“Cash for Clunkers” failed

Here is a new paper by Atif Mian and Amir Sufi:

A key rationale for fiscal stimulus is to boost consumption when aggregate demand is perceived to be inefficiently low. We examine the ability of the government to increase consumption by evaluating the impact of the 2009 "Cash for Clunkers" program on short and medium run auto purchases. Our empirical strategy exploits variation across U.S. cities in ex-ante exposure to the program as measured by the number of "clunkers" in the city as of the summer of 2008. We find that the program induced the purchase of an additional 360,000 cars in July and August of 2009. However, almost all of the additional purchases under the program were pulled forward from the very near future; the effect of the program on auto purchases is almost completely reversed by as early as March 2010 – only seven months after the program ended. The effect of the program on auto purchases was significantly more short-lived than previously suggested. We also find no evidence of an effect on employment, house prices, or household default rates in cities with higher exposure to the program.

Ungated versions of the paper are here.


Politicians are one thing, they answer to the ignorant, but I'd like to see a list of real economists who were in favor of this policy. But, I doubt such a list could exist, if you get my meaning.

Not a fan of the policy, but it's a stretch to say that this translates directly into "it failed". I'm not sure why cities "with higher exposure to the program" would be expected to benefit more than cities where the products were actually manufactured (though admittedly I might bother to read the paper and find out if there's an answer), but I think its supporters would argue that shifting demand forward helped see the auto industry through an extremely difficult time, and may have helped nurture the manufacturing recovery that began around that precise time. After all, the point of demand shifting policies is... to shift demand, to a time when there's a greater need. Not to increase (at least, directly) demand through all time.

The paper seems to have made a key error. It asserts the purpose of fiscal stimulus is to increase aggregate demand but then it proceeds to judge C4C only on auto demand. Let's say it's a given that every C4C purchase was going to happen anyway and all C4C did was move it up a month or two and increased the mpg of the purchased cars. In that case each buyer received $4500. Aggregate demand either increase in the form of people buying a more expensive car than they otherwise would have OR it increased elsewhere (say redoing their decks since they spent $4500 less on their planned auto purchase).

Pulling demand forward was primarily the point was it not? Doesn't the author make an unintentional case for the program having lasted longer?


Do you have any problem with specifically it for the reason that they spent money to destroy working engines?

They could have moth-balled the vehicles.

Timing of cash flows matters. When trying to restart the economy, it, in some ways, functions like a company. Companies don't just care about their total profits over a few years, they care about the timing of their cash flows. I don't think that anyone thought that cash-for-clunkers would significantly increase the long-run level of auto consumption.

C4C, by the merits of this article, was NOT a failure; it did exactly what it was supposed to do, and created a burst of demand in late summer 09. Yes, it captured futures sales, but as the article says that was over by March. It was a temporary infusion.

When you look at aggregate gasoline demand, it is still way down. How much from removing late model SUVs from the road -- and under the rules that all about that qualified. Moving someone from a F150 to a ranger is a tremendous savings is gas.

The one thing I wish C4C did was target heavy users; i.e. a lightly used F150 is no problem. Given the incentives, however, I'd argue that is exactly what it did.

A friend bought a old van in preparation for C4C, did not qualify, and it still sitting in his garage.

God in in the details.

Did the use of the phrase "empirical strategy" sound a bit weird to anyone else? Coming from the natural sciences and engineering, I am used to "empirical" relating to targeted experiments or measurements, not post-facto statistical modeling.

@TED; facts, my friend, are your friend.

Cars destroyed? How many cars qualified - maybe 10 models? Almost all the vehicles destroyed were trucks.

And how is destroying a truck "Weakening" the economy?

Bigger issue: those trucks would have been sold to shipped to South America. Is removing them - as opposed to removing them from the US - helpful at all in reducing gasoline demand there?

Boonton ... please, dude, just give up. Just give up. You're in a position where you honestly believed it would be a good idea to use tax money to pay well-above market prices for crappy cars and destroy them, on grounds that it would appease the aggregate demand god.

It's time to just admit your error, stop trying to hide it, and re-arrange your life around a more rewarding career in digging ditches.

(And losing half of that pitiful income to the taxes you love.)

I already said this, last November: http://zatavu.blogspot.com/2009/11/cash-for-clunkers-contributes-to-crash.html

"Most of the cars purchased would have been purchased anyway, either during the time of the program, or shortly thereafter, so no real effect was made on the economy. In fact, to the extent that it just moved up purchases a few months, it took away from future sales." I say a whole lot more than this, though.


1. What about the program's destruction of usable used cars? ...

That's a valid concern but one I think was overstated. The credit was for $4500. Clearly if your used car was worth $4400 the credit wasn't worth much effort to you but if you had a used car worth $500 the credit would net you a $4K profit. As a result, it seems clear that most of the destroyed cars were at the low end of the value scale.

Second, the mpg increases were on the order of 58%. I estimate savings of 300 gallons a year or more based on 13,000 miles driven per year. If you price gas at $2.50 plus $0.50 for the 'hidden' costs that's nearly a grand per year. In most cases we've probably already made back the cost of the junked capital.

Third, gas is NOT priced to incorporate its hidden costs. As a result at the bottom of the used market there are no doubt some low mpg cars on the road today that wouldn't be if gas was 'correctly' priced. This was, IMO, a case where a gov't distortion (destroying useful capital) actually offset another gov't distortion (keeping gas prices artifically low thereby making junky low mpg cars seem economical to keep out of the junk yard for a few more years). A bit like a small tax on high frutose corn syrup needs to be examined in comparision to the massive subsidies for corn.

3. No one ever talks about the insurance-industry subsidy this was. The insurance industry profit is nearly proportional to the aggregate value of nation's total car inventory, but they get more "comprehensive" coverage gravy from newer cars (less affluent people, and the owners of older cars, usually go for the legal minimum liability coverage).

I'm not sure its a subsidy. Insurance is selling more product (comp. coverage for a $23K car versus basic coverage for a $500 used car). Of course they are going to make more profits but they are making more profits because consumers are opting to buy more insurance to protect their cars (or their banks are insisting they do to protect their investments).


When you look at aggregate gasoline demand, it is still way down. How much from removing late model SUVs from the road -- and under the rules that all about that qualified. Moving someone from a F150 to a ranger is a tremendous savings is gas.

Let's keep things in perspective here. The average C4C car was much more efficient than the one it replaced BUT the program only covered 700,000 vehicles. The US fleet consists of something like 60M registered and 6.5 useable but unregistered vehicles. Even if these cars were the gasolineless 'Mr. Fusion' car from Back to the Future, you won't see much impact in aggregate gas demand....esp. compared to the recession.

Boonton ... please, dude, just give up. Just give up. You're in a position where you honestly believed it would be a good idea to use tax money to pay well-above market prices for crappy cars and destroy them, on grounds that it would appease the aggregate demand god.

Look if you want to debate stimulus go ahead. I'm putting the facts out there. If you're going to claim there was no increase to aggregate demand then where did the money go? Did people take the $4500 and burn it? If they spent it elsewhere then that's increase aggregate demand. If they saved it then that doesn't, but if you're going to argue a classical model then saving the $4500 increases investment.

In fact C4C could have caused harm if it made producers expect demand to rise higher or for a longer term and then left them with excessive inventory.

The estimate I've seen is of 700,000 cars about 125,000 were new purchases that wouldn't have 'happened anyway'. It seems kind of silly to say no new purchases happened. A $4500 price reduction isn't going to cause some people to buy a new car who otherwise wouldn't? That's a pretty inelastic demand curve your're working with there. What econ textbook told you that is a great fit for the auto market?

I don't believe there's been any 'inventory crash' post C4C. C4C cleared out a lot of inventory and production went up but there's been no notable piling up of unsold new cars post C4C. Either the auto industry is doing a great job forecasting short term demand or the post C4C sales slump has been offset by general improvement in the market.


First of all I've used "if" for a reason. My guess that C4C didn't have any real economic effect. Now IF it was something more meaningful and then fell short of expected goal it could have negative consequences.

"A $4500 price reduction isn't going to cause some people to buy a new car who otherwise wouldn't?"
In first part of your comment you've already stated that actual price reduction was lower than $4500.As I understand to buy a car you'd have to dig into your savings anyway, and the state of economy at the moment makes that kind of decision less clear-cut than usual.

Andrey, Do you agree or not with the point that the study should have used a but for analysis rather than comparing two periods within the program?

C4C was never sold as a program to move auto sales from March 2010 to Fall 2009. It was sold as a way to "stimulate" the auto industry. This paper shows that it did not do the latter, and as such was a transfer of taxpayer money from people who had recently purchased vehicles(like I did in 2007), to people with the good fortune of waiting until 2009.

"I am not against engine destruction unless it is my own and I am not participating in the program."


I think you should. Reason #1 is that you even if you think that gas mileage is the rationale, that it's real, and that it is worth the total net short-term cost they could have achieved C4C without destroying working engines and drivetrains. The government is on this destruction of useful things kick as an expedient to regulating a secondary market. Seems pretty lazy and dumb to me, even if stopping a secondary market is a good thing, which it's not of course because the secondary market is the poor. Also, if the "ostensible goal" (read PR) was pollution reduction, is it less polluting on net to have to replace and destroy a working engine and car than to let it die of natural causes? I doubt it. The gas mileage improvement was calculated to be 0.7 mpg. However, I wonder if this factored in that many of these clunkers were probably parked. So, if I'm driving a Civic, but turn in my old Explorer for a Forerunner, yes, mileage is better between the forerunner than the Explorer, but now I'm parking the civic to drive the Forerunner. I might enjoy a simple tax break so that I could have a parked pickup to use occasionally. I wonder how many people drive low-mileage vehicles because it is cheaper to have a single car rather than one commuter and one utility car. From the looks of my neighborhood it's a lot. Government could work to eliminate their contribution (including insurance laws) that force people to buy one-size-fits all vehicles, which of course results in them buying the biggest to match their biggest chores.

So, if you agree with the claimed goals of the program you should not agree with the program. I of course don't even agree with the claimed goals of the program. Here's the fundamental problem. We have a debt-driven balance sheet recession. Keynesians don't seem to believe in this, but it's true. This is why Keynesian policy hasn't worked as expected. So, in response to a lack of assets, they destroyed assets in order to indebt consumers to the auto industry. Recoveries are often lead by consumer durables financed by credit. This increase in aggregate debt in order to make a cash transfer to the auto industry is going to slow the recovery. And the proponents say that it's a success because the auto industry is viable. But the same success could have been had and probably less expensively by just handing them money. C4C is simply political shenanigans.

Keynesians don't seem to understand the double-edged sword of debt. That's what gets them into trouble, but that's not the biggest problem in getting us out of trouble. To Keynesians, labor isn't the main thing, it's the only thing. This is wrong. It is making them wrong at every turn. That Keynesians are wrong and embarrassing themselves is not the problem (and nothing new anyway). It is extending the recession and hurting everyone.


Seems you can't have it both ways. If the most C4C people were going to buy a new car anyway then they were going to go into debt anyway (keep in mind, though, there was no requirement to go into debt....I'm sure some paid cash for their cars). In terms of increasing consumer debt then C4C was actually decreasing debt as car buyers reduced their debt load by the net value of the voucher.

I think most Keynesians do see this as a balance sheet driven recession as such mosts policies are related to increasing demand or stabalizing the balance sheets of the banks which control a portion of demand by the credit markets. "going into debt" to the auto industry isn't a major problem IMO. Those that brought new cars judged their finances to be healthy enough to support a car payment (and since they felt it was in the middle of a deep recession they probably were in reasonably good shape). I'd say a valid criticism of the program as stimulus is that other people are much worse off. I'd rather pay unemployed home builders to put solar panels on homes, for example.

Gas mileage is not in itself stimulus. Stimulus is simply increasing aggregate demand. YOu can do this by 'dropping money from a helicopter'. But it's nice to buy something with stimulus that provides benefit even after the recession is gone. For example, the Hoover Dam still stands today producing power and increasing our production function over a half century after the Great Depression ended. The lower gas mileage two years from now (hopefully we will be closer to full employment) will mean our economy can get more from less.

Finally, a balance sheet depression doesn't mean we have a shortage of assets. Even if we did, though, the clunkers destroyed aren't very good assets to begin with. I think the analogy requiring new home buyers to destroy their old homes illustrates this difference well. Unless you have a very bad location or major problems with the home, 'used homes' are valuable assets that have a very long life. Destroying them today will mean ten years from now our economy will have fewer assets to provide us with goods and services


If you had a well maintained used car that was worth more than $4500 you wouldn't have used the C4C program. If you wanted a new car you would have sold it yourself for more than $4500 and used that money to help pay for it.

The idea that used car prices have shot up because of C4C is really off base I believe for a number of reasons:

1. Few cars were actually destroyed by the C4C program, maybe 100K. Most of the vehicles destroyed were trucks. Why would you see increasing prices in the used car market then rather than the used pickup market.

2. Every driver of a used car in the C4C program got a new car to drive making it a swap. We've seen hurricanes easily destroy the same number of cars that C4C did without a spike in used car prices, those car owners weren't swapped into a new car but had to go into the market to buy replacement vehicles.

3. According to http://www.worldometers.info/cars/ the US has something to the tune of 136 million passenger cars and has annual production of around 6.8 million (and because of the recession and implosion of GM and other car companies there's plenty of spare capacity to make more if the market wants it). Losing 700,000 low end used cars from this giant market does not seem like enough to generate any type of relevant shortage.

I think more likely explanations for higher used car prices are

* Weak economy is getting people who would have brought new cars to buy less expensive used ones. For example the Boston Globe piece cited the used Cadellic Escalade rising in price 30% to $35,000. Certainly C4C didn't motivate people with $30K used Escalades to have them junked resulting in a shortage. These are probably people who would have brought a more expensive new luxury car if the economy was better so they opt to 'scale down' to premium used cars.

* The decline of the pickup....weak economy and housing bubble means fewer people need them as work vehicles and people still remember $5 gallon gas. They want lower mpg cars. For new cars this is easy, auto companies simply put more lines into producing cars and few lines to trucks. But for the used market you're stuck with whatever the production mix was years ago. In the used market people are demanding the relatively few good mpg cars were made in the past and bidding up prices.

Comments for this post are closed