World Bank Reprinft Series: Number 318

Sweder van Wijnbergen

rect P011.yf , h-ilation, andGrowth in a Firni ailyRepressect Economy

Reprinted with permission from Journal of Development Economics, vol. 13 (1983), pp. 45-65,copyrighted by North-Holland.

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

World Bank Reprints

No. 277. Oli Havrylyshyn and Iradj Alikhani, "Is There Cause for Export Optmism? An Inquiryinto the Existence cf a Second Generation of Successful Exporters," WeltwirfschaftlichesArchiv

No. 278. Oli Havrylyshyn and Martin Wolf, "Recent Trends in Trade among DevelopingCountries," European Economic Revieuw

No. 279. Nancy Birdsall, "Fertility and Economic Change in Eighteenth and Nineteenth CenturyEurope: A Comment," Population and Development Review

No. 280. Walter Schaefer-Kehnert and John D. Von Pischke, "Agricultural Credit Policy inDeveloping Countries," translated from Handbuch der Landwirtschaff und Ernaihrung indet Entwicklungslandern (includes original German text)

No. 231. Bela Balassa, "Trade Policy in Mexico," World DevelopmentNo. 281a. Bela Balassa, "La politica de comercio exterior de Mexico," Comercio ExteriorNo. 282. Clive Bell and Shantayanan Devarajan, "Shadow Prices for Project Evaluation under

Alternative Macroeconomic Specifications," 7he Quarterly Journal of EconomicsNo. 283. Anne 0. Krueger, "Trade Policies in Developing Countries," Handbook of International

EconomicsNo. 284. Anne 0. Krueger and Baran Tuncer, "An Empirical Test of the Irfant Industry

Argument," American Economic Revicuw

No. 285. Bela Balassa, "Economic Policies in Portugal," EconomiaNo. 286. F. Bourguignon, G. Michel, and D. Miqueu, "Short-run Rigidities and Long-run

Adjustments in a Computable General Equilibrium Model of Income Distribution andDevelopment," Journal of Development Economics

No. 287. Michael A. Cohen, "The Challenge of Replicability: Toward a New Paradigm forUrban Shelter in Developing Countries," Regional Development Dialogue

No. 288. Hollis B. Chenery, "Interaction between Theory and Observation in Development,"Vrld Development

No. 289. J. B. Knight and R. H. Sabot, "Educational Expansion and the Kuznets Effect," TheAmerican Economic Revieuw

No. 290. Malcolm D. Bale and Ulrich Koester, "Maginot Line of European Farm Policies," ThelArld Economyn

No. 291. Danny M. Leipziger, "Lending versus Giving: The Economics of Foreign Assistance,"UVorld Development

No. 292. Gregory K. Ingram, "Land in Perspective: Its Role in the Structure of Cities," WarldCongress on Land Policy, 1980

No. 293. Rakesh Mohan and Rodrigo Villamizar, "The Evolution of Land Values in the Contextof Rapid Urban Growth: A Case Study of Bogota and Cali, Colombia," V\rld Congresson Land Pblicly, 1980

No. 294. Barend A. de Vries, "International Ramifications of the Extemal Debt Situation," TheAMEX Bank Review Special Papers

No. 295. Rakesh Mohan, 'The Morphology of Urbanisation in India," Economic and PoliticalWeekly

No. 296. Dean T. Jamison and Peter R. Moock, "Farmer Education and Farm Efficiency in Nepal:The Role of Schooling, Extension Services, and Cognitive Skills," World Development

No. 297. Sweder van Wijnbergen, "The 'Dutch Disease': A Disease after All?" The EconomicJournal

Journal of Development Economics 13 (1980) 45-65. North-Holland

CREDIT POLICY, INFLATION AND GROWTH IN AFINANCIALLY REPRESSED ECONOMY

Sweder van WIJNBERGEN*Massachusetts Institute of Technology, Cambridge, MA 02139, USA

Received March 1982, final version received August 1982

We analyse credit policy in an open economy macro-model incorporating stylized facts aboutthe financial sector of LDC's (absence of security markets, existence of a curbmarket). Thesefacts are incorporated in the financial sector part of a Keynes-Wicksell growth model. We focuson a transmission channel of monetary policy into the supply side of the economy via the costof working capital. We show that ignoring this effect when analysing restrictive credit policyleads to underestimates of inflation and overestimation of output and current accountimprovements. It is shown that a credit crunch can be stagflationary in the short run. A cut inthe real stock of credit is shown to lead to a lower steady-state capital stock.

1. Introduction

In short-run models of an open economy supply behaviour is typicallyeither assumed to be totally elastic, which leads to Keynesian multipliermodels [see for example Metzler (1942) or Mundell (1968)], or to be fixedexogenously, as in the models often used by adherents to the MonetaryApproach to the Balance of Payments.

It seems to be generally accepted that the first assumption is inappropriatefor an analysis of what is going on in most LDC's: tariff structures andsticky real wages cooperate in producing capital shortages and surpluslabour at the same time. This has led to the widespread use of theassumption of fixed supply and therefore real income. If one adds to this atarget path for future prices (to be reached by incomes policy) the futuredemand for monetary liabilities is predetermined. This in turn leaves only thecomposition of monetary assets (Net Foreign Assets and Domestic Credit),and not the level, as an instrument of monetary policy. An immediateconclusion is that domestic credit is an effective instrument to control the

*Current afiliation: Development Research Department, World Bank. I am indebted toStanley Fischer and Lance Taylor for encouragement and comment on various drafts, to FrancoModigliani for comments that lead to several improvements, and to participants in seminars atMIT and Harvard. I also wish to acknowledge financial support of the Common Wealth Fundvia a Harkness fellowship and the Niels Stensen Foundation. The views expressed in this paperare those of the author and should not be interpreted to reflect those of the World Bank or itsafliliated institutions.

0304-3878/83/$3.00 © 1983, Elsevier Science Publishers B.V. (North-Holland)

46 S. van Wijnbergen, Credit policy, inflation and growvth

NFA position. This is, incidentally, the analytic framework behind IMFstabilization policies [cf. Polak (1957) or Robichek (1978)].

There is increasing evidence however, that these models fail to captureelements that are necessary to understand and predict the consequences ofcredit policy in LDC's. IMF reviews of their own stabilization attempts showthat they have a persistently poor score on both real growth and inflation,overpredicting the first and underestimating the second [cf. Reichman andStillson (1978)]. Specific country studies [Cavallo (1977), van Wijnbergen(1978, 1982); cf. Bruno (1979) for a theoretical contribution] point to certainfeatures of the financial structure in a typical LDC as important left outfactors. It is very important that securities markets and consumer credit arevirtually absent in most LDC's. Bank credit is mainly used for business loanswhich are needed to finance working capital (and quite often fixed capital).This results in a direct link between credit and the supply side of theeconomy, thereby violating the assumption that real income is independentof the credit policy enacted.

A second issue stems from the observation that, if tight credit is used toreduce expenditure and so redress current account imbalances, investmentoften takes a large share of the adjustment burden. Is external balancebought at the expense of long-run growth prospects?

To analyse long-run aspects, we obviously have to construct an openeconomy growth model. Monetary growth models are virtually absent fromthe International Economics literature, and the few that exist [cf. for exampleDornbusch (1975)] are really open economy extensions of the neoclassicalmodels used in the money-growth debate of the late sixties and earlyseventies. While these models have desirable steady-state characteristics, theirassumption of continuous market clearing and full employment 'assumesaway a host of interesting short run problems', as Dornbusch notes in theintroduction to his JME paper. In view of the fact that we want to analyseboth short-run and long-run consequences of credit policy, these models areclearly inappropriate.

The assumption of a high speed of adjustment in asset markets is probablynot unrealistic, certainly not in the presence of large unorganized moneymarkets. However, there is substantial evidence that goods and labourmarkets adjust slowly (compared with asset markets). Accordingly, we willpresent a model that stresses differential adjustment speeds in goods andassets markets. In doing so, the model resembles the closed economyKeynes-Wicksell growth models that were briefly debated in the late sixties.Fischer (1972) shows that, if correctly specified, KW models have the samesteady-state properties as comparable neoclassical models, but exhibit morerealistic short-run behaviour.

To be able to analyse the first, 'short-run' issue discussed above, we willincorporate a financial sector that satisfies all the consistency conditions

S. van WUjnbergen, Credit policy, inflation and growth 47

outlined by Tobin (1975), but stresses the role commercial bank credit playsin financing working capital in most LDCs.

In section 2 the model is presented. In section 3 the short-runcharacteristics of the model are analyzed, and impact effects of a creditcrunch are discussed. The conclusions emerging from a comparison with afixed supply version of the model are that neglect of the direct credit-supplyside link via the financing of working capital will lead to:

(1) underestimates of inflation,(2) overestimation of real growth,(3) overestimation of current account improvement.

The last point depends on the assumptions made on spill over effects fromgoods market disequilibrium.

In section 4 asset accumulation over time is introduced together with pricedynamics to analyse the steady-state characteristics of the model and long-run consequences of tight credit policy. It is shown that this will reduce thelong-run capital stock. In section 5 we show however that in the admittedlyextreme case of perfectly elastic supply of foreign loans, domestic credit hasno influence on the long-run capital stock and domestic product (it doesinfluence national income: tight credit in this case leads to larger foreignindebtedness and therefore to a higher debt service burden).

2. The model

2.1. The financial sector

The model presented here incorporates several stylized facts typical offinancial markets in LDC's.

Markets for non-monetary government debt are largely non-existent. Anunhappy consequence of this is that government deficits automatically feedinto the money supply. Also non-existent are significant markets for equitiesand commercial paper. Intermediation between private wealthholders and thebusiness sector takes place either via the banking system or the curb market.So firms have to finance their fixed capital stock (and its increment,investment) and the working capital they need (more on that in the nextsection) by short-term bank credit, retained earnings or loans from theUnofficial Money Market.

We will consider retained earning as a loan by the owner of the firm to thefirm, and put the firm owner in the same category as money lenders at theUnorganized Money Market, the category 'Public'. In the model presentedhere we will represent all this by assuming that firms issue bonds to finance

48 S. van WYijnhergen, Credtit policyi, inflation and growth

their capital needs. These bonds are held by banks, owners of firms andmoney lenders at the UMM.1

Commercial banks are assumed not to hold free reserves, not anunrealistic assumption for inflation-plagued LDC's. This makes reserverequirements and ce..ings on the volume of bank loans mutually dependentinstruments. In what follows we will assume that credit ceilings are used, butidentical results can be obtained with the other assumption. On their liabilityside, banks do not have control over deposit rates, these are fixed by themonetary authorities (which, of course, is also the case in most developedcountries). This implies that they are thrown off their supply curve and haveto accept whatever amount the public wants to hoid as deposits at thebanking system [Tobin (1975)].

Private individuals, finally, can hold their wealth as money (cash or bankdeposits2) or they can acquire direct claims on firms by lending on theUnofficial Money Market.

A series of balance sheets, based on the financial structure whose outlineswe just sketched, is presented below.

CB Commercial banks Firms The public (firmowners and others)

NFA C R DD pK pB, M1 pW

ND(,CB R pBh pDr pBP

M = DD + C, MB=NFA +NDC '1,

pW= M4 +pBp= M r pK pI-Dr- pBb =MB+pK +pDf.

For a list of symbols and their meaining, see the list of symbols in theappendix.

The behavioural equations transforming all this accounting into economicsare as follows. The public has a Tobin-type demand for money,

Md = rn(i, y)pW (1)

'Capital gains on the firm's assets create a problem in a model without equity on the liabilityside to 'absorb' the gains. The natural way of solving that is to introduce equity in this model,althouLgh the introduiction of a market in such claims would clearly reduce the realism of themodel, stock markets do not play a significant role in LDC's. I have chosen not to introduceequity because of that reason. Instead I have made the simplifying but admittedly very artificialassumption that firms issue new debt simultaneously with the capital gain and distribtite thecash so raised randonly/ among bond holders. The point that this transfer is distributedrandomly is important because in this way the transfer does not become part of the rate ofreturn on those bond holdings, i can therefore still be initerpreted as the (nominal) rate Or returnon bond holdings.

2For simplicity and without loss of generality, we assume that wealth holders maintain a fixedcash -demand deposit ratio.

S. van Wijnbergen, Credit policy, inflation and growth 49

where i represents the interest rate in the Unorganized Money Market(UJMM from here on). This implies a supply of 'private loans' via the wealthconstraint:

pBp =(1-"ni, y))pW- (2)

Further, firms can borrow from banks. Equity is assumed not to existwhile retained earnings are modelled as if they were UMM loans (from theowner qf the finn to the firm). As to the asset side of their balance sheet,they have a capital stock that is fixed in real terms at any momenit in time.They also have to finance their working capital needs Df (in real terms; moreon Df below). Their total demand for funds (in real terms, deflated by p, theprice of the good they produce) becomes

B,=Df +K. (3)

Equality between demand-supply for funds at the curbmarket wouldrequire that the part of the business sector's demand for funds not coveredby bankloans (B,-Bb)3 is matched by the public's supply of curb-marketloans, Bp (all quantities are deflated by p):

B,-Bb=Bp, or (4)

Df(w, y) + K (1 -m(i, y)) W+ Bb. (4)

Demand for working capital Df(w, y) is derived in section 2.2.Using the balance sheets constraints on banks and private individuals, it is

easy to show that equilibrium in the curbmarket implies equilibrium in themoney market.

The critical question is of course, will this equality in fact obtain, does thecurb-market rate move quickly enough to clear the market. The acceptedwisdom seems to be that it does not [see for example Wai (1957) orMcKinnon (1973)]. The argument is always the same, the persistence of wideinterest differentials fot similar loans is taken to be evidence of a segmentedmarket in disequilibrium, a case of credit rationing (see the two authorsmentioned above). The conclusion of disequilibrium however seems to me anon-sequitur. The point is that the different rates may be charged on similarloans (similar maturity etc.), but not to similar customers. In fact the mostwell known explanation of the persistence of credit rationing in highlydeveloped but regulated financial systems such as the American one is

'The implicit assumption made here is that firms try to obtain bank credit before going intothe curb market Bank credit is typically offered at below market clearing rates in most LDC's,so this is not an unrealistic assumption.

50 S. vani Wijnbergen, Credit policv, inflatiotn and growth

exactly the inability of banks to charge different rates to different customersfor the same type of loans [Jaffee and Modigliani (1969)]. The fact that thisis possible in curbmarkets, and the ensuing interest differentials on similarloans are thus an argument for equilibrium in this market, not fordisequilibrium. For empirical evidence in support of this claim, see vanWijnbergen (1982).

2.2. The real sector

Production is subject to a production function assumed to be separable incapital K and other factors L4

y= f (K)h(L)lla, a> 1 (5)

This can be described alternatively by the restricted cost function:

c-=ya f - ag W) (5 )

with w a vector of real (in terms of home goods) factor payments. (5') canbe derived from (5) via simple cost minimization. We will assume that allpayments to 'other factors' are financed by short-term credit, so the demandfor working capital is'

Df = Df(w, v), (6)

where both partial derivatives of Df are positive. We will assume w, the realwage in terms of domestic goods, to be parametrically determined, asmentioned in the introduction.

Firms will attempt to maximize profits py-pc(l +i-p$). This leads to

I =aV,a If -ag(Wr)(l +ifi) (7)

or price equals marginal costs [inclusive borrowing costs (i-p)c]. We takethe real instead of the nominal rate to represent the marginal cost ofborrowing, which is consistent with our assumption of a fixed real wage.6

4For a similar model of firm behaviour see Bruno (1979).'Although for any single firm c-yuf-g(w) is a flow variable, it is straightforward to show

that aggregation over firms leads to a stock demand for credit n-c with the stockiflowconversion factor tn linearly dependent on the production delay. A discrete delay in productionimplies that output y! cannot change instantaneously to the new short-run equilibrium valueafter a disturbance, but can only do so after a time period equal to the duration of theproduction delay. This will be ignored in what follows,

'This is covered in app. D (available on request).

S. van Winbergen, Credit policy, inflation and growth 51

Planned expenditure on domestic goods is

Ad= Cd(, Yd, W) +I+ G +E(q) (8)

with q-p/(ep*) and Yd=y-T-PeMR. Capital gains are included in thedefinition of disposable income to maintain consistency between thedefinition of private saving and private wealth (i.e., to be able to define wealthas accumulated savings).

The investment function is a conventional Keynes-Wicksell investmentfunction:

I =I(i -p, K). (9)

With i and y determined simultaneously by firms and money lenders, thereis obviously no guarantee that demand for domestic goods Ad will equal thesupply y, assuming that prices cannot jump to clear the market. This comesdown to assuming immediate adjustment to disturbances on the assetsmarkets but slow (depending in A, see below) adjustment on goods markets.

If firms noticing the excess supply (demand) for their products react bytrying to adjust their relative price, aggregation over firms leads to a priceinflation equation:

p =no + A(Ad-Y(10)

For a discussion of this equation see Barro (1972) or Fischer (1972),7Finally we have to specify our expectations hypothesis. While somewhat

unusual for an analysis of a LDC type economy, we will assume RationalExpectations (RE). The fact of a less highly developed financial and industrialstructure seems to md an insufficient argument for persistent suboptimalbehaviour. Moreover, the RE assumption permits one to disentangle theconsequences of expectional errors from other phenomena under stuidy. REin this non-stochastic model comes down to perfect foresight:

=re ', ( 1)

Noting .Lat Xr=yp+(1-y)7r* with it* the 'imported' foreign inflation rate,(10) can be rewritten to get

j5-,g* +; !(Ad-Y), or (10')

7 Firms adjust their output price with respect to the prices of all other goods traded in theeconomy, including imports. Accordingly, expected general inflation 7r

6 rather than expecteddomestic goods inflation appears at the right-hand side of eq. (10).

52 S. van Wijnbergen, Credit policy, inflation and growth

Ad-y=k.'l(P-*) with k=iA(I-y)-' (10")

which will turn out to be convenient later on. It is maybe worthwhile topoint out that the assumption of RE adds an immediate transmissionchannel between jumps in the foreign inflation rate and the domesticeconomy. While any well specified expectations mechanism should have thisfeature in the long run, under the RE assumption changes in the foreigninflation rate are incorporated without delay in the expectations formationprocess.

Finally there is the question of how to allocate output when Ad9#y.Obviously at this level of aggregation development of a theory of allocationunder disequilibrium is not likely to be a fruitful exercise; instead we followFischer (1972) and assume that both investors and consumers will have togive in when Ad> y. In particular, actual investment becomes:

I"=I-(1-z)(Ad-y)= 1-v(p-r*) with v=(1 -z)k-1 .

(1 -z) is the fraction of the total amount of rationing falling on investment,z falls on domestic and foreign consumption of our goods (respectively Cdand E). The model is however not a proper Malinvaud-type dis-equilibriummodel because output is assumed to be always supply determined, not just inexcess demand situations.

Spillover effects due to frustrated demands for domestic goods obviouslycannot cause disequilibrium in either money or loans (UMM) market: thespillover is the difference between two flows and therefore a flow itself andcannot instantaneously change any stock. However, spillover effects doinfluence the current account (another flow variable). This issue will be takenup later on.

3. Analysis of the model

At any moment in time K, MR and q are given, inherited from the past.Given these state variables, the 'short-run' model boils down to

=ayaf fag(W)(1 + i-P), supply, (13)

f +K=(1-m(i,y)) W+Bb, equilibrium in the UMM, (14)

p = * + (Ad -Y (15)

Equilibrium in the assets markets together with supply behaviour gives usone relationship between the inflation rate p and income y. Differentiating(14), the assets markets equilibrium equation, gives

S. van Wijnbergen, Credit policy, inflation and growthi 53

(mDfy +my W)dy+ mi Wdi=dBb. (16)

Eq. (13), supply behaviour, yields

dy di -_p O__(a + l d1 + p(17)

Substituting out di gives a positive relation between y and p:

d p (a 1) (1 -) (mDY + my W) 0. (18)dy LM Y miW

To see what is going on, consider an increase in p, the actual and expectedrate of change of domestic goods prices. Ceteris paribus, this would lowerreal credit costs and therefore induce an additional supply of domestic goods.However, this also increases demand for money and thus reduces the supplyof loans while it increases the demand for them, leading to an increase in thenominal rate i. This partially (otherwise the expansion.would not start tobegin with) offsets th.. initial disturbance. Together this yields a positiverelation between p and y [eq. (18) and fig. 1].

This relation (labeled S in fig. 1) indicates what will be produced atvarious inflation rates, given a market clearing interest rate at all times, andmight be called a 'pseudo-supply curve'.

p S

y

Fig. 1

Differentiating eq. (15) which links the change in domestic prices to foreigninflation and disequilibrium in the market for domestic goods, gives eq. (19):

(I +,(cyMR+Ij))di= -,i(1-cy)dy+J*Ildi. (19)

54 S. van WUjnbergen, Credit policy, inflation and growth

The coefficient of dp, (1 +k)(cyMR+I1 )), plays a crucial role in thestability analysis. If it is positive, all three necessary and sufficient conditionsfor stability of the stock equilibrium will be satisfied (cf app. A.1).8 It tulnsout that the coefficient has a straightforward interpretation: -h(cYMR+fI)

is the contribution of a 'blip' in inflationary expectations to the actualinflation rate p. It is partly positive, because given everything else, it reducesthe real rate and so increases investment demand, and partly negative(- )cyMR), because higher expected inflation implies higher expected capitallosses on nominal assets and so reduces disposable income y-T PeMR.

The condition

1 +(c,YMR+ID) >0. (20)

simply states that if the actual inflation rate changes because an increase ininflationary expectations disturbs goods market equilibrium, it shouldincrease by less than the original disturbance.

Combining the asset market equilibrium condition (16) with (19) gives asecond relation between p and y, which might be called a 'pseudo-demandcurve'. This relation gives the rate of (domestic) price inflation p generated byexcess demand pressure in the goods market, again assuming a marketclearing interest rate,

dp k - [(1-c,y) + IIl(mDfy + my W)/(mi W)] 1dy D l + h(cyMR +LI) < v. (<

The curve slopes downward: an increase in output y leads to an onlypartially offsetting increase in aggregate demand (cy < 1). It also increasesmoney demand which, given the money supply, leads to higher interest ratesand therefore less investment expenditure. Both factors slow down inflation.

Now we can analyse the short-run effects of a lowering of the ceiling ondomestic credit to the private sector, Bb (dc.e to the assumption that theprice level p cannot jump, we may indeed consider the credit ceiling in realterms, Bb, as our policy variable),

The decrease in bank loans tightens credit conditions, thereby notpermitting as high a production at the ongoing inflation rate as before:Working capital has become more expensive to finance (i.e., the S curve shiftsto the left). Formally, this can be shown by solving eqs. (16) and (17) withdp=0 to get eq. (22),

dy m.WLM=[mDfy+myW] i (a-1)(1+i-P). (22)

'Appendices are available on request.

S. van Wijnbergen, Credit policy, inflation and growth 55

Demand however will also be effected. Tighter credit will push up theinterest rate and lead to lower investment given the inflation rate: theD-curve shifts to the left too,

dy >0an1BbodBb I (1 -cy)miW+I(mDfY+mYW) b

The net effect on inflation depends on whether the impact on the supplyside via higher costs of working capital (the shift of the S-curve) dominatesthe traditional negative impact on demand via depressed investment. Thiscase, with strong supply effects and/or weak interest sensitivity of investment,is depicted in fig. 2. The output goes down more than aggregate demand forour good, so that investors and consumers will have to be rationed. Inflationwill increase from it*, the world inflation rate from which we started byassumption to PB It is worthwhile to point out that investment under thisscenario is hit twice: first the real rate rises (see app. A.3), so intendedinvestment I goes down; second actual investment r falls short of theintended level due to the rationing. Our model thus tends to confirm the realworld fact noted in the introduction, that investment takes a large share ofthe adjustment burden.

S,- S

Fig. 2

Obviously even if we neglect asset accumulation over time, to which wewill come back in section 4, this is not a sustainable equilibrium: the terms oftrade will change and competitiveness will decline as long as p>7r* (and theexchange rate remains fixed). This will shift both foreign and domesticdemand away from our goods, shifting the D-curve until domestic andforeign inflation rates have converged at point C (fig. 2). This is not aparticularly interesting dynamic exercise however, the assumption of constantasset stocks is highly unrealistic.

56 S. van WVfbergen, Credit policy, inflation and growth

Without the supply effects or wi-th very interest sensitive investment, thedemand shift will dominate the supply shift and both production andinflation will go down after a credit crunch, a more familiar result (see fig. 3).In this case the drop in supply is demand induced and no rationing ofconsumer or investment expenditure occurs [if eq. (10') applies, and note thatPB <n* now].

D D

v

Fig. 3

The observation that inventories tend to be low in periods of tight credit,and the fact that IM1F stabilization programs, which rely heavily on the useof credit ceilings, nearly always perform poorly on their inflation and realgrowth targets [see again Reichman and Stillson (1978)] might be explainedtherefore by the presence of strong supply effects due to the credit-supplylinlk modelled above.

Before we -finish the short-run analysis by looking at the current account,we will have to say something on spillover effects. The current account is'theonly place where they show up immediately. While it may'be reasonable toassume spillover effects of frustrated consumer demand for domestic goodsinto the foreign goods market, this is not obvious for investment. At least inLDC's, domestic and foreign investment are ffiore 'likely to becomplemental-y than substitutes, tide former mainly consisting ofconstruction, the latter of machinery imports. Accordingly, we will assumespillover of consumer expenditure, but no spillover of frustrated investmentgoods demand into the foreign market. The issue is relevant becauserationinlg in the domestic market without spillover into other expenditurecategories will lead to anx unplanned curtiling of expenditure given incomeand therefore to an improvement of the current account.

T'he current account equals actual income minus actual expenditure (withthe possibility of disequfibriumw in the home goods market, it does notnecessarily equal income minus planned expenditure):

S. van W07bergen, Credit policy, inflation and growth 57

actual income

y-cd +I+ G E(q)-, - -23a)

actual expenditute

AT = Cd + cm+ I + G- (1I - z)* ( *), (23b)

CA =E1(q) -c.(q, YD, W3 - Zk -(f -,* = )CA -Z z1( -n*) (23c)

with

CA- E(q)-cm. (23d)

Cm(=Ct-Cd) iS consumer demand for imports. CA is the ex ante currentaccount surplus. The actual or ex post current account surplus CA equals CAminus the rationing term z -I(p -7r*).

Under both scenario's (supply or demand effects dominate the impact oftightening of credit, see figs. 2 and 3), the ex ante current account surplus CAimproves [cf. (23d); in both cases disposable income Yd goes down while qand W do not change instantaneously]. If the credit squeeze is deflationary,the current account CA will improve even more [see (23c); in this case

< n*] unless all rationing falls on investment (z= 0).However, if the impact of the credit squeeze is dominated by supply effects

via higher costs of working capital, inflation will accelerate and the actualcurrent account will improve less than the ex ante current account, and infact may even deteriorate [now p> *, cf. again (23c)], as long as at leastsome of the rationing falls on domestic or foreign consumers of our goods(z>0).

4. Dynamics

To be able to analyse long-run consequences of stabilization policies, wehave to consider asset accumulation and the way it is influenced by thepolicy under study.

Capital accumulation equals investment I minus depreciation 5K, or, usingeq. (21),

(25)

Accumulation of real fmancial wealth equals domestic credit creation andnet change in foreign assets (balance of payments deficits) minus inflationaryerosion pMR:

N,bCB NPA ^+P,-= MR,. (26)

p p

58 S. van WUnbergen, Credit policy, inflation and growth

NIJA equals the current account in the absence of capital flows, in that caseNPA = Y-CT-I- G, income minus expenditure NJ6CcB can be substitutedout using the government budget constraint G=T+NICgB. If we furtherassume zero taxes for simplicity (T=O), all this leads to

MA=G+Y-CT -G-PMR=Yd-CT-I+V(p 7*)' (26)

Finally as long as A + a*, .the terms of trade change, preventingequilibrium in the goods market. This is expressed in the final equation:

= p1a*- (27)

Long-run equilibrium will be characterized by unchanging terms of trade(4=0) and therefore convergence of foreign and domestic inflation rates, aconstant capital stock (keep in mind we assumed zero population growth;generalization to positive growth is trivial and changes nothing essential) andgross private savings sufficient to cover replacement investment and to 'paythe inflation tax':

y-cT=bK+pMR. (28)

To actually solve this 3 x 3 system, the variables i, p and y have to beexpressed as functions of the state variables K, MR and q:

y = y(K, MR, q; Bb, w, 7r* G), (29)

p = p(K, MR, q; Bb W, 7r*, G), (30)- + -?7 + + + (0

1= i(K, MR, q; Bb, w, rc*, G), (31)

Yi is ambiguous: on the one hand, a higher K increases capacity, whichincreases the supply of output, but on the other hand it depresses investmentI, which, ceteris paribus, will lower p, and it raises demand for loans, thusleading to higher i. Together this increases the real rate, which will lower thesupply of output. In what follows we will assume that the capacity effectdominates. We will mention in footnotes places where relaxing thisassumption would reverse the conclusions.

The ambiguity of p4, dp/dBb, has already been discussed in section 3(cf. figs. 2, 3). The sign of P2 could iiot be determined uniquely from the

S. van Wijnbergen, Credit policy, inflation and growth 59

assumptions made. We assume it to be positive for reasons that will becomeclear in the stability analysis. The remainder is fairly self evident.

The model can be analyzed using fig. 4. Using eq. (27) to eliminate q, weare left with two equations describing equilibrium values of K and MR. TheK-locus gives a set of points in K-MR space at which net additions to thecapital stock are zero. An increment in MR will lower i and therefore i-n*,the real rate (note that in equilibrium P=n*). This will raise investment untila new equilibrium is reached at a higher K(12-<O). So the curve is upwardsloping:

dK |,' --(l2IIi3P2P3 >) >O (32,dMR ocus,,, (111+1 2-11i3P1P3)

K

-locus

K-locus

MR

Fig. 4

The MR locus gives those combinations that lead to zero net additions tofinancial wealth (in real terms). An increase in MR will reduce net privatesaving, which can be offset by a higher K: this will increase income andtherefore saving, and reduce net investment in capital goods, which, together,will increase net financial wealth accumulation. The MR locus, therefore,slopes upward too:

dK -((1 CT,y)(Y2 .7*) Iji2 ((l CT,y)Y3 I1 i 3)P2P3 ") >

dMR |mR.I.Us ((' CT,y)Yl-(,Ii +I2)-((1-CT,y)Y3-I1i3)P1Pf3 1(33)

It follows from one of the stability conditions that the MR locus is steeper

60 S. van WUjnbergen, Credit policy, inflation and growth

than the K-locus:

dK. dK.>- (34)dMR MR-locus dMR (oc34)

Now consider again the credit crunch from section 3, a reduction in theceiling (in real terms) on the volume of bank loans outstanding. As in theanalysis of the impact effects, the results depend crucially on whether thecredit has an impact on supply or not. In section 3 it was shown that supplyeffects could actually reverse the sign of p4, making it negative. Consider firstthe case of P4 <O.

Given MR, a lower level of credit will tighten credit conditions, raiseinterest rates and reduce net investment, till a lower K has offset thedisturbance, so the K-line shifts downwards (cf. figs. 5, 6). Further, a creditcrunch will reduce disposable income and therefore saving, leading to anecessary offsetting reduction in MR: the MR-locus shifts to the left (fig. 5).The final result in this case (fig. 5) will be a lower capital stock and lowerfinancial wealth (in real terms) due to the tight credit conditions. Thisobviously implies a lower production of domestic goods and lower income.

K jMR'

I/

/locus

'A

8/" .- "K'

MR

Fig. 5

The picture will be less clear cut if we look at the other case, P4>0. Whilethe K-locus will shift down as before, what happens to the MR locus isambiguous. It can be shown that the long-run result will still be a lowercapital stock, but financial wealth may either increase or decrease (fig. 6illustrates the first possibility).

A reduction in long-run financial wealth does not necessarily result in areduction in NFA. What happens to Net Foreign Assets depends obviouslyon what the government does with net domestic credit. The governmentbudget constraint tells us that

S. van Wtijnbergen, Credit policy, inflation and growth i61

K MR focusi / MR'

/ /

/

A g K locusA /-

/

MR

Fig. 6

G N15CB/P = (NDe/p) +1P(NDC/p). (35)

If the government aims at a zero current account deficit, all nominal financ,ialasset accumulation n*MR will have to be provided by domestic creditexpansion:

G n*MR =(NDC/p) + *(NDC/p), or (36)

(NDe/p) = *(NFA/p). (36')

In this way the government exactly offsets private net saving, with a zerocurrent account deficit as a result. Now if we assune the government wasactually hitting a zero current account target before the credit crunch (i.e.,G = ir*MRO), it is obvious that in the new long-run equilibrium

G = 7*MRo> n*MR1 , (37)

in the case analyzed in fig. 53 or an unchanged government deficit will in thenew equilibrium lead to a current account deficit. Endbgenous tax receiptsT= T(y) clearly would make things worse. A resulting cut back ingovernment expenditure to get the current account in line will lead to furthercontractionary effects.

5. Foreign lending

In this section we will relax the assumption that firns have access todomestic sources of funds only, In fact we will go to the other extreme andassume that they face an infinitely elastic supply at a fixed foreign interestrate i*. This implies that firms arbitrage between foreign and domestic assetsmarkets, thereby fixing i at i*, the foreign rate (expected exchange tateJ.D.E.- C

62 S. van Wijnbergen, Credit policy, inflation and growth

changes are zero, we are in a fixed exchange rate regime). As will be obvious,a decrease in Bb will induce an equal amount of foreign lending, withoutfurther impact effects?' The NFA position is influenced in two ways."0

(1) A stock change (corresponding to an infinite inflow) equal to minus thechange in Bb: dBf= -dBb-

(2) A deterioration of the current account because of the change in the debtservice burden with i*dBf. See fig. 7 and app. A.3.

NFA

to t

Fig. 7

The long-run analysis is quite straightforward: as credit policy will beoffset immediately by capital in or outflows, it cannot influence credit marketconditions and therefore has no long-run effects on the level of production y.In the long run both i and p are fixed at their 'foreign' levels, so it becomesobvious from eqs. (25) and (13) that K and y will be unaffected in the longrun. In terms of the graphical apparatus of section 4, the K-locus is flat inthis case [cf fig. 8 and eq. (32). Note that all the partial derivatives of i arezero now]. What does change is disposable income. The increaseddebtservice actually decreases Yd: in conventional accounting terms, domesticproduct remains unaffected but national income falls. This reduces privatesaving and, in the long run, private financial wealth MR. So even with theextreme assumption of perfect capital mobility the traditional MonetaryApproach tenet of long-run neutrality of credit policy does not survive.While the capital stock and domestic production will remain unaffected by acredit crunch, real financial wealth and disposable (national) income will bereduced because of the increased burden of servicing the foreign debt. Aquantitative ceiling on foreign lending (imposed by the rest of the world?) orforeign interest rates that rise as total indebtness increases, would reintroducethe stagflationary effects discussed in sections 3 and 4.

9This is not entirely correct. Changes in foreign indebtness result in changes in' the debtservice i*Bf, which in tum changes disposable income. So not all transmission channels are cutoff by the perfect capital mobility assumption, This is taken into account in app. C and in thelong-run analysis in this section.

"'Note that the introduction of international capital flows and foreign debt leads to somemodification of the definitions of the current account and disposable income, both now include(with a minus sign) interest payments on foreign debt.

S. van Winbergen, Credit policy, inflation and growth 63

K /MR,

/ / MRlocus

MR

Fig. 8

6. Conclusions

In this paper we have attempted to give a theoretically rigorous butrealistic (at least for LDC's) analysis of the short- and long-run consequencesof credit policy for inflation, income, growth and external balance in a smallopen economy setting.

The model stresses differential speeds of adjustment in assets and goodsmarkets, contrary to most money and growth models, which typically use theneoclassical assumption of continuous equilibrium in all markets. This doesnot affect the long-run properties of the model, but provides it with morerealistic short-run behaviour.

Another feature of the analysis is the attention paid to the use that isactually made of commercial bank credit. It is an institutional fact in mostLDC's that commercial bank credit is used almost exclusively for businessloans to finance working (and, quite often, fixed capital) capital requirementsof firms. This results in a direct transmission mechanism between domesticcredit and production.

We show that this gives a possibly strong stagflationary bias to creditpolicy. Under conditions that are likely to be fulfilled in most LDC's, tightcredit may actually increase inflation initially, will always slow downproduction, but will probably improve the current account. If it does so,however, it is shown that the crucial expenditure adjustment comes from areduction in investment.

This brings us to the long-run consequences: a sustained policy of tightcredit is shown to reduce the long-run capital stock, and, under the sameconditions under which it leads to an increase in inflation in the short run,to reduce the long-run stock of financial wealth (in real terms) as well. If' thegovernment does not adjust its deficit in that case, this will result in adeterioration of the current account in the long run. If it does adjust itsdeficit via a cutback in expenditure (the only possibility in most LDC's), thecontractionary effects will be exacerbated.

64 S. van Winbergen, Credit policy, inflation and growth

Finally it is shown that a perfectly elastit supply of foreign 1lans (perfectcapital mobility) cures most but not all problems: obviously thestagflationary effects will disappeat, as thete is no need to induce the ptivatesector to hold less money and supply more loans via higher interest rates, otfor firas to cut back on production for lack of credit. It does howover leadto an increased debtservice burden and an accompanying decline indisposable income.

Appendix: List of the symnbels used

Ad real planned expenditure on domestic goods,AT real planned total expenditure,.6t real ceiling oh bank credit to the private sector,B1 real stock of loans from abroad,Bp real stock df loans from domestic non.bank sources (UMM),B. real demand for loans by firms,Cd real private consumption of domestic goods,cm real private consuimption of impotted goods,CT real total private consumption,Df real demand for working capital by firms,I real planned investment,1 real actual investment,i, i* domestic (foreign) nominal interest rate,K real fixed capital stock,m fraction of wealth held as money,M nominal money stock;MB monetary base (norninal),MR monetary base (real),NDCcgB net domestic credit from the central bank to the government

(noiminal),NFA net foreign assets (nominal),p, p*e price of domestic (foreign) goods,q terms of trade, q =p/(ep*),y real production of domestic goods,W real private wealth,w real wage rate,z fraction of total rationing that falls on foreign and domestic

consumers of domestic goods (so u-z) falls on investment,7r* foreign inflation rate,7E, nC actual (expected) inflation rate.

Hats (C) indicate percentage changes. Real variabies are obtained by deflatingthe relevant nominal variables with the price index of domestic goods, p.

S. vmn Wtnbgrgen, Credit policy, inflation and growth 65

lttfNte ces

garro, Robert, 1972, A theory of monopolistic price adjustment, Review of Economic Studies,Jan.

Birkhoff, Gprret mnd Gian-Carlp Vota, 1979, Ordinary differential equations (Wiley, New York).runp, Michael, 1979, Stabilization and stagflation in a semi-industrialized economy, in: Rudiger

Dornbusch and Jacob A. Ezenkel, eds., International economic policy, theory and evidence(Johns Hopkins University Press, Baltimore, MD).

Cavallo, Domingo F., 1977, Stagilationary effects of monetarist stabilization policies, Ph.D.dissertation, Harvard University, Cambridge, MA.

Dornbusch, Rudiger, 1975, A portfolio balance model of the open economy, Journal ofMonetary Economics.

Fischer, Stanley, 1972, Keynes-Wicksell and neoclassical models of money anu growth,American Economic Review, Dec.

Frenkel, Jacob A. and Rudiger Dornbusch, eds., 1979, International economic policy, theory andevidence (Johis Hopkins Press, Baltimore, MD).

Frenkel, Jacob A. and Harry G. Johnson, eds., 1976, The monetary approach to the balance ofpayments (University of Toronto Press, Toronto),

I.M.F., 1977, The monetary approach to the balance of payments.Jgfee, Dwight M. and Franco Modigliani, 1969, A theory and test of credit rationing, Amlican

Eponomic Review.Levhari, David atid Don Patinkin, 1968, The role of money in a simple growth model, American

Economic Review, Sept.McKinnon, Ronald I., 1973, Money and capital iti economic development (Brookings,

Washington, DC).Metzler, Lloyd, 1942, Underemployment equilibrium in international trade, Econometrica, April.Mundell, Robert A., 1968, International economics (Macmillan, New York).Polak, J.J., 1957, Monetary analysis of income formation and payments problems, I.M.F. St4f

papers, Nov. Reprinted in I.M.F.Reichman, Thomas and Thomas Stillson, 1978, Experience with stabilization programs

supported by standby arrangements in the upper credit tranches, Finance and Development.Kobichek, E. Walter, 1978, Financial progranmmning standby arrangements and stabilization

programs, Lecture Notes, I.M.F.Tobin, James, 1975, A general equilibrium approach to monetary theory, Journal of Monetary

Economics.Van Wijnbergen, Swedor, 1978, The offectiveqess of credit policy as an instrument to control the

balance of payments, with an application to Chile, processed, I.M.F.Van Wijnbergen, Sweder, 1982, Stagflationary effects of monetary stabilization policies: A

quantitative analysis of South Korea, Journal of Development Economics, April.Wai, U. Tun, 1957, Interest rates outside the, organized money markets of underdeveloped

countries, I.M.F. Staff papers, Nov.

No. 298. Ame Drud and Wafik M. Grais, "Macroeconomic Adjustment in Thailand: DemandManagement and Supply Conditions," Jounal of Policy Modeling

No. 299. Shujiro Urata, "Factor Inputs and Japanese Manufacturing Trade Structure," The Reviewof Economics and Statistics

No. 300. Dipak Mazumdar, "The Rural-Urban Wage Gap Migration and the Working of UrbanLabor Market: An Interpretation Based on a Study of the Workers of Bombay City,"Indian Economic Review

No. 301. Gershon Feder and Roger Slade, "Contact Farmer Selection and Extension Visits: TheTraining and Visit Extension System in Haryana, India," Quarterly Journal of Inferna-tional Agriculture

No. 302. James Hanson and Jaime de Melo, "The Uruguayan Experience with Liberalizationand Stabilization, 1974-1981," Journal of Interamerican Studies and World Affairs

No. 303. Nancy Birdsall and Dean T. Jamison, "Income and Other Factors Influencing Fertilityin China," Population and Development Review

No. 304. Graham Donaldson,"Food Security and the Role of the Grain Trade," American Journalof Agricultural Economics

No. 305. William F. Steel and Yasuoki Takagi, "Small Enterprise Development and theEmployment-Output Trade-Off," Oxford Economic Papers

No. 306. Oli Havrylyshyn and Engin Civan '"Intra-Industry Trade and the Stage of Develop-ment: A Regression Analysis of Industrial and Developing Countries," Intra-IndustryTrade: Empirical and MvIethodological Aspects

No. 307. Mateen Thobani, "A Nested Logit Model of Travel Mode to Work and AutoOwnership," Journal of Urblan Economics

No. -08. Johannes Bisschop and Alexander Meeraus, "On the Development of a GeneralAlgebraic Modeling System in a Strategic Planning Environment," Mathematical

No. 309. Reynaldo Martorell, Joanne Leslie, and Peter R. Moock, "Characteristics and Deter-minants of Child Nutritional Status in Nepal," The American Journal of Clitical Nutrition

No. 310. Robert 1H. Litzenberger and Jacques Rolfo, "An Intemational Study of Tax Effects onGovernment Bonds," Tle Journal of Finance

No. 3 11. Jere R. Behnnan and Nancy Birdsall," The Quality of Schooling: Quantity Alone is[\,Iisleadinrg," American Economic Revieuw

No. 312. Bela Balassa, "Adjustment Policies in Developing Countries: A Reassessment," WorldDeoelorpmlent

No. 313. Keith Marsden, "Services for Small Firms: The Roles of Government Programmes andMarket Networks in Thailand," International Labour Revieuw

No. 314. Gunter Schramm, "The Changing World of Natural Gas Utilization," Natural ResoutcesJournal

No. 315. Bela Balassa and Carol Balassa, "Industrial Protection in the Developed Countries,"The Vrld Econtomy

No. 316. Nancy Birdsall and Jere R. Behrman, "Does Geographical Aggregation CauseOverestimates of the Returns to Schooling?" Oxford Bulletin of Economics and Statistics

o. 317. Jan Peter Wogart and Jose Silverio Marques, "Trade Liberalization, Tariff Redundancy,and Inflation: A Methodological Exploration Applied to Argentina," Weliwirt-schaftliches Arch1iv

Issues of the World Bank Reprint Seiies are available free of charge fromthe address on the bottom of the back cover.