Garett Jones  

Degree Pollution: Why subsidize it?

PRINT
How Does Belief in the Signali... Sumner Channels Yglesias Again...

People who drive heavier cars are more likely to survive accidents.  Since life is the most valuable commodity of all, shouldn't governments give out grants or loans to help the poor buy bigger cars? 

Even leaving the issue of pollution aside, the typical wonk knows why a big-car subsidy is questionable policy: Part of the reason big cars are safer is because they're bigger than the average car.  You'd be fueling an arms race, where the soccer mom in the heaviest SUV wins. 

According to the signaling theory of education, this is what we're doing with college education: We know that kids with above-average education earn more, so we're trying to make sure that more kids are above average.  Lake Wobegon as our national education policy. 

According to the purest version of signaling theory, college is "smart for one, dumb for all." Every college degree pollutes the value of every other college degree.

Signaling isn't an all-or-nothing proposition, of course: College might be part signaling, part skill acquisition.  But if even 10 or 20 percent of the value of college comes from signaling, the argument for government subsidies weakens dramatically (one relevant citation; see the conclusion).  Why fuel an arms race?

Every good policy analyst knows what to do with polluting industries...and the answer isn't "subsidize it." But for some reason, analysts rarely apply that lesson to education. 

The latest entry: Felix Salmon's critique of Megan McArdle's Newsweek piece (where Bryan is quoted).  Salmon correctly notes that college helps particular people earn more; but since heavier cars help particular people stay safe, that doesn't give us a sound argument for college subsidies.

He then makes this more interesting claim:

[W]e're part of a global economy, where many employers will hire only college grads -- and if they can't find those college grads in the US, they'll find them in Ireland or Israel or India instead.

So even if the U.S. cut back subsidies for degree pollution, other countries might keep up the loans and grants and free educations, making our workers look worse by comparison.  But we know the solution to polluting industries that can slide across national borders, at least in principle: A multilateral agreement.

So perhaps governments in the rich countries should think about negotiating an education arms control treaty: And then perhaps someday, the bachelor's will be the new master's.  


Comments and Sharing





COMMENTS (23 to date)
Tracy W writes:
People who drive heavier cars are more likely to survive accidents

That's the wrong measure of car safety. There's two ways to reduce your risk of dying in a road accident, holding driving habits constant, one is to drive a car that protects its inhabitants better in an accident, the other is to drive a car that's better at keeping you out of accidents in the first place. Which might well be a lighter, more maneuverable car.

Otherwise, I agree with your post.

Ritwik writes:

But this is true of all capital/wealth acquisition, not just human capital acquisition. Replace signaling with insurance. As Interfluidity points out, acquisition of wealth after a point is about insurance, not consumption or production. Capital owners are in an insurance arms race. While the argument is easiest to see for personal wealth, it arguably also holds for business capital.

The theoretical case for removing or cutting back on college subsidies is exactly the theoretical case for higher taxation of capital acquisition, by households or businesses.

Steve writes:

Living in the inner city, it seems to me the poor do drive bigger cars than the wealthy. This is because their cars tend to be older. They spend a fortune (to them) on repairs and gasoline, but choose those cars because the purchase price is low. Also, many of them have larger, and extended, families in the same home.

The wealthy, on the other hand, prefer to spend less on repairs and fuel, and have money to spend, so choose smaller cars no matter the cost. The do not have as many children, they do not have as many extended family members living with them.

I also observe more poor people riding on the bus, or walking, which are safer than any form of automobile travel. (I think.)

Prakash writes:

Like Money illusion, the refusal to accept a smaller amount of money even if it may mean more goods, might not there be a degree illusion with every generation thinking it has to do better than their parents did? If that is so, then the only way may be upward.

Has this phenomenon been seen in a country where there is no legal ramifications to using a straight IQ test for hiring?

Jody writes:

"Part of the reason"

But this is key. If the value is virtually all positional, then it's obvious that subsidizing is a bad idea economically. But if the value is more absolute (eg capital formation) than positional (signaling), then there are subsidy levels that can be cost efficient with the general rule of thumb that the cost of the subsidy should never exceed the increase in absolute value.

Bostonian writes:

I agree with Jones that as a higher fraction of people in the U.S. obtains a bachelor's degree, the value of the credential may be reduced. But does he really think more Germans getting degrees has a substantial effect on the American labor market? I'd like to see evidence for that assertion.

Libertarians understand that we should practice "unilateral disarmament" in trade, eliminating tariffs even if other countries keep theirs. We can do the same with higher education, not worrying about whether other countries are sending more people to college.

RPLong writes:
So perhaps governments in the rich countries should think about negotiating an education arms control treaty: And then perhaps someday, the bachelor's will be the new master's.
Interesting concept, and an excellent post.
Sieben writes:

You had me at "soccer mom" :)

Garett Jones writes:

@Tracy W:

The modal view of the weight arms race is noted here, in a contrarian paper:

http://energy.lbl.gov/ea/teepa/pdf/aps-ppt-wenzel.pdf

The modal view is that weight is the biggest factor (Slide 3); the authors of the PDF are trying to take a "Yes, but" approach, and since they're speaking at a sustainable energy conference, we know what the "Yes, but" is all about....

It's a worthy attempt at debunking the mainstream view, but at best they just show that other stuff matters too. When they control for that other stuff, weight matters more not less (Slides 20 and 21).

They have a nice slide on the arms race aspect (slide 8).

The more conventional "weight matters a lot" view, from the Insurance Institute:

http://www.iihs.org/news/rss/pr041409.html

"The death rate in 1-3-year-old minicars in multiple-vehicle crashes during 2007 was almost twice as high as the rate in very large cars."

NHTSA here:

http://www-nrd.nhtsa.dot.gov/Pubs/808569.pdf

"Large vehicles have historically been more stable and provided more protection for their own
occupants than small ones, but they presented a greater hazard to other road users."

Arms race again.

Correlation isn't causation, but physics is.

Tom West writes:

While Garett makes an interesting point, it elides one aspect of this arms race. Without subsidies, the only people who would likely be able to afford this credential would be the upper middle class on up.

If the credential is fairly close to mandatory in order to obtain a middle-class lifestyle, I would prefer that we make some effort to ensure that the possession of that credential doesn't depend entirely upon one's parent's wealth.

I have to say, given the nature of the educational arms race, I have a pretty hard time seeing a justification for a free public school system that doesn't also apply to at least a first degree/certificate in post-secondary education.

Garett Jones writes:

@Tom West:

The subsidy for college increases the need for college. Reduce the subsidy, you reduce the need.

Imagine if we had massive subsidies for big cars now. If we cut back on the subsidies, the rich would keep driving bigger cars---but for the average person, the road wouldn't become much more dangerous (and might even safer).

Reducing the big car subsidy for everyone reduces my personal need for a big car.

Garett Jones writes:

@Jody:

The optimal subsidy for college is negative whenever the marginal social benefit of one more diploma is greater than the marginal social cost. So on price theory grounds, we should considering taxing college instead of subsidizing it.

This is the conventional story of a polluting industry: The question is never whether the externality is greater than the product's private value, the question is whether the net externality is positive or negative.

So if a college diploma is 90% private return, 10% signaling, then in Pigouvian economics we should tax every college degree by 10% of its cost. Then private return would equal social return.

If college has some positive externalities (e.g., better voters, innovation that creates great spillovers), then we just have to add up the positive externalities with the negative. If the sum is positive, Pigou would have us subsidize college. If the sum is negative, Pigou would have us tax it.

Tom West writes:

The subsidy for college increases the need for college. Reduce the subsidy, you reduce the need.

I fully understand that. However, the impact of de-escalation can still hit the poor.

To pick some numbers out of the air. Let's say a 90% of the good jobs currently require a degree. After de-escalation, perhaps only 70% of the "good" jobs require a degree.

After de-escalation, we're better off as an aggregate, but the cost is borne by the group of those who could go to college with subsidies and can't without. They have seen their share of "good" jobs they can apply for shrink from 100% to 30%.

Now, depending on what numbers you assign, it could turn out to be a better deal for the poor after all.

My point is that the "cost" of de-escalation is borne mostly by one particular group (those from lower income families), and this fact should not be ignored in the calculus used to determine whether subsidies should be withdrawn.

Garett Jones writes:

@Tom West

Since most education subsidies are for the middle class or richer (I say this as a UC Berkeley grad), then we could get rid of the middle class education subsidies and keep them for the poor.

Public universities would be the classic example of intra-middle-class redistribution. Mostly subsidized by not-too-progressive state sales and income taxes, and heading right back to the people who pay them. Right hand/left hand.

So "cutting subsidies" needn't mean "cut for the poorest first".

Bostonian writes:

Getting "rid of the middle class education subsidies and keep them for the poor", as Jones suggests, means hitting the middle with a very steep marginal tax rate. Berkeley costs about $33K per year. If a child from a family with income of $20K goes free and a middle class kid with family income of $80K pays $33K, that comes to a tax rate of 33/(80-20) = 55%.

john hare writes:

A lot of this seems to miss a point I consider important. It should be possible to lease a classroom and hire a professor for a year for $100k or so. If there are 25 students per professor, then basic costs are $4k per student per year. With used books and frugal management, a total cost of $6k per student per year seems quite doable in a free, competitive market.

At costs in this range, it should be possible for the motivated student to work through school with possibly a small bit of help from the parents.
I don't see the need for subsidies or government subsidized loans in the first place given the availability of free market education.

I believe that bigger cars are safer than smaller cars even if everyone gets bigger cars. (Assumptions: 1. hold velocity constant; 2. occupants are well strapped into the interior of the car.) A bigger car suggests a bigger envelope around the occupant. More metal to crumple before the occupant gets injured. And greater distance and time through which deceleration can occur. Therefore a lower maximum to the G forces which the occupant will incur during the timespan of collision.

Silas Barta writes:

@Tom_West: After de-escalation, we're better off as an aggregate, but the cost is borne by the group of those who could go to college with subsidies and can't without. They have seen their share of "good" jobs they can apply for shrink from 100% to 30%.

And per your example, you've tripled the good-job opportunities of those who don't spend the opportunity price of college in the first place. Sign me up! (And I mean the anti-royal "me".)

John Fast writes:

I am a good progressive. Two decades ago, my hero, Bruce Ackerman, proposed (in The Stakeholder Society) that since cost of education a closes proxy for social class and unfair privilege, we should tax education rather that subsidizing it.

David Friedman writes:

I offered a version of your argument to Robert Frank in an exchange on my blog a few years back. He was making a fixed sum argument, where the benefit to me of going to a good school was that I got a job that otherwise someone else would have gotten--not that total production actually increased. I pointed out the implication that governments should tax schooling instead of subsidizing it.

He never rebutted and never agreed.

You can find my argument at:

http://daviddfriedman.blogspot.com/2010/05/my-response-to-robert-franks-reply.html

Tom West writes:

@Silas Barta: And per your example, you've tripled the good-job opportunities of those who don't spend the opportunity price of college in the first place.

That would probably be the biggest positive of removing subsidies. As I said, it's not clear to me that de-escalation would be a net negative for the less wealthy.

My claim was only that because de-escalation would impact one sector disproportionately, that sector should be looked at in specific.

As Garret Jones and Bostonian replied, there are ways of mitigating impacts and real problems with such mitigation. All of which point to the basic fact that economic models are only the very first (valuable!) step in determining what policies we actually want to implement in our society. But sadly, reality is neither as clean or as clear when it comes to determining whether policies suggested by those models are actually a net benefit to society (starting with trying to define what "benefit" means...)

Tom West writes:

By the way, I would like to express my appreciation to Garett Jones for replying to comments here. I realize that reading and replying to comments exponentially increases the time involvement, which is pretty high to make cogent postings in the first place, so I am happily surprised to see Garett's willingness to engage with the posters, at least while his professional responsibilities permit.

Thanks.

Gordon Mohr writes:

A compromise: tax the issuance of credentials (diplomas, certificates-of-completion, official transcripts), while leaving instructional fees untaxed (or even subsidized).

Don't allow schools to make purchase of the credentials mandatory for enrollment. (That is, allow no bundling or combining discounts for credential and instructional fees.)

Perhaps even: exempt from the credentials tax those neutral assessment institutions that don't themselves offer or require classes. (I have to think some more about this one. Such a policy could be a strong nudge towards the 'independent certification' practice that Arnold Kling recently named as a likely 'magic bullet' of educational technology.)

Comments for this entry have been closed
Return to top