Increasing borrowing flexibility and reducing costs
Thomas Hoopengardner and Ines Garcia-Thoumi
The World Bank Group (which includes the International Bank for Reconstruction and Development, the International Development Association, and the International Finance Corporation) made commitments of over $15,300 million to its developing country members in fiscal year 1983 (which ended on June 30, 1983). Of this total, over $11,100 million came from the Bank itself. In sharp contrast to bilateral and some multilateral development institutions, the Bank receives only about 5 percent of its funds from the governments of high-income countries. The balance must be borrowed on competitive terms in world financial markets. In fiscal year 1983, for example, it borrowed the equivalent of nearly $10,300 million.
The fact that the Bank borrows the bulk of its funds means that its financial position is under close and constant scrutiny. Since 1981, the Bank has made several notable changes in its financial structure and operations to assure its continuing financial strength in changing circumstances. In January 1982, it began to levy a front-end fee on loans, in addition to the commitment charge and lending rate arrangements already in effect; in July 1982, it made a fundamental change in the way the interest rate on loans was set; and finally, in September 1982, its Executive Directors authorized selling short-term securities to provide an additional source of funds. (It has also instituted currency swaps; these are discussed in the accompanying article by Christine Wallich.)
There were three reasons for these changes. First, the financial markets on which the Bank depends have been much more volatile in recent years than they had been during the first 30 years of its existence. Second, at the same time, the Bank’s vulnerability to this volatility was increasing because of greater exposure to changing interest rates. Finally, its borrowing requirements were growing, and it needed additional flexibility to secure funds at terms favorable to its own borrowers.
Long-term interest rates tended to rise in the 1970s, but changes were rather gradual (Chart 1). Between 1972 and 1978, year-to-year changes of 1 percentage point or more in the long-term U.S. Treasury bond rate were unusual, and over the entire seven-year period the rate stayed within about a 3 percentage point band. Since 1979, long-term rates have been higher, much more volatile, and less predictable as monetary policies of the major industrial countries tightened and uncertainties developed about the future course of inflation. Changes of 1 percentage point in one month have not been unusual and changes of 3–4 percentage points have occurred several times within 12-month periods. Long-term interest rates have recently seemed somewhat less volatile, but there is no assurance that this indicates a return to the earlier stable rates. The impact of the changed financial environment is that the Bank’s borrowing costs are now perceived to be much more volatile and less predictable than they had been previously.
Chart 1Average yields of long-term U.S. Treasury bonds1
Source U.S. Department of Treasury. Treasury Bulletin, various issues.
1 U.S. Treasury series is the 20-year constant maturity rate read from daily Treasury yield curves and averaged for the month.
Meanwhile, the Bank’s ability to withstand rapid changes in interest rates was deteriorating, as a result of two interrelated factors: the simultaneous rapid growth in its commitments created greater risks from lags, while the declining maturity of borrowings and the decreasing importance of paid-in capital both created greater risks from maturity mismatching.
When the Bank makes a loan, the funds do not change hands immediately, but are disbursed as a project is implemented. Under the old lending rate system, the Bank made a firm commitment at the time the loan was agreed to make these funds available to borrowers at a fixed interest rate. However, most funds were not borrowed by the Bank until just before they were needed. If market interest rates rose between the time the lending rate on a loan was agreed and the time disbursements were made, the Bank could incur a loss—resulting from lag risk. (Conversely, it could gain should the interest rate fall over the relevant period, but this did not happen since rates were rising steadily over the 1970s.) If losses from such lag risks were incurred on many loans, net income could fall and borrowing costs would rise even further in response to market perceptions of the Bank’s financial health. The ultimate losers would be the Bank’s borrowers.
Maturity mismatch risk arose because the Bank tended to “borrow medium” but “invest short” and “lend long.” It borrowed primarily four currencies—deutsche mark, Japanese yen, Swiss francs, and U.S. dollars—medium-term, from five to ten years. Its loan disbursements tended to be in deutsche mark, yen, and Swiss francs, generally at longer maturities than the borrowings underlying them. Meanwhile, the Bank tended to rely on its dollar borrowings as a source of funds for its pool of liquid cash and securities, which have much shorter maturities than the borrowings underlying them.
This pattern of exposure meant that if interest rates rose, the average gains on short-term dollar investments increased faster than the average cost of the existing underlying medium-term borrowings. Thus, in the short run, net income would tend to rise. In the medium term, the cost of borrowings would catch up with the higher return on investments, but at the same time this cost would rise faster than the return on the longer-gestating loans, so net income would fall in the medium term. In the very long term the old fixed-rate loans would be repaid, new loans at higher fixed interest rates would be disbursed, and net income would be restored.
Lag risk and maturity mismatch risk and their evolution over time are summarized in the “gap curve” graphs in Chart 2, which measure the extent to which interest rate “fixity” disappears from one side of the balance sheet before it disappears from the other. Note the pattern: first interest rate fixity vanishes from the assets side as the Bank’s short-term securities mature, then from the liabilities side as its medium-term borrowings are retired, and finally from the asset side once again as loans are eventually repaid. This pattern of exposure is a direct and inevitable consequence of “borrowing medium” while “investing short” and “lending long.” A rough idea of the impact on future net income of a 1 percent change in interest rates may be obtained simply by multiplying the exposure shown on a gap curve by 1 percent. A 1 percent interest rate increase at the end of fiscal year 1981 could reduce fiscal year 1987 income by about $180 million, or by about 25 percent. Larger changes in interest rates would have had a proportionately greater impact on income.
Chart 2The Bank’s interest risk exposure1
Source: World Bank data.
1 Actual fiscal years 1971, 1976, and 1981; hypothetical fiscal year 1986.
2 Assuming no policy change by the Bank.
The front-end fee
At the beginning of the 1980s, it was clear that by mid-decade exposure to interest rate risk would be so great that it might raise serious concerns about the Bank’s financial well-being. Greater flexibility in loan charges was desirable in order to deal with this structural problem. The first tool chosen to achieve this was the front-end fee.
Originally 1.5 percent of the loan amount, but subsequently reduced in two steps to 0.25 percent, the front-end fee is a one-time-only charge levied when new loans become effective. It may be paid in cash, or it may be “capitalized”—added to the loan balance—and repaid on the Bank’s usual lending terms. If it is capitalized, it draws interest at the prevailing Bank lending rate. The front-end fee added about $40 million to net income in fiscal year 1982 and about $100 million in fiscal year 1983.
With the front-end fee, net operating income (net income plus contributions to special programs) rose from $610 million in fiscal year 1981 to $620 million in fiscal year 1982 and $776 million in fiscal year 1983. Without it, net operating income would have declined slightly in fiscal year 1982 to about $580 million, but the increase in fiscal year 1983 would still have been substantial—to about $675 million. However, it might be argued that net operating income was propelled to record levels in fiscal year 1983 less by the front-end fee than by factors that could not have been anticipated. From August 1982, long-term dollar interest rates began a rapid descent that in six months erased about half the increases of the preceding three years. Short-term rates fell even further, and approval of the short-term borrowing program (discussed later) enabled additional savings in borrowing costs. Swaps of Swiss francs for dollars further reduced the nominal cost of borrowing. The sharp decline in interest rates resulted in capital gains on investments that also boosted net income.
In reaction to the better-than-expected income in fiscal year 1983 the front-end fee was cut in half to 0.75 percent nine months after it was applied, and it was cut again to 0.25 percent ten weeks after that. However, the most difficult years for net income may well lie ahead. Interest rates remain much higher than they were during the 1970s. Net operating income in fiscal year 1984 is likely to be lower than in any year since fiscal year 1979 and prospects for fiscal years 1985 and 1986 are unclear. It would be premature to eliminate the front-end fee entirely, and it may even become necessary to raise it once again.
Variable lending rate system
The front-end fee could effect rapid but rather small changes in net income. However, it could do nothing to reduce the growth in underlying interest rate risk exposure. This problem was becoming so large by the early 1980s that the only feasible reaction was a fundamental change in loan pricing. The change agreed was to the pool-based variable lending rate system.
Under this pool-based system, the lending rate is adjusted every January 1 and July 1. For a loan in the system, the new rate prevails for the six-month period starting with the next billing date after the lending rate is changed. The lending rate is set by a formula to reflect the Bank’s average borrowing cost over the preceding six months. The formula adds 0.5 percent to the cost of qualified borrowings, which include all borrowings settled after July 1, 1982, plus, until June 30, 1985, a “proxy borrowing” summarizing borrowing activity in fiscal year 1982. The proxy borrowing was included to provide a basis for the initial lending rate and to stabilize the lending rate in the first three years while the amount of qualified borrowings builds up.
The initial rate under this formula was 11.43 percent from July 1, 1982, but declined to 10.97 percent on January 1, 1983 and to 10.47 percent on July 1, 1983. It declined further to 10.08 percent on January 1, 1984. These reductions apply to all loans in the system, not just those committed after the changes are made. In contrast, loans made just before the new system began will continue to bear an 11.6 percent rate until they are repaid.
When most disbursed and outstanding loan balances fall within the new system, lag risk and that part of maturity mismatch risk resulting from borrowing medium but lending long term will be eliminated. It will then no longer be necessary for the costs of rising interest rates to be passed through to borrowers via large, abrupt changes in the front-end fee imposed on loans committed at the time. Rather, market interest rate movements will be reflected in small, gradual changes in the lending rate; borrowers will be affected in proportion to their loan balances.
The new system is not perfect. First, that part of the maturity mismatch resulting from borrowing medium but investing short will remain. In fact, with loans perfectly matched against debt, investments will be perfectly mismatched against equity, since as interest rates fluctuate they will not be offset—the rate on equity being zero. Second, the Bank’s pool-based variable lending rate will adjust only slowly to changes in market rates. This means that from time to time market rates may diverge substantially from the Bank’s lending rate. Nevertheless, the new system will eventually improve the most critical part of the Bank’s asset/liability management problem in an equitable fashion.
Short-term borrowing program
Both the front-end fee and the variable lending rate system permit the Bank to pass through to borrowers the consequences of adverse interest rate movements. But neither gives it room to maneuver to protect its borrowers from temporarily high interest rates. As the Bank’s annual borrowing requirements have expanded, this room to maneuver has become increasingly important.
As Chart 3 shows, the Bank’s borrowing more than doubled between fiscal years 1979 and 1982, but continues to be concentrated in dollars, deutsche mark, Swiss francs, and yen. A comparison of Charts 1 and 3 confirms that the Bank was driven to borrow rapidly increasing amounts of dollars at precisely the time dollar interest rates were exploding. Because the short-term borrowing program had not yet been authorized, extremely high interest rates had to be locked in for extended periods. In 1981, for example, the Bank borrowed medium-term dollars at about 15 percent. This kind of borrowing might have been avoidable, if the short-term borrowing program had been in place.
Chart 3Currency distribution of Bank borrowing, fiscal years 1971–83
Source: World Bank data.
The Central Bank Facility (CBF)
The recent changes in the Bank’s financial policies have made it possible to offer central banks a more flexible instrument that is specifically designed to satisfy their requirements. To appeal to the bulk of central bank reserve holdings, the CBF—established in November 1983—offers dollar-denominated investments with a one-year maturity. To provide flexibility in timing, new investments are accepted daily by the Bank, at its discretion, subject to a minimum principal amount of $5 million and a maximum of $200 million. Liquidity is provided by guaranteeing prepayment of funds, with accrued interest at par and no penalty, upon two days’ notice to the Bank. The prepayment option is designed to meet the operational needs of the central banks rather than to be used for reinvestment or arbitrage. The yield to the investor is based on the rate on one-year U.S. Treasury bills on a coupon equivalent basis published by the U.S. Federal Reserve, plus a 30 basis point spread. The interest rate on investments in the Facility is adjusted on a monthly basis.
The initial size of the CBF is limited to $750 million. However, to provide management flexibility, the amount outstanding in it may exceed $750 million at any time by the amount of investments in the Facility due to mature within 30 days.
In the past, the Bank satisfied its borrowing requirements by securing medium- and long-term funds at fixed interest rates. In September 1982, the Board of Directors authorized a program of short-term discount notes and interest-bearing notes, both with maturities of one year or less, to be issued in the U.S. market. Initially the amount of this form of indebtedness was limited to $1,500 million, but the ceiling was raised to $2,000 million in April 1983. The program reached its ceiling within the first month and has remained near it almost continuously.
Thus far, all short-term borrowings have been in the form of discount notes, so called because they sell at a discount from their values at maturity. The interest rate earned is determined by the size of the discount. For example, a three-month discount note purchased for $980,000 with a maturity value of $1,000,000 would yield about 8.42 percent.
In fiscal year 1982 the cost of the discount notes averaged about 2 1/2 percentage points less than the cost of intermediate-term U.S. dollar borrowings that were planned if short-term borrowings had not been approved. As a result, the discount note program saved about $40 million in borrowing costs in fiscal year 1982 and added approximately that amount to income. This permitted a more rapid reduction in the front-end fee than would otherwise have been prudent.
Although the short-term borrowing program has achieved its initial objectives, its limited size has been a constraint on its effectiveness. It has not been possible to offer investors a broad maturity spectrum, nor has it been possible to maintain a daily presence in the short-term market. To achieve these objectives a program of at least $6,000 million would be desirable. The ideal situation from the borrowers’ perspective might be a pool of qualified borrowings whose cost was responsive to decreases in interest rates but insensitive to increases in rates. The way to achieve this asymmetry would be for the Bank to vary the ratio of short-term to total debt depending on interest rate expectations, and this would also require an expansion in the maximum permitted size of the short-term borrowing program.
The Bank’s changes in financial policy have significantly increased its flexibility and reduced its vulnerability to financial turbulence. The front-end fee permits small, short-term adjustments in net income when needed. The new pool-based variable lending rate system has reversed the growth of the Bank’s exposure to interest rate risk, and it has provided a mechanism for transmitting changes in borrowing costs through to borrowers in a fair way. The short-term borrowing program has created a badly needed alternative to locking in unfavorable interest rates for extended periods, although it remains constrained by small size. These changes have potentially strengthened the Bank’s ability to weather financial storms and to take advantage of opportunities on behalf of its borrowers.