Organizations and Markets

Macroscope

10 January 2007 at 4:22 pm spostrel 3 comments

| Steven Postrel |

Peter’s recent post about macroeconomics and macro vs. micro approaches in organization theory stirs up not-so-fond memories. From my first exposure to Samuelson’s version of the Keynesian synthesis, I found the whole thing off-putting. This feeling was a mixture of impatience with the aggregation assumptions and frustration with the seemingly backward causality of the Keynesian worldview, where production is a trivial matter and the only problem is making sure there is enough effective demand. The latter frustration was somewhat alleviated by learning about adaptive and then rational expectations, aggregate supply, sunspot and other coordination theories, and the host of microfoundation arguments that either got rid of the Keynesian assumptions or at least made some of them plausible. The impatience about aggregation never really went away.

I think macroeconomics sticks around because, despite some of its intellectual flaws, it benefits from some key features that tend to make macro theories in any domain more useful:

1) Measurable aggregate variables that we care about (e.g. inflation or savings).

2) Accounting identities for these aggregate variables that hold independent of the microstates of the system (e.g. savings equals investment).

3) Collective phenomena (correlated actions of the parts) that can only be observed at a macro level (e.g., business cycles).

4) Macro-to-micro causation, where it each micro part is affected by all the other parts, not just by one or two others (e.g., cutbacks in output due to reduced demand due to everyone else’s cutbacks in output).

5) Robustness of collective phenomena to many micro details (e.g., economy-wide unemployment level probably unaffected (to first order) by firms’ organization structures or by individual market structures.

Do organizational or strategic phenomena have any of these five features? We have 1) and 2) for business firms; accounting data approximate the economic variables we would like to measure. As far as 3) goes, a firm’s strategy or a strategic shift can only be defined at the level of the whole firm. Constructs like organizational culture are a little bit trickier — it’s begging the question (traditional meaning) to simply assume that culture exists independently of particular individuals’ beliefs, but it does seem that cultures persist in the face of personnel turnover. We might see 4) for some things, such as which equilibrium is picked in a coordination game (e.g. a weakest-link effort game), but not for all phenomena of interest. And 5) holds only on a case-by-case basis. For example, sometimes it seems reasonable to treat a department of an organization as a unitary actor, on the grounds that the details of its internal structure result in coherent department-level behavior as it interacts with other departments. At other times, a more-reductionist approach may be called for, as when we want to know how new human resource policies are likely to affect firm performance.

I’m one of those annoying (to Peter and Teppo) modeling pragmatists when it comes to disputes about reductionism. I feel that I can assess the above factors, as well as additional issues about tractability vs. degree of realism, on a case-by-case basis better than any predetermined methodology. I’d certainly be unhappy with a blanket rule mandating that we always use the most reductionist theory possible.

Entry filed under: Former Guest Bloggers, Methods/Methodology/Theory of Science.