The problem of macroeconomics

I’ve been reading an article by Joseph Stiglitz whose title made it sound very promising: Rethinking Macroeconomics: What Failed And How To Repair It. His work on how information asymmetries and uncertainty shape institutions and markets make Stiglitz a strong candidate for setting out a new approach to macroeconomics. And it’s needed. Standard macro has failed, even though many macroeconomists who have invested so much in their human capital are understandably reluctant to agree. Even with modifications such as heterogeneous agents (ie people who are not all identical) and, the standard Dynamic Stochastic General Equilibrium model is not an empirically or theoretically useful representation of an economy.

I was disappointed with the Stiglitz article, however. It’s powerful as a critique, weaker on what to put in place of the standard model, although Stiglitz rightly identifies the institutional reality of money and banking, and credit markets, as essential missing elements. (Non-economists may be surprised to learn that banking does not feature as such in conventional macro models, which are indeed highly abstract).

An alternative tack which some of the most brilliant economists around think is most promising is to look at the complexity models used by physicists and biologists. Paul Ormerod’s is a great introduction to this approach, and I’m about to start Alan Kirman’s Andy Haldane at the Bank of England and Robert May have written in Nature about this modelling strategy as well. I’ve just been having another look at John McMillan’s . It’s a terrific book which looks at the institutions and detailed operations of many specific markets. He too concludes that the complexity approach is the right one at the macro scale, adding:

“The economic system has an additional layer of complexity: its components are people, who react intelligently to it and even reshape it.”

McMillan very sensibly notes that: “[Markets] have impressive achievements; they can also work badly. Whether any particular market works well or not depends on its design.”

Economics has a rapidly improving understanding of market design principles at the sub-macro scale. But this still leaves a huge modelling gap at the aggregate level. Until some future Nobel winner comes up with an alternative workhorse macro model that can be described readily in academic papers and taught to students, whether it is a complexity model or not, we will be stuck with the standard DSGE approach. Nor is it clear to me at what level of aggregation one has to switch from the tools of the market design approach to the tools of analysing complex systems. That is the problem of macroeconomics: not only how should we do macroeconomics, but also where does the boundary between micro and macro lie?

[amazon_image id=”0393323714″ link=”true” target=”_blank” size=”medium” ]Reinventing the Bazaar: A Natural History of Markets[/amazon_image]


6 thoughts on “The problem of macroeconomics

  1. As you note, complexity ideas have a long way to go before they gain enough presence in economics academia to have an impact in the real world.

    Still, I feel that they represent the best hope of the moment. Uniting micro and macro is beyond us for now (as in Physics). It will come one day, perhaps, but no time soon.

    The danger of course is that complexity is a subtle thing – it’s even more easily spun than the current crop of theories. I’d ask you to be very wary of Paul Ormerod in this regard, he seems much more interested in retro-fitting complexity onto his existing economic beliefs than pursuing the evidence where it goes…

    • It’s obviously a bit of a fashion now, so no doubt in for some unsubtle treatments. I’m looking forward to reading Alan Kirman’s book. The thing about Paul Ormerod is that although he has views (as so many of us do…), he is an economist writing intelligibly about this. I haven’t found many others.

  2. I´m following Stiglitz for quite some time and I think that he is a brilliant thinker. But as you pointed out, he is much better as a critique of the current state than big innovator (typical for his concept of Privatizing Profits, Socializing Losses)
    Since I´m a political scientist, I prefer when economy is more related to what it really is – to humanities. As we have seen with in our crisis, the chances to find some working and reliable mathematic model are really low.

  3. Is there any reason to think that we *can* make good models ?
    I’m thinking about the events of the fall of 2008; however you want to characterize the run up, in the fall of 2008 you had sudden, unexpected things going on; in more technical parlance, highly nonlinear events with strong neg and pos feedback loops.
    Is there any reason to think that we could model such an event ?

    I would also like to know why it is important.
    What we care about (I think) are jobs, sex, food, taking care of our kids, etc.
    Most of that is not relevant to economics, so the first step is to be more humble about what econ can do.
    The second step is to focus on what econ can do (say, jobs, inflation) and do that explictily: we want 2% unemployment; the baseline scenario is that the gov’t hires people to dig and fill in ditches till we get to 2%; do you have a better solution

  4. sorry – one more thing
    There is , i think, an implicit assumption: economist are “honest” and want the general welfare to improve.
    Is there any empirical evidence for this 😕
    I’m not being rude, I’m just using your own tools: economist should , according to theory, maximize their utility, which means, as an academic or writer, sucking up to the rich (I could use fancier, less inlfammatory bowdelerized language )
    I mean, the idea that taxing cap gains at a lower rate is “good” – wtf ?
    I always like to say, the day china starts exxporting PhD economists who can speak english and are willing to work for 20K a year, is the day the econ prof suddenly discovers deep theoretical reasons why trade is a bad idea….

    • We might yet find out about the cheap overseas PhD economists…. However, to be fair to my profession, most would not say a lower rate of capital gains tax is good in itself. It’s an empirical matter as with any tax – depends on the incidence of the tax, its effectiveness at raising revenue, its effect on future investment & growth (if any), ease of avoidance and admin cost etc etc. The cap gains question becomes politically polarized – but see the Mirrlees Report and other work from Institute for Fiscal Studies for sober empirical approach.

Comments are closed.