Welfare effects of demand management policies: Impact multipliers under alternative model structures

3ii3

Montek S. Ahluwalia and Fr& J. Lysy

The simplest cas”: of multipliers that take account of the linkages from initial production changes to income and consumption changes, and thence to second-round prod&on changes, is of those derived from a closed Leontief model in which consumption demand is endogenously determined. Several studies have examined such mniatipliersfor economies as a whole and for particular regions.’ This approach is more general since it takes account of the circular flow from productia~ to income and consumption. EM.hermore, it permits quantification of impact multipliers for the incomes of diRerent household groups, thus introducing a distributional eiement into the analysis. However, ia remains limited by the assumption of fixed coeffrciants in defining the production, employment, income distribution, and consumption relationships in the economy. This assumption reflects an extremely restricted view of the economy. Under the usual profit maximizing assumptions, the . fixed o~effrcients assumptions underlying the closed Leontief model are valid only iu a world in avhich all primary factors s.re available in +bnitely elastic supply at fixed prices (together with some homogeneity assumptions discussed below). Such a system is entirely demand driven, being .rnconstrained on the supply side. Once we allow for supply constraints, the fixed coef %ients assumption is difficult to maintain. Increases in output levels lead to increa.ses in the demand for primary inputs, and some of these are likely to L,3 in fixed supdly. In this case, we must allow prices of there inputs to rist, which in turn alters output prices. These price changes will lead to some sr Astitution in botll production and consumption so that the assumption of “Ixed coefici lata is obviously inappropriate. This paper examines the sensitivity of such multipliers to aytemative characterizations of general equilibrium using a recently rorlstructed geaerai equilibrium model of Malaysia. Although the particular focus of this paper is on static multipliers., the results are also 3f more general irierest. The_qprovide an example of the extent to which tnodel behavior may depend crucially upon the “closure rules” adopted--a subject which has received considerable recent attention (Cardo::o and ‘Taylor 1979:, Taylor and Lyr,y 1977). THE STWUCTU

ODEL ANB

The model used in this paper belongs to a class of Icornrutable general equilibrium model of which there are many recent cn i~mples. The model has a-

‘See, for exl mpl :.Pyatr a1d KounC ( ‘.979). For an application 1t1.fthe same methdcalogp to a rrgicnal econorq. see Elel: and I-hell (IWO).

WLFARE

EFFECTS

OF DEM

dC ‘WANAGEMENT

POI,!I;:IES

319

been described in detail elsewhere, and iti main features can be wmmarized as follows.* The model dis.tinguishes 14 production sectors, L labor vpes, and 12 bousehold ups, each household having its own demand svstzm.3 on functions permit varying degrees of substitution among I’actors and among intermediate inputs. Household domand functions permit substitution among commodities. Furthermo~-e,imports may be substibuLed for &>mestic supplies in both intermediate use and in household consumption according to a system simikr to that of Armilgton ( 1969 js World prices for imports are given (and are independent of import levels) and are converted into domestic curren :y at a siven exchange rake. In equilibrium, the model determines domestic output and factor price:; so as to clear output and factor markets. For any given exogenous shock we can estimate impact multipliers Iq, solving the model under each c.fthree different factor supply assumptions. At one extreme, we assume that a91primary inputs into produ.ctioncapital, labor, and imports-are available in infinitely elastic supply at given prices. As shown below, on these assumptioas the model behaves exactly like the closed L.eontief system. Alternati~oly., we assume that capital sr+.xksin each sector are fixed (reflecting sbori-term inflexibility of installed equipment), although the supply of labor ‘to the economy as a whole is unconstrained, so that. eat!; sector can hire labor in any arnC’;It at fixed money wages. We have described this closure as Keynesian. Ow 'thrrd factor supply assumption is that capital stock in each sector is fixed aJd t le available supply of labor to ‘the:economy as a whole is al:;o fixed. W: have described this closure as neoclassic& The Structure of the

ode1 arrctIts Price Sys~k-;M

The price system of our model follows from our characterization nf producers as profit maximizers ope’rating under perfectly comper:iCve conditions. On these assumptions, oiltput prices will equal the marginal costs of r9rodt*ztinnand therefore WA depend upon production functiorl parkmeters and the prices of hpU;iS into production. The production function in ddch sector is a multilevel CES function as ‘F’o a dctailcd statement. see 4hlbwalra and I,ysy (19791 %:.J.amplcsof recent gzleral equilibrbm modr:s nn: Adclman and Rohnson ( 1978). ‘ray lor e:. al. ( I379 ). and Robinson and Dcr\is ( 1978). St:e /Appendix ft. r a stmnt~ary of model spel:ilicatior,. ‘T’he erght lab01 tyl~:s correspl;ICEO five skill types with eaJ. of the three !ow:r skills segmenttd inlo an agn,cul.ural labor orce and a nonagricultural labor for-e. The twli higheskills are a:;sumed to Se mobile acr,,ss all sectors of the economy. Sine: the agricultural component of the labor force at eact of the three lower sk.Us can Gt 64 be emrloyqd in the agricultl ;a1 sectors, it .zan be treatPS a‘, a different Eactor from the nonagricuitur. 1co;nponent.

Montek S. Ahlluwaliaand Frank J. Lysy

320

OutpJt

x,

_--L_, Aggregate

value added

intermediates

-1

I

Ccmposite commodity 1

-

- .,. I 1

Composite commodity

2

i

Composite commodity 14

Aggrerate capital

Aggregate labor

Figure 1. Multilevel CES production structure. Note: 4t each level, two or more co:mmodrties or factors combine into a “CES aggregate” which is an input at the next higher level. The elasticity ofsubstitution Q at each level indicates the degree of substitution between the various lower level inputs producing the aggregate. The vrJues forthe various oparameters vary across selctors hilt their typical values (or ranges where relevant) are as foliows: #ranges between 0.75 and 1.5 for all sectors; uL ranges be,.-ween2 and 4 in different sectors; &, d’, and 19 are set at 0, 0.1 and O.!, respectively, for all sectors.

shown in Figure 1. This formulation is common to all the computable general equilibrium models refeRed to previously in foornote 2 except that we allow for some substitutioni among intermediate goods of different types and sIrubstantialsubstitution between do.mestic and imported supplies of each good used as an intermeciiate. It is clear !?romFigure 1 that the price of each sector’s output uGnate1 y must depend l.tponwages, the cost of using capital, and the prices of ah domestic and i~mportedsupplies for mtermediate use. Tn general, when an output Z is a CES flnnclionof inputs Y, and Y,, and is produced under conditio GSof profit maximization and ,@ect competitior~, the following relationship holds between ouput and input prices”:

4This relationship can be derived from the first-order condGans of producer maximization car equivalently from the cost function by ditfere.qtiating the :ost f&~&ion with respect to
Using this relationship, the price ofea..:h CES aggregate at a giecIen le?/e!in Figure f cm be written in terms of the price of each of the input3 ret‘&enext ?Owerlevel. Since irrput prices at each stage can be further decomposed dCkWllWiNdS, the pk.65of output at thti hzghest let el can be decomposed into . prkes of inputs at thlc:lowest level. T&s @vesa set ofprice equations with the gene& form: & =flW*,---,

i = 1,...,!4,

Ws, .^‘i,P;,*--, Pi& PC”,..., Yty],

(lb) where )yl, . . ., ‘cgare the wages sfthz eight iabor types, and r, is the rekrtal0n regr&ecapitapin exh sector. I’his system of 14 equations can be used to fi>r I4 ou@ut prices .P,‘,given the wages of our different labor types, the rental for ca@ta; in each sector, and the prices of imported goods. Tbs price systelr. is e in a larger s:rstetn of equations whi6:h ensure?: a Wajlrasian gen ibrium. An e;quilibrium solution of the model is one in which factor prices and output prices not only conform with equatickn (lb) but also ensure that all factor and prduct markets are cleared. For this we need a set of equations determining equilibrium in factos- markets a& a set determining equilibrium i.1 the product markets. Factor market ~;quilihti;rm reqllires the facdoTdemands to equal factor supplies. Ir. our model, factor demands are derived from the conditions of producer maximization. In the simple case of a CE.S production function, the derived demad for a I actor Y, is 8 function c$ the tsrice of the factor f3,, the price of outt,)utP,, anJ the level of output Lb: Y, = jcu,PJ.P, J”Z.

(:,a)

Using thiz, relationship at the lowest le\*el of our prcductian tree, we l:an subsitl
c

lj(WI.**mr

#‘H, 1’1***

rlQ( P;,*.**

l’;,,

X ,,..., X,,,), j= i,..., Sirrilarly, the demand for cspit;A in each seeto:’ t it2 be

(2b)

P’y,...,P;“,,

Ki’4Zi(CY1,

am*) Wg, r,q am*, 1’14,

Pf,

es*)

Pi49

P;”

3

***

I”74

M6tten * XI9

* *i

‘l4)*

Mont& S. Ahluwalia and IFrankJ. Lysy

322

This block of 22 equations determines factor market equilibrium in one of two ways, If factor demands must be set equal to fixed supplies, the equations detern+ne equilibrium factor prices, given output prices and output levels. Alternatively, if fa.ctor prices are fixed and factor supply is assumed to be infntitely elastic, these eyuatione determine employment levels. Lnboth cases, factor markets are cleared, although in the latter case this clearance occurs at the intersection of a demand curve with a horizontal supply curve. The third set of equations in our model ensures equilibrium in the product markets by equating domestic output levels w&h demand for domestic output. There are various types uf demand for the domestic output of each sector., including demends for inteme&ate use, household consumption, exports, investment, and for govem:nent consumption. Real demand for investment and government consumption is fixed exogenously, but the other elements of demand are endogenous and price resporsive. Demand for intermediate use can be determined in a manner analogous to that described for primary factors in equation (2b). Export demand is determined by a price elastic world demand curve facing Malaysian producers. Consumption demand is a function of household income levels and prices of domestic and imported goods. Household incomes, in turn, are determined by the factor endowment of the household i;md factor prices received. Aggregating across all these various demands, the demand for the domestic output of a sector can be written as I’oll.ows:

~w,=x,~p,x, p.;4,p;:. .“,

Tlr

. . , Py$, L,, . *--.

rpJ,

..)

L;4,

K,, . . . . I&,

i =-I 1, . . . . 14.

WI,

. . . . 148,

(3)

TEken tog&rher, equations ( I b, 2a,b, 3) represent a set of P4+ 14-t-8-t 14 equAons whij:h can be used to solve for 14 output price variables, 14 output variabh:s, and either 8-‘-14 factor demands given wages and rentals, or 8+ 14 wages and rentals, $iven fixed available supplies of labor for the economy as a whole and cspib&lfor each sector. Alte rnatk

Closure RuEes

We ;1owturn to the alternative clo?;urerules under which the model can IX solved and examine their implications. The Leon&f closure rule amounts to specifying factor prices as given and rlsirlg,rquations (2b) and (2~) to solve for factor demands. Under these conditions, our general equilibrium model, despite its complexity, reduces to z simple closed Leontief’model l&h fixed coefficients. The assumption of linear homogeneity in production ensures that with fixed factor prices the price s:qstF,rncan be solved independently of the level of output. Ifall prices

WELFA

EFFECTS OF DEMAND MANAGEMENT P”X_,!CI~S

323

are fixed, the ratios of all inputs to outputs are also fixed [see equation (2a) so that, irrespective of the scope for subsdtution in the t0:hnobogy, production can be characterized in terms of a set of fiixcd coeRk5snt.s at given prices. In other words, demands for domestic output of intermediate u;se in production D can be described in terms of a matrix of fixed coeffkients: D = AX Simi arly, demand for imports for intermediate use, as well as the demand for l&boxand capital, are linear functior, a:of output: M = mX, L = IX, X = kX. Value added or income generated. from each sector can also be written I/ = PX. There is a silnilar simplification of the model on the side of household income and consumption. Since the model assumes that the mapping from factor incomes generated to household incomes is linear, the linear relati.onship between factors and outputs with fixed prices translates iatc, household incomes which are linear fimctions of outputi Y = BX, where 3 is a matrix. Household demands for domestic outputs in ‘lumare linear functions of incomes whell prices are fixed: C = HY + q, where His Amatrix of fixed coefkierits and q is a vector of constants.’ This can be written as C = JX + q, where J = HB. The material balance equation for the model, equating domestic supplies a:ld equating demands for domestic outputs with domestic supplies, can be written as follows: X=AX+JX+q+E’fI+G. Expoia E art: functions of rlomes,l:icprices given world demand conditions

and are determined once domzsti:: prices are known. Investment demands an.d government consumption derlands are exogenous. This equation can be writkn X = [k--A-J]’

‘Z’,

where F is the vector of all exogenous final darnands for +ziomesticoutpat (including the vector of constants q) and 1 is the identity matrix. Thus, our general tzquilibrium model reduces to the closed Leontief -5

The demand function used in our mocielis a three-level ty pi. First. we c!efine“wants” :XII as kx~I. shelter, clothing, etc. Households demzvrd “waists” vary according lo a lineor expenditure system (LB). Each “want” in turn I~ sarlr,iied by various comtwsite comlnrtiities according to a CES transformation. At t.hr?lowest level, each :oi rposik commodity is a CES :Iggregationofdom :stii. and imported supples.With fixed prices, tie ratio ofdomc-saic and imported supplies of valiou!; types within each ‘wanr” n:main lixed. Thus, be tteuseloc.-ld consumption deman 1 fnr d,>m~estcou!pur is a Ike ir functic~nof income with a vectcI cd constants q corresponding to cglantitir:s of do:mesaicotitpu! required to satisfy the .L,,.S “subsblence” quantities of eact! want.

324

Montek Is. Ahluwalia and Frank 3. Lysy

model if factors are available 31 unlimited supply at fixed prices6 The

impact of any exogenous demand variation on output is directly ob&inable fron the coefficients of the inverted matrix [I-A-4-l. The mapact on er;lp:,cyment imulcf household incomes can &XI be obtained given the fixed input coefftcients I and orand the level of outputs and prices. Moving fiorn the Leontief to the Keynesian assumption about factor supplies-capital in each sector is fixed, bu.t labor of all types iis Ceely ava.ilab!e at fixed money wages-produces very different model behavior. The Keynesian assumption amounts to fixing aggregate capital stocks Ki in each sector and using the II4 equations in (2~) to solve for capital rentals in each sector. This changes the relationship between prices and outputs in the modeI in a fundamental way. With the fixed capital stocks, higher levels of output can only be achieved through increased employment of labor, which raises the marginal product offcapital and, therefore, its rental. However, increased capital rentals will raise out~u.fprices [see equation (1 b)] so ,that increased output is only possible with rising output prices. In other, words, +he supply curve of each sector is upward sloping. Under these assumptions, an increase in exogenous demand will produce a somewhat different response f?om that produced in the Leontief case. As domestic prices rise rlelative to fixed world prices, the demand for domestic output is nduced in mo ways. Since exports are price responsive, the demand for exports dechnes as prices rise. Secondly, the rise in domestic prices prompts a shift:from domeshc to imported supplies. Thus the new eq~uilibriumis reached at a lower level of output expansion and 2 higher balance of payments deficit than woullbbe the case in the Leontief sollution. Our third assumption about factor supplies corresponds to the familiar neoAassica1 assumption in which not only are capital stocks fixed, but also there is a fixed ‘amountof labor available to the economy. This amounts to using equations (2b,c) to solve endogenously for all factor prices. The prticipal difference betw ,zenthis closure and the Keynesian closure is that the former assumes full employment of both capital, and labor, so that exogenous shifts in demand cannot aher aggregate GDP, but only its sectoral distribution. In the neoclassical case, the new equilibrium following an exogenous increase in demand is achieved with no ag;$egate increase in value added, Instead, domestic prices rise relative to world prices so that there is a hIndeed, we can go furtJxzr and add .&at some such assumption is essential if we are to postulate a closed Leontief type characterization for any economv Fixed cc&Ecients dictated by tec!ulok~gy are not a suff&nt CGRd;tiGn for asserting a closed ie@ntief view since prices may change in such a ~9:: that Lhe linear relationships in rea: quantities in production may generate negative profits.

WELFARE Z:FQXTS OF DEMAND MANAG iEMENT POLICIES

315

red*tction in ex,:mrt demand for export sectors and also a shift from prcxtuced goods to imports in the! import competing sector:;. &sources released from contrahctictnin these sectors are redepioyed elsewhere to a.llow the nontradaMe .sectors ‘to expand. Thus the system responds to an increase in aggregate d.emaLdby shifting resources fro,mthe tradable to th.e nontradable sec^uorsto meet increased demonic for these sectors output, and there is an incr’e;wed absorption of imports in the economy. The increase in importi:) as ~~11as the reduction in exports, leads to a substantial increase h the balance of payments deficit. This brings us to an inqxxtant aapcct of the structure: of the structure of our model: the tre!atmenf.of the balamce of payments. The model is not constrained to reach equilibrium with a fiixedbalance of payments deficit. Rather the size of .he balance of payments deficit is determined endogenously and reflects the excess of domestic absorptios over total domestic supply. In other words, investment is not constrained to equal domestic savings. Looked at in aggregate terms? the Leontief solution corresponds to a situation in which an increase in ap&:egatedemand leads to an inciaase in both domestic supply and imports in equal proportions. However, as we move to the Keynesian and neoclassic snl: c !osure!:with constraints O.-Ifactor supply, we limit the a’M,ityto expand d!40mest,coutput in the aggregate, with a con.sequent widening of the balmcr: of payments deficit. To some extent this treatment is ccnsistcnt with established practice in the literature on n:.ulti.pliers,which treats imnshrtaYSa leakage. However, it raises the question of whether lure:could aS=lpt some “ultra neociassical closure” that would ensure a fixed &&::it. We note that such a closure cannot be achieved simply by ailoGrg tit.lec&r,ange rate to vary.. In a world of complete price flexibility, if al! real de:naanda,arehomogeneous of degree zero .UIall prices and incomes (,asis the ca:;e in our model), then a change in the exchange rate (coapled with a full-employmenr assumption for both labor and capital) will only raise (allidomestic prices and i:lcomes proportionately, leaving the real equ&brium unchan,ged. Exchange rate changes providdca basis for improvilrg the baJance of payments only if some prices are fixed in monetary terflrs (or adjusted with a lag) or some ellemeqts of demand! are fixled in monetary tennr;. Bfnder these assumptions a rise in domestic prices arising from an exch;~n~ge rate devaluation would Beadto a +e&tction irr aggregate real d~::m!and,thus providing a mechanism for mstoring equilibrium without a widening ofthe \l~a!anceof payments deficit. A second feature of the SWI.KI;W~ of’WYmode!lthat is worth IWing is the absence of any explicit moneta!3’ sec::or, The model provides En equilibrium relative price structur e in wluizh the level of domestic prices is determined relative to the level of*~or!d prices for impcrts (where the iatter are converted into domestic cur~n~y a&a ~tiwenexchange rate). This domMically

326

Mont&

S. Ahluwalirr

and Frar& 3. Lysy

amounts to assuming that monetary polizy is entirely accommodating, In other wordsI the monetary authorities adjust the supply of money to whatever is required to maintain neutralit.;? as the domestic price level changes, so that there are no feedbacks into t5e real system via the monetary sector. Explicit modeling ai the monetary sector would require speciflcation of the demand for money, the nature of .real balance effects, and interest rate effects. These could play an important equilibrating role in response to exogenous demand shifts. For example, a rise in domestic prices might lead to reduction in consumption demand if there are significant real balance efkcts and may also generate pressure on the interest rate, which may he:lp to reduce in)Jestment demand. These are ignored in our model. IMPACT OF A GENERAL INCREASE IN SECTQRAL DEMANDS This section examines the impact of an exogenous general increase in demand as estimated under each of our alternative model closures. Starting from an equilibrium “base” solution, we assume that total exogenous final demand for each sector’s output is increased by 20%. ’ The model is then solved with the new tixdgenous final demands under each Df the three closure rules discussed shove. The partkular increase in dernand is not intended to have any special relevance for policy. Mowever, i t provides a useful quantification of the very different im:pact upon real \ alue-added: employment, and import shares by sector thha.tis predicted by each of the different closure rules. The percentage changes from the base solution for each of these variables under the three alternative model c:losures are summarized in Table 1. As one would expect, the Leontief solution &UWSthe largest. expansionary impact on output. The exogenous shift in demand is easen:ially a .scalar expansion of the final demand vector which leads to an equal ,proportional expansion in output in each set ior. Because of the linearities involved, there is a similar expansion in value added and employment of labor (as well as capital). Imports rise by 20% in each case; thus impoti shares remain constant. The Keynesian solution diff’ersfrom the Leontief in that, with a 5xed apply ofcapital in each sector, expansion in output can only be achieved at rising marginal costs and therefore requires increases in output prices. The ‘Exogenous demands for investment, government, and exports are all raised by 20%. The increase for exports is implemented by shifting tbe world Jemand curve out by 2U%, i.e.. by an amc;unt such thst if prices cemsined constant. export demand would rise by 20%. The govemment demand increas;t inckldes a 20% increase in direct government employment.

-

12.7 14.0 19.2 32.2 23 4 E7.4 !?.I! 20.0 16.0

to.!! 20.0 20.0 20.0 20.0 20.0 20.0 20.0 20.0 20.0 dO.0 2U.O x.0 20.0 ZO.fJ 200

---12.3 -18.1 -21.8 -19.0 - 2.8 - 3.6 --I!?.0 - 4.8 -- 4.5

6.6 10.5 3.9 38

6.5 51

7.8

47 9.4

20.2

16.9

8.7 16.1 7.3 9.2

IA%Wtkr ..~“i

Neoalassiclal

kpMdUY .

(5)

(4

Fhtployment _~_. ._

(2)

&al Value-Added icvels

abie 1: Sa-%xaS Results: General Demand increase‘

-__

!I.3 11.0

-

24.4 10.7

5.6 si 5.5 ,_

-

5.2 5.8 3.2

7.0 4.1

15.2 16.7

0.7 -12.7 9.

6.2 f.8 4.5 6.3

II.2 k5.5 18.5

1l.i

8.6 -20.7 -36.2 - 35.7 -

(6)

148.7 80.0 1

_E

-

42.8 58.4 41.1

63.3 40.3 91.6 SSl

26.7 11.7 12.6

I_s 14.1

18.k 59.2

55.7

24.4 29.9 23.4

26.0 55.6 44_2

328

Mont& S. Ahhrwaliaand Frank J. Lysy

supply curves are upward sloping., whereas in the Lcontief soIu.tion they

were hoaizontal. The rise in domesr’z prices reduces export deru?.ud!and also shifts demand towards imports in the caiseof household consk mption demand, as well as in intermediate demands.” Thus, there is less e.xpansion than in the Leontief case and, of course, there are inters@ctoral shifts because of relative price changes. The largest increases in real vahle-added levels in the Keynesian solution (shown in column 2) are in those sectors that produce nontradable soods ,that neither .?ace competition from imports nor have significant exports ffacinga price elastic world demand curve. Within the nontradablc category, expansion is greatest where the demand for a sector’s output does not ciepend upon an endogenous dt:narld that is itself price responsive. The premier example of this case is construction, whose output is nontradable ind whose demand largely comes from exogenous investiiAent.The other sectors in the nontradable category (Qtilities, trade and transport, services, ard dwellings) expand somewhti? less. The utilities sector is a special case since, although its product is nontradable, it has an extremely steep marginal cost curve because it is heavily dependent on capital. As a result, its’price rose by 6696 in the Keynekm solution, whereas the largest price increase fctr any other sector was only about 11%. Since this sector sells a su$stantiaYamount of its product to households and fol-intermediate use, these users substituted away to tltit:exteilt possible. The main hectors facing substantial competition from imports are the manufacturing sectors: basic consumption goods, advanced consulnption goods, in&strial intennediates, and caLGt;al goods. The import shares ofthe to~zl supply of these goods in the b;iLsesolatian were 30’36,XI%, 4096, and 67’9&respectively, whereas in no oFherser;:tcrrdid it exceed 13%. As prices rise, users shift away from domest.ic sup$y towards Imports; this can be seen in the rise in ratios of imported ‘todon! estic supply recorded in columns 7 and 8 of Table 1. Real value added in these sectors, therefore, rose by only modest amounts in the Keynesian case, 4.7%-9.4963, compared to the 20% expansion in the Leontief case, T.i:‘he smallest expansion was recorded in the export-dependent sectors: r&her, forestry and logging, mining,,and export processing, which provide Ma.laysia’s traditional exports. Exports accounted for 90% or more of final demands in each of these sectors in the base solution. Mining and export proi;:essi:,lghad particularly smiiiIIincreases in output. It is interesting to note that alt6orrgh the increase in value added for each -_.-__

“In our model the scope for substitution betweeri domestic and imported supply is limiter1to these two categories. Government demands for commodities and investment demand is modeled on the assumption of fixed combinations of domestic and imported goods.

hired. at WWM have to

sectofs the cutback of clitp .I

e Leontief Iwel XI, !&.I the

e.g., A-:&*and empIo;VAm,nerlt may be hk&!F. The necxk- *id cbssufr:rule pro&-es resdks that 8:‘e ewrtiafly an extension ot the Keynesian resuit~irrtnv;extent rtftrleirdcvi&~n Fromthe L_eonoi~!f sdutiun. With lab3f supplies as welt as capitalwpplie:; 1:mited‘lo their b jse soNion values, a wegatereal vu132ii&Id must strq apptox-

330

Montek S. t4hluwrtIiaand Frank J. Lysy

imztely

constant. However, with labor mobile acKtss sectors, there can be quite substantial sectoral changes. The relevance ‘ofour sectoral grouping is evident in the neoclassical solution. The nontradable goods sectors still expand with the increase in exogenous fmal demands, with con&uction expanding by a very sub stantial 16%. But this expansion can only be achieved by a shift.of labor from other sectors, which must contract to release the labor. As shown in Table 1 {columns 3 and 6j, thr3import competing anti exportdependent sectors; contract in terms of bcth employment and value added, despite the fact that there is an increase in their exogenous final demands. As we would expect, the export-oriented sectors, facing the most price-responsive demand:;, contract the most (12%-22%)). The imp8M competing sectors contracted by 3%-5%, generally. Advanced cor’lf;umergoods had a reduction in production of 19%. The reason. is thai in addition to facing substantial competition from imports (imports made up 50% of the tot&I supply ofthis good in the base solution), it is also an export arieded sector, with ~:~p~rtsconstituting 47% of its fiial demands in tie base solution. We have seen that with domestic factors in fixed supply and fully emplq fed, an increase in exogenous demands canncrt increase aggregate real olrtput, However, the increase in er,ogt:nous demandls does have an efrect I pan the system. It raises wage rat:& anclca.pital renhaits,and therefore domes:ic prices will be higher relative to the fixed-,world prices. These changes will lead to some distributional effects within the economy that may b: quite different from those predicted under the simple Leontief closure . These ilssuesare illustrated in the experiments described in the next se&or. EFFElCTS OF AETERlVATwE

DEMAND EXPERIMENTS

In this section we shall exa,nine the impt;ct of two types of exogenous demanlij increase ia the economy: M increasl; in government consylmption and an increase ia total invesmen,‘i. Both trlements of final demaild are subjec! to policy control and their levels are often manipulated a$;part of aggreg;de demand manqgement. It is o5en rellevant to ask what impYztsuch demanl,i manage-merntpolicies might have \Dt:employment, wages, and household incorne:s. The direct imp’nctof thl:se demand increases is very differejlt. Goveinnnent consumption expenditure has 8 very high direct emploq’ment coiltent while invzstrnient exj:)enditure is directed largely toward 5 ccrrastr;c’tionand capital go& with Ihe latter showing a very high ratio of imports ICIdomestic supply.’ Ii:is interesting ‘toexamine whether the ._-91n3~: model, capitsl equipment consis{&of fixed pruponions of imports and domestic output of various se3w..

total &ects are also diRe:reni. aad how o1.v ::sse:ssment on thir: score woab1 depend upon the closure rule: ixdcpted. tJn.like the discussion oi’the previous section, ad not its longer-term impact on poten:.ial output. Impact on Emplsyment and Rdative Wags The impact of our two demand experiments upon employment a.r.dwages by skill type is summarize8 in T’ilbie 2. As one would expec t, the experiment wi,th increased government demand shows a larger total ir..lpact on employment in the ILeontief case, prim;.rrily because of the very substantial p&ion of government expenditure dev’>ted to direct employment Total emp!oyment inzrealses by 5.5 90 cornpart d to 4.1% in the investment expr:riment, and thle comp&tic;n among the skill groups .is quite differe, it. Goven;ment employment is much more inten sivc in the higher skilled wtrr!
Type

3.7 4.8

4.3

4.1

3.0

4.2

3.8 4.6

4.1

3.9

2.g

4.1

Keynesian (empOtymciit)

69

--

-

01

9.3

11.9

IO.9

7.6

7.0 7.9

Total

.-.m

(3.4)

(3.7) (3.7)

Rural

Neoclassical

-.

(12.9)

(12.7) (13.3)

NO~Wil!b

~wagtles)

_-

--

~3.1

9.4

1.5

4.9

3.3 4.1

tAc+mtEcf (ehqlnyment)

(4)

_..” .~~ ---__

4.6

9.2 -

16.0

22.2

20.2

13.2

4.3 7.2

9.5 12.2

Total

(3.1)

(3 5: (3:5)

R!W%!

--

(21.1)

(2C.2 j (20.7)

lS&u.r~~)b

(6) Neoclassicrl (w&qges)

2.5 3.5

Keynesian (employmcd

(3

lcsrcesed Govepnmcni Chsumpth

aPercentage increases tram base soiutii-I. There are two separate wage rated for each of the three lowest skilled labcr types in our model since the total supply of each of these tmes is partitioned into a rurai sector labor force that is mobiie across the three rural sectors and a nommral labor force mobiie across the other sectors. The values shown in the column for the total labor force are the weighted av-erage wages received by each P&or type.

TOtd

Llliterates Some primary education timplet& primary education Some secondary education .fL$ier srhoo! nr.r-F. QF1ullLr\e .?r V. .G.-.-S ..V.”

htbvr

1-11)

Leontief temploywent)

-

bie 2: Impact on Employment and Monetary Wages” .._-____-.- _..... ,____ -I__ Increased ilnwrstm~nt Demand

case ofthe ~‘ivestment experimcnt~ La the first placle, invcstmc at increases Iead to haeases in the domestic output of sectors such as constrlnction,

whose demand is not price elastic and is therel’ore not cut back as pric 3srise< Similarly, investment generates a large voiume of additional imports and the trade aid transport margins on ,these generate demand for trade and transport. Since the import voilumesremain high, and tmde and transport markups are fixed, there is very littie cutback in the output of the trade and transport sector in the Keynesianasoli Ition. This zuns out, quantitativ;: _, to be an important ifactsr. By contrast,, the government consumption experiment ge.iierates large initial increases in outpul’ in certain sectors :such as agriculture and fishing (because of the large direct payment3 to labor and the resuhing rise in conslm~~3~13~), but these are cut back in the Keynesian sdlution as prices rise. Turnin,g ito tie neoclassiszl solution, we obviously cannot compare the results of he two demand experiments in terms of the ircretiise in employment of different labor types since the soluttfcthassumes that total supply of each labor type is fixed2 Xnthis case, the additional &mand for labor generated by the injection of additional exogenous demand into the system is reflected in higher monetary wage rates as the system moves to a new full-employment equilibrium. The percentage incrlease in monetary wage rates under full employment measures the increase in monetary income from wages and, to that extent, can be compared to the percentage increase in employment in the lkontief and Keynesim solutions where, with wa;es fixed, monetary wag:: incomes rise in proportion to emplcyment. We note, however, that these monetary wage comparisons must be qualified by reference to the very different behavior of prices in the three solutions. Prices are constant under t?ileLeontief closure but rise in the Keynesian solution and rise even fu&e:r in the neoclassical case. The most interesting feature of the wagBeincreases under the two experiments (see Table 2, cohunns 3 and 6) is that the increitses obtained by the two highest skill types are substantially h;gher than those obtained by the three lower skills. This pattern is understcndable in the case of the government consumption experiment sine 0 the government’s demand for labor is sharply skewed towards the higher &ill 6 a.te,goril;lls. Thi!, Produces a greater increase in the demand for the hii:he.-skilfs than for the lower sklils tmder the Leontief closure and this is trarlslated into higher monetary wage increases in th:: neoclassical case. Howel,er, this relationship does not hold in the ‘case of the investment demand experim :nt. lh this experiment. the incrEased demsnd for labor under the 1.eontief cklsure shows a s!igMly gPelo,ter increase in the demand for the:10~~ skills th;:rnfor lth~highest !~kill~, but when we move to a neoclassical clcsute, the @zttr::rn of wage jncrtases is very different. The increasr in montetarq,wage rates (and, ti ert:fore, in

334

Montek S. Ahhlwah and Frank J. Lysy

mone;inry wa,l.geincome) for the two highest Skills is now sPl'oStL?Midlly gee;ster t&m chose tar the lower skilled workers. Ihis reven,,al of the pattern of wage increases ‘byskill under one closure conlpared ‘m th’r: employment increases bq: skill un&!r anlother exempli’fies the imp3rtancc of motiel closures in ireadiny, ~onclosions about final hlpac&. It is wm-t!~exankhg the reasons for the reversal in some detail. T,he pr@ciyal adjustident mechanism in response to an increase in demand m the neoclassical closure is a rise in domestic: prices. (of about 6%) and a consequent contraction in world demand for exports. Since the exportiag sectors are more intensive in the use of the lc~werskilled labor force, we would e:xpect, in line with St&per-Samuelscsn et%:&, that the shift of produ&on aw !syfrom these sectors would p&as::e a decl:ne in the relative uqct of the lolwerskills. In 0ur model, this adverse impaet is greatly heightened by the segmentas;ion of the ttta[ labor sq~ply of each of the three lowest skills into an agricultural and a nonagricultural labor force. In each case, the agricucltural component is avdlable for the three agricultural sectors and the nonagricultural component is available for the ether sectors. The model permits flolws,from o,ne component is available for 41c other sectors. The model permits flows fro one component to another only over time, but for short-run analysis, wh.ich is what is relevant in e!xamining impact multipliers, we assum~e that labor is innn~bile a,‘-oss these components. On these assumptions, the lc~werskilled labor released from rubbler and forestry and toggingmust fi*ldecnplo;,rmentwithin the agricultural sector; d,the real wr?.geof the ayzultural labor force in the lower skiils act3)ally declines. The pa?tern of wage changes observed in the ca6e of ibe go\v:mment consumption experiment is very similar, with actual@ ,greater diPIerentials of average ‘Wages. Impact OF.Household

Real Consumption

We now :;lrn tc evaluating the impact of the two aliternative demand incn:a:ies upon the I :vel of real consumption of
WEL,FARE EFFEC

‘S OF DEMAND K4NAGEMENT

IWLICTES

335

Ke!~mssian assumptions and that it increases nkonel:aq wages in the

neoclassical Solution. In all C%ie:i, therefore, there is an increase in total monetary wages piaid out, thus raising monetary income3 of labor households. However, shce prices also rise in the Keynesian: and n:ocIassica? cases, the increase in monetary income dues not necessarily ;-eflect real imy,rovements. This welfare impact is best measured in tc:-ms of the increase in real consumption levels of different household goups in our model under different closure rules. These results are presented in Table 3, which shows percentage increases in rell consum@on of the ii 2 differer=t house.hold groups ic the model. ” Thf. behavior of real consumption in respomie to increased exogenous dcsmandunder different assumptions about facto1 av&l&iIii~ displays some interesting features. In going from the Leontief to rhe Keynesian solutions in each demand experiment, we see that household real consumption declkc, although there is still.1an increase relative to t% base so!ution. This is ttecause d.omestic output kvels, and therefore incomes, in the Keynesian solution are lower than in the Leontiefcase, althorsghthey WF:h&r than in the base solution. However, in going from the Keynesian to Lheneoclassical solutions, i.e., imposiag both a ca$aI constraint and a labor constraint, total household consumption increases in real terms tn, a level higher than the Keynesian solution and substamially above the base values. Since aggregate domestic &put in th: ncock~ssical solution is zonstrained by fixed factor sup-plies,it is necessary to e
336

Mm& S. Prltiuwaiiaa& Frank J. Lysy

It is worth neting that the increase fin household ccr.surnption in the neoclassical c8sle over the Keynesian case is mainly the to the r&e ia, consumption of the non-Malay labor househokis. Cons~umptionof both M&y and non-Malay renGers lises less than i:n the Kc,:mesian solution. This is as one would sxpect, sictce these households de1ive their income. fkor4 capital* tl the IX%X!~Skd case, the SanW Zlmount 0%capitai is combined v th less labor than in tile Keynesian case so t!zaathe marginal product oft pital is low:r, leading to lower incomes and E:onsumptionfor rentiers (rel tive to the #Leynesian sotution). The Mah labor househo~ldsdo not display the consumption gains oft& non-Malay k&or households becaur?e 1Malayhouseholds depend heavily upon the earnings c,f labor in the agricdtunrl secitoxlsand, as we have seen,, segmentation of tk tabor ma?set leads to much lower wage increases, for this group than for u&an labor. Thus the segmentation of the labor ma&et built into the: mock! leads to an asyr metric pattern off distribution of bene !its. It i 5 mteworthy thikt in both uxperkments, the nPAEclassk{ _ Aosure leads to greater differentials among the red consumption growth of differ,errt groups thar either of the other two closur $5~Two types offdifferentid are rekvmt. In the first place, higher skiiU households experience a greater incre,Be in their real consumption than r,doIl;?werskill households. Tks is in part a.reflection ofthe !Stoiper-Samuel~oi process mentionfti above 411din part due to the fat! that higher skilled l&or i,s‘fullymobile amoss all sectors and therefore dots no! suf!er fom t.he depessant effect on wages of llow skilled labor in the agricuh.uraJ labor force. Second, there i:smuch greater differential in consumption increases *&hir11 &lay households compared ’ to non-Malay househok!~:.As shown in 7’ablle3, the differential increase in reai cousmnption acre,% the s!dlEt-angefor npsn-Malay households is about 2: 1, whepzas for the Malay households 1t is about 5: 1 or 6: 1. ‘Ihis is oecause the unskilled Malay households are concer&atcd in export-dependent r3lml sectors, which, as we have seen, dispky very low wage increases, Skilled Malsy househoids, on the other hand9 00 not suffer from sny dis%lrantage because skillledlabor is mobile across sl:ctors and wages r+seuniformk~for this m&gory :moss aH sectors. By cou’:rasst., non-Malay hiouseholds of all skills being concentrated -mthe nor+uraBseczors benefit from the fact that In these sectors wages of ail groups r;isu rr:orc or less comparably.

The specific focus of this paper was 143examine differences in klpi%A multipliers generated by dhanges In rx.ogenous demand;. whesl tkesc are

Totai hoilsehoids

NJon-Malays

M&kyS

Malay Bouse&olda mub?rMec; Some primaq Comjik& #hdii2r$” Some secondary H.ighcr school certificate or more Nonmalry households Illiterates Some primary Completed primz-y Some secondwy Higher school certificate or more Wcrntier househoids

~-____

--

3.7 3.6 2.5

3.8 3,7

3.7

_-..-. 3.1

1.7

2.6

6.7

2.6

3.5

6.0

4.6 4,.6

8.5

5.2 51 5.0 6.9

3.9 3.7 3.6 .S..r?

3.0

3.2 3.1 2.9

9.6

5.4

1.4

2.3

4.2 4.3 4.2 3.9

8.9

3.3

1.7

2.9

5.1 15.3 6.0

1.2 1.3 1.3

._

3.3

6.2 6.1

632

2.5 2.4 2.3 4.3

7.3

6.3

2.2 2.4 3.1

57

4.9 3%

10.5

5.0 4.8 4.8 7.8

!3 3

;:i

2.0 2.2 7 ;

Increased Govsrnmeat Consumption -Lcont?*f r-Q Keyrdan Nedigwiclll ----

2.2 1.6

2.n

Keynesiae

Neoclassical

3.4 3.5 3.0

Leontief

Increased iavestmernt

estimated unldler dilI’erent assumpltions &out model ehwe. The results obtained are also of general interest to Imodel bui!d.ers snd pd.ky makers since they illustrate the sensitivity of model behaulfor to a?ternative assLnmpti0ns about model structure. lvIultipLier effects estimated under the Leontii;cf &sure, which carresponds co lthe assumption that all primary resources are available in unlimited SUgJpIy at ,given prices, provided little guidance to the size and pattern of multipliei~ effects generated under Keynesian and neoclassicA closure rule. In particular, under the u3E;Aassicall clo;ure rule, the injection of exogenous demand praduces substan5al changes in prices and the nature ofthese changes dep;:nds crucially upon the model structure. For example, the p;articular assun+on about labor mmket segmentation built into the model leads to a pattern of wage and prier: changes that produces a h&My af3yxunetic effect OFIdiflerent groups iin the economy. It is important to emphasize that be do not assert any general preference fat any one of the three alternative types of model closure, whether on gounds of realism or applicability. Each of these alternatives represents a h$hly stylized description ofthe conditions under whiich comparative static experiments can be perFormed and any one might be more or less relevant for a particuEar cwdrrtry at a particular point in time. In some situations, there may well be sufficient unemployment of the various labor tms!, as well 9s available unutilized capital stocks in different sectors, for the clolsed l_xor!tief model to provide a ;7alid basis for examining tht-:‘effects of marginal variations in exogenous demands. E3qually, there may be situations in which it is important to recognize that capital stocks in individlual sectors are fixed, or hen that there is no labor surplus for the economy as a whole. In practice, development planners are most likely to face hybrid situsiions in which some ofthese features are present in jiffcrent parts ofthe eco~my. Thus. some sectors may be suffering from substantial underutilized capacity while others ar-e not. SimiMy, there may be very substantial underutilization of the available unskilled liabor force whi’le highly skilled labor is in short supply. Constructing a realistic model for a part&lar cncntry requires an appropriate combination of such relevant features. Our results show that f&e pa:Gcular ch$icres ma.de in this rqeci. 41 afl’ect the conclusions qtl?te su.blstantially. Needless to say, the alteru.ative factor supply ,ass;imp!ions are not the only dimensions in which we find 4altematives about model closure. The treatment of the balance of payments is another dimension in whicn alternative formulations, which ensure a fixed defiicit, would produce very different types of model behavicr.

W’ELFARE EFPECTS

CW DEMAND MANAG EMENT POLICXES

339

APPENm‘?i

Table A 1 ofthis appendix: lists the basic equation of tile :oodel used in tkis study. They, have ‘been simplified somewhat in that corn@ -ii ?ns such as tax rates have been left out, and side equat.ions as well as eqm.tiot.: used only for accounting purpo~s have been excluded. The equations break up into three natural blocks: The first determines prices [equations (l)-(6)], the second determines input ratios [*quations (7)-(1 I )I, and the third determines outp.lt levels [equations (12)-(24)].’ The price equations follow the structure s,hown in Figure 1, The price of each upper level product is a CES price index of the prices of the lower level inputs [except in the case of capital (2), where no substitution ir allowed2). Equation (1) determines the price oi’ the labor aggregate in a se:tor as a CES function of wage mtes, (2) the price of cap&al, equation (3) the price of value added, (4) the price of the dome&.ic/import aggregate used as an interme&; :e input, (5) the price of the aggegate intermediate input, and (6) the price (Jf cutput. The price block must be solved simultaneously, since p enters *asan input in equa;tions (2) and (4) ap7d errlerges from (6) Input .ratios are determined in (7)( 1 I), and are r.lerived by repeated application of Shephard’s Lemma. As should be ooted, these re tios are functions only of p-ices in this constant return to st:alt: system. Equation (7 ) determines the ratio in which each particular laG;drtype is demanded, relative to the artificial CES construct ofa sectoral “labor aggregate” LT. The ratio of this labor aggregate to value added is cietermined in (8), +UI~thus is never directly used since we CZI always pass directly from value added to each real labor type by simp!y ;nuliiplyir!g the two ratios of’(7) and (8). Equation (9) determines the capital value-added .;atio, (l(3) the vrlfueadded/output ratio, and (1 I) the domestic iilterme diate good demand. This !inal ratio in (11) is a compound of several ratios. ‘The determination ofoutputleBrels, given prices arad i&put ra.tios, begins Lviti~(12j. The procedure is to start with a guess at the equilibrium output levels, deterr~~ine what this implies for consumption and intermediate demands, combine these with other demands, and see if this correspondls to that in ti;~lly gx:ssed. Ifn~t, iterations continue. This part of the n:odel corresponds precisely to that normally utilized r:na closed Leontief sys,tem. Eqttations (12) and ( 13) determin :: !abor and capit aildern.mdS, given ;a gues’: at output levels, and (14) and \ A5) dcfrne incomes. As seen in [ 15 ), capit,% incomes can be calculated ,using e2.her their price p, or residurtlly. Tne two are the same in a constant. returns system due !o Euler’s The1xem. E+ation (16) is a mapping to --‘The subscripts i anclj run across all lectom 1 across all iabor types. k across ali sectors groducing 3 capital input, and Et xmss all !!r.\:sel;olds. “This can be viewed as a CES in the limiting,caw where the PM ,c~tyof substitution g,oesto

;‘erO.

Montek S. Ahluwalia and Frank J. Lysy

340

(1) (2) (3) (41 (5, (6) (7) (8) (9) (10)

(11)

(12)

4 = kjv~j

(13)

Yf; ‘= W,L,,

(14)

Y.c:= p”K, = p,“V, - Yt y:;

= fc(Yj,,

. . , Yl-,,

I.

(15) *,

Yi,,

.

I

I

)

u:.

.

.

,

1’5 ., ., Y$?,

.I,? =’ Y:IHh _I

--_-_-----..-__---

(16) (17)

___________I____-

-

-Y-s (continued)

WELFARE EFFECTS OF DEMAND MANAGEhiFNT

POLICIES

341

-

Table Al. (imntinuc~) m-P

Q.

I

-

(18)

=?a':

Xi -Fa$Xi

-+ Cf'i

E$'+

ip -+ Gp

(24)

determine household incomes, which will be a function of the factor nwmes generated. There is a similar mapping for the numbers of earners in the households who are currently employed. Equation (17) determines per eamer household incomes, and ! 18) is a simple linear consumption function, with no introcept in order to maintain the homogeneity of the system. Cons:umption demand is determined in ( 19)-j 22). wher: we have simplrfied the three-level system of the model to a two-level one. Here, we nave treated only the domestic/import and the intergoods choices, which are in fact, the crucia’ ones. At the bottom level, a CES combination of domestic and imported goods is USC&and the price of this aggregate is defined in ( 19). At the ?op level, the choice between the

various goods is determined according to Stone’s LES system, which implies the demand system of (20). Once the demand P-B-tke domestic/in,wrt aggregate is known, the demand for the aomestic component can be found ,uing Shepherd’s Lemma [equation (21)j. , The constant price elnstici*y export demand function is (23). where E is the elasticity, fl the foreign exchange rate, ITthe world price, parameter. Quation (24) is the material balance for domestic g investment an.3 government demands are er.agensus. In the Lcontief closure, prices and input r ^Itiosare fixed solved from ( 12) onward for any gi*lenlevel of exogenous k&or and capital demands are gives in (121 and ( 15 1. In the K~y~~c~i~~s&~tk 7. capital demards are driver) back to their ~~-@nal Icveis, after the ir~ca% in

Montek 3. Ahluwaiia and Fa~nk J. Lys,y

332

1:xogenou.s final demands, by raising 6 (or, equivalently, I$-), VhePn the capital rental rate is rahed, all prices change, and new capital demands ;EUI e csdcullattid. In the neoclassical closure, a similar procedure is applied to wa,ge rati:s to hce hhor demands back to their base! levels.

REFERENCES S. f 1978) Income DistributionPolicy in Dewloping Counties: A C&e Study ofKumw. London: Oxford University Pt$ss. Atiuwalia. Montek S., and Lysy, F. (1979) Mathem&icd Structure of the General Equilibrium Model of Mdaysia. World Bank, Waslhin&on, D.C. (mimeo). Armington, P. (1969) A Tht:tny for Demand for Produ.cts Distinguished by Place of Production, IMI;‘ St@Papers 16, 159-l 78. Bell, C., and Hazeli, P. ( 1980) Measuring the Indirect Effects ofan Agricultural Investment Proj,ect on Its Surrounding Region, Americarr .Iourn~~l~~AgricultwuI Economics (to

Arj&nan, I., and Robinson,

arw=).

Cordoso, E. .A., and Taylor. L. (197’9) Identity-Based Planning of Pdce:s: and @anti&s: CaGb:b;idgeand Neocla.ss:ical h fodels for Brazil, Jou.?zzl ofPo&y P.~fodeh,p I, 83111. Pyatt, G., z:.! Round, J. I. r 1’9’79),&ounting and Fixal Price Multipliers in a Social Accounting Matrix Framework, ECG wmic .hrno~ (forthcoming). RObinsOR,S., and Dervis, K. (1’378) The li:oreign Exchangl,(.:Gap, Growth and Industrial StraIegy in Turkey: 1973--1983, HBRD Working Pap::r No. 306. Taglor, L., Bacha, E., Cardoso, E. A., and Lysy, F. (1979) Growth (and Diswihution in Brazil (forthcoming). T2sy&>r,L., and Lysy. F. (1919) Vanishing Short-Term Incg.bmeRedistribt_rtion: ‘Keynesian Clues About Model Surprises, JOUUIU~of Develops e clt Eccwomics (to n:ppwr).