It reflects a modern concern to establish a more systematic understanding of the process of creative destruction

Growth and Retardation in Progressive Economies
10In his comprehensive review essay, Abramovitz (1989) identified structural change and a tendency towards retardation in the growth of output as two among eight salient empirical generalisations about the process of economic growth. Structural change is, of course, a necessary reflection of diversity in the growth rates of different activities in the relevant population. In any ensemble, that activity that grows more rapidly than the ensemble average will increase in relative numerical importance in that ensemble. Retardation, however, is a different phenomenon, the systematic tendency for rates of growth of specific entities or their ensemble to decline with the passage of time. To anyone brought up on the economics of uniformly expanding economies, whose structure cannot change over time and whose rate of growth is constant, neither of these propositions will have much resonance. Yet they are central to the literature to which Abramovitz is referring, in particular to the work of our two principal retardation theorists, namely A.F. Burns and S. Kuznets. [8][8]Abramovitz (1989), pp. 12-13, pp. 30-31 and p. 90. Both authors are concerned with the measurement and explanation of secular or long time movements in the volume of economic activity. Both emphasise that the introduction of new activities and the disappearance of old activities are an intrinsic part of the development of capitalism. Both understood that the evidence for retardation would depend on the level of statistical definition of an industry or sector, and that the broader the aggregate the more the evidence in favour of economic evolution will be suppressed. Accepting that the modern economic system is “characterised by ceaseless change”, neither could proceed with an aggregate analysis of growth nor accept the idea of uniform progress in all branches of activity. It is worth noting that the diversity of growth rates and the retardation of growth rates are closely connected as dynamic phenomena. [9][9]If si is the share of industry i, in a constant price weighted…
11This we can see immediately if we define the growth rate of some aggregate of economic activities between the two dates in the conventional way as g = ?s_ig_i where s _i is the share of the “i”th component in the aggregate. Then it follows that the rate of change in the aggregate growth rate is
12

Furthermore, since ds/dt = s_i (g_i ? g) it follows that we can write this as

where, V _s (g) is the variance of growth rates within the population and R is the averaged rate of change of the growth rates of the individual components of the aggregate. If R is negative we have retardation on average and if it is positive acceleration. Consequently, whenever the growth rate of the aggregate is constant, we must have retardation on average, and the average rate of retardation is numerically equal to (minus) the variance of the growth rates in the population. More generally, the difference between these two magnitudes is exactly the rate of change of the growth rate of the aggregate.
Let us turn now to the main issue and begin with A.F. Burns detailed study of American economic growth in the period 1880-1937. Like other empirically minded scholars, Burns gathered a great deal of evidence to establish that a central feature of modern economic development is the diversity of growth rates of output across different sub-sectors and commodities in the economy. What might appear to be smooth progress of production and trade in the aggregate, hides a considerable diversity of experience. It is interesting to note that Burns takes to task Gustav Cassel, and by implication modern growth theorists, for conceiving of a regularly expanding economy precisely because this necessarily rules out the mainspring of economic growth, namely an uneven incidence of technical and organisational change. His list of the factors that underpin the diversity of growth rates has a thoroughly modern ring to it. It includes the following: new commodities; new raw materials; changes in methods of production; new methods for the recovery of waste products; changes in forms of industrial organisation; increases in the number of uses of given materials and in the number of materials put to a given use; and, finally, the emergence of what he calls learning products and style goods. In sum, Burns claims that “These changes have resulted in an increasing divergence of production trends for they have served to stimulate or depress but to an unequal extent, the development of various industries” (p. 63). Furthermore, what makes an economy progressive is not diversity per se but a positive skew to the distribution of growth rates. [10][10]Glenday (1938) applied Burns’ method to long production series…
Simon Kuznets had independently explored the same themes (1929, 1954) and from a broadly similar perspective. He stated the problem clearly as follows,

“As we observe various industries within a given national economy, we see that the lead in development shifts from one branch to another. A rapidly developing industry does not retain its vigorous growth forever but slackens and is overtaken by others whose period of rapid development is beginning. Within one country we can observe a succession of different branches of activity in the vanguard of the country’s economic development, and within each industry we can notice a conspicuous slackening in the rate of increase”.

(Kuznets, 1929/1954, p. 254)
Of course, the long secular movements of the shares of agriculture, industry and service sectors in total output provide confirmation at higher levels of aggregation of the enduring presence of growth rate diversity and structural change. As do the shifting rural-urban balance of the population, changes in working hours and changes in the pattern of household consumption. Indeed the long swing of development must have been marked by as much by the transformation of preferences as by the transformation of industry. [11][11]For some interesting commentary see Baumol et al. (1989,…
However, the unevenness of growth experience is only part of the picture. For both Burns and Kuznets focused upon a further regularity in the process of restless growth, namely retardation, the persistent tendency of growth rates to decline over time from the inception of the industry. Their explanations of retardation are remarkably similar, emphasising population growth (a minor element), foreign competition, inter-industry relations of competition and complementarity and, of vital importance, technical progress. Indeed, for both Kuznets and Burns, it is retardation in the within industry rates of technical progress, which is the chief explanation of retardation in rates of output growth. Moreover, their theories of technical progress are essentially the same, an industry being created by an invention, or complex of innovations that offers scope for a myriad of improvements of ever decreasing importance. Progress inevitably slackens unless there is some radical breakthrough in the foundations of an industry’s methods as it becomes increasingly difficult to extract further improvements in performance. This is a view beautifully and subsequently expressed by Hicks (1977) who wrote in terms of the “economic children” that follow the original invention. A sequence of initial inventions creates new activities and a potential design space to explore the possibilities latent in the new concepts, and so provides the stimuli to maintain a trajectory of technical innovation over time within the limits resident in these new concepts. This is a theme familiar to all modern evolutionary minded scholars and students of innovation (Dosi, 1982; Georghiou et al., 1984). Kuznets and Burns where not the only scholars to explore these themes. Fabricant (1942) too found compelling evidence for the retardation of growth in American manufacturing output and employment over the period 1899 to 1939 although he worked in terms of broader commodity groups rather than the individual commodities that were the focus of the Kuznets and Burns studies. While Fabricant does not develop the logic of the retardation thesis, being content to echo the Kuznets/Burns line on technical progress, his study is valuable for its emphasis on structural change and the shifting balance of employment within the overall growth of the economy. Fabricant is also clear in his assessment that technical change in the broad simultaneously creates employment and destroys employment. The emergence and exploitation of new fields of employment is a factor balancing the ebb of employment in mature and decadent industries (op. cit., p. 159). Subsequent studies, after 1945, by Hoffman (1949), Stigler (1947) and Gaston (1961) further explored the empirical basis of the retardation theme in different bodies of industrial data but without any further development of the underlying theory. By then growth theory had taken its macro economic turn as the consequences of the Keynesian revolution where applied to long period problems. In all essentials, the picture of growth as an evolving transformation of an economy was lost in its entirety. The rekindling of interest in the knowledge of growth and the growth of knowledge makes its revival both timely and highly relevant to understanding the endogenous sources of growth.
It is fair to say that while the broad outlines of the Kuznets/Burns explanations of retardation and growth rate diversity are clear, a number of important details are not, and this lack of clarity may have contributed to the demise of their approach. In attempting to clarify the various issues, it is apparent that three major and distinct mechanisms are involved in their accounts. The first concerns the dynamics of demand for a new product, or, more precisely, the dynamics of preference formation in which users learn to assign positive utility value to a new commodity. If micro foundations are to be found for such a process they will not be concerned with rational allocation in the presence of given knowledge but with a process of consumer learning in which established patterns of behaviour are disrupted. Following the practice of modern evolutionary economics this process is best visualised in terms of the mutation of decision rules as new areas of consumption space are opened up and explored (Cross, 1983; Metcalfe, 2001a). Our knowledge of these preference formation processes is, to say the least, sketchy, although the empirical evidence in favour of some form of adaptive learning process, albeit one that is embedded in wider social interactions, is strong (Cowan, Cowan and Swann, 1997). As a rule, it is to be expected that rates of learning will depend on the current pattern of individual consumption and upon observations of the consumption experience of others. We must also expect that the response of consumption to variations in prices and incomes will not be stationary over time but will vary systematically and irreversibly along the lines first made famous by the concept of the Engel curve (Pasinetti, 1981, 1993).
The second mechanism concerns the growth in capacity to supply a new commodity, a mechanism in which profitability plays a central role as source of capital and as a stimulus to investment and entry, and, indeed, disinvestment and exit from an industry when necessary. Placing the profit mechanism at the heart of the growth process necessarily does the same for the processes of price and cost formation. Diversity of growth is closely connected to diversity of profitability, and retardation in growth to retardation in rates of profitability. Under the rules of capitalism, one cannot possibly comprehend growth in the application or consumption of new goods without taking account of the central role of profitability in developing the capacity to supply.
The final mechanism involves the development of the technology with product improvements extending existing and opening up new areas of demand, and improvements in the production process reducing unit production costs. Retardation of technical progress was at the centre of the explanation of retardation of output. Yet it is clear that the Kuznets/Burns explanation of the slackening of technical progress is focussed too much on the supply side of the problem and on the analogy of a technological opportunity as a given mineral vein to be progressively exhausted at a declining rate. For one, this makes the rate of technical retardation exogenous with respect to the rate of output retardation. For another, this wholly neglects the demand side of the technical progress problem. It is only with the work of Schmookler (1966, 1972), that the Kuznets/Burns line of causation is questioned. Schmookler consistently argued for a demand side interpretation of retardation in terms of a declining price and income elasticities of market demand, and that retardation of technical progress is not to be found in the increasing cost of invention but in the declining value of invention (1966, p. 204). However, there is no need for these to be mutually exclusive explanations; a little Marshallian logic suggests that some recognition needs to be given to the interrelation between the two blades of the invention and innovation scissors. Such a deeper sense of understanding is provided in the work of Allyn Young (1928) to which we have already referred. Drawing on roots in Smith and Marshall, he articulated the view that the extension of the market causes and is caused by the exploitation of new technological opportunities. We shall suggest below that this is precisely the insight needed to capture one possible link between retardation of output growth and the slackening of the rate of technical change. Of course, the scope of Young’s argument was much broader than the linking of growth of market and technical progress within a single industry. What mattered was the reciprocal dependence between activities in which “inventions” in one sphere initiate “responses elsewhere in the industrial structure which in turn have further unsettling effect” (op. cit., p. 532). For Young, as for Schumpeter, the idea of capitalism in equilibrium did not hold much appeal. Yet, it is sufficient for our purposes to explore the more limited version of Young’s central theme and confine our attention to the sequences that flow from the creation of a new activity and its associated markets.
Taken together, demand oriented learning, the accumulation of capacity and the rate of technical change provide the basis for an ecological dynamic of industry growth wholly appropriate to an understanding of restless capitalist economies. Each interacts with the others and all three are firmly located within the market process so that the nature of their interaction depends on that process of market co-ordination. How they interact and how they are linked to the idea of retardation is the focus of the rest of this paper.