Controlling any monetary aggregate involves an ongoing process compared with discretionary policy decisions, such as changes in the exchange rate or in the trade and payments system. At the same time, the effects of a monetary disturbance on other variables are not instantaneous; the salient problem therefore in determining the appropriate rate of monetary expansion for the future is to take account of the lagged effects on these aggregates. However, the precise structure and length of the lags involved are in general not known, nor are they easy to determine. Moreover, although governments can control domestic credit of the banking system, experience in countries that resort to indirect controls—for example, discount rate policy, reserve and liquidity ratios, and open market operations—shows that credit objectives cannot be met with precision at every moment in time.1 Furthermore, depending on the individual country, external and internal credit markets outside the banking system may play a major role and influence aggregate demand and the balance of trade and payments.

Abstract

Controlling any monetary aggregate involves an ongoing process compared with discretionary policy decisions, such as changes in the exchange rate or in the trade and payments system. At the same time, the effects of a monetary disturbance on other variables are not instantaneous; the salient problem therefore in determining the appropriate rate of monetary expansion for the future is to take account of the lagged effects on these aggregates. However, the precise structure and length of the lags involved are in general not known, nor are they easy to determine. Moreover, although governments can control domestic credit of the banking system, experience in countries that resort to indirect controls—for example, discount rate policy, reserve and liquidity ratios, and open market operations—shows that credit objectives cannot be met with precision at every moment in time.1 Furthermore, depending on the individual country, external and internal credit markets outside the banking system may play a major role and influence aggregate demand and the balance of trade and payments.

Controlling any monetary aggregate involves an ongoing process compared with discretionary policy decisions, such as changes in the exchange rate or in the trade and payments system. At the same time, the effects of a monetary disturbance on other variables are not instantaneous; the salient problem therefore in determining the appropriate rate of monetary expansion for the future is to take account of the lagged effects on these aggregates. However, the precise structure and length of the lags involved are in general not known, nor are they easy to determine. Moreover, although governments can control domestic credit of the banking system, experience in countries that resort to indirect controls—for example, discount rate policy, reserve and liquidity ratios, and open market operations—shows that credit objectives cannot be met with precision at every moment in time.1 Furthermore, depending on the individual country, external and internal credit markets outside the banking system may play a major role and influence aggregate demand and the balance of trade and payments.

The paper addresses itself to the following questions:

(1) What is the appropriate monetary variable to be controlled?

(2) How can targets or limits for monetary variables be formulated to take account of the lagged effects on expenditures and the trade balance?

(3) What are the implications of short-run deviations of credit from the predetermined limits?

The choice of which monetary aggregate to control—that is, money or credit—depends on the policy objective and on the presence or absence of nonprice credit rationing. The appropriate variable for the control of expenditures, and hence for the balance of trade, is total (domestic and foreign) credit, while the overall balance of payments objective requires emphasis on domestic credit.

The paper shows that, owing to lags, the actual level of credit outstanding at each moment in time is of limited importance for changes in expenditures and trade as it provides only an indirect measure of monetary disequilibrium; the duration of the disequilibrium is of utmost importance. This is particularly true in countries with a wide range of financial assets, and hence a complex money supply process, and where developments of credit outside the institutional banking system, for example, the actions of other financial intermediaries, are important.

From the difference in the financial systems among countries, it follows that short-run fluctuations in domestic credit are of varied importance for the balance of trade; they are of utmost importance for the overall balance of payments if capital is highly mobile. As lag structures differ between countries, a credit disturbance lasting for, say, two months may be of minor consequence in one country and spell disaster in another. Some aspects of the determinants of the lag structure are discussed, and the importance of “credit rationing” (i.e., nonclearing markets) for the lag structure is analyzed.

Section I of the paper sets out the problems in greater detail; Section II discusses a simple model; and Section III discusses simulation experiments, which highlight the importance of time lags.

I. The Problems

At the present, most economists agree that money and credit are key factors in the determination of expenditures and the balances of trade and payments.2 The ongoing discussion between “monetarists” and other members of the profession focuses, with some notable exceptions, on the degree to which “money matters,” on the appropriate specification of the transmission mechanism, and how to take account of financial intermediation outside the banking system. There appears also to be a consensus that movements in money and credit matter only if these changes are sustained for at least some period, or are extremely large, although this element is usually not explicitly treated in the static framework of the models. The relative insensitivity of gross national product to short-run fluctuations in the growth of money and credit is explained by the lags specified and observed in the relationship between monetary variables, expenditures, and income. Thus the longer and more stable the lags are, the less important are short-run changes in money and credit. Most empirical investigations found that it takes well over one year for monetary disturbances to have their full impact on the economy—although the lags are highly variable. This means that economic activity at any point in time is influenced by all the changes in the monetary variables that took place over the preceding year or two. If lags are long, the importance of what happens to credit and the money supply during any particular week or month (and perhaps even during any particular quarter) is therefore relatively small, as far as the development of expenditures and income is concerned, unless the change is sustained in subsequent periods. On the other hand, as financial capital is in many cases mobile internationally (within certain limits), one might expect that short-run fluctuations of credit would be of greater importance as far as capital flows, and hence the balance of payments, are concerned.

Although reference was made earlier to “money and credit,” the investigation will show that it is necessary in an open economy to differentiate carefully between the effects of changes in (domestic) credit and changes in the money supply. In a closed economy, expansion (contraction) of credit by the consolidated banking system necessarily implies a matching increase (decrease) in the quantity of money, broadly defined.3 As creation and destruction of money go hand in hand with the expansion and contraction of credit, it remains a matter of personal preference if one stresses the importance of money (Friedman, Schwartz) or the importance of credit (Robinson, Kaldor).

In an open economy, the accounting identity requiring that changes in money (broadly defined) must equal changes in domestic credit of the banking system no longer applies. Reserve movements that reflect trade and capital account balances have a direct effect on the money supply.4 Pursuing a money supply target compared with a credit target in this circumstance will have totally different implications. The basic findings of the model are that the pursuit of a target on total credit (domestic credit and foreign borrowing) is important as far as expenditures and the trade balance are concerned, while the domestic credit target becomes important for capital movements and the overall balance of payments objectives. The model and its findings are in line with the so-called monetary approach to the balance of payments, which, despite its name, provides a “credit” theory of the balance of trade and payments, as the growth of domestic credit (not of money) determines the balance of payments outturn in these models.5

The paper is concerned with determining when the “short run” is long enough, and a fluctuation of money and credit large enough, to matter for movements of the balance of trade and payments. The importance of domestic credit for reserve movements has long been recognized in principle; the paper will shed some light on the possibilities of incorporating time lags into the formulation of credit targets.

II. The Model

This section analyzes the effects of monetary policy on the balance of trade and payments within the framework of a simple model. It assumes that income remains unchanged in the short run and that deviations of expenditures from income (which are necessarily equal to the trade balance), as well as capital flows, reflect money stock disequilibrium. Changes in credit of the banking system—to the extent that they are not offset by induced capital flows—change the money supply, and this in turn influences expenditure behavior via changes in the interest rate and the change in financial wealth. Moreover, in the case where nonprice credit rationing is assumed, a direct expenditure effect is also present. In contrast to other models, it is not true that exports are fixed and imports are proportional to income.6 On the contrary, as the trade balance equals the difference between income and expenditure, any increase in income would improve the balance of trade for a given level of expenditure. Moreover, a rise in nominal income would increase the demand for money and hence reduce expenditure. However, as the concern is with the short run, it is assumed that nominal income remains unchanged—any change in expenditures will therefore have an equal effect on the trade balance. The justification for this is twofold: income adjusts to changes in expenditures with an often considerable lag, and in small open economies prices are to some extent exogenously determined—at least in the longer run. This simplifying assumption is, however, immaterial for the qualitative results derived: If income would respond to changes in demand, all results would remain basically unchanged as long as 0 < ∂Y/∂E < 1.

As the development of domestic credit affects the stock excess demand or supply of money, it also determines the balance of trade and capital movements, and hence the balance of payments outturn. It will be shown that in the absence of capital mobility it is the average level of domestic credit outstanding over a period that determines the short-run balance of trade and payments outturn.

The model presented here provides an analytical presentation; as it was not the purpose to develop a new analytical framework but to ask new questions within a familiar setting, the present model makes no claim to originality and displays certain similarity to the well-known models specified by Dornbusch (1975) and others. The following notation is used:

article image

identities

QE=BT(1)
BTiC=CAorYE=CA(2)
CA+C˙=R˙(3)
Ms=DC+R(4a)
Y=QiC(4b)
B˙=(DC˙+C˙)(4c)

budget constraints

(flow)E+W˙=Y(5)
(stock)M+B=W(6)

Equations (1) and (2) state that the balance of trade necessarily equals the difference between domestic output and expenditures (absorption) and that the current account equals the difference between income and absorption. If the level of absorption increases, ceteris paribus, the trade balance and the current account will deteriorate. It is thereby irrelevant to what extent a change in expenditures leads to increased imports or to an increase in domestic “absorption” of output that would otherwise have been exported.7 While equations (2), (3), and (4a) are self-explanatory, equation (4b) states that domestic income equals domestic output minus the interest payments on the countries’ foreign debt. If C < 0 (i.e., if the country is a creditor), then income will exceed the value of domestic output. Equation (4c) shows that an increase in the supply of bonds can come about only by a reduction of bank credit, that is, open market sales of bonds by the consolidated banking system, or by a purchase of foreign bonds, that is, through a capital outflow.

There are both flow and stock constraints operating in the model. Equation (5) states that expenditures plus financial wealth accumulation must equal income, while equation (6) shows that money balances plus bond holdings (positive or negative) cannot exceed net wealth. Differentiating equation (6) shows that the sum of changes in money balances and bond holdings must equal the change in wealth:

M˙+B˙=W˙(6a)

behavioral equations

The model contains behavioral relationships relating to flows as well as to stocks. These two aspects are linked, since stocks are changed over time by flows. It is assumed that the long-run demand for assets (i.e., wealth) is an increasing function of the level of income, and of the rate of interest.

Wd=Wd(Y,i)Wyd,Wid>0(7)

However, because the adjustment of wealth holdings is costly and the marginal cost of adjusting increases with the speed of adjustment,8 stock equilibrium is not achieved instantaneously. Saving and dissaving take place at finite rates; it is assumed that the accumulation of wealth is proportional to the difference between actual wealth, W¯, and desired wealth holdings, Wd.

W˙=Φ(WdW¯)0<Φ<1(8)

Since flows are finite, W¯ is given at any moment in time. As it follows from equations (5) and (2) that the current account surplus must exactly equal the flow of savings,9 an increase in the stock demand for assets will improve the current account.

The desired portfolio balance between bonds (debt) and money is assumed to depend on the yield of bonds.

MdBd=F(i)(9)

Fi < 0

This formulation assumes that the income elasticity of demand for both assets is equal to one.10 In an open economy with capital flows and a fixed exchange rate system supported by intervention of the authorities, portfolio balance can be achieved instantaneously by, for example, selling bonds abroad (i.e., foreign borrowing) in exchange for foreign currency, which can then be converted into domestic money, or by purchasing (foreign) bonds abroad (i.e., foreign lending), if a reduction in money balances and an increase in bonds is considered. However, these changes in the composition of the portfolio cannot affect the level of wealth. As discussed earlier, the latter can be changed only through (finite) accumulation or decumulation.

From equations (7)-(9) it is possible to derive demand for bonds and money functions:

Bd=Bd(i,Y)(10)Bid,Byd>0
Md=Md(i,Y)(11)Mid<0,Myd>0

The excess demand for bonds, B˜d, can then be written as a function of i, Y, and—because of equation (9)—of the quantity of money.

B˜d=B˜d(i,Y,M8)B˜id,B˜yd,B˜M8d>0

open market operations

Consider first the effects of an open market operation on perfect capital immobility.11 As there are three assets in the model—real output, money, and debt (bonds)—the analysis can be conducted in any one of the three markets (which are linked by Walras’s law) depending on whether one prefers the formulation of a real (absorption), monetary, or credit theory of the balance of trade and payments. We concentrate on the latter aspect.

Turning to the bond market, an open market purchase of bonds by the banking system (i.e., an increase in domestic credit) will result in an excess demand for bonds at the prevailing interest rate (Chart 1). At constant interest rates, this excess demand for bonds must be matched by a corresponding excess supply of money, as the demand and supply in the market for real output do not change unless movements in the interest rate change absorptions via the wealth effects on expenditures (equations (7) and (8)). In the absence of capital mobility (i.e., the possibility of acquiring and selling debt (bonds) abroad), there can be no net purchase or sale of bonds or money by the private sector in the aggregate ( = 0). The interest rate has to decline to clear the markets, that is, to eliminate the excess demand for bonds.

Chart 1.
Chart 1.

Bond Market: The Effects of an Open Market Purchase of Bonds

Citation: IMF Staff Papers 1977, 001; 10.5089/9781451956443.024.A006

With the decline in the interest rate, the desired level of wealth, and hence of accumulation, decreases (equations (7) and (8)). Consequently, expenditures will increase and, by equations (1) and (3), balances of trade and payments will deteriorate as a result of the expansion of credit and money. In the short run, the absence of capital mobility permits the authorities to control the money supply.

What happens over time? As the money supply is reduced via the trade and payments deficits, the excess demand for bond function in Chart 1 will gradually shift upward, that is, the interest rate will increase and thus result in a reduction of the trade and payments account deficits until these are completely eliminated. The consequences of an open market sale follow from the symmetry of the model.

In the absence of capital movements, changes in domestic credit initially bring equal changes in the money supply and have their effect on the balance of payments via the change in expenditures (i.e., absorption). The importance of the lagged effects of monetary (credit) policy on expenditures and trade is the dominant feature. The precise structure of the lag will depend on a variety of institutional features that determine the speed of adjustment.

Consider now the case of a small open economy where capital is perfectly mobile. Since these assumptions imply that the yield on bonds (interest rate) is fixed and exogenous to the country, the desired ratio of money balances to bond holdings is predetermined. Any attempt by the authorities to increase the domestic money stock through open market purchases of bonds will only result in an equal purchase of foreign bonds by the private sector, that is, the stock of domestically held debt and money will remain unchanged. As the yield on bonds will not have changed, desired wealth, and hence savings and expenditures (absorption), will remain unaffected by the open market policy. Hence, the trade account will remain unchanged.12 This does not, however, imply that an expansionary monetary policy has no effect on the balance of payments. The capital outflow (i.e., purchase of foreign bonds) brought with it an equal loss of foreign reserves. Moreover, the banking system acquired interest income through the purchase of bonds; this flow will be exactly equal to the change in foreign interest payments received, that is, to the improvement in the current account.13 Again, the implications of an open market sale of bonds follow from the symmetry of the model.

This analysis leads to the conclusion found, for example, in Mundell’s “assignment” problem (Mundell (1968)) and in Dornbusch (1975): in a world where capital is perfectly mobile, monetary policy (i.e., domestic credit expansion or contraction) has no influence on domestic activity, but has to be assigned to the achievement of balance of payments, that is, reserve targets. Monetary expansion (contraction) will be offset by (nearly) instantaneous capital flows. The effect on foreign reserves of such a policy will be felt (almost) immediately, that is, with no time lag.

Finally, there exists the intermediate case in which capital is mobile but only to a certain degree.14 Consider a reduction in domestic credit. Firms will be driven to borrow abroad; however, institutional limitations will prevent the complete offset of the effects of the change in domestic credit on the domestic money supply (i.e., there will be some space for an independent monetary policy in the short term provided that the authorities are prepared to accumulate or decumulate the necessary foreign reserves). The change in the money supply and the simultaneous change in the interest rate will lead to a change in expenditures, and hence will have an effect on the trade balance. These effects will not be proportional to the size of the change in domestic credit but to the sum of the change in domestic plus foreign credit, that is, to the initial change in the money supply. The capital account responds, however, to the availability of domestic credit.

credit versus money supply

Credit theorists—although in general agreement with the monetarists that an expansion in the money supply will lower the rate of interest and increase expenditures—claim that the composition of credit itself (i.e., the source of the increase in the money supply) is of importance for the effect on expenditures and income and, by implication, on the balance of trade. In particular, they stress that an expansion of the money supply caused by an open market purchase of existing securities will be less expansionary than an expansion of equal size originating from the purchase of new obligations or from the extension of direct loans to the private sector.15 Moreover, expenditures of firms may be financed through credits outside the banking system—for example, by the floating of bonds and equity in the private market. This argument is based on the assumption that households consider money and existing securities as close substitutes and that there exists a nonprice rationing of credit by the banking system.16 Consequently, the availability of new credit (rather than interest rates—that is, the cost of credit) and the money supply determine expenditures.

Monetarists counter this argument by contending that there exists substitutability not only between money, bonds, and other financial assets but also between money and real (existing and new) assets.17 Therefore, any addition to the money supply will have strong and direct effects on expenditures, that is, on absorption. Moreover, some monetarists stress that the propensity to spend out of an increase in the money supply is independent from how this change was brought about, as it is not so much the effect on the cost of credit but the portfolio adjustment to the desired level of money balances that determines spending.18

The argument is therefore one of certain empirical facts, such as the elasticity of substitution between various asset markets and the existence of nonprice credit rationing. Although the model described earlier was not designed to analyze this particular set of questions in detail, it can be modified to shed some light on the implications in an open economy of nonprice credit rationing and of the various assumptions regarding the substitutability between money and other assets.19 It is useful for expository reasons to impose the most extreme set of assumptions of each school of thought on the model, although their main proponents generally hold a more balanced view. For the specification of the model along the lines of the credit theorists, this means that money and bonds are perfect substitutes, and that firms have to depend on direct loans from the banking system for their financing of additional expenditures, as they are unable to issue bonds in the private market.20 The monetarist specification requires that there be no, or limited, room for substitution between money and other financial assets and that firms can obtain financing by selling debt in the private market.

It is necessary to modify the model in the following way. Introducing into the wealth equation direct loans from the consolidated banking system to the private sector yields

W¯=Ms+BsKR

where KR denotes loans.

If the availability of loans is the limiting factor for expenditures and it is assumed that the total amount of any increase in loans will be spent, changes in expenditures can be written as follows:

dE=dKR+Φ(WsW¯)

This can then be called the credit version of the expenditure function,21 while, under monetarist assumptions, the expenditure function remains unchanged

dE=Φ(WsW¯)

The effects of monetary expansion through an open market purchase of bonds and a direct loan from the banking system will be contrasted under both sets of assumptions.

Consider first an open market purchase of existing bonds under credit theorists’ assumptions. If money and bonds are perfect substitutes, the interest rate (price of bonds) will not change and, consequently, there will be no effect on desired wealth. Moreover, as dMs = -dBs, only the composition of actual wealth, but not its level, has changed, and there will be no effect on real expenditures from an increase in the money supply through open market operations.22

Contrast this with the effects of an increase in direct loans to enterprises. It follows from the specification of the expenditure function that the availability of credit will induce additional expenditures by the same amount: dE = dKR. As the increased money supply will not cause a change in the rate of interest, owing to the substitutability assumption, and as the increase in liabilities must exactly offset the wealth effect of an expansion in the money supply, there are no further repercussions.

Returning to the monetarist specification, if firms consider bank loans and financing through floating of bonds as (perfect) substitutes, the effects of a direct loan by the banking system must be equivalent to firms selling bonds to other members of the private sector and a simultaneous purchase of an equivalent amount of bonds through an open market operation. As the first action leaves both the bond and money supply of the private sector unchanged, only the effects of an open market purchase remain. Clearly, given these specifications, direct loans and open market purchases have the same effect on expenditures.

As in the credit theory specification of the model, the availability of credit limits expenditures, the initial change in expenditures is equal to the total increase in direct loans. Hence, direct loans have a more powerful expansionary effect than do open market operations, even if the assumption of perfect substitutability between bonds and money were to be relaxed.

Although it remains an empirical issue as to which specification is the most appropriate for a particular country, some inferences for the differences in lag structures between countries can be drawn. The analysis showed that the crucial aspects are the degree of substitutability between money and other financial assets and the presence of nonprice (i.e., non-interest rate) credit rationing. It appears that in general the presence of nonprice credit rationing is more characteristic of less developed countries (LDCs), and the limited diversification of financial assets in these countries makes a high degree of substitutability between money and other financial assets unlikely. The simpler structure of financial assets in LDCs implies that a change in credit will not be diffused among various money substitutes but will be transmitted directly to the market for commodities. As stated by Park (1973, p. 395) and others, monetary expansion will more likely “impinge directly on real expenditure in LDCs than in advanced countries.” This implies that there will be a relatively strong impact on expenditures. Moreover, monetary expansion in LDCs will more often reflect direct loans to the government or private sector than open market operations. Hence, there is a presumption that the immediate impact of monetary expansion is stronger, and hence the lag structure shorter, in LDCs than in highly developed countries.

Finally, returning to the original specification of the model, we show that even under monetarist specifications the origin of changes in the money supply may be important to the extent that financial wealth is affected. Compare a situation where the government adds directly to the money supply—either by distributing newly created money via helicopter or by transfer payments—with an open market purchase of bonds. In the first case, financial wealth increases to the extent of the change in the money supply, while in the second case only the composition of financial assets changes. As financial wealth enters into the expenditure function, it follows that the first method of increasing the money supply will have much more powerful effects on expenditures than the second one.

credit controls and financial disintermediation

As was pointed out earlier, in certain countries, especially LDCs, it is often the availability of credit rather than the cost of credit that limits expenditures by firms on intermediate inputs, investment goods, imports, etc. Expenditure behavior will depend on the overall credit extended to firms by the banking system (e.g., direct loans and security purchases), households, and financial intermediaries other than banks, for example, insurance companies and pension funds.

Consider, for example, the effects of a ceiling that requires a reduction in bank credit. The banking system is forced to reduce its outstanding claims on private firms. Rather than hold excess reserves, private banks will try to adjust their liabilities downward by reducing, for example, time and savings deposits of households through the lowering of deposit rates. The crucial point is that if households find corporate bonds and equity close substitutes for demand and time deposits, this will result in an expansion of direct credit from households to firms via the bond and stock markets. The effect of the ceiling on credit by the banking system is then—at least partly—offset by credit extension outside the banking system. This means, however, that the credit ceiling on banks will be of only limited importance—in such circumstances—for the development of private sector demand, and hence for the balance of trade and payments. The extent of financial intermediation outside the banking sector varies between countries; as private capital markets are usually not well developed in LDCs, one would expect that substitution of other forms of credit for bank credit would be relatively less important than in highly developed economies. Moreover, monetarists claim—and the observed relative stability of money demand function appears to confirm their view—that money serves entirely different purposes than do holdings of corporate bonds and shares. Therefore, one would not expect that households—in general—consider bonds and shares sufficiently close substitutes for their holdings of time and savings deposits.

It remains, therefore, an empirical question as to what extent private credit outside the banking system is able to offset the effects on expenditures, and hence on trade and payments, of a ceiling on credit by the banking system.

the implications of short-run disturbances in credit

Based on the model, an expansionary domestic credit policy will lead to a deterioration of the balance of payments, and a tightening of credit to an improvement. For the case where capital is highly mobile, the instantaneous effect on net foreign assets will be large and will come via the capital account as portfolio adjustments take place. If the movements of capital are restricted, credit contraction or expansion will lead to a change in expenditures (absorption) and will result in reserve movements via changes in the trade balance. This effect will be more pronounced in the presence of nonprice credit rationing. Although it can be shown that, given the assumptions of the model, the overall long-term effect on reserves of a credit expansion or contraction will be the same in the presence or absence of capital mobility, these differences in the lag effects of credit policy on expenditures and trade balance are of importance in the short run.

What are the implications for the reserve movements and the balance of trade of a short-term disturbance of credit, on the one hand, and a disturbance that is sustained for some time, on the other hand? Consider, as an example of the first case, an open market sale of bonds with matching repurchase after two to three months.

With perfect capital mobility, the reduction in bank credit implied by the bond sale is completely offset through capital inflows that take place (basically) instantaneously. The quantity of money remains unchanged, and there are no effects on absorption and hence on the trade account. Similarly, the purchase of bonds after a period of two to three months will lead to a capital outflow and matching loss of reserves. If the sale takes place six months later rather than after two months, nothing but the timing of the reserve loss will have changed.

The conclusion of this analysis is that, even in the short run, credit fluctuations matter very much for reserve movements, if capital is mobile. They are, however, of less importance for domestic absorption and hence for the trade account. Hence, to control domestic expenditures (absorption)23 and the trade balance, it would be necessary to control not only domestic credit but also the sum of changes in domestic credit and of (short-term) capital flows.24

Turning to the importance of short-term fluctuation in domestic credit in the absence of capital movements, one notes that the effect on reserves comes only via movements in the trade account that are brought about through changes in domestic expenditures. Analyzing the situation from the expenditure side, rather than the bond market, a tightening of credit leads to a reduction in expenditures and hence to an improvement of the trade balance. The extent of the resulting movements in the trade and payments account depends crucially on the speed of adjustment, Φ. However, the short-run effect will be relatively minor compared with the case where capital is mobile. If the initial tightening of credit is, for example, offset after two to three months through a bond purchase, the cumulative trade surplus and the resulting reserve movements over this time will probably have been small. Only if the reduction in credit is allowed to persist for a longer time will there be a substantial cumulative effect on reserves and trade. This can be better seen by considering the following. Using equations (1), (7), and (8), changes in the trade balance can be written as a function of changes in credit.

BT=gDC˙

As the adjustment speed is finite,25 it follows that:

0 < g < 1

Consider now relatively short periods (e.g., months), and assume that credit is increased by DC in the first period. The balance of trade and payments deficit during the first period then amounts to

BT1=gDC˙1

Assume now that domestic credit is permitted to stay at this higher level for another period. However, at the onset of the second period, the excess supply of liquidity in the economy has already been reduced to some extent by the deficit in the first period, so that

BT2=g(DC˙1BT1)

or,

BT2=gDC˙1g2DC˙1

For short periods, the term g2DC˙1 becomes negligible,26 so that the accumulated deficit over the two periods can be approximated by

(BT1+BT2)=2gDC˙1

that is, the balance of trade deficit is proportional to the length of the disturbance.27

It follows that in cases where an increase in credit is compensated for in the following period by a decrease of the credit level below the long-run equilibrium, that is,

DC˙1=DC˙2

that the deficit in the first period will be matched by a corresponding surplus in the second period

(BT1+BT2)=gDC˙1+gDC˙2=0

Since the average credit disturbances balance, there is no net effect on the balance of trade over the two periods. An excessive expansion of credit lasting for one month or even one quarter will, in the absence of capital flows, generate only minor reserve losses. Only sustained expansion of credit should therefore be of concern to the policymakers, unless the short-term disturbance is of sufficient size as to result in irreversible changes in wages, prices, and expectations. If the latter is true, the lagged effects of credit would be autocorrelated, and it is no longer possible to simply average past credit levels.

We therefore assume explicitly for the simulation study that the lagged effects of credit are not autocorrelated—that is, that past credit changes were within the tolerance limits so that the structure of the economy remained unaffected. It would not be possible otherwise to use the weighted average of lagged credit expansion as a measure of monetary events set in motion.

III. A Simulation Experiment

Any empirical investigation of what constitutes the crucial length of a credit disturbance for foreign reserve movements and the balance of trade must focus on two aspects: (1) the degree to which capital can move in the short run to offset the effects of changes in domestic credit on expenditures, and (2) the time lag between expansion or contraction in domestic credit (net of offsetting capital flows) and changes in expenditures (absorption).

Rather than engage in the cross-country analysis of the differences in lag structures and in capital mobility, we conducted a simulation experiment. Using the findings of Kouri (1975), with respect to offsetting capital flows, and the lag structure derived from our own regression analysis, we simulated the effects of credit disturbances of various lengths on trade and payments balances. The two countries chosen, the Federal Republic of Germany and Korea, are examples of open economies with highly different degrees of development, financial structure, and capital mobility.

It was assumed that changes in imports adequately reflected changes in net absorption; exports were considered to be exogenous. Instead of estimating the structural equations of a simultaneous equation system or following Brunner’s approach (Brunner (1972 b)) in determining the elasticities relevant for the nature and characteristics of the transmission process in a multimarket model, we used a reduced form equation linking changes in imports directly to exogenous changes in credit (Table 1). Such specification was analyzed previously by Polak, Fleming, and Rhomberg and, in a different context, also in the St. Louis equation.

Table 1.

Korea and the Federal Republic of Germany: Regression Results1

article image

The results are based on Almon lags using polynomials of degree 3. Im = imports; Q = domestic credit + exports + capital inflow (net).

Although it is usually more desirable to estimate a simultaneous equation system, such a system cannot easily accommodate any lag structure beyond a Koyck lag. Moreover, the comparison of the regression results presented here with those of various other studies shows a surprising consistency. To be more specific, the results for the Federal Republic of Germany indicate that an expansion of credit has very little immediate effect on absorption (imports) relative to credit lagged two to five quarters. Thereafter, the importance of lagged credit declines sharply, and lagged credit has even a marginal negative influence on imports. These results are consistent with the findings of Knöbl and Solheim (1976), who estimated and simulated an eight-equation model for the Federal Republic of Germany. The slightly negative influence of sufficiently lagged monetary expansion in their model is explained explicitly by the adjustment in long-term interest rates. Moreover, the Frowen and Arestis (1976) results regarding the demand for money in the Federal Republic of Germany are also broadly consistent with our findings concerning the length of lags between monetary expansion and absorption. Turning to the results for Korea, the findings seem to confirm the hypothesis that monetary expansion has indeed a much more direct impact on imports in LDCs. The main effect comes in the first quarter and sufficiently lagged expansion in credit has a negative effect on imports. These findings are consistent with those of Otani and Park (1976), who used a multiequation model. Over all, the single-equation approach appears to give satisfying results and good insight into the structure of time lags.

the importance of capital flows

Using the lag structure derived for the Federal Republic of Germany and Kouri’s (1975) findings that capital flows offset about 70 per cent of the effect of credit expansion and contraction, the movements of the trade and capital account were simulated, using the import equation, for a short-term expansion of domestic credit in the first quarter, which is reversed in the second quarter. Chart 2 shows that the offsetting capital movements28 are the dominating factors with regard to reserve changes, while relatively minor movements of the trade balance linger for eight quarters. It confirms that in a situation where capital is mobile, the control of credit is of utmost importance for the overall balance of payments even in the very short run. On the other hand, as capital flows offset a large part of the effects of a credit expansion on the money supply and hence in absorption, the small changes in the trade balance are not unexpected (Chart 2, “trade balance I”). The trade fluctuations are appreciably larger in the second case, where, in the (assumed) absence of capital mobility, domestic credit expansion leads initially to larger changes in expenditures. Even so, the swings in reserves that are (approximately) proportional to the integral under the curve labeled “trade balance II” are still small relative to the overall reserve loss that occurs if capital mobility is present. It is evident from the chart that in the absence of capital flows a disturbance in credit that lasts for only a quarter or so exercises only a marginal influence on the trade balance and on reserves. However, this conclusion is based on the specific lag structure of the country in question. The shorter the time lag between monetary expansion and absorption, the more will short-run fluctuations in liquidity matter for the balance of trade.

Chart 2.
Chart 2.

Federal Republic of Germany: Effects of a Transitory Addition to Domestic Credit in the Presence and the Absence of Capital Mobility1

Citation: IMF Staff Papers 1977, 001; 10.5089/9781451956443.024.A006

1 It was assumed that credit was increased for the duration of one quarter. The scale measures the balances as a percentage of the disturbance.2 Presence of capital mobility.3 Absence of capital mobility.

To explore in more detail the significance of the duration of a disturbance in credit,29 the following experiment was conducted, which highlights the importance of the lag structure. Using the regression results for the Federal Republic of Germany and Korea, summarized in Table 1, the effects of an increase in the rate of liquidity expansion were analyzed. Charts 3 and 4 show the implications for the rates of import growth of uncompensated disturbances in the rate of credit expansion persisting for one, two, and three quarters, respectively. Although the final outcome must be the same, the charts show that for Korea the shorter lags between monetary expansion and changes in absorption led to a substantial trade deficit within a quarter or two,30 while for the Federal Republic of Germany the longer lag structure meant that the immediate effect of credit expansion on imports was relatively small.

Chart 3.
Chart 3.

Federal Republic of Germany: Deviation of Import Growth Rates from Trend1

Citation: IMF Staff Papers 1977, 001; 10.5089/9781451956443.024.A006

1 A rate of credit expansion of 4 percentage points above trend was sustained for one, two, and three quarters, respectively.
Chart 4.
Chart 4.

Korea: Deviation of Import Growth Rates from Trend1

Citation: IMF Staff Papers 1977, 001; 10.5089/9781451956443.024.A006

1 A rate of credit expansion of 4 percentage points above trend was sustained for one, two, and three quarters, respectively.

The importance of the differences in the lag structures between economies becomes even more evident when the case is considered where the authorities compensate in the subsequent quarter for the excessive growth. For the results shown in Chart 5, it was assumed that the faster growth of credit was compensated for by slower rates of credit expansion in subsequent periods, so that at the end the same level of credit existed that would have prevailed in the absence of the disturbances. Although in both cases the sum of deficits and surpluses equals zero, the (short-run) fluctuations in reserves were more pronounced for the shorter lag structure (Korea). (See Chart 5.)31

Chart 5.
Chart 5.

Korea and the Federal Republic of Germany: Deviation of Import Growth Rates from Trend1

Citation: IMF Staff Papers 1977, 001; 10.5089/9781451956443.024.A006

1 An increase in the rate of credit expansion of 8 percentage points above the trend rate for one quarter was compensated for by a matching reduction in credit growth in the next quarter. Thereafter, credit was assumed to resume its trend growth.

conclusions

The simulation study permits the following conclusions:

If capital is highly mobile, then it is the change in domestic credit net of offsetting capital flows that determines changes in expenditures, income, and the balance of trade; movements in domestic credit will determine the direction of capital flows, which in turn will dominate the overall movement of reserves, that is, the balance of payments outturn. Limits on domestic credit are appropriate if the objective is an improvement of the overall balance of payments; limitation of domestic plus foreign credit is required if price stability and the trade balance are also policy objectives.32

If capital mobility is restricted, by policy or through the lack of creditworthiness and access to international capital markets, the effects of changes in domestic credit on reserves are slower; however, the time lag between changes in credit and absorption may vary substantially between countries, depending on the structure of the financial markets.

The starting point of any deliberation in the formulation of credit policy must be the analysis of past credit and monetary expansion.33 It is important to arrive at some judgment, based on a formal model or other analysis regarding the structure and length of the lag between monetary expansion, expenditures, imports, and income. While it may be necessary in some countries to include in the analysis movements in the monetary aggregates as far back as two years, for many countries a much shorter time horizon (of six months to one year) may be adequate. With regard to the base period, averaging past levels of credit may provide a better base than commencing with the latest month for which data are available, as random factors may have distorted these data; moreover, the latest available statistics are often subject to further revisions.

Having gained a picture of the monetary forces set in motion through past policies and of their likely impact in the near future (i.e., of the prevailing monetary disequilibrium), the scope for monetary policy action must be assessed, bearing in mind that the impact of the new policy measures on various aggregates will be felt only with a certain lag.

BIBLIOGRAPHY

  • Andersen, Leonall C., and Jerry L. Jordan,Monetary and Fiscal Actions: A Test of Their Relative Importance in Economic Stabilization,Federal Reserve Bank of St. Louis, Review, Vol. 50 (November 1968), pp. 1123.

    • Search Google Scholar
    • Export Citation
  • Brunner, Karl (1970), “Eine Neuformulierung der Quantitätstheorie des Geldes,Kredit und Kapital, Vol. 3 (No. 1, 1970), pp. 130.

    • Search Google Scholar
    • Export Citation
  • Brunner, Karl (1976), “Issues of Post-Keynesian Monetary Analysis: A Contribution to the Discussion Opened by Professor Thomas Mayer,Kredit und Kapital, Vol. 9 (No. 1, 1976), pp. 2455.

    • Search Google Scholar
    • Export Citation
  • Brunner, Karl and Allan H. Meltzer (1968), “Liquidity Traps for Money, Bank Credit, and Interest Rates,Journal of Political Economy, Vol. 76 (January/February 1968), pp. 137.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brunner, Karl and Allan H. Meltzer (1972 a), “Money, Debt, and Economic Activity,Journal of Political Economy, Vol. 80 (September/October 1972), pp. 95177.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Brunner, Karl and Allan H. Meltzer (1972 b), “A Monetarist Framework for Aggregative Analysis,” in Proceedings of the First Konstanzer Seminar on Monetary Theory and Monetary Policy, issued as a supplement to Kredit und Kapital (Berlin, 1972), pp. 3188.

    • Search Google Scholar
    • Export Citation
  • Chow, Gregory C.,On the Long-Run and Short-Run Demand for Money,Journal of Political Economy, Vol. 74 (April 1966), pp. 11131.

  • Corrigan, E. Gerald,Income Stabilization and Short-Run Variability in Money,Federal Reserve Bank of New York, Monthly Review, Vol. 55 (April 1973), pp. 8798.

    • Search Google Scholar
    • Export Citation
  • Davis, Richard G. (1968), “The Role of the Money Supply in Business Cycles,Federal Reserve Bank of New York, Monthly Review, Vol. 55 (April 1968), pp. 6373.

    • Search Google Scholar
    • Export Citation
  • Davis, Richard G. (1970), “Interpreting the Monetary Indicators,Federal Reserve Bank of New York, Monthly Review, Vol. 52 (July 1970), pp. 15962.

    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger (1973), “Devaluation, Money, and Nontraded Goods,American Economic Review, Vol. 63 (December 1973), pp. 87180.

    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger (1975), “A Portfolio Balance Model of the Open Economy,Journal of Monetary Economics, Vol. 1 (January 1975), pp. 320.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Friedman, Milton (1968), “The Role of Monetary Policy,American Economic Review, Vol. 58 (March 1968), pp. 117.

  • Friedman, Milton (1972), “Comments on the Critics,Journal of Political Economy, Vol. 80 (September/October 1972), pp. 90650.

  • Friedman, Milton and David Meiselman,The Relative Stability of Monetary Velocity and the Investment Multiplier in the United States, 1897–1958,” in Stabilization Policies, Commission on Money and Credit (Englewood Cliffs, N.J., 1963), pp. 165268.

    • Search Google Scholar
    • Export Citation
  • Friedman, Milton, and Anna Jacobson Schwartz, A Monetary History of the United States, 1867–1960, National Bureau of Economic Research (Princeton University Press, 1963).

    • Search Google Scholar
    • Export Citation
  • Frowen, Stephen F., and Philip Arestis,Some Investigations of Demand and Supply Functions for Money in the Federal Republic of Germany, 1965–1974,Weltwirtschaftliches Archiv, Vol. 112 (No. 1, 1976), pp. 13664.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jianakoplos, Nancy,The FOMC in 1975: Announcing Monetary Targets,Federal Reserve Bank of St. Louis, Review, Vol. 58 (March 1976), pp. 822.

    • Search Google Scholar
    • Export Citation
  • Johnson, Harry G., Further Essays in Monetary Economics (Harvard University Press, 1973).

  • Kaldor, Nicholas,The New Monetarism,Lloyds Bank Review, No. 97 (July 1970), pp. 118.

  • Knöbl, Adalbert, and Jon Solheim,Effects of Fiscal and Monetary Policies in the Federal Republic of Germany—A Dynamic Short-Term Model of the German Economy” (unpublished, International Monetary Fund, March 16, 1976).

    • Search Google Scholar
    • Export Citation
  • Kouri, Pentti J. K.,The Hypothesis of Offsetting Capital Flows: A Case Study of Germany,Journal of Monetary Economics, Vol. 1 (January 1975), pp. 2139.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mundell, Robert A., International Economics (New York, 1968).

  • Otani, Ichiro, and Yung Chul Park,A Monetary Model of the Korean Economy,Staff Papers, Vol. 23 (March 1976), pp. 16499.

  • Park, Yung Chul,The Role of Money in Stabilization Policy in Developing Countries,Staff Papers, Vol. 20 (July 1973), pp. 379418.

  • Polak, J. J.,Monetary Analysis of Income Formation and Payments Problems,Staff Papers, Vol. 6 (November 1957), pp. 150.

  • Polak, J. J., and Victor Argy,Credit Policy and the Balance of Payments,Staff Papers, Vol. 18 (March 1971), pp. 124.

  • Robinson, Joan, Economic Heresies: Some Old-Fashioned Questions in Economic History (New York, 1971).

  • Schwartz, Anna Jacobson,Why Money Matters,Lloyds Bank Review, No. 94 (October 1969), pp. 116.

  • Silber, William L.,Monetary Channels and the Relative Importance of Money Supply and Bank Portfolios,Journal of Finance, Vol. 24 (March 1969), pp. 8187.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • White, William H.,The Case for Moderate Use of Monetary Policy to Reduce Lags and Capital Flows” (unpublished, International Monetary Fund, August 27, 1976).

    • Search Google Scholar
    • Export Citation
*

Mr. Keller, economist in the Stabilization Policies Division of the Exchange and Trade Relations Department when this paper was prepared, is currently in the External Finance Division. He is a graduate of the University of Würzburg and the University of Rochester. For helpful comments, he is indebted to colleagues in the Fund, as well as to Prof. Rudiger Dornbusch of the Massachusetts Institute of Technology.

2

The best-known proponents are probably Milton Friedman, Anna Schwartz, Karl Brunner, and Allan Meltzer; and with respect to the aspects of trade and payments, Harry Johnson, Rudiger Dornbusch, and Michael Mussa. There is, however, dissent, for example, from Nicholas Kaldor and Joan Robinson.

3

Abstracting from Friedman’s “helicopter” method of increasing the money supply.

4

Moreover, as reserve movements affect the monetary base, they will, in the absence of offsetting action by the monetary authorities, lead (under a fractional reserve banking system) to multiple credit expansion or contraction.

5

See, for example, Johnson (1972).

7

The relative adjustment speed on the import and export side depends, inter alia, on whether the export commodity is also consumed domestically and on the time required to reallocate resources from the export sector to the domestic goods sector, and vice versa.

8

As an individual has the choice between spending (consuming) and saving out of current income, the marginal utility of consumption that he forgoes increases with an increase in the rate of savings, and vice versa for dissaving.

9

This compares with the findings of Dornbusch (1973); since he considered only one asset, money, the current account surplus in his model equals the flow demand for money.

10

If this is not the case, the absolute level of wealth must be included in the equation. It is also assumed that people are indifferent between holding foreign and domestic bonds.

11

The assumptions of perfect immobility and perfect mobility are imposed for expository reasons; the more realistic case of limited mobility is treated later on.

12

Private sector income will remain unchanged, as the interest earned on foreign bonds acquired must equal the interest forgone on bonds sold to the banking system.

13

It is assumed that changes in earnings by the banking system will not be reflected in expenditures or wealth of the private sector.

14

This characterizes the situation in the majority of countries.

15

Moreover, the sectoral distribution of credit is also of importance. This aspect will, however, be set aside.

16

See Silber (1969) for a more detailed description.

17

Brunner and Meltzer (1972 a) stress the importance of the relative price process covering all assets and yields for the transmission of monetary impulses. Although this author agrees with the usefulness of such an approach, it is not possible to incorporate such a transmission process fully into the present simple model.

19

To actually determine the value of the elasticities of substitution between various real and financial assets and to gain a complete picture of the effects of nonprice credit rationing would necessitate an approach similar to that of Brunner and Meltzer (1968) and (1972 a), which includes a model of bank behavior. The estimations of demand for money functions generally yield low and often insignificant values for the interest rate, which is interpreted by some monetarists as indicating a low elasticity of substitution between money and other financial assets.

20

It is certainly true that even in countries with well-developed capital markets small businesses have to depend on bank credit, as their size would prevent access to the private capital markets. See White (1976).

21

See also Polak and Argy (1971), who suggested that the availability of credit should enter into the expenditure function.

22

This could be interpreted as a Keynesian liquidity trap, that is, an “unstable” demand for money function.

23

And, by implication, domestic income and employment.

24

That is, changes in the money supply or in total (foreign and domestic) credit.

25

The actual adjustment speed will depend on the structural characteristics of the individual country, and especially on the role money and credit play in the economy, that is, on the development of the financial markets.

26

Empirical money demand studies indicate that the adjustment speed with which the desired level of balances is achieved is roughly 50 per cent in the first year, that is, g = 0.055 on a monthly basis; g2 is therefore negligibly small relative to g.

27

The rough proportionality, with respect to time, comes about as a deficit or surplus, which persists for only a short period, does not affect the initial conditions of the system—that is, the prevailing monetary disequilibrium—in a significant way.

28

All capital flows are assumed to take place during the first quarter.

29

In the absence of offsetting capital flows, an expansion of credit must initially bring about an equal increase in liquidity.

30

This is probably the effect of widespread credit rationing.

31

A caveat is, however, appropriate regarding the generalization of the findings. The strong immediate response of imports to changes in Q (zero lag) implied by the equation for Korea may reflect not only the short time lag between monetary expansion and aggregate demand but also circumstances specific to Korea, such as the tying of imports to exports through the import control system and the fact that foreign direct investment often took the form of imports of capital equipment. Moreover, high import content and short assembly time for certain export commodities may also have contributed to the high degree of correlation observed between imports and Q in the same quarter.

32

A similar view is expressed in Polak and Argy (1971).

33

Even Milton Friedman, who is a strong proponent of the idea of letting the money supply expand at constant rates, indicates that “… it would have been absurd to append a constant growth path to the 1933 money stock, for example, … it was desirable to start the long-run growth path from a position of rough monetary equilibrium, not from a position of significant disequilibrium.” Friedman (1972, p. 913, footnote 4).

IMF Staff papers: Volume 24 No. 1
Author: International Monetary Fund. Research Dept.
  • View in gallery

    Bond Market: The Effects of an Open Market Purchase of Bonds

  • View in gallery

    Federal Republic of Germany: Effects of a Transitory Addition to Domestic Credit in the Presence and the Absence of Capital Mobility1

  • View in gallery

    Federal Republic of Germany: Deviation of Import Growth Rates from Trend1

  • View in gallery

    Korea: Deviation of Import Growth Rates from Trend1

  • View in gallery

    Korea and the Federal Republic of Germany: Deviation of Import Growth Rates from Trend1