The Theory of Imperfect Competition: a Review of the Post-Keynesian ContributionCARLOS EDUARDO SCHÖNERWALD DA SILVAResumo:De acordo com a teoria pós-Keynesiana, cada planta da firma é construída em umaescala menor do que a escala ótima. Então, segundo esta escola, o custo médio de longo-prazo pode diminuir e gerar uma situação de retornos crescentes, esta situação não estavadevidamente explicitada na teoria microeconômica tradicional. A escola pós-Keynesianaficou conhecida por sua contundente crítica a tradicional escola neoclássica, crítica essa queestimulou o desenvolvimento das teorias sobre a organização industrial das firmas. O artigopretende apresentar os principais pontos criticados e discutidos pelos autores pós-keynesianos,bem como, as suas contribuições para o desenvolvimento futuro da teoria da concorrênciaimperfeita.Palavras-chave: Concorrência Perfeita, Concorrência Imperfeita, e teoria microeconômicapós-KeynesianaAbstract:The excess capacity, according to post-Keynesian authors, is the fact that each plant isbuilt on a less than the optimum scale. After that, the long-run average cost can decrease and lead to the situation of increasing returns, so raising the demand will promote an industrialorganization in a direction to long-term plant that surely will generate more outputcomparatively to the short-term plant. Thus, they point out that the theory of perfectcompetition does not provide the groundwork to analyze increasing returns of scale, even in the long run. The paper presents the main points criticized and discussed by the post-Keynesian authors, such as their contributions for the future development of theory ofimperfect competition.Key words: Perfect Competition, Imperfect Competition, and post-Keynesianmicroeconomics theory.Classificação JEL: D01 - Microeconomic Behavior: Underlying Principles. L11 - Production,Pricing, and Market Structure; Size Distribution of Firms. Page 2 2IntroductionPiero Sraffa, in 1926, wrote the article “The Laws of Returns under CompetitiveConditions”, considered the cornerstone of the post-Keynesian microeconomics. Sraffaverified that firms operating under perfect competition must be subject to decreasing returnsof scale and that increasing returns would only exist on the presence of monopoly. Later on,many heterodox authors such as Joan Robinson, Lord Kahn, Nicholas Kaldor, Roy Harrod(Cambridge U.K.) and Edward Chamberlain (Harvard University) contributed to write thehistory of the theory of imperfect competition. Robinson (1934) was so excited about the gainconquered by Sraffa that she wrote: “he was not himself completely aware of the freedomthat he was winning for us”.The theory of perfect competition presumes that resources are fully employed. Thereis no open window to use the concept of excess capacity, so market forces, without anyparticular interference from buyers and sellers, promote full employment. However, thegroundwork of the theory of imperfect competition is the existence of structures thatdestabilize the market mechanism, like monopolies and oligopolies, and generate excesscapacity.The objective of the article is to provide a review of the post-Keynesian contributionconcerned to the theory of imperfect competition. The first section provides a discussion about the concept of increasing returns and how it is defined and related to imperfectcompetition. In the second section, it presents how the notion of normal profits works outunder perfect competition. The third section presents how firms take decision and what are therelevant aspects that guarantee the market power.The Theories of perfect and imperfect competition: the role of the economies of scaleInside the economic science, mainstream economics advanced on the theory of perfectcompetition motivated by a faith on the symmetry between the forces of demand and supply.In traditional textbooks, perfect competition is characterized by the assumption that marketforces drive the price of commodities so smoothly that we can reduce and group themtogether. Therefore, the back and forth movement of the market forces can be represented by a pair of intersecting curves of demand and supply.Post-Keynesian authors began deciphering the code of the mainstreammicroeconomics putting side by side the differences between perfect competition and purecompetition. The first contribution was given by Harrod (1931). Primarily, he points out that Page 3 3perfect competition is characterized by organized exchange and homogeneity of the products.In addition, in perfect competition the demand is infinitely elastic because always themarginal cost of production is, in equilibrium, identical to the price. Secondly, purecompetition has its special case in which the demand elasticity is infinite; it happens when themarket is perfectly organized, and the marginal cost of production is like the price.The second contribution was given by Chamberlain (1933). On the hand, he concludesthat pure competition is a state of circumstances in which the demand for the output of an individual firm is perfectly elastic. On the other hand, perfect competition is deeper than purecompetition because it represents additional conditions such as fluidity or mobility of factorsand absence of uncertainty.Robinson (1934) provided the third enlightenment, and it appears the most concisedescription about the concept of perfect competition. Perfect competition, in mainstreameconomics, involves rational behavior on the part of buyers and sellers, full knowledge,nonexistence of frictions, static conditions, perfect mobility and perfect divisibility of factorsof production. Nevertheless, she defines perfect competition in a way that embraces all theassumptions previously exposed, so perfect competition means a situation in which thedemand for the output of an individual seller is perfectly elastic.In the theory of perfect competition, written by several classical authors, the law ofdiminishing returns was based on the problem of rent, so raising the exploitation of less fertilelands have a tendency to reduce the marginal returns per unit of input. Sraffa (1926) statesthat if perfect competition exists than the commodity can be considered as a homogenousgood and under that concept is possible to establish the role of diminishing returns. Here wehave a cut-off point, mainstream economics assumes diminishing returns because they takefor granted that the economy has uncountable small firms that face homogeneous productsand without sufficient power over the industry supply. However, post-Keynesian economicsconsiders commodities as containing properties much more complexes than just simpleagricultural goods. They consider commodity as a collection of diverse articles, ranging fromfoodstuff to ironware. Consequently, they develop the idea of product differentiation to understand the control of the industry supply by a single firm and there might be the case forincreasing returns of scale and, as a result, imperfect competition.Kaldor (1935) points out the role of perfect divisibility assumed by mainstreameconomics to support the theory of perfect competition. He points out that when mainstreameconomics assumes perfect divisibility and, by definition, economies of scale are completely absent, perfect competition will be established because of the autonomy of the market forces. Page 4 4According to Sraffa (1926), mainstream economics does not make clear the role ofincreasing returns as it does explain the role of diminishing returns. The framework relatesincreasing returns to the division of labor. It did not develop the role of increasing returns asbeing a mechanism that happens when the size of the firm rises. The division of labor must besubject to a more radical transformation in the foundations to verify the increasing productivity in the whole industry as an effect generated by the external division of labor.Differently, mainstream economics accepts as true that the modification is promoted byinternal division of labor, so it occurs only in the local structure of the firms. The importanceof external economies was increasingly emphasized as the advantage derived by individualproducers from the growth, not of their own individual undertaking, but of the industry in itsaggregate.“The result was that in the original laws of returns the general idea of a functionalconnection between cost and quantity produced was not given a conspicuous place; it appearsin fact, to have been present in the minds of the classical economists much less prominently than was the connection between demand and demand price.” (Sraffa,1926, p. 537)In mainstream economics, firms tend to have their plants constructed on a scale in which, regard the technical considerations and the supply price, is the one capable ofproducing most cheaply. However, post-Keynesian economics points out that plants and firms may be coextensive or a firm may include a number of plants, or a number of firmsutilize a single plant. Thus, sources of production in perfect competition cannot be subject tothe optimum size autonomously of the state of demand. It follows that no increase of outputwill give an appreciable influence on price; the firm is free to raise its plant to the scale atwhich it can produce most cheaply.“Short-period variations in demand are met by variations in the extent to which fixed means are exploited. Long-period variations in demand are met by an increase or diminutionin the number of sources of supply. Each source of supply tends to have its fixed means of an optimum size independently of the state of demand.” (Harrod, 1934, p. 443)The time adjustment is not considered by mainstream economics as a clarification ofhow diminishing returns operate. In opposition, post-Keynesian economics points out that in the short-run the constraint of the firm is the size of plant, so the capital is holding constantand the use of more factors of production tends to present diminishing returns. However, in the long run the time adjustment is long enough to promote a reconfiguration of the size of the Page 5 5plant, so firms tend to have increasing returns of scale because of the reduction in the long-runaverage cost, envelope curve.“Reductions in cost connected with an increase in a firm’s scale of production, arising from internal economies or from the possibility of disturbing the overhead charges over alarger number of product units, must be put aside as being incompatible with competitiveconditions.” (Sraffa, 1926, p. 540)Sraffa (1926) concludes that to fully understand microeconomics it is necessary to leave aside perfect competition and move towards the opposite direction, monopoly and oligopoly. In monopoly and oligopoly, the power between the supply and demand is notequal, and the competitive effect is not transmitted such as in perfect competition becausefirms have the market power to delineate the price that the product will be sold. The nextsection will be presenting how the mechanism of ‘free entry’ normalizes profits, so normalprofits are one of the keystones of the theory of perfect competition and it was radically criticized by post-Keynesian economics.Normal profits in a context of perfect competitionIn any one industry, profits are normal when they are the same as profits in otherindustries. Normal profits are simply the profits that prevail when there is no tendency for thenumber of firms in the industry to alter. As soon as profits are supernormal, new producerswill come in until profits are condensed to normal profits. Thus, when profits are more than normal profits firms enter and profits decline; if profits remain more than normal, new firmswill come in until profits are reduced to normal and there will exist no incentive for one novelfirm to go in.Robinson (1934) explains that in a long-period supply curve the dynamic adjustmentmay take the view of a process of reaching equilibrium, and then the existence of two levelsof profits presents time as a key variable. It must be conceded that a persistent gap between supernormal and normal profits is likely to occur, so that in fact the event is likely to be found in many industries where the market is imperfect. Furthermore, the continuation of the gapdepends on costs of movement from one industry to another, and it may occur when competition is perfect. Moreover, competition may be imperfect when there are no costs ofmovement, such as the case of special licenses to operate in an industry. Finally, the level ofprofits that will bring new comers into an industry is usually higher than the level that is justsufficient to retain existing enterprise. Page 6 6A gap between the upper level of reward, necessary to tempt new resources into anindustry, and the lower level, necessary to drive old resources out, will exist wherever there iscost of movement between one trade and another, and the double level of normal profits ismerely one example of a phenomenon which may affect every factor of production equally.(Robinson, 1934, p.108)The figure 1 has two supply curves, the lower one applies only to expansions of theindustry, and the upper one applies only to contractions. The supply curves are determined,under perfect competition, by the marginal cost curves of a given number of firms. Robinson (1934) points out that each point on the upper curve is joined to a point on the lower curve atwhich the number of firms is the same. This new curve is called as ‘quasi-long-period supply curve’.Figure 1 – Normal Profit’s Dynamic 0 PQRS Source: Robinson (1934)Starting from a position in which price is 0P and output 0Q, and considering anexpansion of demand, so the supply price climbs up the quasi-long-period supply curve to R.It proceeds for further increases of demand along the upper long-period supply curve to theright. After that, in perfect competition, new firms come in to the industry increasing thesupply, the supply curve shifts to the right, and reducing the price.The quasi-long-period position does depend on history. There is a continuous series ofquasi-long-period curves, and the actual curve depends on the number of firms in existence at Page 7 7the moment, as far as the familiar short-period curve depends upon the quantity of fixed plantin the industry. Robinson (1934)The adjustment on the region of normal profits is based on how the supply price and the output react, and both depend largely upon the past behavior of the industry. On the hand,if fewer firms had happened to enter in the period of high profits, the actual price of a givenoutput would be higher. On the other hand, if more firms had entered the actual price wouldbe lower. Robinson (1934)Another force that shifts the supply curve happens when discontinuous changes in thenumber of fresh entrepreneurs (each in front of imperfect knowledge of the others’ action)come into the trade. Through this new competition, actual profits are heavily reduced to alevel below the one that attracts new comers, but they are not sufficiently low to run out any existing firms. The industry will continue at this inflated size, and it will be in equilibrium inthe sense that no new enterprise tends to enter or old enterprise to leave. Robinson (1934)“It is true that a high level of normal profits will often be found where competition isimperfect. The fact that an old-established firm enjoys “good will” has the effect both ofgiving it a hold upon the market which enables it to influence the price of the commodity which it sells and of increasing the cost of entry new rivals. And the powerful firm which usesthe methods of “unfair competition” to strangle rivals is highly unlikely to be selling in aperfect market. But this association of high normal profits (not abnormally high profits) withimperfect competition is a purely empirical one.” (Robinson, 1934, p. 107)Finally, perfect competition is a situation in which a particular seller does not controlprice and in which a single seller cannot make more than normal profit. According to Harrod(1934), the entry of new firms into an industry must be difficult when it is operating underimperfect competition, so the problem of normal profits is not considered because there aremechanisms that provide supernormal profits. The next section presents the background of thetheory of imperfect competition written by post-Keynesian authors.The theory of imperfect competition from a post-Keynesian perspectiveThe preliminary point about the theory of imperfect competition is that individualfirms are frequently confronted with a downward sloping demand curve. The possibility of aparticular firm being confronted by a downward sloping demand curve (absolute monopoly)was recognized by economic theory before the post-Keynesian contribution. However, thetwisting point are the aspects related to market power, such as marketing expenses (form of Page 8 8advertising, commercial travelers, and facilities to customers), product differentiation, andincreasing returns of scale.Sraffa (1926) points out that a small firm is held in equilibrium by being subject to increasing marketing expenses. According to Harrod (1931), marketing expenses are all costsinvolved to invade the competitor’s territory, including transportation costs. The marginalcost rises when output increases, but it depends on the strength of the demand in a given area.Thus, a small firm to increase its market share must raise marketing expenses, a situation thatis absent in the theory of perfect competition.Supplementary, when the product is not fully standardized or the market is notorganized, producers may have difficulty in marketing increments of produce. A small firmcan move beyond that difficulty in two ways: by lowering the price or by increasing marketing expenses, so it cannot sell a larger quantity of goods without reducing its price, orwithout having to face increasing marketing expenses. Those circumstances are directly related to the notion of imperfect competition.[…] this necessity of reducing prices in order to sell a larger quantity of one’s own product is only an aspect of the usual descending demand curve, with the difference thatinstead of concerning the while of a commodity, whatever its origin, it relates only to goodsproduced by a particular firm; and the marketing expenses necessary for the extension of themarket are merely costly efforts (in the form of advertising, commercial travelers, facilities to customers, etc.) to increase the willingness of the market to buy from it-that is, to raise thedemand curve artificially. (Sraffa,1926, p. 543)According to Harrod (1931), the marginal competitive marketing costs can berepresented as a function of the output of the individual firm. However, he points out that itappears that marketing costs do not depend on only of the output level, so if the rise on thedemand is evenly diffused over the whole market, firms should be able to maintain theirfrontiers without increase the marketing effort. A higher competitive marketing cost is theprice to march into the neighbor’s territory. If no movement in either direction occurs, no risein this cost per unit at the margin should occur. However, all firms will be producing more inthe new equilibrium.Harrod (1931) brings an important analysis about the total costs been a function of thedemand and marketing expenses. He points out that, in mainstream economics, supply anddemand are independent of each other. On the new view, every demand has its own suitablesupply schedule. To establish equilibrium after a change on the former, the latter also must bechanged. The regular graphical representation of the supply curve is no longer possible. Any Page 9 9given supply curve of the old category is only valid while the demand remains the same. Todraw a unique supply curve to be valid for all states of demand, it is necessary to use threedimensions. Cost becomes a function of two independent variables (quantity of output and state of demand) and, consequently, the traditional analysis breaks-down.Succinctly, Chamberlain (1933) defines that, under imperfect competition, the marketis separated to a degree and sales are limited and defined by three new factors: (1) price, (2)the nature of the product, and (3) advertising outlays.The conventional concept of monopoly was the main part of the field of production outside that of perfect competition, so it was recognized that the monopolist’s position wasnever absolute and the elasticity of the demand for his product was always greater than zero.Harrod (1934) reveals that if products are absolutely homogeneous and marketed byorganized exchange is likely to perfect competition to reign. If differences of design and detailare possible, each producer may be defined as a monopolist of his own goods, but subject to the reaction of his rivals. The degree of monopoly may be measured by the similarities ofcommodities.Figure 2. The Traditional Analysis of Monopoly 0 PQDMRMC M P M QSource: Chamberlain (1933)The notation in the figure 2 is given as such: MC is the marginal cost, D is thedemand, MR is the marginal revenue, Q is quantity, and P is the price. Harrod (1931) figuresout exactly what was written in Robinson and Chamberlain (1933), the volume of output(figure 2) is determined by the intersection between the marginal revenue curve, derived fromthe demand curve, and the marginal cost, but the price M P is defined by the demand curve. Page 10 10According to Chamberlain (1933), the demand curve imposes upon the seller a priceproblem for his product comparatively to the horizontal one from the perfect competition.However, the monopolist decision about the price and the output depends upon the elasticity of the curve and upon its position relative to the cost curve for his product. In thatcircumstance “profits may be increased, perhaps by raising the price and selling less, perhapsby lowering it and selling more.” Chamberlain (1933, pg 71) The position and elasticity of thedemand curve for the product of any one seller depend in large part upon the availability ofcompeting products and prices which are asked for them.The crucial difference between perfect and imperfect competition is related to priceand output, so when an individual firm faces an infinite elastic demand for its product theprice that will be charged is lower than the price under imperfect competition. Furthermore,the mechanism that embraces the idea of market power involves a trade-off between the pricecharged and the output supplied, so the firm that enjoys some kind of market power will tend to use the quantity supplied as a variable to control and establish its price level.Consumers are willing to buy more goods if the price level is lower than the oneestablished under monopoly, but the conditions of imperfect competition imposes restrictionsover the market mechanism to bring in more firms to compete and to reduce the price level.Moreover, firms that have fixed plants and market power will tend to face excess capacity because they can reduce the output produced in order to increase price and, consequently,underutilizing the installed capacity.Firms invest in product differentiation to avoid the competition of new entrants, so thevolume of sales is based upon the method in which his product differs from competitors’product. Differently, under perfect competition, a producer may shift from one sector to another, but the volume of sales never depends, as under imperfect competition, upon productdifferentiation. The producer, in perfect competition, is always a part of a market in whichmany others are producing the identical good. The sales may vary over a wide range withoutchanging the price, so they may be as large or as small as he pleases without the necessity ofaltering his product. Chamberlain (1933)The differentiation is an important aspect but its variation may refer to the quality ofthe product – technical changes, new designs, or better materials; it may mean a new packageor container; it may mean more prompt or courteous service, a different location. Chamberlain (1933)If the economies of scale is totally absent when the demand rises the inflow of newproducers will continue, leading to a continuous reduction in the output of existing producers Page 11 11and a continuous increase in the elasticities of their demand until the latter becomes infiniteand prices will equal average cost. There the movement will stop. However, each firm willhave reduced his output to such an extent that he has completely lost his hold over the market.Kaldor (1935)Harrod (1934) presents a position that conflicts with the other post-Keynesian authors;he assumes that only when a monopolist has the control not only the whole supply of onecommodity but that of a large group of commodities, such as all foodstuffs, it representsabsolutely dominating position. “Otherwise his absolutism is tempered.” Harrod (1934, pg.445)The role of the economies of scope is demonstrated by Kaldor (1935) as a mechanismthat reinforces imperfect competition. Thus, if there is not a sufficiently great demand toproduce one product on an optimal scale, the producer may still utilize his plant fully by producing two or more products, rather than building a smaller, sub-optimal plant or leaving his existing plant under-employed. According to Kahn (1935), the size of a firm depends on two sets of factors: (a) the technical conditions of production, as expressed by its cost curve;and (b) the degree of imperfection of competition, as expressed by the demand curve for itsproduct.The question about market integration can be analyzed using the explanation ofHarrod (1931), so the firms, in order to maintain their market power, may require a licensefrom some controlling authority, or the existing firms may be so strong that they are able torepel new competition employing a price war. They may even resort to violence to preventfresh rivals from appearing on the industry. In such cases, no level of high profits will besufficiently enough to tempt new firms into the trade, and the supply of enterprise to that tradeis perfectly inelastic at the existing amount. For such industry, any level of profits is normaland the term ceases to have a valid application.The meaning of entrepreneurs is presented by Robinson (1934) and Kahn (1935), so in a world in which all entrepreneurs are alike there would be a uniform rate of profit in allindustries in the long period. In the real world entrepreneurship is no more homogenous than land in the real world. It is socially desirable to reroute entrepreneurs from industries in which the firms are naturally larger than the average for industry as a whole, and to attract them into industries where firm are naturally below the average in size. Where the firms are already naturally large under conditions of laissez-faire it is in the interests of society that they should be yet larger; where they are already naturally small it is interest of society that should be yetsmaller. Consequently, trades which require unusual personal ability or special qualifications, Page 12 12such as the power to command a large amount of capital for the initial investment, will tend to have a high level of profits; trades which are easy to enter will have a lower level.ConclusionThe excess capacity, according to post-Keynesian authors, is the fact that each plant isbuilt on a less than the optimum scale. After that, the long-run average cost can decrease and lead to the situation of increasing returns, so raising the demand will promote an industrialorganization in a direction to long-term plant that surely will generate more outputcomparatively to the short-term plant. The relation between these short-period and long-period phenomena may be represented by an envelope curve.The component of the demand in the imperfect competition theory is crucial anddirectly connected to production. Those authors figured out that an expansion of the demand have a tendency to force the producers to use more intensively the current plant. Additionally,if they expect that the present demand is not temporarily, so new and larger plants will beplanned and built in the future, establishing in the long-run a higher level of output.The extent to which excess capacity may be generated as a result of imperfectcompetition (under the assumption that the existence of economies of scale, will prevent thiscompetition from becoming perfect) will depend: (i) on how far monopolists take potentialcompetition into account in deciding upon their price and product policy, (ii) the extent to which institutional monopolies are present, it will tend to prevent the generation of excesscapacity if it leaves the scale of differentiation unaffected, (iii) the extent to which the market-situation resembles a chain relationship, i.e. the extent which the various cross-elasticities ofdemand differ in order of magnitude. Page 13 13BIBLIOGRAPHYCHAMBERLAIN, E. The Theory of Monopolistic Competition: A Re-orientation of theTheory of value. Harvard University Press. Massachusetts, 1933.HARROD, R. Imperfect Competition, Aggregate Demand and Inflation. The EconomicJournal. Vol. 82, No. 325, Special Issue: In Honour of E.A.G. Robinson (Mar., 1972), pp.392-401, 1972.___________. The Law of Decreasing Costs. The Economic Journal. Vol. 41, No. 164(Dec., 1931), pp. 566-576, 1931.___________. Doctrines of Imperfect Competition. The Quarterly Journal of Economics.Vol. 48, No. 3 (May, 1934), pp. 442-470, 1934.KALDOR, N. Market Imperfection and Excess Capacity. Economica. Vol. 2, pp;33-50,1935.KAHN, R. F. Some Notes on Ideal Output. The Economic Journal. Vol.45, pp.1-35, 1935.SRAFFA, P. The Laws of Returns under Competitive Conditions. The Economic Journal.Vol. 36, No. 144 (Dec., 1926), pp. 535-550, 1926.ROBINSON, J. The Economics of Imperfect Competition. Macmillan. London, 1933.__________ . What is Perfect Competition? The Quarterly Journal of Economics. Vol. 49,No. 1 (Nov., 1934), pp. 104-120, 1934.