Macroeconomics: The scientist and the engineer

Over at Econlog, Arnold King is telling us how he lost his macro religion. It is an interesting post, and it gives me the impression that he believes any type of technical macroeconomics (either empirical or theoretical) to be somewhat of a fraud – a fair description given the difficulty of stringing out cause and effect in macroeconomic data.

In the post he discusses a essay by Greg Mankiw called the Macroeconomist as scientist and engineer. Far from presuming that macroeconomists do the same thing as scientists and engineers, the point of the essay is to describe the difference between macroeconomists that thrive in the abstract and those that work in the face of policy and data. Although there is not a clear split between the two – the distinction allows him to discuss how many of the recent (last 20 years) conclusions in macroeconomics do not appear to be useful for forming policy.

The essay by Mankiw is very good, and it provides a much better description of the evolution of macroeconomics ideas then my earlier retort to Chris Trotters claim about Keynesianism. However, I would also beware the partisan element of what he states.

He admits that he is strongly involved with the New Keynesian school, but the bias does not lie there. Fundamentally, he does not seem to realise that he also seems to play down the importance of the scientist in his piece. This is interesting, given that he is an eminent professor, and is more than capable when playing the “scientist” role in macro-economics.

I think the “scientist” does have a substantial role to play. Sure, part of the reason that the macro-economic policy recommendations have not evolved inline with theory is because the discipline has had trouble translating them over to policy. However, there is an important difference between now and Keynes time, when his policies were so quickly implemented – macroeconomics exists, and the old ways of doing things are entrenched in the policy making institutions. The higher technical requirements of the new found science in economics, combined with a relative reluctance by many people in policy to move over must also take some of the blame.

Ultimately, it may take a substantial economic shock to shake applied economists into updating their models – could the current credit and energy shocks be sufficient? I am sure that they will at least force central banks to pay more credence to the supply side than they currently do.

Earlier links to the paper are found:

  1. Initial link,
  2. Poets and plumbers?,
  3. A letter from a reader – the guy mentions Don Brash 😉 ,
  4. Arnold’s initial reaction.
7 replies
  1. Matt
    Matt says:

    I read the essay.

    On the issue of price rigidity. Any firm, like Apple, plans an execution of transactions in its market over time. Its effort yields a price for the longer term value of the company. The price is estimated by industry accountants, who keep both their estimate of the fortunes in the software market, plus an estimate of the error of their estimate.

    So, the resultant rate and value of market transactions yield a most likelihood estimate of discounted firm value. We know much from linear estimation theory, like the bounds on price volatility, over time, to yield an estimate of the error.

    Regarding price rigidity, the firm has a very fixed idea of the rate*value of transactions it is willing to risk over time to yield a specific certainty in longer term firm value. He cannot go beyond that because his measurement error is still to great, and the next increment of transactions, taken again from linear estimation theory, would be N*logN rate of increase for each N of precision “bits”. You pay a power law price to get the next increment of market precision. So, the firm (and market) has to restructure to restore excess capital.

    The model is that things important in the economy are viewed by intelligent people through the lens of an ad hoc Kalman filter, decisions are optimal to reduce costs, and both the differential manifold and measurements (transactions) must converge.

  2. Matt Nolan
    Matt Nolan says:

    Hi Matt,

    I am not really sure what you are asking here. If you are just writing one way of describing price stickiness then that is cool 🙂

    It appears that you are stating that when a firm chooses prices over time and subject to uncertainty then in the face of any small shift in demand they are unlikely to actually change the price they set. This will only work in the sense of a unrequited change in relative prices (relative to the counter-factual) as long as there is some sort of transaction cost (as otherwise they would re-optimise whenever they face a statistically significant change in the distribution of their error term). This is where “menu costs” come in for the New Keynesian literature.

    It is important to note that, empirically, menu costs haven’t been found to be significant in many industries (namely supermarkets) as a result, the nominal price stickiness is often confined to specific industries, which in turn makes the relative price issue a difficult one to conceptualise in an aggregate model. Who knows, this is where true general equilibrium theory may make a come-back 🙂

  3. Kimble
    Kimble says:

    Off topic thread is it?

    A post about ticket scalping would be entertaining. I have had several hearty arguments with coworkers explaining that scalping is natural and the onus of “protecting the fans” rests with the sellers of the tickets.

  4. Matt Nolan
    Matt Nolan says:

    “A post about ticket scalping would be entertaining.”

    Indeed, I have been sorely tempted to do a post on ticket scalping for a while – I’m sure I have a bunch of links I sent myself in an email somewhere. I’ll definitely do it at some point – but it probably won’t be terribly exciting 😛

    Want to do a post on barter and first degree price discrimination at some point – discussing why barter may be more “efficient” until the opportunity cost of time gets to high. I’ll try to do that this week. I’ve got two posts scheduled tomorrow anyways so I’ll probably think about it tomorrow night 😛

Trackbacks & Pingbacks

  1. […] Fundamentally, a better characterisation (which more fairly divides up the discipline) was provided by Mankiw, comparing the groups to scientists and engineers (we have discussed this here). […]

Comments are closed.