Frontier Group


Widgets, Systems and Missing Trends: The National Academy of Sciences Report on Gasoline Savings

Not only did the National Academy of Sciences panel fail to recognize the potential for systems-level changes in the future, but it also failed to account for a seismic change that is going on right now: the unprecedented decline in per-capita vehicle travel in the United States.

David Roberts at Grist once again beat me to the blogging punch, this time with his take on the recent National Academy of Sciences (NAS) study on possible strategies to achieve dramatic reductions in gasoline consumption by cars and trucks.

Roberts gets at an important point that’s long been a frustration of mine in our work to quantify the oil savings potential of transportation policies: specifically, the degree to which conventional policy analysis fails to account for the transformative potential of systems-level changes, as opposed to efforts to improve the individual components of those systems, which Roberts terms “widgets.”

Here’s an illustration: suppose in the mid-1980s you were evaluating policies to dramatically reduce consumption of paper for newsprint by 2013. You might think of ideas such as putting a tax on newspapers to make them more expensive, establishing mandatory newspaper recycling, or requiring publishers to use smaller print or lighter paper. All of which would be reasonable, effective alternatives whose results could be estimated based on economic models or simple math.

Imagine, however, you were to say, “Hey, why not build out a global communications and computing network that will enable anyone to get news from anywhere in the world in a fraction of a second. Oh, and create something called Craigslist while you’re at it.” You’d probably get some funny looks from your fellow policy wonks, but more to the point, you would have no earthly way of calculating or estimating what the implications of that change might be. If you guessed that the result of the invention of the Internet would be a roughly 60 percent decline in newsprint use, you would have turned out to be right, but no “reasonable” person at the time would have believed you.

The NAS study suffers from this problem in spades. As Roberts notes, it fails to grasp – because, within the analytical boundaries of such projects, it simply can’t – the potential for all sorts of systems-level changes in transportation, from changing patterns of vehicle usage to self-driving cars. All of which makes the task of replacing oil in our transportation system appear to be much tougher and costlier than it probably is in reality.

I have a couple of additional beefs with the study. First, not only did the NAS panel fail to recognize the potential for systems-level changes in the future, but it also failed to account for a seismic change that is going on right now: the unprecedented decline in per-capita vehicle travel in the United States. The study uses, as its reference case, the Energy Information Administration’s (EIA) 2011 forecast of increases in vehicle travel through 2035. Not only is that forecast’s projection of future driving too high (as the EIA’s more recent forecasts reflect), but the NAS panel compounded the problem by assuming that vehicle travel continues to grow at the same 1.5 percent annual rate projected for the 2030-2035 all the way to 2050. Yep, that’s right, they extended the line.

My second beef is that the NAS panel, like anyone else looking to estimate the benefits of policies to reduce driving, is forced to rely on data from the past to indicate what might happen in the future.  As investment prospectuses always point out, however, “past performance is no guarantee of future results.”

For example, the panel, citing an earlier study, found that if 75 percent of new development were to be high density, overall vehicle travel in the U.S. could be reduced by 11 percent. The NAS study, however, uses language that suggests that this scenario is well-nigh impossible, calling the shift to higher density development an “extreme reorganization of national economic activity” in the context of development patterns of the last 30 years.

But here’s the thing: that reorganization is already happening. And if you pay attention to real estate market observers like Arthur C. Nelson at the University of Utah, who believe that we are headed toward a glut of sprawl-style housing as a result of demographic and other trends, the idea that 75 percent of new development will be of higher density – even in the absence of strong public policy – sounds a lot less outlandish.

My point here isn’t to bash the NAS study or the people who wrote it, many of whom are among the smartest “widget” guys and gals in the business. After all, it wouldn’t take you more than a few minutes of flipping through the Frontier Group oeuvre on this website to find much of the same thinking at play.

Rather, my point is to show how awareness of the recent fundamental shifts in driving behavior have yet to filter through even to otherwise learned institutions like the NAS – and how the failure to account for those important changes has the potential to throw all sorts of policy decisions off kilter. At the very least, future studies like the one issued by the NAS panel need to include sensitivity cases that posit the possibility that the driving trends of the future won’t match those of the past. Ideally, though, we would all start to develop the knowledge needed to understand how changes in consumer preferences (particularly among youth), technology, demographics and other factors will influence vehicle travel down the road.



Tony Dutzik

Associate Director and Senior Policy Analyst, Frontier Group

Tony Dutzik is associate director and senior policy analyst with Frontier Group. His research and ideas on climate, energy and transportation policy have helped shape public policy debates across the U.S., and have earned coverage in media outlets from the New York Times to National Public Radio. A former journalist, Tony lives and works in Boston.