The Discussion about Equilibrium and Imperfect Competition

In 1930 and 1931, Harrod published two important articles in which he showed that it was possible to have long-period equilibria characterised by the existence of firms whose size is less than the optimum, that is, firms that produce at a higher average cost than is technically possible. In equilibrium, price equals average cost and marginal cost; therefore, if the firm produces an output less than that which could be produced at a minimum average cost, this implies that it experiences decreasing average costs and has excess capacity. In other words, there is equilibrium with increasing returns. In order that such an equilibrium be possible, firms must face a downward-sloping demand curve. The downward-sloping nature of individual demand curves depends on firms’ increasing difficulty in marketing their additional output (Harrod 1931: 566). Harrod expressed in a more formal way what Sraffa had already pointed out in 1926.^[1]

Harrod, like Shove, tried to explain this kind of equilibrium by introducing some dynamic considerations.^[2] However, Joan Robinson who, in 1932, published an article on ‘Imperfect Competition and Falling Supply Price’, did not follow Shove and Harrod along this path. Her analysis was rigorously static.

⁴

In particular, Robinson argued that the equilibrium of individual firms does not imply that the industry as a whole is in equilibrium. In the long period, new firms could enter the industry and cause aggregate output to vary. The industry will reach equilibrium when profits have reached their normal level in all firms.

Such a notion of industry equilibrium raised two issues: Is free entry compatible with imperfect competition? And how is normal profit to be defined? For Robinson, the profit that a firm expects in order to be induced to enter an imperfect market must be higher than under perfect competition because the entry into imperfect markets is more difficult than into perfect markets. This, however, does not mean that profits in an imperfectly competitive industry are higher than normal; in fact, the entry into other imperfectly competitive industries is difficult as well. But, in any case, this problem is not relevant because, for a particular industry, it is sufficient to define the normal level of profits as that level just sufficient to dissuade the entry of new firms (Joan Robinson 1932: 547). The normal profit so defined is then included in firms’ long-period costs. The industry is in equilibrium when price equals average total cost—a condition that can be fulfilled only when the individual demand curve of the firm is at a tangent to its average cost, that is, there is equilibrium with excess capacity.

Robinson ascribed this result to Kahn. Shove, in fact, called it ‘Mr. Kahn’s theorem’ (Shove 1933b: 118) and took issue with it. Shove raised two main questions. The first was concerned with supply functions and, more particularly, with the relation between changes in demand and costs. Shove argued that ‘the cost at which a firm provides any given output depends on the way in which it provides it, and this in turn depends on the character of the demand for it’ (ibid.: 115). Therefore, the cost curve, in general, cannot be drawn, as Joan Robinson drew it, as if it were independent of the demand curve. The second question concerned the concepts of normal profit and equilibrium. Shove argued that the industry equilibrium, defined as the absence of a tendency for output to change, does not require profits to be normal. Profits can be below the normal level provided that they are not below that level which would cause some firms to leave the industry (ibid.: 119). If there were perfect competition and, hence, no costs of entry into an industry, the two limits for profits would coincide. But, given the assumption of imperfect competition, entry costs are positive, so that the two profit levels do not coincide.

As a result, despite below-normal profits, the industry output does not vary. Hence, the stability of the industry output does not require that price equal average cost. The equality between price and average cost is necessary only to ensure the maximum level of profits: ‘Mr. Kahn’s theorem does not mean that the individual demand curve of the firm must be tangent to its average cost curve in order that there should be no tendency for the industry’s output to alter, but merely in order that the firm should be earning this maximum profit’ (ibid.: 121). The equilibrium of the industry requires only that the price lies between the average cost defined to include Joan Robinson’s normal profit and the average cost defined to include the minimum profit below which firms would leave the industry.

The last part of Shove’s article was devoted to his alternative line of approach, which is based on the idea that the complexity of actual markets makes it very difficult to expound the analysis neatly and precisely.^[3] The main difficulty lay in the fact that, in general, costs cannot be regarded as independent of the state of demand. Total cost can be divided into two parts: the minimum cost of output (dependent on technical factors) and the special cost, related to advertising, transport, and so on. While the minimum cost in general depends only on the level of output, the special cost partly depends on the state of demand and partly on the technical constraints faced by the firm.^[4] In this context, a change in the character of demand is likely to alter the special costs associated with different levels of output. Economic theory must develop along such lines in order to account for these aspects of the actual world; otherwise, it becomes a mere intellectual abstract exercise.

In her Economics of Imperfect Competition (Joan Robinson 1933), Robinson essentially maintained the same line of approach. Shove’s review of Robinson’s book, although kind in tone, was in fact quite critical (Shove 1933c). He held that the book, far from being innovative, was a careful and detailed reexposition of the conventional theory (ibid.: 657). Shove argued that the part of the book devoted to imperfect competition was less satisfactory than the part on monopoly, mainly because the degree of abstraction adopted by Robinson was unable to deal with the case of imperfect competition.^[5] The high degree of abstraction in Robinson’s book made the analysis easier but also more distant from the real world.

Both Robinson and Shove were aware of the difficulties that arise in the analysis of the firm and the industry when the assumption of perfect competition is removed. Shove insisted on the need to cope with these difficulties rather than assume them away, whereas Joan Robinson assumed most of them away in order to render the analysis tractable.

Almost 10 years later, in 1942, Shove returned to the issue of the unrealism of the contemporary analyses of imperfectly competitive markets, and he contrasted these with Marshall’s approach. According to Shove, Marshall was aware of the relevance and realism of non-competitive market forms. Sometimes, in discussing non-competitive markets, Marshall left his value theory ambiguous and vague, but this was because he ‘accepted the view that value under monopolistic competition is theoretically indeterminate, and concluded accordingly that pure analysis could not accomplish much in that field’ (Shove 1942: 320). Yet it was Marshall who first developed some of the analytical elements on which the modern theory of imperfect competition rested. He had introduced the device of contrasting the individual supply curve with the individual demand curve; the notion that equilibrium is realised when marginal cost equals marginal revenue; that in equilibrium marginal firms must earn a zero profit (ibid.: 321). Younger economists refined those concepts and made them ‘more precise’, but this was accomplished at a considerable cost to the realism of the analysis:

The attempt to fuse realistic study with theoretical analysis has, on the whole, not been followed up ... Partly no doubt this has been due to the itch for precise results: not all of us are content to act on the late Prof. Wildon Carr’s admirable motto (which might well have been Marshall’s), ‘It is better to be vaguely right than precisely wrong’ (Shove 1942: 323).

Shove’s criticism of the younger generation’s mishandling of the Marshallian inheritance raises the more general issue of how modern economists engage with economists of the past. Once again, Shove’s ideas were developed in the context of a criticism of Joan Robinson’s position. The occasion was the publication of her Essay on Marxian Economics (Joan Robinson 1942). In his review article of the book, Shove (1944) considered also Robinson’s critique of orthodox economics and questioned the usefulness of the very notion of economic orthodoxy. For Robinson, for example, orthodox economists postulated the harmony of interests between classes; accepted the capitalist system as a natural and inevitable form of social and economic organisation; based their theory of value on the assumption of perfect competition; regarded the distribution of income as determined by natural laws; and accepted the validity of Say’s Law. Shove objected that it is impossible to identify a set of economists who share all of these characteristics (ibid.: 50—53).^[6] For Shove, ‘what is wanted...is not an attempt to overthrow or disintegrate, but rather an effort, as detached as possible, to understand, to sift truth from error and to formulate precisely the problems which have been left unsolved’ (ibid.: 60).

Shove stressed the limits of the past generation of economists and the need for theoretical advancement, but he also tried to maintain a ‘generous’ attitude towards the economists of the past. At the same time, he criticised his younger colleagues for having concentrated on those aspects of past economists’ analyses that lent themselves to be developed more ‘precisely’ but were also more unrealistic and removed from the real world.

Perhaps Shove’s emphasis on the need to develop a more realistic approach to economic theory influenced the younger economists in Cambridge in a more general though indirect way. Their enthusiastic taking up of Keynes’s new ideas and their loss of interest in the formal analysis of individual firms and industries can be interpreted as an expression of their quest for a higher degree of realism in economics. They became convinced that the pressing problems of actual economies in the 1930s could be understood and dealt with by developing the analysis of the economy as a whole and by not assuming away essential aspects of the real world such as uncertainty and money.

• [1] ‘The chief obstacle against which [firms] have to contend when they want gradually to increase theirproduction does not lie in the cost of production—which, indeed, generally favours them in that direction—but in the difficulty of selling the larger quantity of goods without reducing the price, or withouthaving to face increased marketing expenses. This...is only an aspect of the usual descending demandcurve, with the difference that instead of concerning the whole of a commodity.. .it relates only to thegoods produced by a particular firm’ (Sraffa 1926: 543).
• [2] Harrod, for example, argued that conditions of increasing returns are normal in an industry where therate of expansion of the optimum plant is higher than the rate of expansion of demand. For a moredetailed reconstruction of the 1930s debate on imperfect competition, see Sardoni (1999).
• [3] ‘So long as we are content with a rough and ready indication of the forces at work, we can keep fairlynear to the facts: but any attempt to make our treatment exact is apt to lead either to a degree of abstraction which renders the apparatus inapplicable to the actual phenomena we set out to explain or to adegree of complication which makes it cumbrous to use’ (ibid.).
• [4] For example, the cost at which a firm can provide advertisements, facilities, transport, and so on.
• [5] Moreover, ‘in this branch of her work the author has not so consistently resisted the temptation to usea “technique” which is admittedly designed “for studying equilibrium positions”.. .in an attempt to analyse the effects of change’ (ibid.: 659).
• [6] When Shove considered Robinson’s criticisms of Marshall and Wicksell, he accused her of being carelessand unfair in her reading of their works. Shove argued that Robinson attributed to Marshall and Wicksellpositions that they never held and, more generally, that her criticism of orthodoxy was driven by her ideological stance rather than by a scholarly attitude towards different theoretical approaches (ibid.: 60).