You are on page 1of 11

American Economic Association

Profit Maximization under Monopolistic Competition Author(s): Stephen Enke Source: The American Economic Review, Vol. 31, No. 2 (Jun., 1941), pp. 317-326 Published by: American Economic Association Stable URL: http://www.jstor.org/stable/362 . Accessed: 20/03/2013 06:54
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

American Economic Association is collaborating with JSTOR to digitize, preserve and extend access to The American Economic Review.

http://www.jstor.org

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

PROFIT MAXIMIZATION UNDER MONOPOLISTIC COMPETITION


Current theory nominally recognizes that the individual firm maximizes profits, not through output adjustments alone, but also by determining price, the attributes of the product and selling effort. In this article the role of product variation and the interrelations of these three variables are emphasized. Diagrammatics analogous to those customarily used in price-output problems are developed and these are later employed to demonstrate that the incidence of relevant techno'logical innovations upon the quality of the product of the firm are independent of abnormal profits or of monopolistic advantage. Finally, the joint determination of output, product and promotion is considered. Normally, the exploitation of every profit possibility until marginal costs and receipts were equated would be expected. Available funds may, however, be limited and certain policies may earn losses unless more intensively pursued. In this case certain profit possibilities must be passed over in favor of others, the marginal net return on working capital will vary with different employments, and as the amount of such funds available changes the manner of their expenditure may differ markedly.

Entrepreneurs must determine, inter alia, the price and output of their product, its specifications, and the amount and type of selling effort.' The logic and illustrative diagrammatics of the entrepreneurial determination of output and price are familiar. Similar methods, developed below, are applicable to the problems of determining the optimum degree of durability and of other product attributes. The procedure used is shown to be relevant to a consideration of the incidence of improved technology on prices and quality. Finally, the joint determination of price, product,2 and promotion3 is attempted. In the following discussion it is assumed that the products of immediate rivals are heterogeneous and are considered by consumers to be close but not exact substitutes: they are "differentiated" one from another either because the product (or service) is tangibly different, or because it has been made to seem so, or because of location. The time period relevant to the considerations of this paper is not the long run, but a short period in which there is, however, sufficient time for some rather fundamental adjustments by entrepreneurs: the product of each firm is held capable of considerable modification, but the possibility of changing over to a radically different type of product with altogether dissimilar uses and attributes is excluded. The location of the enterprise is considered given, but distribution facilities or delivery services may be modified. The reasoning throughout assumes that entrepreneurs desire profits, have perfect knowledge, and act rationally.
1 If the firm has a less than infinitely elastic demand schedule, and one assumes that output and sales will be equated, then a price decision simultaneously determines output (or vice versa) and this theoretically comprises but one decision. 2 "Product" throughout includes the material good plus all associated intangibles except those resulting from promotion and the peculiar site of the place of sale. Thus "differentiation" due to location or advertising is excluded: all other types are included. '"Promotion" throughout includes all selling effort, whether advertising or personal salesmanship. Although advertising is only a particular form of promotion it will be used synonymously.

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

318

Stephen Enke

[June

I The relationship of price, promotion, and product, as it might be viewed by management, is set forth in the schema in Figure 1. Management is seen constantly attempting to alter, in its favor, the balance between total receipts and costs. Output is assumed always to equal sales, and gross receipts to be the product of price and sales. While promotion and piroduct affect total costs directly, they determine total receipts only indirectly via sales or price.

TotalReceipts

FIGURE 1

Theoretically, the relationship between any two of the variables indicated may be considered while the remainder are held temp-orarilyconstant. Suppose the rate of expenditure on promotion to be increased and that this increases the acceptance of the unchanged product by the public so that demand is increased in the schedule sens-eand a higher price or greater sales can be realized. If the entire effect is upon price, both sales and output costs will be unchanged; and the cost of promotion can readily be set against increased receipts. However, if price is unaltered, sales will increase and per unit costs will probably be changed; the additional advertising will then be of advantage to the firm only if there are increasing returns with augmented output sufficient to outweigh the cost of promotion. Whether the increased demand is exploited through sales or price will depend upon the magnitudes and slopes of the marginal costs and receipts curves. A further possibility is that management does not wish to increase either the price or sales

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

1941]

Profit Maximization

319

but prefers to change the specifications of the product so as exactly to offset the increased market acceptance resulting from piromotion and reduce its cost of production. This will be profitable if the savings in manufacturing are greater than the expense of the selling effort. If the functional relationship of every connected pair of variables were known, the most profitable way in which any change in a specific variable might work itself out on total costs and receipts could be calculated. In the following sections some of the relationships are considered. II We are all familiar with the diagrammatics by which the effect of variations in output upon profits, via average receipts and costs, is demonstrated. Variations in output change total costs and total receipits, but not commensurately, and it is this fact which makes the entire conventional apparatus necessary. Exactly this same procedure can be used to find the optimum determination of promotion or of the product. Instead of output we might measure, probably along the X-axis, the number of full-page advertisements run per unit of time, or some quality attribute of the product (e.g., durability, economy of operation, etc.). Let us suppose that a concern is engaged in the production and distribution of silk hose. Promotion and also the quality and number of sizes have been determined. This individual firm may superficially appear in equilibrium if it has chanced upon the optimum price-output relationship. Let us further suppose, however, that it has been decided to make only one color available; and, after market surveys, the optimum single color-flesh pink-has been adopted. The question arises as to whether a different number of color options would not prove more profitable. In Figure 2 a hypothetical relationship between number of color options and total receipts (or price) and total costs (or per unit costs) is illustrated. A special case of individual equilibrium with zero profits has been selected, and in a competitive industry where the products of rival concerns are close substitutes, there is probably a tendency to approach, if not to arrive, at this situation.4 The intersection of the marginal curves confirms OX as the optimum number of color options. The more colors offered, the greater the probability that each individual consumer will find exactly the color she wants for her ensemble. Hence, she will likely prefer this brand to others and be willing to pay a premium for it. In other words, the demand schedule for the hose of this firm will
'This is simply Professor Chamberlin's tangency solution but in a different context. It is, therefore, subject to essentially the same assumptions and limitations. As rival branded and advertised goods are not perfect substitutes, and because all rival firms will not normally have the same scale of plant and average cost curves, a final tangency solution for the group in equilibrium would seem to be accidental rather than necessary, although perhaps closely approached.

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

320

Stephen Enke

[June

be shifted to the right and upwards. This improved market acceptance can be exploited in either of three ways: by charging higher prices, increasing sales, or reducing selling effort. We shall assume that the full effect of changing the number of color options is countered by a price adjustment so that promotion, sales, and output remain constant irrespective of the degree of standardization. For each given number of color choices the experts must decide which colors will maximize the aggregate outlay the market

1()

0~

d: /

F-\

o0

i _~~~~~~~i
COLOR OPTIONS
FIGURE 2

.
100

will make for the constant output sold.5 If the output was 1,000,000 pairs a year, the corporation could perhaps realize $500,000 with flesh pink, $600,000 with flesh pink and rust, $670,000 with black, tan, and beige, etc.6 The offering of 100 instead of 99 color choices, however, will probably be unnoticed by the market. Thus in Figure 2 the receipts curve rises at an ever-decreasing rate and terminates in an asymptote. On the cost side
' This assumption, that the entrepreneur is able to determine the most profitable combination of colors for each different number of color options, is analogous with the supposition that he is able to determine, for each given output, the optimum number and manner of employment of the factors. 6 The realizable receipts will, of course, depend largely upon the prices of close substitutes.

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

1941]

Profit Maximization

321

there is reason to suppose that, as the number of colors increases, total costs of producing the given output will increase at an increasing rate. As sales and output are assumed constant, the receipts and costs curves represent either totals or averages. In this case the costs curve has been assumed convex from the x-axis and the receipts curve concave. fHad the reverse been assumed, there would have been no stable equilibrium adjustment. Had we assumed a cost curve

n
()

A.R0A.R. - 1930

//.

.-

~~1940

14

Lli~l

lx

OF QUALITY INDEX
FIGURE 3

concavefrom below, with marginalcosts falling, there would be a determinate solution only if the concavityof the receiptscurve were more pronounced.7 of an improvedtechnologyon costs, prices, What will be the inicidence and products? We might supposethat becausethe cost of makinga specific productis reduced,priceswill be reducedif competitionis effective;or we might argue that a more acceptableproduct, and one which previously would have cost more to fabricate,will now be sold at the customary price has continuedapacein the level. For example, although technicalprogaress automobileindustry during the last decade, the price of a certain representativecar in the low-pricefield was almost identical in 1940 with that
'The preceding analysis ignores the rare product change which does not affect both costs and receipts in the same direction.

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

322

Stephen Enke

[June

ten years before, the benefits of scientific invention for the consumer having taken the form of an improved rather than a less expensive automobile. There has been controversy over whether this outcome is owing to the allegedly monopolistic position of the largest units in the industry, as evidenced by a high rate of return on invested capital in certain cases, or whether this entrepreneurial reaction is also compatible with keen competition between rival producers. In Figure 3 variations in quality, expressed in terms of some "hedonic" index,8 are measured along the X-axis: aggregate receipts (or price) are measured, together with aggregate costs (or average costs), on the Y-axis. Output and sales are assumed equal and constant. And again, unfortunately, all disturbing factors must be "impounded" in ceteris paribus. Hypothetical receipts and costs curves for Corporation X are shown for 1930. The receipts curve is positively inclined, but, reflecting the decreasing marginal utility of additional quality, rises asymptotically. Aggregate costs will also certainly tend to increase with additional quality, but at first the increase in costs will be moderate. As technical limits are reached, however, each improvement (e.g., another mile per gallon without sacrificing other performance characteristics) will be realized only with increasing difficulty and cost. Profits are shown as being maximized with a quality determination of OX.9 Ten years later in 1940, after further invention and innovation, the consumer has modified his scale of values and now expects rather more for his money, and accordingly the price which the market will pay for a given quality has decreased. There is reason for supposing that this reduction will be greatest for the lower qualities. The market places a premium on those quality features (e.g., sealed-beam headlights, column gear-shifts, hydraulic drives, etc.) which are new or have not yet been universally introduced.10Mr. Jones is also eager to enjoy features formerly to be had only on the explensive car of Mr. Bucks. But previous improvements, such as safety glass, four-wheel brakes, and self-starters, are taken for granted and all cars have them. We can ignore the factor of styling because the 1940 fashions probably looked no better last year than the 1930 ones did then. So we assume that the receipts curve becomes lower and more positive in inclination in 1940 than it was in 1930. Compared with the cost curve
8 Cf. A. T. Court's paper, 'Hedonic Price Indexes with Automotive Examples," presented at a joint meeting of the American Statistical Association and the Econometric Society in Detroit, Michigan, December 27, 1938. Perhaps this index might include miles per gallon, acceleration, retardation, maximum speed, ratio of sprung to unsprung weight, etc. 9 The marginal curves are omitted to simplify the diagram. At OX the vertical distance between the receipts and costs curves is at a maximum. 1 While all new models may have these features the majority of automobiles, which are several years old at any given date, do not have them.

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

1941]

Profit Maximization

323

in 1930, however, the cost curve in 1940 is also lower but probably lejs positively inclined. It is reasonable to suppose that there are diminishing returns in the solution of any problem, particularly after it has been fairly satisfactorily solved; and industrial invention is characteristicallyconcerned, not with making simple things easier, but hard things less difficult. The 1940 adjustment (see Figure 3) shows that Corporation X has now selected OX1 degree of quality, for this will realize maximum profits. It is clear from Figure 3 that, if a consumer representative were given entrepreneurial authority subject to the limitation that price must not be less than cost, the quality determination in the 1940 case would again be OX1 and the price and costs would be Oy, thus eliminating profits. The receipts curve is by definition an indifference curve for the consumer;"1 similarly, if prices must equal costs, the costs curve becomes an indifference curve from the point of view of the corporation. The distance between the curves might be likened to a "surplus," i.e., cash profits for the corporation, if the price exceeds Oy, or free utility for the consumers if the price falls short of 0Y1. If we assume that the corporation always charges the maximum that it can for each degree of quality, its determination of quality per se is of no consequence for the consumer. Injury to the consumer, in so far as it may exist, would seem to result from the fact that the corporation can command a price in excess of costs for each several quality: this may be due to peculiarly low costs or to limited rivalry in the industry. The optimum determination of quality will depend upon the twist given the curves by each new innovation and not upon the resulting mean change in their vertical positions, although the latter will of course determine prices, costs and profits, once the degree of quality is known. If we can assume the receipts curve to become more positive and the cost curve less positive, then, irrespective of what happens to price, costs, and profits, technical progress will result in increased quality of product. lThe new price (0Y1) may be greater than, less than, or equal to the old pirice (OY). Whether or not relevant technological advances result in an improved product has nothing to do with the existence of monopoly elements, or abnormal profits or losses, but depends upon the twists given the receipts and costs curves.12The product will be improved if the receipts curve becomes more positively inclined relatively to the costs curve (or the latter less positively inclined relatively to the former). For a change in quality it
"The combination of price and quality is a matter of indifference to the consumer, provided it can be represented by a point on the receipts curve. The position of this curve is, of course, based in large measure upon the valuation placed by the consumer on substitute goods and the prices charged for them. A reduction in the price of a rival "make" will lower the receipts curve of Corporation X, particularly over the range of quality equivalent to that of the competitor's product. 2 What happens to price is another matter. For example, lessened competition might raise the entire receipts curve in 1940. Such possibilities have been disregarded.

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

324

Stephen Enke

[June

is not necessary, as supposed above in the automobile case, that the changes in slopes be opposite in an absolute sense. The more general requirement is simply that the changes in slopes be relative.

IV
Finally, if the entrepreneur knows the relations of marginal costs and receipts to variations in (1) output and sales, (2) the quality of the product, and (3) the extent and kind of selling effort, this knowledge must be employed in the simultaneous determination of these three strategic variables. As a general principle the entrepreneur will, "taking account of his own means . . . push the investment of capital in his business in each several direction until what appears in his judgment to be the outer limit, or margin, of profitableness is reached."'3 To give concreteness, we can conceive of the management of an existing enterprise, controlling a certain fund of circulating capital, deciding in which of the three directions mentioned above it will invest. Of course there are many alternative means by which these ends may be sought: there may be M ways of augmenting output, N ways of improving the product, and 0 ways of increasing promotion. However, just as we do not usually bother about the less efficient ways in which the factors can be combined to give a certain output, so in this case are we concerned only with the most economical way of improving the product or of increasing promotion by a given amount.14 The marginal net return on an additional dollar of working capital invested to augment output (MNPo) can be readily calculated if the customary relations of marginal costs and receipts to changes in output and sales are known. The increase in output resulting from an additional dollar expended will be the reciprocal of the marginal cost at that output range, and the value of this output will be given by the prevailing marginal revenue: for the investment to be profitable the increment of expenditure (or the cost of the last unit produced) must be recovered.15If marginal costs increase soon and rapidly, and the price elasticity of demand is approaching zero, the marginal output possible for each increment of expenditure will be decreasing and the revenue from the extra output will be falling: thus, in time, additional investment for output will no longer be recoverable.16

In exactly the same way working capital may be expended to enhance the acceptance of the product by altering its specifications. Given the relaAlfred Marshall, Principles of Economics, 8th ed., Book V, iv, 4. In each case a longer run envelope curve is fitted to the subsidiary relationships. Symbolically, MNPo = (MRo/MCo) -I where MR. and MC. are marginal costs and receipts for the appropriate output. It follows that MNPO is zero when MC. and MR. are equal. 16 Interest on the investment can be included as a cost.
13 14
16

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

1941]

Profit Maximization

325

tions of marginal costs and receipts to changes in the product,17 the marginal net profit on investment to improve the product (MNPp) can be calculated. Similarly MNPS, the marginal net profit on investment in selling effort, can be ascertained. In Figure 4, MNPO, MNPp, and MNPB (curves 0, P, and S) are brought together. The X-axis represents the rate of investment and the Y-axis the marginal net profit at different rates of expenditure.18 Profits are maximized, for any given total expenditure, when the greatest possible net area above the X-axis is secured."9If each marginal net profit is initially positive,
+Q

S~~~

:2 -

P~~~~~~

"INVESTMENT"
FIGURE

and declines uninterruptedly, the optimum adjustment will necessitate that all be equal; and, if adequate capital is available, investment will be pushed in each direction until each marginal net profit is zero, at which point MCo will equal MRo, etc. It is very possible, however, that when the index of output (or of quality or promotion) is low, marginal costs may temporarily exceed marginal receipts: in this case the curve of MNPO (or of MNPp or
As were assumed known in Figure 2. In Figure 4 we have, as it were, three parallel planes which are adjusted horizontally by management. (In many cases this adjustment will be such that all three marginal net products coincide at QQ'.) The actual investment in each direction is represented by the distance from the origin of each plane on the left to QQ': as investment is increased, the relevant plane is slid past this line from right to left, the others remaining unchanged. Theoretically these planies continue indefinitely to the right, but only the profitable ranges have been given substance by actual investment. 19 Areas above and below the axis do not represent absolute surpluses and deficits but only profits or losses in the sense that the position of the enterprise is improved or worsened.
17 18

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

326

Stephen Enke

INPS) would at first be negative, but later would rise to a maximum positive value and then commence to descend.20In these instances a certain minimum investment, sufficient to cancel earlier losses, must be made before there can be net additioins to profits, and adjustments will necessarily be discrete instead of continuous. Thus in Figure 4 we have supposed the firm to be in equilibrium when spending $700,000 for output, $300,000 improving the piroduct,and $1,000,000 for promotion. Had the total capital available for investment been only $700,000, this would all have gone for output: investment in another direction would be unprofitable unless on a scale sufficient to offset initial losses incidental to improving the product or undertaking selling effort. If various policies can prove profitable only when supported by a certain minimum rate of expenditure (because the marginal net profit will otherwise be negative), then the availability to each firm of adequate surplus capital furndsbecomes important. It is recognized by the writer that these different curves are in practice interrelated and that any change in the determination of one variable may change the positions of the other two curves. In the simple case here considered the range of variation is of course far too great. For small changes in any one variable, however, this interdependence might frequently be ignored. Another difficulty is that the increased receipts resulting from an improvement in the product, or from promotion, may not arise from higher prices being charged but from increased sales-which in turn means a change in output, unit costs, and price. Notwithstanding the many difficulties, this theoretical procedure illustrates the logic of the entrepreneurial determination of oiutput,product and promotion. It also indicates the limited role which output adjustments may play in profit maximization under conditions of monopolistic competition.
STEPHEN ENKE

Harvard University
20It will cross the X-axis twice--when MC equals MR-but will the equilibrium be stable. See Curve S, Figure 4.

only in the second case

This content downloaded from 203.101.161.82 on Wed, 20 Mar 2013 06:54:49 AM All use subject to JSTOR Terms and Conditions

You might also like