Andrews' Later Work

It has often been noted that Manufacturing Business fell neatly between two stools, being too difficult for the lay (i.e. business) reader and not rigorous enough for an academic audience.

E. H. Chamberlin and to his publisher, Harold Macmillan (see ibid.: 219-232). The first lecture provides his own interpretation of Alfred Marshall, which he contrasts with that of A.C. Pigou, of whom Andrews is a strong critic. The second deals with the post-Marshallian analyses of monopoly, oli­gopoly, imperfect and monopolistic competition in the 1920s and 1930s, with particular reference to Chamberlin and to Joan Robinson, concluding with a discussion of the work of the OERG. The third and final lecture sets out Andrews' own ideas on price theory. The lectures are superior on all counts to both the 1949 Oxford Economic Papers article and the theoretical compo­nent of Manufacturing Business.

‘My work has led me to rather different interpretations of Marshall', he notes early in the first lecture, ‘from that which was normal in the inter-war years, although on several points it is consistent with the position sustained by two of Marshall's pupils - Professors [Dennis] Robertson and [D.H.] MacGregor' (ibid.: 179). Andrews believed that Marshall himself was not very sympathetic to the neoclassical theory of atomistic equilibrium of the firm:

Marshall retained the idea of a supply curve, and of a stable equilibrium between demand and supply conditions because such an idea conformed to reality, in so far as actual prices were stable in given conditions; also because, given changes in demand conditions, changes in the levels of market prices could be explained on the lines of systematic changes in supply conditions according to changes in the underlying cost conditions.

Thus, Andrews concluded, Marshall's ‘definition of competition was not that of perfect competition as we have come to know it in later textbooks or as it was already being presented in continental textbooks' (ibid.: 183). On the contrary, Marshall's definition ‘was not in terms of homogeneous com­modities sold in markets where preferences did not exist, but simply in term of the fundamental assumptions of freedom of entry and of parity of prices' (ibid.: 183-184). From this, Andrews concluded that pure competition ‘should be seen as only a special case of Marshallian competition. The latter is the general case which he differentiates from pure monopoly, where freedom of entry was impossible and so the [price] parity condition could not apply' (ibid.: 184).

This led Andrews to make strong criticisms of many subsequent Marshall scholars. ‘Later generations', he maintained, ‘have interpreted the whole of Marshall in terms of pure competition analysis and have convicted him of error in the one field where he now seems to me to have been so original' (ibid.: 184-185); Pigou was especially culpable in this regard. Andrews pur­sued this theme in the second lecture, noting that ‘[t]he revolutionaries of the 1930s...went a good deal further in the assumptions which they postulated rather than justified. In approaching the problem, they were prisoners of the idea of full equilibrium in the individual business which had dominated the older traditional theories' (ibid.: 192).

Piero Sraffa (1926) had seen this very clearly, outlining what in the later published work especially of Joan Robinson would become the Cambridge theory of imperfect competition.

Andrews also commented critically on contemporary approaches to the theory of oligopoly, which had ‘all been far too much concerned with short­period price-cutting to have general relevance to the problem of normal prices, since price-cutting is not a normal phenomenon' (Andrews 1952 [1993]: 199). ‘At their most abstract level'—here he refers to von Neumann and Morgenstern (1944)—‘they seem to me to have little relevance to price for­mation as ordinarily developed in established industries' (Andrews 1952 [1993]: 200). Similar criticisms applied to the work of Michal Kalecki and Peter Wiles. ‘Other examples could be cited, and it would seem that, like the sorcerer's apprentice, we have become the victims of our own devices' (ibid.: 202).

In the third lecture, Andrews reported on his own efforts to escape from these devices: ‘When I study a business I go to it, stay in it, and work, [so] to speak from the inside, so a wide range of experience may be available to me' (ibid.: 205). He spent some considerable time in each business, and insisted on being able to move around freely: ‘It is from studies of this kind', he con­cluded, ‘that my theories of price formation emerged gradually' (ibid.: 206). He began by assuming that ‘all consumers of the product in question are other business men', so that the issue of consumer irrationality did not arise.

In the long term, Andrews insisted, ‘if there is to be equilibrium in the market, the prices of identical products must be identical'. But this ‘does not imply any infinite elasticity of demand for the product of the individual busi­ness’. On the contrary, the attachment of buyers to one particular supplier ‘implies that, at the given price, each business will at any one time have only a definite demand, which can be increased in given conditions only by the rather slow process of building up goodwill’. What Andrews terms the ‘price line’ for the product of any particular firm ‘represents the maximum price which the business can charge if it is to retain its goodwill in the long run. This price will represent the lowest price any potential competitor would charge for a product of identical specification’ (ibid.: 209).

So much for demand. On the question of costs, Andrews argues that, with given input prices, ‘average direct costs will be constant over the range of out­put which the business is organized to produce’ (ibid.: 210). He ‘finds it dif­ficult to accept’ the ‘simple U-shaped cost curve’ of traditional theory, instead proposing ‘a falling long-run average cost curve because of the influence of technical factors’. However, it is likely that ‘such a curve will fall more steeply for increases from a relatively small scale, than it will for increases from a rela­tively much larger scale. The curve of long-run costs in our model may, there­fore, be drawn as falling even more gently’ (ibid.: 211).

The implications for pricing decisions are clear. The businessman ‘will have no cost deterrent to increasing his scale to meet any permanent increase in demand but, equally, for any likely increase in demand he will not expect a substantial fall in his costs’ (ibid.: 211-212). ‘It will be seen that our model postulates no fine marginal balance, except in so far as the business will be doing the best it can, if it meets whatever demand comes its way to the limit of its capacity’ (ibid.: 212). In setting price, the businessman needs to calculate

the estimated average direct costs of a product, and the margin which he pro­poses to add in order to get his quoted price — the gross profit margin, already mentioned. Since his direct costs will be given quantities, given the specifica­tions, his pricing problem is to estimate the gross margin which it is safe for him to charge, revising his estimates downwards if he is forced to do so in order to meet competition... The costing margins are therefore determined by estimates of potential competition (ibid.: 212).

Andrews insists that this does not entail that all businesses always cover all their costs. This distinguishes his version of mark-up pricing theory from that of Hall and Hitch (1939): ‘There is no full-cost theory. It will be quite normal for any business to have a proportion of products which, for long-run reasons or because of the nature of its market, it wishes to keep offering’, even at a loss (ibid.: 213; italics added). Andrews concludes that, unless factor prices change, the prices set in this way will tend to be stable so that the overall price level will become unstable only in quite exceptional circumstances involving ‘a general fall in demand' or ‘great restriction of supply in the face of...a very strong demand' (ibid.: 214). On this important issue, he does agree with Hall and Hitch.

Andrews ends his lectures with

one last methodological observation. I have been very much impressed by the way in which traditional theory has tended to stultify empirical research, in the field in which I am most interested.

However, it was precisely this ‘fuzzy edge' that Andrews' critics would most strongly object to.

After 1952, he turned away from academia to work with businesspeople, whose company he found more congenial. Andrews' second, and last, impor­tant book did not appear until 1964. In On Competition in Economic Theory, he provided the detailed critical history of the theory of the firm since Marshall that should have come in the early chapters of Manufacturing Business, fol­lowed by a critique. Oligopoly is the most common market type, Andrews maintains, and the static marginalist equilibrium method used by the ortho­dox theory of the firm cannot deal adequately with it. There is no theoretical justification for the downward-sloping demand curve for the individual firm. ‘Joan Robinson's demand functions have no analytical roots', he concluded. ‘Her demand curves fall simply because she tells them to do so' (Andrews 1964: 22). In oligopoly, the firm cannot know its demand curve, which depends on the pricing policy of its competitors. Unless there is collusion between firms, marginalist pricing procedures are impossible. But collusion is also impossible unless entry is blocked (ibid.: 25-30). Potential cross-entry brings long-run considerations into the short run, destroying the short-run demand curves used by orthodox theorists and rendering cost and demand functions mutually dependent on each other. Thus, potential competition ‘removes the ring fence which is necessary for the playing of classical and neo­classical games' (ibid.: 84).

On Competition was widely reviewed, in rather more favourable terms than Manufacturing Business, but with criticism of both the depth of the theoretical analysis and the neglect of similar ideas that had been developed in the United States and Europe. Writing in the journal Kyklos, J.B. Heath described the book as ‘stimulating but endlessly frustrating... Time and again [Andrews] takes us to the brink of new ideas and analyses which offer the prospect of further advances in this difficult subject, and then with masterly self-control restrains himself from taking the plunge' (Heath 1965: 710). ‘Can it be hoped', T.A.B. Corley asked at the end of his review in Economica, ‘that Mr. Andrews, with his long experience and intensive study of the subject, will now lay aside his work on restrictive practices, and give us the “foundations” that are so badly needed?' (Corley 1965: 472). Similarly, Derek Robinson—the third of this ilk to engage with Andrews between 1950 and 1965—regretted in the Journal of Management Studies that he ‘does not, in the critique, directly state his own theoretical position'. ‘The next step', Robinson concluded, ‘is for Mr. Andrews to restate his own theoretical position in detail' (D. Robinson 1965: 237).

In the American Economic Review, E.T. Grether also complained about the absence in On Competition of a general theory that could then be applied to particular cases, making Andrews vulnerable to ‘the pitfalls of rationalizing the status quo'. Grether also criticised the ‘notable lack of reference' to the US literature on industrial organisation and the totally inadequate account of the work ofJoe Bain (see Grether 1966: 1264). The final review, by the Cambridge economist Aubrey Silberston in the Economic Journal two years later, also objected to ‘Mr. Andrews' dismissal of Bain's important work on new entry, in the course of two or three pages, on the grounds that Bain has not suffi­ciently taken into account potential competition from established businesses' (Silberston 1967: 866). Indeed, Andrews had not taken adequate account of the concept of “limit pricing” that could be found in Bain's book on Barriers to New Competition (Bain 1956), and later in a well-received book on Oligopoly and Technical Progress by the Italian theorist Paolo Sylos-Labini (1962). There was really no excuse for the omission, since both had been summarised— Sylos-Labini from the as then untranslated Italian version—by Franco Modigliani (1958) in the Journal of Political Economy.

5