Evolutionary Theory of the Firm

Rewriting traditional economic theory?

Nelson and Winter realize that all models, including traditional ones, are abstractions, but...

For example...

Traditional Theory

Nelson and Winter characterize two main aspects of traditional economic theory: maximization and equilibrium.

maximization: firms are assumed to maximize, given their technology (possibly organizational) and input prices.

equilibrium: economic activity only takes place in states of the world in which each firm is doing at least as well as it can, given what everyone else is doing.


Unlike in real life, individuals are never confused, distracted, or make mistakes.

Evolutionary Theory

Forget maximization and equilibrium -- too rigid.

Firms are characterized by their routines.

Routines are regular and predictable patterns of the firm.

These cover: hiring and firing, each task in production, how workers communicate, how to respond to unexpected delays, deliberative procedures regarding planning and strategy, process through which firms search for and institute new routines, etc.

Firms' routines are analogous to individuals' skills. They are "deep channels in which behavior runs smoothly."

They are maintained by day-to-day use, and atrophy with lack of practice. Firms can maintain or lose ability to apply routines.

Routines depend much on tacit knowledge to coordinate. This means that you can't just blueprint them and transfer them elsewhere. Signals can be environment- or context-dependent. Firm- and individual-specific signals. Not always codified or codifiable.

[example: what happens when you get a new person on the job? Even if you tell them everything you know coordination may be disrupted.]

Routines are heritable (routines today look like routines yesterday) and selectable (firms with routines which are more profitable tend to win out).

Regarding selectablility, it may be that successful firms have routines that are robust to exogenous changes (they work well under a wide range of conditions -- so they need not optimize), or that successful firms have particularly good search routines so that they are able to reoptimize very well.

When placed in an unfamiliar situation, firms have the most trouble because their routines are ineffectual. Critical periods for firms are after large exogenous changes in environment. Invention of new routines are often required -- and soon!

Recall that there are no crises in traditional economic theory that firms can work out of. Firms are assumed to know exactly what to do when they lose a key employee. Or when there is a price increase for an important input.

Firms are not production functions, nor are they nexuses of contracts. They are characterized by their specific sets of routines.

Routines and Change

This makes economic change interesting.

Through search, firms adapt routines to new conditions (although not continuously).

Though they are guided by profit, they do not necessarily profit maximize. Thus they are never in equilibrium. External environment and conditions change; this is accounted for in firms' new search.

Organizational invention is explicit in evolutionarly theory, and the invention process is of interest.

Interpretations/Functions of Routines

Routines as Memory

Firms' knowledge resides in its routines. This is organizational memory. This knowledge stays with the firm even as individuals come and go.

Routines economize on information:

Nobody needs to know why routines are the way they are. They just need to perform. Easier goal than abstract "maximize."

Nobody needs to know others' role in the routine.

Coordination is accomplished via implicit and explicit messages. For routine to work, individuals must be able to receive/interpret cues.

"Teamwork": think of basketball team or jazz musicians. Improvisation within structure.

Routines as Truce

What incentives do individuals have to perform according to the routine?

A: Part of the routine dictates processes that are initiated when an individual's behavior diverges from routine excessively and causes disruption.

Example: At the economic consulting firm I worked at, one of the economists was often boorish and disruptive. Putting up with him was incorporated within the firm's existing routines -- so this behavior was tolerated within limits. Things proceeded as normal. However, there was a point at which he became too disruptive (demanded partnership) and was let go, even though he made a lot of money for the firm, and would have continued to make money for the firm if he were a partner.

Why? Accession would have required significant changes in the firm's routines to accomodate him. Large disruption of normal processes.

So...routines exist to deter conflict between individuals from manifesting itself in an excessively disruptive manner. Existing routines provide guidelines for good behavior, and hiring/firing routines provide enforcement mechanisms.

Routines discourage subgoal pursuit: who decides what and when decisions are made? Can deter individual managers from pursuing private goals to company's detriment. (As in GM 1920)

Conflict exists in profitable firms, but it is beneath the surface.

Dilemma: problem of breaking truce (and the associated costs of disrupting existing routines) can prevent firms from making "efficient" changes that they would otherwise make.

Routine as Target

"Just maximize" does not cut it as an objective for individuals within firm. Many individuals make choices and are uncertain about the extent to which they impact profits. "Just maximize" does not economize on information.

Routines become a target -- maintaining routine becomes an objective for individuals, particularly middle managers.

Routines become heuristic for maximization -- just keep things up and running.

For example, given that objective for training new member of the organization, training is complete when the existing routine is up and running...


To here, we have been applying theories and a framework relying on the notion that if we know the firm's technology, talents of the workers, nature of its assets, we can describe and explain its behavior.

This theory instead implies that knowledge of a firm's routines is necessary instead. Modeling the firm means modeling how these routines change over time.