Showing posts with label nobel prize. Show all posts
Showing posts with label nobel prize. Show all posts

Monday, October 18, 2021

Nobel winners make economics more useful, not a parlour game

It turns out that, in economics, maths – like technology and much else – can be used for good or ill. The three academic economists (and one ghost) who won this year’s Nobel Prize in “economic science” used mathematics to make economics more realistic and thus more useful to society.

The reason economics has become dominant among the social sciences – has had so much influence over the thinking and actions of governments - is the belief that understanding how and why people behave the way they do in the economic dimension of their lives – their producing and consuming – will help our leaders solve problems with the economy and make us happier and more prosperous.

But sometimes I suspect that the bulk of academic economists – whose beaverings won’t go anywhere near winning any prize – have lost sight of the goal of improving economists’ understanding of how the economy works and being more useful to the community and its leaders in improving our lives.

I worry that academic economists have become more inward-looking and more concerned with impressing each other than in serving the mugs who ultimately pay their wages. (I make the same criticism of journalists, by the way.)

In the years since World War II, the greatest project in academic economics has been to make it more scientifically “rigorous” by making it more mathematical. To express economic reasoning not in words or diagrams, but in equations.

These days, you shouldn’t do economics at university if you’re no good at maths (which may help explain why student numbers are down). No one gets to be an academic economist unless they’re good at maths. No one gets to be an economics professor unless they’re really good at maths.

Impressing the other academics with your great maths is the way you get on in academic economics. Maths is just so logical, so beautiful, so “elegant”. But sometimes I think these people love maths for its own sake and are turning economics into a branch of applied mathematics.

In an infamous study economists prefer to forget, economists attending the American Economics Association’s annual meeting were asked to answer a question about opportunity cost. Eighty per cent of them got it wrong. Opp cost is the foundation on which most economics rests. Makes you think all these PhDs know more maths than basic economics.

It’s true that expressing an argument in mathematical equations exposes any flaws in your logic – given the assumptions the argument is built on. That’s why the results of modelling – including the epidemiological variety – should be viewed with caution until you know and accept as plausible the key assumptions on which the modelling’s based.

The other day I wrote that economics’ greatest weakness is its primitive model of human behaviour, based on the mere assumption that people always behave “rationally” – which I defined as acting with carefully considered self-interest.

A couple of economics professors took me to task on Twitter. Oh no, not that old canard. “Rational” is just one of the many words in economics that are used to mean something other than their meaning in common speech.

No, what we mean by “rational” is not that people always think logically, but that we look at people’s “revealed preference” – what they actually do, not just what they say. This, I was assured, had long been part of mainstream economics.

Sorry, not convinced. It’s a circular definition: what people actually do (as measured by the statistical data available) is rational behaviour. Why? Because people are always rational. It’s getting around an implausible assumption by making it even more implausible. By defining non-rational behaviour out of existence.

Why would you do that? To make the assumptions of the neo-classical model mathematically “tractable”. That contrived meaning of “rational” may be longstanding mainstream econometrics, but it ain’t mainstream economics. That’s unconsciously assuming economics is now just maths.

When people were going crazy buying toilet paper last year, Australia’s brightest young economist export, Professor Justin Wolfers, argued it was “rational fear” to join the queue because, if you didn’t, toilet paper might all be gone when yours ran out. That was using “rational” to mean what everyone thinks it means.

You can say the same about former Federal Reserve chairman Alan Greenspan’s famous admission in 2008, after the global financial crisis, that he was mistaken to assume the banks’ “self-interest” would protect them from doing risky things that ended up damaging their shareholders.

The commentator Ian McAuley has observed that both engineers and economists use equations and mathematical models, but engineers check their maths against reality and modify their equations accordingly. Economists? Not so much.

To be fair, predicting the behaviour of bridges and suchlike is a lot easier than predicting the behaviour of human beans. This has led many academic economists not to worry about the plausibility of the assumptions on which their model rest.

Just make whatever nips and tucks are need to mathematise the mainstream model and think of all the fun games you’ll be able to play running different “data sets” through it. Other academic economists will be impressed.

Fortunately, not all academic economists are content with their work having such a tenuous link to real-world problems. Nor are the people who decide who gets the Nobel Prize in economics. The various founders of behavioural economics – which my critics contend isn’t real economics - have received awards, including a psychologist.

And the three academic economists sharing last week’s awards were about trying to make economics more realistic and therefore useful to economic policymakers.

Professor David Card, of the University of California, Berkeley, sought to test the straight-from-theory belief - then almost universally accepted by mainstream economists – that raising the minimum wage would increase unemployment, by searching for empirical evidence to support or refute neo-classical theory.

Until relatively recently, economists believed there was no way they could use experiments to test their theories. But a previous Nobel laureate showed some laboratory experiments were possible. And Card showed how theory could be tested by finding a “natural experiment” – a circumstance where the real world had created a test group and a control group, such as two nearby cities in different states, where one state had raised the minimum wage and one hadn’t.

Professors Joshua Angrist and Guido Imbens have done natural experiments too, and have also developed statistical methodologies for going beyond finding correlations between two variables to being able to demonstrate which caused which – showing other social scientists how it could be done.

The point is that the three honoured economists (plus the ghost of Professor Alan Krueger, who was a co-author with two of the three, but died in 2019) did reams of maths – or, more specifically, statistics – but put it to much more productive purposes. There’s hope for economics yet.

Read more >>