Off Topic but a Fascinating Discussion

In the last few days I've ended up in a very long debate with @unlearningecon on Twitter on the state of economics. My protagonist writes a blog with the same name: unlearningeconomics.wordpress.com/.

It's clear that while Twitter is good for a great many things, it can be constraining when it comes to meaningful debate, as 140 characters gets a bit too few, and sending multiple tweets gives the impression of badgering someone.

unlearningecon essentially writes that economists are a pretty out of date, aloof bunch whose models are so woefully inadequate that they aren't fit for purpose. These models aren't fit for purpose because along a number of meaningful dimensions, it's quite easy to point out where some of the basic models taught to undergraduates fall down. He then at length points out where he believes heterodox economists have pushed alternative theories but been ignored by the mainstream in its pursuit of mathematical beauty at the expense of empirical validity.

In my discussion with him it seems that much of his thrust is against macro theory and DSGE models - the wonderful representative agent (or agents).

What I think he fails to understand is (a) how any discipline must be taught, and (b) the distinction between basic textbooks and research, and (c) the basic process of research. I could of course be totally wrong on this, and I await correction. As he appears to have got particularly angry in recent exchanges, I decided it best to spend some time trying to set out my position in a format that might help make things clearer without fanning the flames.

On (a), he is clearly a very incisive person, however he does not appreciate that the ability to abstract and think about modelling the world around us is not something that comes naturally to most undergraduate students. Furthermore, it would be a very unusual teaching method that launched straight into the most complex alternative theories to explain individual behaviour in the marketplace, as microeconomics tries to. Hence we begin by teaching the simplest models to our students, fully aware of their faults. And we then use those faults to enable students to start grasping at the difference between a model and reality, and the purpose behind modelling - to try and better understand that reality. I suspect he hasn't been faced with a room full of undergraduate students waiting for a microeconomics course, but I'm happy to be corrected on this.

On (b), he appears to think that economics research simply builds on the foundations of stuff written decades and even hundreds of years ago, unquestioningly. I've pointed him in the direction of New Economic Papers in the hope that he looks at a few of the recent papers in, say, microeconomic theory in order to help him realise that even micro theorists are looking at better ways to understand individual behaviour.

The reality is that most research is attempting to empirically validate (or otherwise, most likely) the various theories economists have proposed over the years. He criticises behavioural economics as a minor tweaking of the basic neoclassical framework that he thinks is the source of all the ills in economics. However, behavioural economics quantifies all the biases individuals exhibit in their day-to-day decision-making, and in that sense can hardly be a minor modification of the assumption of rationality on the part of individuals. Eventually, with empirical and experimental work to boot, hopefully behavioural economics will help us better understand how we make decisions and develop theories that fit the data and results we find.

I sense unlearningecon is totally unaware that this is the purpose behind most research carried out in economics.

Finally, for (c), unlearningecon appears to blur the distinction between claims and facts, making regular unsubstantiated comments (subsequently getting angry when challenged on them).

He asserts, for example, that physics has a much better empirical and forecasting record than economics. This may well be true, but it is simply an assertion and without any evidence to support it, remains that. It may well be that physicists have developed some excellent models that forecast very accurately - but how do we compare that accuracy to models developed in economics? unlearningecon appears to do this based on his personal perception of various economic theories, which he finds implausible, and on an unquestioning stance towards assumptions made in other fields.

What is needed in order to compare the forecasting record of physics and economics is a metric upon which comparisons can be made. A metric such as mean squared forecast errors might be one, yet this may well ignore the intrinsic uncertainty (unpredictability) of many processes we seek to forecast in either field. Yet without it, we are comparing apples to oranges and can make no progress. unlearningecon has got frustrated with me making this point, and appeals to it being a supposedly well known "fact" that physics forecasts better than economics. Maybe people have already conducted an exercise like that set out here, and if so I'll be interested to read it. In the meantime however, unlearningecon is simply pushing an assertion and getting angry that I contest it.