Annotations re An Evolutionary Theory of Economic Change

Nelson, R.R. . Winter, S.G . An Evolutionary Theory of Economic Change . The Belknap Press of Harvard University Press . Cambridge, Mass . 1982 . ISBN 0-674-27228-5

Preface: creative intelligence, technological or other, is autonomous and develops erratically. Its future cannot be known. The terms economic change and economic development are presupposed: what is their meaning? Later on: the unfolding of economic events over time.

Backdrop: doubts about disparities in economic development, long-term ecological viability, and the relation between economic success and fundamental human values, and the stresses related to each.


First premise: economic change is relevant and interesting. Second premise: existing economic theory must be reconstructed.

Firms are assumed to be profit oriented (searching for ways to improve it), but not profit maximizers over choice sets. Also this theory does not focus on an industry equilibrium where the non-profitable are driven out until the profitable ones are at their desired size. Firms are not modeled using maximization calculus but as sets of capabilities and decision rules. These change over time as a result of deliberate problem solving as well as because of random events. The market, through a economic equivalent of  natural selection, ferrets out the worst performers.

The section Introduction summarizes some complaints with regards to the general equilibrium theory in economics, including rational agents, information, competition, transaction costs, increasing returns &c. It is concluded that these critiques are confirmed by many economists, the authors feel they are members of a crowd.

The authors complain that there is an orthodoxy not just concerning the commitment to values and norms of scientific inquiry but also concerning commonalities in intellectual perspective and scientific approach, and even that there is a narrow set of criteria that serves as an acid test to establish if an expressed point of view is worthy of respect. They concede however, that it is temporary and ever-changing. The authors illustrate: ‘We should note, first of all, that the orthodoxy referred to represents a modern formalization and interpretation of the broader tradition of Western economic thought whose line of intellectual descent can be traced from Smith and Ricardo through Mill, Marshall, and Walras. Further, it is a theoretical orthodoxy, concerned directly with the methods of economic analysis and only indirectly with any specific questions of substance. It is centered in micro-economics although it is pervasive in the discipline’ [Nelson and Winter p 6].

The orthodox views and approaches have their own self-preserving connotative properties: ‘(concerning standard undergraduate textbooks prefiguring the advanced texts) The best of the texts are notably insistent on the scientific value of abstract concepts and formal theorizing, and offer few apologies for the strong simplifications and stark abstractions they employ. Neither do they devote much space to caveats concerning the theory’s predictive reliability in various circumstances. .. Many of the strong simplifying assumptions commonly employed – perfect information, two commodities, static equilibrium, and so on, are emphasized in such texts for reasons having to do with the perceived limitations of the students, and not because the discipline has nothing better to offer‘ [p 7]. And more specifically a drawback of the focus on equilibrium dynamics (and even more when statics DPB) is:’.. , but it is no caricature to remark that that continued reliance on equilibrium analysis, even in its more flexible forms, still leaves the discipline largely blind to phenomena associated with historical change‘ [p 8]. And with regards to the approach to assumed rationality of economic agents in intermediate and advanced textbooks: ‘As theoretical representations of the problems faced by economic actors increase in realistic complexity and recognition of uncertainty regarding values of variables, there is a matching increase in the feats of anticipation and calculation and in the clarity of the stakes imputed to those actors‘ [p8]

Malthus has made the connection between economics and biology, firstly natural selection: ‘Market environments provide a definition of success for business firms, and that definition is closely related to their ability to survive and to grow. Patterns of differential survival and growth in a population of firms can produce change in economic aggregates characterizing that population, even if the corresponding characteristics of individual firms are constant. Supporting our analytical analysis on this sort of evolution by natural selection is a view of ‘organizational genetics’ – the processes by which traits of organizations, including those traits underlying the ability to produce output and make profits, are transmitted through time‘ [p 9]. This hints at a kind of cultural evolution: all kinds of traits of firms are evolved, including those enabling it to make profits &c.

With regards to the teleological trap: ‘It is neither difficult nor implausible to develop models of firm behavior that interweave ‘blind’ and ‘deliberate’ processes. Indeed, in human problem solving itself, both elements are involved and difficult to disentangle‘ [p 11].

Evolutionary Models

The Structure of Orthodox Models

Common denominators of ‘orthodox theory’ are A) a maximization model: 1) something gets maximized 2) there is a set of things that the firm knows how to do 3) the firm’s action is the result of an actionable choice (given constraints and conditions). It attempts to explain the rules of the firm in this way. B) Equilibrium calculus leads to explanations of the choices hence behavior of firms in specific circumstances. The equilibrium provides another equation to the set in order to determine the value of some variable. The suspicion is that this approach is not aimed at the explanation of economic principles and dynamics but at the wish for results from the calculus. The choice and use of mathematical tools have influenced the thinking about economics.

The Structure of Evolutionary Models

And we see ‘decision rules’ as very close conceptual relatives of ‘production techniques’, whereas orthodoxy sees these things as very different. Our general terms for all regular and predictable behavioral patterns is ‘routine’. .. In our evolutionary theory, these routines play the role that genes play in biological evolutionary theory‘ [p 14]. Some important features are: persistency, determining the behavior of the firm, heritable, and selectable. I agree with this statement because it is very similar to the Rodin Testament, roughly: once an idea exists, it can be considered done. An idea can be anything. Not everything in business can be described with routines as many things are not routinely. This is accommodated by the idea that some decision rules as well as their outcomes are characterized by stochastic rules. I am not so happy with that expression, because I am always hoping for an explicit model without the use of probabilities. The authors explain that stochastic terms are required to cover for what is not predictable; had they been predictable then they would have been an element of the set of routines.

Although the routines that govern behavior at any particular time are, at that time, given data, the characteristics of prevailing routines may be understood by reference to the evolutionary process that has molded them. For the purposes of analyzing that process, we find it convenient to distinguish among three classes of routines‘ [p 16]: 1) activities that cannot be changed at a short notice because of its stock of capital 2) period-by-period augmentation or diminution of the firm’s capital stock 3) firms possess routines which operate to modify aspects of their operating characteristics. Why do we need this classification into three kinds of routines?

These routine-guided, routine-changing processes are modeled as ‘searches’ in the following sense. There will be a a characterization of a population of routine modifications or new routines that can be found by search‘ [p 18]. This is the thought that a population of ideas pre-exists; once they are searched and uttered then they can be actioned. How they search is described as: ‘A firm’s search policy will be characterized as determining the possibility the probability distribution of what will be found through search, as a function of a number of variables – for example, a firm’s R&D spending, which in turn may be a function of its size. Firms will be regarded as having certain criteria by which to evaluate proposed changes in routines: in virtually all our models the criterion will be anticipated profit. .. Our concept of search obviously is the counterpart of that of mutation in biological evolutionary theory‘ [p 18]. This requires the assumption that people are capable of anticipating a profit; that they are capable of interpreting the effects of their anticipations on their actions. It is also left implicit why people do that fundamentally: they look for alternative routines to make more profit but why would they essentially do that?

The models in this book are of industries: situations where similar firms interact in a market context characterized by demand and supply curves. To keep clear of short term dynamics (single price on a market in a single period) a temporary equilibrium is assumed. No long-run equilibrium is assumed. The firms have a current operating characteristic, and exist in exogenously determined demand and supply conditions. This configuration determines the profitability. Profitability operates through firm investment rules as a determinant of expansion and contraction off the size of the firm. Through search and selection do firm evolve over time, every state the basis of the next. Operating characteristics of the firm change because of the searches of thát firm. But this is a stochastic process: at every state (not one but) a distribution of possible next states is generated. This method is easily translated to Markov chains. However, the processes in Markov chains are not very by themselves close to economic models.

The Need for an Evolutionary Theory

1 The awkward treatment of economic change by orthodox theory

Economic analysis deals with change. Its adequacy should be measured against its ability to shed light on change in a firm or an industry as a result of exogenous change and its adequacy to clarify innovation. It does not or in an ad-hoc way. Firm and industry behavior after an external shock is not derivable from profit maximizing and equilibrium formalisms of orthodox theory. ‘The most  intellectually exciting question on our subject remains. Is it true that the pursuit of private interests produces not chaos but coherence  and if so, how is it done?‘ [ Hahn 1970, pp. 11-12 on p 26].

Little progress was made (at he time) to theorize economic change because of the impediments of maximization and equilibrium. Note that those are two obstacles posed here, but others are: perfect markets and rational agents. And the same goes for innovation. In other words: orthodoxy operates better when change is absent. There has not been recognition of the central role and the autonomous character of technological development.

In the last half of the 20th century the central thought was that firms are profit maximizers and the economy (industries) are in a (moving) equilibrium.  Technical advance comes from profit oriented investment of firms. By the standard of thee orthodox models profit is now a disequilibrium phenomena, because profits stem from an advantage a firm has over the competition afforded by innovation. In addition innovation is hard to predict: different competitors make different bets and only ex-post is it clear who was wrong and who was right. Profit maximization and static equilibrium does not explain economic growth. This originates in the adherence to orthodox tradition: it had to turn to abstractions and move away from key features of capitalist dynamic such as uncertainty, transience of gains and losses and unpredictabilities of technological advance.

In this vein it is extremely difficult to develop a model for Schumpeterian competition out of the components of maximization and equilibrium. This development risks getting stuck in verbal theorizing instead of formalization. This gap between orthodoxy and Schumpeterian insights had (at the time) not been bridged: business processes of change and innovation were to be over-simplified and the diversity of firm characteristics and hence of their interactions with the industry were obscured.

Diagnosis an Prescription

Main differences and agreements with orthodoxy:

1) Agreement: agents (esp. firms) have objectives, usually profit; profit is the main criterion for their choices.  Difference: the choice is perfect and absolute and long term versus the choice is imperfect and as per the status quo at the time. ‘… Carrying out a new plan and acting according to a customary one are things as different as making a road and walking along it‘[ Schumpeter 1934 pp. 79, 85 in p. 31-2]. Change present problems that automaton maximizers are ill-equipped to solve and theories incorporating automaton maximizers are ill-equipped to analyze [paraphrased p 32].

2) Agreement: competition drives decision making in firms in their industries and so it is worthwhile to look for a state where the forces of competition would be in equilibrium and would no longer produce change. Difference: orthodoxy goes much further to introduce equilibrium at an early stage and narrowing down the possible states of the firm: ‘Such models do not explicate the competitive struggle itself, but only the structure of relations among the efficient survivors. .. This theoretical neglect of competitive process constitutes a sort of logical incompleteness, … It is only in equilibrium that the model of optimizing behavior by many individual actors really works. Disequilibrium behavior is not fully specified (unless it is by ad-hoc assumptions. But this means that there is no well-defined dynamic process of which the ‘equilibrium’ is a stationary point: consistency relations, and not zero rates of change, define equilibrium ‘ [p 32]. This is an important sub-conclusion: the design of the theory and the models are static in nature. Static situations are the norm, equilibrium is not a special case (zero change) of a system usually in disequilibrium, but it is the norm and disequilibrium can only be treated as a special case. The models do not explain how equilibrium comes about from disequilibrium, because they can only describe equilibrium situations as per their structure. 3) In conclusion the formalizations based on orthodoxy cannot deliver: ‘Increasingly, orthodoxy builds a rococo palace logical palace on loose empirical sand‘ [p 33].

Allies and Antecedents of Evolutionary Theory

Managerialist thinking diagnosis the problem of orthodox theory as a failure to represent correctly the motives that directly operate on business decisions.  These insights are useful for insights into managerial behavior and performance, but they do not explain the problems with orthodoxy as described earlier above.

Behavioralists (Herbert Simon) adress some or all of these issues: bounded rationality because people cannot have complete information and they do not have computational power in some situation to maximize over all possible alternative utilities. Calculations are therefore global and situational, maximization is not absolute. Firms satisfice: their objectives function is limited because they don’t have the complete overview of their utility trade-offs and because they they are coalitions of decision makers associated with the firm. Disagreement with the behavioralists in the search of the authors for an explicit model of industries, instead of the behavior of individual firms. Evolutionary models are simpler but explain the industry compared to behavoralist’s.

Analysts of Firm Organization and Strategy

Such strategies differ from firm to firm, in part because of different interpretations of economic opportunities and constraints and in part because different firms are good at different things. In turn, the capabilities of a firm are embedded in its organizational structure, which is better adapted to certain strategies than to others. Thus, strategies at any time are constrained by organization‘ [p 37]. Assuming that organization is self-organization in an autopoietic sense then it is autonomous. Firm capabilities and strategy must logically flow from it. Strategy must logically be a narrative after the fact explaining the way that the business is conducted.

In the evolutionary models higher order decision rules can serve as strategies. Change  in strategy or policy can be treated in the same way as change in technique. Connections exist between the firm’s strategy and its appropriate (or is it the other way around: the appropriate strategy to go with the organizational structure?) organizational structure and between the technique it commands and its organizational structure.

Views of the Activist Firm

.. to view large firms and relatively concentrated market structures as the typical cases in the ‘interesting’ part of the economy, if not in the economy as a whole. These perspectives converge in an assessment of the larger corporation as a critical feature of the institutional dynamics of modern capitalism, as a relatively autonomous chooser of society’s means and some extent of its effective ends, and as the stimulus for the development of new social institutions for its control and accommodation‘ [p 39]


This theory is evolutionary in order to be able to be Schumpeterian.

Frank Knight and the Modern Austrians

Schumpeter stressed innovation as  a deviation from routine behavior, and argued that innovation continually upsets equilibrium‘ [p 41]. Reality brings new opportunity to bear all the time and the economic problem is how to respond to it and make a profit. It is however, impossible to calculate the effects and only afterwards will become clear what was the right thing to do. This theory is a theory about market processes.

Evolutionary Theorists

The article of Alchian Uncertainty, Evolution and Economic Theory of 1950 is an antecedent of this theory: ‘.. this problem had been anticipated by Alchian who emphasized the ‘reproduction by imitating of rules of behavior‘ [ Alchian 1950 pp. 215-6 in p. 42]. Socio-biology is involved via E.O. Wilson 1975. Campbell 1969 argued for a focus on variation and cultural selection retention theory for sociocultural evolution. ‘Our own work may be viewed as a specialized branch of such as theory .. ‘ [p 43].

Classical, Marxian and Neoclassical Antecedents

This theory is consonant with the original ideas of the classical theory as per Smith and John Stuart Mill. Much of the economic theories of Marx are evolutionary and also in other aspects compatible with this theory: capitalist organization of production defines a dynamic evolutionary system and the distribution of firm sizes and profits must be understood in evolutionary terms. But this theory makes no distinction between profits and wages and also do politics not play a role. Lastly the ideas of contradictions and of class have no place in this theory.

Economics comprises two different styles of theorizing: formal and appreciative. In a broad theoretical framework phenomena are observed in a particular framework of appreciation. A theory defines economic variables and their relations and provides a language and a mode of acceptable explanation. Implicitly a theory distinguishes and ranks the elements considered more and less important, complete and sophisticated, respectively. Formal theory is an important source of ideas for invoked in appreciative theory and sharpens its tools. Experimental data that are not accounted for by existing theory are grist for extensions and amendments to them. Appreciative theorizing in a sense follows formal theorizing, because the odds are larger that it is more fruitful where the other is and in that sense the dependency is a constraint.

Heterodox tradition has not made it into mainstream economics because of vested intellectual interests and parochialism [Galbraith . The New Industrial State . 1967]. Conversely the existing theories are flexible and accommodating. However not everything is deemed important or interesting to cater for it; that obviously is a risk for theory development because orthodoxy decides what should be interesting and what not. It occurs that theories are not found interesting the real reason being that the new ideas could only be married to the existing theories implementing large adjustments.

The Foundations of Contemporary Orthodoxy

Systematic understanding of the activities inside firms is not a high ranking subject of interest of economists, but rather industries and national economies &c. The firm is treated as a functional element in the analysis at hand; it is a black box the input and output are gauged at the convenience of the observer. Firms in real life do not do these actions, respond to price changes, or manage productive input combinations or manage financing in isolation, but all of them in parallel. The emphasis in this book is similar, namely on the aggregate (industry, nation), not the individual firm. Different subjects have to be addressed separately in a laboratory firm because it is not logically feasible to observe them simultaneously. The authors keep in mind the events within the firm, having in mind the development of a model for the larger systems.

Orthodox theory treats ‘knowing how to do’ and ‘knowing how to choose’ as different things, namely the ability to choose from a clear set of options and the ability to choose optimally respectively. In this theory they are treated as very similar things: the range of things a firm can do is somewhat uncertain as is the capabilities to choose in a given situation. The parameters are, in that order: objectives  > howto > optimal choice >< internal and external conditions. In what sense can business firms have objectives?

1 The Objectives of Business Firms

In the simplest orthodox model the objective is profit or business value and the more the better. Two opposite considerations: the business firm with the number of individuals involved, the diversity of their roles and the complexity of their relations VERSUS individualist utilitarian philosophy underlying neoclassical theory and optimality theorems. ‘In this philosophical framework, economic organization in its entirety is appraised for  its effectiveness in satisfying the wants of individuals. ‘A forteriori’, the business firm is viewed as in some sense an instrumentality of individuals, rather than an autonomous entity‘ [p 54]. If Miller’s Mill is the subject of observation than the relation between the interests of the owner and the ‘interests’ of the firm are clear, but this relation becomes strained if the subject is  General Mill. I can’t really believe there is a distinction: Miller has many roles to play and some of those can become mixed. In the case of  the owner (and the employee and the CEO &c.) of General these roles are specialized and less mixed. But in my view there is no fundamental difference between small firms and large ones.

Managerialists argue that not the – often passive – shareholder’s interests are relevant but the manager’s. The manager’s interest is operational and can be associated with size and growth. Another school defines a firm as a nexus of bargaining between parties to get to a coalition (Cyert & March). ‘Goals’ and ‘Objectives’ of the firm cannot impose a coherent structure on a firm’s actions. In their view the matter cannot be resolved, because that would involve more bargaining than possible in practical terms. Instead the firm remains in a state of quasi-resolution of conflict. Even if top management manages to negotiate a common ground then the interests can be different during the implementation and result in a new round of bargaining such that it is an important factor in the behavior of the firm: objectives such as profit and revenue are useless unless they are made specific as to how to achieve them for all involved (ie associated with the firm).

In reality, firms’ goals can  be vague, because even if a main objective is established then the partial objectives may remain unclear, unaligned or not sufficiently specified for their users. If this is true for reality and firms show behavior then it can also be true for the firms of this theory. What is in fact required is a procedure to determine which action is to be taken.

2 Production Set and Organizational Capabilities

In 1 some examples were discussed concerning the motivational aspects of the theory of the firm. There is little theory concerning the capabilities of firms: formally what a firm can do is catered for through production sets: the set Y of production possible for the j-th producer is a production set [Debreu 1959 p34 in p 59]. What is the firms production set? ‘The production possibility is set is a description of the state of the firm’s knowledge about the possibilities of transforming commodities‘ [Arrow & Hahn 1971 p 53 in p 60]. But what is the nature of knowledge? In orthodoxy the connotation is with ‘a way of doing something’ or ’technology’ as in a book with blueprints, the knowledge of scientists and engineers. DPB: I reckon in general it is what people believe to be true and specifically it is what they believe to be true concerning their business environment. If it is related to science or engineering then chances are larger that it is in fact true and not only believed to be so. ‘Implicit in both metaphors (‘blueprinted knowledge and knowledge of engineers and scientists, DPB’, and in other discussions, is that technological knowledge is both articulable and articulated: you can look it up. At least, you could if you had the appropriate training‘ [p 60].