Economists and Sociologists: Can’t We All Just Get Along?
| Peter Klein |
I haven’t blogged much on the Comparative Organizations conference hosted by Dave Whetten, Teppo Felin, and Brayden King. It was a terrific conference and I enjoyed myself very much but, as the lone economist in a group dominated by sociologists, I found the experience a little disorienting. Teppo, Brayden, and Gordon Smith — another non-sociologist participant-observer — have posted their reflections and, when Teppo sent this picture of Gordon and me (riding the chairlift at Sundance and no doubt engaged in deep, philosophical conversation), I remembered that I wanted to write something. So here goes.
1. Organizational economists and organizational sociologists are generally interested in the same phenomena. What are the characteristics and performance attributes of various forms of organization? How do social and market conditions, formal institutions, government policy, culture, and the like affect organizations? How do organizations change through time?
2. We differ profoundly, however, in how we try to answer these questions. Phenomena that sociologists and social psychologists take as explanans economists take as explanandum. For example, Anne Tsui gave a nice presentation on culture. There are two types of studies on culture, she explained, one type in which culture has a direct affect on organizational performance, and another type in which culture is modeled as having a moderating effect on other independent variables. I asked her later, what about studies with culture as the dependent variable, e.g., evolutionary game-theoretic models of the emergence of norms? These aren’t really discussed within social psychology, she said.
3. The participants largely agreed that reliance on cross-sectional, OLS regressions is a poor way to study effects of the institutional environment on organizations. Not because of endogeneity, the economist’s usual worry, but because the proxies are too crude and the complex interactions are not easily captured. (There was quite a lot of enthusiasm for small-n, in-depth, ethnographic investigations.) Fine, I asked at one point, but what about using instrumental variables, panel-data models with fixed effects, natural experiments and differences-in-differences, and other modern quantitative approaches? This didn’t seem to get anyone excited (though, as I recall, Omar of all people agreed with me that these deserved more attention).
4. Speaking of endogeneity, my sense is that economists tend to endogenize many variables sociologists treat as exogenous. One example is firm growth, or the progression from one stage to another. Gordon and I duked it out with Howard Aldrich and Art Stinchcombe in one breakout session on the “life stages” of the firm. I tried to argue — without much success, I think — that generalized life-stages models may give the entrepreneur and management team too small a role in deciding when to move to the next stage. An acorn doesn’t choose whether to become a tree but entrepreneurs do choose if and when to seek additional capital, to add partners, to take the firm public, to make acquisitions, and so on. Biological analogies, after all, must be used with caution. Aldrich and Stinchcombe agreed with me later, in private correspondence, that the modal trowth path varies tremendously from industry to industry. What, then, is the value-added of generalized life-stages models? Anyway, perhaps the problem is that sociologists (except for the rational-choice types) are uncomfortable with methodological individualism and want to give cohort and period effects a stronger role than we economists are comfortable with.
5. Here are three things the modal academic sociologist does not think very highly of: capitalism, economics, business schools.
6. Did I mention that Sundance is unbelievably beautiful in October?
7. Finally, despite being surrounded by these guys, I emerged not only unscathed, but also better educated, challenged, and encouraged.