Sunday, October 6, 2013

Top young economists

Believing that the next generation of economists will be for more rigorous and honest about their research keeps me happy. Today I’m not so happy. 

Here’s a list I came across last year, which recently came across my screen again. It compiles the views of 8 top young economists on where economics is going. And it’s bloody frustrating.

It’s frustrating because the more I read of economic research these days, the more it seems to neatly fall into either ‘the optimal control’ game, the ‘label the residual game, or the “correlation game”. And these young guns are simply rehashing these same games.

The ‘optimal control game’ is about mathematising an idea so that it fits into the structure of the calculus of optimal control. Often it doesn’t matter how you squeeze your idea into this mould; simply pluck a term out of thin air, call it optimism, certainty, talent, or some such unmeasurable or knowable thing and stick it in an equation. You then you solve the equations, finding the optimal point, which you can compare to the solutions to the equations when you ‘shock the model’ by changing a parameter - maybe a magical bout of increased optimism.

This method tells us very little about anything, especially when the terms are often vague unmeasurable concepts, and when there is not clue about how the optimal point can be obtained in reality if you don’t actually start there. 

‘Label the residual’ is the game of writing a model that has a term which captures all the unmeasured variation from the other terms. Noah Smith described it here

The ‘correlation game’ is about mining data for relationships that suit your audience. 

There is no prediction game, despite this being the true test of a theory. 

There is also no reward for exploring new methods of analysis, and much criticism when it is tried. And to my genuine concern, outcomes of the unscientific games being played are taken seriously not only within the profession, but often in the political arena.

With such self-inflicted constraints the discipline struggles to adequately answer even the most basic questions it is tasked with - why are some people and countries wealthier than others, why is wealth distributed the way it is, what is money, why is there a business cycle, and more. 

So it is of interest that we examine what the insiders, the next leaders of the ‘tribes of econ’, see as the direction the field is heading. Are they ready to break free from these silly games, testing new models of cooperation and interaction, bringing dynamic mathematical tools to the economics toolbox, and gaining qualitative input from the real decision-makers in the sectors they analyse? Will they incorporate political elements, experimental regularities, and the mechanics of real institutions into their research? 

I wish I could be so optimistic. 

* I disclose that I don’t personally know any of these economists and haven’t read their work for this article (I lied, I have poked around their work a little). 

Nicholas Bloom 
Why are developing countries poor? It’s a good question and a fundamental one to economics for a long time. But poor Bloom seems to be playing a new version of the ‘label the residual game’. He says

I think the answer is complex and linked to a combination of factors around history, geography, luck, etc. I am personally working on management practices: people in developing countries are poor because wages are low, and wages are low because firms are very unproductive, and firms seem to be unproductive in large part because of bad management.

The one thing I give him credit for is that he is involved in randomised experiments, avoiding falling into the “correlation game”. But do we believe that poor countries are poor mainly because there are bit disorganised? Doesn’t that beg the question of why they are disorganised? Or is it actually about uncertainty fairy?

Ray Chetty
How can we increase the rate of economic growth and overall well-being, and how can we reduce the rate of poverty? Good questions. Not new at all, but Chetty seems not overly certain about where to look, naming just about every policy there is as a potential mechanism for growth. This seems like a very broad program of labelling the residual.

Gauti Eggertsson 
How to use monetary and fiscal policy to eliminate unemployment and control inflation? These fundamental questions of economics are still on the table after centuries of research and policy experimentation. Why is that?

We seem to know that credit creation is the real inflation driver, and we seem pretty sure it is closely related to capital investment and employment. But if questioning banking and money remains taboo in economics, I can’t see change happening fast, especially when the top young guns seem to think introducing financial frictions into models is the way forward. Back to the 'optimal control game' then.

Xavier Gabaix
How to model realistic economic agents? Now there’s a better question. But Gabaix seems very certain that some form of bounded rationality is the answer here, which is of course to say, that no one wants to drop their foundation methods, prefer to play a more sophisticated version of the ‘optimal control game'.
Gita Gopinath
How does one accomplish sustainable growth without large boom-bust cycles? Again, another classic problem of economics still waiting for a decent answer. Gopinath at least raises some important issues routinely ignored in economic research - global imbalances, currency wars, capital controls etc. But again, we here about “understanding the propagation of shocks across economies”, which implies we are playing the “optimal control game” once again and don’t really want to think about the mechanics of economic interactions, investment and so forth. 

Peter Leeson
What is the status of the rationality postulate?

If we view economics as an “engine” for understanding the world, the rationality postulate was that engine in nearly all of economics until quite recently. The rise of behavioral economics has challenged the usefulness and, in a more subtle but radical way, the legitimacy of the rationality engine.

A very vague and non committed response by Leeson. He suggests that beliefs about rationality drive economists views and their interpretations of events. But still, what sort of solution to rationality’s status would Leeson be happy with? 

Glen Weyl
Glen seems to think that the rise of digital information, and the outsourcing of decisions we are making to automated computer systems, is threatening to overwhelm Hayek’s argument that “free markets were necessary in order to allow the incorporation of information held by dispersed individuals into social decisions”.

Which is strange, because every economic model of markets requires that all the information is available to every decision maker before any trade occurs. AS I have said before “if prices can convey all information, then there must also exist an alternative non-price method to convey that information which governments could use for an alternative allocation system.”

Justin Wolfers 
What will economists do with more data? Well, Justin thinks economists will continue expanding their social science empire in the spirit of Gary Becker. Of course, any statistician will tell you that uncontrolled data alone can’t generate any useful scientific findings.  “Economic theory will become the tool we use to structure our investigation of the data”. I hope not. Since economic theory doesn’t usually tell us what to expect in the data at large, given the general failure to consider problems of aggregation.

This is really a sad story. The econ tribes wield so much political influence, yet have so little to offer.  These same questions could have been asked 30 years ago and would have seemed cutting edge.  So much has happened, in the world, and very little it seems in economics [1].

If I was to answer I would say the big question for economics is finding new tools that are able to incorporate dynamic elements and evolving strategies of cooperation, addressing all the aggregation problems that arise from the 'optimal control game'.  I would say that on the policy side the big questions are about the mechanics of real institutions, for example the banking system and credit creation, and of the political games that arise when new institutions are formed.  For me the future of economics includes power, rents, and conflict. 

But today I doubt we will see the future I imagine. 

[1] Many researchers will deny that little has changed.  We do know that the economics community is more empirical now than ever. But without theory, uncontrolled data can’t reveal much of scientific value.  And theory seems stuck in a time warp. 

1 comment:

  1. Hi, I found the list too to be a bit disappointing because they seemed to be built around common solution steps rather than profound questions... the type that turn into a math model or correlation study that may be right but doesn't seem to matter.