1998-05-05 Gunnar Eliasson KTH, Stockholm

THE NATURE OF ECONOMIC CHANGE AND MANAGEMENT IN THE KNOWLEDGE-BASED INFORMATION ECONOMY1 by Gunnar Eliasson Department of Industrial Economics and Management Royal Institute of Technology (KTH) S-10044 Stockholm, Sweden

Abstract This paper has a negative message. When the really important decisions are concerned, the economic world is fundamentally unpredictable. It cannot be revealed by analytic methods, which are rather liable to lead us astray at critical moments. This limitation of insight and foresight should be carefully considered when we discuss and use scientific methods of management. A standard argument of economic thinkers of the not to distant past was that new information technology would soon make the real economy around us fully transparent and accessible for analytical understanding, optimization of individual and aggregate behavior and the circumstances perfectly arranged for informed central planning. Such were the predictions of neo walrasian analysts and their derived believers in the business world. So far no such full information equilibrium real economy has materialized. The development has been exactly the opposite. Not even in the centrally planned economies in the East where all unexpected change at the individual and firm levels was brutally suppressed has it been possible to control the economies from a central point. And would anybody have expected anything else? The economists’ thinking on these matters has been reflected in waves of change in management literature and practice. The analytical mode of management writing and practice (long-range planning) of the 1960s was pushed (by the oil crises of the 1970s) into an experimental and control mode in the 1980s. In this paper I inquire into the nature of information use and knowledge transfer in the business world and conclude, on theoretical grounds, that the mode of operations management of large firms should stay where practice has now taken it; i.e. on the experimental and control mode. But this will not be sufficient for new and innovative small businesses that will have to carry the production of the mature industrial economies into the next generation, and where intelligent choice in a fast changing technological environment matters far more than efficient central control.

2 The Problem Between 1969 and 1975 I carried out some 80 interviews of 60 US and European firms (Eliasson 1976) on their internal information and control systems. Then it was all short-term and long-range planning and a strong belief in a repetitive environment, forecasting and central leadership of standardized production. Between 1985 and 1995 I carried out an additional series of some 70 interviews with 50 firms, several of them the same as in the earlier interview series, and also some 15 firms that had attempted in the early 1980s to establish themselves in the then hot business information systems market (Eliasson 1996a). Two different intellectual worlds emerged. In 1970 it was all planning, In 1990 it was all experiment, control and access to information and people. In between the two observation periods lay the great learning experience of the disorderly 1970s. A comparison of the management practices in the two periods is interesting both theoretically and empirically. It illustrates how dependent we all are on the intellectual fashions of the day. By the early 1970s many managers saw the world as predictable and locally (for the firm) plannable. By the early 1980s, a decade of disorderly market experience had suppressed that view, but the rapidly innovating information technology made many managers still believe that technology and universal business information systems would overcome complexity and provide top level management with general, and highly flexible information tools. By 1995, after billions of dollars of lost development money we have learned again. New information and communication technologies have rather made the world around us increasingly heterogeneous, complex and unpredictable. Information technology has increased our capacity to gather, process and analyze information, but at the same time also diminished our possibilities of being informed about the whole by increasing the total of circumstances that we can be informed about at a faster rate than we have learned about the previous state of affairs. Most likely we are becoming relatively (and increasingly so) more ignorant (Eliasson 1990b, p. 46). Current advance among economic analysts, reluctantly accepting the disruptive notion of non-linear economic behavior, furthermore has made it clear that the state of full information equilibrium by standard definitions is not even theoretically attainable in a realistically represented model of the real economy. I call one such alternative non linear model of the real economy to be discussed in this paper the Experimentally Organized Economy (EOE), Eliasson 1987, 1991a). The management of theoretical firms (hierarchies) in an EOE is radically different from the management of a theoretical firm in the conventional neo walrasian or neoclassical model of an economy. Out goes reliance on detached analytical thinking in the executive quarters. In comes rational, experimental behavior, restoring the knightian (1921) distinction between uncertainty and risk. This brief paper introduces the characteristics of the EOE and derives the particular problems of information processing a manager in such an economy has to face and should have expected all along. Viable economies have always been experimentally organized, characterized by significant non linear properties, and not capable of coming to rest in the lit de parade type equilibrium that underlies most analytical methods used in economic and management theory.

3 The Experimental Nature of Economic Behavior The manager of each firm would prefer to look forward to a long and successful business life, without having to face the hazards of the real business world. Is there such a fixture submerged under the violent seas of market life on which real managers have to navigate that a good manager with access to modern information and communication technologies should be able to uncover. Such is the presumption of the standard economic neo walrasian economic model, still being embraced by most economists, still being the foundation of modern finance theory, once being unknowingly embraced by the scientific management movement and by the strategic planners of large corporations in the West during the 1960s, and, not to forget, by many corporate executives. It was even seriously believed by many researchers that the system would soon successfully take on many dominant top executive decisions (see Survey in Eliasson 1976). People would come and go. The system would embody the management competence. Such was the intellectual foundation of Soviet central planning, solidly founded in neo classical economics. To all central planners of the East and to most keynesian policy makers of the West during the 1960s and the 1970s the unpredictable events in markets and the consequent decision mistakes were negative events, and instances of undesired economic waste that should be eliminated through informed policy (Pelikan 1986). Such was the foundation of management teaching, writing and practice in the late 1960s, based on the idea of a possible full information market economy. When the right intellectual support tools had been made available the business situation would be under intellectual control. It all ended abruptly in the disorderly 1970s. Despite the costly learning experience of the 1970s the engineers in some 20 large IT firms, however, set out in the early 1980s to design and market the universal business information system based on the full information idea. They all failed, some completely, at the tune of billions of lost dollars (Eliasson 1996a). Adam Smith (1776) laid down the principal design of a decentralized market economy in which the division of labor makes economies of scale ”in the small” possible and the realization of large macro productivity effects feasible. This benefit, however, came at a significant information and communication cost, a fact that advocates of the ”modern” mathematical representations of the ”invisible hand” took a very long time to understand.2 The organization of the division of labor is an instance of innovative behavior and entrepreneurship. This organization evolved gradually in the market. Once the necessary choices and selections had been made, however, economic activity had to be coordinated physically (transports) and through communication (Table 1). Once an innovative design has been accomplished, competitors would be on ”your doorstep” to learn (imitate). If your organization is large enough you would want to diffuse the knowledge throughout the organization. You would also want to sell your knowledge at a profit (consulting). Learning, hence, becomes a general and resource-using economic activity. Even very simple tasks (you would soon learn) can normally be solved in a very large number of ways. The higher up, the more complex the decision problem and the larger the number of possible solutions. Some of these solutions are better than others. The problem, however, is that you will never know them all and you will never know how good they are until you have tried them, and even though you may have stumbled upon the absolutely best solution you will

4 never know. You will always have to count on the possibility that there may be several much better solutions soon to be put on the market by your competitor. This is the essence of the experimentally organized economy. The business manager can never feel safe and will have to recognize in his or her management practice (Table 3) the possibility of coming out as a loser. The number of solutions defines the large business opportunity set that faces each agent who has to search his or her way into the opportunity set enacting business experiments, being directed by a limited vision (theory) of all possibilities (bounded rationality). Hence, it becomes important to understand how agents access and interpret information being generated by the ongoing economic process, and to what extent this information can be used to predict the future. Since each agent, furthermore, has his or her particular vision as guidance, there will be strong limitations on communication because of limited and differently composed receiver competence (Eliasson 1990a, p. 1). The result will be, at each point in time, a heterogeneous structure of competence, defined by the organization of people in the economy. Much of the knowledge put to use in a firm, especially high-level knowledge, vested in the top competent team of a firm is difficult or impossible to communicate in coded form as information. It is tacit. Tacit knowledge is acquired through on-the-job learning and filters through the economy (selection) through the acquisition of the whole firm, or parts of firms in the M&A market or through the mobility of people or teams of people with competence in the labor market (Eliasson 1991b). I have introduced the four general information activities of the knowledge-based information economy: innovation, selection, coordination and learning (see Table 1 and Eliasson 1990b). Together they can be defined to cover all economic activity representing the intellectual superstructure (the memory) of the economy that controls all other activities. These four information and communication activities are all present in the firm as administrative processes, and in the market or in any combination of the two forming competence blocs (Eliasson and Eliasson 1996) supporting, and geared towards the manufacturing of some particular set of related products. A competence bloc diffuses previously internal (within the hierarchy) knowledge selectively through the market, and is capable of a more efficient use of heterogeneous competence than the monolitic firm (Eliasson 1990a). Competence blocs, hence, are extended form firm organizations that evolve whenever product development and production complexity surpass the internal limits of one hierarchy. They represent an unstable decentralization of production over hierarchies and markets (Eliasson 1996b) that requires new and different competencies of firm managers.

Growth through Competitive Selection Suppose all agents in a market are ranked along a Salter (1960) curve by some simple performance measure like productivity (Figure 1). All firms except one face a superior competitor capable of paying more for the factors of production. But each firm also faces inferior competitors who all strive to improve their performance to come out on top of their betters. In addition all incumbent firms have to be prepared for innovative entrants. No actor can feel at ease in such a situation. The outcome will be a steady improvement along the curve and an outward shifting of the curve, being pushed from its upper left side. Losers leave far down at the right end. As a consequence growth will be recorded at the macro level through (Eliasson 1996a,c) the four investment growth mechanisms of Table 2: innovative entry, enforced reorganization, rationalization and exit. Exit is not least important since it releases resources for more remunerable and more efficient employment elsewhere. It is obvious from

5 what has been said that free entry, or free access to the market is an important factor behind competition and a driver of economic growth through the four mechanisms of Table 2. As long as we are talking about given firm entities Table 2 exhausts all possible means of growth. Reorganization, however, can also occur across these categories.3 Four important conclusions emerge out of this presentation of the EOE. 1. An economy is populated by a large number of actors, each being grossly but differently uninformed about its environment and the opportunity set they are facing, all constantly making more or less serious business mistakes because of their ignorance, some being forced to leave the market. Business mistakes, hence, are a normal cost for economic development that should not be minimized or eliminated. On the contrary, 2. Business mistakes, if not overwhelmingly large (Eliasson 1992), constitute important instances of individual and corporate learning that are unavoidably associated with experimental behavior that pushes the performance of individuals, firms and entire economies outwards. 3. Together all these actors and their successes and mistakes determine each others’ environment and push the opportunity set outwards. Economic growth occurs through competitive selection. 4. For none of the actors is there a fixture (an exogenous equilibrium) submerged under the disorderly surface of the markets. Because of entry and exit and the multiple market confrontations of incompatible expectations the economy is fundamentally non-linear, and these non-linearities constitute the theoretical character of the environment real life managers have to cope with (Eliasson 1991a).

Managing a Firm in the Experimentally Organized Economy The firm of the neo walrasian world is at most marginally uninformed and can obtain reliable information at a cost to be as fully informed as it finds rational. The firm in the EOE is grossly uninformed about most things that matter critically for its survival. It cannot collect additional reliable information. Normally management is not even aware of critical survival determining events that may unexpectedly occur. It knows, however, that it nevertheless has to form a coherent view about its future environment and decide where in that environment it wants to be. It also knows that if it has a good business idea that it believes in, it has to reckon with the possibility that a number of competitors will be on to something at least equally good. As a consequence each time it has to act prematurely on its business idea. The innovating entrepreneurial firm will also have to reckon with the fact that most actors won’t understand his or her business idea, notably the resource providers. Even though the entrepreneur has personally converted the uncertainty surrounding his or her project to low and subjectively computable risks (see next section), the resource providers are likely only to see great uncertainty and apply very high risk premia in their evaluations (Eliasson 1997). Several conclusions follow from this observation. First, considering its limited awareness of the full content of the opportunity set the firm has to count on the possibility that its decision may be all wrong. The members of the scientific management movement, based their beliefs on the static equilibrium model and felt safe to operate on a predictable planning mode, expecting subjectively to be correct on the average.

6 Such a firm (Eliasson 1976) would place a lot of confidence in the strategic long-range planners on the staff, attempting to map the future and substitute detached analysis for business judgement under item 1 in Table 3. In the EOE, however, intuition, experience and good judgement play the central role under item 1, and the formal information systems are geared to items 3 and 4, namely to identify and to correct business mistakes (Eliasson 1996a). Only when these two checkpoints have been passed will the standard operating mode of the firm model of economic theory, financial economics (rational expectations and efficient markets) and standard investment calculations at best be applicable. In the pronounced non-linear domains of normal markets the analytical learning feed backs of item 6 won’t be reliable. It becomes a critical business competence to interpret the information (signals) emitted by the economy such that the subjective order needed for decisions can be established. In such situations the experimental manager mode of the EOE firm focusing on intuitive business judgement (item 1 in Table 3) and subsequent error identification and correction (items 3 and 4) will dominate. This distinction between repetitive learning behavior on the neo walrasian mode and experimental search is theoretically fundamental for economics. Not only has the last two decades seen a shift away from the analytical planning mode so common in the firms of the 1960s and early 1970s (Eliasson 1976) towards the experimental mode (Eliasson 1996a). The new theoretical insights (to be elaborated in the last section) come out of a micro-based theory of macro behavior making exogenous learnable equilibria theoretically non existing artifacts that will be entirely misleading in most situations where economic advice is asked for. Even so the bulk of subjects on the teaching agenda of business schools, like investment calculation and financial economics rests on the assumptions of the neo walrasian model.

A Theoretical Note on the Impossibility of Reliably Informed Economic Decisions Insurance companies base their business on trading in computable actuarial risks based on experience determined distributions. Such are also the assumptions of rational expectations and efficient market theory. Such theory, and reliable statistical influence are contingent on a learnable set of information. Such theory, as well as insurance business, has a problem if the learnable information set is influenced by learning itself. Knight(1921) realized that real business was rather of such a nature, dealing in certain events, occurrences that could not be predicted as drawings from an empirically determined stable distribution. In general, Knight regarded business risks as principally and practically uninsurable. He furthermore suggested that the particular ability of the business manager was to convert such uncertainty into subjectively calculable risks on the basis of which he or she acted with confidence (LeRoy and Singell 1987, Eliasson 1990a). For the outsider, then the risk level (item 2 in Table 3) might appear prohibitive. For the business man, entrepreneur, being confident in his or her assessment of the situation, the subjective risks are low and the business situation under subjective control. The entrepreneur, so to speak, converts a highly complex non-linear environment onto a computable format. Modern finance literature makes no distinction between uncertainty and risk and uses the terms synonymously. All business uncertainties are reduced to computable and insurable risks. Besides the ridiculous implications of such assumptions, they detract attention from the basic character of business life, namely true and individually experienced uncertainty. One consequence of this has been the plethora of attempts to analyze the results of statistical learning and the diffusion of information in the neo walrasian type economic setting. If you

7 know the functional form of the distribution of risks and if it is mathematically nice (for a survey, see Lindh 1993) statistical learning applies. Repeated observation of the data emitted from the distributions allow you to ascertain its coefficients with any desired accuracy. For this to hold it has to be assumed that (1) the underlying structure is sufficiently stable to allow for repetitive data generation, and (2) that information use carries no cost. If you don’t know the functional form of the distribution, however, you still have a problem. If the mathematical characteristics of the distribution are sufficiently non linear (whether known or not) efficient statistical estimation techniques will not allow you to come up with consistent estimates of the underlying parameters.4 Whether stable or not, under quite normal circumstances, the true underlying structure will forever remain unknown to you. If the environment is sufficiently complex any approximate estimable representation of the distribution will normally be unstable and the same results hold. In markedly non-linear economic environments there will exist no underlying exogenous and tractable structure that you can learn about and position yourself against in a reliable and stable way. Learning will have to be very differently presented in such non-linear and experimentally organized environments, very much as presented in Table 3. This situation can arise in two theoretical circumstances: 1. Complexity 2. Significant learning costs. Complexity was explained above. Learning costs introduce another form of complexity. You expend costs while searching for the underlying fundamental structure. Once you have found it these information costs disappear, and since they have been significant codeterminants of the equilibrium, the equilibrium also shifts away. And you have to go on searching (Day 1993). This peculiar explanation is needed for the simple reason that we are all trained in static equilibrium analysis. Complexity, in fact, is all we need to reject the neo walrasian static equilibrium model and its entire superstructure of literature. The Swedish micro-to-macro model Moses (Eliasson 1977, 1985, 1991a) provides excellent illustrations. We, its engineers, know its highly complex non-linear structure (the fundamentals). Hence, we can generate the data for outsiders to watch. Let me use that model to illustrate two critical points. First, the highly non-linear structure of the micro-to-macro model can be approximated by a simple linear model from the economists’ tool box. Suppose, we impose, very much as Walras did on the ”Smith and Ricardo” model, an equilibrium market clearing constraint, forcing each actor to be in capital market equilibrium with the same rate of return. In the Moses model this can only be achieved through explicit simulation of the dynamic competitive market processes pushing the entire economy, through enhanced competition, closer to the ”approximate static equilibrium”. The assumption is that with efficient competition firms will learn all the hard way, or perish. The results (Eliasson 1983, 1984, 1985, 1991a) are significative. The closer to static equilibrium the economy was being pushed by market forces the more unstable the model economy became and the more unreliable the signals emitted by the market processes. Eventually the economy collapsed.5 The closer to equilibrium the more unstable the underlying structure, the more shifty the equilibrium and the more unreliable the information emitted by the economy. This result accords with the uncertainty principle in physics and with the negative results of statistical learning theory (Lindh 1993). Reliable and interpretable price signals are only emitted if the economy (or the model) features a sufficiently stable quantitative structure, to be revealed by the price signals.

8 If firms act on the basis of unreliable price signals structure is liable to change through the differentiated quantitative performance created by the mistaken decisions, for instance entry and exit. Expressed in other terms; static equilibrium did not exist as an operating point of the economy. Hence, the operating economy could not reveal (through market price and quantity signals) its location. Second, for an outsider to ascertain the unknown parameters of the fundamental structure of the micro-to-macro model some approximate model has to be estimated. Bounded rationality is introduced. Now economic theory enters. Antonov and Trofimov (1993), for instance, estimated a traditional keynesian and neoclassical macro model on the data generated by the micro-to-macro model. These two linear approximations to the full model were constantly updated, as new data were being generated by the Moses model and the firms in the Moses model used the predictions of the econometric model approximations of Moses in their planning. The mode of information generation, data collection and analysis in the Moses model economy thus was designed to influence total economic behavior and performance in the Moses model. In some experiments firms were forced to use the predictions of one or the other econometric model (”central planning”). In other experiments they could choose individually what information to use on the basis of the individual predictive performance of prices, sales etc. predictions of the official econometric models or their own more primitive projections. This is what we learned. The simple econometric approximations to the full model were of course incapable of uncovering anything interesting about the deeper structures of the Moses model economy. When the firms individually used their own crude projection models the error rate of course increased. But some firms, nevertheless, stumbled onto continued success. There are two possible and mutually supporting explanation to this: (1) active broadbased and unrestricted search and/or (2) individual actors who know more than can be understood at higher levels. Pluralism is thus needed to make efficient use of the knowledge base of the economy, not centralized information processing. Since competition weeded out the unlucky actors or reduced their resource use, this non analytical experimental management mode generated faster aggregate growth in the Moses economy than did the planned information diffusion and analytical management mode. This result illustrates the main theme of this paper. Comparing the statistical learning models with the experimental search model not only sets the centralized analytical mode of business management of the 1960s against the experimental mode of the 1980s, and for that reason theory against reality. It also demonstrates the impossibility (non existence) of reliably informed decisions.

Implications The above analysis may seem overly abstract and theoretical. It may be, but the implications are very practical. Every manager tries to overcome frustrating uncertainty through resorting to some simple and analytically comprehensible management method. For reasons we have clarified above, belief in such methods will eventually lead him or her astray in critical situations. If all firms resort to analytical management methods they, furthermore, reinforce the past and the economy, very much as Ballot and Taymaz (1996) show theoretically, will easily get locked into an inferior structure.6 For the economy to develop ahead of what seems ”analytically possible” on the basis of the past, the economy needs many actors exhibiting odd and analytically ”incomprehensible”

9 behavior. This means a viable new establishment, and a forceful exit of low performers. The innovative entry process is truly experimental and has little to do with the management of large firms, engaged in standardized production with the purpose of capturing increasing returns. A dominance of such management competence is likely to be detrimental to the development of competence to manage radically new industry. If this is not taken into account in theorizing about the firm or about rational management methods we won’t understand much of what is going on. The notion of a firm can only be created in the context of the market in which it is supposed to operate. The management teacher as well as the economic theorist needs a realistic model (method) of the firm to support teaching and thinking. Since no realistic theory of dynamic markets exists no good theory of the firm has been created. The moral, hence, is that so far we have excellent firms, not thanks to, but despite management teaching.

10 Table 1. The statistical accounts of the knowledge-based information economy 1. Identifying business opportunities (exploring state space)

The creation of new knowledge - innovation - entrepreneurship - technical development

2. Choice and selection

Filtering - entry - exit - mobility - careers

3. Coordination

Disciplining - competition (in markets) - management (in hierarchies) Knowledge transfer - education - imitation - diffusion

4. Learning

Source: Eliasson (1990b).

Table 2. The four investment growth mechanisms 1. 2. 3. 4.

Innovative entry Reorganization of existing firms Rationalization of existing firms Exit (bankruptcy)

Source: Eliasson 1996a, p. 45

Table 3. The competence specification of the experimentally organized firm Orientation (Innovation) 1. Sense of direction (business orientation) 2. Management of uncertainty Selection 1. Efficient identification of mistakes 2. Effective correction of mistakes Operation 5. Efficient coordination 6. Efficient learning feed back to (1) Source: Eliasson (1990a)

11

Figure 1. Labor productivities distributions in Swedish manufacturing

Source: Moses Database, IUI, Stockholm 1992.

12 Bibliography Antonov, M. and G. Trofimov, 1993. Learning through Short-Run Macroeconomic Forecasts in a Micro-to-Macro Model, Journal of Economic Behavior and Organization, 21 (2), June. Ballot, G. and E. Taymaz, 1996. Human Capital, Technological Lock-in and Evolutionary Dynamics. Paper presented to the 6th International Joseph A. Schumpeter Society Conference in Stockholm, June 2-5, 1996. To be published in G. Eliasson and C. Green (eds.), The Micro Foundation of Economic Growth. University of Michigan Press, 1998. Day, R.H., 1993. Bounded Rationality and the Coevolution of Market and State; Chapter 4 in R.H. Day, G. Eliasson and C. Wihlborg, eds. (1993). Day, R.H., G. Eliasson and C. Wihlborg (eds.), 1993. The Markets for Innovation, Ownership and Control. Stockholm: Industriens Utredningsinstitut (IUI) and Amsterdam: Elsevier Science Publishers B.V. Eliasson, G., 1976. Business Economic Planning - Theory, Practice and Comparison. London, New York etc.: Wiley & Sons. --------, 1977. Competition and Market Processes in a Simulation Model of the Swedish Economy, American Economic Review, 67 (1), 277-281. --------, 1984. Heterogeneity of Firms and the Stability of Industrial Growth, Journal of Economic Behavior and Organization, 5, (3-4). --------, 1985. The Firm and Financial Markets in The Swedish Micro-to-Macro Model Theory, Model and Verification. Stockholm: Industriens Utredningsinstitut (IUI). --------, 1987. Technological Competition and Trade in the Experimentally Organized Economy, Research Report No. 32. Stockholm: Industriens Utredningsinstitut (IUI). --------, 1990a. The Firm as a Competent Team, Journal of Economic Behavior and Organization, 13 (3), 275-298. --------, 1990b. The Knowledge Based Information Economy. Stockholm: Industriens Utredningsinstitut and Telia. --------, 1991a. Modeling the Experimentally Organized Economy, Journal of Economic Behavior and Organization, 16 (1-2), 153-182. --------, 1991b. Financial Institutions in a European Market for Executive Competence; in C. Wihlborg, M. Fratianni and T.D. Willett (eds.), 1992, Financial Regulation and Monetary Arrangements after 1992. Amsterdam: Elsevier Science Publishers B.V. --------, 1991c. Deregulation, Innovative Entry and Structural Diversity as a Source of Stable and Rapid Economic Growtrh, Journal of Evolutionary Economics, (1), --------, 1992. Business Competence, Organizational Learning and Economic Growth: Establishing the Smith-Schumpeter-Wicksell Connection; in F. M. Scherer and M. Perlman (eds.), Entrepreneurship, Technological Innovation, and Economic Growth. Studies in the Schumpeterian Tradition. Ann Arbor: The University of Michigan Press. --------, 1996a. Firm Objectives, Controls and Organization. Boston/Dordrecht/London: Kluwer Academic Publishers. --------, 1996b. Spillover, Integrated Production and the Theory of the Firm, Journal of Economic Behavior and Organization, 6:125-140. --------, 1996c. Endogenous Economic Growth through Selection; in A. Harding (ed.), Microsimulation and Public Policy. North-Holland: Amsterdam. --------, 1997. The Venture Capitalist as a Competent Outsider; mimeo INDEK, KTH, Stockholm.

13 Eliasson, G. and Å. Eliasson, 1996. The Biotechnological Competence Bloc, Revue d’Economie Industrielle. Glete, J., 1996. Entrepreneurs and Social Elites; some reflections on the case of Sweden; Paper presented to the 6th International Joseph A. Schumpeter Society Conference in Stockholm, June 2-5, 1996. To be published in G. Eliasson and C. Green (eds.), The Micro Foundation of Economic Growth. University of Michigan Press, 1998. Knight, F., 1921. Risk, Uncertainty and Profit. Boston: Houghton-Mifflin. LeRoy, S.F. and L.D. Singell, Jr, 1987. Knight on Risk and Uncertainty, Journal of Political Economy, 95 (2), 394-406. Lindh, T., 1993. Lessons from Learning to have Rational Expectations; Chapter 5 in R.H. Day, G. Eliasson and C. Wihlborg, eds. (1993). Pelikan, P., 1986. Why Private Enterprise? Towards a Dynamic Analysis of Economic Institutions and Policies; in The Economics of Institutions and Markets, IUI Yearbook 1986-1987. Stockholm: Industriens Utredningsinstitut. Salter, W.E.G., 1960. Productivity and Technical Change. Cambridge, MA: Cambridge University Press. 1

Earlier versions of this paper were presented at the Pacific Telecommunications Council Conference (PTC), January 20-22 1997 in Honolulu, and at the 14th Nordic Conference on Business Studies, Bod•, Norway, August 1997. 2

When Gérard Debreu received the Prize in Economics in Honor of Alfred Nobel he was told that he got it because of his modeling of the invisible hand.

3

For instance, through its acquiring part of another firm or a small firm in the market, and/or selling off a division in the M&A market.

4

This conclusion has two interpretations: (1) It may be practically impossible to process the data needed to estimate the model, or (2) the model may include specifications that make the process influence the underlying structure. Such profound non linearities, characterizing path-dependent development, remove the repetitiveness needed to estimate the model. There will not be enough data to capture the structure. This time satisfactory estimation is theoretically impossible.

5

This could only be prevented (Eliasson 1991c) through a viable firm entry and exit process.

6

This is what Glete (1996) suggests has happened to Swedish engineering industry over the past century.

Labor productivity and wage cost distribution in Swedish manufacturing, 1991 and 1993 3

2

1

0 0

10

20

30

40 VA/L 1991

Source: MOSES Database, IUI, Stockholm 1992

50 W/L 1991

60 VA/L 1993

70 W/L 1993

80

90

100