LIBRARY OF THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY [ii^ASS. 'iNOT. p 7^ ALFRED P. A WORKING PAPER SLOAN SCHOOL OF MANAGEMENT Dynamic ilodpl of the Regulated Firm under Uncertainty* by Hart I n. fiulirahmanyain **• -[VX'Ti MASSACHUSETTS TECHNOLOGY 50 MEMORIAL DRIVE CAMBRIDGE, MASSACHUSETTS 02139 INSTITUTE OF TECH. KOV 2 74 DEvVtY LIBRARY A Dynamic Model RG,'^ulntG(i of the Firm under Uncor ta i nty->v by Marti "" There are no pages numbered ** niibrdhnianyain n. 1 through ^ m-T^ 6. wish to acknov/l odno tlio guidance and helpful comments of my thesis committee Professors Paul MacAvoy, Robert Merton and ote\;art Myers. I PACE A Dynar^.Ic Model the Regulate'-! of ^ \ 7 rm Under Uncertainty 1. Introduction Much of the work done In the regulated firm has Averch and L.L. employs model) producln;^ a recent years on the theory of been influenced by the work Johnson. The Averch-Johnson model static framev/ork a homogeneous single for good, and facing markets for the two Inputs (labour and capital) monopoly firm uses a is subjected capl tal -labour to ratio (the A-J monopoly a major proposition emerging from this model, of H. firm perfect The It uses. that when the is rate-o^-return regulation. higher than the one It that minimizes cost for the level of output It elects to produce. This result and several corollaries have been clarified by several authors. The A-J model has been analytical simple. of However, omitting some practice. 1. Insights, In In providing added advantage of being extremely fruitful and has the the simplicity Is of the extended and 2 Important achieved at the expense aspects of regulatory Many of these abstractions from reality have been (1). of the See P-aumol and Klevorick (6) for a good survey literature. A more complete and up to date exposition Is provided by Hilley (2). 2. PAGE onrllcr and polntorj out analysis the regulatory such hy lap:," researchers other maximization of maximization, and problems e.r;. soc al -wel I f Implications the Into of for Important several are.' Hov;evor, the repiulatory problem have aspects of Incorporator! the choice of a different objective function revenue as few have boon a 8 received relatively limited attention In the literature. The A-J model and Its Leland, the exception of With uncertainty. the received on the subject. The firm In the A-J model demand the problem the work of Is Myers and theory has surprlslnrjy little problems of uncertain Input V/hen extensions Ip;nore the problem of of adjusting does not face the prices or uiipredl uncertain, the firm Its combination satisfy demand, provided, of course. cta!^ Is 1 e demand. with the faced Is of inputs It to say so as to profitable to do This would require us to dlstlnr;ulsh between relatively so. flexible more capital; this, factor, in turn, labour, has and the Implications for durable tlie Input, allocation 7 dccl s Ion. The static view of the Ignores an r'^rulatory problem taken by A-J Important regulatory Issue. In the A-J model. See Ralloy and Coleman (3). See Bailey and Ma lone (4). For Instance, Klevorick (10) and SheshlnskI (15). 6. In (13) and (14), and (12) respectively. 7. The effect of the greater flexibility of labour relative to capital, on the allocation decisions of the monopoly firm, will be the subject of a separate paper. 3. h. 5. PAHF the firm rGmnlns continuously It There no scope Is related A chan^f^/ Issue constraint Is hlndin^T/ nature the Is This behaviour of Is clearly the rer;ulatory regulated firms, action by earns exactly hase at all inconsistent with the which do some of points In observed earn more causing regulators to than the allowed rate-of-return, appropriate of Implies that the firm rate-of-return on its rate the fair throujrh The A-J assumption that the fair rate-of -return constraint. time. demand conditions for chanr;Ing time and technological once and consequently output and price. Inputs has chosen Its the optlmuin position, In 9 o pricrv., r(?dijcing v/h i 1 e others talce earn less than the fair rate and consequently petition for a rate Increase. An Important feature of the actual captured by the not price the firm A-.J version. can charge based criteria, rather than specifying constraint. Adjustments of the are undertalcen on the basis of of the firm I.e., whether or not than the allowed Is regulatory process that regulators fix the on the an explicit ratc-of-return rate-of-return price downv/ards or upwards, the It rate-of-return. past profit performance has earned more or less Thus, regulators act Is provided by Klevorick (11). The observed practice. of styllzation This Is a "reduction" process consists of denials of furtlier proposed rate Increases. In a period of rising factor costs, this In effect reduces price in real terms. 8. 9. A detailed discussion PACE 10 specification of through the rather than charge, all points rate-of- return that this mechanism Is should be noted fact that the the price time, unlike the regulated here, firm In a revision. specifying by Uncertainty Is Is are as many of 1 deal 1 ."re'!, dynamic a explicitly since contingencies, securities or, today, particular for futiire oversimplification a unit of it Is assumed Thus, the number alternatively, of proposed the this 's an It Is uncertainty. be gained of the behaviour uncertain setting. the A-J theorem Is In problem In an and regulated a received of and output decisions uncertainty. exists there VJhlle of are contingency consumption This note studies the Implications analogue of there future case, the tiiere tiiat, of nature. of the regulated firm. rate for states hoped that some undertanding can decisions formula employed here. is In either the used model Introduced through the possible every for (Arrow-Deb reu securities). the Input The Independent securities available as future price context. price-setting function of the state-preference frnmev/ork that uncertainty binding at Is not study the behaviour of to Is dynamic attempts to capture the regulators, It the A-J model. In The objective of this note ' can earn.. It consistent with the the rate-of -return constraint In firm can that the In of such of the regulated the dynamic proved. and a model for firm. An context, under The capital unregulated firms stock are PAHH 11 compareo, when ref?;u1atlon by the re^^ulators no Is "effective," I.e. the price set Is more than what unre/rulated an monopolist would have chosen, under the same demand and cost conditions. The ex post efficiency of allocation of Inputs of the two firms are compared. Section describes 2 firm's environment the with a characterization of the uncertainty. actions are modelled In Section The regulator's The 3. together section next describes the regulated firm's problem and characterizes its optimal policy. optimal policy Section 5 considers the unregulated firm's and compares It with that of the regulated firm. 2. The Firm's rnvlronment the In describe state-preference framework conditions uncertainty, characterized by listing employed the In the possible here future "states of to are nature" that can occur In the future. Out of the can occur random set of possible states of at a point variables are In be described by 10. For a The realized values time. contingent occuring, and uncertainty listing of all nature, only one on the state of the variables under possible contingent detailed discussion, see Hlrshlelfer (7). of the nature study can values. PAGE 12 that a particular given Thus, relevant variables are known values of the By considering all future periods decision making, and obtain In ranges from 1 the future are to M^ regulated The producing a period the t, and , firm (&,t) time t Is q p ( The revenue obtained by occurs .q is D , If The firm's output p. It ). period regarding the labour to be and can be . monopolist a beginning of a price given that the firm satisfies period L^ be assumed that Is t has at the beginning of the period, hires during the where It Is q' In state (p^XO. state (O, t) the entire demand. depends on K, on p, allowed to the firm Is that the firm, In The states the demand obtaining / the (f),t) At the regulatory commission fixes Given the price In to single homogeneous good. charge. of to T. 1 assumed of past experience, the basis under study. Indicated by Is nature In each complete picture a from t the relevant for purposes of uncertainty regarding the variables of nature occurs, with certainty. the possible states of one can period, time state of nature , the stock It and the labour It assumed that decisions employed are completely flexible made once the state of nature firm's production function remains Is revealed. stable through The tl mev-'-The 11. be straightforward Introduce a variable It would to corresponding to technology (analogous to capital stock), as proposed by Klevorick (11), wl'-ich results from tnvostnont in research (analogous to investment). PAGE 13 production function 01 ^ Incorporating -f oppor tin 1 1 les. In ^O L assumptions marginal productivity convenient more ^O K . positive, of but productivity of both factors, as well as other factor Is Increased. Is as (K,L) usual the declining/ marginal Increasing Is v/rltten to For factor, when of one purposes of our analysis. express It production firm's the the terms of the labour-requirements function. 12 (2) The perfect markets firm faces adding to its stock of capital. at the beginning the firm Investment l^ of period (t-1) at the ; it hiring In of period t, 12. The properties follov/ from those of It Is its is stock of which can be Increased, t. In The decision to be made by has a capital beginning of period labour, and K^._, level of at the end but not decreased, assumed that capital labour-requirements function the of For example. the production function. PAGE Ih 1 ri does not depreclater^Hence/ and a price speclflerj hy the .at the given If ) for given may It required to rate w. price. and the demand firm for the This will happen This level Is a when ^-^ QdC^. , p^ q ) X t" ('^c' However, (p ), It labour at the fixed wage Thus, the maximum and ) (p^ q hire the >p/w. function of K^ defined to be ) Is K, The state fully. to (p^ q . hire labour at high enough values of produce the quantity Is -^^ re;;ulators of p satisfy the demand profitable profitable output w. p stock capital a The firm would have to wished to K not be with t revealed to be (©-/t), of nature Is /Kj- period starts !n The firm >0. I, p /for given .^5 13. It relatively simple to Incorporate fIxe'^ Is a depreciation rate for capital which v/ould modify the Investment constraint from one on net to one on gross Insights would be Investment. However, additional no forthcomi ng. model that between 14. It In this the period Is assumed the prcHuctlon period regulatory reviews Is the same as I.e., the period In which the capital Investment decision Is made. In many cases, the production period Is shorter (e.g. between reviews three one year) than the period (e.g. years). In such a situation, the qualitative conclusions of regulated firm The the present model would bo preserved. price as given , between acts as a profit maximlzor, taking fact that the regulatory In view tlie reviews, but keeping affect Interim will capital stock decisions made In tlie The next review. price specified by the regulators at the relative to longer the period between regulatory reviews, the production period, the less the regulated firm deviates with from tlie behaviour firm that optimizes value, of a the period becomes given price. the limiting case when In production period, the Infinitely to the large, compared firm takes price as given and optimizes value. It should be to the result of Hal ley noted that this is somev/hat similar and Coleman extent of overcapitalization of (3), that the the regulated firm varies inversely with the regulatory lag. PAGE 15 The possible states dIvMed Into demand Is of nature one containing two sets, satisfied and fully \/here excess demand exists. are define a number M^ which demand function demand Is an^ all profitable states output satisfied In Increases. Kj. As ,p^ ) In price p the nature of Increases demand/ one falls, Hemand can stock fij- Kt- Is thus the given by If demand satisfied ^ y/ ,. \ ft-Vc- ^^'^%,r^'<C') , be ' The "operating profit" In state (O/t) „ Is a t maximum the and the number the capital In Increases/, and Increases. maximum profitable output Increases. can the state of nature to more states of nature For those ncl udl ng The number H^ In period . Q(Kj. I that states of nature Increasing corresponding just met. of p^ In of can thus be t those states" where other the Assuming the order arranged In In period (0,t) ^ Is I.e. f-V 15. It Is assumed that_ the wage rate w Is constant and certain through time. 0.( !', function of K p ) Is a concave and At 0, dX(^,k.)/^a.=p/\^ ^ p. price ncroases, '^^5-*>^ As Increases, but since c'-^''('^^^k)/.>& >q^ o Increases less than '^'^ dX(6i,K} /<.Hi proportionately. As K Increases, decreases, but si nee ^^^''^'^''^)/'-''^^^'^<0, the Increase In n to compensate for the Increase In K Is less than proportionate. 16. It Is assumed that the deman'-' functions in the various states of nature do not Intersect i.e. demand functions are monotoni cal y Increasing In the Index the state of of nature, 0, at every price. 17. As K Increases, dX/do, decreases In every state of the OX./dts world, as O'X /'>^ck <,o. As Increases with and consequently v/i th the number of the state of the world, the value of M^ Increases so as to make DX/Jaj =p/w . 1 I PAGE 16 p^SCKt^.pt) -uj (3) X C^C^t,Pf), if demand I.e. /<-(.) met defined to Is heglnning of the he the market value as of first period, of a contract unit of consumption In state O^t). horl zon I gl s period Is t ir (5) time 0, the The value of contlnj^ent the decision ven by The present market value of In not that pays one the beginning of state (0/t) as of Income In Is (0^t)>l'i the firm's net operating Income given by 0o.t Lft V -"" ^ (u.t^Kt)i ^;t where KL demand. stock Is of nature with represents the state It Is unity. assumed that the price of Thus, all prices a the highest unit of capital and costs are scaled to units of the price of capital. 3. The Regulatory nnvlronmont The regulatory commission uses Information on costs and demand and firm's current rate-of- return - the latest available sets prices so that the based on current capital PAHH 17 Stock/ level. demand and cost conditions prices fixed for period (t + 1) t's output at capital earning exactly the In period stock mechanism, and as achieved at any to Implement costs, output, reduced to the "fair" Is followed by the regulators The rule the firm's - t. such, fair -^Thls the the and capital v/I 1 result In 1 rato-of-return Is only chance. regulators examine stock In period t on the adjustment an fair rate-of-return point In time, except by this rule, that period Is Is In not order the firm's and set the price for period (t + 1) according to the followlnfj formula. the period of This Is a CvTse of lagged regulation v/I th Dalley and Colenan (3) show that the lag being one period. Is diminished when the the A-J effect of overcapitalization extent of the regulatory Introduced; lag Is overcapitalization varies Inversely with the length of the Baumol and Klevorick (6) sketch out the implications lag. of regulatory lag. In a model where the firm makes decisions investment in research, to labour, and regarding capital, the model, where Increase the stock of knowledge. In a I'levorlck (11) period between revieivs Is stochastic, Investigates the relation between the investments In capital and the expected period between stock and research , detailed Me without mnlclng reviews. concludes that production and regulatory assumptions regarding the demand, predict the possible to reaction functions. It Is not the firm's lag, on direction of the effect of regulatory this references on For other allocation decisions. Also question, see "onhright (7), Baupiol (5) and V/eln (16). see footnote (I't). Baumol and the views of 19. This spirit to Is similar In regulation. To of the working Klevorick (R), regarding the latest by taking procep'ls, roughly quote, "regulation and .Information available on company costs an^ demands, curmnt firm's resetting that the prices so stock, current capital its rate-of-return - based on technology, demand conditions, and so forth - is reduced to the "fair" rate level." See also Klevorick (11), p. 53. and Davis (8), p. 271. for similar opinions. modelling the action of that, Apart from the fact 18. PAGF. (6) p where ^^ i^yt^KQ the regulator's price for period !s p, 5«ir ^ ^ 18 (t+1), s Is tt-i the fair rate-of-return, and wX(Qj,/'K t. Thus earn the actual represents the actual output In period exactly proportion a course, s of the distribution of the price p the firm to stock K^ Its capital that the output turns out to bo q . for (f^,T)s^M, ft. ^K,4-ujm(Kt,r^J<^^^^^ probability where , by Is given sK^^ajX(%i-,Kt) ^.^^ probahllityTT (7) t labour costs In period represents the price that allov/s p, provided, of Thus, ) Is q TTq ^, Is the probability of occurence of 1 - f.' tt state (f>,t). regulators as setting the price the firm can charge , rather than specifying rate-of-return constraint. Is more a realistic. In the context of uncertainty, there Is another argument In favour of Kx ante price-setting regulation. the regulators cannot set the "fair" rate-of-return , since the rate-of-return is stochastic variable. a / Setting the central the mean) does parameter not (say, precisely constrain the firm's actions; there Is no v/ay of determining ej< post , whether the firm followed the regulator's rate-of-return formula. The firm can "cheat" by actually aiming ex ante for a higher rate-of-return, and claiming ex. post that this was the result of an outcome on the positive side of the distribution of returns. This Is similar to the "cheating" described In Myers (I'O, earn more or price sotting, tho firm can still p. 31'i. In by loss However, than rate-of-return. the "fair" constraining price explicitly, the regulators take away the r\nf^Ton of freedom the firm neofls. In order to be able to "cheat." In addition "effective," the greater than what an I t wM 1 ^' Is re;^u1atIon If firm's re/^ulatpd Is should price unregulated monopolist n to be not be would charge. horizon of the firm he assumed that the Its decision ' formula. this to PACE In making T periods. The Firm's decision At beginning the decision to Is of optimize the "operating profits," net Investment costs of first the period, present market present of the the programme In the firm's value of Its market value succeeding of periods. The firm's objective function Is Max (8) where such that '^rs'- l!9^..i T ^ ^i~^\:-\ [ I > o ^^JL^J-±hLl,lLL with probab -^ii-L^!?a,,ra.^jJ_ ^^i^,^ probabi- '^^<v^f\) lity ('>_:i:;,,c 1 - not one of Js 20. The relationship between JT(j and ''/o t supply of Is determined by the strict proportionality, but securities and the preferences of the various participants In the market. i- PAGE 20 Define ^ iSt,('<t,rt) (9) •^ q CKupi) - ft i cscKt,FO^ Assuming th?t the the above. In nature ^rn finely enough divided/ replaced by Intep;ral The solution to beginning of period the beginning of the above We define future of states of the sunmatlon sipns can be sifjins. dynamic prorrnmnlng. value ^O — ~: b+i and substitute present vo t, cash problem can bo obtalnofi by as the maximum Vj.( K^, p^ flows discounted ) the to given that the state of the system at period t Is fK|,^,p^l . V (K^^^^Pj. ) I s the maximum present value across states of nature, discounted to the beginning of time t period t with the price set for capital stock at the end of V (K of an optimal ,p ) is policy starting at the period being p^ , and the previous period being the K^. . ^ defined recursively by the following equation. Is solution for an optimum In the discrete case 21. The Hov/ever, the continuous case. In the to that analogous necessary conditions in the latter case are less tedious to the "continuous" internret. While calculate and easl-^r to of the delineation of in view assumption Is unrealistic for finely divided large is not states, the discrepency relations obtained In be noted that the It should states. case though into the discrete this case can be translated the calculations are messy. ,/ /,. ^^^^ (10) Is /TS a ( . "^ PAGE 21 r ct) ^''-' ' the beginning of period the market value as of contract that delivers one unit (of consumption) I.e., in state value as of nature period t. <^,^ of period 1 of a In of the beginning delivers one unit with certainty at the beginning of period t. one unit In (0,t) The first two I.e., terms of t the above in market values as of the beginning Income across a the market Is contract that contract that delivers of nature In expression, represent the of period of nature In states all (P,t), ^^^ market value '^ /'^t-i state In of (In every state of nature) 'Pq ^ as of the beginning of period in 1 period t. of operating t, time period t; the third term represents the Investment costs Incurred, and the two terms last beginning of are period market the present t, optimal of values policies as of from the (t+1) onwards, given that the state of the system at the beginning of period (t+1). decision, and it should be present market Is Inherited as that Vj.^^ re^^ l^t-\ to the beginning of period (0,t) After substituting /Pet ^ the beginning of value as of and weighting It by (Kj, the the present result of the uncertainty the outcome of noted a discounts It In period represents the period (t+1), back from state t. reaction function regulators, the functional equation can be written as ^C- t. of the PAGE (11) -(kr^t-,) Deflnin/; X, m^ 4- =K, , Vt,, CkV,Pet.,>oL° Sii- V^^, Ck,. ^ dropping subscripts and 2 2 P(,^.)ci0 except where t needed for purpose of clarity, we have Define (12b) K.CiC.p) '"' Thus, HC»vP) ' (13) If Vt CK,r) the optimum K Mc^x - then K solves the problem Max In making its expression for Vt;(K,p), not nyopic. period t, as (.) Vj-,.1 depends on " r)N Hence, =X. X f K/=Max {x, V, happens ? 0, K' where j it K' and so 1^ = of the clear that the firm's decision is knov\' selects an inspection From (K,p). its its optimal on. he solved usin/r: dynamic pro,";ramni investment investment in period enters the optimization problem. (.), I the firm decision, To determine the firm must V^^ ^ Ha^ KtC^.P) f\-(fC,p). optimal to maximize Kj.-Kj_, ( X + Mt (K, without the constraint that less than X, to be Is ^ r^[6,(^c,p)]o J-^'L6eCK.p)J . V.'hile nr. V^^., an actual methods, it (.) for (t + 1), in turn problem can is difficult PAHE 23 assumptions. firm In period . t, all In the optimal decision of the nature i.e., Kfc,.-^ and K. I In (t+1) Is (t+1), except the that price problem of optimization of optimization between the extent affected by investment In period Thus, one to (.) V,^. of \\^ t -. from period (X,p) becomes from the relation and H,,, (.), we can v/rlte ilk) By definition, SubstI Into H tutln,*! do \ (K,p), = '^f and collecting terms, we have Kck^p) For a maximum, we must have (5k (16a) <3< ^,.. %(., •- K,., t. of V^ (K,p), while In and t K, This assumption del nks the decision made In period that that Is both periods Investment In states of further nj^ assumption made here the specific will undertake the firm (t+1). order to study In without maki results analytical derive any to ^ 1-t-. ife PAGE 2h [%. l.^^ii^:^Ai!^' -_£!!l^^_^[/V,,C<„,r,,)l 4- - Multiplying by throughout '<^_^ and P^(K,p)=B(K,p), we have =p p and , since q MCK,r) ^(n,K)=p/w using , =Q(K,p) for 0=M, rj (16b) „here i, sufficiency = || this of specifying without || relation - the demand marginal unit of capital H = functions, present value represents the ^" difficult Is the uncertainty In possibilities and ) ' ^^ , stock In establish to production the greater detail. the cost of period t, ^he • ('<t~<t--i of adding a for one period. The second and third terms represent the reduction In labour costs discounted to The stock. capital present, due to terms last two values of changes In operating resulting from turn result period ,5.. a from change a In marginal a marginal Increase In Indicate the present cash flows In future periods the regulated price, Increase In capital which In stock In t, Th e l)nre":n atod FIrf''s 1 The unregulated firm nc c ision is free to choose both prices p PAGE 25 and Investment for period given t/ possible the conditions, L states on the price myopic fashion, (17) i Max need not take thus, periods t ^.V^ ^. 7- ^o,i -^ C \ Cf<t, P.>] As In (18) ^t-'V, - I..] <=^t., < ^/O the case of the regulated firm, Max t [ ( - -<„ where 4e K^O^^^r^^V' ^ -^r '^^ this (\, Is I. ^ </o, 'v.t such that k -' ^t -'Hi ^" ''-^^ . written as [^(K^n)]^^ It and (t+l). such that It- a make that Is t Into periods as firm can, In demand making Its provided, of course, [iV,,, Kr(Kt-,Pc)J where In future In would undertake positive Investments The unregulated firm's problem the possibilities. The unregulated regulated firm would. a nature, of the unregulated firm Its effect decisions In the beginning of period at / and the production optimal decision, account t, PAHE 26 This Is dynamic a recursively. define Vie. V/^ ( K^^,, as ) problem nf^ period t, of period (t-1) Is the benlnnlnj; of nature discounted to that ;^Iven K^. solved be to the maximum present value flows across states of of future cash the end profirarml functional equation The . stock at capital the for this problem Is defined by (19) /tvit PC '^t where, as In equations (3), In this capital the case, stock alone, state of the system jK^^_^l., rather than Is defined p^ ^ as In the free to choose , J!'^^,^ ref^ulated case, since the unror;ulated both p . t and K and rewrl te 17 ( We define K . ^t=f^t, f I ^^"^ rn Is '^''°^ ^^^ subscripts ), (20) Defining Jj,(K,p) as to '^(^r) (21) Vx/, we can wr te I '.'^. as (O do by the PAGE 27 (22) w^(x]. Ma. ^x.j;(K,r)V- XV wf^> {j,u,p)^^ X. r,iK%^) ^ p If the optlnun K without tho less than to be then K* =X. X, solves the prohlen constraint that Thus, [ ^ happens K* =Max{x, K'U/hore K' J^(K,p). flax I' As In decision the case Is of myopic, not regulated firm, the due to Investment has to be positive. (t + 1), so that the problem of opt ml I one of optimization of J (K,p) and (23) w^^, (K) By definition, . p4 i<^ "^^ t restriction 11, In both -tat perio'^s ion of (K) \J(^^(X) t can be written as °^t The firm's optimum strategy can be obtained by solving ^•' (AM The first order conditions for 25a) - o ' a and becomes j;„rCri) ~ that the assumption aj^aln make Vie firm undertakes investment that the the firm's the maximum are v/rltton as "^ PAGE 28 (25b) - M.^^i'il { qM_' ^l^b^^l d.0 ^ { ^liL^S^ ^_. o ^i^ or. order condition for The first marginal the present value" of the cost of addin?^ a capital In period for one period, There capital stoci; rej^ardlng (except Insofar as a result Implications no are decisions for as capital unit of In to of this decision the of future cannot be stock that has to be set equal decrease In labour costs the marpjnal decision. t, Implies, stock capital periods reduced). The second equation Indicates that the present value of marginal cost and marginal Mov/ever, demand periods whore In Is unregulated firms first order firms, for conditions for .mCkp) .^y^ (2Ra) (2Eb) Increases and satisfied, hence, price and output decisions of the in order to capital stock, compare them. of the two given price, are v/ritten as a Regulated Firm Unrcgul ated not fully cost. now examine the capital regulated and The demand Is Insensitive to small equals marginal \-ie revenue have to be equated, across states. F I ^^"^'C^ ' * - ap 1— ^c,, r rm (^t-Kt.,1 - f p^^[.o4Cv'^)]'^- k[--C'sC^^r),K)jX^-- PAGE 29 The last term !n the equation for regulatod firm can be the wr tten as I (s/Q-sKQ, But = ^^^Sp) >^ r<^ ^^ '^^'''''^ (27) /a'")>0 implied by is V /,^ J follows from Q,^ Q, /Q-wXQ, /Q" )>0 since (w/, ^^(^>0. = J""^ ^' <S since 0,>0,K, which Similarly, <0. which c^--^ y, Q Hence the last terr dK can be wr tten as i v/here cJC = The expression for the regulated firm can be written as /^^^'f'-'^t'-,r-'""-^^^'^''° dp K-^..,)" . (28) _ 7<^ M '^ The terms of stocl< In decrease anove expression present value, of period, In the together with t, have a marginal for one period has the increment Increase the present value two The changes of in components, both in In capital to be equated to the In period present value of labour costs succeeding periods. periods, Indicates that '>./ a p t, changes In revenue in revenue succeeding arising In through tfie PAHE 30 regulatory price mechanism 1) decrease price the through In a posslhle the decrease In lahour In of nature, states Increase In Is dependent on the signs and, of ^^'^*'(Ktn provided the regulator's what the monopoly period will the and /Pp.tti ) formula Implies vice versa. combination chosen In general, not. t. direction the demand curves equal the of precise knowledge of the it faces by the A-J be price shift Is result In the g ) t-^, In positive less than while the the regulated firm, unregulated firm, Indeterminate, firm's production in by at the beginning Thus, that of the without possi bll I ti os, the various states of nature, and the description of uncertainty. from will a an to of the change ,p firm would have charged (t+1) capl tal -labour period all turn, In a , In t. (.^j ^^+ti(K th p another due 2) The direction In v/i period, period In of the various states of nature of costs capital costs. an sign succeeding price In period (t+1), caused. Increase In the The associated change a This Is quite different static case and Its dynamic extens Ion. The main reason why the behaviour of the regulated firm In tlie above discussion Is Indeterminate application of the regulatory formula In is that the equation (7) could the stationary 27.. See Davis (8) for a demonstration that exhibit'", rep;ulatory process point the dynamic of overcapitalization in the A-J sense. PAGE 31 lead to prices that are "too hlf^h" I.e., set never yields formula regulatory would circumstances. It can be that the shov/n chooses an optimal capital stock, at circumstances, under such ratio Is greater for the the so that an though the Under sucli regulated firm price. a we If what than even charge such chosen by the unregulated firm, "effective/" higher price a monopolist unconstrained Is than v/hat an However, charge. regulatory process assume that the regulators would monopolist unconstrained h!r:hor ^ least as large as that given the same price. Also, ex. nost capl tal -labour regulated than firm, for the unregulated firm, for every state of nature that may obtain. Under "effective regulation, ^Au(..) 0, for all 0. or Equation (28) can be written as (29) - (/^ [-/.(^f(^'.r).K)]['-^i'^^' ^ ." - Mtl-vP) where b^ " r. -J- " "^^ ^<fi,,i C'<ch j ly^t,,; ;^ q .f In the the assumption similar to 23. This Is somewhat regulation Is "hinding," I.o., standard A-.l model, that the less than regulators Is allowe'^ by tlie the rate-of -return the monopoly rnte-of- re turn that the firm could have earno-!, were It not regulate'^. PAGf: The relation In the ahove equation first order condition and K^ for unregulated firms can he comparer! with the the unregulated firm. optimal capital as the 32 stock regulated and of the respectively/ we can Defining K^ write for the same price/ (30) M(K.^;r) ThI Impl s i es that (31) MKwrt For this to he true, has to he greater Kt^ . \!e not so. the capital stock of the regulated firm than that for the unregulated firm i.e., can prove this hy contradiction. I.e. the optimum capital firm exceeds that of the regulated the prices specified are equal. the last greater state of for tlie nature It where unregul .ited firm stock of firm or, Suppose this K^t is the unregulated K^ >K/' . Since follows that the index of demand is than for satisfied the Is regulated PACE 33 firm. M(K^ , p)>M(K,., p) quantity producod I.e. maximum ?;reater. I.e. QCK ^^ , p) >n.( sinco , by as t^ V7e for f><-fUK^,p). , Since Also, the firm Is Since >0. o", For states of nature where demand Q(K^,p) >0. unregulated the ,p) K^^ <2ilX^i:> >0, fully satisfied, Is ^^^ <0, ^^^ q< ^T^Cc^/lv)^ (n(t's^,,p),K^). can wr te I (32) /UX^7-..'<-)]^ /^/['ll^>cokj]dx) ^ •^(•^or) ^^(n /K^) since, (310 Rut, ' .N ^.,„r) . \^f^[t.i^(<^>V)^^^)\d-0 I ?/j:o( K^ '^^'^'r) N , p), K;C'.-^,r^ K^). We con write . ^ PAGE , 3lj , MCK,,f) ^ OK since Hence, ^, ( ^, ;>u. (a (K,p),K))>0. [XJ^^ A'JJ '-^ ^ j/^, [ X.(c^(Ka,p\Kjcix3 j J This contradicts equation conditions for the two firms. false. ^^( Hence Y.^ yK^ ^/u(a (K,p),K))>0, Thus, the ex. ,or It derived (31) Hence the assumption made Is K^ ^ K^^ . Given that K^? K^^ and , follows that post capl tal -1 ahour ratio Is at least as large for the regulated firm as for all optimum from the the unregulated firm. In states of nature. The regulatory process In most cases Is asymmetr I c I. c. 2'». It can he shown that ( ^. (n( K, p) , K) >n for fairly ^^]^ general production functions. See Appendix for the proofs in the case of the Cobh-Pouglas and CFS production f unct ions 25. This term Is defined and discussed In detail In Hailey PAGE 35 productive the of ono ml sal regulatory the of (labour) the constraint celling. entirely from aspect variables result of locat on I In fact. mechanism, A-J the certainty context, as well as In theorem In Is this causes the It that the model efficiency regulated the of unregulated firm, at the same price. the static A-J model where static, the discussed here. The theorem proved above compares the capital Input excluHed Is firm stock and with that of This Is In contrast to the capl tal - 1 ahour decisions are compared at the output the regulated firm elects to produce. comparison The Is present model, since output Is known only made with price ex. specified Is post constant. price held ex. In the ante , but . Conclusion 5, This paper has presented In an uncertain context. - model of the regulated firm. attempt was made An to overcome the Averch- Johnson model some of the shortcomings of regulated firm a rate-of-return of the regulation, problems due to uncertainty, and the static nature of the model. The model presented here, while more realistic than the Averch- Johnson version. Is still not regulation in practice. ( I ) pp. 52-r'7. a complete characterization of PAGE behaviour optimal The examined, and contrasted with tfiat firm Is unrej^ulated firm of the employs the The model describe framework to preference repulatod the similar conditions. operating under state of 36 uncertainty, and attempts to capture the price-setting function of regulators In a dynamic Is The rule followed context. the price fixed for that rate-of-return, stock. It more "quickly," than It can firm decides Its time, on adjiist Investment In and adjusts "labour," to nature state of capital that of order to answer rather than numerically, The firm. under study. was rogulntinn, optimal stock: the aheaH of level, after the conditions, is the compared to Input post also examined. questions, to analytically, assume that the Investments In the period including specific production and prorramnes. shown that '^nse'^ as fuel) numerical solutions can be developed for demand conditions, the firms' qx_ its stock capital necessary it v/as by Hov/ever, is above tlie compared, undertook firms being It capital stock employed by the regulated firm, the unregulated the capital can adjust Under those efficiency of the regulated firm In its its optimal revealed. Is period's that the firm to earning the "variable" Inputs such labour (or other input of previous on the further assumed, Is period, when applied a results in the firm's previous period's output "fair" by the regulators on In a dynamic context, past outcomes of demand under price and cost, the PAGE re/^u1ated firm uses an any given price inefficient conblnntlon of Inputs for Specifically, level. "effective," (i.e. the regulators set unregulated monopolist what an employs a efficient. higher level Is unregulated firm post meant at of capital the same capi tal -labour ratio Is a when regulation than the is conHinatlon price level. at Is price no higher than v/ould), the Input efficient combination. 37 least as regulated firm efficient. used Also, large by Py the the ex the PAGE 38 Append? For (AD i<\^' a the At the Cobb-Douglas , when point It x production function. X,- Is ^'^k-^^J' profitable just to Increase production. 7- (A2) Thus, the maximum production, a (A3) -. derivative The (a.) -'k of the 0. , Is given by, ^ labour-requirements function with respect to capital stock Is given by (AU) Substituting 1 aliour-roqui roment at Into (A3) s (A'l), derivative function with respect the point of rnaximum production. (A5) the Is X,(«0.-,r),K). _-(,._r.)'\>"f to capital given by '^/'-i' of the stock, J The quostlon now. the expression positive respect to (A6) i (A5) with In or at whether the sl^n of Is least, respect to capital stock/ Differentiating^ non-nep;at ve. i Is K, with we have K, ^-^fctl^ (If] [..(«..),.)] since confirms which the derivative of -c «-f $ -^ Cj<4-y-0 /!-):' ^ o K I assumption the /.-y the In text, for the Cohb-Douglas case. case the In general) (more of the CE5 proHuctlon function, we can write (A7) C«^-''.-^L-''j'^J' ^ , Differentiating respect to Q , the 1 .L-l^-LaK-r]"''' ;;._ abour-requi roments and setting It equal maximum profitable production, in to p/w, Q - [-1 " v/I th we can write the this case as ^ (A8) function 'JL ' f/uj< where C = [A J- Differentiating the labour respect to capital stock, K, requirements function with PAHE (A9) 4. fL^_- . I;0 - a. k-i^-aK-n"^"''^^^^-^-?) Substituting the maximum output (AlO) - Into (A9), we have (AB), In --^[-fe--] . K \K b To see the sign of X^~^ at the of maximum profitable point output, we differentiate (AlC) with respect to (All) Thus, 1^ [x.r^a,F)'»-)J aK in production the case of functions, labour-requirements both the function, K ° -' Cobb the second at profitable output. Is non-negative. the Douglas derivative point of and CES of the maximum , Referonces and Leland L. Johnson. "Pohavlour of the 1. Averch, Mnrvey Constraint," Anor can under Rep:ulatory Firm Revi ew Vol. No. 5. EcononI c 52. (December 1962) pp. 1053-69. i . Elizabeth 2. Bailey, Regulatory Constraint Lexington, 1973. Theory of Econoni c Lexington, E. . flass. : Bailey, Elizabeth E. and Roger P. Coleman. "The Effect of an Averch-Johnson Model," The In Lagged Regulation Rel of EcononI cs and flanagement Science Vol. 2. tlo. 1. (Spring 1971). pp. 278-92. 3. 1 . dohn Malone. "Resource C. E. and Pel Regulated Firm," The Elizabeth Ralley, h. Allocation and Journal the o_f. -^anagement S c n c p* 1970) pp. 129-'t2. I f <"^ . 1 EconomI cs Vol. 1, Mo. and (Spring 1. Baumol, IJIllInm vJ. "Reasonable Rules for Rate Regulation: V/. MacAvoy, Plausible Policies for an Imperfect \.'orld," P. Pqf^mI tor y ed., ThJI Crisis of the Com.ml ssions N'ew York: l-J.VJ. fJorton, 1970. pp. 187-206. 5. r> . Baumol, V/Illlam J. and Alvln K. Klevorick. "Input Choices Overview of tne Rate-of -Return Regulation: An Pel Journal of Discussion," The EconomI cs and f^anagemen t ci ence Vol. 1. r.'o. 1. (Autumn 1970) pp. 162-90. 6. and 1 .S Ronbrlght, 7. Pub l Ic Ut i 1 i t ies University Press, . PrI ncl pies James C. Rates New York: Columbia of . 1961. Davis, Eric. "A Pynamlc Model of the Regulated Firm with J o urn a Price Adjustment Mechanism," Th-C ^'^^l Mana'Tonon t of Economi cs and in73) pp. 270-82. Science. Vol. U. (Spring No. 1. (Spring 1973) pp. 270-82. 8. a 1 HI 9. ilDii 19 : -h '' i ntprr's t n v-s tmrnt Jac!<. Pren t ce-!'a Englewood Cliffs, M.J. 1 el Per, t a 1 . 1 , I I : i 1 1 7 0". 10. Klevorick, Alvln, K. "The 'Optimal' Fair Rate of e t . PAGE Return/" The "o 1 1 >!ot.jr nr^ of '^cnnoni c^ Vol. 2. ^!o. 1 and r^nnnp-.onAn t S c oner' (Spring 1971) pjp. 122-53. . i Ij2 1. J of a -irm Sutijcct to 11. Klevorlck, Alvln K. "The ^.ehavlour Journal Stochastic Re.Tulatory Review," The i^el 1 anH r^ana,":en':^n Econoni cs Science Vol. k. No. 1. (Spring 1973) pp. of . 57-33. 12. Leiand/ flayne. "'^.er;ul at Ion of Natural Monopolies and the Journni Return," The Rel of Rate of Fair Economl cs and Mana.f^onen t S c e n c Vol. 5. No. 1 (Spring 197'0 pp. 3-15. 1 . i Myers, Stev^art C. "The Application of Finance Theory to dournal Puhllc Utility Rate Cases," The Pell and flanaren^nt Economl rs of Science Vol. 3. No. 1. (Spring 1972) pp. 58-97. 13. . Stewart Regulation flyers, Ih. Under Jour al of ri ManagcTK'nt 1973") "pp." 15. Simple "A C. and Science. Vol. '.'o. nsl'si , i 1. r>ty, " anri / '' Pehavlour of Firm Tho "ell (Spring ' 30/1-15. Sheshl tlodol I'ncer ta Fcono ml rs "i/elfare Eytan. Aspects -of a Regulatory Am^r cnn '^cnnomi c (f'arch 1971) pp. 175-8. f'ote," Constraint: Review Vol. 61. No. 1. i . Incentives Harold '1. "Pair Rate of Return and 16. IJeIn, A. Considerations," with comments hy Carl Some General Trehing, ed R. Hughes in H.,M. Conrad and William Performance U nder Regii a t on '^ast Lansing,: Michigan State University, 1968. 1 i . lll.~7H 3 TOfiD DD3 7DS SHI 7 3 7-7'^ 1 TOflD DD3 7D2 bl7 /;€'7*^ iliiliiliiiiiiiiiiiiiiiiiiiiiiiiiiiiiii 3 lOflD DD3 7DE h33 .7-^ •71'^ 3 TOfiD 0D3 7D2 bSfi HD28.M414 no.720- 74 Bloom, Gordon /Corporate social respon 800385 D*BK^ 000:37714 TDflO 3 ODD abb 73fl 72l-7tf 3 T06D D03 7DH b74 HD28.IV1414 no,722- 74 and analvsi Madnick. Stuar/Application 721943 3 p*M,ti TOfiO ooy'^ijyy ODO 7bl 7bM