Arnold Kling  

Inefficient Subsidies

Even Bigger Than Arnold Thinks... Econbot?...

James C. Garland writes,

state higher-education budgets are not targeted efficiently. By way of comparison, consider the food stamp program, which in 2004 paid out $27 billion directly to 24 million low-income Americans. Imagine if there were, in its place, a food subsidy program by which the government paid that $27 billion directly to supermarkets. Under such a program needy families would benefit little, because most of the savings would be passed on to customers who didn't need help. That would be an inefficient use of public money.

I think that is a valid analogy. It also applies to public education: if your goal is to improve the education for the poor, then stop subsidizing everyone else and instead give targeted education vouchers that phase out at higher incomes.

Or consider health care. If the goal is to help people who cannot "afford" health care, then national health care is horrendously inefficient compared with a voucher system that targets the poor.

Comments and Sharing

COMMENTS (16 to date)
quadrupole writes:

There is effectively no meaningful way to subsidize the poor going to college. The reason is simple: almost *all* 18 year old adults are poor.

There are those who insist that you consider the *families* wealth when marking a college student poor or rich for purposes of financial aid... but ask yourself this: How is it reasonable to base the college funding options of a student on the whims of people who have no compulsion to do their part. I can't begin to tell you how many people I know who have trouble just getting their parents to fill out the *forms*, let alone contribute their 'expected contribution'. Haven't people with craptacular parents been punished enough without making them have to annually beg just to get a signature on a FAFSA form so they can continue their education?

The way I see it, there are only two reasonable solutions:

1) Base financial aid soley on the individual (not the family)

2) Base financial aid on the family and place criminal penalties on a parent failing to fill out the FAFSA, and make the expected parental contribution recoverable by civil suit.

bluehoo writes:

Unfortunately, Democrats will oppose means-testing for all sorts of government benefits (social security, Medicare, education, etc.), even it would mean initially increasing benefits for the poor, because the believe, not w/o reason, that long term support for those programs will diminish.

Fritz writes:

That creates a moral hazard. Students rich or poor burdened with debt motivates completion. I would rather create disincentives to long completion times, increasing tuition for 5th year attendance, bonus for early completion. Require full tuition payment for course dropping. Get the students out producing GDP sooner. That would be a productivity enhancement. We spend billions for teenage squandering of resources.

quadrupole writes:

Fritz, while I agree with your sentiment (having completed a double major in physics and mathematics while others partied through a 5 year english degree), you have to be very careful of the law of unintended consequences.

Many pharmacy programs are 5 year programs, by design (and hard as hell to boot). I know many folks who did five year dual undergrad/masters programs (all five years counted as undergrad financial aid). I've know a more than a few folks take 5 years because they were triple najoring (physics, math, cs seemed to be popular). If you aren't careful, you will cut off at the knees some of your most ambitious industrious students while you try to sweep up the slackers (who are definitely there).

Bud1 writes:

Directing vouchers/support specifically to those in need is a great idea. We don't need any more overmanaged government programs. Extend the concept to disaster relief as well.

Fritz writes:

I'm talking in terms of 4 year degree programs. If Pharm is 5 years, they would be exempt. Those that double major, ect could be addressed. Giving incentives to attend summer session would be great, the marginal costs are minor, but benefits large..

knzn writes:

Leaving aside the very few who are so rich that they can effectively provide their own insurance, there are only 3 reasons not to get health insurance: 1) because you can’t afford it; 2) because you’re an idiot; and 3) because you expect to get it without paying for it (since in the US, “critical” health care is seldom ultimately denied based on ability to pay). Accordingly, the goal of any reasonable system would be to have 100% of the population insured. That being the case, why is it inherently more efficient to have a lot of little insurers rather than one big one?

I recognize that there are efficiencies gained by competition, but in this case, I think those are outweighed by the massive underwriting costs that could be saved under a single payer system. They are almost certainly outweighed by the market power that a single payer would have, though whether that market power is efficient is a whole ’nother discussion topic.

Higher education is not quite analogous, for a couple of reasons. First, there may be a lot of good reasons that many individuals shouldn’t go to college, so the 100% goal would not be reasonable in this case. Second, there are no obvious advantages to be gained from a single-payer education system, whereas the disadvantages are, if anything, more pronounced than with health care. I agree with you in principle about education, but, as bluehoo observes, the real problem there is political rather than economic.

When it comes to public education, the case is somewhere in between.

quadrupole writes:

Fritz, the problem is you run into exactly the same issues you get with any over tuned government program. How do you think it effects the likelihood of starting a 5 year program (like Pharmacy) if I have to submit it to the Feds for approval (which is what you will effectively have to do, as no fin-aid == no program)? As to encouraging summer school for double and triple majors, it probably would, but then how does that figure into the fact that in MANY , MANY fields you NEED the summer *intership* experience to succeed. Again, you force your most ambitious into a corner in a dragnet for the least ambitious. This also throws a complete monkey wrench at other good programs like coops. In a coop program a student works for a company (like Intel) on alternate semesters and takes classes on alternate semesters. At many schools summer is counted as a semester for this purpose, but you can't squeeze as many classes out of a summer semester as out of winter or fall. Again, a good creative program killed by your lust to get the slackers.

What I am far more concerned about (and have no good solution for) is the moral hazard involved in funding all majors at all schools equally. It just makes no sense to go $120k into debt to get an english degree and start in the $20k range (and not go up much from there). But many, many students do. They make irrational major and school choices because of the availability of easy cheap long term credit. I don't know quite how to fix this problem, but I see it as much bigger than the 5 year issue.

Mark Seecof writes:

quadrupole is right. Moreover, the current "financial aid" process is just the front end of a price-fixing, price-discriminating cartel which ensures that any increase in taxpayers' largess to students is rapidly transferred into college administrators' pockets with no increase in educational opportunity or output.

The best way to fix the whole mess would be to completely eliminate all "need-based" financial aid, and replace it with merit-based aid. We should make available a large amount of taxpayer-funded merit-based aid (payable to students in the form of transferable vouchers), giving larger vouchers to students with higher test scores. We should also offer two or three bites at the apple (say, right out of high school, after junior college, and after undergraduate work for would-be grad students).

For example (I'm sure we could come up with a better array of measures), if you get 1400+ on the old SAT, you get five $20,000 vouchers (good for 5 or 6 years). 1300-1399, four $18,000 vouchers. 1200-1299, four $16,000 vouchers. 1100-1199, four $12,000 vouchers (still enough for two+ years of study at a $20,000/year school, so maybe you should do lower-division work in some JC). SAT 1000-1099, two $14,000 vouchers--take a skilled-trades course, or go to a JC and try to qualify for more vouchers afterward.

The advantage of this scheme would be better support for "poor but deserving" students. The taxpayers would be happier to support bright students, and the students would gain support in proportion to their ability to utilize it.

We should place no restriction on supplementing vouchers with cash but we should *forbid* schools to ask students about their family finances. To prevent invidious price fixing, tuition rates should not refer to students' family finances. At the same time, we should require all voucher-accepting institutions to publish their complete tuition schedules for public inspection, so students can factor price into choice of school.

(We should leave schools free to charge whatever they wish for different courses, or for day vs. evening students, or to give merit discounts, or whatever... We should simply outlaw demands for students' or their families' tax returns. You don't have to show your parents' tax returns to buy a book from Amazon.)

spencer writes:

Yes, but the goal of public education is not to provide education to the poor. The goal is to educate everyone.

Why would your every have such a weird idea?

Maybe it is because you consistently ignore the fact that most of public education does a quite good job --like George Mason university for example. Or do you believe that the middle and upper class students taking economics at the public education facility of GMU are getting a poor education or that their tax payer supported job is only to educate lower class students?

Robert writes:

Mark --

Empirically, subsidizing students for past performance seems to be counterproductive.

Every year, the National Science Foundation awards something like 1000 rather nice fellowships to beginning graduate students. The applications consider chiefly standardized test scores (GRE) and undergraduate GPA. Letters of recommendation and personal statements are also taken, but they are uniformly excellent.

A dirty little open secret is that, compared to other graduate students, insofar as such comparisons can be made (years to completion, whether they complete, number and impact of publications, etc.) NSF fellows are below average graduate students. One possibility is that standardized tests don't mean a thing. A second though, is that having already received the fellowship, NSF fellows don't work as hard as other graduate students. Either way, it doesn't look good for merit-based support on the pre-earned model.

It would be better to pay for performance: offer student loans in which a portion of the loan amount is forgiven if performance goals are met. (This in turn, however, raises the spectre of kickbacks to college officials for granting good grades.)

resigned writes:

I think a good point was raised in that many entering college students are not aware of the correlation between jobs and income. What about a simple weighting of financial aid by employment rate and commensurate salary in the field? For example, let's say that the average salary for someone with a degree in engineering is ~$45,000 after college, then the degree of financial aid should reflect that as compared to that of the english major. Perhaps an additional weighting factor could be given based on the job rate after 2 years for people in the field given income. That way, there is a tie between market demand for a particular field and the money available to study for it. Grants could be provided by the government for areas which it is believed that the market may not currently support, but for which they find a need (for example, arabic language studies post 9/11 might have grants for students who achieve above a certain GPA after their first year--or perhaps an agency issued exam, whatever, the NSA may offer grants to math majors, etc.). Then, if someone wants to study art history, they can continue to do so, if their (or their parents) finances permit it, but they go in with a better realization that there's not money in it. This may have to be fine tuned for newer fields ("nanobiology" and the like).

I think the reason for giving aid is to not waste talent. The problem is that at the collegiate level, much damage has been done. Let's suppose that you have a middle class student in a good school district with involved parents and they work hard. You also have a poor student in a lousy district, again with well motivated parents, that does well in that district. The problem is that the student in the lousy district is likely not challenged as much as the one in the richer district. He's doing well in classes, getting As, but doesn't realize that because of differences in opportunities, he is unlikely to be able to go head to toe with the student in the stronger district and the relative scores of the students on the two exams will likely reflect this. For such students, I think aid should still be available (perhaps even at the community college level, and if they do well there, then they should be elgible for aid at the 4 year college level), but the aid should be biased towards fields which are likely to employ that student.

resigned writes:

One note on grants--one of the more transparent ways of awarding merit scholarships that I've seen was at MSU. They did an initial cull based on PSAT scores (which determine national merit winners). Following this, they invited us to the campus and we took an exam there (ranging from essays to math). Highest scorers got the highest scholarships. This is fine for competition between equals.

James writes:

Sure enough, the goal of public ed. is to educate everyone rather than just the poor. However, providing education to those who would otherwise procure it by their own means is an unnecessary expenditure for accomplishing this goal. If I'm someone who is already willing and able to pay $X for an education for either my self or my child and the state gives me $Y on top of that, where does the extra money go? I might spend it on more education or I might spend it on something else. If I get the $Y in the form of a voucher or in-kind provision, then I spend out of the $X. Whatever the goals and intentions may have been in giving me $Y, it may as well be a tax rebate. Not that I object to tax rebates. I just think that if a policy is to be justified in terms of its goals, it should minimally have the effect of accomplishing its goals.

Mark Seecof writes:

Thanks, Robert. I didn't know of the NSF fellowship story, but I'm willing to believe it. I think it could be fine to switch from prospective to retrospective payment, but I suggest that a good program would have an entrance filter along the lines I suggested (to signal to students what sort of programs they should seek to enter, if any). Once a student begins a program on a promise of "progress based" aid, the amount and timing of that aid should be guaranteed if the student progresses adequately--no one should be cut off part-way through by political changes in the program.

(I experienced something like that years ago, when California cut off merit-based grants to undergrads--awarded yearly, then--and redirected the funds to so-called need-based aid, much of which was in fact awarded by race.)

As for the NSF results you describe, I would bet on explanation number 2: already-funded students slack off. Although I would expect quite a lot of range-compression in test scores of prospective grad students, I doubt that such scores are meaningless (though the grades may be!).

resigned writes:

Say, is there any hard data for the NSF results? Also, for the race based UC results (for example, more cases of poor whites being denied aid relative to middle class blacks?). I've become suspicious of anecdotes as of late (everyone remembers some great teachers, but statistically, they are not the strongest students in college for example and it seems that many are failing basic competency exams.)

Comments for this entry have been closed
Return to top