plan, the cost data should be timely and not proven
wrong, the analysis should be objective and not biased, the analysis should show how
well each alternative contributes to solving the main transportation problem the
community wants solutions for, a decent range of alternatives should be considered, and
so forth. In a nutshell the alternatives analysis should attempt to seek out the best
solution to the problem based on objective performance criteria, not serve as a sales pitch
or biased justification for some politically predetermined outcome.
In other words it should actually and honestly seek to: “delivering the highest value for
Americas investment in public transportation” (Ref 31, p. 3)
5
Using this definition Sound Transit has never conducted a proper alternatives analysis for
the light rail system they plan to build.
The specific reasons that the 1993 alternatives analysis (Ref 1) should not be accepted as
the basis for Sound Transit's light rail planning are detailed in the remainder of Part 1.
They fall into two general categories. First, the 1993 FEIS was in itself flawed and did
not even constitute a proper alternatives analysis for the rapid rail system being proposed
in 1993. Second, about 12 years have passed since the homework behind the 1993 FEIS
was completed. In the interim Sound Transit has switched from rapid rail to light rail,
initial cost estimates have proven wrong, and a host of other things have changed. All
these factors would render the 1993 FEIS obsolete and irrelevant to today’s situation,
even if it had been done correctly back in 1993.
1.5 The 1993 alternatives analysis is irrelevant because it addressed heavy or rapid
rail, not light rail.
The 1993 alternatives analysis that ST has used to justify its light rail plans was not an
analysis of light rail vis-a-vie other alternatives, but rather an analysis of heavy or rapid
rail, which has significantly different characteristics.
There can be no doubt that the 1993 FEIS evaluated rapid or heavy rail since the term
was used so widely and defined so precisely: (underlining by author)
“The fifth plan element is a regional rapid transit system…” (Ref 1: p.xiii)
"This (Rail/TSM) alternative includes about 125 miles of rapid rail running on an exclusive,
grade-separated right-of-way and a 40-mile diesel-operated commuter rail line. For planning and
modeling purposes it has been assumed that the rapid rail system would have operating
characteristics similar to recently built system in other cities and that it would have similar
environmental impacts. Within the range of other North American rail lines, the rail system that is
proposed would fall into the definition of "heavy rail", since it would have a relatively high
capacity and would operate almost entirely on an exclusive guideway completely separated from
automobile traffic." (Ref: 1, page xxxii). In contrast, parts of Sound Transit's light rail will
operate at grade in the Rainier valley thus limiting its speed and capacity. It also has different
safety characteristics than the grade-separated system in the 1993 FEIS.
"A rapid rail line operating in exclusive right-of-way has the theoretical capacity to carry over
22,000 persons per hour in each direction past a single point. These numbers represent the rail
capacity of the downtown Seattle transit tunnel; capacity would be higher in segments with less
frequent station stops." (Ref: 1, page xviii) In contrast, the maximum capacity of Link light rail is
claimed to be 16,000 pph in downtown Seattle and on the north line. South and east lines would
have a much lower capacity.
“The capital improvements analyzed in the system Plan FEIS include a proposal for a rapid rail
system…” (Ref 1: p. 1-2)
“A rapid rail line with 4-car trains….” (Ref 1: p.1-10)
6
“2.2.3 Rail/TSM Alternative The Rail/TSM alternative overlays an extensive rapid rail system
and a commuter rail line onto….” (Ref 1: p.2-24)
“2.2.3.1 Technology This alternative would be based on an rapid rail system on exclusive, grade
separated right-of-way…Maximum speeds would be 55 to 70 MPH, with average speeds
(including station stops) around 35 to 40 MPH.” (Ref 1: p.2-24)
Because the 1993 FEIS compared rapid, not light, rail with bus alternatives it does not
constitute an adequate alternatives analysis for Link, in spite of Sound Transit’s
assertions that it does. It offers no proof that spending $X billion on light rail would
bring more benefit to the region than spending the same amount on other alternatives
such as bus rapid transit.
1.6 Link light rail is much slower than the rapid rail studied in the 1993 FEIS.
The rapid rail in the 1993 FEIS had an average speed of 36 MPH whereas Link will have
an average speed of only 26 MPH. (Ref 38: p.16 and Ref 1: p. 2-58) The rapid rail in the
1993 FEIS also had a higher top speed. “Maximum speeds would be 55 to 70 mph, with
average speeds (including stations stops) around 35 to 40 mph." (Ref 1: page 2-24) In
contrast Links maximum speed is 55 mph. (Ref 2: page 3-9).
As another point of reference, WASHDOT policy requires that HOV lanes be managed
so average speeds equal or exceed 45 MPH.
The fact that Link is slower than the rail system in the 1993 FEIS means that all the
benefits asserted for rail in that study would be diminished if light rail were used instead
of rapid rail. In other words, the 1993 FEIS makes the “rail” alternative look better than
it should be, if rail now means light rail whereas before it meant rapid or heavy rail
Because it is slower Link would attract fewer riders than rapid rail, even if the routes and
stations were the same. With fewer riders the cost-effectiveness measures computed in
the 1993 FEIS would not apply to a fully mature 100+ mile version of Link. Link would
look worse.
In addition, all the other benefits that are related to ridership, such as impacts on traffic,
air pollution, energy consumption, and the degree to which the alternative supports land
use goals would be lower.
The fact that Link is slower will diminish whatever advantage the 1993 FEIS asserted
that rail would have over bus. Perhaps it would disappear entirely. In other words, if
Link speeds were plugged into the models used to estimate ridership in the 1993 FEIS,
perhaps the bus alternatives would have attracted MORE riders than the rail alternative
and left no basis to claim rail was superior.
In sum, by basing today’s decisions on a dated alternatives analysis that makes rail look
better than it should, we are not making rational decisions based on relevant data.
7
1.7 The rail network used in the 1993 alternatives analysis differs considerably in
size and routing from the network options Sound Transit intends for Phase 1 and 2.
A proper alternatives analysis needs to compare the actual configuration an agency plans
to build against other feasible alternatives, rather than compare hypothetical or irrelevant
configurations. If an agency is allowed to compare hypothetical network configurations
it might as well also compare them in the context of hypothetical land use patterns or
even use comparisons from other cities. Clearly there are places in the world where
heavy rail is clearly superior to light rail, or where light rail is superior to bus rapid
transit, or where bus rapid transit is superior to light rail. It all depends on the specifics,
including the size and shape of the proposed networks.
The 1993 FEIS compared one specific 125-mile rail network with several bus-based
alternatives.
This network is not similar in size or shape to any of the light rail networks that Sound
Transit may end up with. Currently, Sound Transit only has FTA approval to build a 14-
mile starter system. That may be all that ever gets built. Or maybe it will be extended to
SeaTac but not to the U District, or vice versa. Or maybe to Northgate. Or maybe it will
be extended across Lake Washington, perhaps with a branch to Issaquah, or maybe not.
The merits of these potential end-state LRT networks vis-a-vie BRT are not addressed by
a prior study of one particular 125-mile rapid rail network. See attached Maps for details.
Common sense argues that there should have been an alternatives analysis comparing
Central Link with other alternatives before ST asked voters to approve it in 1996. They
would have been better informed. And it should have been required before the FFGA
was approved for the starter system. One could also argue that the current FFGA is
invalid because a proper alternatives analysis was never completed. The need for one
remains as long as the community could still change its mind and Phase 1 contracts be
cancelled, or as long as it is still possible to terminate the Initial Segment south of the
DSTT.
Phase 2 planning is now started and Sound Transit must not be allowed to maintain that
the alternatives analysis has long since been completed, as they do in the Draft Long-
Range Plan. Instead, common sense argues that we should now have an alternatives
analysis for each of the main end-state scenarios that may emerge so we can determine
which if any make sense. Thus we should have an apples to apples comparison of at least
the following light rail scenarios vis-a-vie an equal amount spent on other alternatives:
1) a truncated Initial Segment that stops south of the Downtown Seattle
Transit Tunnel so it doesn't interfere with a regional BRT system.
2) 14-mile Initial Segment as currently approved
3) 21-mile Central Link
4) 24-mile version of Central Link going to Northgate
5) ?-mile version of Link from Everett to Tacoma
8
6) the 125+-mile light rail system that constitutes Sound Transit's fondest
dream for Phase 2 (see Ref 8: page 4 and Ref 65) See also PSRC’s MTP
which envisions 125 miles of light rail. (Ref 17: p.73)
All these potential light rail end-state scenarios should be studied since, while a short
starter system may not make sense, perhaps a more extensive system would. Alternately,
perhaps a larger system would exceed the point of diminishing returns, whereas a short
"skim the cream" system would make sense. If light rail makes sense at all, there will be
an optimum balance between that part of the transit network served by rail versus that
best served by bus. This sweet spot can only be located by bracketing it with the
evaluations of various end-state scenarios such as outlined above.
Beside route length, the layout of the 125- mile rail network used in the 1993 Plan differs
substantially from the light rail network ST now envisions for Phase 2. For instance the
1993 rail alternative included a North-South line running east of lake Washington from
Everett to Sumner, whereas Phase 2 "Development Options" include no such route. (Ref
2: Figures 2-10 and 3.20) (Ref 8: page 4) The alignment chosen for estimating ridership
in the 1993 FEIS may also have been the Duwamish alignment, although that isn’t clear.
(Ref 2: page 2-32) If the Duwamish alignment were used, that would be another
difference between what was studied in the 1993 FEIS and what Sound Transit wants to
build.
In sum, the cost, ridership, and impact data collected in the 1993 FEIS are not relevant to
impending decisions about building or extending light rail because the shape and size of
the rail network analyzed in the 1993 FEIS is not equal to any of the Link’s network
scenarios. It is therefore necessary to conduct an alternative analysis, in other words a
mode against mode “fly-off”, for each potential end-state scenarios for light rail to
determine which, if any, make sense.
1.8 The bus alternatives in the 1993 FEIS don't reflect the best design for bus
alternatives. They are obsolete in terms of what can now be done, and may have
been designed to fail.
There are several reasons why the bus designs in the 1993 FEIS do not reflect the best
that buses could do today, and therefore render the 1993 FEIS obsolete.
First, the 1993 FEIS makes no mention of BRT. Therefore the bus alternatives in that
study are unlikely to have incorporated all the recent innovations in Bus Rapid Transit.
As explained in the extensive BRT literature these include a host of ways for making bus
service more efficient and attractive. (see References 9, 10, 11, 12, 13, 14)
Second, there was no knowledge at the time that the Alaska Way viaduct needed to be
replaced due to earthquake vulnerability, nor that there would be a major effort to
redevelop the south Lake Union area. These offer the opportunity to build an elevated
"BRT/HOV way" that would branch off the I-5 HOV lanes at Mercer St., cross the south
9
Lake Union Area, tie into SR 99, and head south along what is now the Alaska Way
viaduct. Just one story up, this "BRT/HOV-way" could probably be architecturally
integrated into the planned new buildings so wouldn’t be visually intrusive.
Providing another major new N-S transit/HOV corridor through the city would remove
much of the capacity limitation held against the TSM option. With a south Lake Union
station this new busway could provide high quality transit service to this emerging
employment center. In addition, by heading west toward the Seattle Center then south
along the waterfront this BRT route could improve transit service to Ballard, Belltown,
the biotech development on the north shore of Elliot Bay, the commercial/industrial area
south of the stadiums, the Duwamish valley, and West Seattle. These areas are all
ignored by light rail. The attractiveness of this addition would not have been apparent in
1993 and is another reason the 1993 FEIS is obsolete in terms of the light rail versus BRT
choice facing us today.
Third, the 1993 FEIS claimed that bus alternatives didn't have enough capacity thru
downtown Seattle. However, the RTP failed to consider a scenario where through buses
would be accommodated by extending the I-5 HOV lanes through downtown, or solve
the capacity problem in a variety of other ways that staff had already identified. In this
respect the bus alternative in the 1993 FEIS was a strawman alternative deliberately
designed to make the bus alternatives look bad. This complex subject is treated in detail
in Part 5 of this report.
The point here is that Sound Transit’s choice of light rail over bus cannot legitimately be
based on an alternatives analysis wherein the bus alternatives were deliberately designed
to fail. For this reason alone the FTA should have refused to accept the 1993 FEIS when
processing Sound Transits request for a Full Funding Grant Agreement (FFGA).
Fourth, the 1993 FEIS did not contemplate the hybrid bus technology that is now
available. The availability of this technology has operational, cost, and environmental
implications different from buses evaluated in the 1993 FEIS.
In sum, the bus alternatives in the 1993 FEIS are dated, biased, and do not represent bus
technology at its best. The 1993 FEIS is therefore irrelevant to impending decisions
about if and how to extend Link light rail. To be relevant in today's environment, new
alternative analyses -- that compare Sound Transit's light rail plans against several
modern well-crafted BRT alternatives-- are needed. This is a most fundamental point.
1.9 The 1993 alternatives analysis is irrelevant because it is based on obsolete and
incorrect cost assumptions
Since virtually any transit technology can be made to work, one of the most important
reasons for doing an alternatives analysis is to determine relative costs.
10
Officials currently leaning on the 1993 FEIS to help justify their choice of light rail over
bus are looking at obsolete and incorrect cost data. History shows that cost per mile
assumptions made in the early 1990s were wildly optimistic. Since time has proven that
the cost assumptions in the 1993 FEIS were incorrect, that document is not relevant to
today's decision making.
The author would like to show the magnitude of this error, in other words quantify the
difference between rail costs assumed in the 1993 FEIS and the latest figures based on
actual construction cost bids, which in turn reflect a much greater level of engineering
analysis than had been done in 1993. Unfortunately this is very difficult for several
reasons. First Sound Transit has not stated all their cost estimates in a consistent manner,
such as 1995$. Second, they have started using YOE$ which appear to vary depending
on the latest implementation schedule. Third, some estimates include reserves and other
relevant expenses whereas others do not. Fourth, the "plan" has changed from the 125-
mile system, to a 21-mile system to a 14- mile system so it is difficult to say what would
now be the cost of the original 125-mile system. What follows are the best indications of
percentage cost escalations since the 1993 FEIS that the author has been able to locate.
First, some data in constant dollars: When voters approved the original 21-mile light
rail plan in 1996 the cost was estimated to be $1.736 billion (in 95$), or $86 million per
mile. (Ref 2: p. S-49) (Ref 15: p.34). Sound Transit's May 1996 Sound Move brochure
said: "Sound Move is based on extremely conservative cost and ridership assumptions."
(Ref 15: p. 31) Voters no doubt relied on that statement.
However, by the time the FEIS for Central Link was published in November 1999 the
cost had risen to $2.1 billion (95$) or $100 million per mile, and by Sound Transit’s
board meeting in Jan. 2001 it was up to $2.6 million (95$) or $123 million per mile, a full
44% higher than what voters approved. (Ref 2: p. S-49) (Ref 27, Section 13, p.1-8)
Then, since the original plan was no longer affordable, Sound Transit decided to shorten
the route and eliminated the most costly portions. In short, these sources show costs
increased 44% (in 1995$) from the mid-1990's to January of 2001. An earlier chart by
the author shows this visually:
Cost History (billions of 96$)
Vote
1996
FEIS
Oct 99
ST Board
Mtg Jan 01
EA
Feb 02
1.7
2.1
2.6
1.5
shorter
easier
route
11
Second is a view in YOE$: On 2/2/01 Sound Transit published a table labeled "Updated
2001 Financial Plan". (Addendum G) The table shows that light rail (the Phase 1 system
to NE 45th) changed from $2.161 billion (YOE$) in the Sound Move plan to $4.0 billion
in the 2001 Updated Financial Plan. The $4.0 billion figure also appears in the Central
Link Board Briefing Book dated Jan 11, 2001. (Ref 27: Section 13, p. 7-8) In sum, this
YOE data shows the costs of Central link jumped 85% during the period from 1996 to
2001.
Third is a June 2004 analysis by transportation planner Jim MacIsaac: Mr.
MacIsaac found that the total cost for the 24.8-mile Central Link has risen from $2.3
billion at the time of the vote to somewhere between $5.335 billion and $6.75 billion.
These are YOE$. This means cost per mile has doubled or tripled from $93 million per
mile back in the 1996 timeframe to somewhere between $215 and $272 million per mile
today. Mr. MacIsaac's table is inserted below.
12
It is clear from the record that between May1996 and January 2001 Sound Transit had
learned enough --from more detailed engineering and input from construction firms-- to
realize their initial cost estimates were off by about 44%. What we don’t know is
whether or not Sound Transit learned anything before May 1996 or after January 2001
that would have also changed their cost estimates. In other words were the underlying
unit cost assumptions used in the 1993 FEIS the same as those used to estimate the cost
of Central Link in the Sound Move brochure, and are the unit costs used in Jan 2001 still
valid today? If there were increases during those periods they would be in addition to the
44%. A new alternatives analysis based on the latest unit cost data is needed in order to
get a reasonably accurate estimate of what the 125-mile system would cost today.
However, in the meantime the author will simply assume --for the analyses in Part 6--that
the only change is 44%. In other words, the author will assume that the cost of the 125-
Link Capital Cost Estimates Excluding all Reserves and Financing Costs
Segment Length Sound Cap Cost MacIsaac Estimates
(miles) Move 2001 Est1 Current2 per Mile 2009$8 2012$9
Link North: (YOE$) (YOE$) (YOE$)
Northgate - U-District3 3.2 $600 $775 $242 $937 $1,070
U-Dist - CPS (Montlake)4 4.9 $1,720 $1,945 $397 $2,353 $2,685
CPS - Lander & Maint Base5 3.1 $1,871 $680 $715 $231 $790 $790
Lander - McClellan5 1.0 13.8 mi $300 $300 $300 $330 $330
McClellan - Boeing Acc5 4.8 $580 $580 $121 $640 $640
Total North 17.0 $1,871 $3,880 $4,315 $254 $5,050 $5,515
Link IS North Only 8.9 $1,560 $1,595 $179 $1,760 $1,760
Link South:
Boeing Acc - S.154th (Link IS 4.9 $480 $610 $124 $670 $670
S.154th - Airport Term7 1.7 $427 $250 $210 $124 $255 $290
Airport Term - S.200th 1.2 $200 $200 $167 $245 $275
Total South 7.8 $427 $930 $1,020 $131 $1,170 $1,235
TOTAL CENTRAL LINK 24.8 $2,298 $4,810 $5,335 $215 $6,220 $6,750
Link IS Extensions only 11.0 (21.2 mi) $2,770 $3,130 $285 $3,790 $4,320
1 ST April 2001Capital cost estimates in YOE$ based upon 2009 opening (excluding all reserves and financing costs).
2 Most recent estimates per 2004 financial plan for Link IS; per footnotes 3-7 for extensions (assumes all completed by 2009).
3 40% of CPS to U-District segment, per ratio of Base Construction estimates presented in North Link DSEIS.
4 May 2004 ST disclosure of total capital excluding reserves = $1775 in 2002$ x 1.095 to YOE (2006)$.
5 CPS - Boeing Access = $1534m in YOE$ per ST 2004 Link IS estimates. Add $61 for DSTT ROW.
Broken down into three subsegments by JWM estimates aided by the ST 2001 breakdowns.
6 The 2001 estimate reflected alignment in median of SR-99. Since changed to mostly elevated "Tukwila route".
7 This segment was originally to be elevated. The revised POS drive plan now makes it possible to build much at-grade.
8 Includes 10% project reserve for all segments plus inflation factor of 1.1 for 3-year delay of all Link IS extensions.
9 Includes an inflation factor of 1.2 for 6-year delay of all Link IS extenxions plus 15% project reserve for extensions.
Capital Cost Estimates
13
mile rail system in the 1993 FEIS is not the $7.9 billion (95$) it was then thought to be,
but --based on what we know today-- is more like $11.38 billion (95$).
To summarize, cost assumptions made in the early 1990's have proven grossly inaccurate,
partly because the difficulties of construction in Seattle's difficult terrain were
underestimated. Light rail is much more expensive than was anticipated in the 1993
FEIS. It is at least 44% higher and perhaps more. This makes the cost-effectiveness
measures in the 1993 FEIS obsolete. A new alternatives analysis based on updated costs
is needed.
Part 6 of this report continues this analysis of cost issues, and ends up with a cost
comparison between building a mature 100+ mile version of Link and building a BRT
system having equal ridership.
1.10 The 1993 alternatives analysis is irrelevant and obsolete because employment
patterns have changed
Since 1993 employment patterns have changed significantly, very likely in ways not
anticipated in land use inputs to the modal split modeling behind the 1993 FEIS. For
instance, employment at Boeing has dropped and shifted. Microsoft employment has
grown dramatically. Paul Allen has relatively recent plans for dramatic growth and
redevelopment in the South Lake Union area. Biotech employment has grown in the
north Elliot Bay area. There has been much residential high-rise construction in Belltown.
There is significant employment and residential growth planned on the Sammamish
Plateau just north of Issaquah. These land use changes affect the relative merits of light
rail versus BRT since light rail would not serve many of the new growth areas whereas
BRT could.
In summary, a FEIS based on 10 or 15-year old land use assumptions is no suitable basis
for today's light rail planning.
1.11 The 1993 alternative analysis is inadequate because it did not quantify the
impact of each alternative on the region's main transportation problem, namely
traffic congestion.
For the last decade at least, virtually every newspaper article about Puget Sound's
transportation problem emphasized that "THE" problem is traffic congestion. Not airquality,
not energy consumption, not inadequate mobility for the poor, but rather the
traffic congestion suffered by individuals and by companies trying to conduct business.
One of the many examples is Seattle Times lead editorial of July 16, 2001 entitled
"Legacy Time: Congestion is the No. 1 Issue".
Polls echo this finding. For instance the PSRC's Puget Sound Household Travel Survey
concluded: "Respondents agree that reducing traffic congestion should be the primary
goal of transportation plans." (Ref 16, p. ii) A Sound Transit Poll said: "It is also very
14
important to the majority (62%) of the respondents that the projects remove single
occupancy vehicles from the roadways, even if they do not plan on using the services
themselves." (Ref 22, p.15). A Mineta Institute Report entitled "Why Campaigns for
Local Transportation Funding Initiatives Succeed or Fail" says: "Results from the
(Evans) poll indicated that a majority of respondents were highly concerned with the
issue of traffic and congestion…" (Ref 21, p.89) The Evans McDonough poll is also
summarized in Ref 27, Appendix B, p.4-8.
The fact that traffic congestion should be the main metric for evaluating transportation
system performance is reinforced by the fact that the Puget Sound Regional Council (our
MTO) used "delay" on freeways and arterials as the main comparison metric in the
alternatives analysis behind the Metropolitan Transportation Plan. (Ref. 17, p. xvii) (Ref
30, p.24) Delay is a good proxy for traffic congestion.
Even RTP (Sound Transit) acknowledged the primacy of congestion reduction on the first
page of its 1997 Documentation of Major Investment Study (Ref 3)
"Chapter I. Planning Process
The problem is traffic congestion, including its effect on transit performance. The goal is to
provide a cost-effective public transportation system in the Central Puget Sound Region that
provides an attractive alternative to the single occupant automobile." (Ref. 3, p. i)
However, in spite of all this, the effects of the different alternatives on traffic congestion
was never measured in the 1993 FEIS, as evidenced by the lack of a congestion or delay
metric in the 'evaluation criteria' listed on pages xxvii and xxviii or in the key results
found in Table 3.9.3. (Ref 1, p. 3-95) Nor is reducing congestion mentioned in the goals
listed for the project on page 1-44 of the FEIS. The 1993 FEIS sidestepped the issue
with a few innocuous statements. For instance, under "Purpose and Need" the report
simply said "Implementing the System Plan by itself will not solve congestion on the
region's freeways and arterials or environmental problems associated with automobile
use." (Ref 1, p. xvi). Nor would one expect it to. The point is to understand the
magnitude of its contribution. This the 1993 FEIS did not do.
By way of context, in it's pre-vote public hearings Sound Transit continually implied that
light rail was a solution for the congestion problem. That became the myth and a
probable reason why voters approved the system in 1996. The myth seemed to linger so
in a December 22, 2000 Seattle Times OP ED the author attempted to widely debunk it
with data from Sound Transit's own documents. (Ref 20). Within days the Chairman of
the Sound Transit Board finally admitted in a responding OP ED that light rail would not
fix traffic congestion.
Sound Transit has continued to avoid the embarrassing issue of how little its rail system
would reduce traffic congestion. Neither the more recent 1999 FEIS for the original 21-
mile Central Link Light Rail (Ref 2), or the 2002 Environmental Assessment for the 14-
mile Initial Segment (Ref 18, p. x) measure the effect of the light rail system on traffic
congestion. What little Sound Transit does say is hidden well back in the report. For
instance page 23 in the 2002 FEIS says: "Overall congestion levels for the downtown
15
intersections would likely remain the same compared to the No-Build Alternative." (Ref
18, p.23)
The way in which Sound Transit uses traffic congestion as a driving rationale yet fails to
say specifically whether or not light rail reduces it is neatly illustrated in Appendix B of
the Dec 2000 Sound Transit Board briefing book. (Ref 27, Appendix B) Here, under
"Why Link", Sound Transit lists its reasons why "Light Rail is critical for Central Puget
Sound". None of the reasons listed include anything about reducing traffic congestion.
Yet three pages later they describe a public opinion poll finding that "Reducing Traffic
Congestion" is the third highest regional priority, outranked only by providing a quality
education and maintaining a strong economy. In short, there is a serious disconnect
between what people want, what Sound Transit would like them to believe they are
getting, and what Sound Transit knows it's delivering.
It is also interesting to note how the "purpose and need", or "goal" of the rail project has
evolved away from trying to help solve the congestion problem. The 2002 EA now says:
"The purpose of the project is to construct and operate an electric light rail system
connected to the region's major activity centers." (Ref 18, p. 2) This is tantamount to
saying the purpose of the project is to build the project.
In sum, the 1993 FEIS cannot be considered a proper or relevant alternatives analysis
since it did not quantify the degree to which each alternative helped solve the region's
primary transportation problem, namely traffic congestion. Additionally, Sound Transit
continues to manipulate thinking by citing traffic congestion as a rational for action while
at the same time protecting themselves legally by not committing in writing that light rail
will do anything about it.
1.12 The cost data in the 1993 FEIS is now obsolete because much of the HOV
network needed for the bus alternatives has since been completed and represents a
sunk cost.
The 1993 FEIS estimated the capital cost for each of the four major alternatives. (Ref 1,
p. xxxiii) The cost to "complete, expand, and enhance" the HOV network was included
in the cost of TSM for each alternative. (Ref 1, p. 2-7)
The three build alternatives each included $1.5 billion to complete the HOV network.
(Ref 1, p. xxx) However, because much of the HOV work has since been completed
some or all of that $1.5 billion can now be subtracted from the cost of the alternatives.
That shifts the relative costs of the rail versus bus alternatives making the bus alternative
--which was half the cost of the rail alternative-- even more attractive. (see Ref. 1, Table
4, p. xxxiii for specific values). Having better data on relative costs is germane. This is
another reason the 1993 FEIS is obsolete and irrelevant in making a light rail vs. BRT
modal choice in today's environment.
16
1.13 The 1993 FEIS, and the FEIS's for Link IS and Central Link, are all seriously
flawed because the stated "Purpose and Need" was either vague or not tied directly
to solving the regions main transportation problem, which is traffic congestion.
1993 FEIS: First, the statement of purpose is not clear. There is no sentence in the
Purpose and Need section that clearly and unambiguously says: 'The purpose of the Plan
is to….'
Instead, under Purpose and Need the 1993 FEIS begins by stating:
"The System Plan presents an effective mass transit alternative to problems associated with
automobile use…" (Ref. 1 ,xvi)
One can get very confused trying to parse this poorly written phrase and determine what
it really means. I would simply say that such a vague and undefined statement of purpose
is not adequate to justify a multi-billion dollar project.
A much clearer expression of purpose could have been made, and should be required.
For instance the Purpose could have been: 'reduce traffic congestion by diverting auto
users onto mass transit' or 'improve mobility for the young, old, and poor by improving
public transit service', or whatever.
Secondly, the Purpose for the region’s largest transportation project is not linked to the
regions major transportation problem. The Purpose and Need section lists three
problems, which are held forth as justification for the Plan. Congestion is the common
element in each:
"Congestion constricts travel on most major freeways…"
"Slower and less predictable travel times result from congestion"
"Transit and HOV modes are caught in congestion…"
Again, if congestion is the acknowledged problem should not the main Purpose of the
Plan be that of reducing congestion? If the Plan is directed toward solving some lesser
problem that should be made clear and the project should be of lower priority.
It should not be acceptable that an FEIS seek justification and support by, on the one
hand, pointing to a problem of great public concern --and thus implying that somehow
the Plan is a solution to that problem--, while on the other hand refusing to make clear
just how the Plan helps solve the problem or officially commit that it will do so.
In sum, the Purpose and Need section in the 1993 FEIS is defective in two ways:
First, the "Purpose and Need" statement does not make clear either the problem the
Plan is intended to help solve or the way in which the Plan (or project) would help
solve it. It says there is a problem, and it says there is a Plan. What it does not do
17
is show the relationship between them. In other words it does not commit that the
plan will solve, or help solve, the problem.
Second, the "Purpose and Need" does not commit the Plan to help solve the regions
most pressing transportation problem, which in this case is traffic congestion. A
project that is not so committed should have a relatively low priority for FTA
funding.
Central Link FEIS: The “Purpose and Need” section in the 1999 FEIS for Central Link
starts out with a very clear albeit useless statement: “The purpose of the proposed light
rail project is to construct and operate a starter electric light rail system designed to
connect several of the regions major activity centers:…” (Ref 2: p.1-1)
This is tantamount to saying the purpose of the project is to build the project. This
ridiculous statement does not commit the project to solving any particular problem or
adding any particular benefit. However, if Sound Transit can get away with such a
statement they are legally or perhaps morally bound to deliver nothing other than a
completed project.
Subsequent paragraphs say the plan “would deliver substantial time savings”, “would
greatly improve transit capacity” “would connect Northgate to the University District”
and so forth. (Ref 2: p. 1-6 thru 1-8) These are purported benefits but they are not stated
as the Purpose for project and presumably Sound Transit is not promising or committing
that they will result.
Link Initial Segment: Under Purpose and Need the 2002 EA for Link IS says: “The
purpose of the project is to construct and operate an electric light rail system connecting
the region’s major activity centers.” (Ref 18: p.2)
Clearly Sound Transit’s lawyers missed this one because it says rather directly the
purpose is to “connect the regions major activity centers”. There are about twenty
officially designated Urban Growth Centers in the Puget Sound region. Link IS does not
even connect two of them together. Per the FEIS, Link stops at South 154th street, which
is north of the SeaTac airport and north of the SeaTac Urban Growth Center. The South
154 street station is within about a half mile of the South Center Urban Growth Center
but separated by a freeway and completely inaccessible. See attached maps.
However even if the purpose statement was just an innocent mistake, the larger criticisms
apply; namely: As it reads the Purpose of the project is to build the project, and the
Purpose does not commit to helping solve the traffic congestion problem.
1.14 The 1993 FEIS mishandled the capacity issue in a way that penalized the bus
alternative.
18
This was a critical deficiency in the 1993 FEIS that has had serious consequences.
However it is a complex subject, which is addressed at length in Part 5 of this report.
1.15 The 1993 FEIS mishandled the cost-effectiveness issue in a way that penalized
the bus alternative and hid important information from officials and the public.
This is also a complex point. Part 6 is devoted to explaining it. Part 6 also takes data
from the 1993 FEIS and uses it to project the cost of building a mature 100+ mile version
of Link versus the cost of building a BRT system having comparable transit ridership and
ridership dependent benefits.
19
Part 2: Sound Transit morphed rapid rail into light rail.
2.1 The 1993 alternatives analysis specifically dismissed light rail as not being
suitable for the region.
Light rail was, along with other technologies like monorail, given cursory consideration
in the 1993 FEIS, then dismissed as inappropriate. The FEIS contained the following
statements:
"A surface LRT system like MAX in Portland would operate at average speeds of 18 to 20 mph,
relatively slow compared to the grade-separated Rail/TSM Alternative, which would average 35 to
40 mph." (Ref 1: page 2-50)
"Surface LRT operating across intersections is typically limited in terms of train length and
frequency. … Conventional transit practice and highway standards suggest that when train
frequencies are under 6 minutes, cross traffic on arterials well be affected to the extent that grade
separation is necessary….These constraints limit the capacity of surface LRT systems, as
compared to grade-separated systems." (Ref 1: page 2-50)
"While surface LRT has been very successful in some systems due to low-cost right-of-way or a
very dense urban setting, its operating performance relative to grade-separated systems is
generally characterized by slower speeds, lower ridership, lower capacity, and lower reliability.
These characteristics mean that surface LRT is unlikely to satisfy the demand of a three-county
system." (Ref 1: page 2-50)
At the time of the 1993 FEIS a group called the Puget Sound Light Rail Society or
PSLRTS proposed a 78-mile light rail system as an alternative to the 125-mile rapid rail
system favored by the RTP. Pages 2-51 through 2-61 discuss that alternative. The RTP
concluded that:
"Surface LRT options were analyzed to the point that it became clear that these options did not
adequately serve the goals and objectives of the Regional Transit Project. Because of the superior
performance of the grade-separated RTP system in terms of consistency with land-use objectives,
level of service, and ridership, it was recommended as the rail technology in the recommended
draft Systems Plan." (Ref 1: page 2-61)
See also the RTP's disparaging remarks about a light rail system -proposed by R2B2-
that included a route in Rainier Valley. (Ref 1, p.2-62)
The term “rapid rail” was used consistently throughout the 1993 FEIS to describe the
Rail/TSM alternative. Thus there can be no question that’s what the 1993 FEIS analyzed
in the form of the Rail/TSM alternative.
It is curious, if not indeed disingenuous, that Sound Transit should so disparage light rail
in their one and only alternatives analysis, only later to say –as they do in the Long-
Range Plan-- that this same analysis now justifies their selection of light rail as the
technology of choice. (Ref 65)
20
The distinction between the "heavy" or "rapid" rail studied in the 1993 FEIS and "light"
rail is not just semantic. Sound Transit will argue that some of the characteristics of the
rapid rail in the 1993 FEIS, such as train length, are identical to that of the light rail now
being planned. That's true. However, the rapid rail in the 1993 FEIS never ran at grade
down the middle of a street whereas parts of Link light rail does. Not all the way, but far
enough to affect headways and capacity on the entire south route and forever shortchange
the entire south end in terms of capacity. Nor are we certain that proposed Phase 2 Link
routes east of Lake Washington would be entirely grade separated as they were in the
1993 FEIS. All the ridership projections, and the cost effectiveness measures in the 1993
FEIS were based on a 100% grade separated network.
Surface running light rail will result in more accidental deaths and injuries, and have
different impacts on street traffic that would the rapid rail in the 1993 FEIS. What was
said or not said in the 1993 FEIS about these topics would not apply to Link. There are
also differences in speed and capacity between the rapid rail in the FEIS and Link light
rail. In short, the "rapid rail" in the 1993 FEIS was not just light rail under a different
name.
In sum, the 1993 alternatives analysis for rapid rail does constitute a valid alternatives
analysis for Link light rail.
2.2 Sound Transit morphed the heavy rail it studied into the light rail it wants.
State, Federal, and common sense all require that ST have an alternatives analysis for
Link. But the way ST has handled this requirement was to begin asserting that the 1993
alternatives analysis for rapid rail was actually an alternatives analysis for light rail.
They morphed an alternatives analysis for rapid rail into one for light rail. This was not
done by actually modifying or redoing parts of the 1993 FEIS to account for the
difference, but rather just by changing what they said.
This has been an effective tactic, sometimes called the “big lie”. History has some
chilling examples. The big lie is that when supposedly knowledgeable and credible
persons or organizations repeatedly assert that something is true, many people will come
to believe it’s true, even if it isn’t. Thus when Sound Transit and RTA assert the 1993
FEIS was an alternatives analysis of light rail, many will accept that as fact. Some
example quotes where the term light rail is being used to describe the rapid rail system
studied in the 1993 FEIS:
“Regional Transit Project (RTP) …The analysis recognized the limited ridership potential because
of the shoreline location of the rail line, especially in comparison with the high cost, fully gradeseparated
light rail alignment along I-5 between Northgate and Everett.” (from RTA report dated
1997, Ref 3: p. 20)
____________________
“1992 Regional Transit Project—Transit Technology Overview. Recommendation: Light rail,
commuter rail and buses.” From a Sound Transit brochure called FACT dated Aug 2002 (Ref 6)
__________________________
21
“3.1 Conclusions of Past Systems Planning Work
All previous work leading to the adopted Long-Range Transit Vision generally concluded that, outside
the corridors with existing rail lines where commuter rail or DMU technologies were possible, the
most viable HCT options for the region were LRT and BRT.” (from a 2004 ST report, Ref 7: p.7)
The most important “previous work” was the 1993 FEIS, which specifically dismissed
light rail. See the actual quote in the following subsection. The deceit continues in a
Sound Transit presentation where one slide said:
“1990-1993 System Alternatives
* Final Alternatives:
* Enhanced bus,
* Transitway/Busway (BRT)
* Rail
* Commuter
* LRT “ (on a slide presented June 24, 2004 by Sound Transit) (Ref 8)
This slide is particularly egregious since it clearly refers to the 1993 FEIS and it clearly
says one of the final alternatives was light rail (LTR). That’s just not true.
There is another deception in the above-mentioned slide where it says
"Transitway/Busway (BRT)" was also one of the final alternatives. In fact BRT, while
similar in many respects, is not synonymous with the Transitway/TSM alternative
constructed by the RTP. BRT can include a host of innovations designed to make buses
more attractive. See References 9, 10, 11, 12, and 13. The 1993 FEIS never mentions
the term “BRT” and thus is unlikely to have incorporated all these innovations. It is
therefore also deceptive for Sound Transit to state in this slide that their one and only
alternative analysis considered BRT.
In short, this 2002 Sound Transit slide says that light rail and BRT were given a
competitive “fly-off” back in the 1990 to 1993 period. Unless there was a phantom
alternative analysis done between March of 1993 and YE 1993, this is an outright lie.
Albeit in light of similar "misstatements" by Sound Transit it's not surprising.
The PSRC has also written deceptively in this regard:
2. What elements were included in the Regional Transit System Plan (RTSP) adopted by the
JRPC in 1993? … Overall, the long-range RTSP, as subsequently modified into Sound Transit’s
long-range “Vision Plan”, proposed the following components for the regional transit system:
* approximately 125 miles of light rail….(Ref 28: p.2)
The Regional Transit System Plan (RTSP) published in March of 1993 did not include
any light rail, it actually denigrated light rail. (Ref 2) What it did propose was 125 miles
of rapid or heavy rail as the quote below clearly states:
“Elements of the System Plan …The fifth plan element is a regional rapid transit system, with either
buses running on a combination of HOV lanes and transitways…or a regional rapid rail system…
(Ref 2: p. xiii)
22
Sound Transit, with some help from the PSRC, has a behavior pattern of slowly
morphing the record into showing what it wants, and misleading officials and the public
into thinking something other than the truth. Its latest attempt, as of August of 2004, is to
change the definition of high capacity transit (HCT) to mean transit running on an
exclusive right-of-way. Exclusive right of way is not a necessary condition for HCT. In
fact, BRT intermixed with car and vanpools could carry far more passengers than Link
light rail. (See Part 5 for details.)
If Sound Transit is successful in so redefining HCT it will be a clever trick for ruling out
any alternative that would mix BRT buses in with car and vanpools on, for example, the
I-90 center lanes. It will nicely dovetail with Sound Transit's other current maneuver to
officially re-designate those lanes for exclusive use by HCT only. The consequence of
these seemingly innocuous maneuvers is clear. If HCT means exclusive right-of-way and
I-90 center lanes are dedicated to HCT, then all alternatives that would mix BRT with car
and van pools are simply taken off the table and never even studied. This will help make
the selection of light rail across Lake Washington almost a foregone conclusion of the
Phase 2 studies. This appears to be Sound Transit’s intent.
2.3 Why did the FTA let Sound Transit finesse the alternatives analysis?
The author has no direct knowledge of the answer and hopes the Congress will
investigate. However, to an outsider it seems that the FTA would be motivated to
increase its statement-of-work and thus budget by establishing and moving applicants
through the most elaborate planning process possible. This keeps staffs happily
employed. However, at the end of the day the FTA would have no incentive to deny
marginal projects and return money to the treasury. All bureaucratic incentives for
agencies like Sound Transit and the FTA would seem to mitigate against finding the most
cost-effective solutions, but rather would encourage pursuing the most expensive projects
sellable to the public. Hence Sound Transit's refusal to conduct a proper alternatives
analysis. Hence FTA's failure to demand one. At root it boils down to pork barrel
politics at taxpayer expense. Harsh words perhaps, but a plausible explanation for
observed behavior.
References 12,13, 21, 24 and 50 discuss this matter from a national perspective. They
show that Puget Sound is not alone in the way rail boosters and transit agencies have
abused the planning process. The author highly recommends Martin Wachs excellent
article on the ethics of forecasting. (Ref 24) Part 7.1 of this report ends with a long and
interesting quote from that article.
23
Part 3: The shortcomings of Link Light Rail
All would agree the vision of a rail system whisking people from place to place is
appealing. However, many other things are appealing too: things like replacements for
the Alaska Way viaduct and the 520 bridge, things like more and better paid teachers,
things like more parkland, things like fixing Seattle’s deteriorating streets or the Mercer
mess. There is never enough money to satisfy all the region’s needs so the main issue is
not whether light rail is appealing, but rather whether it is appealing enough to be worth
its enormous cost. In other words, is light rail the least cost way to help solve the regions
transportation problems? It’s doubtful. Analysis shows there is good reason to question
Link’s cost-effectiveness relative to other alternatives. As former Washington State
Governor Booth Gardner said about Link:
“It costs too much and does too little” (Op Ed, Seattle Times)
Beside being critical of Link per se, the author is of course critical of Sound Transit for
not providing the public with objective information about Link’s performance. In the
sections below the author presents some of that information, using bar charts to
graphically explain the points.
Sound Transit has tried to hide this type of information. The skeptical reader is invited to
try to find the performance measures revealed below (such as cost per car removed from
traffic) in any of Sound Transit’s Environmental Impact Statements or brochures. What
little information ST provides about effectiveness or cost-effectiveness is hidden in
technical tables far back in Environmental Impact Statements. There are no bar charts
communicating the main points so they stand out and can readily be grasped. It is clear
by their absence that Sound Transit does not seek to inform the voter and help him or her
reach rational decisions. Instead most of what ST publishes is marketing material meant
to sell and manipulate.
Link light rail is one of those things where the less you know about it, the better it
appears. What follows is a list of Link’s specific shortcomings. They support the notion
that Link is the failed result of a faulty process.
3.1 Link does almost nothing to reduce congestion
Data in Sound Transit’s EIS’s show that Link light rail system won’t even make a dent in
Puget Sound’s traffic congestion problem.
Given that Sound Transit says “traffic congestion is the problem”, one might expect the
agency to publish data showing how much Link would contribute toward solving the
congestion problem. However, Sound Transit did not publish in their EIS’s any estimate
of Links effect on traffic congestion, or on travel delay, which is a fairly good proxy for
congestion. Instead we are left to infer Links effect on traffic congestion from the only
relevant measure ST does provide; namely, Links effect on regional vehicle miles of
travel or VMT. To the extent Link reduces the load on the road system, in other words
24
reduces VMT, then delay and congestion should drop also. (This is an admitted
oversimplification since congestion is a complex phenomenon involving latent demand,
long term home and work location decisions, and other behaviors.)
Without light rail, PM peak period traffic on area roads and freeways is forecast to be
16,996,000 vehicle miles of travel (VMT) in 2020. With Central Link light rail this
would be reduced to 16,933,000 VMT. (Ref 2: p.3-4) These figures are accurately
plotted in the chart below. Central Link would reduce peak period traffic a little over one
third of one percent. It is equivalent to removing one car of every 370. These VMT
figures are the best available measure of Central Link’s ability to help solve the region’s
vexing traffic congestion problem.
One way to put 1/3 of 1% into perspective is imagine the 2.75- mile long Evergreen Point
bridge, covered with cars stopped bumper to bumper from the Lake’s eastern shoreline to
Montlake Blvd. At 20 feet per car there would be about 725 cars in each lane. Reducing
traffic by 1/3 of 1% is equivalent to removing about two and a half cars from each lane.
Even where rail should have its greatest impact, the numbers are disappointing. For
instance, the number of peak hour vehicles crossing the ship canal in 2010 would only
decline from 45,789 to about 45,740. The number of peak period auto trips leaving
downtown Seattle would only drop from 30,800 to 30,100. Both are incredibly small
impacts.
0
5000
10000
15000
20000
Regional VMT in PM Peak in 2020
Without
Central
Link
With
Central
Link
Source: Central
Link FEIS,
Table 3.1-4
VMT in
000’s/day
16996 16933
25
Figure 3.1.2 is another way to put Link’s minuscule effect on traffic congestion into
perspective. The total “pie” represents the amount by which regional travel, measured in
VMT/day, is expected to grow between 1998 and the year 2020; namely by 22,384,000
VMT. This is due to population growth and other factors. The small wedge shows how
much of that growth increment would be taken away by drivers giving up their cars to
ride Link IS. Basically Link IS would reduce the growth in VMT by 81,000 daily VMT.
Unfortunately, this savings would be wiped out in only 29 days by the normal growth in
regional traffic. In other words, after spending about 13 years (1996 to 2009) and $1.5
billion (02$) building Link IS, travel will be back to pre-Link levels only a month after
the system opens for business.
Central Link would reduce traffic by 271,000 daily VMT, so its effect would take about
three months for growth to overcome.
Some argue Link’s Initial Segment is only a starter system, which will grow and
eventually produce more meaningful results. Here again the data show otherwise. The
1993 EIS prepared for a $11.5 Billion (in 91$) transit plan shows a 125-mile rapid or
heavy rail system would only reduce 2020 traffic volumes 1.9% more than a bus solution
costing a fraction as much. (Ref 1: Table 3.9.3) Figure 3.1.3 illustrates this graphically.
If the reader can’t see the difference between the “rail” bar and the “without rail” bar, it’s
apparent that drivers wouldn’t notice any difference on the roads either.
LRT effect on daily regional VMT
growth between 1998 and 2020
Reduction due
to Link IS = 81
Growth = 22,384
Values are daily
VMT in 000's
14 mile Link IS = 81
21 mile Central Link = 271
Link IS' savings
are wiped out in
29 days by
normal traffic
growth
9
26
3.2 Light rail is not a cost-effective way to get cars off the road.
Sound Transit continually states that Link is cost-effective, as for example in a press
release/handout entitled “Light Rail Benefits” wherein one of the six benefits listed is
“Cost Effective Transportation” (Ref 64) This seems quite deceptive in light of the
following.
Capital and operating costs for Central Link total $291 million/yr in 02$. (See Tech
Notes for details) Central Link would remove about 6320 cars from peak period traffic
in 2020. (See Tech Notes for details). If the main objective of Link is to remove cars
from peak period traffic then it’s fair to assign Link’s full cost to that benefit. Doing so
means it would cost $46,000 per year for each car Central Link removes from peak
period traffic. This per year cost would continue for the 30-odd years it would take to
pay off the construction bonds.
Capital and operating costs for Link IS total $191 million per year in 02$. The Initial
Segment would remove 1890 cars from peak period traffic. This means it would cost
$101, 000 per year for each car Link IS removes from peak period traffic.
Figure 3.2 shows these costs graphically; however, values in the figure came from an
earlier study by the author and differ slightly from those given above. The “vote” bar is
for Central Link and is based on the cost estimates Sound Transit published at the time
the vote was taken in 1996. The “Jan 01” bar is based on the latest cost estimate for
Central Link, which was made in Jan 2001 and which was $2.6 billion in 02$. The “EA”
bar refers to Link IS. The chart shows that the cost of one of Link’s main benefits (i.e.:
taking cars off the road) increased 233% over what voters approved, yet the Sound
Transit Board has never agreed to go back and ask voters if they still wanted to proceed.
Effect of 125 miles of Heavy Rail
on Regional Traffic
60,000 est.
*1993 FEIS, daily VMT 84,189* 82,524*
2000 without rail with rail
11
Figure 3.1.3
In 2020 In 2020
27
The idea of paying $46,000, or even $100,000 per year for each car Link removes from
traffic is clearly a ridiculously expensive way to reduce traffic congestion. It is ridiculous
because the Seattle Times reported on a local company that increased carpooling by
offering their employees just $75 a month to carpool. It is ridiculous because one can
imagine how many people would car or van pool if offered , to pick a number, just $500 a
month or $6000 a year. It is ridiculous because it would simply be cheaper to pay people
to quit their jobs and stay home.
In spite of Link being one of the most expensive light rail systems in the country and
costing tens of thousands of dollars a year for each car it removes from traffic we still
have curious statements like the following from Sound Transit Board member and Wash.
State Secretary of Transportation Doug MacDonald in a memo to Norman Mineta,
Secretary of U.S. Dept. of Transportation:
“Every transportation investments in our region must be tested for its cost-effective long range
contribution to moving people and goods in the region. Link Light Rail meets that standard.” (Ref
58)
3.3 Light rail is not cost-effective as an alternative to driving.
Some officials, like Ron Sims former chair of the Sound Transit Board, trumpet that light
rail provides an alternative to driving. Indeed it does for the relative few that, according
to Sound Transit’s ridership forecasts, would give up driving to ride light rail.
Each car taken off the road takes about 1.25 people off the road. Thus the 6320 cars that
Central Link would take off the road by 2020 would be providing about 7900 peak period
commuters an alternative to driving. This is only two-tenths of one percent of the
4,100,000 people forecast to live in the Puget Sound region by 2020. (Ref 2: p1.3).
Annual cost per vehicle removed
from peak period traffic
$ 30,000
$ 43,000
Operating cost (02$) $ 100,000
Capital cost (02$) + 233%
from vote
Vote in 1996 Jan 01 EA in 2002
28
If the main objective of Link is to give people an alternative to being stuck in traffic like
everyone else then it’s fair to assign Link’s full cost to that benefit. This means it would
cost taxpayers $37,000 per year for each commuter who gives up using an automobile to
ride Central Link.
It is questionable whether the 4,100,000 persons living in Puget Sound would like to
build a $2.6 billion rail system just so 7900 would have a way to bypass congestion. If
so, to paraphrase Winston Churchill: never before will so many, have paid so much, to
benefit so few.
These figures look even worse for Link IS since it has only one third the predicted
ridership of Central Link. The per-person cost for the 2365 auto users that would switch
to Link IS comes to $81,000 per year.
In this, as in many other areas, Sound Transit seeks to deceive the taxpayers. On the one
hand they say a main objective of light rail is to provide people with an alternative to
driving. On the other hand they do not say how many people will take advantage of this
benefit or what this benefit will cost everyone else.
3.4 Link is far more expensive than a bus rapid transit (BRT) system having
comparable benefits
This conclusion is mainly based on the detail provided in Parts 5 and 6 of this report.
Part 5 establishes in considerable detail that an all-bus alternative based on bus rapid
transit (BRT) could attract and carry the same number of transit riders as light rail. This
is true whether the light rail system be a short one such as the 14-mile Initial Segment of
Link, or a fully mature version of Link extending over 100 or more route miles.
Part 6 establishes that a fully mature version of Link would probably cost Puget Sound
taxpayers about $900 million per year more than an all-bus system capable of attracting
and carrying the same number of riders, and thus having comparable benefits.
What follows is some information from a completely different source that confirms this
reports conclusion that BRT is less expensive than light rail.
Studies, such as one by the US General Accounting Office, have concluded that BRT is
usually less expensive than light rail and should be taken more seriously as an alternative
to light rail.(Ref 9) Figure 6 from the GAO report is reproduced below as Figure 3.4.
To build this chart GAO surveyed 18 light rail lines built since 1980, 9 busways, and 8
HOV facilities. It shows the average light rail system costs about $35 million per mile.
29
The following presentation slide compared the cost of Link’s Initial Segment with the
national average for light rail systems and BRT using data from the GAO report. It
included a cartoon by Seattle area’s outstanding cartoonist Dave Horsey, which along
with other articles of that period, called public attention to Link’s high cost.
34.8
13.5
9
0.7
0
5
10
15
20
25
30
35
40
1 2 3 4
Cost per Mile (millions of 2000$)
Light Rail BRT on
busways
BRT on
HOV lanes
BRT on
arterials
Capital cost per mile for light rail
and bus rapid transit
Figure 3.4a
30
This slide simply illustrates that Link IS, at $ 128 million per mile, is far more expensive
than the national average for light rail; perhaps because Puget Sound’s terrain is not
“light rail friendly”. We have water barriers and hills that require extensive tunneling.
Additionally, we do not have the abandoned railroad rights-of-way that make
constructing light rail in other cities relatively economic.
What the Puget Sound region does have, however, is an extensive HOV network. There
were 191 lane miles of HOV open to traffic as of Jan. 2001, and another 100 miles are in
various stages of construction or planning. (Ref 46: p.2)
These completed HOV lanes are the region’s most precious transportation resource.
They provide an already paid for “guideway” for bus rapid transit, as well as a guideway
for car and van pools, school buses, emergency vehicles, and other multi-occupant
vehicles. They can be managed so traffic flows at a relatively high speed, and they have
enormous people moving capacity. Unlike the “guideways” for Link light rail that will
take years to build, these HOV “guideways” already exist, and BRT service on them
could be implemented fairly quickly. In fact, some BRT service already exists in the
form of express bus routes.
Figure 3.4b helps visualize the tradeoff this region is making by going with light rail.
The figure shows that we could have about 199 miles of BRT on the regions HOV lanes
for about what the 14-mile Initial Segment of Link is costing.
Benefit of Switching from LRT to BRT
14 miles of
Light Rail
199 miles of BRT
on HOV lanes**
* Cost of Initial Segment in 2000$
** based on $9 Million/mile
in GAO report
$1.8
Billion*
Figure 3.4b
31
Of course, the $9 million per mile figure cited by the GAO might not apply here, so this
is a very rough approximation. (The data in Part 6 is considered more accurate) It does
however suggest the wisdom of conducting a proper alternatives analysis to see just how
extensive a BRT system could be built for what Link IS is costing. Certainly 199 miles
of freeway BRT would attract far more transit riders and do much more to reduce
congestion, clean the air, control sprawl, etc. than would 14 miles of light rail.
3.5 Link will require more transfers
Express bus or BRT routes will typically circulate through residential areas picking up
riders, then enter an HOV lane for a non-stop trip downtown. In many cases riders do not
need to transfer. If Link is built many of the express bus routes will be redirected to feed
riders onto light rail. This will increase the number of riders who need to transfer.
3.6 Many of the regions employment centers and key destinations would not be
served by Link.
The Puget Sound Region has over 15,000 miles of local roads and streets and 185 miles
of freeways.(Ref 43: p.A4:3) In contrast the Link network would be somewhere between
14 and 125 miles in length. With such a sparse network Link will not be able to provide
convenient service to many of the regions employment and activity centers.
For instance, 14-mile Link IS, which runs from Westlake Station to South 154th Street,
only serves one of the region’s major employment centers: downtown Seattle. 21-mile
Central Link would serve only downtown Seattle, the University District, and SeaTac.
However, Central Link will not serve a host of other important places including: West
Seattle, Bellevue and the entire eastside, Renton, Southcenter, Magnolia, Ballard,
Bothell, the south Seattle industrial area along the Duwamish, Harbor Island, the
emerging south Lake Union biotech center, the north Elliott Bay biotech center, and so
forth. None of Boeing’s plants or offices will be served: not Auburn, not Kent, not
Renton, not Plant 2, not the Longacres complex, not Bellevue, not even Boeing Everett.
Even if Link is someday extended into the mature network Sound Transit envisions, it
will still bypass many of these locations.
NOTE: References 8, 43 and 49 have maps showing where Sound Transit plans to
extend “high capacity transit”, which officially could mean either light rail or BRT. In
practice, there is little doubt that Sound Transit officials have already decided informally
that HCT really means light rail. One piece of evidence is that Sound Transit has been
maneuvering to have the center lanes on I-90 designated for exclusive use by HCT, even
before doing an alternatives analysis. However, if the agency were objective it would not
ask for exclusive use until that alternatives analysis is complete, because if the answer
were BRT, exclusive use would not be needed, indeed it would be counterproductive.
32
Another piece of evidence is that the baseline Technical Report for PSRC’s 2001 Update
to the Metropolitan Transportation Plan makes the assumption that the HCT network will
be almost entirely light rail. (Ref 30: p. 75) Finally, there is the Fleet Management Plan
for Central Link which says: “While design work is progressing for the initial system
from NE 45th Street to S 200th Street, plans for rail system extensions are underway.
...Additional extensions include an Eastside extension to Bellevue and surrounding
communities.” (Ref 38: p.10)
3.7 Investments in Link do nothing to encourage car and van pooling.
Car pooling and van pooling are actually more effective than mass transit as a way of
reducing the number of peak period cars on the road. Figure 3.7 shows modal share in
the 3-county area in 1990 and 2000.
Building light rail does nothing to encourage car and vanpooling. On the other hand,
implementing an all bus alternative would include not only finishing the HOV lanes, but
spending a considerable sum to build HOV to HOV connector ramps at major freeway
interchanges, and to build flyovers so HOV vehicles could enter and exit the HOV lanes
without cutting across several general purpose lanes. All this concrete will not only
benefit BRT, but would also make car and van pooling more attractive.
In other works the investment in HOV lanes, which are the “guideways” for BRT, has a
double payoff. This incidentally, is just one more thing that that the 1993 FEIS ignored
and that a new and proper alternatives analysis should address.
0%
20%
40%
60%
80%
100%
All
Trips
Work
Trips
3
35
62
Transit
Car/van pool
Single
occupancy
vehicle
7
21
72
Mode Share for Trips in Puget Sound Region in 1998
Source: Ref 17:
p.xvii)
Figure 3.7
33
3.8 Placing Link on I-90 would reduce bridge capacity, compromise safety, and
impair car/van pooling.
Right-of-way is perhaps the scarcest resource in Puget Sound, and there is a need to get
the highest utilization out of it.
This topic of I-90 capacity will be treated in more detail in Part 5.16. However, the short
version is that the I-90 bridge will have more capacity if the center lanes are used for a
mixed stream of BRT buses, carpools, van pools, school buses, etc. than if they are turned
over to light rail. This is true because mass transit, whether BRT or rail, will have a
relatively light loading on I-90. Neither can really keep the center lanes fully occupied.
Even in peak periods there will be either the occasional train, or the occasional BRT bus.
Best to use a technology where car and van pools can fit in between transit vehicles.
Sound Transit is currently pushing for a reconfiguration of the lanes on I-90 from
Bellevue into Seattle that would not only preempt the reversible center roadway for mass
transit but also add a pair HOV lanes into the space now occupied by 6 general purpose
lanes. To do this the 6 GP lanes would be narrowed and re-striped into 6 GP and two
HOV lanes. Lane width would drop below interstate standards and shoulders would be
narrowed. This in turn would have two deleterious effects. First, the narrower lanes
would compromise safety for all bridge users. Second, the narrower HOV lanes and
inability to provide a buffer strip between the GP and HOV lanes would slow car and van
pools and thus make them less attractive. (See Reference 33 pages 6-16 and C-1 for
leads into the technical literature relative to HOV lane design and related safety issues.)
3.9 If not extended, Link could result in underutilization of the DSTT
This is a risk with light rail, not a certainty. The issue is getting maximum people moving
benefit from King County's investment in the DSTT. If light rail does not go across Lake
Washington the DSTT will be able to deliver a maximum of about 21,000 persons into
downtown during AM peak hour (or out in the PM). This would be comprised of 16,400
from the north and 5500 from the south. In contrast buses in the tunnel could move about
32,000 persons per hour into the downtown, half from the north and half from the south
and east. This topic, of potential interest to property owners in downtown Seattle, is
treated in more detail in Part 5.17.
3.10 If not fully extended, Link will compromise the regional BRT system.
The Downtown Seattle Transit Tunnel is critical for either a regional light rail system or a
regional BRT system. But the tunnel can’t easily serve both.
If Link’s Initial Segment is all the light rail that ever gets constructed, Link will simply
become a local intra-Seattle people mover. By default the regional HCT solution will
become an extension of the existing regional express bus system, in other words BRT.
However, if Link IS operates in the DSTT, the regional bus system will be compromised.
34
Either BRT buses will be forced onto surface streets, or not as many will be able to use
the DSTT as may be required.
This is a good reason for keeping light rail out of the DSTT until it is absolutely certain
that taxpayers are behind extending Link into a full regional system. The cost of
extending Link into a full regional system is given in Part 6 of this report. It is so large
that voters may not approve the necessary tax increases.
3.11 Light rail would be less effective at controlling sprawl than an all-bus system
Assertions in the 1993 FEIS that rapid rail supports the goal of reducing sprawl seem
questionable. One assertion prominent in the FEIS’s executive summary was:
“The Rail/TSM Alternative would also fully support regional land-use plans limiting urban sprawl
and concentrating new growth in existing centers.” (Ref 1: p.xxxv)
The RTA provided no rationale, cited no sources, to justify the notion that rail transit
helps limit sprawl. What follows is a very brief analysis.
First, there is the question of just why rail is said by some to be a means of controlling
sprawl. History shows that rail can actually cause sprawl. Anyone familiar with the
bedroom communities that have grown up along the rail lines radiating from Manhattan
or Chicago knows that building these lines enabled or encouraged people to move to rural
areas and create the suburbs. The same is true with BART’s extensions over into the east
bay. By the same token building light rail to Redmond or Issaquah would encourage
sprawl by making it convenient for people to move to the Sammamish plateau or
Northbend, then drive to the end of the light rail line for a commute into Seattle. Logic
says that anything that makes road or transit travel easier will probably lead to the spread
of housing and economic activity. The 1993 FEIS admitted this:
“Stations with park and ride lots located near the urban growth boundaries many encourage some
urban sprawl.” (Ref 1: p. 3-130)
The only scheme whereby an improvement in transportation would seem able to limit
sprawl is if road and transit travel within the urban growth boundaries were made easy,
while travel outside the urban growth boundaries were made difficult. Presumably this
would be an incentive to live within the urban growth boundary. Of course, under
current law there would be nothing to stop people living outside the boundary to drive
over the boundary line and enjoy the same travel benefits as those living within. And the
idea of not improving roads outside the growth boundary, or preventing those outside
from driving across the boundary line is politically impractical. In other words, this
whole scheme seems unworkable. What then is the mechanism by which mass transit
might control sprawl?
Rail fans claim that high-density development will occur around rail stations. Research
questions the accuracy of that, finding that high densities are more the result of tax
incentives than the existence of rail transit. However, assuming it were true, why would
35
more development occur around a rail station than would occur around a BRT stop? An
even bigger question is why would high-density development around a mass transit stop
or station be of any significance in controlling sprawl? Presumably the issue with sprawl
is the expansion of housing and other development outside the urban growth boundary
and into rural areas. If so, then whether the new development occurs as high-rises
adjacent to mass transit stations, or as moderate density development spread throughout
the urbanized area, makes no difference.
In conclusion, it appears that any improvement in transportation, including building light
rail, will allow if not encourage sprawl. The possibility that high capacity transit might
encourage development to occur in high rise clusters strung like beads along a rail or
BRT route, as opposed to letting that development spread out within the designated urban
growth boundary, seems somewhat irrelevant.
There is one additional point to make along this line. If society wants development to
cluster in high rises along transit lines, what difference does it make if the transit is steel
wheeled light rail or rubber tired BRT? A mature light rail network would probably have
about the number of route miles, and follow the same corridors, as a mature BRT
network.
This is illustrated by comparing a map of Sound Transit’s vision for a fully mature light
rail network with a map of the region’s HOV network. (The left-hand map below shows
the region’s freeways. HOV lanes are either in place or planned for most freeways. The
right hand map shows the rapid rail network studied in the 1993 FEIS. A fully mature
light rail network would be similar. Ref 49: p. 10-11 maps it, but the map was too large to
scan into this report.)
36
The second line of reasoning has to do with timing. Voters approved light rail in 1996. It
is now 13 years later and not an inch of rail has been laid. It will be 2009 before even the
first 14 miles of light rail becomes operational. There are serious funding problems to
overcome before Link could even be extended to Northgate. (Mr. MacIsaac has written
about the financial impediments to expanding Link.) Yet more years will pass before
Figure 3.11
37
Link could be extended into the full regional network Sound Transit envisions. Even if a
mature version of Link were able to control sprawl, much more sprawl will have occurred
before Link could become that mature network and begin to reduce it.
Here BRT has an advantage. The guideways for BRT are partly in place. BRT-like
service already exists and could be intensified into a fully functional BRT system fairly
quickly.
This leads to the following conclusions:
1) There is no apparent reason why either light rail or BRT would help limit sprawl
outside Puget Sound’s designated urban growth boundaries.
2) There is no apparent reason why a light rail network would control the pattern of
density within the urban growth boundary (e.g.: make beads of high density
development along transit lines) any better than a BRT network having about the
same number of route miles and serving the same corridors.
3) The higher cost for light rail means that the region is more likely to build a large
BRT network than to build a large light rail network, and more likely to build X-miles
of BRT years sooner than X-miles of light rail. Therefore in the foreseeable future
pursuing BRT would probably have more effect in concentrating development along
transit lines than would pursuing light rail, again assuming it has any effect at all.
3.12 Link is inequitable in whom it taxes versus whom it benefits
Sound Transit taxes everyone within the “RTA District”, which comprises most of the
urbanized areas within Snohomish, King, and Pierce Counties, stretching from Ft. Lewis
to Everett. (Ref 49: p.6 has map) Yet Link’s benefits are largely confined to the narrow
corridor along which Link runs. Thus many of the people who are taxed for Link will not
benefit, either directly or indirectly. They will not benefit directly since Link routes will
not come near their communities, West Seattle and Capital Hill being good examples.
They will not benefit indirectly since Link will have almost no effect on traffic
congestion.
3.13 Link is inequitable in where it provides for safe operation and where it does
not.
When the Sound Transit Board decided to provide tunnels for Link’s north line but not
provide tunnels in the Rainier Valley they established an inequity in terms of safety. The
history of light rail has been characterized by collisions between trains and pedestrians,
and between trains and cars. Sound Transit has plans to mitigate such accidents, but they
will still occur, and that harm will occur to Rainier valley residents, not residents of north
Seattle. Transportation consultant John Niles has investigated and written about this
matter, but Sound Transit shrugs it off.
38
3.14 Link is susceptible to terrorism, power outages and earthquakes.
Link is susceptible to single point failures, bus systems are not. Link will have a complex
control system, rails and tunnels all of which could be disabled relatively easily by
terrorists. A couple pounds of explosives or a few cut wires would bring the entire
system to a standstill. An area wide power outage would take the system down. An
earthquake could damage rail and BRT structures, but rail would be more susceptible
because damage to underground tunnels would take a long time to repair and the trains
have no way to drive around problem areas. Any train that stalls on the track, for
whatever reason, blocks the entire line.
In contrast, bus systems are comprised of independent little units called buses. They don’t
rely on a centralized control system. They contain their own power source. They can
drive around obstacles.
3.15 Link is a whole new technology to manage and maintain.
It will take a whole new and different maintenance infrastructure, and a different set of
skills, to support light rail. This will be expensive. In contrast, BRT would simply mean
an incremental expansion of the bus maintenance infrastructure already in place. ST’s
consultants said the following about automated guideway technologies but it applies
equally well to light rail.
“Potential expansion of ST’s various lines of business will require expansion of supporting facilities.
…any new technology such as automated guideway that is beyond what is currently in place within
ST’s service area will require more substantial supporting efforts. These supporting efforts will
include major new facilities such as maintenance bases, layover track, control center, etc. In addition
the interfaces between any new technologies and ST’s existing system will require expensive
infrastructure to allow passenger transfer between the modes. …Introduction of automated guideway
technology would require a new layer of administrative support….” (Ref 7: p.3)
3.16 Light rail is not compatible with Puget Sound’s topography and local
circumstances
Steel wheel on steel rail limits light rail to a maximum of 5% grades. The author has
seen the route profiles for Link. The 5% grade limitation forces Link to be far below the
surface on Capitol Hill so it can climb up from the DSTT then dive deep enough to go
below the unconsolidated glacial till at the bottom of the ship canal. It then struggles to
climb the hill in the University District, but can’t handle the grade so again requires deep
underground stations in the U District. In other words light rail is fighting our relatively
unique terrain.
The turn radius of rail system is larger than bus and thus rail systems cannot be routed
around buildings and other obstacles as easily as bus routes.
39
Puget Sound also lacks the abandoned railroad rights of way that light rail has benefited
from elsewhere.
These factors are probably responsible for light rail costing more here than in other cities.
For instance, according to Joni Earl, ST’s Executive Director: Portland’s light rail cost
$54 million per mile, Denver’s $20 million per mile, Sacramento’s $35 million per mile,
Dallas’ $41 million per mile, and San Diego’s $61 million per mile. (Ref 39) In contrast
Central Link would cost $150 million per mile.(Ref 39). (Reference 9 also contains cost
comparisons.)
In short, just because light rail was an easy fit in some cities does not make it the right
technology for Puget Sound’s unique conditions.
3.17 Dedicating the center lanes on the I-90 bridge would compromise traffic flow
and apparently disadvantage pedestrian and bike users.
Sound Transit is actively maneuvering to have the lane configuration on the I-90 bridge
drastically altered so as to free up the center roadway for light rail. ST has published
diagrams of the various options. (Ref 59) According to those diagrams the westbound
roadway is currently 52 feet wide and has three 12-foot wide traffic lanes. However
Sound Transit wants the R8a option, which calls for the roadway to be 58 feet wide, and
for it to be divided into 4 lanes. The extra 6 feet would apparently come from making the
pedestrian and bike path narrower.
The width of the lanes would not be consistent; some would be 12 feet while others
would drop to a sub-standard 11 feet. That should cause some confusion, especially as
trucks shift from one to the other. Shoulders would also be narrowed.
The ST handout also warns (in very small type font) about traffic problems that may
result from these narrower lanes and shoulders:
“A number of operational strategies to address safety concerns associated with reduced-width travel
lanes and shoulders will be evaluated. These include speed management through variable posted
speeds and/or reduced speed limits, shoulder rumble strips,….” (Ref 59)
What ST is admitting is that automobile traffic will be adversely impacted, both by
forcing speed reductions and by introducing the kinds of safety issues that call for rumble
strips.
40
--blank page--
41
Part 4: Bus rapid transit (BRT) and other alternatives
4.1 Some background
4.1.1 A brief definition and rationale for BRT
Scott Rutherford, Professor of Civil Engineering at University of Washington, along with
others, has offered the following:
“Bus Rapid Transit (BRT) is growing in popularity throughout the world The reasons for this
phenomenon include its passenger and developer attractiveness, its high performance and quality, and
its ability to be built quickly, incrementally, and economically. BRT also provides sufficient
transport capacity to meet demand in many corridors, even in the largest metropolitan regions. In the
United States the development of BRT projects has been spurred by the Federal Transit
Administrations BRT initiative. These projects have been undertaken, in part, because of the
imbalance between the demand for “New Starts” funds and available resources.
Decisions to make BRT investments should be the result of a planning process that stresses problem
solving, addressing needs, and the objective examination of a full range of potential solutions, of
which BRT is only one.
…The FTA defines BRT as a “rapid mode of transportation that can combine the quality of rail transit
and the flexibility of buses”
A more detailed definition, …, is:
BRT is a flexible rubber tired rapid transit mode that combines stations, vehicles, services, running
way, and ITS elements into an integrated system with a strong positive image and identity. … In
many respects, BRT is rubber tired light rail transit (LRT), but with greater operating flexibility and
potentially lower capital and operating costs.” (Levinson, Zimmerman, Clinger, Rutherford, “Bus
Rapid Transit: An Overview”, Journal of Public Transportation, Vol 5, No.2 2002)
4.1.2 Seattle is not alone in ignoring the bus alternative.
Consider the following:
“Strong views exist on the merits of light rail as a preferred alternative to dedicated bus-based
transitway systems. Why did many of these cities supporting and building light rail not consider
having a very flexible bus system on the dedicated alignment, which has the capability of offering
much better door to door service than a very inflexible fixed rail system? The answers are relatively
simple -- the adage that ‘trains are sexy and buses are boring’ (quoted from the Mayor of Los
Angeles) says it all. We have previously described this as ‘choice versus blind commitment’
(Hensher and Waters 1994).
When the evidence suggest that one can move three times as many people by dedicated bus-based
transitway systems for the same cost, or the same number of people for one-third the cost as light rail,
one wonders about the rationality of urban planning. For example, Wentworth (1997) concludes,
from a review of the proposal to extend the light rail system in Sydney between Central Railway and
the Circular Quay that a redesigned bus system would provide a better immediate result at a greatly
reduced cost. He asks:
42
…perhaps the investors themselves may have been taken for a ride by professional promoters… Or is
it just an innocent mistake? The only thing clear is that there is something fishy about the whole
affair.” (Ref 23; p.5)
4.1.3 BRT is proven technology.
The reader is referred to Addendum I for an excellent overview of bus rapid transit or
BRT, and how it would fit in the Seattle context. The US General Accounting Office
(GAO) has published a report, which describes BRT and compares it with light rail. (Ref
9) The Transportation Research Board has a short brochure entitled: BRT: Why more
communities are choosing Bus Rapid Transit. (Ref 11). The Journal of Public
Transportation devoted an entire issue to BRT. (Ref 14) A BRT versus light rail case
study was conducted in Los Angles and reported in the Transportation Research Record.
(Ref 12)
Here in Puget Sound King County Metro developed a “Six year Transit Development
Plan” for BRT. (Ref 48) Unfortunately there must have been political pressure on Metro
to avoid studying any BRT routes that would have competed with Sound Transit’s light
rail. Thus Metro’s report says: “Would the BRT service compete for riders with current
or planned future Sound Transit Express bus, commuter rail or light rail…? Corridors
that do or would compete were eliminated.” (Ref 48: p.4)
As a result Metro only studied a number of innocuous locations for BRT, mostly on
arterials. Still the report is useful because it shows there is local familiarity with BRT,
and because it describes what can be done to make arterials BRT friendly. This is
important because BRT not only provides a high level of service along the high volume
parts of the route where light rail might be used, but it also provides a higher level of
service than traditional local buses after it branches onto arterials and into neighborhoods.
Put another way, a mature light rail network might have 100 route miles, whereas a
mature BRT network might have twice or three times that.
4.1.4 The key advantage of BRT is that much of the “guideway” already exists and
is already paid for.
As of January 2001 Puget Sound had 191 HOV lane miles open to traffic. The total “core
HOV” system will have 297 lane miles when complete.
The central Puget Sound region has one of the most extensive HOV systems in the world.” (Ref 47:
p.6)
If the region implements light rail almost every inch of “guideway” (tunnels, elevated
structures, rails and surface right-of-way) must be constructed from scratch, or taken
from its current users (e.g.: the DSTT, the center lanes on I-90) Even in the DSTT the
existing rails must be replaced because they were designed incorrectly.
BRT can use existing and already planned HOV lanes. Below is a map of those lanes.
43
44
4.2 The BRT alternative
Plan B or the BRT alternative to Link light rail would build on the existing express bus
system. It might employ a large number on the new hybrid buses. It would operate on
existing HOV lanes and be similar in many respects to the TSM alternative studied in the
1993 FEIS.
Plan B would probably start out with a fairly extensive BRT route structure, first because
the express bus network is already fairly extensive, and second because for what even the
Initial Segment of light rail is costing, a rather large BRT system could be built right
away.
BRT would continue to use the DSTT as express buses do today. Chapter 5 describes a
number of alternatives for increasing bus capacity through downtown Seattle if that
becomes necessary. The most extreme of those is to construct a second bus tunnel
parallel to the first. This would probably not be needed for years.
South Lake Union busway?-- The south Lake Union area is poised for major
redevelopment as a important employment/residential center, but Link light rail ignores it
completely. But it could be well served by BRT, and vice versa. One intriguing
possibility would be to construct an elevated bus way for BRT from the Mercer Street
express lane exit, across the south Lake Union area, and tying into SR 99. BRT buses
could then proceed south using the Alaska Way viaduct or its replacement.
Such a busway, including a south Lake Union station, could probably be physically
integrated into the new buildings. As such it would not be a visual blight, and would
definitely make the area more accessible.
In addition, some of the BRT buses crossing the busway might continue west to serve
two other important destinations, which Link ignores; namely, Seattle Center and
Belltown.
Clearly if the region decided to scrap Sound Transit’s light rail strategy, there would be
plenty of money for such a busway. In addition, Link money, being transit money, could
probably be used to fund HOV lanes in the tunnel proposed to replace the Alaska Way
viaduct. That infusion of money might help accelerate the project.
Network Size and shape-- The high speed portion of a fully mature BRT network
would probably use most of the region’s HOV lanes and thus have the general shape
shown in Figure 4.1.4. However, there could be other major extensions such as the
abovementioned south Lake Union busway, BRT on SR 99 in south Seattle, and BRT
over into West Seattle. In addition, the BRT system would extend out along various
arterials. The idea would be to pick up riders in the neighborhoods, then use the HOV
lanes to make fast trips to various employment centers.
45
Because BRT has a different cost structure than light rail the BRT network might start
out being very extensive, whereas light rail would start out with just the 14-mile Initial
Segment and grow slowly thereafter. What might change over time with BRT, as more
money is spent, is not network size but rather network quality and service frequency. In,
other words over time the quality (or speed) of the BRT network would be increased by
adding more HOV to HOV interconnections, and more flyovers that give HOV’s direct
access to HOV lanes. Network quality could be increased on arterials with a variety of
relatively mundane signal and lane improvements that would increase bus speeds. The
frequency of service is just a matter of buying more buses and operating them more
frequently.
What this means in terms of the alternatives analysis called for in this report is as follows.
Per Part 8 of this report, it is proposed to create a number of light rail alternatives that
differ either in cost or in the degree to which they reduce delay, then design a BRT
alterative with comparable cost or benefit to each of the rail alternatives. If the pairs of
alternatives were designed to be equal in cost we would be looking for how they differed
in benefit. If they were designed to have equal benefit we would want to know how their
costs differed.
The rail alternatives are likely to differ in the length of the light rail line, so there would
be one alternative having 14-miles of light rail, one with 21-miles, and so forth up a fully
mature scenario having 100+ miles of light rail. However, we would not be creating a
BRT alternative with 14-miles of BRT to compare against 14-miles of light rail. Instead,
all the BRT alternatives would probably have a very extensive (100+ route miles) BRT
network since the “guideways” already exist. The difference between a low cost or low
benefit BRT alternative and a high cost or high benefit BRT alternative would be in
network quality and service quality. For example, a BRT alternative designed to cost the
same as Link IS ($1.5 billion 02$) might have 150 route miles of BRT but no additional
HOV to HOV connectors and relatively infrequent bus service. In contrast, a BRT
alternative designed to cost the same as a 60-mile version of Link (roughly $6 billion)
might have the same 150 route miles, but include HOV to HOV connectors everywhere,
lots of flyover direct access ramps, and very frequent service.
In other words these alternatives analyses would not be comparing a 14-mile light rail
alternative against a 14-mile BRT alternative. Instead they might compare a $1.5 billion
dollar light rail alternative against a $1.5 billion BRT alternative. The $1.5 BRT
alternative might have 100 route miles of BRT, but perhaps fewer HOV to HOV
connectors and perhaps a lower number of buses and thus less frequent service than
might be possible.
4.3 The “buses will be stuck in traffic” issue
Critics of bus solutions claim BRT will be stuck in traffic and therefore not able to
compete with light rail able to move freely on its own dedicated guideway. Is that
criticism valid?
46
The first thing to note is that transit networks come in two parts, the line haul subnetwork,
and the feeder sub-network. The line haul sub-network for light rail is of course
the light rail system itself, but light rail must be fed by bus routes that operate on arterial
streets. The line haul sub-network for BRT is where buses operate on freeway HOV
lanes. The feeder sub-network for BRT is when the buses depart HOV lanes to operate
on arterials. Actions that can be taken to speed BRT on arterials could be taken to speed
buses feeding light rail. Clearly there is no intrinsic difference between bus and rail
feeder networks, both are equally susceptible to being stuck in traffic.
As to the line haul sub-network, it is true that HOV lanes need to be kept moving at a
good speed to make BRT competitive with light rail. That can be done by policy. It may
mean changing the minimum carpool occupancy from 2 to 3 persons sometime in the
future. If such a change is needed it would not just be needed for BRT, but also to keep
car and van pooling attractive.
Current DOT policy is to manage HOV lanes to maintain an average speed no less that
45 mph during the peak commuting hours. Many of the BRT buses would be express
buses that would make few if any stops once entering the HOV lanes. Thus their average
speed would be about the same speed as the HOV lane is moving.
On the other hand light rail must stop at all stations. Thus while Link has a maximum
speed of over 50 MPH, it will only average 26 MPH. (48 minutes from NE 45th to South
200th per Ref 38: p.16)
In short, the feeder buses for light rail, and the feeder buses for BRT are equally
susceptible to being stuck in traffic. The line haul portion of the BRT network will not be
stuck in traffic because the HOV lanes can be managed to flow at a reasonable speed.
But all this is not something people should be arguing about. It is an argument about
engineering details at the 5000-foot level akin to arguing whether steel wheels are better
than rubber tires. The correct thing to do is go up to 40,000 feet and look at what the
models predict in terms of ridership and effect on travel delay. The models take into
consideration the speed with which HOV lanes operate, bus speeds on arterials, the delay
while loading and unloading at stops or stations, maximum vehicle speeds, acceleration
rates, any inherent preference people may have for rail, and a whole raft of other factors.
These models are calibrated with data from actual systems in other cities. When the
model says X people will ride light rail, or Y people will ride BRT, things like the degree
to which buses are stuck in traffic will have already been included in the analysis. As
will be explained in Part 5, Sound Transit’s predecessor the RTA, did model an extensive
BRT like bus network and found it would attract 518,000 daily riders in 2020. (Ref 42: p.
3-8) This would not have been possible if “buses were stuck in traffic”. In contrast,
Central Link is estimated to attract 133,000 daily riders in 2020. (Ref 2: p.7-11)
In conclusion, the claim that “buses will be stuck in traffic” is an irrelevant red herring.
Run the ridership models, and do it honestly. If any buses are stuck in traffic that will
47
have been cranked into the models, and will have affected the predicted ridership. If the
ridership looks good, one can ignore all the low level issues.
4.4 Plans C, D, and E
If the objective is to increase transit ridership the only alternatives are probably light rail,
heavy rail, monorail, and bus. But if the objective is to reduce travel delay there are
many other alternatives to consider, such as car and van-pooling, transportation demand
management (TDM), and road improvements. Should these be plans C, D, E and so
forth?
There are two ways these other alternatives can be considered. One is to construct hybrid
alternatives, which combine BRT and one or more of these other alternatives. For
instance, construct a $2 billion alternative where $1.5 billion goes to BRT and $500
million goes for van pools. The other alternative is to construct pure alternatives where
all the money in a given alternative goes for one type of remedy, such as van pools.
To make a long story short, the thing the region should really do is apply “least cost
planning” or LCP. The essence of LCP is to take an objective like reducing travel delay
and determine the least cost way to achieve it. Probably that least cost way would be a
market basket of different remedies; for instance an optimized mixture of BRT, car and
van pools, TDM, and selected road improvements. However, in theory, it could be 100%
van pools, or 100% roads. (See description of LCP and end of this report)
If the region decides to force ST into conducting a proper alternatives analysis, one of the
scenarios or alternatives that must be evaluated is Sound Transit’s vision for a fully
mature 100+ mile light rail system. Such a system would cost roughly $15 billion. If it
were decided to do an “equal cost” alternatives analysis that would mean designing a
BRT alternative also costing $15 billion. It is hard to imagine how $15 billion could be
spent on BRT no matter how large the network or how lavish the service. Thus if an
equal cost approach is used, the only practical thing to do would be to construct only
hybrid scenarios where BRT is combined with car and van pools, and other remedies so
as to get up to the cost of the rail alternatives.
In sum, the alternative to light rail or “Plan B”, may be pure BRT, but it is more likely an
optimized combination of BRT and other remedies. The best approach is to use Least
Cost Planning to find the optimum combination of remedies.
4.4.1 Car and Van Pooling
This report focuses on the bus versus rail issue because that has been the focus of most
planning activities to date. However, BRT is not the only alternative. Car/van pooling is
probably even more cost effective than BRT and should be among the alternatives
considered most carefully as the region charts its transportation strategy. This is true
48
because car and vanpools are already far more effective in getting people out of single
occupancy vehicles than is mass transit, as the following chart makes clear.
To put this in perspective consider the following. Central Link is only expected to reduce
PM peak period VMT by 3.7% in 2020. (Ref 2: p. 3-4) This is a reduction of 63,000
VMT during the peak period. The average work trip is a little over 10 miles long so
Central Link would be taking 6300 cars out of the morning commute, and the same 6300
out of the evening commute. One could say Link removes 6300 cars from peak period
traffic.
However, in 1998 there were already 459,690 work trips being made by car/van pool,
presumably half in the morning and half in the afternoon (Ref 17: p.xvii) This means
there were about 230,000 persons commuting to/from work by car/van pool. If we
assume every car/van pool carried two persons this would mean that there were 115,000
carpools in operation, each of which removes one car from the evening commute, and
one from the morning commute. In short, car/van pooling is now taking 115,000 cars out
of peak period traffic. This is 18 times more than Central Link is expected to remove.
Expressed differently every two people deciding to car pool takes at least one vehicle out
of peak period traffic. (Actually it would be more than one vehicle since some car pools
and most bus pools have over two persons. But two is a conservative estimate.) Link
takes 6300 vehicles out, so it would take a maximum of 12,600 additional people
deciding to carpool to have the same impact as Central Link. 12,600 is 5.5 % of the
230,000 people now car/van pooling.
0%
20%
40%
60%
80%
100%
All
Trips
Work
Trips
3
35
62
Transit
Car/van pool
Single
occupancy
vehicle
7
21
72
Mode Share for Trips in Puget Sound Region in 1998
Source: Ref 17:
p.xvii)
49
In short, increasing the number of people who car or van pool to work by just 5.5%
would remove as many cars from the commute as building 21-miles of light rail at a cost
of $3.6 billion (YOE$).
A recent article in the Bellevue Reporter indicates just how cost-effective car and van
pool programs can be, and at the same time indicates the paltry sums being devoted to
encourage them.
“The Commuter Challenge offered by the Regional Smart Commute Program, pays those who
typically drive to work alone to instead ride the bus, join a car pool, bicycle or walk to the office or
work from home.
…The program pays $3 for every day a participant leaves his single occupancy vehicle (SOV) at
home.
…Flexible and profitable. The program pays participants up to $192. It is funded by a $100,000
WSDOT grant.” (Amy Roe, Bellevue Reporter, Nov. 24, 2004)
To put this in perspective, $3 per day times 200 workdays per year comes to $600 per
year for each car this program removes from peak period traffic. In contrast Central Link
will cost $43,000 per year for each car it removes from the road. (See Part 6 of this
report for details) The $100,000 this program gets is just 1% of what Sound Transit
spends on public relations alone each year.
50
Blank page
51
Part 5: The Capacity Issue
5.1 Introduction
5.1.1 Overview: ST and the RTP used capacity as a major criterion in their selection of
rail as opposed to bus technology for the region’s HCT network. However RTP’s
analysis was incomplete and biased. Their reasoning was not valid at the time and is
even less valid today. This is partly because we are now dealing with light, not heavy or
rapid rail. Based on what’ s known today, bus alternatives would not only have adequate
capacity but may in fact be superior in meeting the region’s long term capacity needs.
In one very serious case of bias RTP officials knew capacity was a key consideration and
potential weakness of the TSM (bus) alternative, nevertheless they: 1) chose to assume a
worst case assumption about bus capacity in the DSTT rather than make a serious effort
to determine an accurate value, and 2) failed to fix any remaining capacity shortfall using
any of a number of previously identified remedies. In short, RTP deliberately created a
hobbled bus alternative to compare against rail.
The consequences of putting forward this designed-to-fail bus alternative were pervasive
and long-lived. It led to a lower ridership forecast for the bus alternative, less attractive
economics, and a long series of negative assessments claiming buses lacked adequate
capacity to meet the region’s needs. It was therefore decisive in essentially ruling out
buses as the core technology for the region’s HCT network back in 1993. There has been
no serious reconsideration of the issue since. Thus the current light rail strategy
continues to rest on the flawed and misleading way capacity was treated in the 1993
FEIS.
The essence is as follows. The maximum load ST claims a rail system would need to
handle by 2020 is 15,000 persons per hour through downtown Seattle. About 20% of this
would be used by persons traveling to nearby Capitol Hill. Because a freeway BRT
system would bypass Capitol Hill it would need a capacity of only 12,000 persons per
hour through downtown to accommodate the same overall transit demand. It would take
about 110 buses per hour to provide this capacity. There have been six separate studies
of bus capacity in the downtown Seattle Transit Tunnel (DSTT). Although the results
ranged from 125 to 192 buses per hour, Sound Transit’s predecessor the RTP assumped
that the tunnel could handle only 100 buses per hour. Even if that were correct and the
tunnel could not handle the entire load, there are several other ways to increase bus
capacity through downtown. When all this is taken into consideration it appears ST’s
assertions that an all-bus alternative would not have enough capacity are unfounded. To
the contrary, buses would have considerably more capacity to meet the regions long term
needs than would Link light rail.
Light rail has other capacity-related deficiencies: There is a fixed upper limit on Links
capacity whereas bus capacity is relatively open ended. Placing light rail on the I-90
bridge would reduce the people moving capacity of the bridge relative to having the
center lanes shared by buses, carpools and vanpools. There is an equity issue in that
Link’s design shortchanges the region’s south-end in terms of transit capacity. There is
52
some risk that Link’s south line may not have enough capacity. The ability of the DSTT
to move people in and out of downtown Seattle would be reduced if rail is not extended
across Lake Washington.
Because the 1993 FEIS mishandled the capacity issue in a biased manner, dealt with
heavy not light rail, and failed to address the full range of capacity issues, a new
alternatives analysis is needed. It needs to compare not only the currently planned 24-
mile Central Link against bus alternatives, but also the 100+ mile version of Link
envisioned by ST.
5.1.2 Visualization of capacity issue
Figure 5.1.2 helps visualize key aspects of the capacity issue. The light rail line would
come south from Northgate with a station in the University District and one on Capital
Hill before entering the Seattle central business district (CBD) and going into the
Downtown Seattle Transit Tunnel (DSTT). There are other stations between the U
District and Northgate but they aren’t shown since they aren’t relevant to this discussion.
The Capitol Hill station is shown as a jog because it’s east of I-5.
Max Load
Point
Max Load
Point
Northgate
U Dist
Capital Hill
Surface
streets
2nd Bus
tunnel
Rail BRT
Alaska Way
viaduct
Seattle
CBD
DSTT
Figure 5.1.2
North
corridor
53
Ridership forecasts show that the maximum passenger load point on the entire rail system
would be just north of the DSTT. Maximum load is measured as passengers per hour
(pph) in the peak direction. Rail would need enough capacity to carry this maximum
load.
The main bus or BRT line serving the north corridor would share an HOV lane on I-5,
and would not serve Capitol Hill. There are four different ways to get buses through
downtown. Today some go through the DSTT while others use north-south streets such
as 3td Avenue. The maximum capacity of the DSTT remains uncertain at this time as
will be explained later, but estimates range from 100 to 192 buses per hour. The capacity
of surface streets is also somewhat uncertain, although RTP has identified ways to
increase the number above today’s levels. A third potential way to increase bus capacity
through downtown would be to divert some buses off I-5 at Mercer, cross the south Lake
Union area on a new busway and tie into SR99. Once on SR 99 the buses could proceed
south on whatever facility replaces the existing Alaska Way viaduct, especially if that
replacement includes HOV lanes. A fourth possibility was identified by the RTP back in
1993, namely to build a second bus tunnel. In short, if the DSTT in combination with the
existing number of buses on surface streets does not have enough capacity, there are three
ways to add additional bus capacity through downtown.
A bus alternative would also need enough capacity at the max load point to carry its
projected ridership. The maximum load point on freeway BRT would be just north of
Mercer if some of the southbound bus stream were diverted at that point to some future
south Lake Union busway, otherwise it would be just north of the DSTT. The remainder
of Section 5 assumes the latter since none of the projected ridership scenarios is high
enough to demand a south Lake Union busway. It’s just an option.
The bulk of section 5 deals with the issue of whether rail has an advantage over bus in
terms of capacity at the peak load point, or vice versa. This only becomes an issue with a
fully mature or fully built-out rail or bus system. Shorter systems, such as the 24-mile
Central Link or a BRT system with equivalent ridership, would not have enough
passenger volume to make capacity an issue. A fully built out light rail system would
have 100+ route miles, as did the Rail/TSM alternative in the 1993 FEIS and as ST
envisions Link growing to. A fully built out bus system as I define it here is a BRT
system having the same geographic coverage and ridership as the fully built out light rail
system. In other words it would include 100+ route miles of freeway BRT.
Thus Section 5 will focus on these key questions:
1) Would a fully built out light rail system have enough capacity to handle forecast
demand at the peak load point in 2020?
2) Would a fully built out light rail system have enough capacity to handle the
region’s long term needs for growth in demand well beyond 2020? In other words, is
light rail “strategic”?
54
3) Would a fully built out bus or BRT system have enough capacity to handle the
forecast demand at the peak load point in 2020?
4) Would a fully built out bus or BRT system have enough capacity to handle growth
in demand well beyond 2020? In other words, is BRT “strategic”?
NOTE: The author uses the following terms to describe bus alternatives: “all-bus”, bus
rapid transit or“BRT”, and “TSM”. They all mean pretty much the same thing: namely
a mix of express buses operating on freeway HOV lanes and on arterials where they are
given bus only lanes, signal priority and so forth. The bus alternative would also include
Park and Ride lots. In other words the bus alternative is an integrated system built
around buses. As was the case with the 1993 FEIS’s TSM alternative, the bus alternative
would be an extension to or overlay on the current local bus system. The BRT or express
bus-on-HOV portion of a bus alternative could be thought of as light rail with rubber
tires using an existing right of way rather than carving a new one. Whether BRT or light
rail, the networks would follow the region’s main freeway corridors. The largest light
rail network might have somewhat over100 route miles. A BRT network might extend
over the region’s 200+ route miles of HOV lanes.
5.1.3 Capacity has been used as a discriminator against the all-bus alternatives.
The probability that an alleged capacity limitation was used as a major reason for
dismissing the bus alternatives is illustrated by the large number of times it was
mentioned in various RTP and ST documents. The following quotes are from the 1993
FEIS:
“A rapid rail line operating in exclusive right of way has the theoretical capacity to carry over
22,000 persons per hour in each direction…These numbers represent the rail capacity of the
downtown Seattle transit tunnel…. In comparison the bus passenger capacity of the downtown
tunnel is about 13, 400 persons per hour...” (Ref 1, p. xviii)
“TSM Alternative bus traffic levels would meet or exceed the capacity of the downtown Seattle
transit tunnel and the street network in downtown Seattle at rush hour, resulting in a constraint to
meeting the demand created by the TSM Alternative.” (Ref 1, p. xxxiv) The same is said about
the TSM/Transitway Alternative on page xxxv.
“Bus volumes to meet ridership needs would exceed street capacity in downtown Seattle…” (Ref
1, p. xliii)
“Neither the TSM nor the Transitway/TSM Alternative could accommodate the transit demand
associated with implementing the Visions 2020 land use concept and consequently would not fully
support its land use goals.” (Ref 1, p. 4-5)
“The Rail/TSM alternative would provide enough capacity to meet the high end of projected
transit demand.” (Ref 1, p. xxxv)
The October 1992 “System Plan Technical Report” said:
55
“In the long term, neither of these (TSM or Transitway/TSM) alternatives offers the mobility or
capacity necessary to support growth management policies, control sprawl or provide the mobility
necessary for sustained economic viability.” (Ref 45: p.89)
The Nov 1993 “Central Corridor Project Justification Report” said:
“The TSM and Transitway/TSM alternatives were found to provide only a temporary solution to
the transportation needs of the region. In the long term neither of these build alternatives offered
the mobility or capacity necessary to support the adopted growth management policies, control
sprawl or prove the mobility necessary for sustained economic vitality.” (Ref 19: p.23)
“Finally, only the Rail/TSM Alternative had sufficient capacity to accommodate transit demand in
key areas such as the Seattle CBD…” (Ref 19: p. 25)
The Nov. 1999 FEIS for Central Link reiterated the theme as follows:
“The benefits that led to the selection of the rail alternative were its capacity to meet the high end
of projected transit demand.” (Ref 2: p. P-1)
“Criteria such as capacity, operating speeds, …were considered” (Ref 2: p. S-19)
“1.1.2 How were the alternatives identified and narrowed for the EIS? …Evaluation criteria
included capacity, …” (Ref 2: p 7-3)
The PSRC’s Summary of Prior All-Bus and Integrated Rail/Bus Alternatives Analyses
report dated 2001 quoted the Expert Review Panel as saying
:
“…only rail provides the capacity, speed, and reliability to meet growing demand.” (Ref 28: p.3)
The Summary goes on to say:
“Probably one of the most significant factors that tipped the scales toward the rail/bus alternative
was the long term effectiveness and relative efficiencies that rail services would provide over the
20 year planning period. In addition to faster speeds and greater reliability in direct service to the
high demand transit markets, the rail bus alternative was found to provide a substantially higher
level of long term future capacity than either of the all bus alternatives.”
Finally, the recent “Summary of Prior All-Bus and Integrated Rail/Bus Alternatives
Analyses” dated April 5, 2001 and appended to the June 14, 2001 “Central Link Board
Workbook” says:
“As mentioned previously, the bus capacity constraint in downtown Seattle was one of the major
reasons for selection light rail in the central I-5 corridor” (Ref 36, Appendix Q, p.7)
In sum, it is clear that an alleged capacity shortfall in the bus alternatives was a key factor
in the selection of rail. The sections below examine whether this alleged shortfall was in
fact real.
56
5.2 How the capacity issue was mishandled in the 1993 FEIS
5.2.1 Overview of what RTP did and didn’t do in the 1993 FEIS
The RTP ran ridership forecasting models to estimate year 2020 ridership for the rail
alternative. The models predicted that rail ridership during the peak hour at the peak load
point just north of the DSTT would be 15,000 persons per hour in the peak direction. (Ref
1: p. 2-58) RTP then assumed the rapid rail system could operate at 90-second headways
giving it an ultimate capacity of 22,000 persons per hour. (Ref 1: p.2-58) From this RTP
concluded that rail had enough capacity to meet 2020 demand, plus plenty of excess
capacity for growth.
RTP also ran the models to predict ridership in the DSTT for the TSM and
Transitway/TSM alternatives. The result was that the TSM alternative attracted 518,000
daily riders. (92% of what the rail alternative attracted) In parallel RTP reviewed six
previous studies of tunnel capacity. Those studies had each come to different conclusions,
ranging from 125 to 194 buses per hour. However, RTP decided to assume a maximum
capacity of 100 buses per hour. One hundred buses per hour would not handle the
518,000 riders so RTP reduced the predicted TSM ridership to what they said 100 buses
could handle, namely 473,000 daily riders. (Ref 42: p.5-2) This was an 8.7% reduction.
This “constrained” ridership was the one reported in the 1993 FEIS and became the basis
for calculating all the ridership related benefits and cost-effectiveness measure for the bus
alternative. Obviously the bus alternative would have fared much better in its fly-off
against rail if the RTP had found some way for the bus alternative to achieve its
unconstrained ridership, either by double-checking its assumption about bus capacity in
the DSTT, or if this failed, finding some workaround.
However, the RTP did not attempt to narrow the range of uncertainty about the true
capacity of the tunnel, but instead assumed worst case. Neither did the RTP attempt to
fix the alleged DSTT capacity bottleneck with any of the remedies they had identified
elsewhere, even though the $6 billion difference between the bus (TSM) alternative and
Scanned image of
Table in Ref 42
57
the rail alternative provided plenty of money for such remedies. Thus the RTP
deliberately entered a capacity deficient bus alternative into competition with the rail
alternative. Naturally, the bus alternative didn’t do as well as it might have. Thus in
higher level summaries prepared for policy makers the RTP repeatedly asserted that the
bus alternatives were deficient in terms of capacity, and this was given as one of the
major reasons the bus alternatives were dropped in favor of rail.
In a nutshell, an alleged capacity bottleneck in a 1.5-mile segment of a 100+ mile BRT
network became the tail that wagged the dog.
Specific ways in which RTP mishandled the capacity issue are further described below.
5.2.2 The 1993 FEIS and subsequent studies did not define or quantify the capacity
benchmark the alternatives were being judged against.
The following benchmarks are mentioned in the quotes in Section 5.1: “transit demand
associated with implementing the Visions 2020 land use concept”, “ high end of
projected transit demand”, “ capacity necessary to support the adopted growth
management policies, control sprawl or prove the mobility necessary for sustained
economic vitality”, “ capacity… to meet growing demand”. Unless these benchmarks are
quantified how is it possible to rule that one system meets them and another does not?
5.2.3 The 1993 FEIS assumed a lower bus capacity for the DSTT than all prior
studies had estimated as being possible.
Prior to the 1993 FEIS, six paper studies had attempted to estimate the bus capacity of the
DSTT. (Ref 34) The results ranged from a low of 125 buses (one way) to a high of 192,
and are plotted in Figure 5.2.3. (Other values in the Figure will be described later.)
Study #6, an “internal Metro technical memorandum” cited on page 12 of Reference 34,
says the following:
“ ‘tunnel capacity for outbound revenue service leaving downtown is approximately 100 buses per
hour per direction…’ as one of its major conclusions. The phrase “leaving downtown” is
important as approximately 45 buses per hour per direction also operate in the tunnel for
deadheading (10 per hour), inbound terminating (35 per hour) or unscheduled trips. Thus the
capacity of the tunnel was taken to be 145 buses per hour per direction. The DSTT’s capacity was
based on the DSTT’s stated peak hourly capacity of 145 buses per hour per direction and the
opinion of Metro’s operating staff.” (Ref 34: p.12)
58
In other words this report indicated that the tunnel could handle 145 buses per hour if
managed for maximum useful capacity but only 100 buses per hour if managed
inefficiently.
Nevertheless RTP chose to assume a capacity of only 100 one-way (or 200 two-way) for
the 1993 FEIS. This is documented in a Technical Memorandum dated May 1992. (Ref
26: p.3) (It is also confirmed by applying the same ratio of peak load point traffic to total
daily traffic used for rail to the estimate of constrained daily traffic for the TSM
alternative. For rail the ratio was 15,000/124,000. For bus 11,100/92,000 gives the same
ratio. Source Ref: 1 p. 2-58, 3-108, and 3-115. A peak load of 11,100 would require
about 102 buses per hour, assuming each carried 109-passengers.
In addition, the various paper studies clearly evidenced a great range of uncertainty.
Given the importance of this parameter the RTP should have conducted real world
experiments to find the correct value.
The downstream consequences of this low-ball assumption—on ridership, on all ridership
related metrics, and on qualitative statements made denigrating the bus alternatives for
15,800
15,000
13,400
11,000
7600
5700
122
55
52
50
70+
100
125
144
161
192
0
133
145
150
21,000
16,400
5500
Buses
per hr
Persons
Per hr*
Study 4
Study 3
Studies 1 & 6
Study 2
Study 5
Value used in 1993 FEIS
Figure 5.2.3
Capacity as
estimated in
prior studies
* Assuming 109 persons/bus
59
not having enough capacity-- were significant. If capacity was such a critical issue why
didn’t RTP assume the tunnel would be managed efficiently? This failure proves RTP’s
bias against the bus alternatives.
The consequences of having picked 100 rather than 145 can hardly be overstated. If RTP
had used 145:
• the bus alternatives could have carried the forecast demand, and RTP could not
have dismissed the bus alternatives because they lacked sufficient capacity
• the bus alternative would have had the capacity (15,800) to carry the same 2020
passenger load predicted for rail in the 1993 FEIS, namely 15,000. The 15,800
comes from 145 buses per hour times 109 persons per bus.
• the bus alternatives with a capacity of 15,800 would have almost as much
capacity as the current light rail (16,400) thus largely negating any current claim
that rail has a capacity advantage over buses.
• the TSM ridership estimate would have risen from the constrained value of
473,000 daily riders reported throughout the 1993 FEIS to the unconstrained value
of 518,000 reported in the technical documents. (Ref 1: p. xxxiii) (Ref 42: p.3-8 and
5-2)
• all TSM benefits or metrics for cost-effectiveness that were a function of ridership
would have risen accordingly
In the sections below the author assumes that buses would be operated efficiently in the
tunnel and, for simplicity, assumes the tunnel could handle something in the mid range of
what prior studies indicated, namely about 145 buses per hour each way. If these were 60
foot articulated buses with a capacity of 109 persons, 145 buses per hour translates into a
bus capacity in the DSTT of 15,800 persons per hour each way. In short, from here on
the author will assume the true bus capacity of the DSTT is 15,800 pph in each direction.
However, if the DSTT really could not handle 145 buses per hour, there are other ways to
increase bus capacity through downtown Seattle. These will be addressed presently.
5.2.4 The 1993 FEIS appears to have assumed an unreasonably low headway for
rail thus artificially inflating its capacity.
The 1993 FEIS said:
“A rapid rail line with 4-car trains operating in exclusive right-of-way on 90 second headways has
the capacity to carry over 22,00 persons per hour past a single point in each direction” and “These
numbers represent the rail capacity of the downtown Seattle transit tunnel” (Ref 1: p.1-10)
Later this was cited as a point in favor of rail:
60
“The forecast peak demand in 2020 is about 15,000 passengers per hour. Thus the (rapid rail)
system design incorporates an “expansion factor” of about 50% to accommodate future growth in
ridership” (Ref 1: p. 2-58)
However, 90-second headways have apparently never been achieved on US rail systems
according to rail expert Tom Rubin. It seems especially problematic since the east and
south lines would need to merge with perfect timing.
5.2.5 The RTP failed to fix the alleged bus capacity problem by applying already
identified remedies.
If the bus tunnel can’t carry 145 buses per hour there are a variety of fixes the RTP could
have applied. They involve finding parallel paths through downtown. One set of
solutions places more buses on surface streets, another routes buses that don’t need to
stop in the DSTT through the CBD on HOV lanes that could be constructed on I-5
through the downtown. These are briefly described in a 1992 report (Ref 26) and were
therefore known at the time of the 1993 FEIS. The author has also seen an internal staff
memo dated May 21, 1992 addressed to Bob White (the RTP director at that time), which
mentions some of these other remedies and suggests they could have, but weren’t
applied. Excerpts from this memo follow:
“This note reviews “Analysis of Capacity Constraints in Selected Activity Centers”, Draft
Technical Memorandum, May 13, 1992. …”
“The subject memorandum does not address appropriate margins of bus capacity in the Seattle
CBD for the system planning issues before the RTP and the JRPC.”
“The peak hour bus capacity of the Seattle CBD is critical to the system planning effort.”
“They do not answer the question of how much surface bus capacity would be increased through
the 1993 TSM alternative of transit only use of Third Avenue or contra flow lanes on Second and
Fourth Avenues.”
“Imagine the extremely steep slope of a line on the Ben Porter graph representing the cost
effectiveness of capital projects (contra flow or exclusive use of Third Avenue at only 40 to 50
million 1993 dollars) which removed the capacity constraint from the TSM ridership.”
Underlining by Harkness.
This last quote indicates that staff had determined that inclusion of the relatively
inexpensive contra flow or exclusive use of Third Avenue remedies would have at least
allowed the bus alternative to reach its unconstrained ridership forecast. Consider also
the following quotes from another staff memo written during that time:
“A few of the assumptions that have particularly significant policy and performance implications
are: …What the table (Table 4 in the 1993 FEIS)rh does not show is the unconstrained TSM daily
ridership which was forecast to be 518,081 or 92% of the Rail alternative. … The (ridership)rh
constraint is based on a limit on the number of buses the model indicated could serve the Seattle
CBD. … Perhaps even more crucial is that low cost improvement in Seattle, such as contra flow
transit lanes on Second and Fourth avenues were not included in the analysis (though analysis by
Metro in the 1980’s had shown the validity of the concept). As a result even though the TSM
alternative includes hundreds of millions of dollars for roadway improvement it does not include
61
the small improvements that would have done the most to alleviate the assumed (capacity)rh
problem.” (from transit agency staff memo dated 1993 with underlining and parens initialed “rh”
inserted by Harkness)
“It is particularly unfortunate that the TSM alternative, with all its flexibility (and so many
assumptions open to question) is being set up as a straw man by which to measure the rapid rail
element.” (from same staff memo)
Bus travel through downtown Seattle could be improved by constructing an HOV lane
through the CBD. This possibility is described in one of the State DOT reports on HOV.
(Ref 33: p.16) There is also the possibility of diverting some of the buses to or from the
north-end at the Mercer St. exit and building a busway through the south Lake Union area
over to SR 99 and then south on the Alaska Way viaduct, which is planned for
reconstruction. Still another alternative is a second parallel bus tunnel, which METRO
estimated would cost about $600 million. (Ref 26: p.8)
“The cost of another tunnel and the access and egress to it necessary to successfully operate a two
tunnel system in the CBD have been conceptually explored. The cost of a Forth Avenue tunnel would
be about $600 million dollars.” (Ref 26: p. 8)
Besides increasing capacity, parallel paths have the added benefit of providing walking
access to transit over a wider area than the DSTT stations allow.
If none of the simpler remedies were found sufficient, in the worst case RTP could have
added $600 million to the TSM alternative for a second bus tunnel thus decisively
eliminating the bus capacity problem and all negative consequences thereof. This would
have reduced the $6 Billion difference between the TSM and rail alternatives but still left
the TSM alternative billions less costly than the rail alternative.
Section 6 will show how adding a second bus tunnel affects the cost effectiveness of the
bus alternative. It will be shown that although costs increase, ridership gains more than
compensate. In other words, adding a second tunnel, if in fact one is needed, improves
the cost effectiveness of the TSM alternative by allowing the TSM alternative to carry its
full predicted ridership.
In the Rail/TSM alternative between 16 and 20 miles of the 125-mile rail network were in
tunnels. (Ref 1: p.3-91,92) The extra investment in tunneling bought an increase in
capacity and performance for the rail alternative. It must have been a conscious decision
to include intensive tunneling in the rail alternative, and a conscious decision not to
include any in the bus alternative, not even 1.5 miles in the form of a second bus tunnel.
In short, RTP management knew there were various options to fix the capacity problem
in the TSM alternative, but failed to apply them. In this way they knowingly put forth a
hobbled or weakened bus alternative to compete against the rail alternative.
A related consideration is that “fixes” to bus capacity problems can be applied
incrementally, if and when future transit demand grows to the point they are needed.
Since the DSTT is currently operating at only half capacity this is a long way off. From a
62
cash flow or discounted return-on-investment (ROI) perspective this is an important
point, and should have been addressed in the FEIS.
The cost advantages of creating what might be called an “enhanced TSM” alternative --
that is fixing the bus capacity problem and then simply expanding TSM until rail
ridership was equaled-- should have been explored in the 1993 FEIS. Since it wasn’t,
the author does so in Section 6 by estimating how an “enhanced TSM alternative” would
compare with the Rail/TSM alternative in terms of cost-effectiveness.
5.2.6 Subsection Conclusions:
The capacity issue was skewed in the 1993 FEIS in a way designed to favor rail. On the
one hand the capacity attributed to rail was probably too high. On the other hand the
capacity attributed to the bus alternative was certainly too low. This was because the
RTP make what appears to be an unreasonably low, and certainly unverified, assumption
about how many buses could operate in the DSTT. Then, asserting that buses lacked
adequate capacity, RTP failed to fix the problem using any of a number of remedies they
had already identified. Thus the RTP pitted an overly optimistic rail alternative against a
deliberately hobbled bus alternative. This lead to the bus alternative being rated lower in
ridership and ridership related benefits than it should have been, and to false assertions
that buses didn’t have enough capacity to meet the regions needs. The still lingering end
effect has been to take bus or BRT off the table, as far as the core parts of the region’s
HCT network are concerned. Because the 1993 FEIS mishandled the capacity issue, it
did not then and does not now constitute a valid basis to discriminate against buses from
a capacity viewpoint.
5.3 From today’s perspective there is even less reason to dismiss bus alternatives for
an alleged capacity deficiency than there was in 1993.
What follows is a current view of the capacity issue, especially as it relates to capacity in
the DSTT. It takes into consideration the fact that we are now dealing with light rail,
rather than the rapid rail studied in the 1993 FEIS. Section 5.3 starts by establishing
some basics, then goes on to examine the implications of an apparent conflict between
different ST forecasts and statements of demand at the peak load point. Section 5.4 treats
some additional capacity issues that are less related to downtown Seattle and the DSTT.
5.3.1 Light rail has less capacity than the rapid rail studied in the 1993 FEIS.
The 1993 FEIS assumed rapid rail could operate on as short as 90-second headways thus
giving it a capacity of 22,000 persons per hour. (Ref 1: p. 2-58) The maximum capacity
now claimed for Link light rail is 16,400, which assumes 2-minute headways.
5.3.2 Link’s capacity can never be increased, except perhaps on the east line.
63
The capacity of Link light rail in the DSTT and along the entire north line is limited by
the headway, which can’t be reduced below 2-minutes, and by the length of the trains.
Train length is limited to 360 feet both by the length of the platforms in the DSTT and the
need to keep trains short enough so as to not block street intersections in Rainier valley.
Link’s capacity is therefore permanently limited to 16,400 pph on the north line and
about 5470 pph on the south line, assuming 2 and 6-minute headways respectively.
The east line is limited to 10,930 pph, unless a second rail tunnel is built downtown. The
second tunnel would not increase rail capacity on the north line since capacity along the
entire north line is limited to the 4-car trains that the then existing stations could handle,
and by 2-minute headways. Nor would it increase capacity to the south, where headways
are limited by at-grade operations in Rainer Valley and train lengths are limited by the
length of the then existing station platforms. However, trains from the eastside could run
more frequently if dead-ended into a second tunnel, and if there were no on-street
operations on the eastside.
In short, light rail has systemic capacity limits. In contrast bus capacity can be increased
to almost any value desired by building parallel paths in high demand areas.
5.3.3 BRT has far more ultimate capacity than Link could ever have.
Buses operating on an exclusive busway or occupying an entire HOV lane have truly
enormous capacity. Theory suggests 60, 000 or even 80,000 persons per hour. A BRT
system in Bogotá actually carries 25,000 persons per hour and a busway in New York
moves 44,000. The author has documented this in a separate paper. (Addendum F).
However, even if a single lane of buses could deliver say 60,000 pph into Seattle’s
downtown the existing streets and bus tunnel would not be able to handle it.
However, if a second bus tunnel or parallel busway were someday built through
downtown Seattle – thus doubling the capacity of the DSTT—a single freeway HOV lane
dedicated to BRT could bring in over 30,000 pph from the north, 15,000 from the south
and 15,000 from the east. Again, this assumes the DSTT could handle 145 buses per
hour in each direction thus giving it a directional capacity of 15,800 pph.
In contrast, light rail is forever limited to 16,400 from the north, 5500 from the south.
The east is limited 10,000 unless a second rail tunnel is built.
If, there is a desire to implement technology that will meet the region’s long term
capacity needs, BRT is the better choice.
Figures 5.3.3a, 5.3.3b and 5.3.3c portray the capacity story graphically. They compare
the capacity of Link versus BRT along the north, south and east corridors. Strictly
speaking they compare the capacity of Link with BRT operating on a dedicated freeway
lane. The author topped the BRT bar at 40,000 pph even though the theoretical limits are
much higher. 40,000 is more than enough to make the point. The actual volumes on
several BRT systems are shown to reinforce the point using real world examples.
64
Persons/hr one w ay NYC 44000 **
at peak load point
*Actual current use
** Based on current bus
volumes, all seated
*** Includes standees
Ultimate 5500 ***
Supplied 3300
Used 1700
Link L ight Rail
Plan for 2020:
South line
Pittsburgh 3700 *
Brisbane 9000 *
Ottawa 10000 *
Curitiba 11000 *
Bogata 25000 *
BRT
Ultimate 50000+ Capacity: BRT vs L ink south
Harkness 3/03
40,000
Figure 5.3.3b
Figure 5.3.3a
Pittsburgh 3700 *
Brisbane 9000 *
Ottawa 10000 *
Curitiba 11000 *
Bogata 25000 *
NYC 44000 **
Ultimate 16400 ***
Supplied 5480
Used 5400 ****
Link Light Rail BRT
Plan for 2020:
Persons/hr one way
at peak load point
North line
*Actual current use
** Based on current bus
volumes, all seated
*** Includes standees
Capacity: BRT vs Link north Ultimate 50000+
Harkness 3/03
**** Assumes LRT goes to Northgate
40,000
65
Again, note that BRT capacity in an HOV lane stretching north from downtown Seattle is
not realistically limited by what the HOV lane might carry down across the ship canal
(perhaps over 40,000) but rather what can be handled downtown by the DSTT, surface
streets, and possibly by a second bus tunnel and/or busway over to SR 99. Just the
addition of a second bus tunnel would raise what downtown could handle from 15,800
(assuming the DSTT could handle 145 buses per hour) to nearly 32,000. This is far more
than any forecast for future demand.
It is immediately apparent from the height of the bars that an HOV lane devoted entirely
to BRT would have far more capacity than Link could ever have.
In addition, the forecast ridership on Central Link (SeaTac to Northgate) is shown. For
instance the chart shows that Sound Transit’s Operating Plan for 2020 envisions
operating trains at a frequency that would supply a capacity of 5480 pph on the north
line, of which 5415 pph would actually be used if Sound Transit’s ridership prediction is
correct. It is obvious that the ridership predicted for Central Link is much lower than
what BRT routinely handles in other cities. Again, BRT has more than enough capacity.
DSTT
16,000 vs 50,000+
5500 vs 50,000+
2 min. headway
3 min, headway
6 min. headway
30 Trains/hr
20 Trains/hr
10 Trains/hr
10,500 vs 50,000+
Harkness 3/03
Ultimate capacity: Link LRT vs. BRT
LRT
BRT
Figure 5.3.3c
66
The main message from these charts is that BRT on a dedicated HOV lane would have
far more capacity than Link, and could easily handle the number of passengers predicted
to use Link in 2020.
A second important point is that it would not be necessary to devote an entire HOV lane
to BRT because BRT simply wouldn’t need that much capacity. Instead BRT buses
would share HOV lanes with car and van pools.
5.3.4 BRT does not need as much DSTT capacity as rail to handle the same regional
traffic load.
If bus and rail both had the same capacity in the DSTT the bus alternative would actually
have more effective capacity to serve the north corridor than rail. This is because about
20 % of light rails capacity going north from the DSTT is used up by persons getting off
in nearby Capital Hill (Ref 1: Fig. 3.21) (Ref 38: p.13) In contrast, a freeway based BRT
system would leave Capital Hill patrons on short local bus routes as they are today, thus
reserving all its capacity for the north corridor.
More specifically, light rail could move 16,400 persons per hour north out of the DSTT,
but only about 13,100 pph worth of that capacity would be available to serve patrons
going north of Capital Hill. In contrast freeway BRT would have its entire capacity
available to serve the north corridor. Thus a BRT system able to operate 120 buses per
hour in the DSTT –giving it a capacity of 13,100 pph—would have as much effective
capacity to serve the north corridor as light rail with its DSTT capacity of 16,400.
67
These facts are plotted in Figure 5.3.4. The rail side of the figure plots various values of
rail capacity, while the right side shows how much capacity BRT would need in order to
move the same number of people up and down the north corridor. For instance, a BRT
system able to operate 120 buses per hour at the peak load point would have the same
maximum capacity in the north corridor as Link light rail. Moreover, an enhanced TSM
(BRT) alternative requiring just 110 buses per hour in the DSTT would have been able to
handle the same regional transit ridership as the Rail/TSM alternative with its peak load
of 15,000.
This is an important point. If just 110 buses per hour were able to operate in the DSTT,
as opposed to the 100 assumed by RTP, then:
1) The TSM alternative would have been able to handle its full “unconstrained”
ridership, and negative comments like the following could not have been made in the
1993 FEIS:
“Neither the TSM nor the Transitway/TSM Alternative could accommodate the transit demand
associated with implementing the Visions 2020 land use concept and consequently would not fully
support its land use goals.” (Ref 1, p. 4-5)
0 0
Light
rail
BRT
Figure 5.3.4
Equivalent
capacities at peak
load point
31,500 (with 2nd Tunnel)
15,800 (145 buses/hr)
13,100 (120 buses/hr)
12,000 (110 buses/hr)
16,400
15,000
Maximum capacity of
Link light rail
68
2) An extended version of the TSM alternative could – even without increasing the
number of buses on surface streets or building a second bus tunnel—have handled the
same number of transit riders as the rail alternative. And negative comments like the
following could not have been made:
“The TSM and Transitway/TSM alternatives were found to provide only a temporary solution to the
transportation needs of the region. In the long term neither of these build alternatives offered the
mobility or capacity necessary to support the adopted growth management policies, control sprawl or
prove the mobility necessary for sustained economic vitality.” (Ref 19: p.23)
“Finally, only the Rail/TSM Alternative had sufficient capacity to accommodate transit demand in key
areas such as the Seattle CBD…” (Ref 19: p. 25)
In sum, the effect of Capitol Hill on the capacity issue is not an obvious point and must
be remembered when comparing bus and rail capacities in the DSTT. It means that if bus
and rail alternatives are both designed to achieve the same regional transit ridership, the
bus alternative needs only 80% as much capacity in the DSTT as the rail alternative.
5.3.5 Uncertainty about demand at the peak load point places ST in a bind.
We now turn from discussing capacity to discussing demand, and how demand relates to
capacity.
There are three statements in the RTP and ST literature that address demand at the peak
load point. First, the 1993 FEIS said that a125-mile rapid rail system would have a peak
one-way load in the year 2020 of 15,000 persons per hour just north of the DSTT. (Ref 1:
p.2-58)
Second, the Feb. 2002 EA for Link’s Initial Segment said: “Long range ridership
projections for Sound Transit estimate up to 15,000 Central Link riders in the peak hour
peak direction at the peak load point (assumed to be in the DSTT).” (Ref 18: Appendix
L, p.23).
There are several problems with this statement. First, although ST said “Central Link”
they must have meant a mature 100+ mile version of Link. Nothing else makes sense
since they had previously estimated peak ridership on Central Link was only 5415 pph.
Second, the 15,000 number appears to be just blind repetition of the 15,000 number in the
1993 FEIS, and is probably wrong for a variety of reasons. For one thing Link is slower
than rapid rail so a mature light rail network would attract fewer riders than a mature
rapid rail network. Second, it is very unlikely ST has modeled a mature version of Link,
and even more unlikely that such modeling would have produced exactly the same
15,000 value. Despite these problems ST did publish this number and we need to
examine its consequences.
69
The third data point is the Fleet Management Plan for the 24-mile Central Link, which
states that the peak load would be 5415 pph in 2020. (Ref 38: p.13) This is not directly
useful since it applies to a 24-mile subset of what could be a 100+ mile light rail network,
and we need the value for that mature network. Nevertheless it is a useful number. As it
came from a relatively recent modeling exercise for Central Link it is probably fairly
accurate.
This presents a dilemma. We need an estimate of peak loading on a fully built out
version of Link in order to determine whether Link would have enough capacity, as well
as to see whether or not a BRT system carrying the same overall number of transit riders
would have enough capacity. We can either accept ST’s statement that 15,000 is correct,
or derive an estimate from other data. This is done below, but since the results differ a
great deal and I can’t prove which is correct, I will describe the implications of both.
Author’s estimate of peak load on a mature version of Link-- Rather than accept
ST’s figure of 15,000 there is another way to estimate the leak load point volume on a
mature or fully built out version of Link. That way is to take ST’s forecast for the peak
load on the 24-mile version of Central Link and factor it up based on ridership patterns on
RTP’s 124-mile rapid rail system. The peak load predicted for Central Link was 5415
pph in 2020. (Ref 38: p.13) An analysis of ridership on RTP’s rail system indicates that
peak load point volumes on the full 124-mile rail network would be about 1.58 times the
peak load point volume on a portion of that network equivalent to the 24-mile version of
Central Link. Details for computing that ratio are given in the Tech Notes.
Assuming this same ratio applies to Link, a mature 100+ mile version of Link would
probably have a peak load of about 1.58 times the peak load on the 24-mile Central Link.
This comes to 8450 pph. (5415 x 1.58) The author fully acknowledges that 8450 is a
rough estimate, but it is probably accurate enough to support the points that follow.
These demand figures are plotted in Figure 5.3.5 along with their BRT equivalents.
70
Implications of the two conflicting estimates-- The entire capacity issue hinges on
whether rail and buses have enough capacity to handle projected demand in 2020, and
beyond. The preceding paragraph revealed two conflicting estimates of projected year
2020 demand: 15,000 and 8450. It’s not clear which is correct. Therefore:
If 15,000 is correct:
1) Although light rail with its capacity of 16,400 could handle the 2020 demand of
15,000, it could not meet the region’s long term needs since it would not have enough
excess capacity to handle demand growth much beyond 2020. In other words light
rail is not strategic.
2) Buses do have adequate capacity so they can’t be ruled out as being inadequate to
handle 2020 demand or meet the region’s long term needs. Therefore an all-bus
solution like BRT is strategic.
If 8450 is correct:
1) Light rail could handle 2020 demands with extra capacity for growth.
Light
rail
BRT
Figure 5.3.5
31,500 (with 2nd Tunnel)
8450 ? 6760 (62 buses/hr)
15,800 (145 buses/hr)
13,100 (120 buses/hr)
12,000 (110 buses/hr)
4330 (40 buses/hr)
16,400
15,000 ?
5415
Maximum capacity of
Link light rail
Max demand at peak
load point for 100+
mile light rail network
Max demand on
24-mile Link
70 buses/hr currently
operate in DSTT
71
2) The bus capacity issue clearly disappears, but by an even wider margin. It would
only take 62 tunnel buses per hour to service this demand, and that is fewer than
operate in the tunnel today.
3) Buses clearly can’t be ruled out as being inadequate to handle 2020 demand or
meet the regions long term needs. Again BRT is strategic.
4) The TSM bus alternative in the 1993 FEIS could probably attract and carry more
transit riders than a mature version of Link. The TSM alternative in the 1993 FEIS
had 85% (474,000/560,000) as many riders as the rapid rail alternative. In contrast a
mature light rail network with a peak load of 8450 would have only about 57%
(8450/15000) as many riders as the rapid rail alternative with its peak load of 15,000.
In other words, the original TSM design, without extension and certainly without a
second bus tunnel, could probably out-perform a mature light rail system in ridership.
This intriguing theme won’t be pursued further since doing so would overly
complicate this report and isn’t necessary to support the report’s main conclusions. It
is simply one more reason why an apples-to-apples comparison is needed between a
fully mature version of Link and an all-bus alternative. Until then we can only
estimate the relative merits of a mature version of Link vis a vis an all-bus alternative
by extrapolating from data in the 1993 FEIS for rapid rail.
This “if, then” analysis places ST in an embarrassing bind. If ST sticks with the 15,000
figure it must admit that Link is not strategic and therefore disappoint supporters who
justify Link on the assumption it will meet the needs of our children and grandchildren.
In addition, ST must admit that bus alternatives –if properly designed-- do have enough
capacity and should have never been taken off the table for capacity reasons. On the
other hand if ST accepts that 8450 may be more accurate, ST is forced to admit --even
more assertively and without qualification-- that buses could easily do the job since it
wouldn’t even require more buses in the DSTT than currently operate there.
5.4 Other capacity issues
5.4.1 Even before 2020, Link may not have sufficient capacity to handle demand on
the south line
The peak load point on Link’s south line would be just south of the International District
station. Capacity at this point depends on headway. Although elsewhere ST claims to be
able to operate on as little as 5-minute headways in the Rainier Valley this may be overly
optimistic (Ref 38: p.iii). The 1993 FEIS says: “Conventional transit practice and
highway standards suggest that when train frequencies are under 6 minutes, cross traffic
on arterials will be affected to the extend that grade separation is necessary.” (Ref: 1, p.2-
50) Even then 6-minutes requires merging eastside trains perfectly so as to achieve 2-
minute headways in the DSTT. In any case the author assumes for purposes of this report
72
that 6 minutes is the best Link could do. Six-minute headways would give the south line
a maximum capacity of 5470 persons per hour.
One potential way to provide extra capacity at the peak load point on the south line would
be to run trains on three minute headways down past the peak load point and then reverse
every other train at Beacon Hill so headways in the Rainier valley stay at 6-minutes. ST
mentions this idea but says that any reversal north of Henderson is impossible. (Ref 38:
p.11) Thus capacity of the south line can be no more than 5470 through the peak load
point.
The “Fleet Management Plan” for Central Link shows year 2020 peak loading on the
south line would be about 4300. (Ref 38: p.24) However, it is relatively hard to say what
peak loading on the south leg of a mature network might be based on that figure.
However there are two other ways to estimate it. Again, both are based on ridership
patterns in the 124-mile rapid rail network. Again, both depend on whether the peak load
on a mature version of Link would be 15,000 or 8450.
The first uses the map of route volumes in the 1993 FEIS. Figure 3.21 in that report
shows that daily demand on the most heavily used portion of the north line is 124,000
riders per day, whereas the most heavily used portion of the south line carries 59,000
persons per day. The ratio between them (2.1 to 1) should also apply to peak loads at
their respective peak load points. Thus if peak load on the north line is 15,000 peak load
on the south line would be 7140 pph. If peak load on the north line is 8450 then peak
load on the south line would be 4020.
The second approach assumes the ratio between passengers going into the north corridor
(from the CBD east and south corridors) and passengers going into the south corridor
would be the same as the ratio between the peak loads into those corridors. The RTP's
Travel Forecasting Results Report says that 88,600 daily trips go to the north corridor
while 61,000 go to the south corridor. This is a ratio of 1.45 to 1. Thus if peak load on
the north line is 15,000 peak load on the south line would be 10,340 pph. If peak load on
the north line is 8450 then peak load on the south line would be 5830.
The author has no way to determine which of the four estimates above is best. However,
in three of four cases, demand exceeds capacity at the peak load point on the south line.
Once again, the uncertainty about demand places ST in a bind. If ST asserts that 15,000
is correct, then Link would not have enough capacity to meet even year 2020 demands on
the south line. On the other hand, if ST finds that 8450 is more nearly correct, then once
again ST can’t claim buses lack sufficient capacity.
Given the uncertainty, it is important to conduct a new and proper alternatives analysis
wherein one of the alternatives is a fully built out 100+ mile version of Link, per the
dreams of ST management. This would provide a much better estimate of peak load
point volumes than we have today and provide a more definitive answer to the capacity
questions posed earlier.
73
5.4.2 Link’s design shortchanges the south corridor in terms of capacity
At great expense, officials decided to tunnel and otherwise entirely grade-separate Link
north of the DSTT. Besides improving safety, this allowed 2-minute headways in the
north corridor. In contrast, they decided to save money by running parts of the south line
on surface streets. As a result the south corridor gets only one-third the people moving
capacity afforded the north corridor. This may be an equity issue since taxpayers in the
south corridor pay at the same rate as those in the north.
5.4.3 Sound Transit’s EIS makes the false claim that light rail can move as many
people as a 12-lane freeway.
The 1999 FEIS for Central Link contains the following statements:
“ For example a light rail line can move the same number of people at peak travel
hours as a 12 lane highway,….” (Ref 2 p.1-9)
“The proposed light rail line would provide the same people moving capacity as a
12 lane highway…” (Ref 2: p. 1-9)
Sound Transit board members have further distorted the truth with statements like the
following:
“When this system is up and running - Northgate to SeaTac - in 2020, it will carry
as many people every day as I-5 does today.” (ST Board member Cynthia
Sullivan’s statement on Dave Ross’s March 12, 2003 radio show)
This issue must be dissected in two parts. First, is to compare how many people light rail
can or could carry versus what a 12-lane highway can or could carry. This type of
comparison is essentially irrelevant and meaningless, but that is how the EIS statements
were worded and a response is warranted. The second, is to compare what highways
actually carry with Sound Transit’s predictions of what light rail is actually expected to
carry. The latter is far more relevant.
The first is simple. Link light rail has a maximum one-way capacity of 16,400 persons
per hour. As explained in Addendum F a single highway lane devoted to BRT could
carry well over 50,000 persons per hour. The theoretical maximum is probably over
70,000, but it seems pointless to seek an exact number since no busway in the world
requires much over 40,000. If a single lane devoted 100% to buses could carry over
50,000 pph it is obvious that even a single highway lane devoted to a mixture of buses,
carpools and vanpools could carry more people than Link. Thus Sound Transit’s EIS
statements are patently false, and are simply meant to mislead. They typify Sound
Transit’s habit of using Environmental Impact Statements to sell rather than to inform.
74
Transportation consultant Jim MacIsaac has compared the actual number of people
carried on I-5 today with the number Sound Transit predicts would ride light rail in 2020.
His charts are reproduced below.
Mr. MacIsaac found that in the year 2000 Interstate 5 carried 412,700 persons per day
across the Ship Canal, in contrast to the 72,200 predicted for Link in 2020. (these are bidirectional
daily volumes) Since Interstate 5 has 12 lanes across the Ship Canal it is quite
apparent that Link would not be carrying as many people as a 12-lane freeway even here
where Link is heavily used.
The overall utility of a facility can better be shown by looking at the total number of
people it serves along its entire length. In this respect Mr. MacIsaac found that 1,145,100
persons “boarded” I-5 along the 23-mile stretch paralleling Central Link. In contrast
Central Link would have 157,000 boardings. This clearly shows that the Board
member’s statement “…in 2020, it will carry as many people every day as I-5 does
today” was vastly incorrect.
In sum, Sound Transit’s claim that Link is equivalent to a 12-lane freeway is simply
false. Based on Sound Transit’s predictions for Link ridership the best it would do by
2020 is carry somewhat more than a single lane pair on I-5 carries today through
downtown Seattle. Along most its route Link would carry less on a daily basis than a
single freeway lane pair (ie: 2-lane freeway).
I-5 vs. Light Rail: Average Weekday Person Trips
0
50,000
100,000
150,000
200,000
250,000
300,000
350,000
400,000
450,000
500,000
S.272nd
SR-516
S.200th
S.154th
Boeing Acc
Spokane
I-90
Mercer
NE 45th
Lk City Wy
NE 110th
NE 145th
NE175th
Selected I-5 Interchange Locations (Approximately to Scale)
Person Trips per Day
Interstate 5(a) Link - 200th to Ngate(c) Initial Segment(d)
Can LRT carry as many persons as a 12-lane freeway? Absolutely Not
Ship Canal
Maximum Load Points
North King
South King
Sea-Tac Airport
Downtown Seattle
Rainier Valley
S. 200th Link Terminus
Northgate Link Terminus
I-5 Person Trips - Year 2000
Sum of bus transit trips plus vehicle trips
@ 1.35 average vehicle occupancy
FINDING: By 2020 with full operation from S.200th to
Northgate, Central Link is estimated to carry only 14%
of the person trips per day carried by I-5 in year 2000.
Prepared by:
James W. MacIsaac, P.E.
April 15, 2003
Central Link 2020 Person Trips
Corridor Comparison Section
Trips Accommodated
2000 I-5 - 1,145,000 per day*
2020 Link - 157,000 per day
2020 Link IS - 42,500 per day
* On-ramp counts.
10 Lanes 13-12 Lanes 11 Lanes 10-8 Lanes
75
5.4.4 The planned redevelopment of the Alaska Way viaduct and south Lake Union
area offer a new possibility to establish a second major “busway” through
downtown Seattle.
A quick look at the Seattle map shows that it might be possible to construct a grade
separated “busway” extension from the Mercer St express lane exits, go west a few
blocks through the south Lake Union area, then tie into SR 99. Buses may need to
mingle with cars for a short distance on SR 99 but then if the new Alaska way “tunnel”
contained HOV lanes they could proceed south to serve the Duwamish and West Seattle
corridors, and even loop back into I-5. This would effectively create a second major
route for bus HCT through Seattle. The busway could be aesthetically integrated with the
planned construction in south Lake Union and provide service to that important emergent
growth center, which light rail completely ignores. It could also better serve the growing
Belltown area. This would give buses another major route through downtown and
probably eliminate any capacity constraint imposed by the DSTT. These possibilities
were not apparent at the time of the 1993 FEIS, another reason that document is obsolete.
5.4.5 Placing light rail on the I-90 bridge will reduce the bridge’s usable capacity.
Sound Transit is currently maneuvering to have the center roadway on the I-90 bridge
designated for exclusive use by high capacity transit or HCT. To Sound Transit HCT
I-5 vs. Link Light Rail - PM Peak Hour Person Trips
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
50,000
S.272nd
SR-516
S.200th
S.154th
Boeing Acc
Spokane
I-90
Mercer
NE 45th
Lk City Wy
NE 110th
NE 145th
NE175th
Peak Hour Persons - Both Directions Combined
I-5 (a) Link - 200th to Ngate (b) Link Initial Segment (c Ship Canal
North King
South King
Sea-Tac Airport
Downtown Seattle
S. 200th Link Terminus
Northgate Link Terminus
Corridor Comparison Section
I-5: 11 Lanes Avg (Incl HOV)
NOTES:
(a) 2000 traffic counts x 1.30 ACO plus 2000 transit.
(b) Sound Transit 2020 ridership estimates - 24 miles to Northgate.
(c) Sound Transit 2020 ridership estimates - 14-mile Link IS.
Central Link 2020 Rides
Rainier Valley
Link IS 2020 Rides
I-5 Person Trips
in Year 2000
Daily or Peak Hour -- The 12-lane Myth Does Not Apply
76
really means light rail. Such an action would reduce the capacity of the bridge. The
reason is not complex. According to the 1993 FEIS during peak periods there would be
11 trains per hour on the bridge by 2020. (Ref 1: p. 3-112) At 537 persons per train this
amounts to only 5900 persons per hour. There is little reason that light rail would attract
any more transit riders than would BRT in that corridor (the author knows from
experience express buses provides excellent service) so one can assume BRT would also
carry at least 5900. It would take about 54 buses per hour or roughly one per minute to
carry this load. There would be plenty of room to fit other HOV vehicles between the
occasional buses. Thus if the center roadway were used by a combination of BRT buses
(carrying 5900 pph) + van and car pools + school buses + emergency vehicles + other
HOVs its capacity would be considerably greater than if dedicated to light rail carrying
just 5900. Given that right of way is probably the most precious transportation resource
in an urban area, maximizing its use would seem a high priority.
Sound Transit might argue that it is not removing any lanes—by taking the two wide
center lanes while adding two more narrow lanes on the outside of the bridge—and is
therefore not reducing bridge capacity. However, ST is reducing bridge capacity below
what it would be if the center lanes were put to the mixed use described above.
A proper FEIS would address this issue.
5.4.6 Unless light rail is extended to the eastside, devoting the DSTT to light rail will
reduce the capacity of that facility to handle commuters to and from Seattle’s CBD.
The issue here is getting maximum people moving benefit from King County’s
investment in the DSTT. If light rail does not go across Lake Washington the DSTT will
be able to move a maximum of about 21,000 persons into downtown during AM peak
hour (or out in the PM). This would be comprised of 16,400 from the north and 5500
from the south. In contrast buses in the tunnel could move about 32,000 persons per hour
into the downtown. This assumes tunnel buses are operated efficiently (145 buses per
hour) and can therefore move in 15,800 from the north plus 15,800 from the south and
east.
There is no guarantee that Link’s Initial Segment, much less Central Link, will ever be
extended. This, plus the fact that a partly completed light rail system would compromise
the bus system, are among the type of risks that a proper alternatives analysis would
address.
5.4.7 In the far distant future the region might need heavy rail, not light rail.
There is no capacity crisis that demands or deserves rail in the foreseeable future. BRT
can easily handle the demand for many years to come. In the unlikely event the region
ever gets dense enough to outgrow BRT it would take a completely grade-separated
heavy rail system like BART, not light rail, to solve the problem. BART-like systems
77
use long trains and have one-way capacities in the 40,000 pph range as opposed to Link
maximum of 16,400.
Such a system could always be built later. (Although not across the I-90 bridge for
structural reasons.) The rights of way are not disappearing forever. The subsoil under
Capitol Hill, under the Ship Canal, and along the proposed rail line is not being preempted
for other uses. Tunneling technology will continue to advance.
In short, there’s no urgency to build the ultimate system as an investment for our
grandchildren, and the ultimate system wouldn’t be light rail in any case.
5.4.8 Putting the capacity issue in perspective.
The above analysis has shown that a fully mature version of Link light rail, extending
over 100 route miles would probably have a peak load point volume of roughly 8500
persons per hour in the peak direction by 2020. This is the best estimate we have at this
time since it is based directly on the ridership forecast for Central Link factored up with
ratios from the rapid rail studies. As explained above a comparable BRT system would
need about 80% as much capacity as light rail at the peak load point, or 6800 pph. This
translates into about 63 buses per hour, a small fraction of what an HOV lane could carry.
According to the Transit Capacity and Quality of Service Manual published by the
Transportation Research Board, an HOV lane used exclusively for buses could carry 800
to 1000 buses per hour. Sixty three buses per hour can easily be carried by the existing
express lanes on I-5 and the existing DSTT. This very, very netted down summary begs
the question: Given what we know today, why should there be any concern about bus
capacity?
The following comments from the METRO Six-Year Plan dated Dec.1995 also help put
the capacity issue in perspective:
“The existing transit system has a significant amount of underutilized capacity.
* Less than 3 percent of currently scheduled bus trips run more than 90 percent full.
* During the PM peak commuter period (3:00 to 6:00 pm), only 56 percent of the
seats on the buses leaving downtown Seattle are filed. Utilization is even lower on
buses departing Bellevue and the University District.”
In other words, the problem is lack of demand not the shortage of capacity.
Another key point re capacity is that there is no specific benchmark, target or figure that
represents the region’s need, even in a particular year like 2020. The larger or more
intensive (frequency of service, number of routes) the rail or bus system is, the more
ridership it would attract and thus the more capacity it would need to carry that ridership.
However, the size or intensity of the transit system is simply a variable, which depends
on how much society decides to spend on it. If the RTP had chosen to propose a 50-mile
78
rail system costing say $6 billion rather than a 125-mile system costing $11.5 billion the
maximum load would have been much less than 15,000. And a bus system with the
same overall ridership would have had a max load in the DSTT so low as to raise no
concern. As it was the RTP chose to study alternatives whose ridership was just large
enough to make bus capacity an issue, but small enough to keep rail capacity from
becoming one.
In short the amount of capacity we need depends on the level of demand, which in turn
depends on how large a transit system we intend to build. And how large a system we
intend to build depends on how much additional transit ridership we want to buy. It all
comes down to the cost and affordability of generating additional demand. Thus the two
key policy questions are these:
1) Which is the least expensive way to buy added transit ridership, bus or rail?
2) How much should we buy? (How much is affordable? Given the law of
diminishing return, at what point does the ever increasing cost of buying additional
transit ridership outweigh the benefit?)
The focus should be on answering those questions. Capacity should be relegated to its
proper place, which is just one of the underlying technical details addressed along the
way.
5.4.9 Main conclusions for Part 5
1) The RTP unfairly manipulated the 1993 FEIS in a way designed to make the bus
alternatives appear to have insufficient capacity. This was done, in part, by putting forth
a bus alternative with known but correctable capacity limitations.
2) Statements in the 1993 FEIS alleging that an all-bus alternative would not have
sufficient capacity are misleading and incorrect.
3) A properly designed all-bus alternative would have more capacity than Link light rail.
It would have enough capacity to handle forecast transit ridership in 2020, and far
beyond.
4) There has been no ridership forecast nor capacity analysis done for a mature or fully
built-out version of Link light rail, therefore it is somewhat uncertain such a network
would have sufficient capacity to handle its projected ridership.
5) Although it purports to, the 1993 FEIS does not provide proper justification for
continuing on a rail-centric strategy for capacity reasons. This is one reason the FTA
should not accept the DSEIS for ST’s Long-Range Plan, since the DSEIS is based on the
1993 FEIS.
79
6) The capacity issue is a red herring that has diverted attention from more important
issues, especially the relative cost-effectiveness of rail and bus alternatives as means to
increase transit ridership and reduce travel delay.
7) The faulty findings of the 1993 FEIS appear to have been instrumental in directing the
region toward choosing rail rather than bus technology for the backbone of the region’s
high capacity transit network.
8) A new and proper alternatives analysis is needed in order to officially confirm or deny
these findings, provide more accuracy to the various numbers estimated herein, and
stimulate reconsideration of the regions current rail-centric strategy. To accomplish this
the analysis would need to consider a number of different size light rail networks ranging
from Links 14-mile Initial Segment up to a fully mature 100+ route mile version of Link.
Each of these different size light rail networks should be compared against bus or BRT
alternatives designed to achieve equal ridership.
80
Blank page
81
Part 6: Cost Effectiveness Issues
6.1 The 1993 FEIS masked the poor cost-effectiveness of rail
The voluminous 1993 FEIS contained very little information on cost-effectiveness. What
little it did contain was summarized in Table 4, which is scanned in just below.
Whether by ineptitude or design the way the information was presented in Table 4
masked the fact that rail is much more expensive than buses as a way to increase transit
ridership. Just looking at the $7.74 versus $5.92 values in Table 4 doesn’t shock the
reader into realizing that.
Figure 6.1 takes the information in Table 4 and presents it in a way that does highlight
this information. What Figure 6.1 shows is that if nothing special is done to improve
transit, by 2020 the existing bus system (NoBuild) will attract and serve 389,000 daily
riders at a cost of about $3 each. (Each so caller “rider” is actually a single one-way trip)
However society could increase ridership another 85,000 by implementing the TSM bus
alternative. Those extra 85,000 riders would come at a cost of about $15 each. Then, if
still more riders were desired, society could add the 125-mile rapid rail system in addition
to TSM. This would add a still additional 86,000 riders, at a cost of about $25 each.
NOTE: The data sources and calculations behind the charts below can be found in the
Tech Notes at the back of this report. The author’s No-Build cost per rider differs
slightly from that in Table 4 due to differences in the way capital costs were annualized.
Presenting the information in this way properly calls attention to several important policy
issues. First, it shows that adding riders using rail as about twice as expensive as adding
them using TSM. This begins to call attention to rails high cost, relative to bus, as a way
of increasing transit ridership. It also it shows that increasing ridership by investing in
massive transit projects such as Rail/TSM is quite expensive relative to what society has
82
been willing to spend on transit historically. Finally, it begs the question: “Is it worth
it?.” Is it worth spending $15 or $25 per ride just to increase transit ridership?
The 1993 FEIS was deficient, if in fact not deliberately deceptive, in not having presented
cost and ridership information in a way that highlighted these points.
6.2 The 1993 FEIS failed to explore the possibility of extending TSM to higher
ridership levels.
A plot to visualize key performance versus cost relations can be useful. An FEIS that
genuinely attempts to educate should make use of such graphics.
Thus Figure 6.2a simply plots the data in Table 4 to show graphically how ridership
increases in proportion to money spent. The dot at the end of each line shows the cost
and ridership that would occur if the alternative
0
5
10
15
20
25
30
0 200 400 600
Riders 000's/day
Cost in 91$/rider
$15.50 each for those
added by TSM
$3.30 each with No
Build
$24.70 each for those
added by Rail
389 474 560
Figure 6.1
83
were fully implemented. In other words if the Rail/TSM alternative described in the
1993 FEIS were fully implemented it would involve building 125-miles of rapid rail, cost
$960 million/year more than the No Build alternative and attract 172,000 more daily
transit riders than would the NoBuild alternative.
However, it is possible to implement scaled-back versions of the alternatives, such as a
shorter rail system or a smaller TSM system. These would cost less and attract less
ridership but be incremental steps toward the mature systems.
Thus each of the lines in Figure 6.2a can be thought of as a strategy thrust, something
starting at lower left (the No Build base) and expanding incrementally along the line until
either the money runs out or the ridership goal has been obtained. This is because the
HOV segments, park and ride lots, and increased bus services that comprise the TSM
0
200
400
600
800
1000
1200
0 50 100 150 200
Riders above NB 000's/day
Cost above NB millions of 91$/yr
Rail/TSM
TSM (constrained)
Why not just do
more TSM ?
85 172
Figure 6.2a
84
alternative can be built incrementally, with each improvement bringing more TSM
ridership. Likewise the 125-mile rail system can be built a leg at a time or even a station
at a time, each improvement leading to increased ridership. The general relation is clear,
the more money spent on each of these strategic thrusts, the more ridership it will attract.
A glance at the arrows in Figure 6.2a shows why extending TSM seems attractive;
namely it appears TSM could reach the same ridership as rail, but at less cost. The RTP
acknowledges that extending TSM is possible. In one report RTP addresses the
“Incremental Cost to Expand Capacity” saying: “The cost to expand the No Build, TSM,
and Transitway/TSM Alternatives is identical at an estimated $2.1 billion per additional
100,000 passenger carrying capacity.” (Ref 45: p.46)
However, the FEIS didn’t consider this possibility. Instead, as was explained in Section
5, the FEIS claimed TSM ridership was capped at 474,000 (85,000 over no build) due to
an alleged 100 bus-per-hour capacity limitation in the DSTT. (An internal Metro memo
from 1992 obtained by the author indicated a level of staff concern about this, and might
be relevant if there is any investigation of wrongdoing. See Section 5.4.3 for quotes
from this memo.)
The author has therefore constructed Figure 6.2b to show how the cost picture would
have looked had the RTP removed the capacity constraint using what is probably the
most expensive option; namely building a second bus tunnel costing $600 million. This
would have allowed TSM to achieve its “unconstrained” ridership level of 518,000 daily
riders (130,000 above No Build), rather than its constrained ridership of 474,000.
The end result in Figure 6.2b is a scenario wherein TSM -- with 518,000 riders-- has
gotten much closer to rails ridership without a dramatic increase in cost. The cost of the
tunnel appears relatively insignificant in the big picture.
85
At this point, and with the second tunnel in place, TSM has no further capacity
limitations. However, since TSM is doing more, there is less remaining for rail to
contribute. The number of additional riders rail could add above and beyond TSM has
diminished from 86,000 to just 42,000. In other words adding the 125-mile $7.9 billion
rail system would only increase ridership by 42,000 riders. Above and beyond what the
unconstrained TSM alternative would attract.
Figure 6.2c shows what that means in terms of cost per rider. Figure 6.2c is essentially
Figure 6.1 updated to show the impact of adding the second bus tunnel or some other way
to relieve the capacity bottleneck so TSM can achieve its unconstrained ridership. Figure
6.2c shows it would cost $48 (in 91$) per ride for each of the 42,000 additional daily
transit riders rail attracts. Again Figure 6.2c begs the question: how much is it reasonable
to pay to expand transit ridership? Today bus fares average 80 to 90 cents, perhaps rail
0
200
400
600
800
1000
1200
0 50 100 150 200
Riders above NB 000's/day
Cost above NB millions of 91$/yr
TSM if constrained
TSM if ridership is
unconstrained using a 2nd bus
tunnel
Rail/TSM
130
$43 million/yr
for 2nd tunnel
Figure 6.2b
86
fairs would be higher, but the remainder (probably $45+per ride) would be a public
subsidy. (Later Figure 6.3b will show how Figure 6.2c looks when updated to current
conditions.)
0
5
10
15
20
25
30
35
40
45
50
0 200 400 600
Riders 000's/day Cost in 91$/rider
$47.50 for
Rail/TSM
$11.50 for TSM
with 2nd bus
tunnel
$ 3.30 for
No Build
390 560
Figure 6.2c
518
87
Figure 6.2d, completes the story that the 1993 FEIS should have presented. It shows
what it would have cost to extend the TSM alternative until its ridership was the same as
the rail alternative. This equal ridership comparison is the type of apples-to-apples
comparison needed. It is much more revealing than RTP’s 1993 FEIS, where the
alternatives were neither equal in cost or in benefit.
The cost of the “TSM extension” in Figure 6.2d was based on the previously cited RTP
statement that it would cost $2.1 billion in capital for 100,000 additional riders. The
TSM extension needs only 42,000 to equal rails ridership, and thus would cost $882
million. After O&M costs are added, the total cost of using TSM to reach rail ridership
levels is $556 million per year. See Tech Notes for details. This analysis shows that
0
200
400
600
800
1000
1200
0 50 100 150 200
Riders above NB 000's/day
Cost above NB millions of 91$/yr
TSM (unconstrained)
Rail/TSM
TSM extension
$ 404 million/yr
difference
Figure 6.2d
88
using TSM rather than the Rail/TSM alternative would save taxpayers about $400 million
(in 91$) per year over the 30-year life of the bonds.
The 1993 FEIS was deficient in not having developed and presented this type of analysis.
It’s critical for public officials and taxpayers to understand just what they would be
paying to expand transit ridership, and to judge whether that cost is reasonable.
6.3 Using the latest information, the difference between the cost of rail and bus is
greater than it would have appeared in 1993.
Figure 6.3a brings Figure 6.2d up to date by correcting for known changes since 1993
and converting costs from 1991$ into 2002$. Again it shows the relative cost of the
TSM versus the Rail/TSM alternatives as they would look today.
Strictly speaking the rail ‘curve’ in Figure 6.3a is an update of how the rapid rail/TSM
alternative in the 1993 FEIS would look today if the cost information were updated but
the original ridership predictions were left alone. Nevertheless, Figure 6.3a is a fair and
useful indication of the situation facing us here in the fall of 2004 as officials press
forward, not with rapid rail but rather with light rail.
In other words, Figure 6.3a is a conservative estimate of the relative costs of pressing
ahead with light rail as opposed to switching to an all-bus alternative centered around
BRT. It shows that taxpayers could save over $900 million per year by switching to an
all-bus alternative.
The qualifier “conservative” is used because there are some differences between the rapid
rail system laid out in the 1993 FEIS and Sound Transit’s light rail. First, we know
light rail would have fewer riders since it is considerably slower than the rapid rail in the
1993 FEIS. Central Link would have an average speed of 26 MPH whereas the rapid rail
in the 1993 FEIS had an average speed of 36 MPH. (Ref 38: p.16 and Ref 1: p.2-58)
Any eastside extensions of light rail may or may not operate on streets, but if they do,
they would also run slower than rapid rail. Second the capacity of Link’s south line
might constrain ridership.
When these factors are considered the rapid rail/TSM curve in Figure 6.3a probably
presents an overly optimistic view of what we can expect from Sound Transit’s light rail
strategy in terms of cost versus ridership. For these reasons this Figure is conservative.
Most likely the rail curve would be steeper than shown and the cost difference greater. Of
course we can’t create a more accurate version of this chart without a new and proper
alternatives analysis to provide the data.
89
.
Figure 6.3b is the updated version of Figure 6.2c. It puts the costs in 02$, and accounts
for the fact that rail costs in the 1993 FEIS were underestimated and half the HOV
network is now complete. In other words, Figure 6.3b is based on the same capital and
operating costs as were used to construct Figure 6.3a.
The solid line in Figure 6.3b represents the costs of the Rail/TSM scenario or strategy.
First, it shows the No build riders cost $4.35 each. Then it shows it would cost $13 for
each rider the TSM component of the Rail/TSM alternative can attract above No Build.
And it shows it would cost $ 93 for each rider (or ride) which rail could attract above and
beyond what TSM could attract. Again it raises the question as to whether the last 42,000
riders which rail could attract are worth society spending about $90 to subsidize each one
way ride.
0
200
400
600
800
1000
1200
1400
1600
1800
0 50 100 150 200
Riders over NB 000's/day
Cost over NB millions of 02$/yr
Rail/TSM
TSM (unconstrained)
TSM extension
$ 917 million/yr
Figure 6.3a
90
In contrast, the hard to see horizontal dotted line shows how much it would cost if the
region decided to scrap the rail strategy and simply extend TSM until TSM reached the
same 570,000 ridership level as rail. In that case all the additional riders --above what No
Build had already attracted-- would come at an average cost of $12.70 each.
6.4 Results when VMT is the metric for benefit
The 1993 FEIS uses transit ridership as the main metric for evaluating the cost
effectiveness of the different alternatives. Yet, almost no-one actually cares about transit
ridership per se. As noted earlier, what people are concerned with is traffic congestion or
perhaps some inadequately defined measure of “mobility”. The PSRC uses “travel
delay” as their effectiveness metric and of course the RTP and ST should have done
0
10
20
30
40
50
60
70
80
90
100
0 200 400 600
Riders 000's/day
Cost 02$/ride
570
$ 4.35 No Build
$13.17 TSM
$93.16 Rail
$12.70 Extended
TSM
Figure 6.3b
91
likewise. However, since they didn’t the author will use what’s available, namely total
daily “Vehicle Miles of Travel” or VMT. VMT is a proxy for congestion and delay. The
logic is simply that if the demand on the road network (as measured in VMT) is reduced,
there will be less congestion and less delay. (Peak period VMT would have been a better
proxy for congestion; however, that measure was not reported in the 1993 FEIS so could
not be used in Figure 6.4)
Figure 6.4 is therefore a plot of how well the rail/TSM and TSM alternatives perform in
terms of reducing VMT relative to the NoBuild baseline. Figure 6.4 is essentially Figure
6.3a redone with a different variable on the X-axis. The performance of Sound Transit’s
“Initial Segment” and “Central Link” projects are also included. This is possible because
ST did cite the effect of these projects on VMT in their respective EIS and EA. (Ref 18:
p.20) (Ref 2: p3-4). The Extended TSM line assumes that if the rail/TSM and extended
TSM alternatives have the same ridership –which they do by definition--they will also
achieve the same reduction in VMT.
Figure 6.4 is the most important chart in this report because it summarizes the current
situation facing the region as it begins implementing a rail intensive strategy as opposed
to a bus-only strategy.
Figure 6.4 conveys several messages. First, it reinforces all the prior charts in showing
that an all-bus strategy is much more cost-effective than a rail intensive strategy. Second,
it shows that Sound Transit’s current Initial Segment (of Link light rail) will be even less
cost effective than the hybrid rail/bus alternative studied in the 1993 FEIS. When
thought about carefully this simply means that the rail element --in any strategy that
improves both bus and rail in concert—will be less cost effective than the bus element.
Given that information, it again highlights the obvious question: Is the rail element cost
justified? Third, the relatively poor cost-effectiveness of Sound Transit’s light rail
projects is apparent because their plot-points are above the Rail/TSM line. In short, they
achieve relatively little reduction in VMT for their cost.
Fourth, the performance of a light rail only strategy (in other words extending the Initial
Segment until it reaches 100+ miles, but doing little or nothing for bus) can be implied by
constructing a line from the origin through either of the Link plot-points and thence out
until light rail achieves the same VMT reduction as Rail/TSM. It is clear that the cost
would well exceed even the high cost of the Rail/TSM alternative. This means it would
be ruinously cost-ineffective to pursue light rail without at least improving the bus system
at the same time, and with sufficient funding.
In short, Figure 6.4 teaches that bus-only is the most cost effective strategy, combined
rail/bus is a distant next best, and rail-only is worst of all. It shows that an Extended
TSM alternative could probably achieve the same reduction in VMT at a far lower cost
than would the rail/TSM alternative. At the point where VMT is reduced 3.75% the TSM
alternative is $917 million per year less costly than the rail/TSM alternative.
92
6.5 The need to consider other alternatives beside bus and rail.
When reducing congestion, reducing delay or improving mobility becomes the objective
–as opposed to simply increasing transit ridership—it invites consideration of other
alternatives besides just improving mass transit. It suggests adding to the list of
alternatives being considered such things as improving roads, increasing car and vanpool
use, building bike paths, and a host of “demand management” programs that seek to
reduce the need to travel.
Some of these might actually be more cost-effective in reducing delay than investing in
mass transit. Figure 6.5 illustrates this by showing –conceptually, in the form of hollow
circles-- a host of other alternatives some of which are likely less expensive than even the
TSM alternative.
0
200
400
600
800
1000
1200
1400
1600
1800
0 1 2 3 4
Reduction in VMT % under NB
Cost millions of 02$/yr over NB
Rail/TSM
TSM, extended
Initial Segment
(14-miles)
Central Link
(21-miles)
Figure 6.4
$ 916 million/year
93
Clearly, a truly proper alternatives analysis would include and evaluate these other
alternatives. FTA and SEPA regulations may not require that of alternatives analyses,
but the public interest clearly does.
6.6 Least cost planning: the final refinement.
So far the metric for cost has just included the direct capital and operating costs.
However, there are associated environmental, safety and other costs that should be
included to get a true measure of ‘social cost’. In addition, any ancillary benefits should
be subtracted from the cost to arrive at a net social cost. These non-dollar costs can
usually be converted to dollar equivalents using standard or semi-standard techniques.
0
200
400
600
800
1000
1200
1400
1600
1800
0 1 2 3 4
Reduction in VMT % under NB
Cost millions of 02$/yr over NB
Rail/TSM
Initial Segment
(14-miles)
Central Link
(21-miles)
Figure 6.5
TSM, extended
Circles represent other
possible alternatives
94
The end result is a long list of alternatives each rated in terms of its cost-effectiveness in
achieving the primary objective. The results could be plotted on a chart such as Figure
6.6, which again is conceptual. The Y-axis becomes a measure of net social cost
expressed in dollars. The X-axis is the measure of effectiveness or benefit. Alternatives
about which we have some data are shown as solid dots. Hypothetical alternatives are
shown as hollow circles. A line connecting the least cost alternatives defines the least
cost way to way to achieve any desired level of benefit along the X-axis. This line will
undoubtedly bend up according to the law of diminishing returns. Thus, once the easy
solutions are exploited and run out of steam, it will get increasingly expensive to make
additional progress against the main objective. At some point the added benefit won’t be
worth the added cost. This point can be determined using classic cost-benefit techniques.
Thus something like Figure 6.6 –if it were filled out with real data—would allow
decision makers to determine which projects would achieve the most benefit at the least
cost. It is a powerful and logical way to help prioritize projects in the face of limited
resources. This technique is sometimes called “Least Cost Planning” or “Integrated
Resource Planning”. Least Cost Planning (LCP) is a requirement of Washington State
Law which applies to the PSRC. PSRC should have used LCP to evaluate each major
project in the Metropolitan Transportation Plan (Destination 2030), including Sound
Rail/bus
curve
All-bus
curve
Least cost curve
Net Social
Cost
Reduction in Travel Delay Large
High
none
Figure 6.6
95
Transit’s light rail plans. (Ref 43) However, PSRC has not yet complied with that law so
there is no data available to build a diagram like Figure 6.6 or to know how the various
light rail and road projects now being proposed (as part of the RTID package, or the
Metropolitan Transportation Plan) actually perform.
The Tech Notes contain some added detail on “Least Cost Planning”.
6.7 Concluding comments
The Puget Sound Region has inadequate information upon which to make intelligent
decisions about which programs or projects would best solve the region’s transportation
problems.
The alternatives analysis conducted by the RTP in 1993 is a small and very imperfect part
of the homework needed. In future, the primary goal or objective for evaluating all
major transportation proposals should be something that people actually care about, such
as reduced congestion or delay. In addition, it is necessary to compare a full range of
alternatives, and to compare them using net social cost so that environmental and other
side effects are taken into account. This story about the 1993 FEIS and events since has
shown how far short we are from having good answers.
The analysis in Section 6 shows that a bus-only strategy is probably much more cost
effective than a rail or hybrid rail/bus strategy in reducing travel delay; however, it also
suggests looking for other alternatives that could be better yet.
96
Blank page
97
Part 7: Evaluation and recommendations
Parts 1 through 6 of this report have been relatively narrowly focused. First they
described the deficiencies of Sound Transit’s one and only alternatives analysis, the heart
of the planning process. Then they showed that Sound Transit’s current light rail strategy
would be far more expensive that a comparable all bus alternative. In short they made
the point that Link and the present rail-centric strategy, is the failed result of a faulty
process.
Part 7 includes several specific actions that should be taken to reassess, and if need be
change, the region’s current mass transit strategy before additional hundreds of millions
of dollars are committed. These are followed by asking, and answering, whether this
report is accurate enough to justify taking the actions recommended.
However, before making those recommendations several broader concerns should be
examined. One is to take a closer look at why we came to be in this situation. Why did
the planning process go astray? Another is the current lack of trust in government’s
ability to spend transportation tax dollars wisely. That’s relevant because
recommendations to fix the process must at the same time fix the deficit in public trust.
Sound Transit has become the poster child for why people distrust government to spend
tax dollars wisely, and that is adversely affecting not just Sound Transit’s prospects but
the ability to raise funds for other transportation projects as well.
7.1 Genesis of current situation, or why things went wrong.
Because source data is hard to find, every statement in Section 7.1 is not backed up and
footnoted in the same scientific manner used elsewhere in this report. Therefore the
statements in Section 7.1 could be called opinions. Nevertheless they are based on the
author’s frequent interactions with the rail planning process over a period of 10+ years,
and his struggle to explain why transportation plans and decisions that just don’t seem to
make sense --from the perspective of someone stuck in traffic, or from a taxpayer’s
perspective-- happen anyway.
It has become clear that explaining what drives officials and decisions in this arena is not
a relatively straightforward engineering or business analysis which starts with a statement
of the problem, goes through an objective evaluation of options, and results in some
rational conclusion as to which option would best solve the problem. Instead, what has
happened is best explained by some complex mix of myth, fact, wishful thinking,
uninformed opinion, altruistic and not so altruistic motives, hard-ball politics, ego,
psychology, bureaucratic maneuvering, and most of all, money. When all this enters the
mix it is not surprising that actually spending taxpayer money wisely so as to make the
most progress against traffic congestion falls by the wayside.
How did it come to pass that the planning process was abused in order to promote light
rail and that Puget Sound is now embarked on a light rail plan whose implementation
98
would result in the region wasting billions of dollars relative to an all-bus alternative?
Why is it that taxpayers still lack objective information about the alternatives?
To begin with one must acknowledge the superficial appeal of rail transit. Everyone can
envision walking aboard a train and being whisked effortlessly from place to place. Rail
lovers, like the author, know there is no other form of travel that is quite so soothing and
worry free as we relax, watch the world slide by, and need only remember to get off at
the right station. Those who have ridden subways in New York, Paris, San Francisco
and other cities find them far less hassle than driving in a strange city. Rail just seems
neat.
Environmentally minded folk favor rail’s electric propulsion, since it moves air polluting
generating plants elsewhere. In addition, they feel rail systems encourage high-density
development and reduce sprawl. Civic boosters claim Seattle can’t be a world-class city
without having rail transit. People see the region growing, traffic growing, and, at least
some feel the solution just can’t be more freeways, especially not through Seattle proper.
But if not freeways, what other than mass transit? However for most, rail is probably
perceived in some vague way as a solution to our serious traffic congestion problem. If
they wouldn’t ride it themselves, they hope it will get others off the road thus making
driving easier for themselves.
Emotionally, rail is (seemingly) a simple “silver bullet” solution for a complex and
intractable problem. People like simple solutions. Elected officials especially like simple
solutions. They are easy to explain, easy to sell.
Elected officials have another set of reasons to favor rail. They revolve around power
and money. If it proceeds, Link light rail will be among the largest public works projects
in Puget Sound history. Billions of dollars will be spent. Officials can feel important
making decisions about how billions are spent. They can proudly leave a concrete
legacy. They can please certain powerful parties on the receiving end of those billions.
They can be seen as providing jobs and stimulating the economy. Perhaps most of all
they can be seen a bringing in “free” money from Washington D.C. It has been said that
years ago Atlanta got the money that should have gone toward building a rail system
here. Now it’s time to get our fair share.
Transportation as messy as politics Things sometimes are connected in ways we don’t really see,
such as the “blanket primary” and the transportation votes.
…. Regard the tyranny presently exerted over public decision making by single issue Tim
Eyman/Permanent Offense vigilantes and by the single-interest constellation of law firms,
contractors and other vendors who dine at Sound Transit’s taxpayer financed light rail trough. The
diners reciprocate with a reverse flow of campaign contributions and political support to King
County and Sound Transit Chairman Ron Sims and other office holders who fill that trough
With new federal funding still uncertain and funds on hand low, Sims recently has sought back door
money from City Light ratepayers and has given King County Metro Transit an ultimatum to
commit now to operating the light rail line when it’s scheduled to carry its first passengers in 2009.
99
Seattle Mayor Greg Nickels, Sound Transit’s former finance chairman, will ask the City Council to
approve next week a fresh $63 million in scarce city funds to help with cost associated with tossing
Rainier Valley people out of their homes and businesses to make way for light rail.
Additionally, as Thomas Shapely pointed out on Tuesday’s Post-Intelligencer Editorial Page, Sims
and Sound Transit could use another back door device hidden in the regional transportation package
on this fall’s ballot to generate new billions to keep light rail alive. (Ted Van Dyk, Seattle P-I, April
4, 2002)
A recent Seattle Times article about Senator Patty Murray nicely summarized how this
pork barrel game is played:
She never wavered in her support of a controversial third runway at Seattle Tacoma International
airport, and she stuck with Sound Transit when the politics looked dicey, helping the agency
eventually to score $500 million to build light rail
(Seattle Times, Oct 25, 2004)
Rail transit has long been favored by at least a significant number of elected officials in
the Puget Sound area. This goes back at least to the Forward Thrust effort in the 1960’s,
which envisioned a BART like rapid rail system for Puget Sound. It continued though
several efforts to sell the voters on building a rail system. Voters rejected a $6.7 billion,
69-mile rail plan in March 1995, but in May 1996 voters finally approved the much
smaller 21-mile $1.8 billion “Central Link” light rail project. However by 2001 Sound
Transit found it had grossly underestimated Central Link’s cost and, in the wake of much
public concern, cut the project back to the 14-mile “Initial Segment”. The FTA provided
some funding for the Initial Segment, and construction has begun. (See reference 3 for
more history)
In trying to zero in on what went wrong, consider first the huge disconnect between the
problem (traffic congestion) and the proposed solution (light rail).
The first page of the glossy “Sound Move” brochure, which Sound Transit widely
distributed at the time of the 1996 vote begins as follows:
Why regional transit?
There’s a old saying that advises “if it ain’t broke don’t fix it.” But if you are one of thousands of
people traveling on our region’s overburdened and clogged highways each day you can probably
relate to a modified version of that proverb- it’s broke, let’s fix it.
The problem is traffic congestion. Our region rates one of the worst traffic …
(Ref 15: p.1) (underlining by Harkness)
From this the naïve, gullible voter might have gotten the impression that Link light rail
was being put forth by knowledgeable and responsible officials as an answer, or at least a
partial answer, to the congestion problem. Certainly Sound Transit intended to convey
that message. As anyone who attended the public hearings before the 1996 vote can
attest, Sound Transit did everything it could to imply, and lead voters to believe, that
100
Link was an answer to traffic congestion. However, this was done artfully so as not to
leave anything in writing.
Unfortunately, as was documented in Part 3, Link would have virtually no beneficial
effect on traffic congestion, something which had been known all along by transportation
planners, and something which David Earling, then chair of the Sound Transit board,
finally acknowledged publicly on December 26 of 2000:
Light rail: There will never be a better time Richard Harkness got one thing right in a guest
column (The Times, Dec 22): Light rail will not ease traffic congestion. Yes, that’s a fact. (Seattle
Times, Dec 26, 2000)
Of course, ST’s first public admission that light rail would not ease traffic congestion,
came a bit late since the vote had occurred four years earlier.
And in spite of statements hidden in the EIS and VMT data in the EIS’s showing that
Link wouldn’t have any noticeable impact on congestion, and in spite of Dave Earling’s
public admission to that Link would have no effect on congestion, key officials continued
to mislead the public. The following is taken from a speech King County Executive Ron
Sims made to the Sound Transit Board on January 11, 2001.
Mr. Chair, Board Members, I would like to move the adoption of Resolution 2001-01 and speak to the
motion….Today, we are voting to take a positive step forward to relive transportation congestion along
Interstate 5….We must today move ahead to build a regional transportation system with light rail.…I
urge a yes vote from all members of this board. (Ref 62) (underlining by Harkness)
That particular speech contains a number of other unfounded assertions that are
interesting to read. Later when he was chairman of the Sound Transit Board, Mr. Sims
said:
…we cannot bus our way out of congestion along I-5…That is why we are moving forward with light
rail.. (Ron Sims, in article entitled “Second Opinion: Light rail coming round the bend, King County
Journal, 2/9/03)
In a Sound Transit News Release celebrating FTA’s announced intent to fund Link,
Senator Patty Murray made the following misleading statement:
This is yet another important milestone in the effort to address the transportation needs of the Puget
Sound region. We cannot and will not stand by and watch some of the country’s worst congestion
choke our state’s economy. (Sound Transit news release dated July 11, 2003)
How could the planning process have gone so far astray as to propose spending billions
on a non-solution to the region’s most pressing transportation problem? And perhaps
more sinister: how could the public officials in charge of this process have gone forth to
the voters implying that light rail was a solution when they knew, or should have known,
it wasn’t?
These questions been asked about rail projects in other cities. (See References 21,23, and
24). As a participant in various public hearings going back before 1996, attendee at
101
Sound Transit Board meetings, and student of this long standing effort to sell rail, the
following seems the best explanation of observed behavior.
Key officials have long wanted a rail transit system in Puget Sound, period. In the early
1990’s consultants were hired to prepare the necessary paperwork in order to get federal
funding and meet state law. Although alternatives analyses are theoretically objective
studies, consultants are generally fairly astute at figuring out what the client really wants,
and repeat business means delivering it. As a result the 1993 FEIS and alternatives
analysis became a pro forma exercise meant to justify a preordained conclusion. The
elected officials in charge created the “culture” within which the 1993 FEIS was
completed, and almost certainly they did not demand an objective evaluation of
alternatives. It must have been a don’t ask, don’t look, don’t tell culture. Officials
probably didn’t ask why the bus alternative in the 1993 FEIS couldn’t be fixed and
probably didn’t ask probing questions about cost effectiveness of the rail element.
Certainly the 1993 FEIS did not address these topics. The technicians doing the actual
work no doubt understood they weren’t supposed to look for ways to fix the bus capacity
problem, present cost data objectively, or to tell management if they had ideas that would
have worked against the preordained outcome. Nothing strictly dishonest is needed in
this type of culture. It’s just a matter of not looking into things where the answer might
be unwelcome. If you don’t look you don’t have to know, and if you don’t know you
don’t have to decide whether or not to tell.
The end result was that the 1993 FEIS is a sham. The document is meant to satisfy State
and FTA requirements that call for an alternatives analysis, but is almost totally useless in
providing objective information that would help officials or the public make a wise
choice between rail and bus technologies.
However, assume for the moment that public officials were motivated solely by what
they sincerely believed was best for the region. Is it adequate that they make decisions
on the basis of belief alone, no matter how well intended? Certainly the answer is “no”.
Transportation is an exceeding complex matter about which the average citizen or elected
official has very little understanding. Just as the author, admitting no knowledge of
police and fire must entrust public officials to find good solutions, so must the average
voter trust public officials to find good solutions for transportation problems. We expect
officials to do the necessary homework so that knowledge replaces uninformed belief and
naive opinion. Thus taxpayers have given elected officials large staffs and have given
Sound Transit a virtually unlimited consulting budget. We expect due diligence. We
expect objective homework to be done, and we expect public officials to act on the basis
of sound analysis not on the basis of their gut feelings or personal beliefs. To the extent
this is not done there is a betrayal of the public trust.
Perhaps the worst abuses to the planning process occurred years ago in preparation of the
1993 FEIS, but the current Sound Transit Board of Directors has inherited the problem by
virtue of not having corrected past abuses. In particular the Board continues to lead the
region into a multibillion dollar rail-centric strategy with no proof it’s the best alternative,
102
and knowing it doesn’t address the congestion problem. They represent to the FTA that a
proper alternatives analysis has been done, and they continue to shove the project ahead
without making sure the public really understands the key facts about Link’s cost
effectiveness, or the full implications of embarking on a billion dollar-a-year strategy for
extending Link into a regional system.
It would appear that the FTA colludes with Sound Transit in these abuses of the planning
process. For instance, the FTA could have rejected Sound Transit’s reliance on the 1993
FEIS as the study which legitimized the choice of light rail. The FTA could have seen
the obvious faults in that old study as outlined in this report. The FTA could have forced
ST to do a proper alternatives analysis comparing what they now propose to build against
bus and other alternatives. Unfortunately it appears there is little incentive for the FTA
to do so. No bureaucracy wishes to under-spend its budget. So the FTA sets up elaborate
requirements, which keep staffs happily employed and cause agencies like ST to spend
millions on studies. But at the end of the day the FTA doesn’t say no, providing the
politics are right.
As a result Link light rail has nothing to do with solving the region’s transportation
problem, but everything --or almost everything-- to do with money.
Link is quite a clever way to transfer money from the pockets of many to the pockets of a
few. Not enough is taken from the pockets of the average taxpayer (in the form of sales
tax and auto registration tax) to cause him or her to ‘fight city hall’, while enough goes to
those relative few on the receiving end to ensure their behind-the-scenes pressure to keep
the project going. Had it been proposed to fund Link by raising the B&O tax on
businesses, the pain would have been focused on those with sufficient muscle to protest;
and there is no doubt Sound Transit’s plans would have been scrutinized much more
closely.
Another reason Sound Transit’s a clever mechanism for wealth transfer is that the
taxpayers have practically no way to stop it. The State Legislature created Sound Transit
as a legal entity, but they forgot to make the Sound Transit Board directly elected for the
sole purpose of solving transportation problems, and they forgot to give voters in the
Sound Transit taxing district any practical control over Sound Transit via initiative.
Instead Sound Transit Board members, such as the recent chair, King County Executive
Ron Sims, are elected to other offices wherein transportation is only one of their many
concerns. As a result the only way local voters can control Sound Transit is by mounting
a state-wide initiative. This self confidence in its own untouchability was probably one
reason Sound Transit’s attorney had the hubris to say in court that Sound Transit
recognized no limits on how much it could spend or how long it could take building light
rail.
“Sound Transit says voters set no money limit Facing legal challenge, agency cites brief reference
on ballot. Sound Transit argued in court yesterday that it has the power to collect taxes for as many
years as necessary, without a spending limit, to complete the entire light rail line voters approved in
1996.” (Mike Lindblom, Seattle Times, Sept. 28, 2002)
103
It has become increasingly clear that voters don’t trust Sound Transit and Sound Transit
doesn’t trust voters. Voters overwhelmingly voted to rescind Sound Transit’s tax on
vehicle registration, but ST still fights to overturn that expression of public will in the
courts. ST changes the project greatly from what voters approved in 1996 but vigorously
fought a lawsuit that would have put the much altered plan back on the ballot for a vote
of confidence.
Additionally, ST wants more money. This was foreseen and the legal mechanism already
exists. It calls for ST to ask voters for an increase in the sales tax. However, while
claiming they have a public mandate and public support for light rail, ST is clearly
deathly afraid to go back to the voters.
Thus ST supporters have resorted to political strong-arm tactics to force light rail into a
package of regional projects (the RTID), which was originally intended for the specific
purpose of reducing congestion.
…King County Councilman Dwight Pelz, D-Seattle, a Sound Transit board member, has lobbied to
include $1 billion for light rail in a possible regional transportation ballot measure.” (Mike Lindblom,
Seattle Times, March 10, 2003)
As a result the following was placed on the November 2004 ballot:
King County Local Voters Pamphlet
Nov. 2, 2004 General and Special Elections
Advisory Measure No. 1
ADVISORY MEASURE NO. 1
LOCALLY FUNDED TRANSPORTATION PLAN
The King County Council passed Ordinance No. 14995 concerning an advisory
measure on a locally funded transportation plan. This advisory measure asks
whether the voters in King County support development and placement on the
ballot in 2005 of a locally funded transportation plan designed to relieve traffic
congestion and increase safety through a mix of road and transit projects in King
County in the Interstate 405 and State Route 509, 522, 167 and 99 corridors,
including replacing the Evergreen Point Floating Bridge, beginning to replace the
Alaskan Way Viaduct, and extending light rail to SeaTac Airport and the University
District. Do you want a locally funded transportation plan to relieve congestion and
increase safety to be developed and placed on the ballot in 2005?
YES
NO
Source:
http://www.metrokc.gov/elections/2004nov/pamphlet/measures/kcadvisory1.htm,
underlining by author
104
Again, note the contrast between this ballot measure and what David Earling, then chair
of the Sound Transit board, acknowledged publicly on December 26 of 2000:
Light rail: There will never be a better time Richard Harkness got one thing right in a guest
column (The Times, Dec 22): Light rail will not ease traffic congestion. Yes, that’s a fact. (Seattle
Times, Dec 26, 2000)
Or the following statement hidden back in the Transportation Technical Report for
Central Link:
Therefore, based on traffic forecasts, the Link system will not result in a significant difference in
traffic volumes. (Ref 23: p. 95)
The fact that Sound Transit supporters forced light rail into the RTID package knowing it
would do virtually nothing to reduce traffic congestion, and that Sound Transit approved
the ballot measure’s wording implying that light rail will reduce traffic congestion, is a
remarkable example of cynical manipulation and deception.
What Sound Transit management understands is the power of the “big lie”. If something
is said often enough by those in positions of authority many will believe it, whether it’s
true or not. As voters marked the ballot above, few would have read or now remember
Dave Earling’s Op ED of December 2000, and far fewer still would have discovered
those places in Sound Transit’s EIS’s which state that light rail will have minimal effect
on traffic. In short, Sound Transit will fool most of the people, and for their purposes
that’s good enough.
There are two reasons why all this background is relevant. First, if we are to fix a broken
process we need to understand not only how that process is broken, but why it is broken.
That is, we need to understand who the decision makers are and what their motivations
might be. Second, before any remedy can be applied successfully we must deal with the
issue of trust. It is pointless to assume that a set of players --who have abused the process
in the past and who now lack the public’s trust-- will simply behave better in future.
One final comment is appropriate even if it seems redundant. It is clear that Sound
Transit’s board does not place a high priority on finding the lowest cost ways to improve
transportation in the Puget Sound Region. In other words achieving the “most bang for
the taxpayer buck” is not of great concern to that group. It is, however, one of the
public’s highest priorities. Thus, there is a fundamental disconnect between what the
Sound Transit board, and the PSRC board for that matter, see as their highest priority and
what the public wants emphasized.
This is a fairly harsh criticism but it seems justified. If the Sound Transit Board really
wanted cost-effective solutions they would have scrutinized the 1993 FEIS in the way the
author has done and probably chosen BRT, since it appears BRT would save billions.
And after finding that Link’s costs had been underestimated and needed to be increased
by 44%, they would have revisited their choice of rail, rather than simply looked for
additional money. They would have published and agonized over, rather than hidden, the
high cost-per-rider data for Sounder and Link. They would have calculated the
105
approximate cost of fully implementing their current rail centric strategy and compared
that against BRT. These are the minimum things a board really concerned with spending
tax dollars wisely would have done.
The Puget Sound Governmental Council (PSRC) must also take some blame. If that
organization were sincerely concerned with getting the most bang for the taxpayer buck it
would have eagerly embraced and practiced Least Cost Planning as State Law requires
rather than trying so hard to avoid it. It would have analyzed Sound Transit’s light rail
plans and ensured they were cost-effective relative to other alternatives before simply
downloading them intact into the Metropolitan Transportation Plan. It would be
considering van and car pooling as a major alternative to investing billions in mass transit
and highways, rather than just as window-dressing attached to the main transit and
highway alternatives.
Sound Transit’s abuses are apparently not unique. In an excellent article in the Journal of
Business and Professional Ethics, Martin Wachs says:
“Forecasts are presented to the public as the results of unbiased scientific procedures yet they are in
reality often highly subjective exercises in advocacy. Professionals who must prepare forecast are
frequently confused by the mixed signals which they get. According to law, and in the eyes of the
public, their forecasts are expected to provide analyses aimed at clarifying choices among courses of
action. But their direct superiors and clients expect them to produce forecasts which will become part
of the supporting documentation justifying a course of action which had already been chosen for
political reasons….
…Public policy rewards this dilemma by requiring through laws and regulation forecasts which are
supposedly neutral and objective, while distributing political reward in the form of grants, contracts,
promotions, and recognition to those whose forecasts are effective at proving their agencies political
position most emphatically.
…Investigation into the bankruptcy of the Washington Public Power Supply System, for example,
shows that forecasts which failed to materialize of low nuclear power costs and a growing demand for
power were prepared at the behest of corporate managers who insisted that the forecasts be optimistic
in order to influence investor and government officials. Technical experts who argued that the
forecasts were misleading were told to be silent or leave the project….
Similar experience is now being amassed with respect to urban rail transit systems. …
I have interviewed public officials, consultants and planners who have been involved in these transit
planning cases and I am absolutely convinced that the cost overruns and patronage overestimates
were not the result of technical errors, honest mistakes, or inadequate methods. In case after case
planners, engineers, and economists have told me that they have had to “revise” their forecasts many
times because they failed to satisfy their superiors. The forecasts had to be “cooked” in order to
produce numbers which were dramatic enough to gain federal support for the projects whether or not
they could be fully justified on technical grounds. One young planner tearfully explained to me that
an elected county supervisor had asked her to estimate the patronage of a possible extension of a light
rail (streetcar) line to the downtown Amtrack station. When she carefully estimated that the route
might carry two to three thousand passengers per day, the supervisor directed her to redo her
calculations in order to show that the route would carry twelve to fifteen thousand rider per day
because he though that number necessary to justify a federal grant for system construction. When she
refused he asked her superior to remove her from the project and to get someone else to revise” her
estimates.” (Ref 24)
106
7.2 The Importance of Trust
The role that trust, or lack thereof, plays in reaching any solution for Puget Sound’s
transportation problems is hard to overstate. Sound Transit’s refusal to conduct a proper
alternatives analysis, and give voters objective information about Link and alternatives to
Link, is one major factor that contributes to this mistrust. The story about trust is best
told through an extended series of quotes, mostly from local newspapers:
On Sept. 6, a group of 88 prominent citizens –including Booth Gardner, Dorothy Bullitt, Matt Griffin,
former King County Assessor Norwood Brooks, and a dozen elected municipal and county council
members – called for an independent review to get a realistic estimate of the cost of the project.
Within hours, their appeal was summarily and rashly rejected by the Sound Transit board. (Emory
Bundy, Op Ed, Seattle Times, Dec 12, 2000)
_______________________
Sound Transit: a matter of trust The 10-year plan for increasing transportation system capacity in
the Central Puget Sound area was dubbed ”Sound Move” by its creators at the regional transportation
authority, Sound Transit. …
….Half of those ten years are now past, and the Sound Move plan has fallen well short of its billing.
The most visible culprit of course is the light rail project, which is $1 billion over budget, three years
behind schedule and the subject of a federal audit.
…It’s the sort of indecision that makes one wonder, despite the agency’s official denial, just how
much of the 10-year plan was completed in a vacuum, without input from the very people is meant to
serve. It’s the kind of day to day waffling and mismanagement that wastes time, overruns budgets
and over time, causes people to lose trust. Some of us in the Legislature have noticed the loss of trust
in Sound Transit.
… Why should taxpayers support long tem financial commitment to transportation when Sound
Transit provides such a convenient example of a commitment gone sour?
…The second offers revote on the grounds that the Sound Transit board’s actions have significantly
altered the proposition citizens approve in 1996.
…Because of Sound Transit’s lackluster performance so far, any taxes directed toward transit and
transportation projects from her on out need to be the best spent money in state government.
….At the heart of the public trust , President Abraham Lincoln once wrote, is trusting the public.
Government should not be – and cannot afford to be- afraid of letting the people judge how well their
money is being spent. Lets. Vote. (Op Ed by 15 members of the State Legislature, Seattle Times,
Feb. 20, 2001)
____________________
Light-rail critics numbers growing …The Bellevue Downtown Association has piled on the
growing ranks of light-rail critics. … ‘Frankly we have no confidence the light rail plan will provide
meaningful help to the regional mobility crisis’, the association wrote in a letter to Dave Earling,
chairman of the Sound Transit board. The letter, released yesterday comes only a few weeks after a
similar dispatch sent to Earling by the Downtown Seattle Association. And the Greater Seattle
Chamber of Commerce, while not directly opposing light rail, said in January it has ‘grave concerns’
about the project. In addition, two former governors, Booth Gardner and John Spellman, and a large
coalition of elected officials and community leaders oppose light rail. Earling says he’s tired of all
the griping. (Andrew Garber, Seattle Times, Feb. 27, 2001)
__________________
107
Light rail labeled ‘problem’ project Key congressional panel to scrutinize program. … The
agency has come under increasing attack by business groups, community leaders and even its own
citizen oversight panel since it revealed in December that the 21-mile rail project is $1 billion over
budget and three years behind schedule. (Andrew Garber, Seattle Times, March 10, 2001)
________________
Despite dying dot.coms, a sagging stock market and a national economy on the verge or recession,
Sound Transit is banking on unending good times to pay for its $3.8 billion light rail project.
A close look at the agency budget show it is counting on a recession-free economy, low interest rates,
stable costs, throngs of transit riders and federal largesse that would set a record. …..
…It clearly a house of cards says Robb McKenna, vice chairman of Sound Transit’s finance
committee and a critic of light rail. ‘The assumptions are unrealistic.’
…Richard Steinmann, a top Federal Transit Administration official, says Sound Transit would set a
record if it gets all the money it is asking for, and ‘nobody has pulled that off.’ (Andrew Garber,
Seattle Times, March 27, 2001)
_________________
Sound Transit troubles, Day 2 Sound Transit’s financial woes got $50 million worse yesterday.
The agency learned that amount of money on top of an additional $75 million in federal funding is
being withheld because of concerns raised by the US Inspector General’s Office. The hold …could
delay construction… and ultimately make the $4.1 billion venture –already the most expensive in the
country-- even more costly. … Metropolitan King Councilman Rob McKenna, vice chairman of the
Sound Transit finance committee and a light rail critic, says the agency and its board of directors need
to start taking the problems more seriously. ‘There is no introspection here’, he said yesterday. ‘There
is just spin control’. (Andrew Garber, Seattle Times, April 6, 2001)
______________
Note to politicians and transit planners: The analogy of “silver buckshot” is an excellent description
of how solving regional transportation problems must be addressed. But even silver buckshot won’t
work if out incumbent politicians and the Sound Transit Board don’t quit dishing out “silver b.s.” …
In light of the dynamic growth in King, Pierce and Snohomish counties, how can we trust Sound
Transit to produce a practical system … Can we trust Sound Transit management to have the correct
answer? Or do we need a flexible, expandable, affordable transit network? It is not too late to
abandon the light rail proposal. (Tobey Wilkins, Soapbox, Seattle Post Intelligencer, June 30, 2001)
____________________
Legacy Time: Congestion is the No. 1 Issue Several lawmakers preparing for this weeks triple
overtime session in Olympia say if the states daunting transportation problems were so easy to tackle
they would already have been solved.
Legislators are letting themselves off easy. For too long they have been afraid to level with the public
about transportation.
… Voters do not elect lawmakers to spend entire terms calculating their re-election chances. Between
elections they are supposed to do the public’s work. (Editorials The Newspapers View, Seattle
Times, July 16, 2001)
_______________________
A near majority (48 percent) of King County voters now has an unfavorable opinion of Sound
Transit, compared to a one-third minority with a favorable opinion. (December 2001 King County
Transportation Poll, by Olympic Institute)
108
____________
Circumstances today present a major opportunity for someone to provide the bold leadership our
community so desperately needs. That someone is Ron Sims.
As both king county executive and chairman of the Sound Transit board, Sims is in the strongest
position to address the fact that the Link light rail project as promised to voters cannot be done.
He should fully disclose Link Light Rail’s current fiscal condition and realistically address its
limitations.
Sound Transit’s own data, particularly its 2002 financial plan produces no evidence that the agency
can fund any extensions of the initial 14-mile segment of light rail either north or south, in the next 20
years. By Sound Transit’s own estimates after shelling out $2.9 billion for this “initial” segment,
we’ll be left stranded in Rainier Valley completely out of gas. For at least two decades.
… At every turn, Sound Transit has failed at the Link light rail effort. Sure everyone is feeling
desperate to “do something” about transportation and congestion but throwing good money after bad
is not just “doing something”, it’s doing something dumb.
… The defeat of referendum 51 shows that the public is seeking smarter transportation solutions that
will use public resources prudently.
… We must stop before taking on such an obligation because if the 14 miles south from the state
Convention and Trade center is all that will be built, why bother at all? We should, instead, free up
those financial resources to make way for a better solution, one that will have an honest, significant
impact on reducing traffic congestion. Sims could lead such an effort. (Booth Gardner, former
governor, State of Washington, Op Ed, Seattle Times, Nov. 14, 2002)
___________________
Sound Transit, an agency in search of credibility, hurt itself anew in the way it picked former Mayor
Norm Rice to head an independent review. Rather than consult broadly with critics or even the entire
board, a small group at the agency selected Rice, a former board member who is regarded as a
defender of Sound Transit’s overall direction.
…Huge budget changes and revisions to the plan voters approved in 1996 require an honest
reassessment of the fundamental rationale for the light rail program. (Editorials: The Newspaper’s
View, Seattle Times)
__________________
Cranky state reeling from leadership void The Good Ship Washington is leaderless, rudderless,
adrift. The one thing voters agree on is we are cranky --cranky at Sound Transit-- , miffed about
sweeping transportation proposals, unwilling to buy almost anything our leaders try to sell…. With R-
51 down, and I-776 potentially cutting into road and transit funding, the message on transportation is:
Don’t do much; we don’t trust you to do it correctly. (Joni Balter, Seattle Times, Nov. 10, 2002)
_____________
That’s similar to what happened in Washington State where polls indicated that the main issue for
voters opposed to Referendum 51, the transportation package, was their lack of trust in the
government. (Susan Gilmore, Seattle Times, Dec. 9, 2002)
_________________
State’s transportation problems loom large over legislative session Before they can find money
to fix roads and build new ones, may political leaders say they have to repair voters’ trust. Polls
indicate a major reason Ref 51 --which included a 9-cent-a-gallon gas tax-- went down is because
voters don’t believe government spends their tax dollars wisely. (Seattle Times, Jan. 12, 2003)
109
___________________
Sound Transit’s difficulty is not only the management of the project or the Sound Transit Board’s
reluctance to pursue accountability. The project was flawed from its very inception
The projects history is typical of many public projects where utopian ambitions fail to acknowledge
contingent realities of place. Seattle’s unique topography and settlement pattern have created
technical and fiscal difficulties due to the deep and potentially unsafe tunnels and costly property
acquisition necessary for the project. More fundamentally, the revelation that the Sound Transit plan
does not expect to relieve traffic congestion raises the question: What, then is the Sound Transit lightrail
plan about? (Folke Nyberg, OpEd, Seattle Times)
_________________
State Rep. Glenn Anderson, R-Fall City, accused the agency of stonewalling on cost figures until the
legislative session is over. Lawmakers are considering hostile bills to require that Sound Transit
board members be publicly elected, or to put light rail to a public revote. ‘That’s information the
public needs to know’, he said. ‘The impression I’ve gotten from a number of sources is, they know
what it is and they’re afraid to tell people’. (Mike Lindblom, Seattle Times, March 10, 2003)
________________
8th District Democrats Gather … Former radio talk show host Dave Ross, the third Democrat in
the field, had a more general goal in mind: to build people’s trust in government. ‘One of the things I
hope to do is not just be your representative but also be your eyes and ears into what’s going on,
who’s making he trades, where the money is going and whether it’s being wisely spent,’ he said.
(Warren Cornwall, Seattle Times, 2004)
________________
Initiative 776 is one particular instance where voters approved a measure Sound Transit
didn’t like. Rather than accept the will of the people, Sound Transit chose to fight the
issue on legal grounds. They may yet prevail on narrow legal grounds, but if so that
victory will simply erode public trust further.
What’s the plan for light rail now? Sound Transit light rail is in trouble. The overwhelming
success of Initiative 776 in Tuesday’s election is yet another watershed in the agency’s difficult battle
to build a regional light rail system. …Where will the agency and its governing board, chaired by
King County Executive Ron Sims, go from here to protect the light rail project? The first step is to
challenge the constitutionality of I-776… (P-I Opinion, Seattle Post-Intelligencer, Nov. 10, 2002)
____________________
Regarding the passage of Initiative 776, ‘the department of transportation told Sound Transit in
September that it would continue to assess and collect the tax --even if I-776 passed-- unless a court
ordered it to stop’.
Is there any wonder why people have no trust in their elected officials? They have stated explicitly
that they don’t care what the results of the election were, and would not respect the will of the
electorate unless forced by lawsuit. Do the result of an election have no meaning whatsoever?
The hubris of the elected officials in this region is astonishing. How can politicians expect to be
perceived as being accountable when they so blatantly demonstrate the exact opposite behavior?
(Randy Blaylock, letter to editor, Seattle Times, Nov. 14, 2002)
________________
110
Why doesn’t anything stick anymore? That’s what people are asking as the thunder from the
election debacle rolls over the state. Anger, palpable and measurable, followed the news that King
County government sold some bonds to pay for Sound Transit just ahead of the Nov. 5 vote on that
very issue. (James Vesely, Seattle Times, Dec 8, 2002)
_________________
How many no’s will it take? Our local politicians again appear to be clueless when it comes to taxes
and the voters ‘mixed messages’. To the contrary, our message is loud and crystal clear: “Tell us
where the money is going. Assure us every single agency has a performance audit. Move to a zerobased
budget and stop trying to hide who gets what”. We are simply telling our elected leaders that
we don’t trust what they are doing with our money. (Scott Carty, letter to editor, Seattle Times,
Dec 14, 2002)
____________________
After Referendum 51 the only choice is a better plan grounded in a higher level of public trust (Doug
MacDonald, Secretary, Washington State Department of Transportation In Ref 54)
____________________
The state Supreme Court upheld the tax cutting Initiative 776 yesterday but leaders of Sound Transit
said they expected to keep collecting its car tab taxes for another quarter-century.
Sound Transit officials say they believe they can keep collection car taxes because yesterday’s ruling
did not order them to stop
Chief counsel Desmond Brown said Sound Transit will sell more bonds backed by the taxes. (Mike
Lindblom, Seattle Times, Oct. 31, 2003)
__________________
Sound Transit should live with the law Initiative 776 is now the law. The law says Sound
Transit’s 0.3 percent car-tabs tax is repealed….It cannot mean that Sound Transit can go on collecting
the whole tax for the full 25 years, collecting several times the amount needed to cover the bonds, and
spending the money on anything it likes. Nor can it mean the Sound Transit can pledge the car-tabs
tax to a new group of bonds to keep the game going.
Sound Transit attorney, Desmond Brown, said Thursday that his client could sell a new batch of
bonds in just that way.
…The counties of King,…., all of which lost taxes to I-776, have to live with the law. So does Sound
Transit. (Editorials: The Newspaper’s View, Seattle Times)
7.3 Requests that Sound Transit consider alternatives
There have been numerous requests that Sound Transit reconsider its light rail plans, as
the following quotes confirm:
The Sound Transit board is struggling with its next move, but is handicapping itself in its eagerness
to keep its $500 million federal grant for Link light rail. It would be better to give that up, take a
fresh look, and go back next year with a defensible plan.
….Sound Transit has spent million of dollars over the past five years saying and implying that Link
light rail is an answer to congestion, a response to congestion, will fix congestion, zap congestion,
provide an alternative to congestion, bypass congestion, and replace the need for 12 lanes of
freeway. That’s the basis for its public support. But, as admitted recently by Sound Transit’s
executive director, Link won’t solve, or even reduce congestion (Times, Oct. 5). It’s a waste of
money.
111
….When its board reconvenes Thursday, Sound Transit can plunge ahead into a troubled future, or it
can give up on the federal grant for this year, go back to the drawing boards, and put together a
better plan. (Emory Bundy, Op Ed, Seattle Times, Dec. 12, 2000)
_____________________
Halt Sound Transit’s Ever More Costly Plans One billion dollars. That’s how much Sound
Transit’s light rail project has gone up in price. Taxpayers from Everett to Tacoma are expected to
pay.
….Light rail was supposed to be the backbone of a system that also includes regional buses. That
concept is now very much in doubt. Sound Transit’s directors need to compare the repriced light rail
with an all-bus alternative on a backbone of dedicated road lanes. (Editorials The Newspapers
View, Seattle Times, Dec. 14, 2000)
____________________
Full speed ahead for light rail In a hasty attempt to secure $500 million in federal funding before
the Clinton administration steps down, Sound Transit will move forward with light rail despite
mounting objections about costs, the agency’s leaders say.
…But Sound Transit Executive Director Bob White and board Chairman Dave Earling say they
already know what the board will decide on that pivotal day: The agency will not explore
alternatives to light rail… (Chris McGann, Seattle P-I, Jan. 5, 2001)
______________________
Ron Sims’ free ride ….Sims is a very strong executive.
…Why does he have so much confidence in Sound Transit’s light rail ? As Sound Transit struggles
to find the right configuration for its project, when is he going to offer a smarter alternative?
(Opinion, Seattle Times, March 25, 2001)
______________________
Sound Transit Board: It’s time to do your job After the overwhelmingly negative wave of recent
events, you would think a board that calls itself “Sound” would stop shelling out our money to move
forward on such a monumental undertaking as a $4.2 billion light rail plan. Your would think they
would call a timeout not just to patch up the holes that have been revealed, but a timeout to actually
rethink whether this flowed vessel, light rail, is going to get us to the vital goal of reducing traffic
congestion in central Puget Sound.
Yet I response to the damaging report issued last week by the US Inspector General’s office… all we
get here at home is more patch-up.
… My response is this: Is there any event or combination of facts, any misgivings about cost,
funding, ridership, or concerns over the Inspector General’s criticisms, anything at all that would
finally cause the Sound Transit Board to ask, “Is proceeding with light rail still a good idea?” It
seems no issue exists that’s significant enough to prompt the board’s serious review of alternative
solutions.
…The board’s irresponsibility is found in the decision to remain silent despite their growing
awareness of that misinformation, including possession of significant evidence that light rail may not
be a cost-effective transit alternative.
Civic groups, critics and other elected officials are calling ever more loudly for a complete review of
the project including available alternatives… (Booth Gardner, Op Ed Seattle Times, April 11,
2001)
112
Sound Transit could use some healthy skepticism …But Sound Transit was created using
misleading projections to win its enabling election. If we may finally treat the immaculate status of
the agency with some good, healthy skepticism, let us now see if it might reduce its ongoing expenses
by sharply reducing its hordes of consultants, spinmeisters, attorneys (is it necessary to have all the
big Puget Sound law firms on retainer?), $200-per-hour “temporary” big wigs and other recipients of
political sinecures.
Then once it’s down to essentials, let it (or better, let some independent investigator) furnish the
public with realistic appraisals of the costs and benefits of light rail… Let it look at alternatives….
(Hank Bradley, letter to editor, Seattle Times, April 15, 2001)
____________________
The hidden election over Sound Transit ….When challenged, we say we know the 14-mile
streetcar is fairly ridiculous, but you have to start somewhere, With other $2-billion, $4-billion, $6-
billion wads of money, paid by generous taxpayers somewhere, we shall build a vast system that
someday goes where we want to go. “It’s not for us,” we say. “it’s for the children. They will thank
us for it.”
They won’t. We will build far less than we promise, because we cannot afford it. Our children will
look at our pitiful production and laugh at us: “you expect us to use this?”
And we will say, “but we spent billions on it”.
And they will not care.
There is still time to prevent this embarrassment. Light rail may not be listed on the ballot, but if you
look for it, it is there. It’s time to till it, and get serious about things that will actually work: buses,
van, carpools and --dare we say it—roads. (Bruce Ramsey, Seattle Times, Oct. 24, 2001)
_____________________
Before tunneling begins, the region expects honest answers abut cost, benefits and alternatives. If
Plan A is the best approach, it will withstand the scrutiny. If a smarter Plan B exists, give it an airing,
and equal scrutiny. (Editorials The Newspapers View, Seattle Times)
____________________
Its time to consider alternatives to light rail, which is going nowhere fast Sound Transit has revealed
that it cannot fulfill its key promise to voters…. So what now for the deeply flawed Link light rail
plan? …We need a Plan B worthy of the innovative and imaginative people who live here; however,
we don’t know what that is yet because Sound Transit refuses to seriously consider alternatives to
light rail. … The elected officials who make up the Sound Transit board pronounce that the voters
wanted light rail and that is what they are going to get. In my opinion, what voters want is for you to
use your best judgment and improve the traffic situation –whatever it takes. (Booth Gardner, TNT
6/21 newspaper article)
Sound Transit may argue that this is all old news, and that after a number of personnel
changes they are now a “reformed agency”. However, their recent attempts to push Link
forward using strong-arm tactics and/or deceptive practices (e.g.: RTID ballot, I-90 center
lane preemption, and legalistic response to I-776), plus their continued refusal to
seriously consider alternatives, all indicate this isn’t true. Sound Transit has still not
leveled with the voters, nor have they regained voter trust.
113
7.4 Key recommendations of this report
The following actions should be taken at this time:
1) The FTA should withhold any additional money for Link until a proper
alternatives analysis has been completed, its results fully communicated to the
public, and a public vote of confidence confirms voter support for continuation
of Sound Transit’s light rail strategy. Part 8 of this report outlines key
requirements for conducting a proper alternatives analysis.
2) The Sound Transit Board of Directors should voluntarily undertake the
above actions with or without the FTA requiring same. They should do this to
confirm they are on the right path and to restore public trust.
3) Congress should investigate the manner in which mass transit grants are
approved to ensure that FTA controlled planning processes are not abused in
the ways chronicled throughout this report. They should investigate Sound
Transit as one case example of abuse. Congress should insure that the process
is redesigned to obtain maximum transportation improvement at the least cost
to Federal and local taxpayers.
The Sound Transit Board has no good reason to balk at taking these actions. If Board
members are confident they are on the right path, a proper alternatives analysis can do
nothing but confirm it, silence the critics, and help restore public trust. If they are not
sure, then they should welcome having the better information that a proper alternatives
analysis could provide, welcome the opportunity to reaffirm public support, and welcome
the opportunity to best serve the region by changing course if necessary. There is no
hurry to build Link since it won’t make any difference in traffic congestion, and since the
rights-of-way are not disappearing.
Vetting-- Although this report finds Sound Transit’s rail centric strategy unjustified, the
author does not feel that this report is sufficient in its current state to justify halting
construction of Link. Instead the author feels that authorities should immediately cause
this report to be reviewed by an objective team of experts to confirm or refute the logic,
calculations and conclusions herein. This could be done in about six weeks. If those are
upheld by the team of experts then there would exist reasonably credible, but still not
conclusive, evidence that an all-bus strategy could achieve much the same benefits as
light rail, and do so at a far lower cost. Such vetting would justify temporally halting
construction on Link until the full-fledged alternatives analysis is complete in 12 to 18
months. This may mean paying millions in contract delay or termination fees, but this is
small change compared to the billions that would be wasted pursuing the wrong strategy.
Finally, Link could be restarted if the full-fledged analysis is favorable, or terminated if it
is not.
114
7.5 How accurate does this report need to be to justify the above recommendations?
The author recognizes the import and controversial nature of the above recommendations.
It is appropriate to make a kind of reality check to see if the key findings of this report are
accurate enough to justify taking the actions recommended. A first cut at this follows,
with hope it will soon be followed by expert team review.
7.5.1 How accurate is the author’s finding that an all-bus alternative would have
adequate capacity and could attract as many riders as a mature light rail system?
The author did not simply assert that a mature all-bus alternative would be able to attract
the same ridership as a mature rail alternative. This conclusion is based on the RTP’s
own ridership forecasts for the rail and bus alternatives, which the author believes are
probably correct, at least one relative to the other. Cheating at this level is not likely, but
if there were any it would have been designed to overstate the rail ridership and
understate the bus ridership. However, the only way the conclusions of this report would
be invalid is if RTP did exactly the opposite by underestimating rail ridership and/or
overestimating bus ridership.
The 1993 FEIS found that the all-bus TSM alternative, as it was designed, would attract
518,000 daily riders or 92% of the 560,000 attracted by 125-miles of rapid rail. This is so
close that there is little doubt that ridership on the TSM alternative could be bumped up
to equal rail ridership just by doing simple things like adding routes or making the buses
come more frequently. Additionally, light rail is slower than rapid rail and is unlikely to
achieve the same ridership. Thus there is little doubt that a bus alternative can be
designed to achieve the same ridership as a light rail system, even a fully mature one.
But would it also have sufficient capacity to serve this ridership?
As Part 5 explained in detail, a properly designed bus alternative would have sufficient
capacity to carry its full predicted ridership. The only place capacity could become a
problem is in the downtown Seattle transit tunnel. However, a range of remedies was
identified for that potential problem. If one remedy doesn’t solve the problem the next
could. First, the tunnel buses should be managed so as to achieve full capacity in the
tunnel. If that doesn’t solve the problem additional buses could be placed on surface
streets according to schemes Metro has already identified. If that isn’t enough then a
through HOV lane could be constructed on I-5 allowing buses not actually needing to
stop downtown to bypass the DSTT. If that doesn’t suffice a busway across south Lake
Union and along the waterfront could handle some of the load. And if none of that is
enough a second bus tunnel could be built. It seems there are remedies in depth. In
addition, there is evidence that the absolute level of ridership may be lower than
originally expected making the whole capacity issue largely irrelevant.
In short, if one believes the RTP ridership modeling exercises were reasonably accurate,
at least in a relative sense, then there seems little doubt that a properly designed bus
alternative could attract, and carry, the same number of transit riders as light rail, even a
fully mature 100+ mile version of light rail.
115
7.5.2 How accurate is the author’s finding that an all-bus alternative would be
much less expensive that a light rail system attracting the same ridership? And how
accurate does it need to be?
The key findings in this report ---that light rail is far more expensive that an all-bus
alternative attracting the same ridership, and that ST’s rail centric strategy would
probably cost taxpayers almost a billion dollars per year more than a bus centric
strategy—are based directly on cost estimates in the 1993 FEIS and related documents.
If the RTP got those numbers wrong then the cost estimates in this report will also be off.
But they would need to be very far off to invalidate the author’s general conclusions.
Again, the author is inclined to accept the cost estimates in the 1993 FEIS as being
honestly arrived at and reasonably accurate given what the engineers knew at the time.
And if there was bias, such as inflating the cost of the TSM alternative by including such
things as improving the waterfront trolley –which was done-, it probably acted to make
the TSM alternative more expensive than it should have been, and/or make the Rail/TSM
alternative less expensive than it should have been. In other words the real cost
difference between the bus and rail alternatives would –if RTP cheated—be even greater
that shown in this report. This report is therefore safe, if not indeed conservative, in
accepting the RTP cost estimates as being accurate for the time. If the reader is willing
to accept the RTP’s cost estimates, there is only one other place for error. That is in the
way the author adjusted those costs to produce an estimate of what a fully mature version
of Link might cost, and what an extended version of RTP’s original bus alternative might
cost. The results all came together in Figure 6.3a.
The author began by assuming a fully built out version of Link would have about the
same size network as the 125-mile rapid rail system in the 1993 FEIS, would follow
about the same routes, involve about the same amount of tunneling, have about the same
station designs, use the same trains and therefore cost about the same as the rapid rail
system in the FEIS. The only significant adjustment the author made to RTP’s cost
estimate was to correct for a probable underestimation of rail construction costs. Thus he
escalated the cost of rail by 44% to reflect the difference between what Sound Transit
originally said light rail would cost and what contractor bids showed it would actually
cost. But just because Sound Transit got their cost estimates wrong it does not
automatically mean the RTP cost estimates were too low, however it is a strong
indication they were. Still the author may have erred in applying this 44% so a reality
check is in order.
Here are two sanity checks: First, the plot-points in Figure 6.3a show that Central Link
and Link IS are both costing more than the rail trend line (cost vs. benefit line) in that
figure. If a trend line were plotted through either of those plot-points and extended out to
show the cost of a fully mature light rail system it would go up at a steeper slope and
show a cost differential even greater than $900 million per year.
The second check is to look at the cost per mile for Central Link and Link IS and see how
that compares with the cost per mile of the rail trend line in Figure 6.3a. This is not
116
perfect either, but it’s close enough. Per details in the Tech Notes, Central Link will cost
$ 157 million (02$) per mile, Link IS will cost $ 138 million (02$) per mile, and the rail
trend line in Figure 6.3a is based on a cost of $ 120 million (02$) per mile. Therefore the
rail trend line in Figure 6.3a is probably overly conservative since it assumed a lower per
mile cost than Link is actually experiencing.
The author also made assumptions in calculating the cost of extending the TSM
alternative until it reached rail ridership. This involved making the conservative
assumption that the bus capacity issue was real and that nothing short of a second bus
tunnel would solve it. So the cost of that tunnel was included in the bus trend line in
Figure 6.3a. The cost of that tunnel came directly from an RTP estimate. Likewise the
cost of extending TSM to achieve more ridership was an RTP estimate. Finally, the
author did reduce the cost of the TSM alternative to account for the fact that about half
the HOV lanes were completed in the 1993 to 2004 interim. Since this all relies on RTP
data it is not obvious how the author’s cost estimate for the TSM trend line in Figure 6.3a
could be very far off, unless of course the RTP was wrong.
Finally, one must consider the extreme difference between the cost of the all-bus
alternative and the light rail alternative as shown in Figure 6.3a. It would take a massive
error in the cost estimates to erase such a big difference. It is pointless therefore to be
overly concerned whether the true difference is $900 million per year, $1.3 billion, or
$600 million. It is huge in any case. And it strongly suggests Sound Transit is leading us
in the wrong direction.
In sum, if one assumes RTP cost estimates were anywhere near accurate, the author’s
conclusion that light rail is far more expensive than an all-bus alternative is correct. It
also safe to say that building a mature version of Link would cost taxpayers hundreds of
million of dollars per year more than building a comparable all bus system.
The numbers in Figure 6.3a are large and the message is shocking. There’s an emotional
urge to say ‘it just can’t be true’. If you, the reader, are still in denial, please go back
through every step and every assumption. If you can’t find any errors you may need to
accept the conclusions of this report.
7.6 The funding going to transit appears to be out of balance. The PSRC’s
Metropolitan Transportation Plan calls for about $ 105 billion to be spent on
transportation in this region between now and the year 2030. Of that, $45 billion is for
mass transit: $19.4 billion for HCT plus $25.4 billion for local transit. (Ref 61: Table 7)
The HCT being 125-miles of light rail. Thus about 45% of total transportation dollars are
planned for mass transit. However only 2.8% of trips in 1998 were made on transit and
only 5.8% would be in 2020 if all the transit improvements envisioned in the MTP were
implemented. (Ref 17: p. xvii) This imbalance is shown graphically in the following
chart. A chart by WSDOT Secretary Doug MacDonald makes much the same point. (It’s
attached as an Addendum)
117
7.7 Does it matter if the region spends billions more on mass transit than it needs
to?
This report concludes that pursuit of Sound Transit’s light rail strategy could end up
costing the region roughly a billion dollars a year more than switching to an all-bus
alternative capable of achieving the same transit ridership and ridership related benefits.
So what? Does anyone care?
The answers are not obvious. It will be interesting to see if anyone gets concerned
enough to actually take action. There are lots of government programs that cost in the
billions. The word “billions” appears in the papers on a daily basis. Ordinary people
have a hard time relating to something that abstract. Perhaps the best way to make the
cost of Sound Transit’s rail plans meaningful is to explain what else could be done with
that amount of money, in other words, what we are giving up in order to build light rail.
For some indications consider the following:
Guiding King County’s budget through the health-care crisis Last month, I unveiled a $3.2
billion budget before the Metropolitan King County Council as a blue print for healthy people, a
healthy environment, and healthy communities. …
… Unless we do something to change the status quo, King County can expect a doubling of its health
plan premiums within six years, from $154 million to more than $300 million.
…. Two years ago, we made changes to our health-care plans… as a result we saved $6 million in
2003. (Ron Sims, Op Ed, Seattle Times, Nov. 3, 2004)
_________________
Trips Cost
100% 100%
Transit
6%
Roads
Roads
Transit Transit
45%
118
Transportation needs require forceful action The Blue Ribbon Commission on Transportation has
laid out a blueprint for meeting the goal of moving people and goods efficiently. Building a
transportation system on that blueprint will be expensive and will mean more taxes.
….The commission estimates that the gap between what we need to spend on transportation over the
next 20 years and what we’ve budgeted is about $95 billion –nearly $5 billion per year. (PI Opinion,
Seattle P-I, Dec. 17, 2000)
______________________
Council offer ‘kinder, gentler’ budget Seattle plan raises fees, fines to restore some of mayor’s
cuts
The Seattle City Council today will unveil a spending plan that softens some of the most controversial
budget cuts proposed by Mayor Greg Nickels, restoring money for community health clinics, a Fire
Department medic unit and library books.
… the city faces a $60 million general-fund shortfall between expected tax revenues and the amount
needed…
… The revisions include: $1.6 million in restored funding for community health clinics…Restoring a
medic unity at a cost of $800,000 … Spending $1.1 million more on social services, including food
banks…. (Jim Brunner and Bob Young, Seattle Times, Nov. 14, 2002)
____________________
Wanted: Brave Souls with clear ideas …. As if the general fund budget shortfall were not
headache enough, a separate but essential fund, the Health Services Account, also faces a substantial
budget shortage of $500 million or more… (P-I Opinion, Seattle P-I, Sept. 22, 2002)
__________________
School-renovation fund $11 million short A Seattle School District review shows that its schoolrenovation
programs are running deficits that could mean some projects will be delayed, trimmed or
eliminated. (Sanjay Bhatt, Seattle Times, Aug. 4, 2004)
__________________
More money for parks endorsed The city (Bellevue) in 2003 identified $175 million in parks
projects over the next decade that it would like to complete. (Warren Cornwall, Seattle Times, Nov.
2, 2004)
___________________
Special-ed funding prompts lawsuit Lawmakers accused of neglecting duty Benzel said there is
a gap of more than $100 million between what the state pays for special education services each year
and what those services cost districts. (Cara Solomon, Seattle Times, Oct 1, 2004)
_______________________
$878 million more sought by Bergeson for schools Terry Bergeson, state superintendent of public
instruction, yesterday asked for an additional $878 million for public schools over the next two years,
an amount she says is essential to reach the goals of the states decade old education reform law.
(Linda Shaw, Seattle Times, Sept. 23, 2004)
________________________
Lack of Money Hampers Superfund site cleanups Facing a record budget shortfall of about $250
million and about 475 uncompleted sites, the nearly 25-year-old program aimed at protecting
Americans from industrial contamination is in crisis. Program managers are scaling back their
spending requests and slowing cleanups. (Juliet Eilperin, Seattle Times, CloseUp, Nov. 26, 2004)
________________________
119
Financial news gets worse for Seattle schools Insolvency possible, finance chief says (Seattle
Times, Dec. 16, 2004)
_____________________
Locke eyeing big tax boost Gov. Gary Locke, with only a few weeks left in office, is expected to
propose a tax increase of about $500 million tomorrow as a way to soften the blow of a projected $1.5
billion budget deficit. (Andrew Garber, Seattle Times, Dec. 15, 2004)
Just weeks ago Washington State voters rejected a ballot measure that would have
increased the state sales tax by one percentage point to fund statewide educational
improvements. Sound Transit is already collecting a 0.4 percent sales tax, which is
barely enough for the Initial Segment. To extend Link to Northgate, much less extend it
into a full regional system, would obviously require raising the sales taxes going for light
rail to more than the one percentage point sought for education, or raising other taxes
instead.
Light rail is obviously competing with education for scarce taxpayer dollars. Underfunding
education is one of the “opportunity costs” associated with pursuing light rail.
Sound Transit’s 2004 Financial Plan says that $2.437 billion will be spent on Link’s
Initial Segment between 1997 and 2009. In just 2004 Sound Transit will collect $271
million in taxes.
The amount of money going to even the Initial Segment of Link light rail would make a
big dent in the funding needed to reconstruct the Alaska Way viaduct and rebuild the 520
bridge. But the Initial Segment is just the first step in Sound Transit’s ambitions for light
rail. Clearly the $900 million per year difference between embarking on Sound Transit’s
rail-centric vision for mass transit versus switching to an all-bus alternative would be
more than enough to pay for both projects.
It would not be surprising if every environmental program proposed for Puget Sound (all
the bike paths, all the parks, all the stream clean-ups, etc) could easily be paid for using a
small fraction of what Link is costing. Someone should compile a list and find out.
In short, the money that might be wasted on Sound Transit’s rail-centric strategy is not an
abstraction, it comes at the expense of opportunities foregone, and of other good ways to
spend tax dollars.
Nor is the impact on the Federal budget, of questionable projects all across the country
like Link, something to ignore.
Congress lift debt ceiling New borrowing to avert default
Congress last night sent President Bush an $800 billion boost in the federal borrowing limit,
spotlighting how the budget has lurched out of control in recent years and how difficult it will be to
afford future initiatives. … “I want someone to explain to me how it can be moral for a father to stick
his kids with his bills,” said Rep. Gene Taylor, D-Miss. (Seattle Times, Nov. 19, 2004)
120
7.8 Concluding comments on RTA’s 1993 FEIS
The 1993 FEIS is an impressive looking document, about an inch thick and perhaps 300
pages in length. The author has been drawing information from it for years, and has
studied it intensely over the 4 months it has taken to write this report. The 1993 FEIS is
backed by a two or three foot thick stack of technical reports on such things as the cost
buildup for the Rail/TSM alternative and ridership modeling results. Some of these are
used as references in this report.
Presumably the purpose of these EIS’s is to help officials and voters make intelligent
decisions regarding massive investments in public transportation. In this sense they are
decision aids somewhat akin to the “business cases” used in private industry. They look
authoritative and the reader starts with the assumption they are competent and objective.
Unfortunately, after trying to dig useful information from the 1993 FEIS the author has
come to realize the shortcomings of this report as a decision aid.
Although it is quite a brutal thing to say the author has concluded that the 1993 FEIS is
essentially a sham. Yes, it does contain a lot of useful “raw data” kinds of information in
the form of cost and ridership estimates and route alignments. However, rather than
being dedicated toward thoroughly and objectively informing and educating the reader so
he or she can make a wise decision, the 1993 FEIS is apparently dedicated to two
different objectives. One is to satisfy FTA and state laws that require projects of this
nature to have had completed an alternatives analysis. The other is to sell and justify a
preordained conclusion, namely that rail is the better alternative.
The reader not interested in a sales pitch but actually trying to understand the merits of
rail verses bus is forced to piece together bits of information from different places in the
FEIS and resort frequently to technical backup documents not readily available. Much
searching and many calculations are required. The careful reader finds that the FEIS is
most interesting not for what it says, but rather for what it hides or ignores. Thus the
reader will note that none of the charts appearing in this report –as tools to clarify key
findings-- can be found in the 1993 FEIS, nor is the data to plot them found directly in
the report.
As an example of what was hidden, the author was long unaware that the RTP actually
made two ridership estimates for the TSM or bus alternative. The one featured
prominently in the FEIS –and used to estimate all the ridership related benefits of the
alternative- was their so called “constrained” estimate, although that label was not used
and thus did not signal that there might be another. The indication a higher estimate
existed was hidden back in the middle of a paragraph on page 3-93. Discovery of this
fact was the Rosetta stone which allowed the author to unravel how the RTP had fatally
biased this entire study. It was what enabled him to unravel the capacity story in Part 5
and produce all the cost charts in Part 6. Even then the actual value for “unconstrained”
ridership did not appear in the 1993 FEIS. It was finally located in one of the technical
backup reports. (Ref 42)
121
The instances of things not addressed or assertions made without adequate backup are not
obvious to the casual reader, and the author missed many of them until actually starting to
write this report. It is hard to stress this point enough. It is only when the reader really
gets into it, that these problems become evident. But in retrospect they stand out. Some
but by no means all are addressed in this report.
The challenge is to find the adjectives that seem most appropriate for describing the 1993
FEIS. The author has tried to apply them fairly and accurately. They include terms like:
inadequate, irrelevant, and biased. RTA’s creation of a deliberately hobbled bus
alternative goes beyond simple bias. Manipulation and deceit are perhaps the proper
terms.
The word “sham” is used just once in this report, here in this subsection. Websters
defines “sham” as “an imitation or counterfeit purporting to be genuine”. The1993 FEIS
purports to be an objective decision aid containing the information needed to make a wise
decision. It isn’t. Local officials and the FTA should never have accepted it in the first
place. Clearly it does not provide a valid basis for Sound Transit’s selection of light rail
over a bus alternative. Clearly it is not an acceptable basis for Phase 2 planning.
122
Blank page
123
Part 8: Guidelines for a proper alternatives analysis
8.1 General guidelines
Clearly state the main goal- What is the primary problem we are trying to solve? How
can we measure the extent to which any given alternative might solve it? For instance if
the main problem is traffic congestion, then the main goal should be a quantifiable and
relevant metric, such as " reduce average travel time" or "reduce travel delay" as
measured across the Puget Sound area. Obviously there are secondary goals, and they
should be measured as well.
The author recommends that “reduction in travel delay” (measured in hours per average
week day for all transit and auto trips) be the metric used. It would be calculated by
subtracting total daily travel hours (for all daily transit and auto trips) in the build
alternative, from total daily travel hours assuming all trips moved at the posted road
speed limits.
Objectivity-- Obviously great pains must be taken to ensure that the overall alternatives
analysis process is managed objectively by a party with no vested interest in any
particular alternative winning.
Have advocates design the alternatives- Just like in the competition for some new
military aircraft, we want each alterative to be as well designed as it can be. We want the
alternative analysis “fly-off” to compare the best light rail design against the best all-bus
alternative. Thus each alternative should be designed by someone who is skilled and
motivated to produce the best possible design. We know Sound Transit favors light rail
and would produce the best possible light rail alternative. A different party needs to
design the all-bus alternative, and still a different party is needed for any car/van pool,
TDM alternatives.
This paper has made it abundantly clear that an agency which has a bias toward one
particular alternative can design the other alternatives to fail. The author has seen some
of the subtle ways this can be done without attracting much attention. The fact that
Sound Transit is designing the bus alternatives for the Phase 2 planning is very
dangerous. Even if Sound Transit sincerely plans to be objective, it has an apparent
conflict of interest or bias.
Have the alternatives evaluated by a neutral third party- The reasons for this are
obvious. It will probably take some effort to find neutral parties due to the cozy nature of
relationships between transit agencies and consulting firms. We will need to find a
consulting firm that has no close relation with this region. We will need to supplement
their work with help from universities and other independent consultants. The role of the
evaluator is to estimate the degree to which each alternative would achieve the objective,
and to estimate all the direct and indirect costs.
124
Use constant dollars- It is very hard to compare the cost of plans when they are costed in
YOE dollars because the same project can cost different amounts in YOE$ depending on
its construction schedule. The cost of alternatives should be stated in constant dollars
such as 02$ or 04$. This would make it easy to compare the cost figures in the Central
Link FEIS and the 1993 FEIS with those in any new alternatives analysis.
Discount cost and benefits- Both the cost of an alternative, and its primary benefit such
as delay reduction should be stated in “Present Value” terms. In support of this, a
spreadsheet should be delivered as part of the alternatives analysis final report that shows
how much would be spent on the project each year over the next 30 years. That cost
stream should be discounted to a single present value or present cost if you will.
Likewise the benefits, such as hours-of-delay-reduction per year, should be estimated for
each year, tabulated on the spreadsheet, and discounted to a present value. The present
value of the benefits compared with the present value of the costs is the best measure of
cost effectiveness.
The reason for doing this is obvious. The general notion is that a dollar spent later is
worth less than a dollar spent today, and a benefit received later is worth less than a
benefit received today. Sound Transit has been collecting taxes for Link for almost ten
years but construction has just begun and no actual benefits will occur until the line opens
for service in 2009. Thus the bulk of the cost of a rail alternative is incurred years before
any benefits accrue. In contrast, a bus alternative can start providing benefits shortly
after moneys are spent, and at least part of the investment in a bus alternative can be
staged so improvements aren’t made until they are really needed. Thus the cost stream
for an all-bus alternative would not be as much front-loaded as the cost stream for a rail
alternative.
Estimate ALL the costs- Include not only the dollar cost of the project per se but also
costs to the environment, etc. Reduce everything to a dollar equivalent to the extent
possible. Any cost that people feel is relevant should be included.
Consider ALL the alternatives- If the primary objective is to reduce travel delay for
auto and transit users alike, then there is a long list of potential solutions beside building
light rail or BRT. This full range of solutions should ideally be included in the
alternatives analysis. If political realities restrict the new and proper alternatives analysis
called for in this report to just a rail/bus comparison, then some other organization like
the PSRC should be tasked to undertake this broader comparison of alternatives. PSRC
is already mandated to do so as part of their legal requirement to implement Least Cost
Planning.
To be more specific. For every rail+bus, or all-bus alternative considered, one or more
non-transit alternatives should also be considered since they may be more cost-effective
in reducing delay than expanding mass transit. Alternatives that should be considered
include road improvements, car and van pools incentives, and Transportation Demand
Management or TDM. Altogether there are about 50 alternatives that really should be
125
considered, although that means a more through process than officials may be willing to
support at this time.
Communicate the results widely- Describe the alternatives, the degree to which they
achieve the objective, and all direct and indirect costs to everyone with an interest. Do so
with the intent of educating, not selling. This is key to building trust.
Present the cost-effectiveness information in a way that maximizes understanding-
The1993 FEIS presented cost-effectiveness information in a way designed to mask, not
illuminate, rail’s relatively high cost. Clearly this was not objective and did not foster
understanding. A better use of graphics is also suggested.
Never let the alternatives differ in both cost and benefit- This was a major failure of
the 1993 FEIS. In theory, use of cost-effectiveness should make it easy to compare
alternatives that differ in both cost and benefit; however, in practice it is fraught with
chances for confusion, especially in regards to benefits that are not quantified, such as
impact on sprawl.
Design alternatives to be equal in cost, or in primary benefit- The alternatives
analysis should consist of comparing a series of pairs or sets of alternatives. For instance
we should compare the 14-mile Link IS with a bus alternative having the same present
cost. Then we should compare the 21-mile Central Link with a bus alternative having the
same present cost, and so forth until we compare a fully mature 100+ mile version of link
with a bus alternative of equal cost. OR alternately, we should design rail and bus
alternatives so they have the same impact on travel delay, then see how their costs differ.
Define the cost baseline carefully- The question of which costs are sunk and therefore
should not be included in the cost of the “build” alternatives is important. Sound Transit
might argue that multi-year contracts already negotiated or signed constitute sunk costs
and therefore should not be included in the cost of the rail alternatives. A more sensible
approach would be consider as sunk only costs that could not be recovered under contract
termination clauses. The general idea is that rail build costs should include all the costs
which could be avoided if we decided not to continue with rail. The same idea with bus.
In theory, we can get the same overall result, that is some picture that resembles Figure
6.3a, either way. It will take careful thought to see whether design-to-equal-cost, or
design-to-equal-benefit, is the better approach.
Determine the costs and benefits for different degrees of alternative implementation,
not just one arbitrarily sized project- All of RTP’s and Sound Transit’s FEIS’ have
just computed costs and benefits for a single point along what is really a continuum of
project sizes. In other words the 1993 FEIS looked only at a 125-mile rail system, the
Central Link FEIS considered only 21and 24-mile versions of Link. As explained in Part
6 we need to understand the cost benefit picture for a full range of different size rail and
bus systems. There are three primary reasons for this. One is that we are not sure
126
whether the rail and bus cost-benefit lines cross over at some point. One might be the
least expensive way to achieve a modest increase in transit ridership, but not be the least
expensive way to achieve a major increase. (I doubt it, but we need to know). Second,
because of the law of diminishing returns, we need to understand the maximum degree to
which a given alternative might make sense to implement. In other words it might not
make sense to build more than 20 miles of light rail, or is it 100? By understanding the
entire curve we can identify where the law of diminishing returns takes over and the
marginal cost of building more line exceeds the marginal benefit. This theme will be
further amplified in Part 8.2
The third reason is that we need to see where we are going. We are currently embarking
on a light rail-centric strategy with no clear idea where it is leading. Other than for this
report, we have no idea how much it might ultimately cost, and no idea how much it
might ultimately be able to reduce traffic congestion. If voters knew where Sound
Transit’s strategy was leading they might have second thoughts about its affordability and
its ability to really make a difference, or they might vote to accelerate it.
Be careful about what is considered a de-facto starting point-
Nothing should be taken as a fait accompli or irrevocable decision unless that’s
completely true. In particular Sound Transit should not approach the alternative analysis
assuming that Link IS is a done deal and can’t be stopped. Clearly if society wants it
stopped it can be stopped, and no more need be spent on Link than what will already have
been spent when the alternatives analysis is complete and the system is placed to a public
vote of confidence. This will be contentious.
Sound Transit clearly intends to assume Link IS, indeed all of Central Link, is a done
deal. Clearly ST plans to limit any future alternatives analysis to these narrow questions:
Should the HCT link across Like Washington be an extension of Central Link light rail or
BRT? Should the extension from South 200th to Tacoma be light rail or BRT? Should
the extension north from Northgate be light rail or BRT? This is not adequate. The
region needs to determine if even the Initial Segment is justified. This means comparing
the Initial Segment, and Central Link against all-bus alternatives.
Provide cost and benefit metrics that can be related to prior studies-- The FEIS’ for
Central Link and Link IS report costs in YOE$. These are very difficult (for anyone
beside Sound Transit) to convert into constant year dollars so they can be compared with
what was said in, for example, the 1993 FEIS. Some of the EIS’s give ridership for the
entire alternative but not the rail component per se. All the existing EIS’ quantify
ridership and other impact for the years 2010 and 2020. Phase 2 plans to use 2030 but
they must also use 2010 and 2020 or the results will not be comparable with these prior
studies. The PSRC uses “delay” as the main dependent variable. Sound Transit should
also include estimates for impacts on delay. In short there should be a template of cost
and performance metrics, which all studies conducted over the years must fill out in a
consistent manner.
127
8.2 Scenarios and alternatives that should be studied
8.2.1 The concept
Figure 8.2.1 shows what one of the most important outputs of a proper alternatives
analysis should look like. It compares the costs and benefits of following a rail-centric
strategy versus following an all-bus strategy.
Figure 8.2.1 requires explanation. First, it is a chart relating cost to benefit, where cost is
the cost of building and operating a transit system, and benefit is the reduction in travel
delay resulting from that system. As noted earlier both metrics should be in terms of
“present value”.
The solid line represents a rail+ bus strategy where the first step would be building the
14-mile Link IS along with the needed bus feeder lines. It’s similar to the Rail/TSM
alternative in the 1993 FEIS in that rail always needs to be combined with bus. This
could be followed by extending light rail to the 21-mile Central Link, then perhaps to a
60-mile system, and eventually to a 100-mile system. Society would probably choose to
implement this strategy a step at a time by extending Link in stages, but in theory we
could jump directly to any point along the line, such as building a full 100-mile system as
one big project. The dotted line is the all-bus strategy.
Figure 8.2.1
100-mile Link
Link IS
Central Link
60-mile Link
0
0
Cost
(present value)
Reduction in Delay (present value)
Rail + Bus
All-bus
128
Each circle represents an “alternative”, scenario, or project that should be evaluated in an
alternatives analysis. The size of these projects (in route length, cost, or benefit) is
somewhat arbitrary. For instance, Central Link could have been 24 instead of 21 miles in
length. It could have been a $1 billion, $1.5 billion, or $2 billion project. The circle at
60-miles is just an arbitrary point picked by the author. The circle labeled 100-miles is
meant to represent what Sound Transit is considering for a fully built-out light rail
system; namely, light rail from Tacoma to Everett and east to Redmond and Issaquah.
The point is to evaluate enough alternatives, scenarios or projects to define the line.
The rail+bus curve in Figure 8.2.1 is shown as always being more costly than the all-bus
alternative because that’s what Part 6 of this report indicates will be the case. However,
that’s not for certain. Perhaps the curves cross each other at some point. The only way to
find out is to compute the costs and benefits for enough points along the curves to make
sure we understand their shapes. In other words, the study team needs to design a
number of rail alternatives (or scenarios), which differ in terms of length, then compare
each with an all-bus alternative designed to have the same delay reduction benefit.
Figure 8.2.1 shows what the results of an “equal benefit” alternatives analysis might look
like. The circles line up vertically along the dashed equal benefit lines. In an equal cost
study the rail and bus circles would line up horizontally.
Sound Transit already has most of the data to define two of the eight data points in Figure
8.2.1, and that’s why they are shaded black. The EA for Link IS obviously has the cost
data. The mode split models have been run for Link IS and from their results Sound
Transit should be able to compute the impact on regional travel delay fairly easily. The
same is true for Central Link.
What the new and proper alternatives analysis would involve is getting data to plot the
hollow circles.
It should be clear then that we need something more than was done in the1993
alternatives analysis. It just compared one rail alternative against one bus alternative.
We need a number of rail alternatives to be compared against a number of equal benefit,
or equal cost, bus alternatives.
Again, the first reason we need to see the entire curve is because we need to see where
Sound Transits rail-centric strategy is leading. Could we afford to follow it to a full 100-
mile system? If we did, would it reduce delay enough to make the whole program
worthwhile?
The second reason we need to see the full curves is see if they cross over. It is
theoretically possible that a short light rail system in the most heavily traveled part of the
network would be the least expensive way to get a modest amount of delay reduction, but
that an all-bus network would be less expensive to achieve a large reduction. A crossover
seems unlikely, but we need to see the entire curves to be sure.
129
The third reason we need the entire curve has to do with cost-benefit. It is almost certain
that both curves will gradually get steeper per the law of diminishing returns. What that
means is that at some point the cost of getting an extra measure of delay reduction will be
higher than the value of having that reduction. For example, we could, in theory, extend
light rail from Issaquah to North Bend. This might save a few commuters a few minutes,
but the cost would probably exceed the benefits. Right now, ignoring the bus alternative
entirely, we have absolutely no idea how big a light rail system it might make sense to
build. That is, how far out the curve it would make sense to go.
8.2.2 Specific alternatives that should be included in a proper alternatives analysis
The light rail alternatives:
There are at least five light rail+bus alternatives or scenarios that should be evaluated.
Each alternative is comprised of a HCT sub-network embedded into a much larger (in
route miles) local bus network. The HCT sub-network in each scenario would be
comprised of light rail and BRT. The non-HCT part of the transit network would be
essentially the existing bus system, except some routes would no longer run into
downtown, rather they would terminate at light rail stations. The scenarios are:
1) Link IS+BRT, except with Link terminated south of the DSTT - The only
part of the high capacity transit (HCT) network that would be light rail in this
scenario would be light rail from South 154th to somewhere just south of the bus
tunnel. The rationale for this scenario is mainly political. Rail fans could still have
light rail, but it would stop short of the DSTT so as not to interfere with BRT
operations.
2) Link IS+BRT-- This scenario would build Link IS as currently planned, but that
14 miles would be the full extent of light rail. The remainder of the HCT network
would be BRT.
3) Central Link+BRT-- This scenario would build a 24-mile version of Link
from South 200th to Northgate. The remainder of the HCT network would be BRT.
4) 60-mile Link+BRT- This scenario would build 60-miles of light rail and
complete the rest of the HCT network using BRT. The idea here is to have a data
point between a 24-mile system and a 100-mile system. The figure of 60-miles is
just suggestive. The light rail system that ST plans to study for Phase 2 might be a
good candidate. The author thinks it might be about 60 miles long.
5) Fully mature version of Link+BRT-- This scenario would have light rail
going anywhere and everywhere that it might conceivably make sense to go in the
next 20 years. Then BRT would be used for the remainder of the HCT network.
This fully mature version of Link might have 100 or so miles of light rail.
The All-Bus Alternatives—
130
There would be five all-bus alternatives. Each all-bus alternative would be designed to
be equal in cost, or in benefit (e.g.: delay reduction) to one of the rail alternatives.
The HCT parts of the network would use BRT. The non-HCT parts of the transit
network would essentially comprise the existing bus system. In other words, the bus
alternatives would be similar to the TSM alternative in the 1993 FEIS, except they would
be updated to include the latest improvements in BRT, such as off-bus fare collection.
It is anticipated that existing and planned HOV lanes or express lanes would suffice as
“guideways” for BRT. It is not envisioned that any lanes would be converted into
exclusive busways. Thus BRT buses would intermix with car and van pools on those
HOV lanes.
The HOV lanes would be managed so as to achieve a good speed differential relative to
the GP lanes. If and when it looked like the HOV lanes might slow down from over use,
the requirement for car pool occupancy would be increased from 2 to 3, or 3 to 4.
The cost of the All-bus alternatives would include the cost of completing any HOV lanes,
HOV to HOV interchanges, or HOV center access ramps that would be needed to
optimize the alternative.
The all-bus alternatives would include the most cost effective fix to the downtown Seattle
capacity problem discussed in Part 5, if such a fix were needed.
The other alternatives— As pointed out elsewhere in this report, the truly proper way to
approach this problem of reducing congestion or delay is to consider additional
alternatives beyond just improving mass transit in the form of BRT or light rail. Car and
van pools are particularly cost effective.
“During the past two decades, vanpooling has proven to be one of the regions most successful TDM
products. There are now more than 1200 public vanpools operating in our four county regions,
carrying more than 5.5 million passenger trips every year.” (Ref 47: p.7)
Thus, ideally, there should be entire alternatives where all funds are devoted to car and
van pool incentives. There should be other alternatives where car and van pool
incentives are combined with other Transportation Demand Management (TDM)
concepts. And there should be alternatives wherein part of the money goes for BRT and
the remainder goes for car and van pools.
If for some political or procedural reason these other alternatives can not be included in
the “new and proper” alternatives analysis recommended in this report, then they should
be done in parallel by the Puget Sound Regional Council in a way that would allow the
results to be directly compared against the bus and rail alternatives. (e.g.: equal in cost or
in benefit to the rail and bus alternatives)
131
8.3 Roles and responsibilities
8.3.1 Sound Transit cannot be trusted to conduct an objective alternatives analysis.
Therefore Sound Transit’s role should be limited to two things:
1) Sound Transit should pay for the entire analysis including funding the tasks in
8.3.2 or 8.3.3 below.
2) Sound Transit should design the light rail alternative and estimate its costs.
Much of this work has already been done for Central Link and Link IS; however,
ST will need to design the mature 100+ mile version of Link as well as one or two
intermediate steps toward it. Also ST should prepare cost estimates for the various
light rail alternatives.
Neither Sound Transit nor the Sound Transit Board should have any management
responsibility over the tasks in 8.3.2 or 8.3.3.
8.3.2 The all-bus alternatives should be designed and cost estimated by some
organization that has the skill and motivation to design the best possible bus alternatives.
METRO should be considered for this role.
8.3.3 A third organization should run the mode-split models and produce ridership
estimates, estimates of impact on travel delay, and most if not all the estimates of
environmental impacts. This organization might also provide guidelines for doing the
cost estimates so they are done in consistent manner. This same organization should
probably author the alternatives analysis report. The PSRC should be considered for this
role, assuming they make a formal commitment to be objective.
8.3.4 Several expert panels should be formed to monitor the above mentioned players,
and ensure an objective and through job. They should have sufficient power to force
correction of any problems. Members of these expert panels should have no conflicts-ofinterest.
These expert panels should include known critics of Sound Transit’s light rail
plans. There should be at least one expert panel knowledgeable about mode-split models,
and another expert in economic analysis.
8.3.5 Ideally, there should be still another organization that can design and cost one or
more alternatives that rely mostly on car pools, van pools, and various forms of TDM.
8.4 Early indicators as to whether or not Sound Transit plans to continuing abusing
the planning process for Phase 2.
The author does not expect Sound Transit’s Board to conduct a new and proper
alternatives analysis as described in Sections 8.1 and 8.2 unless there is outside pressure
forcing them to do so. How would we know until it’s too late? As Sound Transit begins
132
Phase 2 planning there are several litmus tests that will give early warning as to whether
or not Sound Transit plans to continue abusing the planning process. These tests are:
1) Sound Transit designs the all-bus alternative by itself rather than lets outside bus
experts and advocates do so. (It would be easy for Sound Transit to design the all-bus
alternative so it’s deliberately inferior to the light rail alternative. There are a host of
ways to cheat. One way is to inflate costs by including expensive things like freeway
HOV to HOV interconnects until and unless it is proven each of them is cost effective.)
2) Sound Transit refuses to include alternatives suggested by its critics. In specific,
Sound Transit refuses to include alternatives that do not have any light rail, or that
truncate Link IS south of the DSTT. In other words, all its alternatives take Link IS or
Central Link, and Link in the DSTT, as givens.
3) Sound Transit refuses to give its critics meaningful roles in the design and evaluation
of alternatives. Taking “input” at public hearings is not a suitable substitute.
4) Sound Transit does not structure the alternatives analysis so all the alternatives are
either equal in cost or equal in benefit.
5) Sound Transit does not commit publicly and officially to use reduction-in-travel-delay
or reduction-in-congestion as the main metric for evaluating the alternatives. Instead ST
uses something like transit ridership.
6) Sound Transit does not commit publicly and officially to discount the cost and benefit
streams when arriving at a summary measure of cost-effectiveness.
7) Sound Transit appears to be in a hurry with Phase 2. (Since Link will do virtually
nothing to relieve Puget Sound’s pressing traffic congestion there is no apparent reason to
hurry commitments to extend it, like a used car salesman coercing customers into hasty
decisions. Additionally, it would be foolish to lock in any plans to extend Link across
Lake Washington until Link IS is successfully operating and until funding to extend Link
to Northgate is assured.)
133
Part 9: Conclusions
Main conclusions of this report:
1) The existing planning process is not producing the kind of information needed by
officials and the public to make intelligent decisions about major mass transit
projects. Important information is missing or obscured. Promising alternatives are
ignored. Reports seem intended more to sell than to inform.
2) Sound Transit and its predecessor agency the RTA have abused the planning process
in order to promote light rail. They biased key studies by making inappropriate
assumptions and masking key information. They compared a robust rail alternative
against a deliberately hobbled bus alternative. They disseminated misleading
information to the public.
3) ST justifies its choice of light rail on the one and only rail vs. bus alternatives analysis
conducted here since the 1980s. However, that study was deliberately biased to favor
rail. When that bias is removed the underlying data shows that an all-bus solution
could probably achieve the same level of benefit at far lower cost.
4) ST and RTP dismissed bus alternatives largely on false claims that buses lacked
adequate capacity. Their analysis was deliberately manipulated to support these
claims. BRT has more than adequate capacity to meet the region’s long-term needs.
Light rail has less capacity than BRT and is therefore less strategic.
5) As construction begins on Link there is still no study which compares the benefits of
spending $X billions on light rail plan versus spending the same amount on bus rapid
transit.
6) If the money now intended for light rail were instead redirected toward other projects
such as BRT, car and vanpool enhancement, and other transportation projects the
region could probably make considerably more progress in solving our transportation
problems, because these other alternatives are more cost-effective.
7) By objective measures Link does not seem like something worth pursuing. Among
other faults it would have almost no effect on traffic congestion and is not costeffective
as an alternative to driving.
8) BRT is a viable alternative to light rail in the Puget Sound Region. It could achieve
the same benefits at a much lower cost and has more than adequate capacity to handle
long term growth.
9) Link is the failed result of a faulty planning process. The process can and has been
manipulated to favor preordained outcomes. It is not objective. It does not produce
the type of information needed to make intelligent decisions. It fosters distrust. It is a
process that needs to be fixed. This would take local and Federal action.
134
10) The region is embarking on a rail-centric mass transit strategy, which could result in
over 125 miles of light rail. If fully implemented, that strategy will probably cost the
region about a billion dollars per year more than an all-bus (BRT) strategy having the
same level of transit ridership and related benefits. Meanwhile Link IS and Central
Link are probably costing over twice what comparable all-bus alternatives would
cost.
11) Link light rail is an example of the “waste, and abuse” that is driving up the Federal
budget deficit, because it was sold on the basis of misleading information and because
there are more cost-effective alternatives.
12) There has been insufficient public discussion about the merits or consequences of
committing the region to this multi-billion dollar rail-centric strategy, and there is no
solid analysis demonstrating it’s the best strategy.
13) Transportation planning in the Puget Sound region has not placed a high priority on
spending taxpayer money efficiently or in finding the lowest cost solutions.
14) Link should be put on hold and further Federal funding withheld until and unless a
new and honest alternatives analysis is completed, and that analysis demonstrates that
light rail is superior to BRT and other options. On the basis of available evidence,
such a conclusion seems unlikely.
15) The FTA should review, and if appropriate withdraw, its prior acceptance of the 1993
FEIS as meeting FTA requirements for a proper alternatives analysis, since that
particular analysis is faulty in so many respects and contains no apples-to-apples
comparison between Sound Transit’s light rail plans and all-bus alternatives. By the
same token FTA should not allow Sound Transit to proceed with any Phase 2
planning until a new and proper alternatives analysis has been completed.
16) Sound Transit’s planning process for Phase 2 will not provide the information needed
for this region to make intelligent decisions about massive investments in mass transit
unless the recommendations listed in Part 8 of this report are adopted.
17) The PSRC should be forced to obey the State Law that requires Least Cost Planning
because imposition of that planning technique is the single most important thing that
can be done to help ensure that scare transportation tax dollars are spent wisely.
18) Sound Transit’s latest Long-Range Plan should not be approved because it’s very
foundation, the 1993 FEIS, is obsolete and corrupt.
135
Brief Introduction to Least Cost Planning
Least cost planning is a planning methodology to identify the least cost way to achieve a stated objective.
An example of a stated objective would be a 10% reduction in traffic delay in the Puget Sound region.
Within Washington Administrative Code (WAC 468-86-030 and WAC 468-86-080) least-cost
planning is defined as “a process of comparing direct and indirect cost of demand and supply options
to meet transportation goals and/or policies where the intent of the process is to identify the most
cost/effective mix of options.”
Washington State Law requires that regional planning agencies, such as the Puget Sound Regional Council,
use Least Cost Planning as the basis for their regional transportation plans.
No single remedy will suffice-- The methodology recognizes that no single remedy, such as synchronizing
traffic lights, adding HOV lanes on I-5, or building 20 miles of light rail is likely to comprise the entire
solution. Instead the least cost solution will probably be a carefully designed mix of remedies, for instance
spending: $200 million on remedy A + $500 million on remedy B + $400 million on remedy C, and so
forth. In short, the least cost solution is probably a “package” of different projects.
Include all the costs-- Least cost planning also recognizes that finding the least cost solution in pure dollar
costs is not sufficient. In other words it is not sufficient to consider only the construction and operating
costs of some option like light rail. The side effects, that is the social, safety and environmental costs are
also important. Thus LCP considers these kinds of costs and converts them into dollar equivalents. For
example a ton of air pollution may be assigned a dollar equivalent cost of $10,000, or a serious injury may
be assigned a cost of $50,000. Economists and insurance adjusters have well established ways to derive
these dollar equivalents.
Besides having negative side effects that are considered as costs, most projects have positive side effects
that are in addition to helping achieve the stated objective. An example might be a roadway improvement
not only helps reduce delay but also reduces accidents. Positive side effects are also converted into dollar
equivalents. The objective is to estimate what I would call the “net social cost” of each alternative. The
net social cost would be the capital and operating cost of the alternative, plus the dollar equivalent of the
negative side effects, minus the dollar equivalent of the positive side effects.
For the purposes of LCP a remedy is any project or policy that might help achieve the stated objective. A
remedy can also be considered as an alternative or an option, because during the LCP planning process the
different remedies, alternatives, or options will essentially be competing with each other to see which
survive and become part of the least cost package.
Consider all the alternatives-- A key aspect of LCP is to consider any and all alternatives that might help
achieve the objective. If the objective is to reduce travel delay this means not only considering different
ways to increase the supply of transportation (such as building new roads and transit routes) but also ways
to reduce demand (such as telecommuting, and shifting traffic to off-peak hours). And it includes ways to
increase the efficiency of existing transportation investments (such as ramp metering, and van pooling).
Each alternative is a specific project-- Least cost planning can not deal with open ended alternatives
such as “build more roads”. Instead LCP deals with specific projects whose cost and benefit can be
quantified. Thus LCP needs alternatives stated in terms like this: build 20 miles of light rail from SeaTac
to Northgate, replace the 520 bridge, or institute a $20 million car pool incentive program.
Compute the cost-effectiveness of each project-- The essential task in LCP is to compute the net social
cost and benefit (effectiveness) of each project. Again, the benefit is the degree to which the project helps
achieve the stated objective. For instance the 520 bridge replacement project might have a net social cost
of $500 million and reduce regional travel delay by 1.2%. The cost-effectiveness of this project would be
$417 million per 1% delay reduction.
136
Create a prioritized list of projects-- One of the main outputs from LCP is a list of all the alternatives
that were considered rank ordered in terms of their cost effectiveness.
What is the least cost solution?-- The least cost solution is found by selecting the most cost effective
project first, the second most cost effective next, and so forth until enough have been accumulated to reach
the stated objective in terms of beneficial effect. Implementing this market basket of projects would be the
least cost way to achieve the stated objective. This idea is illustrated in the diagram below, which arranges
a number of projects (labeled A through H) in order of decreasing cost-effectiveness. In other words
Project C is the most cost effective while F is the least cost effective. The projects can all be of different
sizes. In this particular example a scaled down version of Project B might be needed whereas project F is
simply not cost-effective enough to survive.
Least cost planning does not dictate the solution—The “least cost solution” identified by LCP is
essentially a suggestion. Decision makers will always have the ability to accept, modify or reject it.
Hopefully, provides objective information that helps society make more rational decisions.
Disagreements over valuation-- Not everyone may agree with the dollar value assigned by the planning
team to things like a ton of air pollution, or a life lost to accident. Some might feel strongly that a higher or
lower value is more appropriate. This difference in values can not be resolved analytically and in that
respect LCP is not a perfect exercise. However, what can be done is to run sensitivity analyses to see if
assigning a different value would alter the final results. For example if an environmental organization felt a
ton of air pollution was worth $80,000 rather than $10,000 it would be very easy to put that value into the
spreadsheet and see if the order of the alternatives on the priority list changed. If not, perhaps the issue
would die out. If it did change the order of the projects, the dispute over which projects to implement
would need to be settled by other means.
A
B
C
D
E
F
G
H
Cost
Effect
Stated
objective
137
The law of diminishing returns-- Every type of remedy such as widening roads, building light rail,
coordinating traffic lights, or incentivizing people to carpool follows the law of diminishing returns. The
first dollars spent on any of these remedies will probably be spent where “bang for the buck” is highest.
For instance the first 24 miles of light rail from SeaTac to Northgate would certainly attract more transit
riders than 24 miles between say Northgate and Marysville.
Thus each type of remedy, such as Sound Transit’s current rail-centric strategy, will follow a curve like that
below. The most productive 20 miles of light rail would achieve relatively high benefits per dollar spent,
whereas the least productive 20 miles would achieve relatively low benefits per dollar spent.
The only way to actually construct a light curve representing Sound Transit’s light rail strategy would be to
hypothesize a number of different light rail projects, evaluate each in terms of cost and benefit and plot the
results. A line between those plot-points would define the light rail curve. The same logic applies to all the
other remedies, and reinforces the idea that least cost planning must evaluate a whole range of specific
projects, as opposed to a type or class of projects.
The same law of diminishing returns applies to the market basket of remedies that constitutes the least cost
way to achieve the stated objective as illustrated below.
Benefit (e.g.: reduction in delay)
Net
Social
Cost
High
Heigh
Zero
Most
productive
20 miles
Next most
productive
20 miles
Least most
productive
20 miles
138
It is important to understand the concept behind these curves because there are practical implications. For
one thing just because the first 20 miles of light rail might be fairly productive it does not follow that it
would make sense to add another 20 miles. That next 20 miles might be far less cost effective than
switching over and doing something different with the next $X billion. So perhaps after building 20 miles
of light rail we should next implement 100 more miles of HOV lanes. Perhaps the next 20 miles of light
rail would make sense, or then again maybe the next best project would be some road widening.
The next major implication of the curves is that at some point the curve gets so steep that going further
makes no sense. That point occurs when the cost of gaining another increment of benefit exceeds the value
of that benefit. A cost-benefit analysis can determine that point. This point may or may not be reached
before achieving the originally stated objective. In other words if the original stated objective were to
reduce travel delay 10% we might find it too expensive to go that far. We may find anything over 8% does
not make sense. This means we would just implement enough projects on the prioritized list to achieve that
8%.
Benefit (e.g.: reduction in delay)
Net
Social
Cost
High
Heigh
Zero
Most
productive
projects
Next most
productive
projects
Least
productive
projects
139
Tech Notes
Vehicles Removed From Peak Period Traffic in 2020:
Central Link- Central Link will reduce regional PM Peak period VMT by 63,185 in the year 2020. The
source data (16,996,434 – 16,933,249) is found on page 3-4 in the FEIS. (Ref 2: Table 3.1-4) Per PSRC
staff, the average one-way work trip length is about 10 miles. Dividing the reduction in VMT by the
average trip length indicates that Central Link would remove about 6320 cars from peak period traffic in
2020.
Using the change in VMT to estimate the number of cars removed is the honest way to get that number
because the intent is to provide some indication of how much rail would reduce traffic congestion that is
easier to visualize than something called “VMT”. However, Sound Transit has chosen to use a different
approach, which gives a larger, albeit misleading, number.
“Link Light rail will remove 16,000 vehicles from the daily commute-“ (Ref 27: Appendix B,
page 1-8)
This 16,000 number is not cited in the FEIS for Central Link and the author has not found its derivation.
Sometimes a “new transit rider” is said to equate to one less car trip. However that is misleading for two
reasons. First, Central Link trips are shorter (5.3 miles) than the average auto trip. Second, if someone
drove 20 miles to reach a light rail station then took Link the last two miles that would count as a new
transit trip, and thus one less auto trip.
Link IS- Data in the EA shows the Initial Segment will reduce peak period VMT by 18,899 VMT, which
is equivalent to removing 1890 cars from peak period traffic. (Ref 18: Table 3.1-1)
__________________________
Figures 6.2a: Source of all data: Ref 1: Table 4. Capital costs converted to annualized value by assuming
30-year 6% bond. Thus annual cost = .072* capital cost. Annualized capital cost added to O&M cost to
get total annual cost used on Y-axis. All costs in 91$. Total annual cost of No Build (NB)= $360 M;
TSM= $737 M; Rail/TSM= $1320 M. Chart plots costs and ridership above and beyond or over the cost of
the No Build alternative so $360 M is subtracted from the TSM and Rail/TSM totals to get the values
plotted. ( $377 M/yr for TSM; $960 M/yr for Rail/TSM) Ridership metric plotted is total daily transit trips
minus the 388,500 attracted by No Build.
Figure 6.2b: Rail line is same as in Figure 6.2a. TSM plot point (the end of the line) is moved right to the
higher “unconstrained ridership” of 518,000 daily trips or 129,500 over NB. (Ref 42: Table 3.2) It is also
moved vertically to account for the cost of adding $600 M for a second bus tunnel. $600 M is from Ref 26
page 8. Thus total cap cost of TSM with the second tunnel was $5.3 B. The value plotted is it’s total
annual cost of $420M/yr over NB
Figure 6.2 d: This is the same as Figure 6.2b except for the extension of the TSM line. The cost of the
extension had a capital and O&M component. The RTP had estimated it would cost $2.1 billion to expand
the TSM alternative enough to attract another 100,000 daily riders. (Ref 45: p. 46) This cost was prorated
to account for the fact only 42,000 added riders would be needed to increase unconstrained TSM ridership
to where it equaled that of rail (i.e. 560,000). This brought cap cost of a TSM alternative able to equal rails
ridership up to $6.193 B. I assumed the O&M cost for this extension was the same, on a per rider basis, as
that used to plot the TSM line in Figure 6.2b. The total annual cost plotted for extended TSM was
$515M/yr over NB.
Figure 6.3a: This is an updated version of Figure 6.2d meant to indicate what the Rail/TSM and TSM
alternatives would cost today. The capital cost of the rail component in the $11.5 billion Rail/TSM
alternative costs $7.9 billion. This was multiplied by 1.44 to account for ST’s finding that light rail was
140
costing 1.44 times more than previously expected. Also the HOV lanes are now half completed. The
Rail/TSM and TSM alternatives both contained $1.5 billion for HOV lanes, so $750 M was subtracted from
each alternative. Finally all costs were multiplied by 1.32 to inflate them from 91$ into 02$. The capital
costs in 02$ were: Rail/TSM= $18.778 B; extended TSM=$7.184 B; NB=$1.584 B. Total annual costs
were Rail/TSM=$2001 M/yr; extended TSM=$1084 M/yr; NB=$476 M/yr. The costs above NB were
plotted. The $917 M/yr difference is $2001 M-$1084 M.
Strictly speaking this is an updated cost comparison between the rapid rail/TSM alternative and the TSM
alternative in the 1993 FEIS. However it gives a good indication of the difference between the light
rail/bus alternative that Sound Transit is planning and an all-bus or BRT alternative. That is true because
from a capital cost viewpoint the cost of the rapid rail system in the 1993 FEIS would not be much different
than the capital cost of a light rail system following similar routes and having similar stations, which seems
to be the intent. What differences we know about, such as at-grade rather than tunnel along Rainier Ave,
would save a $100 M or so, but that’s not enough to noticeably alter these results. As to ridership, the light
rail is slower and would probably attract fewer riders, but saying this chart gives a good indication of the
performance of light rail is a conservative approach since if light rail attracts fewer riders than indicated in
this figure the rail line would get steeper and the difference between rail and bus grow wider. In other
words the $917 M/yr difference is probably too low.
Figure 6.4: This is simply figure 6.3a with a different X-axis. It would have been preferable to plot the
effect of the alternatives on Peak Period VMT since that is the best proxy for impact on traffic congestion
that can be obtained from the data provided in the EIS’s. However, the 1993 FEIS did not report on Peak
Period VMT so the reduction in Total Daily VMT relative to No Build –which all the EIS’s did report--
was used to dimension the X-axis. The performance of Central Link and Link IS is also shown.
Figures 6.1, 6.2c, and 6.3b: These charts all show the incremental cost of adding incremental riders. The
first plateau is the total annual cost of the No Build alternative (i.e. the current bus system with incremental
expansion as population grows) divided by the number of annual riders it carries. The annual costs are
calculated as described above. For instance the No Build alternative costs $360M/yr (91$) and carries
109M riders per year. In the transit industry a “rider” means one person taking a one-way trip by transit.
Thus riders= rides. Thus the cost per ride under the No Build alternative is $3.29. But the No Build
alternative only attracts 389,000 daily riders. If we want more we could implement the TSM alternative.
That would cost another $ 377M /yr above the $360 million we are already paying for NoBuild. However,
it would add 85,500 additional daily riders or rides, which is equivalent to 24.4 million additional annual
riders. Dividing the added cost by the added rides shows it will cost $15.40 for each additional ride. But
this still does not get to the 560,000 ridership level of the Rail/TSM alternative. Adding rail on top of the
TSM would cost an additional $583M/yr. The cost for each additional rider it attracts comes to $24.70 per
ride. If the benefit of attracting each of these added riders does not exceed that amount then rail is not
worth building. Put another way, at some point the law of diminishing returns takes over. It is simply not
worth the cost of attracting still more transit riders.
Figures 6.2c and 6.3b are constructed using the same approach. When, in Figure 6.2c, the TSM capacity
restraint is removed with a second tunnel its ridership increases so there is not much left that implementing
rail could add. This causes the cost per ride for each ride provided by rail to rise dramatically to $47.50.
Figure 6.3b is based on the updated costs used to construct Figure 6.3a. The marginal cost of the rail riders
now exceeds $90 per one-way ride.
______________________________
Cost and ridership for Link and Rapid rail:
NOTE: The unit costs used in the 1993 FEIS and those used in the EIS’s for Central Link and Link IS
appear to be identical except for adjustments for inflation between 1991 and 1995. In other words the cost
for a given station or for a given segment of track appear to be identical in constant dollars whether for the
rapid rail in the 1993 or the Central Link system costs as presented to the voters in 1996. Evidence for this
is found in a 1996 RTA memo from Bob White, then head of the RTA.
141
“…the ERP concluded that RTA’s capital cost estimates were appropriate and concluded that “the
capital cost methodology used for the JRPC system plan (i.e.: the 125-mile rapid rail system in the
1993 FEIS) was appropriately updated to reflect 1995 construction costs and was then applied to the
new options …” We are continuing to use these same tunnel cost estimates.” (Ref 63)
…First, as I have discussed, to remind you that the RTA numbers still provide a sound basis for
upcoming decisions. (Ref 63)
What this means is that the author is probably correct in assuming that the cost of a 125-mile rapid rail
system as reported in the 1993 FEIS would, after increasing it by 44% to account for known errors, provide
a good estimate of what a 125-mile version of Link light rail would cost. Of course that estimate is not
perfect since a few miles along Rainier Ave are now at grade whereas before they were in tunnel, and no
doubt there are other details that have changed. However, this approach is the best available way to
estimate the cost of a fully mature version of Link, and the capital cost estimate would need to be off by
literally billions of dollars to invalidate the main finding of this report; namely that a fully mature (100+
mile) version of link light rail would be hundreds of millions of dollars more expensive than a BRT system
having the same ridership.
Central Link- SOURCE DATA: Cap cost = $2.599 billion in 95$ as of Jan 11, 2001 (Ref 27: Section 13)
O&M cost = $42 million 95$ per year. (Ref 2: p.5-11)
Year 2020 daily rail riders= 133,000, of which 55,000 (or 0.414) are new transit riders. (Ref 2: p.3-21 and
7-11) Ratio of new riders to total riders was slightly lower (0.373) in New Starts Report. (Ref 40: p.10)
Another ST document says: “Almost one third of the 100,000 plus daily light rail riders will be former
automobile commuters.” (Ref 27: Appendix B, p.2-8)
No build=72883000 total daily VMT; with IS VMT=72802000; with central link VMT= 72612000. (Ref
18:p.20)(Ref 2: p.3-4) ESTIMATES: total annual cost = .072*2600 million + 42= $229 million/yr in 95$.
02$/95$= 1.27. Thus Central Link costs $291 million/yr in 02$. Reduction in total daily VMT= 0.37%.
NET: $291m/yr for 0.37% reduction in VMT.
Initial Segment- SOURCE DATA: Cap cost=$2.1B in YOE$ (Ref 18: p.45) O&M cost = $52 million
Year$ assume 01$ (Ref 18:p.47) $3602 in YOE$ = $2600 in 95$ (Ref 27: Section 13, p.1-8) Year 2020
daily boardings= 42,500 (Ref 18: p.20) New transit rider data not located. VMT data per Ref 18: p.20.
ESTIMATES: $2.1 B in YOE$ = $1.52 in 95$. Cap cost of Initial Segment =.072*1520 million = $109
million/yr in 95$, or $139 million per year in 02$. Total including O&M= $191 million per year in 02$.
Based on Central Link’s ratio of total to new transit riders: 55/133* 42,500= 17,575 new transit riders on
IS. Reduction in total daily VMT= 0.11%. NET: $202m/yr for 0.11% reduction in VMT.
Rapid Rail in 1993 FEIS- Cap cost for alternative =$11.5 billion. Cap cost for just rail =$7.9 B in 91$.
(Ref 45: p.44 and Ref 45: p.40) O&M cost= $399 million/yr in 91$. (Ref 1: p.xxxiii) Ratio 02$/91$ = 1.32.
Total daily 2020 rail boardings = 330,000. (Ref 42: p.3-26) New 2020 transit riders due to Rail/TSM
alternative= 172,000 per day. Year 2020 total daily VMT for NB= 84,189,000, VMT for rail=
81,030,000.(Ref 1: p.3-95) ESTIMATES: Total annual cost for alternative=.072*11500 million +399=
$1227 million/yr in 91$, or $1619 million/yr in 02$. New transit riders on rail= 55/133*330000= 136,500
per day. VMT percentage reduction for entire alternative= 3.75%. NET: $1619 million/yr for 3.75%
reduction in total dailyVMT.
_________________________
Estimate peak loading on Link if it were extended north, south and east: The approach used was to
start with RTP’s forecast of 15,000 pph --as the one-way peak load on the 125-mile rapid rail system-- and
then estimate how much that would decline if the system were truncated to Central Links 24-mile route. If
the result comes close to the 5415 predicted for Central Link then the two forecasts are consistent.
142
Step one was to see how much effect removing the east-side route from the 125-mile system would have.
According to the Travel Forecasting Results report there were 88,600 daily trips to the north corridor from
the south and east corridors plus the Seattle CBD. This volume translated into a peak load of 15,000 just
north of the DSTT. (Ref 42: Table 3.8) Of that, 12,300 came from the east corridor. Per Ref 42, Figure 3.3
some east to north riders use the I-405 corridor but I assume all went via the CBD. Eliminating all riders
from the east would leave 76,300 trips to the north corridor via the DSTT.
Step 2 removes some of those riders to the north corridor who come from the south corridor. Of the 88,600
total, 23,000 came from the south corridor. Per Table 3.15, of the 21,300 persons who depart south
corridor stations in the PM peak period 7900 use stations south of SeaTac. I assume half of those 7900
would not use rail at all if it ended at South 200th, but the other half would transfer to bus. Thus the south
line would loose 4000 PM departures, which can be assumed to comprise 19% of its riders. 19% of 23,000
is 4400 riders. 76,300 less 4400 leaves 71,900 riders going into the north zone assuming it extends to
Everett.
Step 3 is to remove half of all those that would be destined to stations north of Northgate. Again the
remainder are assumed to transfer from rail to bus, or simply park at the Northgate station. Per Table 3.15
18,700 of the 43,300 PM departures from north corridor stations occur north of Northgate. Assume half the
18,700 or 9400 would no longer use rail if it ended at Northgate. 9400 is 21% of the 43,300 which means
that the north line would loose 21% of its 71,900 remaining riders. The end effect of eliminating the east
line and truncating the north and south lines is that daily ridership into the North Corridor would drop from
88,600 to 56,800 riders. This would cause loading at the peak load point to drip proportionally, from
15,000 to 9600. This is admittedly a very crude approximation.
In short, this exercise establishes the approximate ratio of peak load point volumes between a full 100+
mile rail system and 24-mile South 200th to Northgate rail system. That ratio is 15,000 divided by 9600 or
1.56 to 1. This ratio was determined from examination of RTP’s rapid rail system but the same ratio should
apply to a light rail system, since any difference in speed between rapid rail and Link should not affect the
ratio.
There now exist two forecasts and one statement. First, we have RTP’s forecast that peak loading on the
124-mile rapid rail system would be 15,000pph. in 2020. (Ref 1: p.2-58). Next we have ST’s statement
saying the long range ridership projections estimate Central Link would have a peak load of 15,000. (Ref
18: Appendix L, p.23). One can only assume they meant that for a fully built out 100+ mile version of
Central Link. Finally, we have ST’s forecast that peak loading on a 24-mile version of Central Link will be
5415 pph in 2020.
Thus if ST was right in saying a fully built out version of Link would have a peak load point volume of
15,000, then the peak load on the 24-mile version should be roughly 9600. However this conflicts with
ST’s forecast showing the peak load on the 24-mile system would only be 5415. Since 9600 is much
different from 5415 there appears to be a conflict between these two forecasts. Either the 15,000 figure is
correct, or the 5415 figure is correct.
The author suspects that the 5415 number is the correct one because it was a direct forecast made for
Central Link, whereas the 15,000 attributed to Central link was probably just taken blindly from RTP’s old
1993 FEIS for rapid rail. However, since the author can’t be certain it is interesting to consider the
implications of each being right.
If 5415 is the correct peak load for the 24-mile version of Link then a fully built out version of Link would
probably have a peak load of only about 8500, not the 15,000 claimed. This comes from applying the 1.56
to 1 ratio to 5415.
If the 15,000 figure is correct then Link would not have enough capacity on the south line by 2020 and not
enough on the north line to handle growth much beyond 2020.
________________________________
143
Cost per mile sanity check: Central Link: SOURCE DATA: Cap cost = $2.599 billion in 95$ as of Jan
11, 2001 (Ref 27: Section 13) 02$/95$= 1.27. Length = 21 miles. Calculation: 2600 x 1.27 / 21 = $157
million (02$) per mile. Link IS: SOURCE DATA: Cap cost=$2.1B in YOE$ (Ref 18: p.45) $2.1 B in
YOE$ = $1.52 in 95$. 02$/95$= 1.27. Length= 14 miles. Calculation: $1520 x 1.27 / 14 = $138 million per
mile. Per mile cost of rail trend line in Figure 6.3a: 125-mile rapid rail in 1993 FEIS cost $7.9 billion in
91$. Ratio 02$/91$ = 1.32. Cost per mile in 02$= 7,900 x 1.32 / 125 = $83.4 million/mile. Correction
factor is 1.44. Thus cost per mile of the trend line in Figure 6.3a is $83.4 x 1.44 = $120 million per mile.
In sum, the rail line in Figure 6.3a, and the finding that a full build out of light rail would cost about $900
million a year more than a comparable BRT system, are conservative because they are based on light rail
costing $120 million per mile, whereas Link IS is costing $138 million per mile and Central Link was
expected to cost $157 million per mile.
____________________________________
Capital cost of Sound Transits Rail centric strategy versus an all-bus alternative: The capital cost of
the rail centric strategy (which is essentially the Rail/TSM strategy but with 125-miles of light rail rather
than 125-miles of rapid rail totals $18.8 billion in 02$. Derivation was: Take original cost of $11.5 billion
in FEIS. Subtract one half of the $1.5 billion that was for HOV lanes. Multiply the $7.9 billion that was
for rapid rail by 1.44 to correct for underestimates. Then multiply by 1.32 to convert from 91$ to 02$.
Capital cost of the all bus alternative in Figure 6.3a (which is identical with the TSM alternative in the
FEIS) is $6.8 billion in 02$. Derivation: Take original cost of $4.7 billion and subtract one half of the $1.5
billion that was for HOV lanes since these have now been built. Add $600 million for a second bus tunnel.
Multiply total by 1.32 to convert into 02$.
__________________________________
144
Blank page
145
References
Abbreviations:
PB/KE = Parsons Brinckerhoff/Kaiser Engineers
RTP = Regional Transit Project
RTA = Regional Transit Authority (predecessor to Sound Transit)
PSRC = Puget Sound Regional Council
1 Final Environmental Impact Statement, Regional Transit System Plan, Vol.1, Regional Transit Project,
March, 1993.
2 Final Environmental Impact Statement, Vol.1, Central Link Light Rail Project, Sound Transit, Nov.
1999.
3 Documentation of Major Investment Study, Regional Transit Authority, March 12, 1997.
4 Transit Technology Overview, Technical Memorandum, Final Draft, by PB/KE, for Regional Transit
Project, Aug, 17, 1992.
5 Transit Technology Review, Sound Transit Link light rail project, LTK Engineering, 2/2/99
6 FACT: The study of light rail alternatives, Sound Transit, Aug. 2002
7 Sound Transit Long-Range Vision Update/Phase 2 Planning, White Paper on the Review and
Qualitative assessment of HCT technologies, Final Report, by PBQD for Sound Transit, June 2004.
8 SEIS/Long-Range Plan Scoping Summary Report and Definition of Phase 2 Alternatives, Office of
Policy and Planning, Sound Transit, June 24, 2004. (slide presentation)
9 Mass Transit, Bus Rapid Transit Shows Promise, Report to Congressional Requesters, US General
Accounting Office, Sept 2001
10 Bus Rapid Transit for the Puget Sound Region, Current Status and Future Potential, Nelson, MacIsaac,
Niles, Integrated Transport Research, Inc., Dec 2001. (attached as Addendum I)
11 BRT Why more communities are choosing Bus Rapid Transit, Transportation Research Board, Wash.
D.C. ,2001
12 “Bus Rapid Transit or Light Rail Transit- How to Decide? Los Angeles Case Study”, Transportation
Research Record 1793, Paper No. 02-4171, John Stutsman.
13 “A bus-based transitway or light rail? Continuing the saga on choice versus blind commitment”, Road
and Transport Research, Vol 8 No 3, David Hensher, Sept 1999.
14 Special Edition: BRT, Journal of Public Transportation, Vol 5, No 2, 2002. (contains 8 papers on BRT)
15 Sound Move, The Ten-Year Regional Transit System Plan, Regional Transit Authority, May 31, 1996.
16 1999 Puget Sound Household Travel Survey, Final Report, by NuStats Research and Consulting for
Puget Sound Regional Council, Dec 1999.
17 The 2001 Metropolitan Transportation Plan Alternatives Analysis and Draft Environmental Impact
Statement, Puget Sound Regional Council, Aug. 31, 2000.
146
18 Environmental Assessment Initial Segment, Central Link light rail transit project, Sound Transit, Feb.
2002.
19 Central Corridor Project Justification Report, by PB/KE for the Regional Transit Project, Nov. 1993.
20 “The System is Warped; put light rail on hold”, Seattle Times, OpEd by R. Harkness, Dec. 22, 2000.
(Attached as Addendum A)
21 “Why Campaigns for Local Transportation Funding Initiatives Succeed or Fail: An Analysis of Four
Communities and National Data”, Mineta Transportation Institute report 00-1; by P. Hass, K. Massey,
L.Valenty, and R. Werbel; June 2000.
22 “Sound Transit Marketing Research Summary Report” by Pacific Rim Resources, Inc. for Sound
Transit, Oct. 27, 2000.
23 Transportation Technical Report, FEIS, Central Link Light Rail Project, Sound Transit and RTA, Nov.
1999.
24 “Ethics and Advocacy in Forecast for Public Policy”, Business and Professional Ethics Journal, Vol 9,
Nos 1 & 2, by Martin Wachs.
25 Development of Busway and Transitway Alternatives, Regional Transit Project, by PB/KE for
METRO, Jan. 1993.
26 unused
27 Central Link Board Briefing Book, Sound Transit, Jan 2001.
28 Summary of Prior All-Bus and Integrated Rail/Bus Alternatives Analyses, by Puget Sound Regional
Council’s Transportation Policy Board and Growth Management Policy Board, Puget Sound Regional
Council, April 5, 2001.
29 Light Rail or Buses in the Downtown Seattle Transit Tunnel: Assessment of Benefits to King County
Metro and Regional Public Transportation, IRT Tunnel Team Report to King County Council
Transportation Committee, by J. Niles, J. MacIsaac, Integrated Transportation Research, Inc. July 25,
2001.
30 2001 Update, Metropolitan Transportation Plan, Baseline Technical Report, Puget Sound Regional
Council, June 2000.
31 Advancing Major Transit Investments Through Planning and Project Development, Version 1.1., by
Office of Planning and Office of Program Management, Federal Transit Administration, Jan. 2003.
32 December 2001 King County Transportation Poll, by The Olympic Institute for the King County
Council.
33 Puget Sound HOV Pre-Design Studies, Final Report, Office of Urban Mobility, Washington State
Department of Transportation, May 5, 1997.
34 DSTT Joint Bus/Rail Operations Capacity Assessment, Technical Memorandum, by PB/KE for
METRO, April 2, 1992.
35 Strategies for Managing Traffic Congestion, PSRC, Nov. 1996.
36 Central Link Board Workbook, Book II, June 14, 2001, Sound Transit
147
37 Evaluation of Joint Operations in the Downtown Seattle Transit Tunnel, Sound Transit, Aug 21, 2001,
revised 9-17-01.
38 ST Link Light Rail Project, Central Line Fleet Management Plan, Final Draft, by LTK Engineering
Services for Sound Transit, July 10, 2000.
39 Sound Transit MEMO, from Joni Earl, Executive Director to Sound Transit Boardmembers, Jan. 4,
2002.
40 Link Light Rail, New Starts Report, submitted to FTA by Sound Transit, Sept. 3, 1999.
41 Transitway Alternative Report, Final Draft, DF-8, by PB/KE for RTP, Nov. 12, 1991.
42 Travel Forecasting Results Report, Draft, D-22, by PB/KE for RTP, Feb. 28, 1992.
43 Destination 2030, Metropolitan transportation plan for the central Puget Sound region, technical
appendices, PSRC, May 24, 2001.
44 General Managers Quarterly Management Report, Year End 2003, King County METRO
45 System Plan Technical Report, Draft, D-82, by PB/KE for RTP, Oct. 1992.
46 Washington State Freeway HOV Program: Status, Performance, Questions and Answers, Washington
State DOT, Jan 2001. (16 page brochure)
47 Regional TDM Action Strategy for implementation in the Central Puget Sound Region, The Regional
TDM Advisory Committee, Nov. 1998. (sponsored by WSDOT, and PSRC)
48 Bus Rapid Transit Partner Jurisdiction Briefing Materials, Six Year Transit Development Plan for 2002
to 2007, METRO, Nov. 2001.
49 The Regional Transit Long-Range Vision, Sound Transit, as adopted May 31, 1996.
50 “The Mythical Conception of Rail Transit in Los Angeles”, J. Richmond, University of Sydney, Nov.
11, 1996, for publication in Journal of Architectural and Planning Research.
51 Summary Report and Recommendations, Regional High Occupancy Vehicle(HOV) Policy Advisory
Committee 1998-1999, sponsored by PSRC.
52 Not used
53 letter from Charles Royer, Chair Project Review Committee to Dave Earling, Chair, Sound Transit
Board of Directors, June 27, 2001.
54 Straight Talk About Transportation, Freight Mobility Roundtable, Jan 10, 2003 slide presentation by A.
Davis, Chair Transportation Commission and D. MacDonald, Sec. WSDOT.
55 TSM Alternative Capital Cost Estimate Results Report, Draft, by PB/KE for RTP, Sept. 1991.
56 Capital Cost Estimate: Rail Alternaative- Volume 1, Draft, by PB/KE for RTP, Jan 1992.
57 Regional Transit System Master Plan, as adopted Oct. 29, 1994, RTA, 1994.
58 WSDOT Memo from Doug MacDonald to Norman Mineta, dated July 3, 2003.
148
59 Sound Transit handout, a set of diagrams of alternative lane configurations on I-90 bridge including R1
(existing) and R8a (preferred by ST), not dated but distributed in 2004 and perhaps earlier
60 “Federal Transit Administration Grant Agreement”, a legal document from FTA to Sound Transit
approving the FFGA, dated Oct. 1, 2000.
61 Destination 2030, Metropolitan Transportation Plan for the Puget Sound Region, PSRC, May 2001.
62 Speeches, “King County Executive Ron Sims’ remarks to the Sound Transit Board”, Jan 11, 2001.
published by King County.
63 RTA Memo, to Dick Ford from Bob White, Executive director of RTA dated Jan. 26, 1996; and RTA
MEMORANDUM to RTA Board from Bob White dated Feb. 26, 1996.
64 ST NEWS, October 3, 2000, a Sound Transit handout dated 2/2/2001
65 Regional Transit Long-Range Plan, Draft SEIS, Sound Transit, Dec. 2004.
End of Report