Stealing From The Future

Early in 2009 the Initiative On Global Markets website published updates of then current research projects including the following description by two economists from the University of Chicago Booth School of Business.

Fiscal Stimulus and Macroeconomic Fluctuations: Evidence from Cash for Clunkers

Atif Mian, Associate Professor of Finance
Amir Sufi, Associate Professor of Finance
Summary: We seek to estimate the economic consequences of short‐term boost in consumer spending on cars driven by the Obama Administration’s “Cash For Clunkers” program. We exploit cross-sectional variation (across counties) in the fraction of automobiles that qualify as “clunkers” to isolate the impact of boost in auto-sales on outcomes such as unemployment, housing market, consumer defaults, and household expenditure. Our research hopes to shed light on one of the most important questions facing policy makers at a time of recession: Do Keynesian policies of government stimulus work?

On September 1, the results of their innovative study were released.  The study sought to test the Keynesian model inherent to the program by examining short and longer term impacts on consumer behavior.  In short, they wanted to understand if a consumer stimulus of this kind would actually stimulate the economy, or if it would merely shift buying patterns.  The Cash for Clunkers program provided the rare opportunity to test out a yet long-beloved, economic model.  By comparing consumer behavior in large states with little response to the program, against consumer behavior in all states where they followed every sale made under the program, Sufi and Mian were able to run the study in real-time, as well as examine the resultant behavior in the months after Cash for Clunkers ended.

A look at the history behind the proposal provides an interesting perspective on Mian and Sufi's final results.  During the first Clinton term, Alan Blinder, an economics professor at Princeton University, proposed that the government could boost its efforts to control carbon emissions by influencing car purchases through a trade-in stimulus.  His proposal at the time died a quiet death.  Flash-forward to 2008, when the economy was starting its long dive, and Prof. Blinder saw his chance to ressurect his 'modest proposal,' this time emphasizing the potential economic benefits to unemployed auto workers and sales reps.  In a Talk of the Nation interview one year ago, Blinder explained why he thought the program as enacted was failing:

'Well, certainly not. Things that come through Congress never come out the way you expect them to. I mean, my proposal focused much more on, for example, emissions than on miles per gallon, although they're related, but they're not the same. And also, importantly to me, one of its objectives was to assist poor people who tend to own the clunkers. The way the law was written, you could only get the cash for the clunker if you bought a new car. And a lot of poor people can't afford to buy a new car.'

Apparently Blinder thought (and presumably still does) that the car stimulus was not stimulus enough.  Indeed, this is evidenced by his statement a moment later in the same interview.

'Well, I think the stimulus part has worked swimmingly. And I should have mentioned, when I wrote that piece in July '08, I mentioned four objectives. We already talked about three: stimulus, anti-poverty and pro-environment. Right at the end, I said, And by the way, it would be good for the automobile industry. Well, that's been the objective, of course, that's been the guiding light of this legislation, this national legislation. I think it's been quite effective. I mean it's been a big stimulus to sales.'

Blinder has been arguing for some time that all of the stimulus and bailout programs have been outright successes, albeit with very little evidence to back up his claims.  In July of this year, he and Mark Zandi released another modest dud, "How The Great Recession Was Brought To An End."  In it, the authors assert the now familiar mantra that the government's stimulus programs averted a far worse crisis, based on (you guessed it) economic modeling, and cite as evidence the rise in GDP of 3.4%.  The paper was published one month before the government drastically revised its estimated growth down to 1.6%, and released the August unemployment rate, which ticked up to 9.6%.  Arguably, Blinder is falling into the classic trap of failing to 'test' a model under actual conditions in order to prove its robustness and usefulness as a predictive tool.

Unlike Blinder, Mian and Sufi tested the proposal against actual sales and found that the program merely 'stole' future sales by simply advancing sales by a couple of months.  Within less than a year, as the Planet Money interview with Sufi so eloquently put it, the program was a total wash.  The abstract neatly sums up the paper's conclusions: 'We also find no evidence of an effect on employment, house prices, or household default rates in cities with higher exposure to the program.'  In the interview, Sufi summarized the problem with assuming that government can induce behavior:

'We don't take a stand on whether that's a useful thing. It's hard to answer that question. You could argue that if you're in the depths of a really bad recession, maybe shifting purchases of automobiles by even a couple months, maybe that's a useful thing to do. But I think it's incredibly important for policymakers to understand whenever you induce the purchase of any good, whether it be homes, whether it be a washing machine, you are in some sense stealing those purchases from the very near future. And I think that's the key result of our paper that ends up mattering.'

All of which is to say that while government may have a role in both boosting the economy, and in encouraging behavior, it can not do either by trying to force consumer behavior down a particular path.  Doing so merely steals from the future.