Interest rates, credit, and economic adjustment in Nicaragua

POLICY RESEARCH WORKING PAPER 1529 Interest Rates, Credit, Nicaragua's dollar-equivalent and real interest rates are not and Economic Adjustment unusually high by regional in N icaragua standards. A sustained reduction of interest rates below the regional average Ulricb Ldcbler may be possible, but would require further major structural reform. The World Bank Latin America and the Caribbean Region Country Department II Country Operations Division November 1995  POLICY RESEARCH WORKING PAPER 1529 Summary findings The high commercial lending rates Nicaragua is currently rates in other Central American countries. These high real experiencing, together with a perceived scarcity of credit, rates are attributable entirely to a real currency have often been blamed for the country's slow growth depreciation that has been taking place since 1992, and and have been considered a major failing of the are not greatly different from rates observed in other Latin adjustment program initiated in 1991. American countries that underwent similar adjustments. Lachler suggests that such blame is largely misplaced. Lachler explains the link between real interest rates Current interest rates are indeed higher than historical and adjustment in Nicaragua and, in that context, levels or international benchmark rates (such as LIBOR explores policy options for reducing interest rates. or the U.S. treasury bill rate), but those are not the His main conclusion: A sustained reduction in real appropriate comparators for Nicaragua today. interest rates to below those observed in neighboring On the other hand, Nicaragua's real interest rates have countries would require further major structural changes, risen significantly in recent years and currently exceed real such as the adoption of a foreign currency standard. This paper - a product of the Country Operations Division, Country Department 11, Latin America and the Caribbean Region - is a self-standing report prepared as a contribution to the Bank's ongoing policy dialogue with Nicaragua on important economic issues facing the country. Copies of this paper are available free from the World Bank, 1818 H Street NW, Washington, DC 20433. Please contact Gerry Carter, room 14-308, telephone 202-473-0603, fax 202-676-1464, Internet address gcarter@worldbank.org. November 1995. (30 pages) The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the excbange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be used and cited accordingly. The findings, interpretations, and conclusions are the authors' own and should not be attributed to the World Bank, its Executiv/e Board of Directors, or any of its member countries. Produced by the Policy Research Dissemination Center  INTEREST RATES, CREDIT AND ECONOMIC ADJUSTMENT IN NICARAGUA Ulrich LAchler The World Bank Country Department II Latin America and Caribbean Region   INTEREST RATES, CREDIT AND ECONOMIC ADJUSTMENT IN NICARAGUA Ulrich Uchler' The World Bank CONTENTS I. Background II. The Evolution of Interest Rates since 1991 Dollar-Equivalent Rates Basic Decomposition of Risk Real Interest Rates III. Real Interest and Exchange Rates IV. Why Do High Real Interest Rates Constitute a Problem? V. Alternative Policy Options for Reducing Interest Rates Reverse the Process of Financial Liberalization? Reverse the Stabilization Process? Maintain Current Policy Settings Deepen Financial Sector Reforms VI. Concluding Summary References Annexes Annex A: A Theoretical Framework for Understanding Real Exchange Rate Determination Annex B: Data Sources and Definitions Annex C: Consolidated Balance Sheets and Income Statements of the Nicaraguan Banking System Annex D: Lending Rates in South America & Mexico in 1992 1. The author is grateful to Mario Flores, Armando Navarrete and Pablo Miranda of the Central Bank of Nicaragua for their help in obtaining the data needed for this report. Thanks are also due to Joe Ryan, Julio E. Revilla, Armando Caceres, Gustavo Arcia and Noel Sacasa for their thoughtful comments on earlier drafts.  I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  INTEREST RATES, CREDIT AND ECONOMIC ADJUSTMENT IN NICARAGUA 1. The high commercial lending rates currently observed in Nicaragua, together with a perceived scarcity of credit, have been frequently blamed for the country's slow growth and pointed out as a major failing of the adjustment program initiated in 1991. This essay suggests that such blame is largely misplaced. While current interest rates are indeed high when compared with historical levels or international benchmark rates such as LIBOR or the US Treasury Bill rate, these rates are not the appropriate comparators for Nicaragua today. A look at rates prevailing elsewhere in Latin America reveals that Nicaragua's dollar-equivalent interest rates are not unusually high. Nicaragua's real interest rates, on the other hand, have risen significantly in recent years and are currently above real rates in other Central American countries. These high real rates are entirely explained by a real currency depreciation that has been taking place since 1992 and are also not unusual compared to the rates observed in other Latin American countries that experienced similar adjustments. This essay seeks to explain the link between real interest rates and the adjustment process in Nicaragua, and in that context discuss alternative policy options for reducing interest rates. The main conclusion of this analysis is that a sustained reduction in real interest rates below the rates observed in neighboring countries would require further major structural changes, such as the adoption of a foreign currency standard. I. Background 2. Interest rates have risen throughout Latin America since the onset of the debt crisis in the early 1980s and subsequent liberalization of financial markets. Many factors contributed to this phenomenon: adverse terms of trade shocks and rising intemational interest rates, combined with overambitious public spending programs, led to a rapid expansion of fiscal deficits that initially were financed with foreign borrowings. As foreign borrowing opportunities dried up, rampant inflation and dollarization ensued. Attempts to stem this tide through interest rate and capital controls discouraged domestic financial savings and encouraged capital flight. Repressive financial policies, sometimes accompanied by the nationalization of banks, left financial institutions severely weakened. At the same time, financial innovations greatly increased the mobility of capital across national frontiers, so that the controls previously applied to maintain interest rates artificially low became largely ineffective. As these events unfolded, interest rates increased because (i) savers, once hurt by inflation, subsequently demanded higher returns to compensate for the risk of renewed instability, (ii) interest controls could not be maintained lest they encourage further capital flight, and (iii) the cost of intermediating credit increased as financial institutions became weaker. Nicaragua's financial system also evolved according to this stylized pattern, with additional distortions induced by civil war and centralized planning. By the end of the 1980s, it had experienced one of the worst hyperinflations observed in Latin America and its nationalized financial system was bankrupted. 3. In March 1991, the Government of Nicaragua initiated a stabilization and market- liberalizing adjustment program designed to promote faster output and export growth. This program has been very successful in bringing down inflation and arresting the continuing decline of GDP that took place after 1983. A key element of this program is the liberalization of the financial sector, which included (i) the unification of the exchange rate system into an official market for current account transactions and a parallel market for financial and non-government service account transactions, (ii) progressive removal of interest rate controls, which were  2 completely eliminated by 1993, and (iii) eliminating the practice of directing credit to specific subsectors. The monopoly on banking services held by the state-owned banks was abolished in 1991 and, by the end of 1994, nine private commercial banks had begun to operate. 4. With the liberalization of the financial sector, active markets developed for loans and deposits in both US Dollars and Cordobas. All Cordoba transactions except sight deposits, however, continue to include a "maintenance of value" provision, whereby the rate of return in local currency is indexed to the official exchange rate. Even though the stabilization program succeeded in bringing down inflation rates near single digit figures and in maintaining the spread between the parallel and official exchange rates below 5 percent, the market has continued to express a general preference for indexed transactions. II. The Evolution of Interest Rates since 1991 5. A discussion of interest rates, especially in a comparative context, only makes sense when their nominal levels are adjusted for differences in inflation or in the rate of currency depreciation. Accordingly, this analysis focuses on dollar-equivalent and real interest rates, using a decomposition method applied by Rodriguez (1994): for any nominal lending rate, IA, the corresponding ex post dollar-equivalent rate is given by IADOL s IA - DEV, where DEV stands for the rate of devaluation of the domestic currency.' The real interest rate, in turn, is derived as LIREAL - IA - INF, where INF represents the rate of domestic inflation. In the case of Nicaragua, since virtually all loans and deposits are indexed to the US Dollar, interest rates are automatically quoted as ex post dollar-equivalent rates based on the official exchange rate. To calculate real interest rates in that case first requires that the dollar-equivalent rate be transformed into nominal Cordoba rates (using the official devaluation rate) and then deflated by the domestic inflation rate. Dollar-Equivalent Interest Rates 6. The top half of Taible 1 compares the average dollar-equivalent lending rates in Nicaragua with those prevailing in the other Central American countries during 1991-94. These lending rates refer to short term loans with maturities of up to one-year. As described in Annex B, these lending rates do not always refer to the same loan modality or aggregation procedure. For some countries they refer to weighted-average rates and for others to arithmetic averages. In spite of these differences, however, meaningful comparisons across countries can still be made. Based on Table 1, Nicaragua's dollar-equivalent lending rates do not appear to be systematically higher than those observed in the other countries in the region. Although somewhat higher than the Central American average in 1991-93, Nicaragua's rates are clearly below the average in the first half of 1994. In this sample, only Honduras stands out with very negative interest rates in 1993-94, suggesting the onset of disorderly macroeconomic 1. This derivation is an apprximation of the corect conversion formula, 1+IADOL = (1+IA)/(1+DEV). Applying the natural log operator to both sides of the equation, this expression converges to IADOL = IA-DEV, as IA and DEV approach zero; since lim Ln(1 +x) = x as x- >0. The same approximation method also applies to the calculation of real interest rates, 1+IAREAL = (1+IA)/(1+INF). The dollar-equivalent and real interest rates presented here are aU based on these approximations in order to permit comparisons with the results obtained for other countries by Rodriguez (1994), who used this approximation method to decompose interest rates into naturally additive spreads. To maintain appmximation errors within acceptable bounds, al interest rates are first transformed into monthly rates before applying the approximation procedure and then reconverted into annual rates.  3 adjustments. Excluding Honduras from the sample, the average lending rate over the three year period covered in Thble 1 is 19.8 percent for the region and 20.5 percent for Nicaragua. TABLE 1 Average Lending Rates in Central America (on Local Currency Loans) 19911 1 1992 1 1993 19942 Dollar Equivalent Lending Rates (in percent per annum) Costa Rica 25.3 30.2 19.9 26.1 El Salvador 15.1 8.7 19.8 18.6 Guatemala 18.5 15.2 14.7 30.9 Honduras 19.4 12.3 -1.0 -14.5 NICARAGUA 22.3 22.7 20.5 14.1 Real Lending Rates (in percent per annum) Costa Rica 14.2 17.3 19.3 15.1 El Salvador 16.8 -3.8 6.4 9.8 Guatemala 15.1 5.7 9.7 14.3 Honduras 14.0 13.9 4.4 -7.7 NICARAGUA 3.6 20.5 23.2 21.4 1 Rates for 1991 refer to July through December. 2 Rates for 1994 refer to January through June. Source: See Annex B. 7. Using a sample of six South American countries and Mexico,2 Rodriguez (1994) calculated the average dollar-equivalent lending rate for local-currency operations to be 35 percent in 1992 (or 32 percent when Peru is excluded from the sample). These rates appear significantly higher than those reported on average for Central America. The author also notes, however, that a dollarization process has been taking place in these countries (similar to that observed in Nicaragua), which has reduced significantly the relevance of unindexed local- currency transactions. In the case of Argentina, for example, he calculated that only 43 percent of credit transactions in 1992 were made in unindexed local currency (at an average lending rate 2. The countries included in Rodriguez' (1994) sample are: Argentina, Bolivia, Chile, Colombia, Mexico, Peru and Uruguay. In 1992, Peru experienced sharp exchange rate fluctuations that temporarily resulted in highly negative interest rates that distort the sample average interest rates for that year. lb account for this distortion, the average interest rates are presented both with and without Peru.  4 of 36 percent), while 57 percent of credits were indexed or dollar-denominated (at an average rate of 13 percent), yielding a weighted-average, dollar-equivalent lending rate of 23 percent. When all sources of credit are included in this manner, Rodriguez (1994, pg. 22) estimates that the average dollar-equivalent lending rates in his country sample was close to 21 percent in 1992. This last figure approximates the average rates calculated for Central America.3 TABLE 2 Average Dollar-Equivalent Borrowing Rates in Central America (on Local Currency Deposits) l 19911 1992 1993 19942 (in percent per annum) Costa Rica: 1 month dep. 9.4 17.1 7.2 12.8 3 month cert. 14.5 18.8 10.0 14.6 El Salvador: 2 month dep. 9.6 4.0 15.0 12.9 Guatemala: 1 month dep. 10.5 6.4 3.4 17.2 Honduras: 1 month dep. 6.8 1.0 -11.6 -24.6 3 month cert. 8.9 6.8 -8.0 -20.5 Nicaragua: 1 month dep. 16.1 14.9 12.2 5.9 XRates for 1991 refer to July through December. 2 Rates for 1994 refer to January through June. Source: See Annex B. 8. A similar observation also applies to dollar-equivalent borrowing rates, which are shown in Tible 2. The average dollar-equivalent interest rate on 1-month savings deposits in Nicaragua is 12.7 percent, which is within the range of rates observed in other Central American countries; again excepting Honduras. Nicaragua's borrowing rates are highest in 1991 and then decline over time. As described in the next section, this interest rate behavior can be attributed to declining risk premia. The average rate calculated by Rodriguez for South America and Mexico in 1992 is 11.3 percent (excluding Peru). These rates are close enough to reaffirm the earlier finding that interest rates in Nicaragua are not unusually high by regional standards. 9. The preceding finding -- that dollar-equivalent interest rates appear to be roughly similar across the region -- should not come as a surprise in view of the liberalized conditions that currently characterize most financial markets in Latin America. As predicted by the interest parity hypothesis, under conditions of sufficient capital mobility, interest rate arbitrage equates 3. Data on the extent of dollarization in the Central American countries is very limited. In Nicaragua, about 50 percent of an commercial bank deposits are denominated in US Dollars, but only 19 percent of loans are dollar-denominatd. Lending rates in late 1994 on these dollar-loans are quoted at 14-16 percent. Based on this information, the weighted average dollar-equivalent lending rate in Nicaragua for 1994 (Jan-Sep) is about 18.9 percent.  5 domestic interest rates to the corresponding "world" interest rates plus the expected rate of devaluation and a premium to compensate for differences in country and currency risk. Therefore, since capital controls have been largely eliminated and assuming that exchange rate expectations are not systematically biased, Nicaragua's dollar-equivalent interest rates should on average turn out similar to the rates prevailing in neighboring countries, except for differences in risk. In view of Nicaragua's reputation as an extremely high-risk country (Euromoney, September 1994), perhaps the most surprising finding so far is that Nicaragua's dollar-equivalent rates do not appear to be significantly higher than those observed elsewhere in the region. Basic Decomposition of Risk 10. This subsection identifies two elements of risk embodied in the interest rate. As described by Rodriguez (1994) and shown in TFble 3, lending rates can be additively decomposed into the following components: Dollar-Equiv. Lending Rate = Lending Spread + Borrowing Rate = I-ending Spread + Borrowing Spread + TBILL, where TBILL denotes the interest rate on (3-month) US Treasury Bills, which serves as a benchmark to calculate risk premia. The coexistence in Nicaragua of dollar-denominated and Cordoba-denominated deposits permits a further decomposition of the Borrowing Spread into two sources of risk as: Dollar-Equiv. Lending Rate = Lending Spread + CRED + RISK + TBILL, where RISK represents a measure of country risk and CRED represents a measure of credibility in the economic program, sometimes referred to as currency risk. (Increases in CRED indicate declining credibility.) 11. The variable denoted RISK is defined as the difference between the average interest rate on Dollar-denominated deposits in Nicaragua and the US Treasury Bill rate. Since Dollar deposits in Nicaragua and US Treasury Bills are both denominated in the same currency, they should command the same rate of return except for differences in risk associated with the country/institution that emits each financial instrument. Based on similar reasoning, CRED is defined as the difference between the dollar-equivalent interest rate on Cordoba-denominated deposits and the interest rate on Dollar-denominated deposits in Nicaragua. Since Cordoba- denominated term deposits in Nicaragua are indexed to the US Dollar, their interest rate should be the same as the rate on a Dollar-denominated deposit if it were certain that the Cordoba would be maintained fully convertible vis-a-vis the US Dollar. Full convertibility in this context would mean the maintenance of a unified exchange rate (or constant exchange rate spread) and the absence of exchange controls. The difference between both interest rates, therefore, measures the market's lack of confidence in policymakers' ability to maintain unified exchange rates, which ultimately depends on the ability to maintain adequate fiscal and monetary discipline. Both sources of risk have been calculated for Nicaragua with the results shown in lable 3.  6 TABLE 3 Main Determinants of the Interest Rate in Nicaragua Dollar-Equiv. Lending Borrowing Lending Rate Spread Spread TBILL l ~~~~CRED_t RISK NICARAGUA annual average rates (in %) 1991 (Jul-Dec) 22.3 6.2 10.7 0.4 5.0 1992 22.7 7.8 10.0 1.5 3.4 1993 20.5 8.3 7.0 2.2 3.0 1994 (Jan-Sep) 19.8 8.7 5.5 1.9 3.7 Source: entral Bank of Nicaragua 12. According to the preceding decomposition method, the lending rate is determined by (i) the rate on US Treasury Bills (as a proxy for international credit conditions), (ii) the market's perception of country risk, (iii) the market's confidence in the macroeconomic program, and (iv) the average lending spread of commercial banks. Thble 3 shows that among these four factors, country risk (RISK) appears to be the least significant in determining borrowing spreads in Nicaragua. Much more important is the perceived risk of destabilization, (CRED). In any case, both CRED and RISK appear to be declining in 1994, suggesting a gradual return of confidence in the Nicaraguan economy. The other important component in determining the lending rate is the lending spread. Although the spreads shown in Table 3 are not unusually high by Central and South American standards, they are high compared to those observed in other regions with competitive banking systems. Real Interest Rates 13. The lower half of Table 1 describes the real lending rates observed in Central America during 1991-94. The average real rate for the Central American region over the three-year period turns out to be 11.7 percent (or 12.8 percent when Honduras is excluded). In contrast to the earlier finding on dollar-equivalent rates, Nicaragua stands out this time by exhibiting the highest real lending rates in Central America during 1992-94; with average rates exceeding 20 percent. By way of comparison, Rodriguez (1994) calculated the average real lending rate to be 30 percent for credit transations in local currency during 1992 for his sample of Latin American countries (or 22 percent if Peru is excluded from the sample). When indexed and dollar-denominated credit sources are also included, however, he estimates the weighted-average real lending rate to be around 20 percent (or 12 percent when Peru is excluded).  7 III. Real Interest and Exchange Rates 14. The contrasting behavior of real and dollar-equivalent interest rates is entirely due to changes in the real exchange rate. Recall that the dollar-equivalent lending rate, LADOL, and the real lending rate, IAREAL, are defined as: IADOL = IA - DEV, and IAREAL = IA - INF. The difference between both rates can then be seen to represent an index of real exchange rate changes, denoted DRER: DRER = IAREAL - IADOL = DEV - INF, such that DRER > 0 indicates a real devaluation of the local currency and DRER < 0 indicates a real appreciation.4 Observe from Table 4 that real lending rates in Nicaragua quickly increase after 1991 and then exceed dollar-equivalent rates after 1992, indicating that a real devaluation has been taking place. This real devaluation coincides with the stabilization and adjustment program initiated in 1991 with the objective of promoting efficient export-led growth and reducing the country's external imbalances. Among the other Central American countries, only Honduras experienced a similar real exchange rate depreciation in 1993-94. 15. The need for structural adjustment arises when balance of payments deficits have become unsustainable and a country's growth is constrained by a lack of foreign reserves. This has been the case with Nicaragua, whose resource balance deficit in 1991-92 averaged 30 percent of GDP, while its exports had declined to less than half the value exported a decade earlier. While foreign donors have been willing to finance such high external deficits on a temporary basis, they have also indicated that Nicaragua cannot count on such aid inflows indefinitely. Due to budget constraints within the donor countries and the emergence of new claimants on donor funds (especially in the former Soviet Union and Middle East), aid flows to Nicaragua are expected to decline gradually toward the per-capita aid levels received by other low-income countries in the region.5 Under these circumstances, a real devaluation would be desirable in 4. The most common definition of the real exchange rate for analytical purposes is the price of tradables divided by the price of non-tradables. Since such price series are difficult to obtain, a common procedure is to construct a proxy for the real exchange rate as, RER = EPfIP, where E is the nominal exchange rate (Cordobas per US$), P is the US Wholesale Price Index, and P is the domestic Consumer Price Index. Since consumer price indexes include the prices of non4radables, while wholesale price indexes only contain the prices of traded goods, this proxy variable tracks changes in the relative price of tradeables to non-tradeables. Thking a proportional derivative of RER, denoted DRER, yields: DRER = DEV - INF + INF, where DEV = DE/E, INF = DP/P and INF' = DPY/P'. This formulation is the same as that in the text under the assumption that changes in the US Wholesale Price Index can be ignored, since they are negligible compared to changes in the nominal exchange rates and the domestic CPI. An important byproduct of ignoring INF, however, is that this formula yields a downwardly biased measure of real devaluation rates. The estimates of real devaluation rates in lkble 4, therefore, are useful for comparing real devaluation rates aeross countries in the region, but are less accurate for measuring the absolute amount of real devaluation taking place in any one country. S. While total aid inflaws to Nicaragua (including donations and loans) have been declining gradually, the composition of that aid has changed substantially in favor of project (tied) aid. In particular, the liquid (untied) portion of aid declined by more than half, from USS 477 million in 1992 to US$ 217 million in 1994. This aid component is the most relevant for the issues of economic adjustment addressed here, given that reductions in project aid are automatically associated with an equivalent reduction of imports (and, thus, do not immediately create an unfinanced external deficit), whereas reductions in liquid aid require deliberate policy responses to discourage imports and promote exports, or adjustments through reduced domestic absorption.  8 order to encourage the compensatory adjustments needed to close the balance of payments gap created by the aid decline. TABLE 4 Real Devaluation Rates in Central America (DRER = LAREAL - LADOL) (in percent per annum) 19911 1992 1993 19942 Costa Rica -11.1 -12.9 -0.6 -11.0 El Salvador 1.7 -12.5 -13.4 -8.8 Guatemala -3.4 -9.5 -5.0 -16.6 Honduras -5.4 1.6 5.4 6.8 NICARAGUA -18.7 -2.2 2.7 7.3 Rates for 1991 refer to July through December 2 Rates for 1994 refer to January through June Source: Own calculations based on data in Table 1. Note that negative figures indicate a real appreciation of the local currency. 16. It is also important, however, to recognize that Nicaragua does not have much choice about reducing its external deficit in the face of declining foreign aid because it does not have unlimited reserves to draw down and lacks the necessary financial creditworthiness to borrow abroad from non-concessional sources. That is, Nicaragua cannot borrow at will from international financial markets to close a balance of payments gap. Rather, this gap is mainly determined by the amount of net aid supplied by foreign donors, be it in the form of fresh disbursements or debt service relief. As the supply of aid is reduced, therefore, Nicaragua has to decrease its trade deficit, either in an orderly manner through a deliberate reduction of domestic absorption, coupled with supply-side incentives, or in a disorderly manner through an inflation tax and protectionist trade policies. Either way, the trade deficit has to come down and this invariably implies a real exchange rate devaluation; i.e., an increase in the relative price of tradeable goods versus the price of non-tradeables.6 17. Real devaluations play an important role in reestablishing external balance by encouraging the production of exportable goods and import substitutes, while discouraging the domestic consumption of both products. This shift in economic incentives is reinforced by the rise in real interest rates above dollar-equivalent rates: for producers of tradeable goods, the dollar- equivalent lending rate is of greater relevance when making investment decisions than real rates because their receivables are calculated in dollar-denominated terms with prices determined in 6. Since the prices of tradeable goods in a smaU open economy are determined in the wvrld market, domestic supply and demand conditions only serve to determine the prices of non-tradeables. Tb reduce the trade deficit by a target amount in this context, aggregate domestic absorption has to decline, which exerts downward pressure on the prices of non-tradeables and, thereby, raises the relative price of tradeables. This argument is spelled out with greater analytical rigor in Annex A.  9 world markets. Conversely, the real interest rate is more relevant for producers of non-tradeable goods, the prices of which figure more prominently in the calculation of domestic inflation. Therefore, an increase in real lending rates above dollar-equivalent rates -- reflecting the onset of a real devaluation -- renders investments in the tradeables sector relatively more attractive than investments in the non-tradeables sector. This is precisely the desired incentive pattern when seeking to promote an outward-oriented economic adjustment. 18. Any systematic divergence between real and dollar-equivalent interest rates can only be temporary, however, since that divergence is due to the rate of change, rather than IW1, of the real exchange rate. Once the real exchange rate reaches its equilibrium level, the pressure to depreciate or appreciate further is removed and, hence, both interest rates wwuld again be identical. 19. One final observation is that Nicaragua's experience with high real interest rates is not unique and not nearly as disruptive as in some other countries facing adjustment needs. In fact, the rates observed in Nicaragua appear modest compared to the truly exorbitant real rates (in excess of 70 percent) reached in Argentina or Peru at the onset of their adjustment programs in the early 1990s; see Annex D. IV. Why Do High Real Interest Rates Constitute a Problem? 20. The surge in real interest rates observed worldwide in the early 1980s has raised widespread concern about their possibly detrimental economic effects. In response to these concerns, numerous studies were carried out to measure the impact of high interest rates on key economic variables such as output growth, investment, factor productivity and relative factor returns. An empirical regularity observed in several cross-country studies (e.g., World Bank (1989) and Galbis (1993)) is that countries with higher real interest rates generally tended to exhibit faster output growth, but not higher investment rates. This finding suggests that higher interest rates discourage investment, but encourage a more efficient allocation of resources which raises overall productivity, such that the net impact on growth is positive. While other studies (e.g., Khatkhate, 1988) have questioned the empirical robustness of these findings, a basic lesson from this literature still holds, namely that higher interest rates do not automatically constitute an obstacle to growth. Rather, the impact of higher interest rates on investment and growth mainly depends on what has caused interest rates to rise in the first place.7 21. Earlier sections indicated that the increase of real interest rates observed in Nicaragua is primarily caused by two factors: one is the elimination of domestic interest rate and capital controls, which permitted Nicaragua's financial market to become reconnected to the international financial system. This, in turn, enabled Nicaragua's dollar-equivalent interest rates to reach international levels, plus a risk margin. The second important factor is the onset of a real devaluation, which coincided with the adoption of stabilization and adjustment measures needed to reduce Nicaragua's unsustainable external deficit. Since the presence of interest 7. On theoretical grounds one would not expect a simple, invariant relationship between real interest rates, investment, sAvi and growth, since all these variables are simultaneously determined by other more fundamental factors. For example, an increase in real interest rates due to a decline in financial savings on account of an expected dealuation is likely to be neptively related to investment and growth. In contrast, an increase in real rates occasioned by an investment surge triggered by the discovery of precious natural resources would be associated with higher investment and growth.  10 controls generally tends to contract the size of credit markets, their elimination should lead to an overall expansion of credit.8 On the other hand, the introduction of stabilization and adjustment measures designed to reduce the extemal deficit would be expected to restrain the expansion of credit triggered by the elimination of controls. Judging by the extremely rapid growth of private sector credit during 1991-94, this restraining force appears to have been minimal; see Figure 1. While real GDP has only grown by about 3 percent during 1991 through 1994, the volume of outstanding credit has grown in real terms by 93 percent between September 1991 and September 1994. This rapid expansion in credit took place, moreover, in spite of a major portfolio clean-up in the public commercial banks that reduced the volume of outstanding credit by 32 percent between March 1992 and June 1992. Clearly, the total availability of credit does not appear to have been a major constraint on growth. 22. The notion that higher real interest rates lead to higher productivity is only expected to apply for economies where the system of credit allocation functions efficiently in channeling savings to those activities with the highest expected rates of return. The combination of rapid credit growth, rising real interest rates and slow Evolution of Commercial Bank output growth observed in Nicaragua Credl during 1992-94, however, casts doubt on the efficiency of its credit allocation L 3000.00 mechanisms. It suggests, instead, that Q 2W0.0I) much of the expansion in credit has