Friday, November 8th, 2013

Dude, Where Are My Cars?

Clark Williams-Derry, a blogger with the Sightline Institute, has been running a blog series called “Dude, Where Are My Cars?” which examines the increasing disconnect between traffic projections and traffic reality. The Sightline Institute is into sustainability advocacy, so would naturally have an anti-car POV, but there’s some interesting stuff in there. I was particularly struck by this graphic that’s been making the rounds:

In other words, the Washington DOT continues to use basically the same upward slope for future traffic predictions despite the fact that traffic has been on a steady downward slope since 1996. They just slide the origin point down the actual curve. Williams-Derry straightforwardly calls this “B.S.” He then proceeds to give a point by point rebuttal of the DOT’s claims regarding their projection.

He has strong words and in my experience they are somewhat justified. I try to give people the benefit of the doubt, but that’s becoming harder and harder to do in these cases. However, lest we be too harsh, keep in mind that transport is highly politicized, and staff bureaucrats, many of whom are no doubt not particularly well-paid, are under enormous pressure to toe the party line set by the political overlords. It’s not realistic to expect most of them to stick their necks out, particularly when there’s so little chance it would actually affect the outcome in any case. Like with Jay Carney, the job of a spokesman is to aggressively promote and defend the boss and his policies, not to go on some Socratic quest for the truth.

This also shows how incredibly difficult it can be refute DOT claims. Their traffic models are basically black boxes. As we all know, garbage in, garbage out. It’s very easy to tweak parameters in any model to get the results you want. Also, it’s not always obvious what the parameters in fact are. While some items like population and job growth can be fairly easily analyzed and critiqued, it’s usually hard to get a grip on much of the model. If they say the number in 2040 is X, how do you know if it’s too high or too low? Most of the time you don’t.

Thus the person who wants to try to understand if these are legitimate versus manufactured numbers has to read through reams of dry, technical material to accomplish one or both of two things (assuming there’s a problem): show inconsistency between numbers, or draw an easy to understand picture for the public of what the numbers actually imply if you believe them. One the latter point, much like some companies that disclose troubling information right in the prospectus on the assumption no one who happens to read it will notice or care, sometimes the dirt is right there for anyone who wants to read through some material. I’d put the recent Louisville bridge study into that category.

Unfortunately, few people have the time or inclination to do this. And with the state of newspapers as they are, there are very few transport reporters who are able to do a real analysis. This means that when you read the average article about transportation in the newspaper, it’s usually little more than a re-written DOT press release. Given that the papers will print whatever the DOT spokesman tells them uncritically, why not take advantage of that to the max? No surprise, that’s exactly what they do.

Topics: Transportation
Cities: Seattle

13 Responses to “Dude, Where Are My Cars?”

  1. Jonathan P says:

    For someone who has worked with transportation models before, it’s a challenging issue. Models must be based largely on empirical data, which are inherently backward looking and slow to react to changing trends due to the time needed to collect and analyze long term, sustainable trends. Yet there is growing evidence that we are perhaps at or past an inflection point where the past economic and travel trends (predating the Great Recession) are taking hold. Yet, still models and the information they produce, are in greater demand, as officials (rightly) request more precision of potential benefits and impacts of these projects.

    I think the larger problem is that there is too much power in the hand of these large state DOTs. The failure to address real transportation reform has nearly all federal money flowing to them, rather than the metro regions, where real solutions are being developed. It often takes decades to plan and build a very large transportation project, so you have these legacy highway projects that the DOTs are dusting off the shelves (now with tolls added to help the bottom line). And the DOTs are led by Governors’ simply looking for ways they can show they’re creating jobs through “shovel ready” public works projects. It’s a perfect match. The models are just the tool that the DOTs use to show that a “3rd party” (a computer) confirms that their project will be a success.

    I don’t favor full transportation devolution, because I believe there is a role for the federal government to play. However, if Congress cannot act, we might see fewer destructive, wasteful projects if we simply have the federal government establish national policies, set performance targets, and restrict funding to maintenance, some regional dollars to MPOs, and perhaps a few mode neutral discretionary programs in the TIGER and New Starts mold.

  2. Jonathan P, thanks for the comment. I can appreciate the difficulties. However, I frequently encounter assumptions used in the forecast, such as job and population growth numbers, that are significantly off-trend from the historical average with no explanation.

    What’s more, many of these DOTs easily walk back high projections to much lower values – after the project they want is approved. I’ve highlighted recent examples including Louisville, Portland, and Maryland where there were significant reductions in the forecasted traffic level released post Record of Decision.

  3. Nathanael says:

    Clark Williams-Derry has pretty conclusively documented bad faith in WSDOT’s projections.

  4. Rod Stevens says:

    There’s a governance issue and there is a disclosure issue.

    The governance issue is that unless there is a compelling regional or national interest, the federal government should give out transportation money on a per-capita basis and let cities and states decide how they use it. Right now our system awards places with the most political skill, regardless of the inherent need. Since 90% comes from the federal government, local jurisdictions have little incentive to use it well.

    The disclosure issue is linking projections to actual experience. Showing the future and the past on the same graph would reveal the hockey-stick nature of these projections. Like this graphic, they should also show previous projections, so that readers can judge for themselves how good the agency has been at making projections.

  5. Ben Leis says:

    Although I’m not that impressed with WSDOT’s traffic projections some special circumstances surround the 520 floating bridge. Tolling started back on the bridge in 2010. This had the initial effect of shifting traffic from it to the I-90 bridge. Neither earlier forecast had that built into its model. I’m not certain but the 2011 model may have the assumption built back into that I90 tolling will soon be instituted as well (which is fairly likely) and at that point some of the aritficial traffic distortion that current tolling situation engenders will reverse.


  6. Mike says:

    I think the key question is, have we really reached “peak car” in major cities, and why. Is it a trend of certain demographic groups eschewing car dependence, which could be reversed as they age and their families grow, or is congestion pusching all demographics off the roads. In the latter case, it signals a clear need for better transit, particularly rail with higher capacities. In the former, there is also a case for better transit, although it’s a more tenuous case where more frequent buses and “BRT features” may be adequate to meet this cyclical incrase.

  7. Mike, I think part of the issue that VMT was historically increasing at a rate greater than population. I think a lot of people just assumed that would happen. But now that we’ve got as many cars as drivers in the US, women have already entered the workforce, etc. many of the things driving that growth are played out. I’m not sure we’ve hit “peak car”, but possibly with short periods, I don’t see us going back to hypergrowth in auto traffic.

  8. Chris says:

    Also remember, these models are based on 30 or 40 years of past actual data, which would swamp the current trends. Assuming that current trends continue, this would mean that our forecasting models will start to reflect those changes over time.

    Additionally, remember that the DOTs are not responsible for the creation of the employment and population forecasts that are used in their models, thos3 come from other sources. They can, and should, be challenged on the VMT per household and cars per household assumptions that are used on top of those variables, but we’re notoriously bad on all of our employment and population forecasts because we assume that what is happening today will continue 20 years out. Unless we come up with better ways of forecasting the baseline assumptions, the GIGO principle will continue to apply.

    Also, the political point is well taken. Any model has to be viewed with an eye towards “is this something people would really put up with?”. If the answer to that question is no, then the model is not accurately depicting human behavior, which in the end, is what this all comes down to…

  9. @Chris, again, I’ll bring up that these models seem to change their output depending on which side of the approval divide they are on. In the case of major bridges in Louisville and Portland (CRC), there were large traffic increases forecasted in the FEIS. In the Louisville case, a supplemental EIS lowered the traffic forecast significantly though it did not reevaluate any alternatives in light of that (only the original preferred alternative versus their modification). In both cases also the investment grade toll study showed vastly lower projected volumes. In the admittedly controversial case of the Intercounty Connector in Maryland, it was the exact same scenario, only that road is built and traffic is below even the investment grade toll study. It’s now said to be exceeding expectations based on an even lower forecast made right before opening.

    Bent Flyvbjerg at Oxford seems to have demonstrated such a bias in over-projecting traffic (and, in fairness, transit ridership) in major projects that he believes it could only be accounted for through dishonesty:

  10. Phineas Baxandall says:

    Clark has done great work on this issue. There will be ongoing uncertainty about VMT trends. A good standard would be if DOTs didn’t prioritize projects that only make sense if there is a reversal in the decade-long trend toward lower VMT. A cautious approach would be to prioritize projects that make sense in a range of scenarios, including stagnant VMT and further declining VMT as well as back-to-the-future scenarios of rapid VMT growth.

  11. @Phineas Baxandall, sounds like a good idea.

  12. Nathanael says:

    There are certainly some projects which make sense with declining VMT as well as with increasing VMT. Complete streets projects (everyone needs sidewalks), grid reconnections, bridge structural projects, etc… perhaps these simply don’t involve enough contractor pork, or perhaps they’re not “sexy” enough, so most DOTs just don’t give a damn.

  13. TRANSDEF says:

    Phineas is being all too modest here. His team at US PIRG did a great study on millennial preferences and peak driving that recommended: Don’t assume the VMT trend of the past will return!

    Highly recommended. Amazingly sane!

The Urban State of Mind: Meditations on the City is the first Urbanophile e-book, featuring provocative essays on the key issues facing our cities, including innovation, talent attraction and brain drain, global soft power, sustainability, economic development, and localism. Included are 28 carefully curated essays out of nearly 1,200 posts in the first seven years of the Urbanophile, plus 9 original pieces. It's great for anyone who cares about our cities.

About the Urbanophile


Aaron M. Renn is an opinion-leading urban analyst, consultant, speaker, and writer on a mission to help America’s cities thrive and find sustainable success in the 21st century.

Full Bio


Please email before connecting with me on LinkedIn if we don't already know each other.



Copyright © 2006-2014 Urbanophile, LLC, All Rights Reserved - Click here for copyright information and disclosures