Sunday, June 15, 2008

Close But Not Quite in Modesto

NOTE: We've moved! Visit us at the California High Speed Rail Blog.

It's HSR editorial season, it seems, and the latest entry comes from the Modesto Bee. It's certainly not the sad joke of an editorial the Contra Costa Times put out, but it still demonstrates many of the common flaws the state media uses to assess the project - particularly an inability to examine the project in context. The Modesto Bee at least opens the editorial mentioning $4.50 gas, but unfortunately they don't consistently apply that all-important point in their commentary. A quick examination follows.

the Senate Transportation Committee conducted hearings on the rail concept in December and January, and recently released a report with six recommendations -- none of which are yet addressed in Galgiani's bill. Perhaps the most important is to require the commission to produce a business plan "consistent with a standard financial prospectus." We also agree with the committee that construction needs to take place on regional segments before the long-distance rail is built.

I've consistently been arguing against using that report as a guide for assessing the HSR project - the report is very deeply flawed, especially in its total failure to assess HSR in the context of permanently high fuel costs and the financial risk that poses to the state. I also have argued against lowering the priority of the long-distance segments unless there are ironclad guarantees that those WILL be built after the regional segments. Without such guarantees there is no point to approving the bond.

Quit low-balling. The cost for this all-new railroad has been set at $33 billion, with $7 billion more needed to add service from Merced to Sacramento and from Anaheim to San Diego. These estimates haven't changed in six years -- and no one believes them. A 2005 World Bank policy study shows that major rail projects usually cost 44 percent more than budgeted.

This is not a useful statement. What did the World Bank study point to as the cause for the increased costs? Rail projects in California such as the Metro Gold Line are being delivered on-time and on-budget. Without an understanding of the mechanisms behind those cost increases it is dishonest to use that study to predict the HSR project will go over budget.

Who will pay? Proponents say the federal government will be an important partner, but Wednesday the House of Representatives approved only $1.75 billion -- $350 million a year over five years -- for all the nation's high-speed rail projects. The rest, say backers, will come from private investors. But large banks and hedge-fund investors are reeling from the mortgage crisis. That leaves taxpayers and riders.

This is deeply misleading - and I'm putting that generously. The House bill was an Amtrak bill. The HSR money was gravy, unexpected and unplanned. It was thanks to the work of Jim Costa, Fiona Ma and Nancy Pelosi that the money is there at all - but it is NOT intended to be the final amount of funding Congress will give HSR. In April we reported that Congress is anticipating $60 billion for HSR in the 2009 transportation bill. Barack Obama is a vocal supporter of HSR and would likely support and help pass such an investment. There is NOTHING to suggest $1.75 billion is all Congress will give to HSR. The Modesto Bee surely knows this, or they should have done their research before publishing that part of the editorial.

As to the banks and large investors, the credit crisis won't help. But neither is it crippling. There is a lot of overseas capital looking for long-term stable investments, and CAHSR would be a natural site for them to park their money. And if they don't then we can and should sink more money into it. Of course the Modesto Bee, like every other media outlet that has pontificated on HSR, has refused to ask what the cost of doing nothing is - it is certainly higher than the cost of building HSR.

Get real about ridership. Proponents say that by 2030, they expect to generate $1 billion in profit based on 100 million riders a year. That means every Californian would have to ride the train three times a year. When asked about the validity of such numbers, a person familiar with the details whispered, "crazy." Another called them "black box" figures.

This is more of the anonymous sourcing and vague reporting that the media is way too often passing off as journalism. Who claimed the numbers were "crazy"? How is the public to assess the validity of such a statement? The Modesto Bee is irresponsible in not sourcing or explaining that claim.

Further, who says HSR will only be used by Californians? We do still get a lot of tourists, many of whom will use the system. Business travelers will use it far more often than three times a year.

But the most important aspect of the ridership question is fuel costs. The Modesto Bee led off the editorial mentioning high gas prices - so why ignore it here? Ridership may not hit exactly 100 million by 2030 - though I am certain that it will, as ALL HSR lines around the world have rapidly hit their ridership expectations - but with permanently high fuel prices and the resulting airline crisis it strains credibility to not expect ridership to be sufficient and high.

Be careful where you dig. The commission has hired Parsons Brinkerhoff to manage the program. Earlier this year, the company was forced to pay $458 million for problems with its "Big Dig" project in Massachusetts, whose final cost ($14.6 billion) doubled original estimates.

I was not under the impression Parsons Brinkerhoff had been given any permanent contracts. We plan to watch this aspect of things quite carefully.

The Modesto Bee concludes "High-speed rail is a good idea that should be pursued." We strongly agree. But we wish the Modesto Bee would take a more informed approach to assessing the project - and stop trafficking in unsourced, unverified, specious claims in order to raise doubts about the project.


Kevin Gong said...


I suppose it is only natural for most citizens to assume that large complicated infrastructure projects like CAHSR will not meet expectations. After all, infrastructure is often boring and we only hear about it in the news when something goes wrong with it.

As a vocal CAHSR blogosphere proponent, are you writing letters to the editors to these various papers to share your views?

Rafael said...

@ robert cruickshank -

it's true that some transportation infrastructure projects come in on or even under budget. Sadly, they tend to be the exception rather than the norm. It's understandable that many observers are taking CHSRA's estimates of construction costs with a large grain of salt, though it is unfair to dismiss them a priori solely on the basis of negative experience with other projects.

In a seminal 2002 study, Flyvbjerg et al. provided some insight into just why so many transportation infrastructure projects all over the world do go over budget. Project size and duration were found to be the strongest indicators of cost escalation risk. That's mostly because they are much harder to plan.

Once ground is broken, planning mistakes, unforeseen turns of events and especially, feature bloat - i.e. backers saddling a project with additions they could otherwise not get funded - lead to expensive engineering change orders that drive up the final cost. Success depends on quality assurance processes and aggressively defending the scope of the project.

Another important finding was that cost overruns have less to do with who puts up the funds than with how project managers are supervised and held to account.

Finally, politicians and contractors do occasionally - far from always - collude to present the public with low-ball estimates, because they are convinced they would otherwise never win approval to break ground: "You want the truth? You can't handle the truth!" This mentality is especially prevalent when the technology involved has little or no track record.

Case in point: the long-planned maglev line from Munich to its new airport some 26 miles to the north. This was supposed to be a public display of confidence in locally developed high-tech. The country's former prime minister Stoiber felt building one in Germany was critical to developing a brand-new export-oriented industry. Almost immediately after he was forced to resign after an unrelated scandal, the contractors involved in the project doubled their cost estimates to bring them in line with reality. This refreshing bout of honesty led to the rapid cancellation of the entire project.


Implications for the California HSR project:

- California has actually done a lot of things right on the HSR project. It set up a formal planning body that was given ample time develop the route and station choices. Proven technology was selected. Gov. Schwarzenegger avoided aggressively championing the project, keeping it non-partisan.

- however, the state also got a few things wrong. CHSRA was kept at such arms length that no-one made sure it expended enough energy on preparations for raising funds from non-state sources. This lack of project maturity was one reason it was kept off the 2006 ballot, in the full knowledge that this second delay would entail cost inflation. It also led the Governor to cut the Authority's planning budget down to almost nothing for a couple of years.

Meanwhile, CHSRA optimistically assumed that state politicians - especially in places like Modesto - would evaluate their proposals broadly on the eventual benefits to the entire state rather than narrowly on their impact on their constituencies.

In particular, CHSRA did the project no favors at all when it yielded to behind-the-scenes pressure by SF peninsula politicians and prematurely abandoned further study of those Altamont Pass options that do not involve a new Bay crossing. Instead of forcing politicians to tackle head-on the unspoken but legitimate tax base issues posed by facilitating growth in long-distance commuting, it has permitted project bloat via an expensive overlay network. A creative modification to the fare structure would not increase HSR construction cost yet generate funds for local transit services in Central Valley towns.

With funding so tight and uncertain, now is definitely not the time to select any but the cheapest construction option that meets project goals. That means some variation of Altamont Pass via SPML instead of Pacheco Pass.

- it might be wise to heed the State Senate's concerns about project phasing. If each individual construction project is relatively modest in scope and results in an immediate, tangible improvement at the local or regional level, it will be much easier to sustain voter and investor support for the duration of the entire project. Building confidence is especially important early on when funding for the later stages has not been secured yet. That is why tunneling at e.g. Tehachapi may not be the best place to break ground, though there is obvious value in exploratory drilling.

- the saga of HSR in the UK shows that it is essential to think big from the outset, as California has done. Getting the HSR project off the drawing board requires broad voter support. Switching from a single SF-Anaheim starter line to two smaller ones (SF-Sacramento via Altamont and Anaheim-San Diego via LA Union Station) could make a lot of sense. Each of these would cost about half as much as the SF-Anaheim link. Phase 2 would then be expansion down to Bakersfield and up to Palmdale, respectively. Phase 3 would be crossing Tehachapi Pass.

Statewide interest in a north-south HSR link will always be strong, both because of growth constraints at the I-5 grapevine and airports and because it will bind the two halves of the state closer together. That said, tunneling should not be endlessly deferred as it is always a difficult and protracted process. It may be necessary to begin phase 3 before phase 2 is complete.

- retaining the option to pull the rip cord and terminate a grand project like HSR prematurely gives investors, includes taxpayers, a tool to keep costs in check. Ironically, that actually makes it more likely that the project can be completed without breaking the bank.

- however, actually pulling that rip cord does carry high opportunity costs, because the identified transportation demand will then have be met some other way. The State Senate report did indeed neglect to emphasize this essential point. Once planning has reached an advanced stage, the objective should be to keep the project alive - just not at any cost (cp. BART extension to Santa Clara).

- Gov. Schwarzenegger's belief that private investment will automatically increase the chances of success is at odds with Flyvbjerg's findings. While the private sector participation is essential to raising funds in this case, the key is making sure they are well spent requires that CHSRA, regardless of personalities, be more tightly supervised.

The State Senate report suggested handing that task to state bureaucrats, but Caltrans is still a highway-centric organization. An alternative would be to set up a new organization with representatives from all who invest in the project, including some planning and engineering experts with extensive HSR experience in other countries.

This should not be seen as a lack of trust or confidence but rather, as a quality assurance process. Throughout the life of the project, someone needs to challenge assumptions, double-check calculations and ensure that decisions are based on an assessment of their impact on the entire project as opposed to just whatever the next phase is.

Beyond attention to detail, it is essential to keep the project aligned with investor priorities - which may change over time. Of course, CHSRA should resist such changes that would lead to delays and/or cost overruns. Some creative tension is actually desired, but of course, he who pays the piper calls the tune.

Anonymous said...


PB is the project manager for the HSR project. See:

I think the original RFP or RFQ was on the old website but I can't quickly find it on the new website.

Robert Cruickshank said...

kevin, I have been so focused on the blog that I haven't been composing such letters. But I invite readers to do so themselves. Feel free to use my arguments, though not my exact words, in such letters.

michael, thanks for that info. It certainly bears close scrutiny.

rafael, excellent points as always. I agree with most of them, especially that sound project management is vital to bringing a project in on time and on budget. Political meddling, of the sort we've already seen, is what adds such project costs and overruns. My point wasn't so much that overruns don't exist, but that they aren't a guaranteed or natural thing. They can be avoided.

You're obviously more willing to embrace a regional HSR-first approach than I am. My concern with it is that unless there are firm guarantees that such regional routes are but the first phase of a project and that we WILL build the intercity connections, we run a very high risk of never actually getting around to building those connections.

I'm also unconvinced that voters will see regional HSR and understand how it relates to intercity HSR - they serve different rider types and are viewed as different kinds of systems by voters.

Further, how would such a regional HSR system work? Would the CHSRA run it? Or would Caltrain and Metrolink run it? That could lead to compatibility issues down the line.

There aren't any easy solutions here, I recognize, but I still believe that HSR is politically stronger by adhering to the core mission of intercity travel that can double as regional commuter rail as well on its various segments.

Rafael said...

@ robert cruickshank -

my perception is that there is strong demand for HSR service between Sacramento and Bakersfield, not just into the Bay Area. It seems inconceivable to me that an SF-SJ/SC-Sacramento Northern California starter line would not be extended south.

Similarly, Los Angeles is mustard-keen on running HSR tracks up to the fast-growing Antelope Valley and on leveraging Palmdale airport to relieve LAX. Combined with geological and environmental reasons, this is why the route runs across Tehachapi rather than the Grapevine. The associated SF-LA line haul penalty is 10-12 minutes.

Once the network reaches Bakersfield and Palmdale, only a relatively short - if challenging - connector across Tehachapi Pass is required to implement the north-south link. Capacity bottlenecks at the Grapevine and Bay Area airports will ensure that it gets built.


Also, as you have pointed out in the past, commuter and intercity rail aren't necessarily separate entities. It simply isn't true that the Altamont Pass option would only serve regional traffic. CHSRA's claim of a 40 minute line haul penalty relative to Pacheco for SF-LA express trains is completely bogus. The time penalty would be more like 14 minutes if it sticks with SJ Diridon and just 8 if Santa Clara/SJC is used instead.

This would come on top of the penalty already incurred due to Palmdale, but still unlikely to severely impact ridership. Right now, CHSRA projects that downtown-to-downtown travel between SF and LA would be 2 minutes shorter with the train. Switching from Pacheco to Altamont would make it 6 minutes long. Either way, it's a wash.

The beauty of HSR over both cars and airlines is that it can provide fast, safe and convenient service at a wide range of length scales, e.g. 10-600 miles. Obviously, trains that have to stop every 10 miles could not run at high speed, but there are three strategies to address that.

1. Every secondary station on the HSR network will feature bypass tracks for express trains. Where space is at a premium, you need to summon the creature from the third dimension. With appropriate architecture and civil engineering, express tracks can run underneath a central platform served by the slower trains. Note that HSR trainsets can be as long as 1/4 mile (e.g. Eurostar at 394m), but train scheduled to stop at secondary stations can be shorter, especially if bi-level HSR rolling stock (cp. TGV Duplex, Shinkansen E4 MAX) is used. Note that these designs are rated for top speeds of 186 and 150mph, respectively.

Road grade separations in the immediate vicinity of stations with vertically stacked tracks would have to be implemented using over- rather than underpasses. Passengers will also have to access the platform from above. The ticket hall and the forecourt for dropoff/pickup, taxis and connecting buses would usually be next to the tracks, passengers would use the sidewalk on the overpass to reach the platform.

If need be, the ticket hall(s) and forecourt(s) could also be located above the tracks and platform. The forecourt(s) would then be reached via turnoffs located on the crest of the overpass(es) - in locations with high road traffic there might be one at either end of the platform.

2. CHSRA's year 2000 business plan calls for multiple service levels. Express trains will only stop at SF and SJ/SC in the Bay Area and then proceed non-stop to either Sacramento or LA. Semi-express trains will stop at some but not all of the secondary stations in-between.

SNCF implements semi-express service by running trains non-stop for half the route and as locals for the other half. Passengers traveling between secondary locations have to change trains in the middle.

Caltrain has developed a different strategy for it's "baby bullet" trains. All of those stop at only a subset of a all stations, but different trains serve different subsets. This works because of a carefully designed schedule.
There is no reason why you could not apply this concept to HSR trains, at a larger length scale.

For example, one train could run non-stop from SJ/SC to Sacramento, the next one could also stop in Fremont Warm Springs and Tracy, the one after that in Pleasanton and Stockton, the next one at Fremont Warm Springs and Tracy etc.

If required, a dedicated HSR network can support headways as short as 5 minutes without compromising safety. Combined with long and/or bi-level rolling stock, HSR offers massive capacity.

3. During rush hour, there would probably be fewer express and more semi-express trains to better serve medium-to-long-distance commuters. At other times, the ratio will be reversed.

Note that during rush hour only, long-distance trains could run as semi-express between SF and Tracy before continuing non-stop to LA. This would make it easier to fit them into the schedule. My point is, there is plenty of scope for schedule optimization.

Note also that AB3034 as passed by the Assembly requires that there be local trains that stop at every station of the defined sections of the network. This allows passengers traveling between secondary locations to avoid a change of trains. However, the bill does not specify how often such local service would have to be offered, nor at which times of day. Again, plenty of wiggle room.


IMHO, the biggest issue posed by operating HSR primarily as a mix of express and semi-express trains through Altamont Pass is that it would encourage long-distance commuting on a massive scale.

CHSRA's own analysis suggest Tracy to San Jose would take an express train just 25 minutes. A semi-express with one stop in-between would still take less than 30, with excellent on-time performance in any weather. That is a very attractive proposition to anyone who works in Silicon Valley but cannot afford to buy suitable property there.

To give you a sense of scale, a single TGV Duplex trainset offers 500 seats. Run one every 5 minutes between 6:20 and 9:40am and you're up to 20,000 seats capacity - a whopping ten-fold increase over ACE, which would be canceled once HSR is available.

To minimize dwell times, optimize pedestrian traffic on the platforms and avoid overcrowding on the trains, you'd have to require reservations on HSR trains during rush hour. Durable paint marks on the platform would instruct passengers waiting to board exactly where they should stand. Other markings would establish one-way lanes along the platform. If need be, you could even designate some train doors as entrances and others as exits to establish one-way traffic along the aisles on the train.

Commuters typically purchase monthly or annual passes, so their reservation would be free. They'd still need to make it before every journey, using either a web browser or an SMS from a cell phone.

In theory, taking so many people off the roads - e.g. I-580 in Livermore Valley - would be godsend in terms of air quality, energy security and global warming.

In practice, California's tax code would force SF peninsula counties to massively expand their transit services, which are heavily subsidized by locally collected property and sales taxes. However, all those long-term commuters would be paying those taxes primarily where they live, not where they work. Moreover, simply by choosing to live far away, they would reduce demand for housing in Silicon Valley - which means realty values there would fall or at least no longer rise as quickly.

This issue can be mitigated by adopting a two-tier fare structure. Morning rush hour passengers boarding in Alameda, San Joaquin, Stanislaus or Merced county and headed for Santa Clara, San Mateo or San Francisco county would be required to present a combo ticket to the conductor. This would entitle them to use HSR plus connecting local transit at either end. Alternate evidence that they are paying for local transit would also be acceptable. That means ticket dispensers and validators at every HSR station in Northern California should feature touch screens and software that support every transit system, i.e. you would be able to buy a BART ticket in Merced or a THEBUS ticket in San Francisco.

During evening rush hour, passengers in the other direction would be subject to the same fare restriction. At other times of day and on weekends, regular HSR-only tickets would be sufficient.

Legally, combo tickets could be implemented via a special fixed-value sales surtax on top of the prevailing rate. Surtax revenue would be split 50/50 between the counties of origin and destination. Note that it would not be earmarked for local transit but rather go into the general fund. This would give counties the option of investing in (electric) bicycle infrastructure as well as buses, light rail etc.

A similar fare structure could be used within Southern California if desired. The beauty of requiring HSR commuters to purchase combo tickets is that it generates both demand and funding for local transit. This is a prerequisite for successfully zoning areas close to the HSR stations for medium-to-high-rise residential and commercial property.

Finally, consider this: enabling long-distance commuting will kick-start the economies of Central Valley towns. Some Silicon Valley companies will choose to set up back offices and data centers there, because electricity supply growth is not constrained and the earthquake risk is lower. Green collar jobs tied to biomethane and renewable electricity generation will follow. And last not least, encouraging population growth close to sources of water reduces demand for pumping even more of it hundred of miles south and across the Grapevine.