In my initial jottings on the paper, I wrote the following:
Seems to me that there are three (well, 2 1/2) fundamental assumptions, but don’t know that I agree with them: 1) that micro-foundation means that the level of analysis should be the individual, and that individual behavior aggregates up to organizational outcomes; 2) that the aim is ultimately to explain macro-outcomes – so routines as DV is helpful only insofar as it gets us to routines as IV and macro-outcomes as DV; 2.5) this outcome is really not ‘macro-outcome’ but performance in the interest of competitive advantage.

I guess I’m saying that I flagged the MI as part of the issue as well. I also think there is more than performance at stake, though this is the main outcome variable of interest in Strategic Management (incidentally, this is a key distinction between SM and Sociology more broadly).

When I see routines, I don’t see habit versus decisions, or routines vs. capabilities (broadly) in your way of putting it. I see routines as something more like crystallized actions – Berger and Luckmann would say sedimented action.

That said, many really good analyses have played off of this distinction, to the benefit of understanding culture, for example. Ann Swidler wrote about culture’s use during ‘settled’ and ‘unsettled’ moments. And Barley’s work on routines-as-scripts, which are altered by technologies and then potentially rebuilt to accommodate or change relations of authority in workplaces, this is another pivot on the same idea.

My initial bit of snark maybe was about the fact that you seemed to engage with a pretty narrow set of needs for understanding routines (must have micro-foundation in order to build macro-foundation) – this is a metaphor, theory is not really a house. And second, there seemed to be a primacy given to micro-foundation as being equivalent to individual motivation and action, which I still find not totally helpful.

Of course, I take your point that I should actually have engaged with your argument as made and actually have followed up on the papers before being so dismissive.

There is a really nice and poignant story of organizational routines (not practices, mind you, routines) in Michael Lewis’ experience in a children’s ward of a Berkeley hospital. I guess I don’t know what you are aiming at (i.e., the ‘real progress’ you attribute to TCE).

My other impulse, perhaps not surprisingly, is to say that with respect to routines changing (point 4, maybe 5 as well), individual actors are overrated. Routines change because of shifts in legislative environments (make it illegal to ask about whether a job candidate has kids/is married and you’ll change at least to some degree who gets hired); they shift because of individual activism; they change when changes in technology, broadly speaking, make possible new routines or obsolete old ones (Yuval Millo tells interesting stories about how routines around margin requirements in clearing houses shifted with incorporation of language of Black-Scholes-Merton).

So I guess I’m a little puzzled at the lack of utility you’re talking about. Is it the lack of a tightly cohesive theoretical agenda? Lack of agreement about outcome variables? That there is heterogeneity in empirical object and approach? All of the above?

Arguments about policy and law turn on whether they support or undermine a “predictable environment” for decision making. Critiques of decisions sometime put a great deal of weight on that law not getting ahead of the political process. The formation of common law is essentially a process of distilling and codifying current (legal) practice in resolving disputes. Etc.

To be sure, rules can establish a new equilibrium by establishing a stable coordination point (in Schelling’s sense), or in some cases can move a situation from one equilibrium to another. So I agree that simply saying rules codify existing practices is too limited — they can form the context for new or altered practices.

Expectations are not inertial, because they coordinate action, and action is the opposite of inert. They are constantly violated by error, testing, seeking advantage, etc. and must be constantly rebuilt. Their relative stability is evidence of a strong equilibrium in the interactions.

Even when expectations are stable and not threatened, publicly codifying the expectations can greatly reduce coordination costs, because rules can make it much easier to maintain the equilibrium.

Conversely, however, if an organization tries to establish and enforce rules far from any self-sustaining equilibrium, the effort will be very costly and unsuccessful.

“an event history of rules matters” should be ended with a period. the word ‘matters’ is a verb.

“rather a reflection” should be “rather THAN a reflection”, i.e., not a reflection of an internal status.

taking all that into account, should one characterize rules as really codifying expectations, when the expectations being codified are
ones threatened by change and unlikely to be stable when rules emerge? whether a rule represents an expectation long after it has emerged seems to me accidental or just a deflection of later external threats, rather a reflection. even a restatement of a rule may be a timely action, not a stable description.

my critique doesn’t apply to rules as teachings. but i’d decouple a teaching from a rule.

I find most treatments (including Bourdieu, John Seely Brown, etc.) suggestive but lacking in the analytic tools to get a grip on the phenomena and move us toward something as strong as the Coasian framework.

H. Peyton Young has analytic tools that are applicable here, though I’m not sure if they are quite strong enough. Of course a huge amount of work is required to build an adequate bridge between the qualitative accounts (Bourdieu etc.) and the more analytic models.

However the basic concept required for good analysis is easily stated. People act based on their expectations of others actions. Over time, people converge on expectations that are relatively stable, given everyone else’s expectations. These relatively stable and self-sustaining expectations ground routines, practices, dynamic capabilities, etc.

Rules are typically a codification of the expectations in given cases. They may be useful in guiding analysis, and may play a constitutive role in shaping behavior, but they always have to be understood in terms a much richer ground of expectations about interpretation, enforcement, etc. An approach based on rules that are literal and “self-implementing” is absurd.

Note that this approach works fine with both symmetric and asymmetric roles, and even different populations with radically different expectations or “models”. It handles convergence on new practices, stability, drift, and the collapse of practices. I think it lets us develop analytic accounts for all of Bourdieu’s ideas. Etc.

I’ve written more about this a long time ago in the context of Turner’s critique of “practice talk” at my blog. This was before I discovered Young.