Appendix IV: Growth of International Bank Assets: Empirical Analysis
- International Monetary Fund
- Published Date:
- August 1981
Introduction and Summary
Data published by the Bank for International Settlements (BIS) show that gross external assets in both domestic and foreign currencies of deposit banks in major industrial countries increased from $39 billion at the end of 1965 to $964 billion at the end of 1979.84 When allowance is made for some minor conceptual breaks in the time series, the average annual rate of increases during the 14-year period was 25.7 per cent. This Appendix presents a simple framework for interpreting the growth of external bank assets over the last several years and, on the basis of econometric analysis for the period 1970–79, discusses developments in 1980 and makes projections for 1981.85
The main conclusions from the econometric work can be summarized as follows: (1) the high growth rate of international banking is related in an important, if complicated, way to the rapid expansion of nominal world income and trade; (2) other factors, possibly including increased recognition of the advantages of these markets in terms of higher return and lower cost to participants, may have also contributed in a systematic way to the exceptional growth of international bank assets during the 1970s; (3) while increases in the degree of imbalance in world trade positions may have led to increases in international bank assets and liabilities during particular periods, their contribution to the overall growth of international bank assets over the entire period 1970–79 was only marginal because the degree of trade imbalance rose considerably during certain periods and declined very sharply during others.
The actual rate of growth of banks’ gross external assets (as defined above) was 20.1 per cent in 1980, compared with 24.9 per cent in 1979. Adjusted for exchange rate changes,86 the underlying pace of expansion of external assets was broadly maintained in 1980. The slower growth in 1980 of world gross national product (GNP) and trade tended to depress the rate of market growth, whereas the significant increase in global payments imbalance—and attendant increase in demand for balance of payments finance—was a major expansionary counterforce. These seem to have been roughly offsetting.
For 1981, the expectation of slow rate of growth of world trade and income, coupled with the expected decline in world trade imbalance, could exert downward pressure on the pace of expansion of banks’ external assets. This suggests that the growth rate of banks’ gross external assets in 1981 could decline by 5–6 percentage points from the 20 per cent recorded for 1980. This projection relates to the growth in banks’ gross external assets, including certain interbank claims (re-depositing among the banks), which on average account for some 40 per cent of the total. It has been observed that this interbank segment grows more rapidly during periods of speculation against the U.S. dollar,87 and it may be that the opposite result would obtain during periods when there is upward exchange market pressure on the dollar vis-à-vis other major currencies. During the first half of 1981 the exchange markets have increased strongly in favor of the dollar. If this has not been reversed by the latter part of the year, the interbank segment of banks’ external assets could grow relatively more slowly during the year than the segment representing claims on final borrowers. Thus, the rate of increase in net external assets, which excludes interbank transactions and therefore is a better measure of international lending to final borrowers, might well be higher.
Determinants of International Bank Lending
Increases in the growth rates of world trade and income, as well as the size of the world’s trade imbalance are all believed to have a positive influence on international bank lending. The demand for external bank loans tends to increase along with world imports because trade financing is often provided to the buyer by a bank located abroad. On the other hand, for a given level of world imports, international lending will be larger the more imbalanced trade is between countries. Larger trade deficits—or more generally, current account deficits—may increase demand for international bank credit, while larger current account surpluses are likely to result in larger official deposits with international banks.
An increase in wealth allows economic agents to expand their assets and liabilities, including international bank deposits and loans. Large increases in the price of oil and other primary commodities during 1973–74 led asset holders in many countries to re-evaluate their wealth and to increase their external public indebtedness accordingly, in anticipation of a future income stream that was larger. Also, the supply of international credit by banks can be assumed to reflect demand for deposits by bank customers, which one would expect to be positively related to the volume of wealth and income.
The rate of change of international bank credit can thus be expressed as
R(Q) = αR(M) + βR(I) + γR(W) + u
where the R-operator stands for growth rate, Q for international bank assets, M for imports, I for trade imbalance, and W for wealth and where α, β, and γ are each expected to have a positive sign. The degree of world trade imbalance is measured as the sum of the trade deficits and surpluses of individual countries in absolute value, divided by the level of world imports.
Changes in interest rates are not explicitly included in the above equation. The negative impact on international lending of a simultaneous rise in all major interest rates would, however, be captured in its negative impact on nominal wealth and world trade.88 An opposite effect arises from the fact that higher interest rates encourage lending and deposit taking to move through international channels, which are generally less subject to controls—such as interest-free reserve requirements and interest rate ceilings on bank deposits.
The major growth rate of international bank claims was estimated by using quarterly data for 1970–79. Most of the explanatory variables discussed above are not readily observable, so that proxies had to be constructed.
GRQ, the dependent variable, is the quarterly growth rate of gross external assets, expressed in U.S. dollars, of deposit banks in nine European countries, Canada, Japan, and the United States. The average value for 1970–78 was 5.96 per cent, which corresponds to an annual compound rate of growth of 26 per cent. GRWIMP is the quarterly growth rate of world imports also expressed in dollars. The average value during 1970–78 was 4.73 per cent, or 20 per cent on an annual basis. GRSUM is the quarterly growth rate of the sum of GNP of the United States, Canada, Japan, the Federal Republic of Germany, and the United Kingdom, all expressed in U.S. dollars. This variable is used as a proxy for the growth rate of wealth. Data limitations made it impossible to consider the GNP of a larger group of countries on a quarterly basis. The trade imbalance is the ratio of the sum of the absolute values of trade deficits and surpluses of 56 countries to world imports. The quarterly growth rate of this ratio (GRIMB), which has fluctuated between minus 15 per cent and 62 per cent during 1970–79, is entered as an explanatory variable.
Using the ordinary-least-squares method, the following equation was estimated for the period from the second quarter of 1970 through the fourth quarter of 1979.89
GRQ = α0 + α1GRWIMP + α2GRSUM + α3GRIMB
with the following results:
|D-W = 1.91 R2 = 0.74|
The explanatory variables are significantly different from zero at the 95 per cent confidence level. This would seem to lend empirical support to the above hypotheses. Chart 9 depicts actual and estimated values of the quarterly growth rates.
Chart 9.Quarterly Growth Rate of External Bank Assets1
1 Dotted line represents estimated data for the period from the second quarter of 1970 through the fourth quarter of 1979 and forecasts for the calendar year 1980.
The positive constant term could be interpreted as reflecting the discovery of the advantages of international bank deposits and loans by some new borrowers, depositors (including central banks), or financial intermediaries in terms of higher risk-adjusted returns or lower costs. On the other hand, this coefficient could reflect a shift by developing countries toward preferring resource transfer through the intermediation of foreign banks over direct foreign investment. Also, the advent of greater foreign exchange rate flexibility in the 1970s has boosted demand for forward foreign exchange contracts; it is likely that this has stimulated the growth of the Eurocurrency market, since banks tend to offset the exchange risk they assume when selling such contracts to their customers by borrowing in the interbank market.
The coefficients on the three economic variables may be interpreted as implying that a 1 per cent growth in world imports and GNP causes international bank assets to increase by 0.17 per cent and 0.78 per cent, respectively, while a 1 per cent increase in the percentage of trade imbalance causes international bank assets to increase by 0.13 per cent. These findings are consistent with the fact that the market for international bank credit has two major roles: the provision of financing for trade and the provision of a channel for portfolio diversification by investors. The amount of financing needed for trade is influenced both by the level of world trade and by the degree of trade imbalance. It would thus be expected that measures of both these variables (in terms of rates of change since rates of change are being explained) will have a positive effect on the size of the market. Regarding portfolio effects, market size is likely to be positively related to world nominal wealth.
Analysis of 1980 and Projections for 1981
Out-of-sample forecasts for 1980 were not very accurate on a quarterly basis (see Chart 9), but were reasonably accurate when cumulated into an annual total. Growth during the second quarter was severely underestimated, probably because the proxy variable for wealth is so strongly influenced by quarterly changes in GNP of the United States. Compounding the four estimated quarterly growth rates, however, resulted in an annual rate of change of 21.7 per cent, compared with the actual rate of 20.1 per cent.
No forecasts have been made for the individual quarters of 1981 because of the unavailability of reliable forecasts of the growth rates of world trade and trade imbalance on a quarterly basis. It may be useful, though, to examine the available forecasts for the annual growth rates of these variables. The annual growth rate of total nominal GNP of the five major industrial countries is estimated at 8.7 per cent in 1981 (using average exchange rates prevailing during the first quarter of 1981), compared with 9.6 per cent in 1980. World trade expressed in nominal U.S. dollars is expected to increase by only 8 per cent, compared with 21.5 per cent in 1980. Finally, the trade imbalance variable is expected to decline to 21.5 per cent in 1981, after increasing to 22.3 per cent in 1980. One reason is that the surplus of oil exporting countries is expected to diminish slightly and that the combined deficits of industrial countries and non-oil developing countries are also expected to decline, while nominal world imports, which function as a scale variable, continue to increase.90
Given the above forecasts, the annual growth rate of gross international bank assets is projected to be 14.2 per cent for the year 1981, compared with 20.1 per cent in 1980.91 If the above projection proves accurate, 1981 would register the lowest growth rate since 1965, the first year for which data are available. The decline with respect to 1980 would be partly due to a lessening of the degree of real trade imbalance, as defined for the purposes of this Appendix, but more importantly, would be due to the effect of the appreciation of the U.S. dollar in terms of other currencies on world imports and GNP.
This paper has focused on international bank lending net of interbank transactions. In 1980 the average annual rate of growth of net international bank assets was 21.8 per cent, which compares with a rate of increase of 20.1 per cent for the gross series. The latter is expected to increase by about 14 per cent in 1981. But, as the share of interbank transactions in the total is quite volatile and tends to decrease when the U.S. dollar appreciates in the foreign exchange markets, the net series could grow more rapidly than gross international assets in 1981, possibly at about 18–19 per cent. This projection incorporates the valuation effect of the appreciation of the U.S. dollar in the early months of 1981; adjusted for such a valuation effect, the growth rate should exceed 20 per cent, compared with 25 per cent in 1980.