worker management and the modern industrial enterprise

19
WORKER MANAGEMENT AND THE MODERN INDUSTRIAL ENTERPRISE * A. B. ATKINSON I. The behnvior of the labor-managed cnterj)ri.se, 377,-11. The capitalist twin, 381. —III. The separation of ownership and control, 385. —IV. Self- financing and investment incentives, 390. —V. Conelusions, 391. Despite the long histor>' of interest in the idea of worker-man- aged enterprises or producer cooperatives, it is not a subject that has attracted a great deal of attention from economists. With a few notable exceptions, there has been little analysis of the behavior of a worker-managed firm and how this might differ from that of a corresponding capitalist firm. Yet such analysis is clearly essential for any assessment of the performance of alternative economic sys- tems and of the impact that increased worker participation would have on the economy. Among the small number of papers written on this subject, reference should be made particularly to the pioneer- ing work of Ward and more recently to that of Domar and Vanek.^ However, while the analysis of these authors is of great interest, the models on which it is based exclude a number of important factors that are relevant to a modern industrial enterprise. The models con- sidered relate much more happily to the case of an agricultural co- operative than to that of a large industrial enterprise, and indeed Domar's paper was explicitly concerned w^ith the Soviet collective farm. Many of the interesting questions about labor management arise much more sharply, however, in the context of the large in- dustrial enterprise. It is often argued that worker control may work satisfactorily in small-scale agricultural cooperatives, but would have quite undesirable effects in a large-scale industrial concern such as General Motors. How, for example, would the existence of substantial economies of scale affect the behavior of the labor- managed enterprise? Would it be able to plow back funds in order * I am grateful to G. M. Heal, M. A. King, J. E. Meade, and D. M. G. Newbery for their criticism of an earlier version of this paper. It has also benefited from comments made at seminars at the universities of Essex, Man- chester, and Warwick. 1. B. Ward, "The Firm in Illyria: Market Syndicalism," American Eco- nomic Review, XLVIII (Sept. 1958), 566-89; B. Ward, The Socialist Economy (New York: Random House, 1968); E, Domar, "The Soviet Collective Farm aa a Producer Cooperative," American Economic Review, LVI (Sept. 1966), 734-57; J. Vanek, The General Theory of Labor-Managed Market Economies (Ithaca, N.Y.: Cornell University Press, 1970).

Upload: varga-anita

Post on 18-Dec-2015

218 views

Category:

Documents


1 download

DESCRIPTION

Worker managed enterprises

TRANSCRIPT

  • WORKER MANAGEMENT AND THEMODERN INDUSTRIAL ENTERPRISE *

    A. B. ATKINSON

    I. The behnvior of the labor-managed cnterj)ri.se, 377,-11. The capitalisttwin, 381. III. The separation of ownership and control, 385. IV. Self-financing and investment incentives, 390. V. Conelusions, 391.

    Despite the long histor>' of interest in the idea of worker-man-aged enterprises or producer cooperatives, it is not a subject thathas attracted a great deal of attention from economists. With a fewnotable exceptions, there has been little analysis of the behavior ofa worker-managed firm and how this might differ from that of acorresponding capitalist firm. Yet such analysis is clearly essentialfor any assessment of the performance of alternative economic sys-tems and of the impact that increased worker participation wouldhave on the economy. Among the small number of papers writtenon this subject, reference should be made particularly to the pioneer-ing work of Ward and more recently to that of Domar and Vanek.^However, while the analysis of these authors is of great interest, themodels on which it is based exclude a number of important factorsthat are relevant to a modern industrial enterprise. The models con-sidered relate much more happily to the case of an agricultural co-operative than to that of a large industrial enterprise, and indeedDomar's paper was explicitly concerned w^ ith the Soviet collectivefarm. Many of the interesting questions about labor managementarise much more sharply, however, in the context of the large in-dustrial enterprise. It is often argued that worker control may worksatisfactorily in small-scale agricultural cooperatives, but wouldhave quite undesirable effects in a large-scale industrial concernsuch as General Motors. How, for example, would the existence ofsubstantial economies of scale affect the behavior of the labor-managed enterprise? Would it be able to plow back funds in order

    * I am grateful to G. M. Heal, M. A. King, J. E. Meade, and D. M. G.Newbery for their criticism of an earlier version of this paper. It has alsobenefited from comments made at seminars at the universities of Essex, Man-chester, and Warwick.

    1. B. Ward, "The Firm in Illyria: Market Syndicalism," American Eco-nomic Review, XLVIII (Sept. 1958), 566-89; B. Ward, The Socialist Economy(New York: Random House, 1968); E, Domar, "The Soviet Collective Farmaa a Producer Cooperative," American Economic Review, LVI (Sept. 1966),734-57; J. Vanek, The General Theory of Labor-Managed Market Economies(Ithaca, N.Y.: Cornell University Press, 1970).

    AnitaLine

    AnitaLine

    AnitaLine

    AnitaLine

    AnitaLine

    AnitaLine

    AnitaLine

  • 376 QUARTERLY JOURNAL OF ECONOMICS

    to maintain the growth of the enterprise, or would the rate of in-vestment be lower than in a capitalist firm? How would it be af-fected by the separation of ownership and control in a large enter-prise?

    The aim of this paper is to examine the microeconomic impli-cations of labor management in the context of a model more ap-propriate to the modern industrial corporation. In particular, it at-tempts to do the following:

    1. Provide an explicit treatment of the growth of the enterpriseand of decisions about the rate of investment. The models consideredby earlier writers are essentially static in nature and do not allowone to consider whether a labor-managed enterprise is likely to growmore slowly than its capitalist counterpart.^

    2. Examine the way in which economies of scale affect the be-havior of the labor-managed enterprise. There is strong evidencethat these are an important characteristic of a number of major in-dustries, and they are one important feature distinguishing the in-dustrial cooperative from the collective farm.

    3. Allow^ for the finance of capital formation by workers plow-ing back income. In the models of Ward and Domar, capital is as-sumed to be supplied from outside, and they therefore preclude dis-cussion of the question of insuring adequate self-finance, which hasarisen frequently in the Yugoslav debates.^

    4. Allow for the separation of ownership and control in thetypical capitalist corporation. In comparing the performance oflabor-managed and capitalist enterprises, it has been assumed thatthe latter maximizes profits in a traditional manner. As has beenstressed in the recent theories of the firm, however, there may be animportant element of managerial discretion, which affects the per-formance of the capitalist corporation and should be taken into ac-count in the comparison.

    5. Allow, correspondingly, for the separation of ownership andcontrol in the labor-managed enterprise. In any large-scale enter-prise the direction of policy would be in the hands of managers, andwhile accountable to the workers, they would (like their capitalistcounterparts) enjoy a certain degree of discretion. We therefore

    2. After this paper was essentially completed, an article appeared by E. G.Fnrnbotn, "Toward a Dynamic Model of the Yugoslav Firm," CanadianJournal of Economies, IV (May 1971), 182-97, discussing the dynamic behaviorof the Yugoslav firm; the approach, however, is quite different from thatadopted here.

    3. See D. D. Milenkovitch, Plan and Market in Yugoslav EconomicThought (New Haven: Yale University Press, 1971), pp. 214-16.

  • THE WORKER-MANAGED ENTERPRISE 377

    bave to examine whether the behavior of a managerial labor-ownedfirm would be very different from that of a managerial shareholder-owned firm.

    The plan of the paper is as follows. Section I sets out the basicmodel and examines the behavior of the labor-managed enterprisewith regard to pricing and investment (its rate of growth). SectionII compares its behavior with that of a corresponding capitalistfirm of the "Fisherian" profit-maximizing type. Section III consid-ers the issues raised by the separation of ownership and control inboth capitalist and labor-owned enterprises. Section IV discussesthe question of self-finance.

    T, THE BEHAVIOR OF THE LABOR-MANAGED ENTERPRISE

    A. The Basic ModelThe basic assumptions of the model, which are common to both

    labor-managed and capitalist regimes, are the following: *1. There is no substitutability in production between capital

    and labor.^2. There is labor-augmenting technical progress at rate y:

    (1) Lt^aKte-y',where Lt denotes labor employed and Kt denotes the capital stockat time t.

    3. There are economies of scale, and output Yt is given by(2) Yt^b Kt'^ ^ > 1 .

    4. The firm faces a downward-sloping market demand curvefor its output. This demand curve shifts outward over time at arate Go; however, by expenditure on sales promotion the firm canexpand its sales (for any given price) at a faster rate G. The netrevenue will depend on the price chosen and on the rate of growth(since selling costs have to be subtracted).

    5. It is assumed that the firm makes a once-for-all decisionabout its price (which remains constant over time) and about itsrate of growth (which is a constant proportional rate). The firm istherefore in "steady state growth." "^

    4. The model owes a great deal to the paper by R. M. Solow, "Some Im-plications of Alternative Criteria for the Finn," in R. Marriw and A. Wood,eds., T'he Corporate Economy (London; Macmillan, 1971), although it differsin the introduction of in(;rea.sing return.s to scale.

    5. This a.ssumption rules o\it any discussion of alternative choices oftechnique under different regimes, but this has already been adequatelytreated by Ward and others.

    6. This characterization is chosen for its convenience. The analysis could

  • 378 QUARTERLY JOURNAL OF ECONOMICS

    0

    T(G)

    -*- GFIGURE I

    6. Since output grows at a rate G, the capital stock grows at a

    (3)7. The net revenue of the firm can be represented as a function

    of the initial output YQ (or equivalently Ko] and the rate of growthG:

    in which the function F represents the gross revenue and the func-tion T{G)=l~s{G), where s{G) are the selling (or expansion)costs as a proportion of sales revenue. The assumed shape of T{G)is shown in Figure I.^

    8. It is assumed that the gross revenue per unit of capital attime zero {F{Ko)/Ko) has the shape shown in Figure II, For thisto be true, we require that the elasticity of demand be a declining

    function of Ko, falling from above to below this value. This

    does not appear unreasonable: for example, with 1/^ = 0.6 (the 0.6law), we would require e to be above 2.5 for small Ko and below itfor large Ko.

    B. The Labor-Managed {LM) FirmThe basic assumption underlying the work of Ward, Domar,

    and Vanek is that the labor-managed enterprise aims to maximizebe generalized to the case of firms choosing an output policy Y(t), but con-centration on steady state behavior does not seem unreasonable at this stage.

    7. In particular it is assumed that T(G) = l, T(Gm.i)-0 and that{ T'lT) is an increasing function that takes the value zero at Go and tendsto infinity as G-*Gn,^j..

  • THE WORKER-MANAGED ENTERPRISE 379

    K,

    the income per worker (the rate of dividend), an assumption de-scribed by Vanek as "the natural and rational concern of all par-ticipants in an enterprise." While it clearly does not capture allaspects of the motivation of the cooperative firm, it does not seem anunreasonable "stylization" on which to base the analysis. However,we have to consider the modifications of this objective required bythe introduction of dynamic considerations. In contrast to theearlier, essentially static models, we have to make some assumptionabout the way in which the time stream of dividends would bevalued by the workers. The assumption made here is that the firmaims to maximize the present value of dividends at a discount rate5. There are good reasons for expecting this discount rate to be high,and these are discussed further in Section II.

    In the first part of the paper it is assumed that all capital isfinanced by borrowing at an interest rate ;. There are no interme-diate inputs or hired workers. The net income of the LM firm istherefore given by (for simplicity, depreciation is ignored)

    Ro e'^^-jpkKt,where pk denotes the price of capital goods (assumed constant). Thedividend rate is equal for all workers, giving per worker ^

    8. For discussion of "egalitarian" and "inegalitarian" LM firms, see J. E.

  • 380 QUARTERLY JOURNAL OF ECONOMICS

    (5) dt = -aAo a

    The present value of dividends over an infinite horizon is given by(6) Do - 7r . / . xs -^

    aK,y I I 1 \ \ a$

    ( it is assumed that ^ ^ 8 y> I 1 )G,nas )The LM firm will choose its scale of operations (Kf,) and its

    growth rate [G) to maximize Do. Let us consider first the scale ofoperation, which is the aspect discussed by Ward and others. Themaximization of Do simply requires that gross revenue per unit ofcapital be maximized, i.e., the choice of K*o in Figure II. In thedetermination of the scale of the LM firm, there is no differencetherefore from earlier results.^ The determination of the growthrate is less straightforward. Differentiating with respect to G, weget the first-order condition,^

    which may be written (where a = 9/{l l/f>.)),

    T{G) / a-G'The LM firm equates the marginal cost of an extra 1 percent ofgrowth (in terms of revenue foregone) to the present value of thegain arising from the economies of scale. If there were no economiesof scale (/i.= l ) , the LM firm would choose Go; it is only the econ-omies of scale that make it at all interested in growth. I t is interest-ing to see that the rate of growth chosen is independent of the rateof interest charged on capital (j). I t is also independent of thechoice of KQ. The growth rate chosen, however, is a decreasing func-tion of 9. The higher 6, which may be seen as an "effective" discountrate, the lower the growth rate (as would be expected) ; and in theextreme case, where the workers were concemed only with the currentdividend do, the growth rate chosen would be GQ.^

    Meade, "The Theory of Labour-Managed Firms and of Profit Sharing," Eco-nomic Joumal, LXXXII (March 1972). 402-28.

    u

    9. At the point K*o, the elasticity of demand^ , which is the resultM - 1

    obtained by Vanek, op. eit., p. 102.1. It is assumed that the shape of the functions is such that the second-

    order conditions are satisfied.2. For the enterprise to continue operation, it is clearly necessary for

  • THE WORKER-MANAGED ENTERPRISE 381

    II. THE CAPITALIST TWIN

    In the work of Ward, Domar, and Vanek, the behavior of theLM firm is compared with that of a capitalist counterpart. This"capitalist twin" is defined by Domar as having "the same produc-tion function and prices as the coop, and with a wage rate initiallyequal to the coop's dividend rate." The aim of such a comparison(which is not made very explicit by these authors) is to separatethe differences due to the form of organization from those attribut-able to changes in the distribution of income. The introduction oflabor management can in this way be decomposed into two parts:a change in organization with no change in the reward to labor,and a shift in the distribution of income within a given organiza-tional framework. It is with the former that we are concerned here.

    In this section the behavior of the LM enterprise is comparedwith that of a corresponding capitalist firm facing the same marketconditions, which aims to maximize the present value of profits. Wefirst set out the basic model and then specify the precise basis forcomparison.^

    A. The Capitalist Profit-Maximizing {CPM) FirmThe capitalist firm has the same net revenue function, but has

    to pay a wage itv to its employees. It finances its capital formationout of retained earnings.^ The dividends paid to the shareholdersat time t are equal to the net revenue minus wages minus retainedearnings:

    ' K

    The wage rate is assumed to rise at a rate y : Wt^Woe'''. The stockmarket value of the firm is assumed to be equal to the present valueat the market rate of interest i of the dividend payments to share-holders;

    lG ilG/ix) L M -^where it is assumed that i > Gmax (and hence > Gmas/;^ ) it to "break even." It may be noted that Do"'">O does not imply that thecurrent dividend is positive, since in maximizing Do. the enterprise balancescurrent against future dividends. It seems reasonable to suppose, however, thatthe current dividend must be nonnegative, so that the growth rate cannot be in-creased beyond the point where

    R,{K,*, G)=j p*Xo*.3. The model corresponds to that of Fisher, Inc., in Solow, op. cit.4. In the model considered here, issuing loan capital at the market rate

    of interest makes no difference to the firm's behavior, and it can be shownthat new issues cannot increase the finance available.

  • 382 QUARTERLY JOURNAL OF ECONOMICS

    The CPM firm is assumed to maximize tbe difference betweenits stock market value and the value of the capital employed:{10} Zo-Fo-PfcXo

    The scale of the firm is determined simply by

    From this we can see that if market conditions are such that theCPM firm can achieve a strictly positive value of Zo, its scale islarger, and therefore its price lower, than with the LM firm (seeFigure II). This result corresponds to that of Vanek. '^

    In choosing its rate of growth, the CPM firm sets ^Ro Koii-G)

    (12) ( _

    Where Zo>0, the growth rate is greater than Go even if tbere areno economies of scale, so that we can see at once one significantdifference between the two types of firm. By combining (11) and(12), the first-order condition for G is

    i-T'jG)) 1 J / Ko ZRQ \ 1TiG) i-G

    1 1-iAi-G i

    where e is the elasticity of demand (which has been assumed to bea declining function of Ko).

    B. Does the Capitalist Firm Grow Faster?In the present context the basis for comparison between the

    LM and capitalist firms is less straightforward than in the staticmodels considered by Ward and Domar. The dynamic model con-sidered here introduces two new elements: "^ (1) the comparison ofa time path of payments to labor rather than just the current re-

    5. Op. cit.. Ch. 6.6. Again it is assumed that the second-order conditions are satisfied.7. A further important factor not discussed here is uncertainty and the

    relative riskiness of different types of income.

  • THE WORKER-MANAGED ENTERPRISE 383

    muneration; and (2) the specification of the rates of time discountapplied by shareholders and workers. The assumptions made hereare that the present value of wages paid by the capitalist firm isequal to the present value of dividends: iDo = Wo/{^ y)), and thatthe rate of discount used by the CPM firm is equal to the rate ofinterest paid by the LM firm ii-j). From (6) we may see thatthese assumptions imply that ^

    (14) H

    In the case of the discount rate applied by the workers (S), it mightbe felt that this should equal {; on the other hand, it has been arguedthat the absence of any property rights in the firm may lead 8 tobe greater than i. Since the return to investment accrues to thoseemployed in the firm and not necessarily to those who made the in-vestment at an earlier date, it may be expected that workers will bereluctant to save in tbis form if other savings media are available.As Furubotn and Pejovich have pointed out, with a horizon of tenyears, a retum of 13 percent is required on nonowned assets inorder for the worker to receive the equivalent of 5 percent on ownedassets; and that for a horizon of twenty years, the equivalent rateis 8 percent."* This suggests that there are good grounds for ex-pecting that S>{. However, we saw in Section I that the "effective"discount rate is not S but O=B y (y being the rate of technical prog-ress). It may therefore be possible that the effective discount rateapplied by the LM firm when we allow for technical progress ia lessthan i. In what follows, the dependence of the comparison ona{6/{1 1/fi)) is examined.

    In comparing the growth rates of the LM and CPM firms, let usstart wath the value of a applicable to the former. This determinesfrom (8) the rate of growth of the firm and hence from (14) thevalue of H, since KLif.=Ko*. It may be noted that GLM and H aredeclining functions of a. The value of H in turn influences GCPM,through its effect on Kcpn and hence in (13). It may be shownthat GCPM is a declining function of H and hence an increasing func-tion of a, SO that we have a situation such as that shown in the upperpart of Figure III.^ There is therefore a value a such that for a

  • 384 QUARTERLY JOURNAL OF ECONOMICS

    -a

    FIGURE III

    the first-order conditions indicate a faster rate of growth for theLM firm, and for a>a' the CPM firm grows faster. Moreover, acan be seen to lie between i and i/(l l//i,) from an inspection ofthe right-hand sides of equations (8) and (13). It should also benoted that ZQ is an increasing function of a (as shown in the lowerpart of Figure III), so that Zo is negative for lower values of a, andat these values the capitalist finn would not produce. In particular,at a, ZQ has the sign of

    The first term is less than or equal to unity (from the definition ofKQ*). From the definition of aiGLj^^Gcr^) the right-band sides of(8) and (13) have to be equal, i.e.,

    1 _ 1 (1- lA)a-G' ~ i-G' i-G'/p. '

    / cannot, both be positive; and using the assumption about e, we candeduce from (13) that they have the same sign.

  • THE WORKER-MANAGED ENTERPRISE 385

    which can be rearranged to give

    i-G'

    i-G'/1.Now Zo>O implies \/t-\-\/\x.>\, which in turn implies from (16)that the second term in (15) is greater tlian unity, and hence that2o

  • 386 THE WORKER-MANAGED ENTERPRISE

    A. The Majiagerial Capitalist {MC) FirmThe model of the managerial firm discussed here is based on the

    work of Marris as developed by Solow. According to this model,managers aim to maximize the rate of growth subject to a take-over constraint.^ This constraint is usually formulated in terms ofthe valuation ratio,(17) Vo^mp.Ko,where m is a constant ( m ^ l ) . In other words, if the stock marketvalue of the firm falls below a certain proportion of the value of itsassets, it is subject to the risk of a take-over bid.* This constraintwill, in fact, hold with equality, and can be written as Zo =- (1-m) pk Ko or(18) - - - = [i-Gj mpfc+ : - - - .

    Ao "- lli/ix -"

    For any Ko we can find the largest G that satisfies (18), as illus-trated in Figure IV. From this it is clear that the managerial firmchooses Ko to maximize R,)/Kn, i.e., the same scale of output as theLM firm.*^ The output is different from that of the CPM firm unlessin that case Zo = 0.

    That the scale of output is the same for both the MC firm andLM firm is explained first by the fact that both types of firm have aratio maximand. In effect the MC firm maximizes the average rateof profit, since it is the average rate that would interest a potentialtake-over bidder he cannot (in general) bid for part of the firm'sactivities. In this respect, therefore, the MC firm resembles the LMrather than the CPM firm it could be seen as a "capitalist co-operative." Secondly, the assumption of fixed coefficients meansthat maximizing income per worker is effectively the same as maxi-mizing profit per unit of capital as far as the scale of operation isconcerned.

    In order to eompare the growth rates of the LM and MC firms,let us assume as before that the present value of the payments per

    3. R. L. Marri.s. The Economic Theory oj 'Managerial' Capitalism (Lon-don: Macmillan, 1964).

    4. The form of the valuation ratio constraint depends on the assumptionsmade about the extent to which a take-over bidder could exploit the existenceof economies of scale. The assumption made here is that the bidder wouldassess the potential profits on the basis of running the firm as a separate unit.

    5. The solution obtained here is different from that of Solow, since thecombination of increasing returns to scale and the assumptions about the priceelasticity of demand insure that Ro/Ko has a maximum.

    6. It should be noted that these conclusions depend on the assumptionthat managers are interested in growth and not in size. If the managerial utilityfunction included both g and Ko as arguments, then the scale of operationcould clearly be greater than Ko*.

  • THE WORKER-MANAGED ENTERPRISE 387

    R.H.S. of equation (18)

    FIGURE IV

    worker are equal and that i = i. Let us consider first the case wherem = l. If we denote by ZQ* the value for a CPM firm with K = KQ*,then it is clear that for H such that Zo*>0, the growth rate for themanagerial finn is greater than that of the CPM firm. Hence, wemay deduce from the results of the previous section that GMC>GLMwhere H is such that the capitalist firm can break even (seeFigure III). However, if m< l , then it is possible for low values ofa that Gi_M may exceed GMC but that the MC firm continues opera-tion (with Zo

  • 388 QUARTERLY JOURNAL OF ECONOMICS

    imposed on them. In the latter case, it seems reasonable to supposethat the workers are concerned with the level of the dividend andthat there is a minimum dividend below which the workers dismissthe managers (or leave the enterprise). The situation in Yugoslaviahas been described by Pejovich as follows:

    The institution of the Workers' Council is reminiscent of the role playedby stockholders in a free-market economy. . . . The workers, thougli they donot know much about the art of management, know surprisingly well howto compare their earnings with the earnings of other workers. . . . If theyfind their earnings lower than those of other workers, the director . . . mayfind himself fired.^

    This constraint could relate to current dividends or to the presentdiscounted value, but from the description given by Pejovich, theformer seems the more plausible assumption.^ The objectives of themanagers are clearly very difficult to specify, but there are somereasons for expecting that they will he interested, like their eapital-ist counterparts, in the rate of growth of the enterprise. It seemslikely that remuneration would be related to the size of the firms,hence providing the same incentive for growth as in the capitalistcorporation. As Marris has remarked, "the world over. Presidentsare paid more than Vice-Presidents, Commissars more than Deputy-Commissars, . . . , and middle managers more than junior man-agers." It may also be reasonable to suppose that growth is animportant political objective and that managers would be more re-sponsive to political pressures than workers as a whole. Accordingto Furubotn and Pejovich,^ "since [the director] represents the cen-tral government, he is particularly aware of national purposes andpriorities; at the same time, he knows his long-term career is likelyto be influenced by the degree to which his guidance of the firmcontributes to the realization of national goals. On this logic, themaintenance of relatively high investment rates by the firm mayseem essential to the director." In what follows it is assumed thatthe managers of a LM enterprise aim to maximize the rate of growth,an assumption similar to that adopted by Furubotn and Pejovich.

    On these assumptions, the aim of the managerial labor-ownedfirm can be seen as maximizing G subject to

    8. S. Pejovich, The Market-Planned Economy of Yugoslavia (Minneap-olis: University of Minnesota Press, 1966), p. 91.

    9. The constraint could also be seen as a supply function for labor in theway described by Domar: a single LM firm in a capitalist economy could notpay a dividend less than the going wage rate unless the workers are willingto accept less for idealistic or other reasons.

    1. Furubotn and Pejovich, op. cit., p. 439.

  • THE WORKER-MANAGED ENTERPRISE 389

    (19) ado = --^i Pk'^a d'.As before, the firm chooses Ko such that Ro/Ko is maximized. Thegrowth rate, however, is increased to the point where (19) holds withequality, and if d' is less than the current dividend paid by the LMfirm, the MLO firm grows faster. To this extent, allowance formanagerial discretion tends to narrow the gap between capitalistand cooperative enterprises.

    With the objectives assumed here, the separation of ownershipand control leads to a rise in the growth rate in both labor-managedand capitalist firms. Whether the MLO enterprise grows faster thanthe managerial capitalist firm depends on the degree of managerialdiscretion. If the MC firm pays a current wage Wo, and m 1, thena necessary (but not sufficient) condition for GMr.o>Oifc is that

    _a^ ( ^ _ ^ 1 j ( 1 '! ~ ,^ ) - l .If C denotes the ratio of capital to labor costs in the MC firm

    ( 1, this may be written^ au'o'

    (20) < 1 - -The right-hand side of (20) is evaluated for different values of Go,l/fi, and C in Table I. From this it is elear that where there aresubstantial economies of scale, for the MLO firm to grow faster, it isnecessary that there be a considerable degree of managerial discre-tion; for example, where l//x = 0.6 and Go^4 percent, the MLO firmwould only grow as fast if the dividend requirement were less thanthree quarters of the wage paid by the MC firm.

    TABLE I

    c

    c

    l//. = 0.9

    10.50.25

    l//. = 0.6

    10.50.25

    2%

    0.950.960.97

    0.820.860-89

    Rate of Growth Go

    4%

    0.870.910.92

    0.580.680.74

    6%

    0.740.800.84

    0.250.440.53

    Note: i = lQ%

  • 390 QUARTERLY JOURNAL OF ECONOMICS

    IV. SELF-FINANCING AND INVESTMENT INCENTIVES

    Throughout the foregoing analysis we have retained the as-sumption made in the work of Ward and Domar that all capital em-ployed by the enterprise was supplied from outside. In reality, how-ever, the enterprise is likely to have to rely in part on self-financeand a proportion of the funds for investment will have to comefrom the reinvestment of profits. The institutional background tothe plow-back decision in Yugoslavia has been described by Milen-kovitch as follows: There arctwo possibilities, one based loosely on Yugoslav experience to date, and theother drawing from Yugoslavia as it is scheduled to be after 1970. In the formerversion, the chief source of new investment funds is the tax on socially ownedcapital, which may be regarded as the interest paid to society for the usethereof. The proceeds of this tax are allocated, through the banking system,to the enterprises seeking investment loans. The latter variant has no capitaltax and the principal source of investment funds is enterprise profits.^

    In order to examine the consequences of these two systems, letus assume that a fraction A of total eapital is financed by borrowingand (1 A) by retained earnings. In this situation the present valueof dividends becomes

    Ro 1 P'^^' ^0 i0-Gil-l/,)) a

    The scale of operations is unaffected. The first-order condition fordetermining the growth rate for the LM firm discussed in Sections Iand II becomes(21) ( -^^A igp p,Ko(l~\)U~G)

    Insofar as capital has to be financed internally, this reduces theright-hand side and hence the growth rate chosen; investment islower in the second of the situations described by Milenkovitch thanin the first. Moreover, this is independent of the terms on whichloan finance is made available a reduction in A decreases thegrowth rate whatever the interest rate ; (assuming that the enter-prise can continue to break even). However, the eonclusions reachedfor the managerial MLO enterprise are rather different. As set outin Section III, the firm chooses the highest G consistent with

    In this case, the terms on which loan finance is made available are2. Milenkovitch, op. cit. p. 222.

  • THE WORKER-MANAGED ENTERPRISE 391

    important: a reduction in A only decreases the growth rate wherej

  • 392 QUARTERLY JOURNAL OF ECONOMICS

    that each of these factors is important. The main conclusions maybe summarized as follows:

    L The rate of growth of the labor-managed enterprise dependson the extent of economies of scale and on the workers' discount rate.If there were no economies of scale, the enterprise would not beinterested in growth. The growth rate is independent of the rate ofinterest.

    2. If the premium over tiie return on owned assets required bythe workers to invest in the enterprise exceeds the rate of technicalprogress, the labor-managed enterprise chooses a lower growth ratethan its capitalist counterpart. If the premium is less than the rateof technical progress, then either GcpM>Gt3i or the correspondingcapitalist finn cannot break even.

    3. The separation of ownership and eontrol in the capitalistfirm makes a major difference to the basis for comparison. If labormanagement replaced not a Fisherian capitalist but a Marris-stylemanager maximizing the rate of growth, the difference in pricingpolicy would disappear (given fixed coefficients of production).

    4. The separation of ownership and control leads both labor-owned and eapitalist firms to grow faster, but if there are significanteconomies of scale, a considerable degree of discretion is requiredfor the managerial labor-owned firm to grow as fast as the mana-gerial capitalist firm.

    5. If the government wishes to increase investment, reductionsin the price of investment goods are likely to be more effective thanvariations in the rate of interest or in the availability of externalfinance.It should be emphasized, however, that these conclusions are derivedfrom a highly stylized model of the behavior of a cooperative enter-prise and, while the formulation is a natural development of earlierstatic models, alternative approaches may lead to rather differentresults. Two qualifications in particular should be stressed: thespecification of the objectives of the LM firm may well not be anadequate representation of the goals of a cooperative enterprise; andno account has been taken of the distributional changes that wouldaccompany the introduction of labor management.UNIVERSITV' OF ESSEX