How increasing tuition fees can increase university participation rates

The Quebec government has just released a collection of studies on financing education, including this one (121-page pdf, in French) written by Valérie Vierstraete, a professor of economics at the Université de Sherbrooke. It addresses exactly the sort of questions that need to be asked and answered in this debate: what would happen to university enrollment if we increased/decreased/held constant tuition fees, and what are the implications for government finances?

After a certain amount of establishing the context, the paper starts with a simple probit model of university attendance decisions. Since she was obliged to work with data from Quebec, identifying variations in university participation rates that were due to variation in tuition fees is somewhat problematic: tuition fees were frozen throughout most of her sample. Although we might want to take the precision of some of these estimates with a grain of salt, her results are similar to what we've seen elsewhere, so there's no reason to dismiss them out of hand.

The next step is some policy experiments around the base case scenario of keeping tuition at $1617.76/year. She considers several scenarios, but I'll talk about three (the other scenarios involved only minor variations around the base case and generated negligible changes):

1. Free tuition
2. Increase tuition by 50% to $2456.56, which is half the average in the rest of Canada, and 
3. Tripling fees to $4893.13, the average in Canada outside Quebec.

Professor Vierstraete is careful to track the effect of these changes through the tax system: among other things, tuition fees generate tax credits, and they also affect levels for financial assistance. Using a student population of 230,000 as a base case, she obtains the following effects on enrollment:

1. Free tuition: 17,993 more students (7.8% increase)
2. 50% increase: 5788 fewer students (2.5% decrease)
3. 300% increase: 22,120 fewer students (9.6% decrease)

Again, since all three scenarios involve going way outside the observed range of variation of tuition fees, these point estimates should definitely not be taken as definitive. It would have been nice to see some error bands.

The next issue is what effect that these would have on government finances. Once again, she does the hard work of (among other things) tracking funding arrangements between Ottawa and Quebec City, and between Quebec City and the universities. For the cases in which tuition fees are raised, it is supposed that only half of the increase is clawed back in the form of lower transfers from the provincial government to the universities.

1. Free tuition: Government expenditures increase by $153m; university funding levels stay the same.
2. 50% increase: Government expenditures reduced by $7.8m; universities receive an extra $67m.
3. 300% increase: Government expenditures reduced by $22.4m; universities receive $246m more than before.

Professor Vierstraete doesn't make any policy recommendations (that wasn't her job here), but her results make it possible for us to make some rough estimates about the effect of these sorts of policy initiatives.

Eliminating tuition fees would increase university enrollments, but at the cost of an additional $8500 for each of the 17,993 new students. Since more than 90% of those extra expenditures would be going to the 230,000 students who were attending university anyway, this is a remarkably wasteful way of increasing PSE enrollment.

Although the link between PSE participation and tuition fees is small, it is not zero:  tripling tuition fees would force out almost 10% of the students, particularly those from low-income families in small towns. But tripling tuition fees would also mean that universities would have an extra $246m: more than enough to pay the $108m it would cost to offer free tuition to those 22,000 students whose financial situation is too precarious to handle the tuition fee increase. And if they wanted, they could even afford the $88m cost of waiving fees for the 18,000 potential students who would have come if tuition were free. And there would still be $50m left over for other things, in addition to the $20m freed up in the provincial budget.

In principle, a policy of raising tuition fees to the national average and then helping those who are in financial difficulty could have the same effect on post-secondary enrollment as a policy of free tuition. But while free tuition would involve increasing public expenditures by $150m, a policy of higher tuition could actually reduce public expenditures.

Since these estimates are subject to a certain amount of error, they can't be used to justify an immediate increase of %300 in tuition fees. But they certainly make it plain that the appropriate policy path involves increasing tuition fees and using these revenues to help students in financial need.

Comments

[Geoff Wozniak]

Sometimes, I wish the Canadian Federation of Students would be more receptive to (or at least acknowledge) such arguments. Maybe I wouldn't have scoffed at them so much when I was an undergrad. (I still do, as a grad student.)

As a student, I don't oppose raising tuition, but it does cause me to apply for more scholarships and bursaries. I just don't want to see those go away.

[tom s.]

One paragraph I don't understand:
"Eliminating tuition fees would increase university enrollments, but at the cost of an additional $8500 for each of the 17,993 new students... more than 90% of those extra expenditures would be going to the 230,000 students who were attending university anyway"

And one other thing I wonder about:
What about the cost and difficulty of means-testing. Many public programs have been made universal because means testing is divisive/introduces bad incentives/is therefore difficult to administer fairly. Is this an issue with PSE funding if you take the route of increasing fees for those who can pay and handing out the extra to those who need it?

[Stephen Gordon]

The first point refers to the fact that tuition would be free for everyone, not just the 18,000 students who would be taking advantage of lower tuition fees and deciding to go to university. For the 230,000 students who were already going to university, free tuition is free money. Although it certainly affects their welfare, it doesn't change their behaviour.

And I emphatically reject any claim (eta: I realise you didn't make it, but I've heard it often enough to make it a sore point for me) to the effect that the costs of figuring out which students are in financial difficulty are so big that it's cheaper and more efficient to give money to rich kids who don't need it than to determine who does. This is not a hard problem; think of the GST credit, the Child Tax Credit and any number of programs that are designed to help those with low incomes. Identifying which families are poor is the least of our problems.

[tom s.]

Ah, I see what you mean. Thanks.

[Mandos]

The fact that it would be free for everyone is a sociopolitical feature, not a bug.

[Stephen Gordon]

Giving free money to rich kids who don't need it is a bad sociopolitical feature.

[Mandos]

No, it isn't. If it increases the political likelihood that the program will be retained, then it's a positive sociopolitical feature. That's one of the benefits of a middle-class entitlement. You don't want it to be like welfare, which becomes an easy political target, since most people don't think they'll ever use it.

[Stephen Gordon]

Free tuition is a regressive policy.

I can understand median-voter model arguments to the effect that progressives should accept a policy that is less progressive than they would have otherwise liked in order to secure broad support. I don't understand why progressives should use this argument to support policies that are actually regressive.

[Mandos]

I'm dubious about your claims that student voter participation is the reason why tuition subsidies survive. I'm also largely in agreement with Travis Fast on the second thread about the politics of the situation, and appreciate his mention of potential fixes to the regressivity you point out.

I can understand median-voter model arguments to the effect that progressives should accept a policy that is less progressive than they would have otherwise liked in order to secure broad support. I don't understand why progressives should use this argument to support policies that are actually regressive.

I'd be more willing to agree with you if you could present an argument that involves the long-term economic advantages of getting an education, the class flattening effect, etc, etc.

[Stephen Gordon]

The advantages for broadening access to PSE are pretty clear, don't you think? And isn't it equally clear that my preoccupation is to drag in as many credit-constrained kids into the system as possible? If that means that rich kids have to pay more, well, that's a consequence I'm willing to live with.

[Mandos]

The advantages for broadening access to PSE are pretty clear, don't you think? And isn't it equally clear that my preoccupation is to drag in as many credit-constrained kids into the system as possible? If that means that rich kids have to pay more, well, that's a consequence I'm willing to live with.

But we're back to the original question. Will programs targeted to credit-constrained kids survive the end of higher-income subsidies?

Medical tuition has been largely deregulated in Ontario. Anecdotal evidence tells me that the income bracket has gone up. I wonder if anyone has done a study on it.

[Stephen Gordon]

You're focusing on the wrong coalition. Right now, the winners from free tuition start at the top of the income distribution and stop somewhere below the median. Low-income households are getting a disproportionately small and smaller share of the PSE budget.

[Ian King]

I think that's what's often missed is that tuition is not the largest cost of post-secondary education. Hell, Quebec students shell out about the same for books as they will for tuition with rates as they are. The cost of living while going to school (or to look at it another way, the foregone income) is far greater -- and cheap tuition doesn't address this.

The problem with tuition cuts is not just that they're regrassive, as Stephen has explained, but that their effects are piddling in the real world. A five-percent cut will give the rich student a might or two at the bar, but won't come close to closing the four-figure gaps that would-be students from poor families face. Cutting tuition in Canada means one end of the income spectrum gets a break it doesn't need, while the other gets a break it can't use.

Tuition in B.C. nearly doubled between 2001-02 and 2004-05, but headcounts actually went up because schools added more capacity during those years. Headcounts have fallen by a few thousand since, but that may be due to a hotter job market; tuition has grown at CPI during that time. Where the B.C. government fell down was that it didn't offer more generous financial aid for students who needed it -- aid limits increased by only a few hundred a year for single students, and undergrad grants for less well-off students were eliminated. That's the other side of the argument for increasing tuition -- you increase help for those who need it.

[reason]

I still think the system that Australia used (at least for a while) of a graduate tax is a good solution. Like Tom S. I'm wary of means tests (mainly because they tend to produce income brackets with perverse incentives) and fail to cover all individual circumstances (family black sheep for instance, or girls in some families). Surely the issue of progressivity (as you have argued elsewhere) is a question for the entire system, not just education.

[Martin Haworth]

Does it not make sense that where there is a bigger financial commitment, there will be a better chance of students actually delivering better performances?

Regards

Martin

[Marcel]

GOOD DAY,
Please i wish to appply for the canadian free tuition in to the university but did not understand where actually, am from cameroon and will really love if you could help me with the link thaks for concern.
THANKS.