How relevant is ‘market failure’ to policy analysis?

The National Centre for Vocational Education Research this week published a series of papers on competition in the vocational training market, including one by me, and another by Australia Institute boss Richard Denniss.

The Denniss paper questions the role of markets in vocational education because of ‘market failure’ – defined as real-world markets not matching the theoretical perfectly competitive market of economics textbooks.

I would not contest the basic observations on which this claim is made. For example, there is ‘imperfect information’ in education markets (eg it is hard to know which education provider is best). There is scope for ‘inter-temporal mismatch’ (eg students may take courses for which there turns out to be no employer demand when they graduate; providers may offer courses only to find that student demand has changed by the time they start).

But describing issues like these as ‘market failures’ is not a good analytical approach. Rather, these are inherent problems in coordinating the delivery of education, with which any coordinating mechanism, market or state, has to contend.

So the policy question is not whether the education market meets theoretical standards rarely observed in the real world, but whether better long-term results are likely to be achieved through education providers and students interacting through free exchange, or through central direction and control.

This was the approach I tried to take in my paper, using empirical data rather than pure theory so that we compare, so far as possible, what actually happens in markets, quasi-markets, and central planning.

I think the information problems in education are, if anything, under-stated by Denniss. He focuses only what students don’t know enough about. But knowledge about students is important as well. Student aptitude, ability, motivation and effort are all vital inputs into the final educational product.

A problem with central planning approaches is that they by their nature put low value on information about students, with the partial exception of prior academic performance. The point of central planning is that the planners think that they know better than both the students and the providers. This doesn’t preclude some effort to discover student preferences, though typically Australian central planners have ignored student applications data and forbidden price signals.

However in Australian post-school central planning the state has only controlled supply of courses. It can’t actually force students to study courses that don’t interest them. So demand does de facto have some influence, though in an inefficient way – the system only adjusts when it cannot fill places at all, rather than trying to optimise the initial match between students and courses.

While students may not be expert on the relative merits of different courses, in practice Australia’s central planners neither know nor care. There have been very limited quality evaluations – there was no organisation dedicated to assessing university quality until earlier this decade, and there are no implications in places or funding from its reports. Students may have imperfect knowledge, but the planners have no knowledge.

The problem of ‘inter-temporal mismatch’ provides no strong case for central planning either. I report studies showing the woeful record of labour market forecasting, and in higher education they haven’t even attempted to solve this problem. I argue that history shows the central planning is highly reliant on both the competence and capacity to act of the central planner. Both have been lacking in higher education, though less so in vocational education, where a federal system, strong employer input, and competition from a large private sector have helped maintain focus.

In higher education, we have good demand data for undergraduates. This shows that demand has responded in the appropriate manner to labour market trends. The ‘inter-temporal mismatch’ has been on the state-controlled supply side, particularly in the health professions. In vocational education, I was able to show responsiveness to labour market trends (ie places falling in courses leading to occupations in low demand, and rising in courses leading to occupations in high demand), though I was not able to isolate the causal mechanisms.

A market system of vocational and higher education would get some things wrong: nobody can know the future for sure, knowledge of institutions is often hard to get and usually historical, students don’t always know what they should do or their own capacities, etc. There is certainly better (vocational) and worse (higher education) central control. But the Denniss paper does not show that in practice a central system could out-perform a market system.

A theoretical ‘market failure’ is at most the first step in an argument against markets. It’s an easy step, because few real markets won’t have some departure from ‘perfect competition’. The hard part is showing how inevitably imperfect state institutions will systematically outperform inevitably imperfect market institutions.

12 thoughts on “How relevant is ‘market failure’ to policy analysis?

  1. Andrew the problem I can see with a market based system is that there would be even more over-enrollment is very popular/trendy courses such a Marine Biology was in the 90s. Yes 1000 kids each year want to be Vets/Zoologist, but we only need 50 of them.

    However flooding the world with dentists might bring the cost down for a check-up or filing so I’m all for that.


  2. M – Significant over-supply is the normal state of the graduate labour market. Indeed, the centralised system has largely avoided graduate shortages not by good planning but by making education so cheap to students that the higher education sector floods the labour market with graduates. It’s only where they have tried to plan (like medicine) that we see chronic shortages.

    So it is not really clear that over-supply of popular courses is any worse than over-supply of any other course.

    While faddishness is a possibility, my paper argues that there is little evidence that it is a significant issue in practice. Applications tend to be quite stable between fields of study, so if fads really occur they are within broad fields of study. And I point out that there are various reasons why supply won’t necessarily rise to meet faddish demand.

    Perhaps the biggest issue here is that we have been reluctant to use price to regulate demand. While there are still significant opportunity costs, if you don’t use prices you are not using a major method markets use to control demand.


  3. That’s an incredibly weak argument to come from the head of the Australia Institute. Is that seriously the best they can come up with? Has anyone ever asserted that real-world markets conform to asumptions made for theoretical purposes? Talk about a strawman.


  4. I don’t see why they can’t just have a set, conservative number of commonwealth supported places and then let anyone else who doesn’t make the cut have the option of paying full fees like international students.


  5. Mitch – Because that still requires the central planner to make a judgment about how many places are required and who should receive one (or the criteria by which they receive one). So this still assumes a high level of centralised knowledge on the part of the government, which it does not have.


  6. Who cares about most fads anyway? A lot of undergraduate degrees have nothing to do with particular vocational skills for particular jobs, but are still useful to have. I imagine that, for example, a degree in marine biology would be quite useful in other areas (chemistry, statistics, some maths, biology, geology etc.) — you could do a postgraduate degree in medicine after that if you wanted something vocational, and if you happen to know a bit more about fish than, say, someone with a degree in philosophy, is that really a problem?


  7. Incidentally, does anyone know where the Australia Institute’s money comes from? The CIS does not publish an annual report on its website but does say broadly what its sources are, but with TAI, not a word.


  8. If no one can accurately predict where the economy will be in 3 years time, how can the government accurately predict what kind of higher ed courses are required? The idea that central planning can prevent too many people taking ‘fad’ courses is similar to the idea that governments can prevent asset bubbles. Was it so clearly wrong (at the time) for lots of people to study IT in the late 90s when the Nasdaq was booming? Or train to be a real estate agent in 1988? Or to enter mortgage broking in 2007?


  9. “Incidentally, does anyone know where the Australia Institute’s money comes from?

    A branch of the Kantor family (related to the Murdochs) has bankrolled the Australia Institute and other similar causes. Presumably there are others. You can donate on their website.


  10. I think Denniss’ paper, along with Tom Karmel’s Overview highlight a major problem reformers face when trying to argue the case for market reforms – the appalling rubbish taught in most economic degrees.
    Denniss is quite right in pointing out that the theoretical perfectly competitive markets of economics textbooks don’t match how real world markets works.
    I think we need to trash the perfect competition model and show that real world markets don’t need to deliver an ‘optimal’ or ‘efficent’ allocation of resources to deliver the best outcomes. Hayeks’ Use of Knowledge in Society would be a good place to start, but I don’t need to tell you that, Andrew – it underlies your paper in this collection. Maybe we need to make the arguement more explicitly?


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s