I. Direct Contribution of External Financial Markets to Domestic Price Inflation

The external depository financial intermediary in the form of branches and wholly or partially owned subsidiaries of major national banks has emerged as the dominant institution, with time deposits and syndicated loans as the most popular financial instruments. The offshore bond market and the other direct offshore markets, such as the commercial paper market, are of little more importance now than they were ten years ago. Concerns about the inflationary potential of external financial markets had their origins in the disequilibrium models of the money and credit supply process, in which the growth and the composition of the liabilities of banks are constrained by the growth of the stock of base or reserve money. Since external depository financial intermediaries hold only minimal reserve balances with their domestic correspondents, their potential to increase the money and credit aggregates in external financial markets, through the creation of external substitutes for reservable domestic liabilities, appeared to be limitless.⁷ The fact that periods of rapid growth of external financial markets tended to coincide with periods of high rates of price inflation provided the empirical evidence for such views. The more recent attempts at modeling the interactions between external and domestic financial markets differ in such details as the number of different types of transactor and financial instrument and assumptions about exchange rate determination. Nevertheless, they are akin in spirit and yield similar conclusions. The structure of these models consists of the portfolio preferences of the nonbank public, domestic and external banks, and of the central banks in the domestic and external regions; wealth and balance sheet constraints; and asymmetric financial regulation in the form of reserve requirements on some classes of deposit in the regulated region. The quantities of central bank liabilities (i.e., high-powered money), together with a reserve requirement on domestic deposit liabilities, determine the size of the monetary aggregates and the price level. The essential factors that determine the equilibrium distribution of assets over the two regions and the equilibrium interest rates are the gross-substitute asset demand and supply functions and the asymmetric financial regulations.

The differences in the conclusions concerning the money and credit creating potential of the external financial markets, which emerge from these models, can be explained fully by the differences in the assumptions about the adjustment behavior of interest rates and quantities in the relevant financial markets. In particular, market disequilibrium models give rise to multiplier effects if the speed of adjustment of interest rates is lower than the speed of adjustment of quantities.⁸ In the case at hand, the assumption of fixed or slowly adjusting interest rates in financial markets leads to the fixed-coefficient money and credit multipliers obtained in the analysis of an exogenous shift of domestic bank liabilities to an external market.

A review follows of the money and credit creating potential of external financial intermediaries in the temporary equilibrium models in which interest rates adjust slowly relative to the speed of adjustment of the quantities of financial assets demanded and supplied. In particular, the early literature on offshore financial markets assumed that asset demand and supply functions are infinitely elastic.⁹ In this case, the amount of deposit liabilities is determined entirely by the reserve requirement and the volume of central bank liabilities. The uncontrolled external intermediary holds reserves in the form of demand deposits with a domestic correspondent bank in the amount of k_e(ED), where k_e is the ratio of the reserves (ER) held by the external intermediary against its nonbank deposits (ED). Only a fraction of the loans made by external intermediaries is redeposited with external intermediaries; the remainder is leaked back to the domestic intermediaries. This fraction is assumed to be approximated by the ratio of external deposits to the sum total of external and domestic demand deposit liabilities, that is, $g=\frac{ED}{DD+ED}$. An exogenous shift of domestic demand deposits in the amount of $\overline{\mathrm{\Delta }ED}$ to the external intermediaries leads then to an expansion in external deposits of $\mathrm{\Delta }ED=\frac{1}{1-(1-k_e)g}\overline{\mathrm{\Delta }ED}$. The multiplier $m_e=\frac{1}{1-(1-k_e)g}$ can assume a mminimum value of unity and has no extremum. Such an exogenous shift of deposits can originate from a change in depositors’ evaluations of the risk associated with the liabilities of external intermediaries, as might occur when the intervention rules of the lenders of last resort change. A decrease in the transaction costs of international financial transactions also will produce a shift of deposits toward external intermediaries. Empirical investigations have sought to obtain estimates for g and k_e and thereby obtain values for m_e. Such calculations produced values for m_e ranging from 1.02 to 2.0, depending on the definition of reserves and the size of the Eurodollar market.¹⁰

A second method for estimating m_e attempts to identify ex post a reserve base for the Eurodollar market and to compare it with the total outstanding external liabilities to obtain a multiplier. The estimates obtained in this manner range from 2 to 18, depending on the definition of a reserve base and the size of the external market.¹¹ Since it is not possible to identify empirically the initial deposit shift to which to apply this multiplier, this method is not very useful.

It is possible to determine similar multipliers for other classes of deposit liability, once the ratios among other items in the external balance sheet and the equilibrium reserves applicable to such items are specified.

In the disequilibrium quantity-adjustment models, the exogenous shift of deposits from domestic to external intermediaries results in excess reserves onshore if the shifted deposits are reservable. The domestic intermediaries will expand loans and deposits until excess reserves are again fully utilized. Since the total stock of high-powered money is unchanged, the stock of onshore monetary liabilities will contract only by the amount of the correspondent balances held domestically by external intermediaries, that is, by k_e(ΔED). Hence, the exogenous shift of reservable domestic deposits to external intermediaries will result in an expansion of total deposits (i.e., monetary assets) held by the nonbank public of (1 – k_e)m_eED = ΔM. Thus, the money creating potential of an exogenous shift of deposits from onshore to offshore intermediaries can be measured by M in the expression above.

The total amount of onshore and offshore deposits supported by the predetermined stock of central bank liabilities is then given by

where DD and ED are onshore and offshore demand deposits, k_d and k_e are onshore and offshore reserve ratios, and R is the amount of high-powered money.¹² The multiplier $m_r=\frac{1}{k_d(1-(1-k_e)g)}$ has a minimum value of unity and is unbounded. Since the deposit expansion multiplier in the absence of an offshore market is given by $m_d=\frac{1}{k_d}$, that is, $DD=\frac{1}{k_d}R$, it is observed that m_T = m_dm_e and hence $\frac{m_{T}R-m_{d}R}{m_{d}R}=(m_e-1)$ represents a once-and-for-all percentage increase of m_dR in the monetary aggregate M = DD + ED that can be achieved by economizing on high-powered money. Thus, (m_e–1) is a second measure of the money creating potential of the offshore markets. With the empirically plausible range of 0.02 to 0.08 for g and 0.01 for k_e, the external financial markets can bring about a 2 to 9 per cent once-and-for-all increase in the monetary aggregate consisting of domestic and external deposit liabilities.¹³

The disequilibrium analysis of the inflationary potential of external financial markets, described above, thus yields unambiguous analytical results under the assumption that in-period interest rate changes are absent, that is, demand and supply functions of financial assets are infinitely elastic.¹⁴ However, the assumptions concerning the relative speed of adjustment of prices and quantities in financial markets, which define this adjustment process, are subject to doubt. The assumption that interest rates remain fixed or change more slowly than the quantities of financial assets demanded and supplied prevents offsetting changes in the composition of portfolios. For example, the domestic intermediaries are assumed to be able to lend the excess reserves that result from a shift of reservable deposits to an external location without inducing a change in loan and deposit rates. Such assumptions are justified only by the presence of imperfections in capital markets in the form of interest rate ceilings, imperfect information, or other price-adjustment costs. With the exception of interest rate ceilings, it is hard to find empirical evidence that supports the existence of such imperfections.

If interest rates vary so as to equate the supply and demand of financial assets, then it becomes necessary to determine the effects of interest rate changes on the composition of the portfolios of wealthowners and intermediaries. In the portfolio models that seek to incorporate such general equilibrium consideration, the portfolio behavior of wealthowners is captured in their demand for loans and supply of deposits functions. The demand for loans and the supply of deposits depend on market interest rates in the onshore and offshore markets and are now assumed to possess finite interest rate elasticities, and deposits as well as loans are assumed to be gross substitutes. All credit is intermediated, and the portfolio behavior of intermediaries is passive in the sense that it is determined by their balance sheet constraint. Desired reserves, then, are a constant fraction of deposits, and the markup of loan rates over deposit rates is assumed to be constant.¹⁵

By applying comparative static methods, it is again possible to derive the equilibrium effect of an exogenous shift of reservable domestic deposits to an external region on the volume of deposit liabilities of external intermediaries. In particular, a shift of $\overline{\mathrm{\Delta }ED}$ from onshore to offshore deposits will raise the equilibrium level of the offshore deposits by less than $\overline{\mathrm{\Delta }ED}$, that is, the initial deposit multiplier is less than unity. In the disequilibrium quantity-adjustment model, the offshore intermediaries were constrained from expanding deposits only by their desired reserve ratios and a fixed stock of base money. A flow of domestic deposits to an external region at constant interest rates led to a multiple expansion of external deposits. In the equilibrium model, the exogenous flow of domestic deposits to external intermediaries results in higher domestic market interest rates and lower interest rates in the external market. Hence, a fraction of $\overline{\mathrm{\Delta }ED}$ returns to the onshore market.¹⁶ Thus, the equilibrium increase in ED, the stock of offshore deposits, is less than $\overline{\mathrm{\Delta }ED}$, while the decrease in DD, the stock of onshore deposits, is also less than $\overline{\mathrm{\Delta }ED}$; onshore interest rates increase and offshore interest rates decrease. The total money stock, M = DD + ED, increases at most by $\overline{\mathrm{\Delta }ED}$. The money creating potential of the offshore markets in such general equilibrium models is, therefore, much less than in the disequilibrium models. As before, however, this expansion in M is due to the economies in high-powered money that the multitier banking system affords intermediaries.¹⁷ Empirical estimates for the initial deposit multiplier in the equilibrium model range from 0.6 to 1.0,¹⁸ and hence they are in accord with the upper bound derived analytically.

However, while the general equilibrium multistage banking models represent an improvement over the temporary equilibrium fixed-coefficient multiplier models, they still only characterize the money and credit potential of a one-time exogenous shift of deposits of a given size to an external intermediary. In the models described above, the size of the initial shift of deposits to external financial markets is given exogenously or is determined by the exogenously given interest elasticity parameters of the asset demand functions of nonbank wealthholders. To obtain an adequate assessment of the money and credit creating potential of external financial markets, it is necessary to investigate the determinants of the magnitude of the flow of domestic deposits to external locations. The possibility that an initial shift of deposits to an external location results in a multiple expansion of money and credit aggregates has been ruled out by the analyses reviewed above. Hence, the volume of domestic deposit liabilities that is moved to an external location in response to changes in incentives is much more important in determining the money and credit creating potential of external financial markets than the size of any external money or credit multiplier. It is, therefore, necessary to explore the portfolio adjustments of nonbank wealthholders and external intermediaries in response to changes in the transaction and risk costs of external transactions.

The set of financial instruments available to private wealth-owners consists of the liabilities of domestic and external depository intermediaries and of the directly issued liabilities of domestic and external borrowers. Similarly, the financial deficit unit can borrow from domestic and external depository intermediaries and directly from domestic and external lenders. The risk and return characteristics of any particular financial asset are then dependent on the type of issuer (direct or intermediary), the location of the issuer (domestic or external), and the currency of denomination.

In particular, the depository intermediary will arrange its balance sheet so as to maximize its own market value subject to its investment opportunity set and its production technology. The risk and return characteristics of its deposit liabilities are then endogenously determined by the default risk of the intermediary and its expected profits. The default risk in turn is determined by the investment decisions of the intermediary, the extent of its open position on the term structure of interest rates (its maturity transformation), and by the possibility of lender-of-last-resort protection from a central bank.¹⁹ The characteristics of the joint distribution of the return on the liabilities of the financial intermediaries and direct borrowers define the opportunity set for nonbank financial surplus units, and private wealthowners will select a portfolio that is mean-variance efficient with respect to opportunity set. Such a general equilibrium asset-pricing model will then yield equilibrium prices and quantities under the assumption that individuals select a mean-variance efficient portfolio and have the same ex ante beliefs about the expected returns and covariance structure of the assets in the opportunity set.²⁰ It is conjectured that such models generate an explicit equilibrium valuation formula in which the market value of financial assets is approximately equal to its expected return less a correction for its nondiversifiable risk and discounted by a risk-free interest rate.²¹

External intermediation and direct external investment can be included in such models of asset pricing by including the liabilities of external intermediaries and the liabilities of external direct borrowers in the portfolio opportunity set available to the individual lender. An external intermediary can offer a higher expected return on some types of asset, owing to net savings in production costs in international financial centers and to savings in regulatory costs. The risk characteristics of liabilities issued externally are affected by the possibility of access to a lender of last resort and the legal connection of the external intermediary to a domestic intermediary.²²

The possibility of denominating the return on financial assets in different currencies introduces an important complication in asset market models. Returns on assets are measured in nominal accounting units, while preferences are defined over bundles of real goods. Hence, nominal returns must be deflated by a price index, and the holder of a nominal asset must be compensated for any uncertainty in the price index, that is, for purchasing power risk.²³ When the rates of exchange of various currencies are flexible, the change in the purchasing power of these currencies may differ. In this world, purchasing power parity and unencumbered capital flows are sufficient for the real commodity returns on a particular security to be the same for all residents in all countries. The nominal return on a given security will vary across countries because of changes in exchange rates, but these changes offset exactly any changes in the purchasing power of different monies.²⁴ If the purchasing power parity condition does not hold, there is exchange rate uncertainty in addition to purchasing power uncertainty. Once the joint distribution of the changes in the exchange rate and the purchasing power is known, it is possible to derive an international portfolio opportunity set, and equilibrium prices of the available financial assets will take the same form as in the model restricted to domestically issued financial instruments. This type of asset-pricing model of international financial markets outlined above would have to be modified when restrictions on international investment are present. Such a model would, however, indicate approximately how the equilibrium price and the currency composition of a domestic or external financial asset are determined by the asset’s expected rate of return and its non-diversifiable risk from various sources; and it also would reveal the amounts of funds invested in each asset.

Any changes in the determinants of the risk and return characteristics of domestic and external financial instruments, such as a reduction in the cost of some types of financial transaction, increased lender-of-last-resort protection, increases in default risk of borrowers, changes in purchasing power risk, and redistribution of financial wealth, will alter the domestic and external financial market equilibrium in accordance with the portfolio preference of nonbank wealthholders.

The past decade, for example, has witnessed secular changes in financial transaction technology and in the default risk of external depository intermediaries. The decline in the cost of long-distance financial transactions has permitted external depository intermediaries to offer a higher rate of return on their deposit liabilities. Increased lender-of-last-resort protection and a better definition of the legal responsibilities of domestic parents toward their external subsidiaries have lowered the risk associated with external deposit liabilities with more favorable risk-return characteristics,²⁵ resulting in a sustained secular growth of external deposits. Furthermore, the subsidy to external intermediaries that is implicit in the extended lender-of-last-resort protection has given external banks a competitive advantage over directly issued external debt, resulting in a slower growth of the external bond markets. Also, the pattern of trade flows among oil exporting countries and the developed oil importing countries during the past decade has resulted in a redistribution of financial wealth toward oil exporting countries. The political risk associated with the liabilities of external intermediaries is presumed to be lower for nonresidents than for residents.

Thus, the secular changes in risk and transaction costs, and the redistribution of financial wealth, have caused adjustments in the portfolios of nonbank wealthowners and in the portfolios of domestic and external financial intermediaries that have resulted in a secular growth of the liabilities of external financial intermediaries. The equilibrium increase in the stock of liabilities of external intermediaries and the changes in their composition are determined along the lines of the general equilibrium financial markets model described above.

While it is possible to establish analytical limits on the expansion possibility of the external financial markets owing to an exogenous shift of deposits, as in the portfolio equilibrium banking model, it is not possible to establish similar bounds on the rate at which deposits move toward external intermediaries. This movement is determined by the rate of technological innovations in the financial transaction technology, the reduction in risk of external liabilities, and by the way in which the general financial market equilibrium is established. Consequently, the implied money and credit potential of external financial markets that arises from the possibility of economizing on base money may be quite large. Recent developments such as the introduction of same-day settlement and international banking facilities in the United States are likely to contribute substantially to the flow of deposits to un-reservable balance sheets, with an accompanying expansion of the total domestic and external monetary aggregates.

In addition to such secular changes in the stock of external deposits, there are cyclical changes brought about by changes in domestic monetary policy. The short-run increase in domestic interest rates arising from a restrictive monetary policy increases the opportunity cost of noninterest-bearing reserves held against domestic bank deposits. External intermediaries are thus able to offer liabilities with more favorable risk-return characteristics, and the equilibrium stock of external liabilities will increase. However, when exchange rates are flexible or when the effects of exchange market intervention on the domestic money stock are offset by changes in the currency composition of the central bank assets, residents cannot directly evade a restrictive domestic monetary policy by increasing their net borrowing from external financial intermediaries.²⁶ The balance of payments constraint implies that any increase in net foreign borrowing has to be met by an increase in the deficit on the current account. But in the short run the current account is unresponsive to a restrictive monetary policy, and in the longer run a reduction in economic activity will reduce the deficit, forcing a reduction in net borrowing from abroad. Hence, the balance of payments constraint implies that a restrictive domestic monetary policy can be thwarted only if residents are willing to hold more liabilities of external banks, thereby increasing the ability of external banks to exploit the money and credit creating potential of base-money economies.

The conclusion that the possibility of economizing on high-powered money endowed the external financial markets with a significant inflationary potential is undermined by doubts about the appropriateness of including all or even some external liabilities in the domestic monetary aggregates. In the literature reviewed above, the inflationary potential of a shift of domestic deposits to external intermediaries is approximated by the increase in the total of domestic and external deposits. Since the returns on onshore and offshore deposits have different mean-variance and maturity characteristics, it is not immediately apparent how one would construct an appropriate monetary aggregate that would best measure the inflationary potential. As an extreme example, if it is found that all liabilities of external banks are to be excluded from domestic monetary aggregates, then the growth of external financial markets has no inflationary potential, while a maximum inflationary potential exists when all external deposits are to be counted as part of the domestic monetary aggregates.

The possibility of constructing an appropriate monetary aggregate from the liabilities of domestic and external banks remains largely unexplored.²⁷ It is nevertheless possible to use ex post data on domestic and external monetary aggregates to establish bounds on the potential contribution of the growth of external money and credit aggregates to domestic price inflation. A very narrow definition of a total monetary aggregate is arrived at by including only that part of the liabilities of external intermediaries that matures in less than eight days and that represents transactions among residents only (see Tables 2 and 3). For example, from 1975 to 1980, the growth of this portion of the Eurodollar balance sheet has been about twice as high as the growth in the U.S. monetary aggregate M1-B, the growth rate of which is used as an intermediate target of monetary policy. In particular, M1-B grew by 35 per cent during this period, while round-trip external liabilities of less than eight days’ maturity grew by 74 per cent. However, since this class of external liability constituted less than 3 per cent of the total monetary aggregate, the growth in the total monetary aggregate exceeded the growth of M1-B by only 0.3 percentage point during 1975–80. The liabilities of external intermediaries that mature in less than eight days and that represent transactions among residents and of residents with nonresidents grew by 44 per cent over the period 1975–80. The growth of the corresponding total monetary aggregate exceeded the growth of M1-B by 2 percentage points.

Table 2.

Maturity Transformation in External Financial Markets, November 1980

(In per cent)

Sources: Bank of England, Quarterly Bulletin (various issues); and data from U.K. and U.S. banks.

Table 2.

Maturity Transformation in External Financial Markets, November 1980

(In per cent)

	Assets	Liabilities
Less than eight days	9.6	37.8
Eight days to one month	9.4	22.9
One month to six months	17.2	33.6
Longer than six months	63.8	5.7
	100.0	100.0

Sources: Bank of England, Quarterly Bulletin (various issues); and data from U.K. and U.S. banks.

View Table

Table 2.

Maturity Transformation in External Financial Markets, November 1980

(In per cent)

	Assets	Liabilities
Less than eight days	9.6	37.8
Eight days to one month	9.4	22.9
One month to six months	17.2	33.6
Longer than six months	63.8	5.7
	100.0	100.0

Sources: Bank of England, Quarterly Bulletin (various issues); and data from U.K. and U.S. banks.

Table 3.

Type of Dollar-Denominated Transaction in External Financial Markets

(In billions of U.S. dollars)

Source: Bank for International Settlements.

Table 3.

Type of Dollar-Denominated Transaction in External Financial Markets

(In billions of U.S. dollars)

	Assets		Liabilities
	1975	1980	1975	1980
Central banks	5.1	17.8	30.4	78.0
U.S. residents	6.7	11.7	8.3	42.0
Rest-of-world residents	78.6	189.1	41.3	90.8
	90.4	218.6	80.0	210.8

Source: Bank for International Settlements.

View Table

Table 3.

Type of Dollar-Denominated Transaction in External Financial Markets

(In billions of U.S. dollars)

	Assets		Liabilities
	1975	1980	1975	1980
Central banks	5.1	17.8	30.4	78.0
U.S. residents	6.7	11.7	8.3	42.0
Rest-of-world residents	78.6	189.1	41.3	90.8
	90.4	218.6	80.0	210.8

Source: Bank for International Settlements.

Under the additional assumption that the marginal velocity of the combined monetary aggregate has not exceeded the marginal velocity of the narrow domestic monetary aggregate,²⁸ one can conclude that the potential direct contribution²⁹ by the external financial markets to the U.S. rate of inflation has been between 0.06 and 0.4 percentage points per annum during the period 1975–80. The relative insignificance of the direct contribution of the growth of external monetary aggregates to domestic price inflation has been due to the low ratio of external to domestic monetary aggregates. Further increases in this ratio will increase the direct contribution to domestic price inflation of a growth rate of external monetary aggregates in excess of the growth rate of domestic monetary aggregates.

It is frequently claimed that the inflationary potential of external financial markets may have been enhanced by the extensive maturity transformation in the balance sheets of external depository intermediaries (see Table 2). Net liquidity creation occurs when the intermediary’s stock of liabilities of a given maturity exceeds its stock of assets of the same maturity. In particular, net liquidity creation is positive if the intermediary borrows short and lends long. The maturity transformation by external intermediaries became more pronounced in the latter half of the 1970s. There is, however, no analytical or empirical evidence that maturity transformation is in fact inflationary in the absence of changes in the size of monetary aggregates. Furthermore, by including gross liabilities of external banks,³⁰ instead of net liquidity creation, in the appropriate monetary aggregates, we obtained estimates of the inflationary contribution of the external markets that include the effects of a 100 per cent maturity transformation and hence overstate the true inflationary contribution of external financial markets.³¹

II. Indirect Inflationary Potential of External Financial Markets

It is recognized that external financial markets possess an important indirect inflationary potential through their effects on domestic monetary and exchange rate policies, in addition to the direct inflationary potential of base-money economies afforded by bank liabilities issued externally.

First, the internationalization of markets for bank liabilities has greatly increased the difficulties of defining and interpreting money and credit aggregates.³² In such an environment, the choice of a narrow domestic monetary aggregate as the instrument of monetary policy may prove to be inflationary. During the past decade, it has been increasingly recognized that appropriately conducted monetary policy should seek to attain an intermediate target in the form of a predetermined rate of growth of some monetary aggregate, using bank reserves as the instrument of monetary policy. For example, the central banks of the Federal Republic of Germany, the United States, the United Kingdom, and Canada have moved in this direction.³³ Such a monetary policy paradigm, however, requires a reliable definition of monetary aggregates that is empirically measurable. The internationalization of banking has made it necessary to decide whether the geographical location of bank liabilities should affect their inclusion in the monetary aggregate to be targeted.³⁴ The measurement problems encountered with a more comprehensive definition of the money stock, rather than analytical or empirical evidence, are responsible for the narrow definition.³⁵ Since a proportion of the balance sheet of external depository intermediaries constitutes a substitution of external for domestic financial intermediation, that is, it represents round-trip transactions among nonbank domestic transactors (see Table 3), such an exclusion results in an underestimation of the relevant money and credit aggregates to be targeted. A comprehensive measure of money and credit aggregates should not be affected by shifts of liabilities among different geographical locations, but only by risk, maturity, liquidity, and transaction characteristics of the financial asset included in the aggregate.

Moreover, the narrow domestic aggregates are not reliable proxies for the appropriate aggregates. The secular rate of growth of external financial markets, which has been due to the reduction in the risk associated with the deposits of external financial intermediaries and to the reduction in the transaction costs of international financial transactions,³⁶ has been in excess of the growth of the narrowly defined domestic aggregates. In addition, changes in monetary policy itself alter the relationship among domestic and external money and credit aggregates. The short-term increase in domestic deposit rates accompanying a restrictive monetary policy increases the opportunity cost of reserve requirements against domestic bank liabilities. As a result, externally issued liabilities are substituted for domestic liabilities. Thus, secular growth and cyclical variation in the size of external financial markets change the ratio of domestic to externally issued bank liabilities. The secular growth of external markets in excess of the growth in domestic financial markets imparts an inflationary bias in monetary policies targeting narrowly defined domestic aggregates. This bias is strengthened by the cyclical variations of the size of external financial markets caused by changes in domestic monetary policy. The expansion of external financial markets during periods of restrictive monetary policy undermines the deflationary potential of such a policy. By issuing deposit liabilities externally, the domestic banking sector can exploit the base-money economies and is thus able to evade partially the effects of a monetary restriction. This effect is at least partly offset, however, by the decrease in the growth of external financial markets during periods of domestic monetary expansion.

Second, the external financial markets possess a significant inflationary potential when monetary authorities choose a domestic interest rate as an intermediate target of domestic monetary policy.³⁷ Under such a policy regime, the effect on domestic interest rates of a shift of domestic bank liabilities toward external depository intermediaries will be counteracted by the monetary authority, with an expansion in domestic base money. Consequently, any reduction in the risk, transaction, or regulation costs associated with the liabilities of external financial markets not only produces an increase in external money and credit aggregates but also results in an increase in domestic base money and hence an increase in domestic money and credit aggregates over their original level. The growth of domestic monetary aggregates is thus partly determined by the growth of the external financial markets. Hence, such a policy regime does not satisfy the necessary conditions for long-run stability in the expansion of domestic and external money and credit aggregates.³⁸

Third, it is frequently presumed that the inflationary potential of the external financial markets is enhanced by the redepositing of dollar reserves in external markets by foreign central banks.³⁹ This presumption led in 1971 to an agreement among the central banks of eleven major countries to refrain from such practices.⁴⁰ It was argued that the deficits in the U.S. balance of payments caused foreign central banks to acquire dollar reserves, which were re-lent to the rest of the world when deposited in the external currency markets, thus requiring further intervention in the exchange markets by foreign central banks. Therefore, a deficit in the U.S. official settlements balance results in a multiple increase in the official settlements balance of the rest of the world, as well as a multiple increase in reserves. The monetary base of the rest of the world increases by the same multiple unless the effects of exchange market intervention in the monetary base is sterilized. This analysis depends crucially on the assumption that the external banks relend the newly deposited dollar reserves of foreign central banks to the rest of the world and on the assumption that the initial shift of official reserves from U.S. financial markets to the external market does not induce offsetting flows of deposits, such as might arise with the increase in domestic deposit rates and the decrease in external deposit rates accompanying the shift of official reserves. Such assumptions are, however, not warranted, since the external and domestic financial markets for dollar-denominated instruments are closely linked. Hence, the redepositing of official reserves by foreign central banks can be dismissed as a factor contributing to the inflationary potential of external financial markets.

Fourth, it has been argued that the increase in the mobility and volume of international capital flows, as well as the increase in the substitutability of similar financial assets denominated in different currencies, induced by the growth of the external financial markets, has exerted significant inflationary pressures on the affected economies.⁴¹ The presence of asymmetries in the balance of payments policies of the capital importing and capital exporting countries or asymmetries in the response of domestic prices to exchange rate changes is a necessary condition for capital flows to be inflationary.

An increase in the volume of international capital flows that adds to external settlement imbalances implies that a larger change in the exchange rate or a larger volume of exchange market intervention is needed to restore equilibrium. The devaluation of the exchange rate in the capital exporting country will exert an expansionary effect, while the revaluation of the exchange rate in the capital importing country will exert a contractionary effect. The presence of downward rigidity in the price level of the capital exporting country, owing to long-term contracts, implies, then, that increased variability in the exchange rate may be inflationary. This argument is correct in the short run only; in the long run, such an inflationary tendency can be sustained only by such increases in the monetary aggregates as validate the price increase.

While there is ample casual empirical evidence that various countries have adopted an accommodating monetary policy in the presence of downwardly rigid prices, there is little or no econometric evidence to quantify this contention. More importantly, there is no substantive empirical evidence to support the hypothesis that the external financial markets have notably increased the mobility of international capital or the substitutability of financial assets in different currencies.⁴²

Alternatively, asymmetries in the ability of national policy authorities to sterilize the effects of the increased exchange market intervention on the monetary aggregates can be inflationary. If, for example, the decrease in the monetary base of the capital exporting country is compensated, while the increase in the monetary base of the capital importing country is only partially sterilized, then any increase in capital flows will exert inflationary pressures. In particular, before the change in U.S. policy in November 1978, capital flows from the United States added to the official reserves of the receiving countries, which then reinvested these reserves in U.S. financial markets; that is, foreign central banks acquired the U.S. assets sold by U.S. residents in exchange for foreign currency assets. Hence, the U.S. monetary base remained unchanged. Since the increase in the monetary base of the receiving country was frequently not fully sterilized, such capital outflows may have been inflationary.

On the other hand, an increase in the volume of international capital flows that offsets external settlements imbalances implies that a smaller change in the exchange rate or a smaller volume of exchange market intervention is required to restore equilibrium. It is argued, however, that it is more likely for deficit countries to employ contractionary policies to restore balance of payments equilibrium than it is for surplus countries to employ expansionary policies. In this case, increases in capital flows, which lessen the deflationary tendency of the adjustment process, are viewed as inflationary.⁴³

III. Conclusions and Some General Observations

The synthesis and assessment of the existing theoretical and empirical literature on the money and credit creating potential of external financial markets and the inflationary implication of this potential, as presented above, permit some general conclusions.

Despite their popularity in the literature, one is forced to reject the early disequilibrium fixed-coefficient models of multistage banking systems, as well as their later extensions to general equilibrium portfolio models, as inadequate for an analysis of the money and credit creating potential of external financial markets. Models of this type are by design restricted to a partial equilibrium analysis of the expansion in money and credit aggregates made possible by the economies in base money that an exogenous initial shift of a given amount of reservable deposits to an external location affords the banking system.⁴⁴

The money and credit creating potential of external financial markets attributable to base-money economies is, however, determined not so much by the size of the various multipliers, but more importantly by the magnitude of the deposit shifts toward external markets, that is, the quantity to which the multiplier is applied. In fact, the main conclusion that emerges from the existing literature is that the ultimate expansion in the total of domestic and external money and credit aggregates caused by an initial exogenous shift of bank liabilities to an unregulated region is smaller than the size of the initial shift of deposits. Hence, the volume of the initial deposit outflow, in response to a favorable change in the incentives to hold and issue external deposits, is the more important determinant of the money and credit creating potential of external financial markets.

The volume of the initial deposit shifts is governed by the changes in the relative returns of domestic and external financial instruments. Such changes in relative yields in turn are due to the secular reduction in the risk associated with the liabilities of external intermediaries and the reduction in the cost of international financial transactions. Furthermore, changes in domestic monetary policy result in a cyclical variation in deposit rates, which alters the opportunity cost of holding noninterest-bearing required reserves against domestic bank liabilities and so contributes to variations in the relative yields of external and domestic bank liabilities.

The absence of analytical models of the secular and cyclical changes in the incentives to hold and issue external deposits, that is, models of the changes in risk and transaction costs of external deposits, and the lack of models of international financial market equilibrium capable of relating such changes to equilibrium market returns and quantities make it difficult to determine accurately the money and credit creating potential of the external financial market.

It is likely, however, that a better definition of the lender-of-last-resort role of central banks and of the legal responsibilities of a domestic parent bank toward its external subsidiary will result in a rapid reduction in the risk associated with external deposits. Also, the competitive reduction of financial regulations applicable to external banks, such as the establishment of the international banking facilities in the United States and of offshore centers in many other locations, may significantly reduce the cost of external banking. In either case, the growth of external intermediation is likely to accelerate further.

The implications of the money and credit creating potential of external financial markets for domestic price inflation have remained largely unexplored in the literature. The contribution of the expansion in the liabilities of external depository intermediaries to domestic price inflation is determined by the fraction of their monetary liabilities that is to be included in the domestic monetary aggregates. It is concluded here that at least that part of the monetary liabilities of external intermediaries that represents round-trip transactions among domestic residents should be included. If one assumes, in addition, that the marginal velocity of the combined monetary aggregate did not exceed the marginal velocity of the narrow domestic monetary aggregate, then it is possible to obtain some upper bounds for the potential contribution of external financial markets to domestic price inflation. Historically, the growth rates of external money and credit aggregates have exceeded those of their domestic counterparts. As a result, the combined domestic and external aggregates have grown faster than the domestic aggregates. By assuming that the marginal velocity of the combined aggregates did not exceed that of the domestic aggregates, one can obtain some ex post estimates of the direct contribution of external financial markets to domestic inflation. For example, from 1975 to 1980, the growth rate of this portion of the Eurodollar balance sheet has been twice the growth rate in the U.S. monetary aggregate M1-B, which is used as an intermediate target of monetary policy. In accordance with this definition of a total monetary aggregate, it is concluded that the external financial markets have at most contributed directly 0.3 percentage point to domestic price inflation in the United States over the period 1975–80. If, however, the definition of a total monetary aggregate is extended to include that fraction of external monetary liabilities that represents transactions among residents and nonresidents, then the growth of the external monetary aggregate exceeded the growth of M1-B by 2 percentage points during the period 1975–80. Such an assumption about the relation between domestic and external monetary aggregates implies, then, that the contribution of the growth of external monetary aggregates to domestic price inflation in the United States may have been as high as 0.4 percentage point per annum from 1975 to 1980. Similar estimates can be obtained for other countries. Further increases in the ratio of external to domestic monetary aggregates could increase the direct contribution to domestic price inflation of a growth rate of external monetary aggregates in excess of the growth rate of domestic monetary aggregates.

The inflationary potential of the money and credit creating activities in the external financial markets is enhanced by the external depository financial intermediaries’ practice of sustaining a greater degree of maturity transformation in their balance sheets than do the domestic depository financial intermediaries. While the maturity transformation in external balance sheets became more pronounced in the last half of the 1970s, there is no empirical evidence on its inflationary effect. However, by including the gross liabilities of external banks in the appropriate monetary aggregates, as was done above, instead of net external liquidity creation, one obtains estimates of the inflationary potential of base-money economies inclusive of maturity transformation.

It is concluded, further, that the increase in the volume of international financial intermediation attributable to the more efficient matching of borrowing and lending units achieved in the external financial markets is by itself not inflationary. Under a flexible exchange rate regime, variations in the exchange rates and interest rates will ensure that the excess of the domestically originating aggregate supply over domestic absorption in the capital exporting countries is matched by an excess of investment over saving in the capital importing countries. In a fixed exchange rate regime, the central banks will alter the currency composition of their liabilities to offset any increases in private capital movements that are inconsistent with the given exchange rate.

The conclusion here is somewhat less sanguine with respect to the potential of external financial markets to add to domestic price inflation indirectly by influencing the conduct and effectiveness of domestic monetary policy. In particular, the growth of external financial intermediation has made the definition and interpretation of money and credit aggregates much more difficult. The questions concerning the inclusion of the liabilities of external intermediaries and the liabilities of resident foreign banks in the appropriate domestic money and credit aggregates remain largely unanswered. Furthermore, even if such questions can be resolved satisfactorily, the difficulties encountered in measuring the liabilities of external intermediaries, in particular of foreign intermediaries, remain.

The observed secular growth of the liabilities of external banks at a rate in excess of the rate of growth of domestic bank liabilities and the cyclical fluctuations in the volume of external liabilities caused by changes in domestic monetary policy together rule out the possibility of a stable quantitative relationship among the domestic and external money and credit aggregates. The failure to include external aggregates in the appropriate domestic money and credit aggregates thus implies that the traditional domestic money and credit aggregates are less reliable as intermediate targets of monetary policy. For example, any particular target rate of expansion of a domestic monetary aggregate will be more inflationary the greater the rate of growth of the appropriate external monetary liabilities. Since an increasing number of countries have adopted such a method of conducting monetary policy, this problem is of some importance.

Furthermore, it is concluded that an increase in the mobility of international capital flows or an increase in the substitutability of financial assets across currencies attributable to the growth in external financial markets may exert inflationary pressures. Any increase in international capital flows that increases external payments imbalances must be met by exchange market intervention or exchange rate changes. The source of the inflationary potential of increased capital flows is seen to originate from an asymmetry in the ability or willingness of monetary authorities to sterilize the changes in monetary aggregates brought about by exchange market intervention, as well as from the asymmetry in the contribution of exchange rate changes to domestic price inflation when prices are downwardly rigid. In particular, the proportion of the decrease in monetary aggregates that is sterilized in periods of net capital outflows is assumed to exceed the proportion of the increase in monetary aggregates that is sterilized in periods of net capital inflows. Alternatively, the exchange rate depreciation in the net capital exporting country is assumed to induce monetary authorities to expand monetary aggregates by more than they contract the aggregates during an exchange rate appreciation in periods of net capital inflows. Similarly, in the presence of downwardly rigid prices, the monetary authority is assumed to validate the price increases in the traded goods sector that arise with changes in the exchange rate.

While there is a good deal of casual empirical evidence that supports such assumptions about the conduct of monetary policy, there is little or no empirical evidence that the growth of external financial markets has notably increased the mobility of international capital or the substitutability of financial assets denominated in different currencies.