The Problem of Increasing Returns

Shove always emphasised the extreme complexity of the economic system and the need for analytical tools to cope with such complexity. An illustration of his approach to economics is his contribution to the Economic Journal's 1930 Symposium on increasing returns and the representative firm.

The main issue discussed in the Symposium was the compatibility of the theory of competition with the existence of significant internal economies of scale that give rise to increasing returns. As Robertson (1930) put it, both Pigou and Sraffa were convinced that competition was incompatible with internal economies, but the former responded by essentially ignoring internal economies, while the latter made recourse to the theory of monopoly. As for Shove, he argued that competition and internal economies could be compat?ible, although he did not make the traditional distinction between internal and external economies and considered a kind of competitive equilibrium that was different from equilibrium under perfect competition.

Shove (1930: 94-116) started his analysis with a long quotation from an unpublished study of his.^[1] The quotation is concerned with those characteristics of long-period equilibrium that, for him, Marshall wanted the notion of the representative firm to display. The most important of these characteristics are: (i) if an entrepreneur has a special aptitude for the work in which he is engaged, his earnings are higher than those he could get anywhere else; (ii) the actual earnings of individual firms may also differ from one another due to random causes; (iii) a firm’s profit varies over the different stages of its life; and (iv) equilibrium for the industry as a whole does not imply equilibrium for all individual firms operating in it.

For Shove, however, the use of Marshall’s representative firm is appropriate only to describe equilibrium as the outcome of the rise and fall of individual firms. Marshall’s representative firm is not useful when, as in Shove’s article, equilibrium is considered as ‘arising from the ebb and flow of all kinds of resources from one occupation to another’ (ibid.: 96). In this case, a different approach is required, and Shove proceeded to his description of equilibrium.^[2]

Long-period equilibrium in an industry is realised when it is not profitable for any agent to modify the existing allocation of resources among firms—in other words, when it is not profitable either to concentrate resources in a smaller number of firms or to distribute them over a larger number of firms. This sort of equilibrium position is naturally affected by the initial distribution of resources and, hence, there exists an infinite number of equilibria, each corresponding to a particular initial distribution (ibid.: 98). This approach to equilibrium makes it possible to allow for any degree of heterogeneity of resources without any need to aggregate them. For Shove,

the economic problem presented by the real world seems to be much more a question of sorting out and fitting each into its appropriate niche a vast number of heterogeneous individuals and activities than of regulating and directing into the proper channels large homogeneous streams of standardised productive agents: a jig-saw puzzle rather than a problem in hydrodynamics (ibid.: 99).

The more traditional approach to equilibrium, based on the equality between supply and demand at the industry level, encounters a significant difficulty. In order to construct the supply curve of a certain good, it is necessary to construct the supply curves of all the factors of production that, to be aggregated, must necessarily be measured in quantitative terms, that is, independently of their value. But all attempts to find this measure have failed (ibid.: 99-100).^[3]

Shove then considered the ‘more difficult matters’, that is, the issue of increasing returns and their compatibility with competition. Having defined internal and external ‘economies of large-scale industry (ibid.: 104; italics in original)^[4] and pointed out that the expansion of an industry as a whole can give rise to both types of economies in individual firms, he distinguished between ‘(1) the improvements in organisation which are made possible or profitable by an increase in the total output of the industry as a whole, and (2) the improvements in organisation which an individual firm...would obtain by increasing its own output while that of the industry as a whole remains unchanged’ (ibid.: 101).

Shove used the term ‘economies of large-scale industry’ for the effects that the expansion of the industry has on the productivity of individual firms’ resources, and ‘economies of individual expansion (ibid.: 104; italics in original) for the effects of an increase in the size of a single firm, the size of the industry as a whole being given. He regarded this distinction as more useful for trying to ‘unravel the relations between the size of an industry and the efficiency of the businesses which compose it’ (ibid.).

Equilibrium is compatible with both economies of large-scale industry and of individual expansion. But it is incompatible with some types of economies of large-scale industry, which Shove called economies of concentration, that is, those advantages that derive from an increase in the average size of the business and can be enjoyed at the given degree of knowledge and intelligence of the operating entrepreneurs. Since, in equilibrium, no redistribution of resources is possible, economies of concentration must be ruled out (ibid.: 104-105).^[5]

Having introduced this classification, Shove went on to consider the compatibility of economies of individual expansion with competitive equilibrium. If these economies exist and firms do not reap these advantages, there must be obstacles that prevent them from doing so. The two most important obstacles are increases in the costs of both transport and marketing, which a firm would bear if it tried to advance into its competitors’ markets (ibid.: 105). Such obstacles make equilibrium compatible with increasing returns due to the expansion of individual firms. These obstacles, on the other hand, would not exist if the industry as a whole expanded. In this case, in fact, single firms would not have to struggle to conquer a market share belonging to others.

The notion of competitive equilibrium adopted by Shove is not that of a perfect competitive equilibrium,^[6] but he also considered the case of perfect competition.^{[7] [8]} Again, equilibrium would be compatible with increasing returns. Shove, following Marshall, introduced the time element into his analysis:

It may be true—to take an extreme case—of every firm in the industry that its average cost would be lower if, other things being equal, its share in the aggregate output were larger. But expansion takes time, and further time is required before the newly-installed equipment can yield all its fruit. During this interval other things are not likely to remain equal ... If a firm could enlarge its output instantaneously, then indeed the predominance of internal economies would, on our present hypothesis, be incompatible with competitive equilibrium; but since it cannot, the two conditions can be reconciled (ibid.: 109—110; italics in original).

Shove’s attempt to set the analysis of returns within a dynamic context was followed by Harrod (1930, 1931) in his early works on imperfect competition, but it was rejected by other participants in the debate.¹¹

At the time of the Symposium, Sraffa had already introduced his notion of imperfect competition (Sraffa 1926), and the debate on market forms was taking place in Cambridge as well as elsewhere. In his article, however, Shove continued to refer to competitive markets, even though his idea of competition was different from what was commonly meant by perfect or free competition. Shove’s view was in several respects closer to Marshall’s idea of competitive markets. Marshall never gave a precise definition of competition, but he described in several well-known instances the behaviour of firms in terms that were more compatible with a hypothesis of imperfect competition than with the prevailing (in the 1930s) definition of a perfectly competitive market.^[9]

• [1] This study concerned the relationship between costs and output. It was never published and the manuscript must have been destroyed after Shove’s death.
• [2] Robertson (1930: 87), instead, was convinced that it was possible to allow for internal economies without having to abandon the theory of competition. For this purpose, he used the Marshallian notion ofthe representative firm.
• [3] Although, in his Symposium article, Shove referred only to labour and land, he was also aware of thespecific difficulties related to the measurement of aggregate capital. He dealt explicitly with the problemof the measurement of capital in his review of Hicks’s Theory ofWages (Shove 1933a: 470—471).
• [4] Internal economies are those advantages that a firm obtains ‘from the efficient organisation of theresources which it employs directly (Shove 1930: 101; italics in original). External economies are the advantages that a firm derives from ‘the organisation of those outside resources, not in its own direct employment,whose services or products it uses’ (ibid.: italics in original).
• [5] Shove also introduced the term ‘economies of rationalisation (ibid.: 105; italics in original), that is,increases in efficiency due to a redistribution of resources within the industry made possible by improvements in the degree of knowledge, expertise, and so on.
• [6] He was aware of this: ‘Mr. Sraffa will say, perhaps, that the equilibrium reached under these conditionsis not competitive but monopolistic ... Whether a given situation is to be called competitive or monopolistic is, of course, a question of words ... We have simply to inquire whether, and if so why, substantialeconomies of mass production are consistent with the survival of a large number of competing firms. Ishall therefore merely observe that if competition implies not only a perfect market but also a completeabsence of transport charges, it never exists in practice. We are no longer being asked to explain whathappens in the real world but to solve a purely abstract and hypothetical problem’ (ibid.: 108—109).
• [7] That is to say, an unrealistic situation in which neither transport nor marketing costs are significant.
• [8] Two years before Shove and Harrod, Young had assigned the time element great importance for theanalysis of increasing returns (Young 1928).
• [9] On Marshall’s notion of competition and its relationship with the Cambridge debate on imperfectcompetition, see Whitaker (1989: 172—184).