Discussion Paper Series

Transcription

1 Discussion Paper Series # Growth Theories Revisited: Enduring Questions with Changing Answers C. V. Vaitsos October 2003 United Nations University, Institute for New Technologies, Keizer Karelplein 19, 6211 TC Maastricht, The Netherlands Tel: (31) (43) , Fax: (31) (43) , URL:

2

3 GROWTH THEORIES REVISITED: ENDURING QUESTIONS WITH CHANGING ANSWERS* C. V. Vaitsos Professor of Political Economy Economics Department, Athens University Abstract Three epistemologically distinct scientific approaches have dominated the relevant theoretical and empirical analysis on the determinants and dynamic processes of aggregate economic growth. These refer to (i) formal growth modeling approaches, (ii) evolutionary economics constructed on elements from appreciative theorizing, and (iii) holistic approaches of reasoned history about growth dynamics. Each one of these approaches is founded on quite distinct scientific paradigms and working hypotheses. They thus reach a different class of scientific conclusions and of applied interpretations. Explaining the reasons and nature of such differences as well as the areas of their complementarity constitutes a major challenge in reviewing the state of our understanding as to how socioeconomic systems experience alternative growth paths, reproduce themselves and change over time. Furthermore, the closer these three approaches come to reflect the actual complexity of aggregate growth dynamics, the more difficult it becomes to establish analytical tractability and to reach more general conclusions. UNU/INTECH Discussion Papers ISSN Copyright 2002 The United Nations University, Institute for New Technologies, UNU/INTECH UNU/INTECH discussion papers intend to disseminate preliminary results of the research carried out at the institute to attract comments * Review research project on the state of the art sponsored by the Research Directorate-General of the European Commission

4

5 TABLE OF CONTENTS PROLOGUE 7 INTRODUCTION 9 A. THE INTELLECTUAL ANTE-SALA OF MODERN GROWTH THEORIES 15 B. FORMAL GROWTH MODELING 21 B.1. THE SAVINGS-CAPITAL ACCUMULATION NEXUS TO GROWTH: 21 B.2. NEOCLASSICAL GROWTH THEORIES 24 B.3. NEW GROWTH THEORIZING AND RESULTING DEVELOPMENT PATTERNS 29 C. EVOLUTIONARY THEORIES WITH SOME ADDITIONAL REFERENCES TO INNOVATION ECONOMICS 43 D. ECONOMIC HISTORY AND GROWTH THEORIES 53 E. EQUITY AND GROWTH CONSIDERATIONS 57 E.1 THE EFFECTS OF INEQUALITY AND REDISTRIBUTIVE POLICIES ON GROWTH 57 E.2 POLITICAL DECISION MAKING, EQUITY AND GROWTH PATTERNS 62 E.3 HUMAN CAPITAL FORMATION AND EQUITY CONSIDERATIONS 63 E.4 POLITICAL STABILITY AND MACROECONOMIC VOLATILITY 64 CONCLUDING REMARKS 67 BIBLIOGRAPHY 69 THE UNU/INTECH DISCUSSION PAPER SERIES 77

6

7 PROLOGUE Economic theories do not pretend to mirror reality. Through an abstraction cum selective simplification process, they only attempt to isolate significant strands in economic causal relations. At the very least, they endeavor to provide a way of organizing thoughts and ordering concepts by explaining how such strands interrelate with each other. Provided, therefore, that they maintain logical integrity, analytical thoroughness and internal consistency, economic theories are rarely right or wrong. Instead, they are more or less useful in our understanding of economic realities at different aggregation levels. Their predictive capacity is also a matter of debate and verification. In turn, the interpretative strength of theoretical reasoning in economics has to be judged in a relative context. The latter reflects the acknowledgment that deviations of such reasoning from real life phenomena and conditions are to be expected in view of the very abstractions and simplifications inherent in theoretical constructs. In extreme cases, though, the assumptions posed in particular analyses could call for such a distortion of evidence or of stylized facts, so as to fit them into preconceived theoretical models, that conclusions reached on the basis of theory completely contradict observed realities. Such examples have been encountered in a significant part of mainstream economic growth theorizing. In these situations, the burden of proof is shifted and put upon the content of reality so that it be consistent with theory. Otherwise it is ignored and negated or, at the very least, questioned as far as its empirical reckoning is concerned. Even more, though, by organizing thoughts and concepts, the interpretative value of economic theories can be judged by the degree and persistence of the influence they exercise in shaping our own perception as to what constitutes reality. This influence on perceived reality about often quite complex, and not only complicated, matters is an issue of fundamental scientific as well as of political interest. The latter has special significance since widely held economic ideas, often derived from theoretical constructs, can assume a major role in either defending vested interests or in shaping historical changes. Consequently, even theories which plainly and totally contradict facts have a special interpretative relevance to the extent that they lead us to believe that reality is completely different from the available evidence or from alternative interpretations we have at our disposal. Economic theory is not, therefore, sociopolitically neutral. Instead, it influences perceptions, attitudes and hence, preferences. Furthermore, it shapes or, at least, it serves as a rationalization of and it often influences crucial decisions and commitments in economic policy making. 7

8 An area of major theoretical interest in economic reasoning has to do with the way distinct schools of thought address issues and interpret the underlying processes and main determinants of socioeconomic reproduction. Such a reproduction of socioeconomic space and its path over time takes the form of four inherently interdependent expressions in terms of performance outcomes: the degree of sustainability required so as to ascertain, through consistency and stability conditions, the viability of a socioeconomic system, growth in pivotal performance fronts at various levels of aggregation, issues of distribution of income and of wealth as well as of relative access to opportunities, resources and power, economic or otherwise, and transformations leading to development and the advancement of capabilities. In the following pages we will focus mainly on issues which address growth concerns in economic theory. Nevertheless, in view of the highly interdependent nature of all four expressions of economic reproduction noted above, selected references will also be made to sustainability, distributional and development issues. In all such cases the main concern will be on how these three highly interrelated expressions of socioeconomic reproduction affect paths of longer term aggregate growth. 8

9 INTRODUCTION Over the longer run, empirical evidence, of diverse origins and reliability, tends to confirm that growth rates and their path over time register significant differences between countries and sectors 1. At the same time, growth performances over the past 200 years have, on the whole, meant that economies increased their output more than before yet, very often, with increasing disparities in per capita income. Diverging growth paths have been evidenced while catching-up processes were also noted with very few examples, though, of forging ahead from others 2. In interpreting such diverging performances among and within economic systems, the central aim in growth theorizing in the industrially more advanced economies has focused on questions of formalizing the causal factors and explaining the reasons behind some widespread critical economic phenomena. The latter involve longer term tendencies of rising output per worker, growing overall capital intensity, rising average real wages over long time spans yet with important variations and inequalities in the process, together with a tendency for relatively more constant rates of return to capital. For economies with relatively less advanced industry, additional preoccupations of a more structural nature have also been incorporated in the theoretical armor. These place relatively more emphasis on the determinants of structural economic transformation and the causal parameters for diverging/converging development paths. The scientific challenges which need to be confronted so as to address adequately and interpret the intertemporal diversity of such continually changing aggregate economic growth performances has led growth theorists permanently to pose some fundamental questions. Such questions have to do, to use Abramovitz s terminology, with the proximate sources of growth over time together with the deepest causes of growth 3. The former involve processes of accumulation and the contribution of factors of production (what John Stuart Mill called the 1 See, for example, Madison, A. (1991), Monitoring the World Economy , OECD Development Center Studies, Paris. 2 For example, Britain overtook Holland in the 18 th century, the US and Germany overtook Britain and others in the late 19 th and early 20 th century, and Japan surged ahead in the second half of the 20 th century. See Dosi, G., Freeman, Chr. And Fabiani, S. (1994), The Process of Economic Development: Introducing Some Stylized Facts and Theories on Technologies, Firms and Institutions, special issue of Industrial and Corporate Change, Oxford University Press, p.p See Abramovitz, M. (1989), Thinking about Growth, Cambridge University Press, Cambridge, UK. 9

10 elements of production 4 ) as well as the forces which shape their productivities. The deepest causes of growth address the impact of attitudes and aspirations, of incentives and conduct since all these factors have a determinant influence on the accumulation and performances of the proximate sources of growth. Critical in such performance patterns is the role of institutions and the effective introduction and implementation of key policy initiatives 5. In response to these enduring questions, though, economic theorizing has provided changing and highly divergent answers. Such differences arise not simply in view of the distinct analytical tools being used and the varying empirical references being employed. Instead, they originate from much deeper and quite dissimilar critical assumptions and methodological approaches. The latter are based on conceptual foundations which differ in fundamental ways among them and refer to quite distinct paradigms in their attempt to interpret processes of socioeconomic reproduction and change. Therefore, theoretical differences might have a lot to do with the divergence noted in perceptions about growth processes and their main determinants. Consequently, such differences could turn out to have a critical role in evaluating policy options and, in the final analysis, in influencing political reasoning. Three epistemologically distinct approaches have dominated the relevant economic interpretations as to how societies aggregate production and income performances grow over longer time periods. In addition, such interpretations combine conclusions about convergence/divergence patterns in per capita income and/or other key distributional indicators. The three scientific approaches refer to (a) formal aggregate growth modeling, (b) evolutionary growth economics constructed on well-founded inputs from what has been referred to as appreciative theorizing, and (c) interpretations based on descriptive or quantitative economic history. Explaining the main differences and areas of possible complementarity between the three above mentioned key scientific approaches in growth theorizing constitutes a major challenge in reviewing the state of our understanding as to how socioeconomic systems grow and reproduce themselves over time. Such concerns address issues whose influence extends beyond oscillations in economic performance to be expected during the course of business cycles. More specifically: I. Formal aggregate economic growth modeling, as dominated for decades by neoclassical economic formulations under very limiting basic assumptions, related increases of output to two main explanatory headings in the context of moving general equilibrium patterns: 4 See Mill, J. S., Principles of Political Economy, Ashley Edition, p For the role of such determinants of growth in the S.E. Asian experience during past decades see Wade, R. (1990), Governing the Market: Economic Theory and the Role of Government in East Asian Industrialization, Princeton University Press, Princeton, NJ. 10

11 a process towards optimum allocation of inputs, as captured by production functions which express very particular causal input-output relations; and a time drift which is supposed to represent all the action with which economic historians are concerned and which is referred to as the technological level or, in more generalized terms, the total factor productivity of a society and which, in traditional growth accounting terms, appeared as the all too important residual. Three decades after the initial formulation of the Solow-Swan model 6, formal neo-classical growth theory had reached the point of facing not only the continued and strong criticism by economists who employed a different theoretical approach or by economic historians. More than that, this particular vintage of growth theorizing also had come to confront the severe skepticism about its relevance by mainstream theoreticians of the neo-classical tradition. For example, in the late 1980 s a leading contributor to traditional formal modeling analysis acknowledged that: Neo-classical growth theory is not a theory of history. In a sense it is not even a theory of growth. Its aim is to supply an element in an eventual understanding of certain important elements in growth and to provide a way of organizing one s thoughts on these matters 7. Such an acknowledgment assigns a very modest role indeed to this line of growth theorizing as compared to the way established economic orthodoxy used to view the significance and general applicability of neoclassical thinking in this field. The growing dissatisfaction with traditional formal neo-classical growth theorizing led, only fifteen years ago, to the emergence of new formal modeling approaches. In most cases, again within a moving general equilibrium framework, these approaches attempt to: endogenize the time drift by modeling technological progress, often as a result of an optimal allocation process8, and take into account the impacts of economies of scale, of varying competitive or oligopolistic/monopolistic conditions as well as of the dynamics of linkages and networking in provoking important externalities, particularly those which result from knowledge spill- overs as well as from quality improvements in human capital and in intermediate inputs. 6 See Solow, R. M. (1956), A contribution to the theory of economic growth, Quarterly Journal of Economics, No. 70, p.p Also Swan, T.W. (1956), Economic growth and capital accumulation, Economic Record, No. 32, p.p Hahn, F. (1987), Neo-classical growth theory, in Eatwell, J. et al (eds.), The New Palgrave: a Dictionary of Economics, vol. 3, Macmillan, London, p For an early and clear statement of this growth modeling cum equilibrium approach to technological progress see Lucas, R.E.B. (1988), On the Mechanisms of Economic Development, Journal of Monetary Economics, 22, p.p

12 These new modeling approaches prompted a revived interest and dynamism in the formal growth literature. It is of interest to acknowledge that such approaches often borrowed and put into a modeling framework ideas and empirical evidence from innovation studies. The latter were for some time marginalized from mainstream economics and were rarely cited in the new formal literature on growth which followed 9. II. The thrust of theoretical interest in evolutionary economic theorizing 10 commences from a completely different conceptual challenge, namely the need to interpret and to model the functioning of economic systems under conditions of continuous disequilibrium. In turn, the corresponding areas of analytical concern focus on the complex interactions existing between: (i) random elements (perturbations) in economic systems, (ii) mechanisms which systematically winnow (separate and select) extant variations through productivity/profitability differences and the market survival of the fittest, and (iii) inertial forces which induce a path of continuity and of change in the selection process 11. Recognizing that overall economic performances depend on the prevailing variations beneath the aggregate level and that in a market economy changes take place in the context of continuous disequilibrium, evolutionary theorizing is often based on an analysis of micro and meso-economic phenomena. At more aggregate levels, the evolutionary nature of change and of selection practices is formulated, as in the case of the new growth models, on the basis of innovation related processes. Here, though, innovation processes are explicitly addressed with respect to concerns about the coevolution of technology, of industry structures and of key supporting institutions. The relevant analysis concentrates on three thematic areas 12 : first, incentive structures affecting the performance of economic actors, second, the growth implications of competencies, strategies and of organizational capabilities, and thirdly, the 9 See Sala-I-Martin, X. (2001), Fifteen Years of New Growth Economics: What Have We Learnt?, mimeo, Columbia University and Universitat Pompeu Fabra. 10 For an early statement of the pioneering work undertaken by Nelson and Winter during the 1960 s and the differences noted if compared to neo-classical growth theorizing see Nelson, R.R. and Winter, S. W. (1974), Neoclassical vs evolutionary theories of economic growth, Economic Journal, No. 84, p.p For a more comprehensive statement by the same authors Nelson and Winter (1982), An Evolutionary Theory of Economic Change, Harvard University Press, Cambridge, Mass. 11 See Nelson, R. R. (1995), Recent Evolutionary Theorizing about Economic Change, Journal of Economic Literature, vol. XXXIII, March, p.p See Dosi, G, Freeman, Chr. And Fabiani, S. (1994), The Process of Economic Development: Introducing Some Stylized Facts and Theories on technologies, Firms and Institutions, special issue of Industrial and Corporate Change, Oxford University Press. 12

13 critical relevance of the functioning of institutions since the latter guide economic coordination and change at all levels. As a consequence, such analysis explicitly recognizes the fact that economic agents and their behavior are embedded in institutions 13. In turn, the latter s evolution is strongly path dependent. Furthermore and in contrast to formal growth theories which tend to be organizational-free, innovation economics rely heavily on the determining influences of organizational structures 14. In this respect, governance concerns (corporate and/or public) affect incentives and information flows. Such factors have a major impact on decision making and on conduct related to growth performances. On the other hand, the modeling of institutions the latter being a critical element of innovation economics is still at a rudimentary stage despite the contributions of appreciative theory and of evolutionary economics. Considerable assistance, though, has also been provided during recent decades by the extensive use of simulation models supported by improvements in computer processing technology. III. Finally, economic historians stress the joint and/or co-evolutionary relevance of key variables and of processes as these affect both the quantitative performance and the qualitative development patterns of societies. Contrary to the methodological approaches of formal growth models which base their interpretations on generalized, invariant and often uni-directional causalities captured by production functions, economic historians focus on two-way causal relations. In terms of growth dynamics such multidirectional causal relations have come to address and acknowledge in explicitly historical terms the fundamentals of technological and organizational innovations. The determinants of these innovations and their productive use are, in turn, critically related to country specific characteristics and the particularities of their institutions and of circumstances as they evolve over time. Some of these characteristics are directly or indirectly linked to the functioning of economic systems and related indicators (e.g. capital accumulation, educational attainment, etc.). Others involve political evolutions and broader 13 On the meaning of embeddedness see Granovetter, M. (1985), Economic Action and Social Structure. A Theory of Embeddedness, American Journal of Sociology, 19, p.p See Chandler, A.D. (1990), Scale and Scope: The Dynamics of Industrial Capitalism, Belkmap Press of Harvard University Press, Cambridge, Mass. 13

14 social phenomena. Therefore, economic historians adopt a much more comprehensive scope in their analysis and acknowledge the relevance of multi-directional causalities 15. This more holistic approach enables economic historians to focus their interpretations on a wider spectrum of phenomena that might otherwise escape adequate explanations, especially in formal growth theorizing. On the other hand, there exist an almost infinite multitude of events which have yet to be classified, described and analysed, a task which makes a brooder framework of interpretations inevitable. Drawn from historical evidence, effort is therefore invested in putting forward more comprehensive explanations so as to interpret the causal reasons behind varying economic performances. For example, such performances might involve the relative decline experienced by some economies or the economic catching-up and leapfrogging in other more successful experiences. Very exceptionally, special instances exist in which economies not only catch-up but also push-ahead from others and achieve technological leadership and higher living standards. In the sections which follow we begin with a brief, critical review of the history of economic growth theorizing that preceded the three epistemologically distinct approaches referred to above. Subsequently, each one of these three methodological approaches is presented more fully with an emphasis on illuminating their main intellectual contributions and limitations as well as on making explicit areas of complementarity and of contradictions between them. We conclude with some key references to the underlying reasoning in the interplay between economic growth and distributional considerations. 15 For a seminal work on these matters see Landes, D. (1969), The Unbound Prometheus, Cambridge University Press. Also see Rosenberg, N. (1976), Perspectives on Technology and by the same author (1982), Inside the Black Box, both by Cambridge University Press. 14

15 A. THE INTELLECTUAL ANTE-SALA OF MODERN GROWTH THEORIES Up to the first half of the 19 th century, differences in (very approximate and of varying consistency) estimates of per capita income levels between now developed economies and representative groups of presently less developed ones were reported as not being all that significant 16. A major inflection point, though, came with the first industrial revolution. This radical change provoked major differences in the technological base of a number of Western European economies as compared to the rest of the world at that time. In view of such technological differentiations which transformed the very foundations of productive systems and of institutions in distinct societies, increasing divergences begun to be registered in aggregate growth rates as well as in per capita income levels. Experiencing at first hand such profound socioeconomic transformations provoked by the first technological/industrial revolution, classical political economists presented highly diverging interpretations about the processes by which economic systems evolve and reproduce themselves over time. Despite such differences, though, they all shared a common fundamental characteristic: they viewed economic growth, as well as the existence of limits to such growth, as a natural process 17. Adam Smith s perception of growth presented few preoccupations: People will make appropriate choices on the basis of self-interested behavior, markets will guide resource allocations by providing rewards and penalties while the invisible hand will enhance general growth and augment the wealth of nations. Deviations from this path will take place when interventions, by man and politics, get in the way 18. Distributional issues also turned out to be viewed in a complacent form. Of the three factors of production which were considered during Smith s period (namely land, labour and capital), the latter two were interpreted as having sufficient mobility and homogeneity so as to lead, in the longer run, to equalizing tendencies in the performances of distinct economic spaces. 16 See relevant comparative statistics in Bairoch, P. (1981), The Main Trends in National Income Disparities since the Industrial Revolution, in Bairoch, P. and Levy-Loboyer, M., Disparities in Economic Development since the Industrial Revolution, MacMillan, London. 17 See relevant exposition in Landes, D. S. (1991), On Technology and Growth in Higonnet, P., Landes, D. S. and Rosovsky, H. (eds.), Favorites of Fortune: Technology, Growth and Economic Development since the Industrial Revolution, Harvard University Press, Cambridge, Mass. 18 For explanations along these lines as to why China s economy had entered a stationary state during the latter half of the XVIII century in contrast to the take-off of the British economy, see 15

16 Differences will arise, nevertheless, due to the unequal availability of fertile land. This, though, was God s work and not a matter for policy making. About two centuries later and after the intervention of two additional technological/industrial revolutions, mainstream neoclassical growth thinking maintained largely unchanged the basic premises of Adam Smith s interpretation. These include the convergence and/or equalizing tendencies of the growth creating process through the pursuit of a natural steady state longer term growth path together with an overall lack of concerns about distributional outcomes. Also, a limited role is attributed to growth inducing policies provided that two general conditions are met: the unleashing of competitive market conditions and the assurance of macroeconomic stability. Contrary to Smith s optimism regarding long term growth, other classical thinkers were distinctly more pessimistic. They were the first to express views on limits to growth. This theme resurfaced again more than a century later, mainly in the prewar period and during the 1970 s as well as more recently. The latter has mainly to do with critical environmental concerns and the requirements posed by sustainable development. In the early versions of limits to growth, Malthus based his conclusions on the effects of longer term tendencies in population growth. Such tendencies will eventually, according to Malthus, surpass output growth, thus suppressing subsistence levels. In turn, this will provide a check on population explosion. Therefore, in this context convergence patterns were founded upon the attainment of basic subsistence levels derived from the relation between population growth and food production growth. Ricardo had a different interpretation. He viewed limits to growth as originating from the constraints imposed by the availability of fertile land. His argument was that, as demand for food increased, countries will be forced to seek less fertile lands. This will raise the cost of food production as well as of food prices which, in turn, will increase wage costs (higher consumption wage ). Consequently, profit rates, especially in industry, will eventually tend to fall, thus reducing the availability of investible resources and of productive capital. Therefore, a stationary state will be bound to emerge. Ricardo overlooked, though, the opportunities arising in his time due to novel technological means as these led to the incorporation of new land rich territories. The latter were made accessible through radical technological evolutions in transport and communications. Also, further progress in productive know-how in other fields throughout the last two centuries proved significant in greatly improving productivity levels in existing land usages. Such developments Smith, A. [1776], (1937), An Inquiry into the Nature of the Wealth of Nations, Modern Library, New York, p

17 overshadowed the importance of limits to growth imposed by constraints on food supplies whose production was based on previous generation techniques. Instead, emphasis was to shift eventually on issues concerning entitlements and capabilities, as explained much later in the seminal work of Amartya Sen. In fact, actual growth performances contradicted the pessimism of these classical thinkers. Annual output and income growth rates in countries like Great Britain were reported to have increased from a fraction of one percent during the period to more than one percent from 1780 up to the turn of the century and by two and a half percentage points in the period 19. Thus, an important kink in the growth curve took place at that time, indicating new conditions and novel ways of organizing production and overcoming its limitations 20. Marx s materialistic conception of history interprets the process and paths of economic growth in the context of the inherent tensions between forces of production, relations of production and superstructure. In this respect, Marx proved to be an important forerunner of subsequent theorizing on relations between technological and institutional change. The driving forces behind economic growth express, according to Marx, the continuous pursuit of capital accumulation through the rising organic composition of capital. This follows as a natural tendency of the capitalist system in order to assure and extend control over the means of production. Yet, and this proved to be an important interpretative limitation, technological innovations and their application in productive activities were perceived almost as being automatic and as a direct offspring of science. Drawing from the latter s advancements, technological progress evolved without calling for any further particular analytical treatment on special institutional or policy issues 21. Marxist theoretical extensions at the international level focused on distributional issues. These involved the understanding of the monopoly phase of capitalism with further extensions elaborated later (e.g. by Sweezy-Magdoff) on non-price competition leading to a clash between transnational enterprises for control of markets world-wide. Furthermore, distributional issues were examined between countries as different per capita growth patterns included increased 19 Crafts, N. F. R. (1989), The Industrial Revolution: Economic Growth in Britain, , in Digby, A. and Feinstein, C. (eds.), New Directions in Economic and Social History, Lyceum, Chicago. 20 Landes, D. S. (1991), On Technology and Growth in Higonnet, P. Landes, D. S. and Rosovsky, H. (eds.), Favorites of Fortune: Technology, Growth and Economic Development since the Industrial Revolution, Harvard University Press, Cambridge, Mass, p It is, therefore, generally the most worthless and miserable sort of money-capitalists who draw the greatest profit out of all new developments of the universal labour of the human spirit and their social application through combined labour. See Marx, Karl [1894], 1962, Capital, vol. I, sec. 5, Modern Library, New York, p. 103 and analyzed in Boder, M. M. [1927], (1968), Karl Marx s Interpretation of History, 2 nd edition, Harvard University Press, Cambridge, Mass. p.p

18 income levels from work in some capitalist economies supported by backwardness and mass poverty for others. Thus, growth patterns and distributional outcomes, both in the domestic economy as well as internationally, were the natural outcome determined by the internal workings and contradictions of the capitalist system as it evolved over time. Extending from the conclusions of classical political economists, pessimism about longer term growth returned again several decades later through the writings of the stagnationist in the 1930 s. Their interpretations on the persistence of unemployment, underinvestment and, mainly of the exhaustion of the capacity of technological corrections to reverse conditions of stagnation were, nevertheless, buffled by the war experience and the reconstruction which followed. At the same time, the widely influential Keynesian contributions on aggregate demand management provided a conceptual base and a policy platform to confront shorter term growth requirements through multiplier effects. These developments in managing shorter term macroeconomic performances effectively left growth models free to return to longer term preoccupations. The latter reintroduced concerns about the supply side of inputs and their productivity determinants. Such inputs were introduced in a variety of production functions in order to explain the process of achieving expected output attainments. The underlying assumption of all the corresponding aggregate modeling efforts had to do with the expectation that growth was universally achievable. This was to be attained not only through the application of novel techniques 22. It was also to be advanced by the overall upgrading of the sectoral composition of an economy. Such restructuring created opportunities for significant productivity increases with corresponding income implications and distributional effects 23. These processes were even presented by some as following an almost biological and deterministic progression of development stages 24. Thus, in such cases the quest for the sources and determinants of the wealth of nations lost intellectual appeal. In the immediate pre and post Second World War years, empirical research concentrated on reviewing issues of classical thinking on growth while generally lacking stimuli from more 22 For a review of the relevant formal literature see Hahn, F. H. and Matthews, R. C. O. (1964), The Theory of Economic Growth: A Survey, Economic Journal, 74, no. 296, p.p See Kuznetz, S. (1965), Economic Growth and Structure: Selected Essays, Heineman, London. For more recent elaborations on these issues with particular reference to their income distributional implications see Jha, S. K. (1996), The Kuznets Curve: A reassessment, World Development, vol. 24, no. 4. Also see Bruno, M., Ravaillon, M. and Squire, L. (1996), Equity and Growth in Developing Countries: Old and New Perspectives on the Policy Issues, World Bank Policy Research Paper, No. 1563, Washington, DC. 24 Rostow, W. W. [1960], (1971), The Stages of Economic Growth, Cambridge University Press, Cambridge, UK. 18

19 novel formulations 25. Major exceptions, though, also existed and originated from the works of Schumpeter and Kuznets who provided path-breaking insights on growth 26. The latter was perceived by these two pioneering authors as basically expressing disequilibrium processes. Such an approach, though, meant a departure from the prevailing economic orthodoxy. Backed by neoclassical rigour, the mainstream of economic thought had the completely opposite orientation based on premises about equilibrium trends and steady-state longer term paths. Similarly, a small number of individual researchers foreshadowed analyses which were to appear three or more decades later as to why factor accumulation could explain only a fraction of output growth. Such contributions to growth analysis explored important concerns on technical advancements, reallocation of resources from lower to higher productivity activities, changes in the composition of the work force and economies of scale 27. Looking back over the whole period, Chr. Freeman reminded us more recently that the routes of the serious divide in the economics profession over the issue of the role of history in economics goes back to the Methodensteit debate in Germany more than one hundred years ago 28. The neo-classical school under the leadership of the Austrian Carl Menger, stressed the need to apply logical deductions based on abstract models and test them utilizing empirical data in order to reach broad generalizations. The historical school, led by Gustav Schmoller, questioned the adequacy of broad generalizations in view of the uniqueness of evolutionary processes and the particularities of socioeconomic systems over time. Interestingly enough, Schumpeter s attitude in his early years appeared to be equivocal, if not siding with the neoclassical school. This attitude was reflecting his life-long inability to cut the Walrasian umbilical cord and at the same time his recognition of the crucial importance and complexity of evolutionary change 29. The Methodensteit ended in almost total worldwide victory for the neoclassical theorists. Yet, seen in the longer term, strong indications exist that this proved to be a pyrrhic victory: fundamental problems have been shown to exist within the mainstream models. Leading neo- 25 See critical review in Abramovitz, M. (1952), Economics of Growth, in Haley, B. (ed.), A Survey of Contemporary Economics, vol. II, Irwin Inc., Rotterdam p.p For early pioneering studies on economic growth linked to phenomena of long swings see Kuznets, S. (1930), Secular Movement in production and Prices, Houghton-Miffling, Boston. Also see Burns, A. F. (1934), Production Trends in the United States since 1870, National Bureau of Economic Research, New York. 27 For example see Schmookler, J. (1952), The Changing Efficiency of American Economy, Review of Economics and Statistics, 34, p.p Also, see analysis and references in Kendrich, J. W. (1956), Productivity Trends: Capital and Labor, Review of Economics and Statistics, Aug. 38, p.p See Freeman, Chr. (1995), History, Co-evolution and Economic Growth, IASA Working Paper, WP-95-76, Laxemburg, Austria. 29 Idem at 4. 19

20 classical growth analysts have themselves acknowledged such severe limitations and inadequacies in the models they advanced over preceding periods 30. In exploring the nature of the mainstream s pyrrhic victory, we need to review three waves of growth modeling efforts. It is noteworthy that each successive presentation was marked by a fundamental departure from the underlying assumptions of the previous generation of models. It turned out that the core of the main conclusions in each case were critically dependent on these assumptions. In other words, the latter proved decisive as to what each generation of models had to say about the growth process. Thus, although a great deal of intellectual effort was invested in model building and the functional technicalities of the different models, the essence of their interpretative capabilities and relevance was hidden in the underlying assumptions upon which each particular modeling approach was based. We now turn to the three generations of modern formal growth theorizing which refer to: 1. The savings-capital accumulation nexus to growth. 2. Traditional neoclassical growth models. 3. The new formal growth interpretations and extensions/rebuttal of the neoclassical thought in mainstream formal theorizing. 30 See Hahn, F. (1987), Neo-classical Growth Theory, in Eatwell, J., et al (eds.), The New Palgrave: A Dictionary of Economics, vol. 3, Macmillan, London, and by the same author (1991), The next hundred years, Economic Journal, vol. 101, p.p Also see, Arrow, K. (1994), Problems mount in application of free market economic theory, Guardian, Jan

21 B. FORMAL GROWTH MODELING B.1. The savings-capital accumulation nexus to growth: A Keynesian approach to growth under primitive assumptions on economic dynamics Centered on the savings-capital nexus of macrodynamics, the starting point in modern formal growth theorizing is to be found in the Harrod-Domar model formulated at the end of the first half of the 20th century31. Its policy application was claimed to be relevant for economies experiencing a long transitional phase of high structural unemployment. This was not an unlikely situation at the time when the model was first formulated given the prevailing conditions in the advanced industrialized countries during the 1930 s. More recent experiences, though, in both more and lesser developed economies have brought again into the forefront the harsh realities of large open unemployment and structural high underemployment32. Thus, despite serious limitations noted below, this early growth model deserves a closer attention due to the proximity of its policy implications to key socioeconomic needs which have resurfaced, within a quite different context, during more recent years. Given the presence of excess labour supply relative to the aggregately available employment opportunities, the main overall conclusions drawn from the Harrod-Domar model link the pursuit of high aggregate output performances to a number of specifically identifiable macro conditions. These can be reduced into four main policy fronts conducive to high growth: high savings/investment rates; high capital productivity; high rate of technical progress which, according to the model, is embodied in tangible capital goods; and as far as aggregate per capita income growth patterns are concerned, a low population growth. It is noteworthy that all of the above determinants of aggregate output growth rates have, under a diversity of formulations, been also present in a number of non-neoclassical approaches. 31 See Harrod, R. F. (1948), Towards a Dynamic Economics, Macmillan, London and Domar, E. (1947), Expansion and Employment, American Economic Review, March 37:1, p.p

22 During more recent periods, these seek to identify causal relations between key macroeconomic parameters and aggregate output performances, namely, the profit-investment nexus, productivity contributions to output growth, embodied and disembodied technological progress and issues on population policy management. A notable limitation, though, of the model had to do with its exclusive concentration on tangible capital as the main source of growth. It has correctly been argued that the contribution to growth by capital is a function not only of its quantity but also and mainly of its direction and quality, of the skills of both entrepreneurs and of labour in putting new investments in productive use, of the presence and quality of social overhead, etc.33. More severe criticism, presented as early as four decades ago and arising from detailed historical work, [reached] the conclusion that the accumulation of capital is in itself by no means the central aspect of the process of economic growth 34. Returning to the model and the key lessons to be drawn from its use, growth performances in an economy during the transition period of high structural unemployment will seem to follow a set trajectory. In summary form, this trajectory follows a set pattern which, in turn, is based on the model s assumptions35. Suppose that: i. aggregate output is proportional to the stock of physical capital, a. ii. realized savings and investment are also proportional to output and income, s. Then, by simple calculations it can be deduced from the above that, since investment is proportional to output (s) and there is an exogenously set capital/output ratio (1/a), the trend of growth of both capital and output is fixed at the rate sa, provided that capital utilization does not vary widely36. Now suppose also, that: iii. labour productivity is rising, for technological or other exogenous reasons, at the rate m. This means that the labour input per unit of output is falling at the 32 For the early seminal work on this subject and its growth and development implications see Lewis, W. A. (1954), Economic development with unlimited supplies of labour, Manchester School, vol. 22, p.p See Freeman, Chr. (1995), p Supple, B. (1963), The Experience of Economic Growth: Case Studies in Economic History, Random House, New York, p See a critical and succinct presentation of the main elements of the Harrod-Domar model in Solow, R. M. (1994), Perspectives of Growth Theory, Journal of Economic Perspectives, vol. 8, no. 1, winter, p.p If the exogenously determined technological change affects capital productivity, as Kaldor hypothesized subsequently in his vintage capital presentation, and this takes place at an average annual rate of?, then the output growth rate will be equal to (sa +?) if full utilization of capital is maintained. 22

23 same rate, thus putting added pressure on problems of structural unemployment in a society. iv. labour force growth is set at a rate n due to corresponding population growth. As it has been amply shown in the relevant literature, the combined dynamics of the above will lead to an impasse. Unless total output rises at the knife-edge rate of (m+n), then either unemployment will continue to rise indefinitely (since output growth is too slow compared to the warranted rate) or the economy will run out of the available labour force (if growth is too fast). A few countries have confronted in their economic history the latter experience by promoting, over several decades, an active immigration policy while their aggregate output increased at very high rates. (e.g. Argentina, Australia, Canada in the late 19th and early 20th century). In terms of the Harrod-Domar model, the impasse will be reconciled only if sa=m + n. Yet, this will occur only accidentally since all four critical parameters, namely s, a, m and n, are assumed to originate from four wholly unrelated sources. Consequently, as far as model building is concerned, the basic technical problem of the Harrod-Domar approach stems from the fact that it is founded on a number of exogenous processes. The latter s different dynamics give rise to a fundamental incongruity of longer term permanent disequilibrium outcomes which, in turn, lead to unsustainable and explosive situations. In the more generalized presentations, such outcomes imply permanent unemployment of either capital or labour, except in the purely accidental case in which both factors of production grow at the same rate and with the initial situation being one of full employment at full capacity. Two different solutions to this problem were proposed in economic theorizing, both appearing in In one of these, Kaldor argued that the savings rate, far from being given, is endogenously determined in an economy. This reflects the fact that savings propensities are dependent on income levels and on income distribution. Aggregate savings can also be the subject of government macroeconomic policy initiatives. Thus, equilibrium growth can be attained in the longer run through appropriate incomes, fiscal and monetary policies. This reinforced the Keynesian character of the Harrod-Domar model with its original emphasis on investment induced aggregate growth performances. An alternative solution to the longer term disequilibrium phenomena of the Harrod-Domar formulation was proposed by the immensely influential Solow-Swan model founded on solid neoclassical thinking. It was argued that the productivity of capital, a, rather than the savings rate, s, is endogenously determined in an economy. For this to happen two additional critical assumptions were required. Both are consistent with and mutually supportive of the neoclassical interpretation regarding the organization and performance of modern capitalist societies. The 23

24 first one is of a technological nature. It views factor productivities as endogenously determined by their relative use on the basis of complementarity, thus altering the foundations of the Harrod-Domar approach. The nature of productivity performances assumed by the traditional neoclassical model under conditions of substitutability between quantifiable factors, as discussed in Section B.2 below, leads to behavioral patterns which are consistent with the second set of neoclassical assumptions, namely the presence of decreasing returns and of competitive markets. In sharp contrast to the permanent disequilibrium outcomes of the Harrod-Domar approach, the underlying logic of the neoclassical one implies a specific equilibrium process for the longer term evolution of a market based capitalist system. According to this latter approach, as capital accumulation takes place in an economy, the ratio of capital to labour grows at the positive rate of (sa n). The rising capital intensity, though, leads, according to neoclassical thinking, to a decreasing capital productivity. This brings down the rate of capital accumulation, sa, which progressively converges to the equivalent rate of labour growth, n. Consequently, in the long run, factor intensity is stabilized and aggregate output growth continues thereafter at a rate which is exclusively determined by exogenous factors, (n +?), that is by population and technological developments. The standard of living in societies, as expressed by per capita income, is improved in the long run only by what happens in the exogenously determined front of technological progress. Consequently, neoclassical thinking comes back full circle to Adam Smith s complacent conclusion that the relevancy of policy preoccupations is reduced once the steady state growth path has been attained. On the other hand and during the transition period, the critical variables, (i.e. either the savings rate (s) in the Kaldorian model or the productivity of capital (a) in the neoclassical one) are endogenously determined and thus subject to policy initiatives. Furthermore, these parameters influence the level of long run steady state equilibrium growth performance while its slope is solely determined by exogenous factors. From the theoretical point of view, this constituted the staring point of the new growth models introduced since the latter half of the 1980 s. Before turning to them, though, we need first to account for the role of neoclassical thinking in modeling growth theorizing. B.2. Neoclassical growth theories At the foundations of all formal neoclassical growth models one finds the conceptual construct of moving general equilibrium. The conditions for such an equilibrium underpin the way neoclassical theorizing interprets the very essence of socioeconomic reproduction and the resulting path of quantitative performance over time. Within this specific context, two 24