What do returns to education say about graduate mismatch?

Andrew Leigh agrees that some people have more qualifications than they need for their jobs. But he’s not convinced that over-education is a problem:

The returns to education have stayed very stable over the past 20 years. If anything, there’s a bigger economic benefit to going to university today than in the past.

It is true that there is no evidence that average returns to higher education have gone down over the past twenty years. But I would not expect the statistics I have been citing to affect average returns as there have always been similar proportions of over-qualified workers, who would have consistently dragged down the averages over time. Though the statistics in my graduate mismatch paper (pdf) only go back to 1991, the time of the last enrolment boom, I also checked some earlier data.

It gets a little complicated because the job categories used by the ABS have changed over the years, but matching as much as I can we get very similar over-education statistics through the years. The earliest data I could find was from 1979, and at that time the proportion of graduates in non-graduate jobs (with the caveat in the first sentence) was 18.7%, remarkably similarly to the 19.2% I calculate for 2006. For 1986 I arrive at a figure of 19.8%. In 1996 it was 22%, but that was a temporary aberration, the unfortunate consequence of the Dawkins enrolment boom graduating into the Keating recession. It was back down to 18.9% in 1998.

There are no signficant shifts in the numbers that can’t be readily explained by the general economic situation and/or passing periods of rapid growth in student numbers. The stability of the percentage over-educated does suggest that there may be attributes of a significant minority of graduates that mean substantial over-education is a permanent aspect of the labour market.

It does, however, highlight the fact that the published average returns to education can be misleading. From my reading of the over-education literature, over-qualified people in the same job categories tend to earn slightly more than the average appropriately-qualified person in the same job category, but they earn significantly less than graduates in professional and managerial jobs. So for the graduates who do find jobs matching their qualifications, the returns are higher than advertised, but for those who do not the returns are below the average.

13 thoughts on “What do returns to education say about graduate mismatch?

  1. Why is having more qualifications than you need for your job considered being overeducated? This assumes that the purpose of education is job training.


  2. That a given graduate’s returns from education can vary from the mean is no surprise. But it’s interesting that the proportion of ‘over-educated’ graduates has remained similar despite a growth in student numbers and higher fees. Does this mean we cannot assume that increasing government subsidised places will increase the proportion of over-educated graduates? Or the reverse?


  3. Rajat – It’s a good question. To some extent, the knowledge that there are lots of graduates will encourage industries that use them to set up or expand, so supply creates demand. Also, some jobs are upskilled; non-graduates could do them but graduates do them better.

    Expanding the higher education system at the margins means taking people who have much weaker school results than the average of those going to uni now, but how are they placed with the other attributes that distinguish graduates who do well in the labour market with those that do not? As I say in the paper, I don’t know for sure but I suspect that they are not as skilled in non-academic areas as well.

    Damien – Though there are debates about how to measure over-education I think your point blurs two issues: an empirical one as to whether someone has qualifications that exceed the requirements of the job, and a policy/normative issue as to whether this matters and if so in what circumstances. We can deal with the first issue fairly confidently, but a conclusion that someone is over-educated does not lead automatically to the conclusion that it is a problem, as I’ve covered in several posts this week (including for the reason you state). But I think there is evidence that for some people it is a serious problem.


  4. At first, my reaction to the post was the same as Damien’s, but after reading the comments I realised that the post is just discussing the (mis)match between jobs and education.

    In terms of the apparent mismatch between education and the requirements of positions, I can see how one might think that there’s a misallocation of resources (in that people are being over-educated for particular positions). It could also well be true that people who are in positions for which they are over-educated may be unhappy – but they can, of course, find new positions in which they will hopefully be happier.

    While there *may* be a misallocation of educational resources in terms of there being many people over-educated for their positions, I’d like to raise the possibility that a person’s overall “contribution” to society (in terms of their unpaid and paid work and other activities) may be increased through more education. This is just an idea, I have no figures about this – but perhaps what this isn’t completely nebulous – maybe it could be measured in some way. For example, from the end of the post:

    “From my reading of the over-education literature, over-qualified people in the same job categories tend to earn slightly more than the average appropriately-qualified person in the same job category,”

    Eg, perhaps a person who is “over-educated” for their position will, on average, be more efficient/productive.


  5. “Eg, perhaps a person who is “over-educated” for their position will, on average, be more efficient/productive.”

    It could be, but it could also be (as someone pointed out in the comments to a previous post on returns to higher education), that graduates are on average more intelligent than say the average non-graduate clerical worker (for example) and employers are rewarding this rather than specific skills acquired through education.

    Some people also argue that graduates in jobs that don’t need degrees encourages credentialism among employers, in that they artificially raise the job requirements. This is generally thought bad for society, because it makes it harder for people who do not have strong educational qualifications, often because of social disadvantage, to progress through the workforce.

    As I have conceded (several times) I am not saying that over-education is always bad, or that even if it is bad we can easily identify those individuals who will end up over-educated. In my paper I am making a narrower point: that if we have more than half a million graduates not using their educational qualifications it casts doubt on the labour market argument, made by Bob Birrell and Craig Emerson, that we need more graduates. On the available evidence, we need more graduates in a number of fields, but not overall.


  6. Andrew, your points make sense.

    Creeping credentialism seems quite a negative thing, leading people to obtain unnecessary credentials if only to compete with others.


  7. “Some people also argue that graduates in jobs that don’t need degrees encourages credentialism among employers, in that they artificially raise the job requirements. ”

    I don’t see why this should happen in the private sector. Employers are generally looking for the least expensive persons that can do the job. In my field companies are now often happy to hire our BSc graduates even though I don’t think they are qualified until they do honours.

    I think it is more likely part of a more compex phenomenon of devaluation of degrees where people doing the same degrees are progressively less and less qualified. It all starts from school.


  8. I think there are three factors, all of which have the potential to explain some current trends, but it isn’t clear to what extent each one is occuring. It would be interesting to know what thoughts and perhaps data people have on this.
    1) credentialism (i.e., pointless qualifications — Andrew & Sacha)
    2) degree devaluation (i.e., for numerous factors, like less well educated high schools students entering — Boris)
    3) More education is needed to do the same named jobs (hence the need for more educated people).


  9. Hi Andrew, coincidentally I’ve been working on a paper with PC which touches on some of these issues, and which PC’s supposed to be giving at the AFR conference on Wed. I don’t see over-education as a particular problem in itself – given the roughness of definition and looseness of data, and the rough data tend to suggest we’re about on a par with the UK and its more of a problem for immigrants than Australian-educated people etc. A lot of it comes to down to views about human capital vs screening views of education, where hard evidence is not abundant. I agree re the graduate output argument – but the main point PC and I were getting at was that looking at the graduate supply end will not tell you everything, or even much, about labour mismatches. Teacher education is a good case in point.


  10. Lawrence – Certainly labour markets in other countries show over-education as well, which is an argument against our higher ed policies being particularly bad in this regard (all Western countries subsidise higher ed, presumably encouraging over-supply). As one of my posts last week showed, OE certainly looks like a problem for some people – but how much can be done at the policy level to reduce this I do not know.


  11. The fact that higher ed is subsidised does not mean that there is oversupply. It might be the case that a free market would undersupply higher ed, possibly due to the presence of positive externalities. Now, this issue is of course subject to debate, but the key point is that the existence of subsidies does not imply oversupply. This gets back to the point I raised in my earlier comment on this thread. When using the term “over” (or “under” for that matter), it is necessary to be precise about the benchmark. In most cases in economics, it is relative to some “optimal” level of the variable in question.

    In terms of the human capital versus job-market signalling debate, I suspect that higher ed plays both roles. However, I think that for most degrees (especially the traditional ones), the human capital effect dominates. I have no empirical evidence for this. However, there are probably some exceptions.

    John Quiggin wrote a journal article some time back suggesting that he thought the human capital story was more relevant thatn the signalling story. I think it was published in the Australian Economic Review, although I’m not sure about that.


  12. The second last sentence in the second paragraph of my previous comment should be followed by a sentence that is missing. It should read as follows:

    I have no empirical evidence for this. It is just a feeling that I have. However, there are probably some exceptions.


Comments are closed.