In Which I Rise to the Bait

16 January 2007 at 5:53 pm spostrel 6 comments

| Steven Postrel |

So I’m browsing some old posts on orgtheory.net, and I see a little discussion about whether organization theory has done much since 2000, and it mentions neo-institutionalism. And since neo-institutionalism kind of gives me the hives, I toss in a little comment about how what little I know about it seems to have been falsified by an empirical paper by Kraatz and Zajac.

What do I get for my troubles? Not exactly a gauntlet to the face, but since their post has such tempting flaws, I’ll bite.

First of all, the liberal arts college isn’t just some random example pulled from the ether. The original DiMaggio and Powell paper used universities as its paradigm of constricting institutional fields. If the iron cage is ever going to work in a straightforward way, it would have to do it here. It doesn’t.

Second, Kraatz and Zajac showed a lot more than Fabio claims. a) While the iron cagers argued that innovations and adaptations would have to come from the highest status organizations in the institutional field (e.g. Harvard or Swarthmore), actually it was liberal arts colleges low in the food chain who innovated earlier. b) The iron cagers argued that firms that violated norms would be punished severely; Kraatz and Zajac showed that the colleges that moved more aggressively to add business curricula did better than those who didn’t. c) Adding business majors at liberal arts colleges was extremely counter-normative. The ideological hostility among faculty and administrators was palpable. And yet the innovators were not punished and gained ground on the traditionalists.

Third, the competing hypothesis, that organizations react to technology and demand shocks as in economic theory fits the data simply and effectively. As demand for pre-professional degrees grew and demand for liberal arts degrees shrank, there was an industry supply shift toward more of the former and less of the latter. This appears to have worked even though these are non-profit organizations, suggesting that, as Alchian and Friedman argued long ago, competitive survival pressure makes even organizations not inclined to profit maximization display comparative statics behavior much like neoclassical firms.

Fourth, Fabio’s attempt to save institutionalism by suggesting that there are multiple fields and you can switch among them whenever you want drains the theory of predictive and explanatory power. So if Ford gets into cash flow problems it can just declare itself a church and start taking up collections? If a hospital is killing too many of its patients it can declare itself a mortuary? Okay, those men have more than a little straw in them, but still — saying that there are different resource streams for schools to tap is just saying that there are multiple sources of demand. It has no institutional content. The whole point of norms and identity is that they have fixed boundaries which penalize straddlers. Letting organizations mix and match and flit from one category to another transmutes the cage from iron to mercury.

Fifth, the multiple-fields immunizing stratagem doesn’t even apply to the liberal arts schools. None of them have jumped out of the liberal-arts field and become “multiversities.” They all see themselves and advertise themselves in traditional terms. Small classes, teaching emphasis, leafy campus, nurturing atmosphere, chance to find yourself, yada, yada, yada. Many of them don’t even have graduate programs, and they get no research grants. So the “multiversity” category is a long way off for these organizations.

I submit that the Kraatz and Zajac paper is about as clean a refutation of a sociological theory as you could imagine. If that doesn’t do it. what could possibly suffice?

Let me stress that I don’t think that all of the mechanisms identified by neo-institutional theory are chimerical. They may well operate. It’s just that they get swamped by strong technological and demand forces in an unregulated competitive environment. And it is the quantitative claim of DiMaggio and Powell — that norms, etc., rather than demand and technology are the important forces — that has been falsified.

Entry filed under: Former Guest Bloggers, Management Theory.

StrangEconomics Coase and Hayek in Law School

6 Comments Add your own

1. brayden | 17 January 2007 at 10:37 am

This comment is cross-posted at orgtheory.

It’s interesting the Postrel frames Kraatz and Zajac (1996) as a refutation of institutional theory. Refutation is too strong a descriptor for my taste and is not one that the authors used in their paper. If we refuted every org. theory for which we found lack of evidence in some study or another, we’d have no theories left. Rather than refute institutional theory, the findings of their study suggest some serious limitations to institutional theory. These limitations were, at least in my reading of Meyer and Rowan (1977) and DiMaggio and Powell (1983), previously acknowledged, although never sufficiently tested. Institutional isomorphism or decoupling are not processes that we should observe in every established industry or institutionalized setting. In fact, D&P are pretty clear in stating that mimetic isomorphism will most likely occur when the outcomes of organizational practices are ambiguous, difficult to measure, or uncertainty is so great that mimicking becomes a cost-effective shortcut for dealing with contingencies in the environment. Thus, in situations where organizations can clearly measure the effectiveness of a change in policy or practice, decision-makers are much more likely to ignore institutional pressures for conformity. In fact, being blindly conforming when there is mounting evidence that this is an irrational thing to do will likely lead to a loss of legitimacy! Who will see an organization as credible that knowingly violates the central premise of its organizing logic?

Meyer and Rowan similarly argue that when efficiency criteria exist that allow outsiders to actively assess an organization’s activities, the organization is less likely to adopt practices or policies symbolically. In sum, institutional theory was meant from the beginning to be a theory about how organizations respond to their environment when uncertainty or ambiguity impedes rational assessment.

The cleverness of the Kraatz and Zajac article is to actually take the limitations of institutional theory seriously, rather than just pretend that institutional theory can explain everything. We need more studies such as this. In a persuasive recent AMR article, for example, Lieberman and Asada suggested other limitations to standard institutional isomorphism arguments that are in need of testing. Someone needs to do it. The problem is that org. scholars are much more likely to try to push a theory past its limits than to work within them. Ezra Zuckerman wrote in a comment responding to an ASR article by Zajac (no less) and Westphal in 2004, “sociologists tend to be self-congratulatory rather than self-critical and to attack economic research without seriously engaging it…The key is to realize that [institutional or social constructionist] powers are limited in characteristic ways….the recognitions of such limits raises interesting questions that could truly advance neoinstitutional theory” (pp. 463-464). This is the real challenge for institutional theory. Only by recognizing and theorizing more precisely the limitations of institutional theory can we truly harness the predictive power that it has. But at the same time, to argue that one study flatly refutes a theory is just as wrongheaded. No theory has the power to explain everything, but institutional theorists have provided lots of evidence suggesting that institutional theory has much to say about certain kinds of organizational processes and behavior. Ignoring all of the contributions because the theory itself has some inherent limitations seems a little silly to me.
2. spostrel | 17 January 2007 at 11:54 am

The iron cage was supposed to apply specifically to colleges and universities in the United States. That was the lead example. If we now say that the degree of ambiguity and uncertainty in THAT setting is too weak for mimetic isomorphism to be a dominant force, then it is fair to assume that it is not that important in more-competitive contexts. The mechanisms of institutional pressure may exist, and they may be measurable as second-order or third-order effects, but the bold claims made for the theory should be abandoned.

The focus of my interest is for-profit businesses competing in markets with little or no “economic” regulation (e.g. entry restrictions or price controls). While these businesses often face huge uncertainty in the relationship between what they do and how they perform, but there is little ambiguity about their ultimate results. So by the standard propounded in the above comment, we should mostly ignore institutional theory in this domain. But that’s not what I see happening, even by Zajac.

Note that many of the ideas expressed in institutional theory overlap with fairly crisp explanations from economics. And the economic arguments don’t even require ambiguity of outcomes. For example, if there is high uncertainty in measuring an agent’s ability (because of a lot of randomness in going from actions to outcomes), competing agents may have an incentive to herd together in their actions, even if there is evidence that deviating from the herd would raise the expected value of organization performance.

I am willing to believe that herding is a strong factor, for example, in the behavior of football coaches, who have very unambiguous outcomes but high uncertainty in the mapping from what they do to who wins. Otherwise, we might see more variation in offensive and defensive systems in the NFL (e.g. the run and shoot or the spread). And Stein’s orignial argument applied to financial managers, who wouldn’t be blamed if they crashed and burned along with all their colleagues, but would be punished if they failed in a unique way. That seems reasonable to me, and perhaps measurable as a first-order effect (and, for all I know, may already have been done).
3. brayden | 17 January 2007 at 1:32 pm

But the uncertainty inherent in most for-profit markets meets the conditions of institutional theoretical analysis. As long as its not easy to assess the outcomes of any particular policy or strategic change, managers will seek confirmation of their activities by seeking “social proof”. Over time, the propensity of herding behavior (as you describe it) may cause certain features, independent of their contribution to the technical operation of the organization; but given enough contrary evidence, the same practice may lose its legitimacy. The point of institutional theory is not necessarily to show that policies and practices endure when they demonstrably lead to inefficient outcomes; it is that policies and practices endure when it’s difficult to assess what kind of effect they have on the organization. And this kind of uncertainty is ever-present in many for-profit markets.

Yes, there may be some markets in which institutional theory applies less (e.g. financial markets where the wisdom of the crowds crowds out legitimacy as a driving force), but in many for-profit markets, managers and other corporate agents are still very uncertain about how to create the kinds of changes that would lead to efficient outcomes.
4. spostrel | 17 January 2007 at 4:32 pm

1. How can you claim that the competitive pressures on liberal arts colleges are too great for the iron cage to be applicable (the attempted defense to Kraatz and Zajac’s findings), but competitive pressures on for-profit businesses are too weak to rule it out? That seems inconsistent on its face.

2. Herding behavior in games is not institutionalism. It is an opitmizing response by individuals in a game with asymmetric information. Legitimacy has notihing to do with it. The two theories are competitors, not complements. There’s a big difference between ambiguous final outcomes (the supposed condition of applicability for the iron cage) and noisy causality with clear final outcomes. Blurring them together doesn’t help. The Iron cage theory relies on the former, not the latter, and would certainly have trouble winning any kind of shootout with alternatives (such as rational herding) for explaining the latter.
.
5. brayden | 17 January 2007 at 4:52 pm

I didn’t say the competitive pressures were too great on liberal arts colleges. I argued that, in this case, rational criteria existed that allowed college administrators to detect that significant benefits could be gained by adding professional programs. When no such criteria exist or when administrators can’t agree on what those criteria are, you would expect legitimacy concerns to dominate.
6. spostrel | 17 January 2007 at 6:21 pm

Now we’re in circular land. Back to the first, third, fourth, and fifth points of my post.

It was Fabio who claimed that the lack of competitive pressure was a key enabler for the iron cage. Sorry if I cross-attributed.