Postrel on Dynamic Capabilities
| Peter Klein |
Former guest blogger Steve Postrel weighs in on the future of the dynamic capabilities approach (reprinted, with permission, from a thread on Academia.edu). Steve responds to the question, “Is the dynamic capabilities approach outdated?” with some typical insightful remarks.
Since DC is primarily an ex post facto construct measured by sampling on the dependent variable — i.e., if the firm successfully adapts, then it had DC — its prominence is not a sign that it is doing much intellectual work. . . .
[T]o a first approximation, arguments for the importance of DC have tended to be of the form “We know a priori that firms need to be able to change their operational capabilities from time to time; we have examples of successful firms that have adapted in this way and examples of less-successful firms that haven’t; therefore we can say that the successful adapters had more of this valuable thing we will call ‘dynamic capability.'”
Certainly there have been empirical papers that do better than that, by, for example, trying to look at firms that have adapted multiple times, or by identifying specific organizational structures and practices that might enhance adaptability. The difficult issue with looking at a “precursor” like experience is that theoretically experience could reduce DC by causing specialization and lock-in. Other putative precursors suffer from the ex post measurement problem — how do we know if a firm has the right knowledge for adaptation until we see whether it succeeds?
I suspect there are also deeper conceptual problems because DC is equivocal even with perfect measurement. It would be pretty hard to specify what one meant by the “amount” of DC a firm has or to compare the “amounts” that any two firms have. DC is certainly not a completely ordering relation and I’m not sure it’s even a partial order. Without presenting formal models and going back and forth between those and peoples’ intuition about what DC is “supposed” to mean, however, one really can’t pin these problems down enough to tell if they are serious. . . .
[This points] at conceptual, not empirical, problems with DC. The first question is what, if any, useful work does the DC concept do for us in thinking about strategy problems? Occam’s razor (or just a desire to reduce intellectual clutter) ought to make us seek parsimony in our theories — management theory is overrun with so many fuzzy overlapping concepts that we should greet critical arguments with relief more than trepidation. . . .
The second sort of question to raise is whether DC is an ordering relation, even conceptually, and if so, of what kind (preserved under monotonic transformations, affine, linear, or what?). [Another commentator] mentioned “flexibility” as a synonym for DC, which gives a useful entry into the ambiguity of the concept. A system can be flexible in the sense of having a wide repertoire of responses or in the sense of being able to shift easily among those responses that are in its repertoire. These are two different things, and firms might be ranked differently on each criterion. Within either of these two criteria of flexibility, there are further distinctions that could also reverse the ranking of two firms; for example, one firm might be very good at transitioning between response A and response B but poor at all other transitions, whereas another might be moderately adept at all transitions. Which has greater DC?
Note that these two questions represent independent critiques of the DC concept. The second one implicitly accepts that the first has been answered or sidestepped. . . .
Helfat (2007) certainly improved the situation. It is a matter of individual judgment how much it did relative to the two issues I mentioned above.