Friday, May 21, 2010

What did abolishing university fees in Ireland do?

This is the abstract for a working paper of mine just released by the Geary Institute (& the UCD Economics School), a less technical summary is at the end.
Abstract:
University tuition fees for undergraduates were abolished in Ireland in 1996. This paper examines the effect of this reform on the socio-economic gradient (SES) to determine whether the reform was successful in achieving its objective of promoting educational equality. It finds that the reform clearly did not have that effect. It is also shown that the university/SES gradient can be explained by differential performance at second level which also explains the gap between the sexes. Students from white collar backgrounds do significantly better in their final second level exams than the children of blue-collar workers. The results are very similar to recent findings for the UK. I also find that certain demographic characteristics have large negative effects on school performance i.e. having a disabled or deceased parent. The results show that the effect of SES on school performance is generally stronger for those at the lower end of the conditional distribution of academic attainment.

http://www.ucd.ie/geary/static/publications/workingpapers/gearywp201026.pdf

Non-technical summary


1. The paper shows how the abolition of university fees in 1995/96 did not help the chances of poorer children getting into university.

2. The paper also explains why this is the case:

  • There was (& still is) excess demand for places: there is a shortage of places not students.
  • The fee reduction benefitted well-off students, low income ones would have been exempt.
  • Most importantly: the paper shows that it’s how students do in the Leaving that matters. The fact that the low income kids do worse in the Leaving is why they are less likely to progress. Changing fees didn’t change that.
  1. The paper documents precisely how students from better off backgrounds do better in the Leaving.
  • If your father is a professional, count on getting about 90 points more than if your father is a manual worker.
  • If your father is “other white collar” count on getting about 50 points more.
  • If your father is unemployed that “costs” you about 30 points.
  1. The paper also shows
  • That it’s the difference in Leaving Cert performance that explains why girls are more likely to progress to university
  • That if a student’s father is disabled that their points are about 50 points lower
  • That if one of their parents is deceased that their points are about 40 points lower.
  1. A clear policy implication of this paper is that attempts to tackle inequalities in university access that do not address these performance differences at the Leaving Certificate won’t solve the problem.


13 comments:

Peter Carney said...

Interesting stuff Kevin but I am surprised by your findings.

I will have to read the paper in full but from the outset the result seems to contradict basic economic theory: lowering the price of something will positively increase the demand for it. And since participation did increase substantially I find it difficult to believe that all the new participants were ex ante would-be participants.

One reason why the effect of the 'free fees' initiative might not be as salient in the data as we might expect is due to the window examined -- rather than expect an immediate effect, I would imagine that the full effect only actually comes 4-5 years after the reform was introduced i.e., for the kids who started secondary school with awareness of a viable academic trajectory. For these kids, the real and "psychological" barrier of fees was removed and removed at a time when it could have an impact on their career choices, aspirations, and leaving cert points.

Kevin Denny said...

Yes you will have to read the paper in full & revise your basic economic theory too! When there is excess demand, does lowering the price increase quantity? Eh, no.
So is there any evidence of these psychological barriers? Which kids are you referring to, the ones who weren't going to be paying fees [clearly not affected by the reform] or the ones who were paying them anyway [likewise]?
Well if you think the reform had an effect later I think the onus is on you to show it. But given that excess demand has continued to be the case (Figure 1) that would make no sense.
Moreover, as the paper shows clearly the SES gradient is due to the Leaving Cert performance so the reform doesn't speak to that.

Peter Carney said...

I'm talking about the categories in-between the means-tested poor and the trustafarians, the kids most likely on the margin of participation and therefore most likely to have been impacted by the reform.

I'm interested in the topic and really pleased to see that you are working on it. My view, and that's all I can afford to offer at this point, is that there is strong reason to believe that the 'free fees' initiative was not an ideal (cost-efficient) way of achieving the stated objectives of educational equality but it must (logically) have had an impact on the kids at the margin. The chronic excess demand argument seems to be camouflaging the perennial increase in the supply of places over the period that were filled by kids..

Anonymous said...

Peter, it is precisely at the margin that 'free fees' had least effect, because those in that position had other supports available to them, which in turn have been asset stripped in order to make free fees for the middle classes affordable for the taxpayer.

Anonymous said...

Kevin,

Thanks for producing this research: it's a valaubale public service and a welcome addition to the debates on educational inequality and university finance that so often overlap on the issue of the fee-abolition that was initiated in 1996.

From a wider perspective, I would like to emphasise that, irregardless of empirical findings from within Ireland, we can look across the water. Writing on the LSE blog, Nicholas Barr says that:

"Tax finance does not widen participation. Between 1960 and 1998, when there were no tuition fees, access hardly improved." Here's the link:


http://blogs.lse.ac.uk/politicsandpolicy/?p=698


Barr also outlines strong theoretical arguments (again, in the post linked above), as follows:

-BEGINS:

# Taxpayer finance can harm access by leading to a shortage of places. Even without the economic crisis, universities will lose out to the NHS, nursery education and school education. And if places are short, it is predictable who gets left out.

# The people who go to university continue to be mainly from better off backgrounds. Why should the taxes of the truck driver pay for the degree of the old Etonian?

- ENDS

The rationale in Barr's two points above should make it clear why fee-abolition had dubious merit.

I have discussed these ideas (and your research, Kevin) with some friends, and most appreciate the empirical result that educational inequality had not been reduced at third-level (by fee-abolition). Most also see the logic in arguing that fee-abolition was:

(i) at best (sic) a subsidy to the families of the students who would go on to higher education anyway

(ii) at worst, a political strategy to win votes (which didn't work anyway)

The question that is put to me though, is "So we re-instate higher education tuition fees, and then what? What about the people at the margin, who have the Leaving Cert. points, but whose parents have modest family-resources?"

It is absolutely imperative that this question be answered just as well as you have addressed the issue whether fee-abolition changed the social gradient of university attendance.

This is, from what I can understand, where one of Peter's points come in. I take President Von Prondzynski's point that fee-abolition was a mechanism which likely justified the abolition of other supports to those on the margin. There is no getting away from the simple example that follows though:

Consider a large family: five siblings with ten years between the eldest and youngest. Let's assume that the five siblings all aspire to go to university, all get the grades to do so, and all would benefit from the experience. Let's also assume that they would study 3-year degrees and that the annual fee would be 5,000 euro. For the parents, this amounts to a total cost of 75,000 euro. This is not a small amount of money.

It is probably quite clear what I am driving at: that a student loan scheme should be introduced. This would completely remove the possibility of making an argument that tuition fees might prevent someone going straight to university after school. There are of course other possibilities such as night-courses or attending as a mature student. But I am personally in favour of a student loan scheme. Barr has written papers on this issue; and there is a UK scheme which is quite well designed, in my opinion. Kevin, I think that addressing this issue might be a useful addition to the conclusion in your paper.

cont. in next comment...

Anonymous said...

cont...

There are some issues around loan schemes that I think need to be discussed more. Without a securitisation-strategy (to securitise the value of future loan re-payments) being in place, then the taxpayer would still gave to provide funds upfront so that universities could finance current expenditure. Furthermore, if the loan scheme is interest-free (as it is in the UK) then there is a subsidy (in relation to the time value of money) which is payed for by taxpayers.

Despite these burdens on the taxpayer (which would not occur if there was a securitisation-strategy and interest charges), there would be solace in knowing that students are more likely to achieve higher standards and receive better teaching (one could reasonably assume that these developments would benefit the wider economy). Students would achieve higher standards and receive better teaching beacuse of their awareness that they are paying for a service. (I've discussed this to some extent with Peter, I should state).

I have also discussed with Peter the argument that fee-abolition eradicated the potential effects of (a social gradient in) debt-aversion. My main thought in this regard is that a well-funded Access program (which began at second-level) could eradicate debt-aversion... and well-funded Access programs are desirable in their own right. As Kevin's research shows, even if fee-abolition reduced some psychological barrier (related to a social gradient in debt-aversion), the social gradient in school performance needs to be addressed with Access programs. The most binding constraints to university attendance are matriculation requirements, and the "points" requirement.

Another point that I want to raise is the issue of retention. Hypothetically assuming that fee-abolition reduced the social gradient in university attendance (which we know it did not), what about drop-out? Getting the customer through the door is a waste of time if you cannot make the sale. I am concerned about the social gradient in retention, but fee-abolition is irrelevant to that concern. Access programs are crucial in tackling drop-out.

It is also worth reflecting on the fact that there is a cut-off point at the end of second-level, after which many economists recommend third-level student loan schemes. Again, I have discussed this issue at some length with Peter. My sense of things is thet the economic costs of higher education are, on average, higher. Furthermore, the age of 18 is ofte viewed as the age of independence and responsibility. (Of course, we know that 18-year old students cannot make an independent choice to attend university unless there is a loan scheme to avail of, or if parents are willing to lend support/finance).

The point is: it is worth noting that second-level fees were abolished in 1968, and that this is broadly viewed as a positive development by economists.

Finally, as Kevin notes, there is an interesting line of enquiry as to whether fee-abolition increased the demand for fee-paying secondary schools. In today’s Irish Times: there are 26,000 students in fee-paying second-level schools paying at least €6000 per annum. This may buy a better Leaving Cert: the exam that we use to allocate places for higher education.


http://www.irishtimes.com/newspaper/ireland/2010/0421/1224268773855.html

Anonymous said...

"I take President Von Prondzynski's point that fee-abolition was a mechanism which likely justified the abolition of other supports to those on the margin."

* I should say "for those on the margin"

* Also, the "justification" is from a 'substitution' point of view

* I am a full supporter of Access programs

Peter Carney said...

@ Martin

Just to be clear, you should recognise that Kevin's work does not show that fee-abolition did not change the social gradient of university attendance. From my reading, it points out that from a period from 1994-1998 there was no statistically significant increase in the probability of less well off kids attending university. This is important; as I have already argued - there is good reason to expect that the reform might not have an immediate effect but a stronger one over time. Remember that this has been in place for 15 years, Kevin has assessed it's first four years. We would need different data to make the claims you are assuming here.

My argument for re-introducing fees along with a generous loan system comes in two parts. The first part concerns the student; if they are paying for a service they will want to receive it. Rather than fulfilling some bureaucratic obligation they would actually engage with learning and thinking - and actually want to be educated. The second part concerns the professors, or their university. At the moment all the incentives in academia are for anything other than educating. If fees were reintroduced, universities would be encouraged to reconsider their educating roles and the service they offer; new competition would emerge between the universities as students become more discerning, incentives will realign and and the quality of education will increase. Essentially, you get what you pay for. And since the whole country get spillovers one can justify the government's involvement in a loan scheme.

Finally, I would like to pick up on you your N.Barr quote: "..if places are short, it is predictable who gets left out"

This is crucial - and perhaps essential to some of the arguments being floated around on the issue of excess demand. In Ireland, the people who "gets left out" of higher education are those who have not shown an interest, application or aptitude for academic learning -- those without sufficient leaving cert points.

And if, as Kevin's work shows, poor kids are less likely to gain high points then we a debate about investing in primary and secondary education --

@ universitydiary

What are "other supports available to them" ?

Kevin Denny said...

Contrary to what Peter says I interpret the findings as indeed showing that the reform did not change the SES gradient w.r.t university. First lets define what we mean. By SES gradient I mean the marginal effect or association between SES and the probability of going to university. I think that's pretty standard.
The paper shows that conditional on Leaving Cert results there is no SES gradient. I also show that the reform doesn't change that. Hardly surprising since the reform is not directed at the underlying cause of the gradient, the Leaving Cert.
Focusing the analysis in a window around the reform is normal. You can argue that "oh well these things take time" but that's a bit of a cop-out because the longer you look the harder it is to identify an effect. I am not interested in theories that cannot be tested.
In any event how would we expect the effect to differ over a longer time horizon? Well its well known that private schools expanded as a result. The parents who gained from the reform re-invested some of the money in secondary schools. There are some academic benefit to this - otherwise why bother?
So the Leaving Cert/SES gradient would probably be greater in the long run and hence the university/SES gradient greater i.e. the reform could well have made things worse, ceteris paribus.
The argument put forward that the reform broke "psychological barriers" is unconvincing. First of all we need some evidence [other than one taxi driver]. Secondly if it refers to low SES kids, the assumption has to be that they are stupid. After all, it didn't change their position materially. I find this unconvincing.

Kevin Denny said...

Contrary to what Peter says I interpret the findings as indeed showing that the reform did not change the SES gradient w.r.t university. First lets define what we mean. By SES gradient I mean the marginal effect or association between SES and the probability of going to university. I think that's pretty standard.
The paper shows that conditional on Leaving Cert results there is no SES gradient. I also show that the reform doesn't change that. Hardly surprising since the reform is not directed at the underlying cause of the gradient, the Leaving Cert.
Focusing the analysis in a window around the reform is normal. You can argue that "oh well these things take time" but that's a bit of a cop-out because the longer you look the harder it is to identify an effect. I am not interested in theories that cannot be tested.
In any event how would we expect the effect to differ over a longer time horizon? Well its well known that private schools expanded as a result. The parents who gained from the reform re-invested some of the money in secondary schools. There are some academic benefit to this - otherwise why bother?
So the Leaving Cert/SES gradient would probably be greater in the long run and hence the university/SES gradient greater i.e. the reform could well have made things worse, ceteris paribus.
The argument put forward that the reform broke "psychological barriers" is unconvincing. First of all we need some evidence [other than one taxi driver]. Secondly if it refers to low SES kids, the assumption has to be that they are stupid. After all, it didn't change their position materially. I find this unconvincing.

Peter Carney said...

I have already outlined my arguments on the issue of the data window - and I believe the transmission effect is completely reasonable. I'm not sure there exists any reform capable of transforming a kid in pre-leaving cert who doodled for the past four years into a university point-scoring student. To expect such an immediate effect is erroneous, in my opinion. That is not true of the same kid who is treated early in their junior cycle. This idea is in no way fantastical; it's testable.

Indeed, from what I can see, the data window viewed in this analysis is narrower than what the available ESRI School Leavers Data allow for - that is if the ISSDA's website is correct. According to their information the survey was also collected in 1999/2000 but perhaps there were other difficulties in using these data.

Despite these genuine empirical difficulties I totally accept that the 'free-fees' initiative was not an appropriate (cost-effective) way of achieving educational equality but I still believe it helped level the university playing field.

BrendanH said...

A thought about LC points -- to what degree are these an intermediate outcome rather than a meaningful predictor?

They are both an expression of a desire to go to university, and an admission requirement.

Anonymous said...

@ Peter:

I'll put in the qualification of "by 1998" if that is more acceptable. But I should emphasise that I am skeptical about looking at effects further out in time, because other things will be coming into the mix.

As Kevin says, "focusing the analysis in a window around the reform is normal." Kevin could incorporate the year 2000 in his analysis as a robustness test, but going beyond that date would be problematic, in my opinion.

I also think that Kevin makes an excellent point: that private schools are one thing that would differ over a longer time horion.
Whether or not private second-level schooling confers an advantage in the points race (and I would like to see some research on this), I think we need a real debate on whether the state should continue to pay the salaries of teachers in private schools at second-level.

This especially merits consideration since the abolition of third-level tuition fees, given the possibility that fee-abolition has seen a reallocation of family resources from university fees to second level fees.