Saturday, November 10, 2007

Defective teaching ethos in university economics

One prominent academic I know dislikes scientific exhibitions in museums (such as ScienceWorks in Melbourne) on the grounds that it makes 'science' seem too easy. The same academic dislikes the fact that I take my environmental economics students on fieldtrips to look at specific environmental problems. He thinks it is all a bit 'down-market'. He prefers Lagrangians to faeces in a pond or to birds nesting on remnant scraps of native vegetation.

These thoughts occur to me because I increasingly notice a trend among teachers of economics and finance in Australian universities towards what I would call educational sadism. It is the idea of deemphasising fact and institutions, accelerating the presentation of material so that the latest theory is used as exposition and generally making academic work as difficult and abstract as possible so that students have to jump really high intellectual hurdles to pass final exams. The secondary motivation these sadists come up with when pressed is that (i) education is a screening device designed to separate out an elite and (ii) that ultra-demanding quantitative courses that include huge volumes of material and vast numbers of concepts are necessary to provide intellectual fodder for doctoral programs. Doctoral graduates, in the main, re-enter academic life and become part of a new generation of educational sadists.

Indeed these new graduates often turn out to be worse than their teachers since they remember most clearly the most recent material they have learned and forget earlier motivational material. Hence they press advanced theory onto neophyte students and skip the motivational material on the grounds that the former is obvious to any ‘fool’ and can easily be omitted. Thus you can teach students about poverty by showing them a Lorenz curve or explain demand theory by solving an optimisation problem with Lagrange multipliers. You can teach students how firms finance themselves by means of counterexamples to the Modigliani-Miller theorem and so on. And if, as an older academic, you question any of this you are given a snappy impatient look that has a ‘take him off to the knackers yard’ feeling to it.

The difficulty with educational sadism is that it leaves out the most important component of a university education – namely learning. Don’t we all want students to understand what they are on about – not just to be able evaluate them? This amounts to asserting that education should impart human capital. It should not just be a screening device.

My view is that a lot of economics graduates from major Australian universities often do not understand much economics. Graduate level theses are often plodding applications of quantitative technique that show very little economics understanding. Graduate students operate often using only very restricted areas of their brain and are a rather frightened about encountering reality and indeed in considering applications of theory. Somehow we have sidelined the issue of thinking. Indeed understanding much about anything these days in a graduate thesis is problematic because understanding typically does not come into it. The result is a cut-and-dried world of lobotomised intelligence involving Lagrangians and Hamiltonians often with very little attempt to understand anything. Markets and market processes are often supressed.

Students who have spent 5 or 6 years studying economics are fearful of an elementary supply and demand argument but have no difficulty at all with a mechanical co-integration argument or an elaborate introspection using game theory. Time series econometrics and game theory have been among the destructive forces in modern economics. Applications of time series analysis in finance have the further distinction of being some of the most boring and thoughtless material you are ever likely to encounter in a university. They neither provide insight or illustrate intelligence.

Am I guilty of teaching sins myself? You bet I am – or at least I was. But I am a reformed character and I do try these days to think about my audience’s capabilities and my main desire is to instil learning. I am also primarily interested in the world – in the meat – not in models per se but in models designed to help us understand the world. My interest is the world. I like models that illuminate things and despise those models that don’t illuminate anything other than narrow, technical prowess.

For my sins - and my loud-mouthed claims on teaching - I have been put in charge of a first year microeconomics unit at my university for 2008 – a challenging assignment. It has made me think hard about what I think students should get from a first unit in microeconomics. I have examined a number of standard textbooks – they are all pretty much clones of the early text by Paul Samuelson - but am stunned with obvious sillyness in more modern approaches.


One of the texts has Powerpoint slides conveniently delivered by the book’s publishers and treats the theory of firm costs involving twenty new concepts in a one hour class with the really critical idea of marginal cost submerged in their somewhere. It is impossible for students to effectively internalise this amount of material in one hour.

I want a less cluttered introduction to microeconomics and will look for one.

The sickness I see in Australian economics instruction is something I will comment on in my blog over the coming months. I think things have gone right off track but I am prepared to listen to contrarty claims and to publish such as posts if there is interest.

One former Department chairman recently told me his honours students are told to pick a ‘paper’ and work on it for their honours dissertation. Why not ‘pick a problem?’ The same department rejected someone I would regard as the best young economics prospect in Australia on the grounds that he was too ‘policy-oriented’. The ideal it seems in these morgue-like departments is dry, uninteresting rubbish.


Our major local economics journal the Economic Record has lost its credibility and rejects anything that has policy relevance almost as a matter of principle. I recently did a literature research on the ‘Record in the 1950s and was stunned with the high quality and interest of the work in these early years. When I remarked on this to one of Australia’s most eminent economists he gave me an embarrassed look and said ’yes it was more interesting then’. And indeed it was. What has gone wrong?

Economics is an exciting discipline. I still believe that, despite my complaints where are attacks at the margin, there is no finer undergraduate university course a student can undertake. Theory, institutional understanding, history, politics, maths and statistics you get it all to at least some degree. A good economics courses should develop in a student the capacity to gaze at the world and to understand. This does involve theoretical understanding but it also involves knowledge of the world and institutions.

While I admire economics as a field that does not mean everything is fine about teaching it or about current research directions. Nor do the good aspects deny me the right to exercise a grizzle at current directions that I particularly dislike in economics teaching. Over the coming months I’ll try to offer some more constructive suggestions to improve things.

10 comments:

Anonymous said...

how about cutting out the keynesian bullshit as a first step, harry. that way you may actually get the kids to understand what markets are all about.

Professors don't even seem to understand markets and the market process.

And get them to appreciate just how important monatary policy is to the whole tool kit. Do that and they're well on their way to undertanding the difference between crap and reality.

Anonymous said...

Hi Harry,

I would be interested to see a more focussed version of your critique. Maybe focus on one issue like teaching firm behavior. Here are a few off-the-cuff comments:

1. Saying game theory is destructive is too broad. Sub-game perfection (for me) is a very useful concept for understanding the real world and plays a key role in a book like Dixit and Nalebuff which is very applied (and useful). Mixed strategies is another useful concept for understanding behavior (again illustrated nicely in Dixit and Nalebuff) - this musn't be what you have in mind ...

2. Institutions and issues come and go. There are surely diminishing returns to studying these e.g. would a student really be better off with another lecture on the details of Work Choices than a good lecture on monopsony (or game theory applied to union-firm bargaining for that matter - again the concept of a threat point in Nash bargaining is a useful one)?

2. My (10 years) experience is that teaching quantitatively material (Lagrangians and all) with an eye on applications improves understanding of the model. I have only had maybe (and this is a big maybe) of one student who had a good understanding of issues without simultaneously being on top of things quantitatively. There were plenty the other way who were on top of things quantitatively and could then go the other way.

3. Another reason for not spending too much time on institutions is similar to the argument for teaching microeconomics and not management and marketing. We teach general/abstract reasoning that can be applied in a range of situations. Institutional knowledge and policy knowledge can be taught or learnt more easily outside of universities - our advantage is in teaching how to use instrumental variables properly or a good understanding of Cournot oligopoly. For example, at the ACCC it used to be the case that they would send new people in the telco unit to a course to learn about the industry - the teacher there would almost surely do a better job than we could.

4. I agree with honours student working around a paper - even replicating - they don't have the skills to really tackle a policy problem properly - again have seen a few policy oriented theses and they generally seem thin and do not demonstrate deeper understanding that a nice piece of simple structural modelling or a nice piece of applied game theory would have.

5. My feeling is that econometrics is a lot better than say 10 years ago. More sophisticated structural modelling, more attention to identification arguments and focussing on what is actually being measured (rather than regressing and spinning a story off signs and significance (though sometimes this is all you can do)) to me has made for much more interesting empirical work (at least in applied microeconomics). Unfortunately we get too little of this work in Australia...

This is enough for now - will be interested in a response and further posts

Anonymous said...

P.S.

Two books that I thought had some nice discussions of core micro theory that you could draw on (though both are almost surely out of print)

Paul Heyne "Economic Way of Thinking"

Alchian and Allen (micro part) of University Economics (I think - it is a famous book anyway)

hc said...

Second anonymous (why not use can identifier if you wish to remain anonymous),

I think game theory is useful and can be introduced early on - with minimal formalism - to explain cartel behaviour etc. I think overall it is pushed too far.

I think lots of institutions and facts live on. Corporations, limited liability and the way firms finance their operations are important institutional data. I doubt studentsd from high school have a clue.

I think for undergrads computing demands from a Lagrangian offers little more insight than drawing indifference curves and shifting a budget line.

Econometrics and time series work are increasingly becoming a substitute for thinking. I think their impact has been fairly disasterous on the profession.

I disagree with your remaining points but responding in detail would amount to another post.

Both Heynes and Alchian & Allen I am looking at carefully. They are excellent books. I also think the original text by Samuelson & partner - now split into macro & micro bits is good. The text by Robert Frank also very well-written but he gets bogged down in detail - the exact issue he seeks to avoid.

Ted Sieper once told me that he thought Alchian & Allen was one of the two greatest books written in economics - the other being Wealth of the Nations. It certainly emphasises economics understanding but it is hard going for first years.

Finally I agree my comments are too general. I'll try to tie them down to specifics in future posts.

conrad said...

There's some balance to this. You can certainly go far too easy, and that point certainly brings higher student satisfaction than an acceptable level (probably far higher). We have pressure to do this where I work (even if implicit). I'd love it if the students I taught could simply think through a problem step by step and understand simple graphs (which a lot of undergraduates can't these days). Alternatively, I used to try and teach some more cutting edge stuff that is hard to understand, but I gave up on that when I realized approximately .1% of the students understood anything from it.


THere's also some difference between knowing how to use some technique to solve a problem and how to press buttons vs. understanding the basics of technique so they can understand papers that use it, but not be competent about doing it. We teach our 4th years to use SEM, but almost none use it in their theses. So whilst I think its a doddle, the actual application is very hard for most students (of course, I'm not in econometrics). However, I'm sure many at least get the main concepts.

I'd love to do field trips incidentally (and the students would probably love it too), but in my area it would be too hard to organize, I couldn't get through ethics, and there are too many students. C'est la vie I guess. They'll have to go to a good university next life, which doesn't exist in Australia.

rabee said...

Well why don't you design an economics curriculum
at your home university that you think is good. I for one am certainly in the market for a good curriculum.


If you're not going to learn difficult and abstract materials in you early twenties, then when are you going to do it? After you leave University?

On the other hand students will almost certainly have many opportunities to gather market and institutional experience after they leave university.

So it make sense to teach students difficult materials when we know that in all likelihood
their few short years with us is the only opportunity that they have to learn this material; and when we know that they will be picking up institutional and market experience after they leave university.

I find that twenty something students who are intelligent are absolutely enthralled with difficult abstract work.

conrad said...

Unfortunately Rabee, whilst 20 somethings might like a challenge, 19 year olds don't. Using game theory :) many people teaching end up in a prisoner's dilemma type situation, where as long as enough students want an easy course and are willing to complain (which of course they do), they'll get it. People never complain that a course is too easy, even if they don't learn anything. Many students also confound the effort with learning something hard with a poor course. They also confound finding everything easy with a good course. If you want to cynically manipulate your teaching scores, this is the easiest way to do it.

I had the interesting experience this year of having anonymous comments on an electronic board, and I actually some say people say that the typical complainers who complain about everything and more (generally 19 year olds kids that have never thought for themselves and don't understand that they need to some day) were just that. So perhaps there are more people like you that hanging around universities these days.

Anonymous said...

Hi Harry (Anonymous again)

Thank you for your response. I would like to follow up two of the points you raised.

1. With the example about Lagrangian's and demand, my experience has been as follows:

a. To draw the budget constraint (and examine effects of it changing) they need to know the budget equation.

b. They should have some idea of a utility function.

If by seeing the diagrammatic version of it most could understand (as opposed to replicate) the next step to set up the problem as a Lagrangian should not be difficult as the object to be maximised and the constraint follow straightforwardly from the graph. At least they should be able to write it down. My experience is that weaker students cannot do this. Then we go to model the efficient outcome (just a different constraint) it should also not be a large leap - but it is. So I conclude that for some reason they only get a limited understanding from the partial explanation (or else I can't teach!).

2. With econometrics it is important to distinguish between different types which have differing degrees of connections to economic theory.

1.a Certain branches of econometrics such some fields within time series econometrics are openly atheoretical or at least propose to focus solely on forecasting.

1.b An intermediate class (in applied microeconomics) is the measurement oriented (natural experiment based) approach popular in labor and influential elsewhere - the link to theory depends on the question being asked and how the data used. The minimum wage work by Card and Krueger while not explicitly based on theory was addressed at an issue directly related to theory (some more freaky applications less so)

1.c Structural modelling such as in Agricultural Economics and modern empirical industrial organization is (in varying degrees) tightly linked to theory and at its best can be used to illuminate problems where theory cannot resolve (how many firms are required for competition - effect on consumer surplus of water restrictions).

I used Frank once and found it too wordy to the point of being unclear - in contrast to Varian's intermediate text (though a former colleague told me that at her previous university in Spain they used Varian as an intro text but just did it very slowly and thoroughly - another approach). Varian sometimes misses the big picture stuff though (which Alchian and Allen is good on)

cheers

hc said...

There are quite a few good comments here. This was a blog post and it is a bit rough at the edges. I am not opposed to theory - I just think things are being pushed too far.

But Rabee's defence of 'difficult abstract work' for 'intelligent' students makes me wonder. This is not a defense based on usefulness but on the fact that it is 'difficult' and designed to appeal to the 'intelligent'. This come close to the core of the problem I have.

But yes I have drafted a syllabus Rabee and the core difficulty is what to omit. Its a problem that Frank has grappled with but not overcome.

Anonymous, in introductory classes on microeconomics most demand theory can be suppressed and attention focused on market demands. Later courses can articulate preferences and utilities as a way of representing them.

I have used the Varian text for 10 years at second year. Its an inspirational book that is exceedingly well written. It doesn't provide applications that get people to the point where they can do much economics.

I need to asrticulate some of these ideas in more specific settings. I don't want top get backed into a corner where I am portrayed as 'anti-theory'. But I do think graduate students are missing core ideas in economics through too much clutter and a bit too much pretension.

gradstudent said...

This is just another example of the older generation putting the boot into the next. You hear this complaint every generation. Why can't people hang up their technical academic boots with pride, instead of insisting that the world follow their intellectual path into academic senior years and policy economics? The old are better at that, the young are better at abstract stuff. Live with it.