You are hereHome ›
Peak Car, Revisited
Posted by: Tony Dutzik on
I’ve never liked the term “peak car.”
First, it was always unclear exactly what was supposed to be peaking – total vehicle travel, per-capita travel, car ownership, or all of the above? Second, like peak oil before it, “peak car” applies a catchy name to a collection of concepts that are important to understand – taking a useful perspective and turning it into a parlor game or prediction contest.
When we addressed the issue of long-term trends in vehicle travel in our 2013 report, A New Direction, we argued that America had reached the end of what we called the “Driving Boom.” We chose our words carefully, and what we meant by them was this: America had experienced a historical period from the end of World War II until sometime in the early 2000s in which an array of big societal forces had aligned to drive consistent, rapid increases in vehicle travel. That historical period, we argued, was over. What was going to come next was uncertain.
But we suspected that, whatever came next, vehicle travel over the long-term was unlikely, under then-foreseeable conditions, to exceed the level of per-capita vehicle-miles traveled (VMT) that prevailed in the peak year of 2004.
Fast forward to 2016, and we now find ourselves at the end of a second year of blistering growth in VMT, even by the standards of the “Driving Boom” era. (A good summary of recent trends is available from Doug Short here.)
It now seems like a good time to revisit the peak car debate – what it means and where (if anywhere) it is going. In this discussion, I can only present our own views on the subject – if you’d like to read other thoughtful perspectives on long-term trends, I would highly recommend David Levinson and Kevin Krizek’s The End of Traffic and the Future of Transport; the International Transport Forum’s summary of its 2012 roundtable on Long-Run Trends in Car Use; Todd Litman’s The Future Isn’t What it Used to Be, and David Metz’s Peak Car blog. Not all of these documents argue that a peak in motorization has occurred or is imminent, and I don’t endorse everything each of these authors have to say, but they all include useful data and perspectives to bring to the discussion.
Why Do We Care Anyway?
The peak car discussion got rolling in the late 2000s (though its antecedents stretch back to nearly the beginning of modern transportation studies, see the ITF report cited above) in part because people working with VMT data began noticing that growth in vehicle travel was not keeping up with earlier projections.
We were among those people (though I would give Robert Puentes, now of the Eno Center, and Adie Tomer of Brookings credit for bringing the VMT growth slowdown to our attention with their important 2008 paper). My colleague Elizabeth Ridlington and I spent much of the 2000s working with organizations to develop state policy action plans to address global warming, and one of the key inputs for those analyses was VMT projections. Around about 2009, Elizabeth began to notice that those state projections were consistently overshooting reality. We first began writing publicly about shifting trends, and incorporating them into our analysis, around 2010 – at least six years after the bending of the VMT curve flagged by Puentes and Tomer had begun.
As analysts, we cared about this because we didn’t want to put our names behind dubious forecasts, even if they could be sourced back to a state DOT. As people concerned about global warming and transportation, we cared because the assumption that steady, rapid growth in VMT would continue indefinitely was shot through both the transportation profession at large, and people on both sides of the global warming debate (including many advocates for climate action). If consistent, rapid growth in VMT was not inevitable – as the data suggested – that was big news, with implications both for how we spent transportation money and for the policy strategies that might prove to be effective in the fight against global warming.
This, on some level, is what the debate about “peak car” has always been about: Whether the development and motorization trends that exhibited themselves in the postwar period in the United States will inevitably continue - either for political or cultural reasons, because of path dependencies, or because the desire for a suburban house with a two-car garage is hardwired into humans’ lizard brains and only requires the right sprinkling of household income to activate – or whether those trends were the result of temporary economic and demographic forces, conscious choices, ephemeral cultural preferences, and/or particular policy regimes, any or all of which are subject to change.
“Peak Car” Is Not (Entirely) About the Millennials
In the early 2010s, a parallel narrative began to emerge about travel trends, one that we also wrote about extensively: the idea that the Millennial generation exhibited transportation attitudes and behaviors that differed from the generations immediately preceding them.
The two narratives are obviously related – if Millennials are driving less, that will reduce overall vehicle travel – but the strongest arguments for a long-term slowing in vehicle travel growth have less to do with the Millennials than they do with a set of observations about what happened during the postwar Driving Boom era.
The postwar years in the United States saw the construction of the Interstate Highway System (and corresponding increase in average travel speeds), mass motorization, mass suburbanization and the simultaneous decline of cities, the entry of large numbers of women into the workforce, and – toward the end of that period – a surge in the share of the population of prime working age.
None of those factors are repeatable in the same way in the 21st century. The big travel speed gains of the Interstate era have been realized. Cities are, if not universally resurgent, at least holding their own against suburbanization in a way they did not in the late 20th century. The employment-to-population ratio, while up a bit lately (and with some modest room for future growth), will never, ever surpass the peak year of 2000, or even the pre-recession peak of 2006, due to the aging of the Baby Boomers. The vast majority of those who want and can reasonably afford a vehicle already have one (or more).
When it comes to automobility, our read of the long-term trends was that there was just not that much more room to run in the growth drivers that were important in the 20th century. And given that the United States is already a freakish outlier in per-capita VMT by international standards (to a degree that I don’t think even most transportation nerds fully comprehend, see chart) and that many other western countries were experiencing similar slowdowns in vehicle travel growth, the idea that the U.S. would see a return of the Driving Boom in the face of stiff demographic headwinds, among other things, seemed unlikely.
The United States Is a Freakish Outlier with Regard to Per-Capita VMT (Source: Australian Department of Infrastructure and Transport)
Now, it was certainly possible, as we acknowledged at the time, that other factors would emerge to change the situation. One of the factors we discussed – gas prices – already has shifted dramatically since 2013 (though whether that change is permanent is uncertain), and in a footnote to the text of A New Direction (published in the spring of 2013), we suggested that driverless cars might be one innovation that could conceivably increase the speed of vehicle travel, which had stagnated in the late 20th century.
We believe that the evidence is strong that Millennials are in fact different – they are more multimodal, more interested in urban living and walkable communities, more open to new transport technologies and modes, and have different opinions about transportation policy – than their immediate forebears. But even if, as we assumed in the “Back to the Future” case in A New Direction, Millennials were to return to the transportation habits of their predecessors, there would still be reason to believe that long-term growth in vehicle travel might be slower than in the past.
“Peak Car” Is Not About the Short Run
So, what about the rapid surge in driving over the last two years?
First of all, we need to be clear about what has happened. Per-capita VMT did not begin its resurgence with the economic recovery. In fact, from the end of the recession in 2009 through early 2014, per-capita VMT continued to fall. Only in mid-2014 did per-capita VMT begin to increase modestly, and only in late 2014 – timed nearly exactly with the collapse of gas prices – did the current period of torrid growth in VMT begin.
The last two years have seen a confluence of three factors: steady economic and employment growth, low gas prices, and easy access to automobile credit that has resaturated the market for personal cars. It has never been cheaper to buy an expensive car, it hasn’t been cheaper to fuel it (in real terms) in more than a decade, and it has been many years since as many people have been in as good an economic position to afford the high costs of automobility as they are today.
Is it possible that this set of circumstances represents a durable “new normal”? Perhaps. But it is important to note that both the growth of fracking (which was the precipitating factor, along with slowing growth in global demand, that sparked the ongoing global oil price war) and the growth of car sales have been underwritten by the incursion of huge amounts of debt.
The oil and gas industry faces a massive and growing debt overhang (following years of record capital inflows – PDF) that is causing firms to go belly-up throughout the oil patch and will likely have long-term repercussions for the oil industry’s capital investment strategy going forward. (Look at the stats on oil industry capital expenditures and think about what they might mean five, 10 or 15 years down the road.) The credit quality of car loans is beginning to deteriorate as the amount of outstanding auto debt hits all-time highs and the market for vehicles hits a saturation point, with no lesser lights than Ford Motor Company now believing that the recent run of record new car sales is coming to an end and used car prices beginning to show signs of weakness for the first time since the recession.
At some point – whether it’s in two months, in 2017, in 2018 or whenever – it is quite possible that the music stops. Someone on Wall Street stops securitizing car loans that will never be repaid. The global oil market “rebalances” and the excesses of today’s auto lending market are curbed. Gas prices stop falling (gas prices have now been lower than the same point in the previous year for 110 consecutive weeks, or since July 2014.) Where we will be at the end of that process is anyone’s guess. But it is worth asking whether the summer of 2016 – a time during which the nation and the world faced a historically unusual gasoline glut and auto lenders continued to fire low-interest (if you’re a prime borrower) or no-look (if you’re subprime) loans, along with generous purchase incentives, into crowds of would-be car buyers like costumed mascots with T-shirt guns at a minor league baseball game – is really the time we want to be declaring a return to normalcy.
There are some legitimate research questions that the experience of the last couple of years raises to me, and places where we need to question whether our understanding of the world from 2013 needs to be revisited. Perhaps the most salient one is the relationship between driving patterns and fuel prices.
A future of consistently high (by historical standards) gas prices was generally assumed back in 2013, but circumstances have changed – dramatically. Real gasoline prices are projected to have fallen by 46 percent between 2012 and 2016, according to the Energy Information Administration. Estimates of elasticity of travel demand with respect to gas prices are all over the map, but estimates from the Driving Boom-era suggest that vehicle travel would have increased by somewhere in the ballpark of 3 to 11 percent in response to the near-halving of gasoline prices, with more substantial impacts over the long run. Recent increases in VMT are within that very wide envelope.
Recent studies, including this National Bureau of Economic Research working paper and this 2012 white paper by the esteemed Todd Litman of the Victoria Transport Policy Institute, suggest that the response of vehicle travel or gasoline consumption to price changes is either greater than previously estimated (NBER) or has increased in recent years (Litman). Intuitively, this makes some amount of sense: with a smaller share of the population in the workforce and commute travel continuing to decline as a share of total travel, a greater share of automobile travel may be falling into categories that we might call “discretionary” and that may be more price-sensitive than commute travel. There are, to take one example, quite a few Baby Boomers out there looking to take cross-country RV trips to visit their grandkids who pay very close attention to the price at the pump. Understanding how travel patterns are continuing to evolve – and how they might respond to prices or policy initiatives – is important and worth attention.
Last, it is worth asking whether some of the X factors that might alter the limiting conditions we identified in A New Direction might be arriving sooner rather than later. Autonomous cars and electric vehicles have both made leaps forward in the last few years and both – autonomous cars through their ability to speed travel or encourage people to extend their travel time budgets, and electric vehicles because of their lower running costs – could alter the economics of driving in ways that increase VMT … though if those new technologies are successful in cutting congestion and emissions, we might not care quite so much about aggregate VMT statistics anyway.
The Bottom Line
So, what can we take out of all this? Here are a few thoughts:
1) If “peak car” is dead, we wouldn’t know it yet.
If the case against peak car is that the decline in per-capita driving from 2004 to 2014 was the result of the unique economic dislocation caused by the Great Recession (even though the decline started three years earlier and continued nearly five years into the recovery), then it also becomes important that the unusual characteristics of the last two years of the recovery – namely, historically low borrowing costs, loose credit markets in the automotive and energy sectors, and the ongoing global oil price war – be considered as well. Is it possible that the events of the last two years herald a lasting return to go-go growth in vehicle travel? Maybe. Is it certain? It’s way too early to tell.
Nor is it adequate to arbitrarily lump the last 10 or 15 years together, call it a “cycle,” draw a straight line through it and assume that that line represents a long-term trend with predictive value. History moves in lurches and fits, and its underlying direction is unknown to the people who happen to be living it. As analysts, we need to have the courage to realize that the knowledge we gain from “rear-view mirror” analysis – while “safe” – is unlikely to be of much use in a world that is shifting radically in all sorts of ways: technologically, economically and culturally. Almost any American could sit down and, in five minutes, rattle off a dozen ways that life in 2016 is very different from how they imagined it might be a decade ago. To imagine that transportation, of all things, is the one area of human endeavor in which the trends of the post-war era will carry on ad infinitum seems off-base. The 21st century is going to be different. How it will be different remains to be seen and is, in many ways, for us to decide.
2) “Predict and provide” is a bad way to make transportation decisions.
One of the key conclusions of peak car-style analysis is that our traditional transportation models are less effective at predicting the future than we’d assumed they were. If anything, the recent surge in VMT – which was just as surprising as the preceding fall, and came more suddenly – validates, rather than undercuts, that conclusion. If, every five years or so, forecasters find themselves saying “boy, we didn’t see that coming,” you really have to wonder whether forecasting trends 10, 20 or 30 years into the future has much utility at all in policy setting (even though scenario analysis still has use as a decision tool).
Historically, those forecasts have been used to justify or assess the benefits of highway capacity expansions under the “predict and provide” model. The events of the last two years have reinforced the need to get away from that model, and to embrace other notions of planning, such as the radical concept of envisioning and investing to build the transportation system we want to see – one that achieves our common goals of access, equity, efficiency, and environmental and fiscal sustainability. Which brings me to the final point …
3) Policy matters.
Somehow, it became an established idea that those of us questioning the inevitable continued rise of VMT were arguing that everything was going to be OK, the Millennials have it all in hand, and policy doesn’t really matter.
In fact, our point was and always has been the precise opposite: changing conditions in society and shifting attitudes demand a new approach to transportation policy that reacts to emerging 21st century imperatives and respects the desires of many Americans – some of them Millennials and some not – for transportation options and development patterns that enable them to live without dependence on a car.
Here is what we wrote at the conclusion of A New Direction:
We may not know the exact shape of the future, but it is increasingly likely that it will look very different from the past. By retiring Driving Boom-era assumptions and policies that no longer serve the nation’s needs, we can build a transportation system that is more affordable, more efficient and more sustainable for the long haul.
The need to revisit old assumptions in transportation planning, and to bring a clear-eyed sense of priorities and fresh perspective to transportation policy is as urgent today as it has ever been – regardless of month-to-month or year-to-year fluctuations of VMT.
If those of us who want a better, more sustainable transportation system want to succeed, we need both to articulate – and to build support for – a compelling 21st century vision for our communities and transportation system and, at the same time, to stick a fork in 20th century modes of thought and policies that have outlived their usefulness but continue to shape transportation policy today. The peak car debate has, if nothing else, exposed those modes of thought to long-overdue scrutiny and critique. It is important to keep that scrutiny going.
 Note: “attitudes” are shaped by both an individual’s personal tastes and preferences and their perception of available alternatives. It is not – or not just – that Millennials came out of the womb wanting to live in cities. It is, in part at least, that the conditions of city living are better now (and, in many cases, the benefits of moving to the suburbs less compelling) than they were for someone coming of age in, say, the late 1960s.
 “Millennials” is an oversimplification. The inflection point in most surveys of transportation attitudes and studies of trends tends to occur in the middle of Generation X.
 Assuming low-end short term elasticity of demand of -0.055 and high-end estimate of -0.25 from sources cited in Todd Litman, Victoria Transport Policy Institute, Changing Vehicle Price Sensitivities, November 2012.