Are Routines Necessary for an Evolutionary Theory of the Firm?

24 June 2006 at 9:53 am Nicolai Foss 5 comments

| Nicolai Foss |

The seminal and in many ways founding contribution to the evolutionary theory of the firm, and its numerous relatives in management, such as the knowledge-based , the competence , the capabilities, etc. views, is without much doubt Sidney Winter and Richard Nelson's An Evolutionary Theory of Economic Change from 1982.

The book is very heavily cited in management (e.g., it is among the top 10 cites in Strategic Management Journal), but has made less of an impact in economics (it was originally intended as an economics contribution rather than a contribution to management). The reason can be found in chapters 4 and 5 that develop notions of routines and organizational capabilities, and try to do so from the notion of individual skill. Since Nelson and Winter (1982) organizational routines have become a primitive in the definition of "higher-level" constructs, such as capabilities or "dynamic" capabilities.

In a number of recent papers (beginning with this paper, but mainly with Teppo Felin), I have identified and discussed various problems with the notion of routines. Very briefly, there are still no clean definitions of routines (which evidently makes it somewhat problematic to define capabilities etc. in terms of routines), the routines construct often implies a denial of methodological individualism, the empirical basis for asserting that routines are so strongly prevalent in real world firms that it is meaningful to think of firms in terms of routines is questionable, routines draw attention away from conscious, rational choice, etc. etc.

Given all these problems, one may ask why evolutionary economists and management scholars remain so infatuated with routines. Perhaps they just don't see the problems (or deny that there are the problems I claim). However,I believe the reason is that evolutionists believe that they must claim a firm-level counterpart to genes, and the relevant counterpart can only be found in features of organizations that are traits of the entity itself (i.e., the firm), transmit information/knowledge over time, and which are selectable. Some evolutionists talk about "organizational replicators."

However, it is far from clear to me whether the routines construct is strictly necessary to an evolutionary theory of the firm. What is needed for such a theory is sluggish, rather than perfect, adjustment. In biology, evolution simply means that gene frequencies change. Applied to domain of firms and industries, evolution means that firms have different rates of growth. However, there may be all sorts of reasons for differential growth that have nothing to do with routines. Firms can fall victim to political deadlocks among their various coalitions. Management may suffer from cognitive biases. There may be path-dependencies of all sorts. Etc. Such things are not easily placed under the rubric of routines. However, they seem to be just as, or more, real causes of sluggish adjustment as the presence of semi-mystical entities, called "routines."

Entry filed under: - Foss -, Management Theory, Strategic Management, Theory of the Firm.

New Issue of Industry and Innovation Does Creativity Harm Innovation?

5 Comments Add your own

1. brayden | 24 June 2006 at 10:34 am

I think path dependence in the management literature has everything to do with routines. Evolutionary org. theorists and others claim that routines bind organizations to certain paths they have taken in the past even when those paths no longer prove profitable. Path dependence may arise from other factors too, but I see it as first, and most importantly, grounded in the concept of routinization.
2. Peter Klein | 24 June 2006 at 3:14 pm

Isn't it important to distinguish between adjustment costs at the level of individuals (e.g., cognitive biases), at the level of groups of particular individuals (collective-action problems), and at some kind of broader, extra-individual level? Any of these can explain organizational path dependence. The evolutionary theorists seem to be seeking to explain the latter, though, as Nicolai says, the phenomenon to be explained is not very clearly articulated.

My problem is that it isn't clear how the evolutionary theorists' notion of routines is identifiable, in the econometric sense. For instance, if inertia is explained solely by individual cognitive biases, then replacing individuals with other, less biased, individuals should reduce the level of organizational inertia. Similarly, if collective-action problems are to blame, then changing group size or reward mechanisms or whatever should lead to changes in observed organizational behavior. What, then, would be a discriminating test of the sort of amorphous, extra-individualistic "routines" the evolutionary theorists are talking about?

One can imagine a test for the importance of, say, brand-name capital: replace the entire workforce with a new one while keeping the brand name and possibly the physical infrastructure the same; to the extent that firm behavior or performance is the same as before, then these outcomes are driven by the brand. With routines, however, this gets tricky, because routines are held to exist outside of individuals, but yet are embodied in those same individuals, such that replacing the entire workforce would also eliminate the routines. How then can we know if such routines exist?
3. brayden | 26 June 2006 at 10:53 am

I agree with the importance of the testability of hypotheses. I don’t think that this is as big a problem for examining the stickiness of routines as you think it is though. Lots of studies have been done that look at changes in organizational rules and other kinds of stickiness in organizations. Most of these studies provide evidence that rules have an extra-individual nature that makes them fairly persistent.

But independent of this, I think routines have an ethnographic importance that can only be understood by visiting an organization and observing how it works. One of the things that makes complex organizations different from other sorts of contractual settings is the existence of and adherence to routines, rules, and other structural elements. Robin Leidner’s book about the routinization of service work in McDonald’s is a great example of this. McDonalds is an organization with high employee turnover, yet they maintain consistent quality because of routines. Even the amount of ketchup splattered on a hamburger bun is routinized. Taking the routine out of the analysis and simply reducing McDonalds to a complex set of contractual elements would seem to de-organization the organization.
4. simon | 26 June 2006 at 10:05 pm

Very interesting question … Before defining a routine, I would like to briefly define activities that are employed by firms. An activity is a collection of behaviors either performed by an individual or a collection of individuals directed toward some end. The end is understood but not completely specifiable before it is achieved. Further, the behaviors themselves are known but their order and exact character are not specifiable a priori by the actors (cause and effect are ambiguous). That is, the actor or group relies upon an intuitive understanding to traverse the problem space in search of a solution. A routine on the other hand, is a collection actions performed by an actor or collection of actors where the end is clearly specifiable and the means of searching the problem space is known (cause and effect understood).

As a consultant I have observed many firms perform activities they could not reproduce systematically. The group was able to succeed once but they did not understand how they navigated the problem space and thus were unable to reproduce the outcome in a subsequent attempt (e.g., introduce a new product). Likewise, I have observed firms that can successfully launch new products because they have made explicit the key activities and their inter-relationships (e.g., design a new product). McDonalds is a great example. They are not promised success just because they have routines (e.g., McLean).

Now I come to capabilities. Capabilities are routines that possess two characteristics. The first is the firm understands both cause and effect (within the routine) and can subordinate the routine to organizational objectives. Second the routine provides a material competitive advantage. While the second is not strictly necessary, it does aid in reducing the set of capabilities to a more manageable number.
5. organizational routines « orgtheory.net | 14 November 2007 at 11:43 pm

[…] piece gives you the most succinct summary of the issue, here’s some more on choice, and here’s Nicolai discussing the same issue last year at […]