Public Interest Transportation Forum -

The Sound Transit Initial Segment Light Rail Project: "Not Recommended"

by John S. Niles

An essay from 2003

Click here for the version published as a Policy Note from Washington Policy Center
Data and sources are below, hot linked to end note references.

Federal, state, and local officials who support Sound Transit’s $2.4 billion dollar plan to build a 14 mile light rail starter line from downtown Seattle to Tukwila frequently trumpet its "highly recommended" rating. For example, one elected official declared on July 24 that light rail in Seattle "will help solve the region's transportation mess and is ‘highly recommended’ for federal transportation dollars[1]." The rating is based on data submitted by Sound Transit to the Federal Transit Administration (FTA) New Starts funding program.

However, close examination reveals that Sound Transit and FTA are using tainted information to make the Seattle light rail project look far better than it really is: Sound Transit light rail should actually be rated "not recommended."

FTA relies on data submitted by agencies like Sound Transit to help decide which construction projects get federal funding. The ratings are then reported to Congress[2]. A project rating of at least "recommended" is needed for federal grant approval, although it is not a guarantee of funding. A rating of "not recommended" means no New Starts money will be made available, and usually prompts a local transit agency to redesign or drop a proposed project.

In other words, if Sound Transit light rail were rated "not recommended" as it should be, that would be a project stopper. This rating would foreclose federal funding for the first phase of a 24 mile, seven billion dollar fantasy system[3], the most expensive light rail project ever served up in America. Congress would have the opportunity to send Sound Transit back to the drawing board, and shift the $409 million remaining in the grant allocation to more deserving projects.

Sound Transit has been able to produce a high rating for its light rail by manipulating the data and calculations that stand behind a measure called "transportation system user benefit[4]." In order to explain this output from a complex computer model of 2020 travel behavior, FTA approximates this measure with traveler time savings.

Under FTA rules, to obtain federal funding the high cost of Sound Transit light rail must be justified by forecasting significant door-to-door time savings for all regional travelers, compared to the same travelers having an improved bus service[5]. The greater the average time savings that can be shown for rail service compared with a strong bus alternative, the higher a new project is rated for cost-effectiveness.

In 1999, before multi-billion dollar light rail cost overruns were revealed, Sound Transit calculated that average travel time savings would be at most seven minutes per trip for light rail versus an all-bus system[6]. However, by 2001, when a more affordable and lower performing light rail initial segment was selected, Sound Transit downgraded the design and performance of the all-bus alternative[7]. The scaled-down Initial Segment rail plan still came out looking good compared with a low-performing all-bus alternative. This makes light rail appear more attractive in the federal project rating.

In February 2003, the Sound Transit section of the FTA New Starts Report to Congress provided data based on the new, lower-performing, all-bus baseline[8]. It yielded the "highly recommended" rating for Sound Transit light rail. The average time saving per light rail rider per day in 2020 now computes as 22 minutes[9], three times the previously seen seven minutes. (This can be derived from the Cost per Transportation System User Benefit of $16.27, a cost-effectiveness measure.) The claimed 22 minute savings is confirmed by its correlation with the assertion from the FTA Administrator in a July 11, 2003 letter to members of Congress[10] that "the daily travel time savings for the projected 42,000 daily light rail commuters will be equivalent to nearly three work-weeks each year."

At best, the jump claimed by Sound Transit in the rail-over-bus time savings from seven minutes to 22 minutes is a technical error not caught by FTA. At worst, it is the result of collaboration between FTA and Sound Transit [11] that bypassed Congress’s responsibility to protect taxpayers from wasteful and fraudulent spending.

Either way, the 22 minute savings allocated to each train rider is crucial to Sound Transit’s "highly recommended" project rating. If the 22 minutes were reduced to the seven-minute average savings estimated earlier, "cost-effectiveness" and "mobility improvement" measures would drop to "low" and Seattle light rail’s overall rating would fall to "not recommended" under FTA’s own regulations and guidelines[12].

The misrepresentations don’t end there. Sound Transit also claims that 15 light rail trains serving a dozen stations would somehow provide the region with more transit capacity and higher ridership than 232 additional express buses using the downtown tunnel and over 200 miles of freeway HOV lanes[13]. But mounting experience from around the country, reported by U.S. General Accounting Office (GAO)[14] and others, shows that buses organized smartly into Bus Rapid Transit (BRT) can provide the capacity of light rail.

This growing reality casts further suspicion on the validity of Sound Transit’s negative assessment of buses. Indeed, to protect the light rail project rating, Sound Transit explicitly excluded BRT as an alternative[15]. Instead, Sound Transit has deployed the baseline buses poorly to give light rail a small daily ridership advantage.

There is nothing new about local officials setting up a low baseline for buses in order to make a gold-plated rail project look better. In 1992, Harvard economist John Kain wrote a paper published in American Economic Review titled "The Use of Straw Men in the Economic Evaluation of Rail Transport Projects[16]." He noted, "Nearly all, if not all, assessments of rail systems have used costly and poorly designed all-bus alternatives to make the proposed rail systems appear better than they are. In some cases, the use of badly designed alternatives is intentional, while in others a lack of interest in developing better bus systems may account for the inadequacies of the all-bus alternative."

Before Congress considers devoting half a billion dollars to begin a questionable project, it should require an independent audit of the baseline data submitted by Sound Transit. FTA regulations require an honest assessment of "the best that can be done for mobility without constructing a new transit guideway[17]." How close is Sound Transit’s straw man to the worst that could be done?

With reform of the FTA New Starts program looming in an upcoming reauthorization of transportation spending, the Seattle light rail project stands as a troubling example of what’s wrong. A costly boondoggle has almost slipped through, despite FTA’s responsibility to be a steward of the nation’s federal transportation resources. Based on FTA’s own criteria, the Sound Transit light rail project should actually be "not recommended," and not a project worthy of federal support.

Washington Policy Center is an independent, nonprofit, 501(c)(3) research and education organization. John S. Niles ( is Senior Fellow for Technology and Transportation at Discovery Institute, and Technical Coordinator at Coalition for Effective Transportation Alternatives (CETA).


1. The quote is from Senator Patty Murray, reported by Jeff Switzer in the King County Journal of July 25, 2003, "House budgets just $15M for Sound Transit light rail," available at

2. The section of the Annual Report on New Starts of February 2003 that describes project evaluation and rating is at  

3. The estimated cost of the full 24 miles has not been announced by Sound Transit. The Coalition for Effective Transportation Alternatives (CETA) seven billion dollar estimate is detailed here.

4. Transportation system user benefits is calculated by a computer program called SUMMIT, which is partially described at

5. Development of the all-bus alternative is described on the FTA web site on a page called Definitions of Alternatives.

6. The seven minute travel time savings comes from inspecting Table 4.2i, "Comparative Analysis of Year 2020 Average Door-to-Door PM Peak Travel Times for Riders Residing in Each Cluster (Work-Place to Home)" on page 101 of the Transit Ridership Forecasting Technical Report, October 1999, part of the Final Environmental Impact Statement for the Central Link Light Rail Transit Project of Sound Transit.   Seven minutes is the travel time difference between the Baseline and Locally Preferred Alternative (LPA) in the Beacon Hill/McClellan Neighborhood Line.  (40 minus 33 equals 7.)  This is the largest travel time savings that can be seen in this table.  Other travel time savings are smaller, or negative; that is, light rail provides a slower average trip than the Baseline alternative.  A pdf excerpt of selected pages from this report, including the Table cited, is posted here

7. The low performing character of the all-bus baseline alternative developed in 2001 is seen in the travel times expected from this system as reported on page 40 of the New Starts data submission by Sound Transit, October 2001. Sound Transit reports that in 2020, on-board transit travel time between downtown Seattle and Rainier Valley would be 17 minutes on light rail, and 35 minutes by bus.  From Rainier Valley to Sea-Tac Airport, the in-vehicle and transfer travel time would be 28 minutes by light rail, and 66 minutes by bus.

8. The "highly recommended" rating for Central Link Initial Segment in the Annual Report on New Starts is at

9.  The FTA New Starts Report of February 2003 reports the incremental cost per incremental trip is $15.58 and the November 2001 New Starts data submission reports the 2020 annual incremental transit trips is 4.87 million. Thus, $15.58 X 4.87 million = $76 million incremental annualized cost, which is also shown in the Sound Transit New Starts data submission of November 2001.  The New Starts Report also has the cost per transportation system user benefit as $16.27/hour. Thus, $76 million / $16.27/hour = 4.7 million travel hours saved. Then, 4.7 million hours / 42,500 daily boardings (per New Starts Report) = 111 hours.  If we allocate 111 hours to the 305 days that each of the 42,500 riders is said to travel (per the New Starts data submission), we get 22 minutes saved per light rail trip, each way.

10. FTA Administrator Jenna Dorn's July 11, 2003 transmittal letter of Sound Transit's FFGA to members of Congress, is posted here.  The Administrator's claim of light rail providing an annual average benefit of "almost three work-weeks," or almost 120 hours, for every light rail user is very comparable to 111 hours or 22 minutes per trip. This claim is critical to establishing the equivalency of traveler time savings to transportation system user benefit, and the argument in note 12 below.

11. Sound Transit frequently describes itself as in a "partnership" with FTA and other Federal organizations to qualify Central Link Light Rail for Federal funding.  An example of this language is seen in Sound Transit Executive Director Joni Earl's memorandum to the Board of Directors on November 21, 2001, where she wrote under the heading of  Federal Funding Partnership:   "The Board would also designate the initial segment as a new Minimum Operable Segment (MOS), a necessary prerequisite to renegotiating the federal funding commitment with our federal partners. We have maintained close contact with our federal partners throughout the year – including the Secretary of Transportation, the Federal Transit Administration, the Inspector General, our congressional delegation and congressional staff of the authorizing and appropriating committees. They are well aware of the Board’s extensive work to select which Link segment to build first. With the new MOS, our formal efforts to rework the technical details of the federal partnership begin." Emphasis added. The Sound Transit partnership with FTA is the closest of all those named.

12. The Summary Rating of highly recommended for Link Light Rail Initial Segment is a combination of Project Justification Rating and Financial Rating. Both need to be medium-high in order for Link IS to be "highly recommended."   The inaccurate rating of Central Link Initial Segment is the Project Justification of medium-high, now calculated as follow:

Project Justification Rating is a 50-50 combination of Cost Effectiveness and Land Use, with Mobility Improvement as a tie breaker if necessary. Cost effectiveness is the Cost per Transportation System User Benefit of $16.27/hour (smaller is better). Mobility Improvement is Transportation System User Benefit Per Project Passenger Mile (in Minutes) of 3.2  (bigger is better). Other data submitted by ST to FTA show that these two number translate to 22 minutes saved per boarding. At present, as reported in the New Starts Report, the Cost Effectiveness is medium-high and Land Use is medium-high. The average of these two is medium-high, and thus the Project Justification is medium-high.

However, based on traveler time saving dropping from 22 minutes to seven minutes, which signifies a proportional change in Cost Effectiveness and Mobility Improvement, the Cost Effectiveness triples to above $48/hour, which is low, and Mobility Improvement drops to below 1.1 minutes, which is low or low-medium. The average of low Cost Effectiveness and medium-high Land Use drops to between medium and low-medium.  However, the tie-breaker of  Mobility Improvement is now low-medium or low, and thus, Project Justification drops to low-medium, which makes Link Initial Segment "Not Recommended." 

13. This comparison is documented in the CETA Report to Congress, May 2003, posted in full text at the CETA web site.

14.  A September 2001 report from U.S. General Accounting Office examines Bus Rapid Transit and finds it a promising alternative to light rail.  Report is available to download at

15. King County Metro Transit, in a November 2001 policy paper proposing BRT plans for 2002-07, noted that corridors that do or would compete with light rail were eliminated.

16. American Economic Review, Vol. 82, No. 2, Papers and Proceedings of the Hundred and Fourth Annual Meeting of the American Economic Association (May, 1992) , pp. 487-493.

17. This quote is from the FTA web site, on the page Definitions of Alternatives.

Return to the Public Interest Transportation Forum home page.

Hit Counter   Last modified: February 07, 2011