Friday, June 10, 2011

William Grindley Brief Note #2: HSR Ridership forecasts and why they lie

There's been an enormous amount of discussion about ridership projections for the California High-Speed Rail project.

Before and during the election season of 2008, when the high-speed rail bond issue was going to be on the ballot, the voters were lied to in a number of ways, ridership being one of them.  They (we) were told that there would be 117 million annual riders. That's one third the population of the US.  The annual ridership of ALL of Amtrak throughout the entire US annually is today at its highest;  29 million annual riders. So, you can see where a forecast of 117 million is pretty silly.

After the election, the CHSRA started to reduce those numbers, in each case citing the new and improved research of what they liked to call the best and most expensive research organization in the world, Cambridge Systematics, Inc.  To make a long story short, today's ridership forecasts are down to around 40 million annually.  And even that number is being rejected as far too exaggerated.  Consulting firms are notorious for telling their clients what they want to hear, not what they should hear, which is the truth.

Therefore, we still don't have plausible or realistic ridership projections for the California project.  Are these numbers important?  Grindley tells us that they are critical to the "financial credibility" of the rail authority and the project.  We learn that pumped up numbers is common practice and that in reality, when trains are finally operating, there are always far fewer riders, less revenue and greater costs to the state to operate those trains.  

China is a great example.  They have been sending trains whizzing across China nearly empty because so few can afford the expensive tickets.  Why will it be any different here? Shouldn't we know that in advance so that we can ask if there is a reason to build this luxury train at all?

Brief Note #2 – June 6th 2011

From the authors of The Financial Risks Of California’s Proposed High-Speed Rail Project and six Briefing Papers. Available at
Finding: Reasonable ridership forecasts are crucial to financial credibility
Background: There are consistently far fewer riders than ridership forecasts. In the 2003 seminal study of megaprojects, the authors stress: ”. . (rail) forecasts were overestimated on the average by 65%.”1 The World Bank’s 2010 high-speed rail report concluded that, “High- speed projects have rarely met the full ridership forecasts asserted by their promoters . . .”2
In 1992, the Eurostar forecasted “15 million passengers per annum in 1995 and growing”. In 2009 Eurostar carried 9.2 million passengers, only 60% of that forecast. 3

A US DOT study found that actual rail ridership forecasts were only 39% of forecasts. Bay Area Rapid Transit’s (BART) forecast for the SFO airport was for initial (2003) daily ridership of 39,500 – and 68,600 by 2010. The opening year daily ridership averaged 16,500, 42% of projections. Daily ridership in 2009 was still less than 17,000, only 25% of the 2010 projection.4

In 2008 the California High Speed Rail Authority (CHSRA) asserted there would be over 90Million train riders annually on the LA to SF route.5 They reduced their 2009 forecast for 2030 to 39Million riders.6 The more than a 60% decrease was challenged by, among others, a 2010 Senate-commissioned study by UC Berkeley’s Institute For Transportation Studies.7 

Even the lower 2009 CHSRA projection inflated some stations’ boardings, artificially increasing revenues, justifying specific routes and therefore financial credibility.8

A pragmatic forecast might use the USA’s cousin to high-speed rail, Acela. In 2009, Acela’s total ridership (Boston-NYC-PHL-WDC) was 3 million, about 11% of the 28Million in nearby metropolitan areas.9 To attract 11% of California’s 48Million residents in 2030, the train would have to carry about 5Milion riders, only 13% of CHSRA’s forecasted riders.10 The Megaprojects book’s findings suggest the ‘overshoot’ indicates attracting about 11 Million riders in 2030.11

These ‘forecasted-to-actual’ examples show actual results to be 30-40% of forecasts; suggesting 2035 ridership of about 14 Million, not 41 Million passengers.12 The CHSRA 2009 Plan shows that about a third of their 2035 passengers would come from the airlines and two-thirds from autos and local rail; suggesting about a total of 14 Million from both, not that Plan’s 41 M. That range, from 12M to 16M, is consistent with actual results and pragmatic analyses.

Conclusions: Something is amiss. The present CHSRA-captive ridership reassessment study group will likely produce results similar to the 2009 version. This cannot go unchallenged. The Senate must require and commission an independent-of-the-Authority ridership reassessment. To regain credibility the project must have believable ridership numbers as its entire financial plan presently depends on a false and misleading ridership forecast.


1. Flyvbjerg, Bent; et al: Megaprojects And Risk, An Anatomy of Ambition; Cambridge University Press, 2003; pg. 26.

2. Paul Amos, Dick Bullock and Jitendra Sondhi; World Bank Report No 55856; July 2010; pg.14.

3. Op.cit. Flyvbjerg et al. for both the Eurostar quote at pg. 22 and page 25 for the DOT quote

4. See:

5. 2008 California High-Speed Train BUSINESS PLAN, November 2008; pg 7.

6. California High-Speed Rail Authority (CHSRA): Report to the Legislature; December 2009; page 73

7. Samer Madanat; Director, UC ITS Berkeley; found at and CARRD Ridership Comments; April 26, 2010 at

8. Both the 2009 SF and Anaheim boardings include, that is, ‘double count’ passengers from Oakland and San Diego where stations are to be constructed in later phases. See Bay Area to Central Valley High-Speed Train: Revised FINAL Program Environmental Impact Report; Volume 2: Response to Comments; August 2010, California High-Speed Rail Authority Page 1082. At  

9. Table in “Amtrak Fiscal Year 2009” Oct. 2008-Sept. 2009.

10. See:

11. Op.cit Flyvbjerg et al. pg. 25. 12 Op.cit Report to the Legislature; December 2009; Figure 1, page 71

No comments: