Using different means, the monetary authorities in Latin America have generally regulated the interest rates paid and charged by banks and other financial institutions. Yet, despite the continuing policy efforts in this area, there are few general studies analyzing the goals and scope of such regulations, much less their effects on those economic variables that they are supposed to influence. 1 Indeed, despite the improvement in general financial statistics, no major effort has been made to reconstruct interest rate series for extended periods of time, either officially or privately. 2 Meanwhile, however, economists have formulated a number of hypotheses on the effects of interest rate policies in developing countries that apply controls on interest rates. 3 But empirical research on these matters in Latin American countries has lagged behind.
This paper attempts both to reconstruct policy measures and data on interest rates on savings instruments in 19 Latin American countries during the decade 1967–76 and to provide a preliminary assessment of the likely effects of those policies in the light of a broad theoretical financial framework. 4
From the point of view of economic theory, an interest rate may be judged to be high or low only in relation to the circumstances existing in the economy concerned. It is generally accepted that one of the most important considerations is the relationship of interest rates to inflation rates, or, more precisely, to the inflation rate expected by the public. For this purpose, Table 1 shows the rates of change in the consumer price index during the decade, together with illustrative figures on the expected rates of change of this price index, based on the adaptive or autoregressive expectations hypothesis, which relates the expected rate of inflation to past inflation rates. 5 The six-year unweighted average used here illustrates the minimum adjustment of interest rates to inflationary conditions that might be expected to be needed normally, whereas the actual inflation rate indicates the maximum adjustment. 6 However, while this range is likely to be realistic in most cases, it may not necessarily comprise the expected rate of inflation (to which interest rates should theoretically be adjusted) in the event of sharp reversals in the rate of inflation owing to exceptional circumstances. Other important considerations concern the relationship of domestic interest rates on instruments denominated in domestic currency to rates on instruments denominated in foreign currency and to foreign interest rates.
Section I presents a general picture of policies on interest rates on savings instruments that were followed in the countries covered in this study. 7Section II first discusses the extent of “financial repression,” that is, the extent to which interest rates were below the actual and the expected rates of inflation. A comparison is also made between domestic interest rates and Eurodollar deposit rates. This section also discusses the relationship between financial repression and “dollarization” and between these phenomena and capital movements. 8Section III summarizes the main conclusions.
I. General Policy Trends
By the end of 1976, in the 19 countries included in this study, the monetary authorities had the legal power to regulate interest rates in one way or another. The general tendency was toward increasing recognition of the need to delegate this power to monetary authorities and to extend it to a broader range of financial institutions.
Regarding the application of regulatory power, the tendency was toward increased willingness to experiment with new forms of policy. This willingness to experiment, however, might to a large extent have been due to the need to cope with the increasing external exposure of some economies and the accelerating rate of inflation during the 1970s. Indeed, while interest rate theory suggests that interest rates must be positive in real terms in the long run to approximate equilibrium, many of the countries included in this study continued to apply policies not conducive to that end. Furthermore, some countries did not take into account that if policy reforms were not properly instrumented, they might create difficulties for the stability of financial institutions. The following discussion distinguishes the experience with regard to discretionary controls, indexation, and interest rate freedom. A summary of the position of countries according to that classification is presented in Table 2.
|I. Discretionary Controls (17)|
|Interest Rate Ceilings on Liability Rates and Related Measures (17)|
|Partial ceilings across institution (6)||Elimination|
|Comprehensive ceilings throughout the decade (5)||Without indexation (3)||With indexation (3||Partial ceilings across instruments (3)||Introduction of ceilings (1)||of ceilings (2)||Interest Rate Floors (3)||Exceptions (2)|
|Honduras||Guatemala||Colombia||Dominican Republic||Venezuela||Guatemala||El Salvador|
|ii. indexation (6)|
|Fairly Widespread Indexation (2)||Partial Indexation (3)||Special I||Indexation (2)|
|Brazil, until 1975||Chile, until 1974||Bolivia,||since 1974|
|Chile, until 1974||Colombia, during 1972-74||Costa Rica, 1974|
|Paraguay, since 1973|
|iii. interest rate freedom (8)|
|Nonregulation (1)||Complete Liberalization (2)||Liberalization of Specific Instruments (4)||De Facto Liberalization (1)|
|El Salvador||Brazil, since 1975||Argentina2||Uruguay3|
|Chile, since 1975||Guatemala|
During most of the decade under study, the monetary authorities in 17 of the countries used their discretionary power to control interest rates. Subject to some qualifications, the two exceptions were Brazil and El Salvador. Two patterns of discretionary regulation can be distinguished during the period 1967–76. The most common pattern—one largely inconsistent with the goal of promoting positive real interest rates during periods of inflationary pressure—consisted of imposing a ceiling on interest rates paid by financial institutions. In a few cases, the authorities indirectly limited interest rates on obligations of financial institutions through interest rate ceilings on loans. A less frequent pattern that appeared during this period, in contrast to past patterns, consisted of imposing minimum rates, which invariably referred to the liability side of financial institutions. While this policy pattern is consistent with promoting positive real rates, it was in practice rendered ineffective by the simultaneous imposition of maximum limits on lending rates under inflationary pressures.
Interest rate ceilings on liabilities of financial institutions and related measures
Ceilings on liability rates were applied—although with various degrees of coverage—in 17 of the 19 countries. In 5 of them—Ecuador, Honduras, Mexico, Peru, and Uruguay—ceilings were comprehensively applied throughout the decade, whereas in the remaining countries ceilings were either combined with other policies, newly introduced, or abandoned.
Ecuador, Honduras, and Mexico were typically noninflation-ary countries before 1973, so that their interest rate ceilings were not necessarily tantamount to repressive measures. But inflation uncovered the weakness of ceiling rate policies because—although they reacted to it by raising the level of ceilings—measures typically fell short of permitting full adjustment. Mexico offers perhaps the best example of the paradigm that the coexistence of rigid control of domestic interest rates and unrestricted capital movements, in addition to discouraging domestic financial savings, can subject the economy to disturbances derived from the divergence between foreign and effective domestic yields as a result of movements in the former or expectations of changes in the exchange rate. The Mexican policy of fixing interest rates worked fairly well throughout the period of substantial internal economic stability and of the stable external environment that lasted until 1973, resulting in a pattern whereby Mexican interest rates were normally 2 to 3 percentage points above comparable U. S. rates. However, one problem was the occasional short-run departure from the appropriate differential caused by rapid interest rate movements abroad. After 1973, the combination of a two-digit rate of domestic inflation and a diminishing confidence in the value of the Mexican peso tended to create a problem of capital outflows, despite the increase in the uncovered interest rate differential in Mexico’s favor. The depreciation of the peso in September 1976 and the uncertainties associated with the floating exchange rate regime triggered further massive withdrawals of peso deposits in the last four months of 1976. 9
Interest rate ceilings were largely repressive in Peru and Uruguay, although measures taken, especially in the latter country, tended to ameliorate this situation. In Peru, ceilings on liability rates were structured in 1962 by the Central Reserve Bank within an overall constraint of 14 per cent per annum for any kind of financial operation imposed by usury legislation dating back to the beginning of the century. The structure of rates remained virtually unchanged until July 1976, and real interest rates were negative. In the face of accelerating inflation from 1973 onward, a decree of June 1976 raised the usury ceiling to 19 per cent and a central bank circular increased ceilings on liability rates by about 4 percentage points on average. 10 The further acceleration in inflation that was taking place required new measures to increase the rates in 1977 and 1978.
Uruguay—a typically inflationary country—started in 1967 from a very repressive position. Although ceilings were diversified by maturity and increased on four occasions between 1967 and 1974, the inflation rate remained high and repression continued. Then, in April 1976, at the same time that an effort was made to reduce inflation, all graduated interest rate ceilings were replaced by a uniform ceiling of 62 per cent, signaling the intention of the authorities to move toward complete liberalization. This was achieved de facto one year later when inflation continued to decline, and, in addition, the authorities raised the interest rate ceiling to 90 per cent.
In the remaining countries, ceilings on liability rates were applied to only a subset of the structure of interest rates on obligations of financial institutions or were applied only throughout part of the decade. Partial coverage arose for the following reasons: ceilings were not extended to all kinds of financial institution or financial instrument; some institutional groups or some liabilities were subject to indexing; the authorities chose to liberalize the rates on selected liabilities; or, on occasion, because of a combination of these alternative policies. Changes out of ceilings occurred either because of a move toward complete liberalization or because countries switched from ceiling to floor rates.
Partial application of interest rate ceilings across institutions, resulting from lack of regulations on some types of financial institution, was a distinctive feature in Costa Rica, Guatemala, and Nicaragua. In Costa Rica and Nicaragua, the authorities failed to regulate private investment banks (financieras) while imposing ceilings on interest rates paid by other financial institutions, but the acute inflationary pressures of the mid-1970s prompted measures to correct institutional imbalances. In 1974, the Costa Rican authorities extended ceilings to investment banks, a measure that solved the institutional problem but increased financial repression. By contrast, in 1976, the Nicara-guan authorities freed most rates of private commercial and state banks to reduce both their excessive disadvantage, compared with investment banks, and financial repression. In Guatemala, while all rates on deposits in commercial banks and savings and loan associations were subject to ceilings throughout the decade, savings deposits in state banks were subject to a minimum rate of 5 per cent until June 1972, when this rate was freed; the system was therefore designed to give an advantage to state banks over private institutions.
Partial application of interest rate ceilings with simultaneous partial indexation and/or interest rate freedom on some types of financial institution occurred in Chile, Colombia, and Paraguay. In all three countries, the traditional policy was based on interest rate ceilings, while other experiments were of more recent vintage. The coexistence of ceilings with partial indexation or interest rate freedom created difficult problems of institutional imbalance during high-inflation periods. 11
In Chile, the policy of automatic determination of ceilings by past average rates of interest was abandoned in 1966 in favor of direct determination of ceilings by the Central Bank for private financial intermediaries, but ceiling rate regulations continued to cause liability interest rates to be sharply negative in real terms. 12 Exceptions were the indexed instruments of the state-controlled system of savings and loan associations, Sistema Nacional de Ahorro y Préstamo (SINAP)—a system created in 1960—and a few other indexed instruments. This lack of a uniform policy consistently tended to favor SINAP versus commercial banks, and financieras, especially when the rate of inflation increased sharply in 1972 and 1973, prompting the authorities in 1974 to take measures to eliminate interest rate ceilings by extending indexation and freeing interest rates.
In Colombia, the Bank of the Republic (the central bank) traditionally determined ceiling rates on all bank and nonbank operations. During the upturn in the rate of inflation that began in 1972, this system became largely repressive, a problem only partly reduced by the creation in that year of a system of savings and loan associations with indexed instruments. In 1974, the authorities extended ceilings to indexed operations.
The structure of maximum deposit rates of commercial banks in Paraguay remained virtually unchanged from January 1959 through 1976. However, the Paraguayan financial system was largely unrepressed because inflation was perceptibly high only in 1973 and 1974; in addition, financieras began to develop rapidly during that period because of freedom from regulations, and a new system of savings and loan associations with indexed deposits began operating in 1973.
The experiences of other countries were largely peculiar to each one of them. Interest rate ceilings in Argentina went through several stages, moving in general toward liberalization. Before February 1971, the Central Bank of the Republic of Argentina kept ceilings on all liability rates of nonbank intermediaries and on the banks’ savings deposit rate but not on time deposits; in some cases, these rates were adequate to compensate savers for the then moderate increase in prices. In February 1971, the Central Bank increased the ceiling on the banks’ savings deposit rate and freed the rates paid by nonbank financial intermediaries in a first attempt to move progressively away from ceiling regulations, but supported an agreement reached informally among the banks to observe a maximum limit of 17 per cent for all time deposits of 180 days and over, whereas these rates had previously been free. This experiment lasted only for one year because the rapid acceleration of inflation to 35 per cent that took place in 1971 threatened to impair the liquidity of banks as nonbank rates began to climb, so that interest rate ceilings were reimposed on all financial intermediaries, thereby causing financial repression. In 1975, under a rapidly escalating hyperinflation, the Bank first increased all ceilings and then freed all rates, except for a 40 per cent ceiling on banks’ savings deposits (raised to 55 per cent in September 1976). This remaining ceiling was still quite insufficient to compensate savers for the severe hyperinflation, and so, in 1977, the ceiling was finally removed, resulting in a substantial increase in short-term interest rates without creating institutional problems.
The Dominican Republic offers an example like that of Peru in which usury legislation played a basic role in keeping a lid on deposit rates. But, unlike Peru, interest rates on deposits were not regulated directly (with a few exceptions); they were limited indirectly by a 12 per cent usury ceiling applicable on all lending operations.
In Panama, the authorities maintained a dual interest rate system whereby interest rates on resident operations were generally subject to ceilings, while interest rates on nonresident operations were completely unrestricted. This policy was devised—in conjunction with capital restriction measures—to insulate the domestic financial system from the influence of offshore banking business, while encouraging the development of Panama into a regional financial center. The latter objective was accomplished, but the rigidity in the structure of domestic interest rates that resulted from regulations tended to discourage savings by domestic residents.
Haiti was the only country in which formal interest rate ceilings were introduced anew during the decade. However, even before June 1972, when the National Bank of the Republic of Haiti (which operates as both a central bank and a commercial bank) formally established maximum rates for private commercial banks, the scheduling of its own rates greatly influenced the decisions of other banks. In contrast to Haiti, Bolivia and Venezuela abandoned the tradition of imposing ceilings on liability rates in favor of floor rates.
In most of these countries, interest rate ceilings were imposed in the belief that a policy of stable and relatively low interest rates would help to maintain a high level of investment as well as to strengthen the stability of financial institutions. This view neglected to adequately consider the detrimental effects of financial repression, such as the deceleration in the growth of financial assets, the reduction in investment efficiency in the productive sectors, the reinforcement of the oligopolistic power of financial institutions, and the deterioration in the distribution of income.
Traditional attitudes in favor of usury-type regulations influenced other countries. These attitudes extended beyond those seen in Peru and the Dominican Republic, to which reference was made earlier.
In other countries, interest rate ceilings on liabilities were the result (intentionally or not) of the commitment of the monetary and credit authorities to keep low interest rates on some types of loan in order to redirect credit flows toward priority sectors, such as agriculture and housing. However, it was forgotten that, in promoting specific sectors, it is counterproductive to maintain the average level of rates below equilibrium because this reduces the volume of funds for any activities, including those to be promoted. It is therefore more appropriate to pursue a policy of offsetting low rates on preferred loans by increasing the rates on less preferred loans. In addition, it was insufficiently recognized that loans provided at low rates often had to be financed by deposits at correspondingly low rates, which paradoxically tended to dry up the funds for the preferred activities.
Finally, in many countries that imposed interest rate ceilings to prevent the increase in interest rates, this policy was concomitant with or even propelled by a keen desire of the fiscal authorities to have access to subsidized credit for use by the governments, the public enterprises, and other public entities. These policies, however, did not in the end help to finance the public sector, as low interest rates discouraged acceptance of public offers.
Interest rate floors
An interesting development during the decade was the switch in Bolivia and Venezuela toward legal floor rather than ceiling rates on obligations of financial institutions, a policy designed to encourage financial savings. 13 Bolivia maintained a minimum rate of 10 per cent on savings deposits denominated in pesos from October 1969 to October 1972 (when this minimum was turned into a repressive maximum rate of 10 per cent) and from October 1973 onward, and a graduated structure of minimum rates on peso-denominated time deposits was added from August 1975 onward. In Venezuela, all existing ceilings on liability rates were replaced in June 1969 by floor rates, and their reference levels were raised.
These experiments were not, however, effectively designed to allow interest rates on financial sector obligations to rise to equilibrium, because minimum rates on obligations were accompanied by maximum rates on the asset side of the affected financial institutions. As interest rate ceilings on loans effectively constrained the maximum rates that those institutions could pay on their deposits, full adjustment to inflationary pressures could not be achieved. However, a desirable by-product of this system was the maintenance of a maximum spread between asset and liability rates that may have pushed financial institutions to increase their internal efficiency.
Experimentation with various indexation forms and techniques spread throughout several countries; however, a general indexation system could not be found. Instead, one can find only two countries—Brazil and Chile—in which indexation played an important role, together with a number of others in which “partial indexation” or other special forms of indexation measures were undertaken. 14
As is well known, among Latin American countries Brazil was the most important champion of indexation policies from the time of their inception in 1964 until 1975, when interest rates were freed (although indexation remained freely negotiable). However, even in Brazil, in contrast to popular belief, less than half the volume of financial assets was ever indexed, and the formula applied for monetary correction that was based on the wholesale price index was altered quite often and even temporarily suspended when its results were considered to be undesirable. 15 Nevertheless, indexation played a prominent direct role in maintaining the real value of some selected key financial instruments, including government bonds, and indirectly maintained the real value of virtually the whole spectrum of obligations of financial institutions and instruments outside the financial system.
Indexation in Chile went through two different stages. Partial indexation from 1960 to 1974 was discussed earlier and was indicated to have caused a severe institutional imbalance. The second stage—one of widespread indexation—began in May 1974, when indexing became applicable to any operation of more than one-year maturity, unless otherwise specified by the contracting parties. (Although readjustment of operations of less than one year was made illegal, the minimum maturity required for indexing was reduced to three months in 1976.) This widespread indexation was instrumental in sharply increasing interest rates toward equilibrium, despite the fact that, unlike that in Brazil, it was not compulsory (in accordance with the basic goal of the reform, which was to liberalize interest rates). Unfortunately, however, the inherited imbalance caused serious institutional problems of adjustment during the transitional period.
Colombia’s partial indexation experiment from 1972 to 1974 led to institutional imbalances similar to those observed in Chile. The response to the interest rate differential created in favor of indexed deposits of the savings and loan associations, compared with the deposits in other institutions subject to ceilings, was dramatic. With the further escalation of inflation in the beginning of 1974, the authorities were confronted with a choice of allowing bank rates to rise toward the level paid by the savings and loan associations or of reducing the rates paid by the latter. The authorities met the problem halfway by introducing a ceiling of 20 per cent per annum on the effective interest rate that savings and loan associations could pay on their deposits, while raising the interest rate ceilings on bank deposits. The effect of these measures was to improve the realignment in the structure of rates and to reduce the imbalance, but at the expense of a failure to maintain the rates at a level sufficient to compensate savers for continuing high inflation. 16
The experiment with partial indexation in Paraguay was similar in its framework to that in Chile and Colombia. (Incidentally, all three were based on the consumer price index.) However, in contrast to Chile and Colombia, the relatively moderate rate of inflation that prevailed in Paraguay (except during 1973 and 1974) and the lack of complete readjustment for the cost of living, coupled perhaps with its less sophisticated financial system, made the problem of institutional imbalance much less serious, and the authorities were not as urgently compelled to undertake corrective measures to ensure adjustments.
Special measures that indexed financial instruments to variables other than general price indices, either for the whole or part of the financial system, were undertaken in a number of countries. These measures presented difficulties, owing to the inappropriateness of particular indices and, when applicable, to their partial nature.
In June 1974, the Bolivian monetary authorities established a dollar-value guarantee clause on commercial bank savings deposits to encourage their increase by protecting savers in case of devaluation; the same system was extended to term deposits beginning in August 1975. This policy, in addition to being ineffective in protecting savers against inflation, put the Central Bank in the dangerous position of having to provide the funds necessary to honor the guarantee, thereby ensuring that the resulting monetary expansion would automatically tend to offset the desirable effects of devaluation.
In September 1974, the monetary authorities of Costa Rica allowed local institutions to offer some foreign currency deposits at rates ranging up to 1 per cent over the respective London interbank offer rate (LIBOR), and one month later all certificates of deposit in finance companies were also tied to LIBOR. Although this linkage of rates permitted a reversal of the trend toward increasing capital outflows, the disparity in the treatment of banking rates subject to ceilings and finance company rates pegged to LIBOR created difficulties, so that the authorities suspended the relationship to LIBOR by mid-1975. In addition to the specific difficulties found in Costa Rica, the policy of indexing to LIBOR (or any other foreign rate) can create serious problems when domestic and foreign rates of inflation are not closely aligned and/or there are expectations of changes in the exchange rate.
During the decade, there were two seemingly opposed tendencies with regard to liberalization of interest rates. While the monetary authorities in some countries strove to extend regulations to all financial intermediaries in order to prevent imbalances in the institutional framework, other countries undertook deregulation measures in relation to a broad range of financial instruments, and in Argentina, Brazil, Chile, and Uruguay, complete liberalization of interest rates was eventually accomplished.
The following discussion is limited to Brazil, Chile, and El Salvador. The experiences with partial interest rate freedom or liberalization of countries that retained some effective interest rate ceilings, such as Guatemala, Nicaragua, and Paraguay, and of other countries that achieved liberalization beyond the period of this study (Argentina and Uruguay), have already been discussed. Summing up, the policy of restricting interest rate freedom to a few instruments was inadequate to permit interest rate adjustment to equilibrium under inflation and, in addition, it produced institutional imbalances.
Throughout the decade 1967–76, among Latin American countries, El Salvador was an exception to the general prevalence of interest rate regulations. Indeed, until 1971 the Central Reserve Bank of El Salvador lacked the power to regulate interest rates. Nevertheless, there was little effective freedom for rates to be determined in the market because maximum rates paid on deposits were fixed by an understanding among the banks. As a result of this understanding, the interest rate level and structure remained unchanged until February 1974, when a minor upward adjustment was effected, followed by further minor upward adjustments in April 1976. This inflexibility in interest rates contrasted with the rise in the consumer price index from about zero in 1967 to a double-digit figure during 1974 and 1975. The experience of El Salvador suggests that interest rate freedom cannot be expected to be a panacea and automatically produce interest rate equilibrium in countries that have a relatively concentrated and unsophisticated financial system, which is the typical case in small, less developed countries. Positive policy actions to avoid market distortions might be required.
Virtually complete liberalization of interest rates was accomplished in Brazil through measures undertaken in 1975 and 1976. One of the reasons for abandoning mandatory indexation was the insufficient increase in interest rates, owing to the rapid escalation of inflation and the lag in the indexation formula. Furthermore, liberalization was considered to be more consistent with a new monetary policy framework in which open market operations would play a larger role. The transition to interest rate freedom was an orderly and successful operation from the start, owing to the existence of appropriate conditions. Throughout the long Brazilian experience with indexation and discretionary actions, the authorities had consistently guided the market toward both maintaining a level of interest rates that could compensate the majority of savers for inflation and aligning the structure of interest rates to reflect the preferences of the public and to reduce discrimination among financial institutions. Furthermore, prior to undertaking liberalization measures, the authorities had appropriately signaled their desire to see interest rates rise once again (from the trough that they had reached in 1973) by taking discretionary measures twice in 1974 to increase ceiling rates by several percentage points and to align the structure of rates to reflect market preferences.
In Chile, interest rate liberalization was accomplished through a sequence of measures that spanned one year from May 1974. Unlike that of Brazil, the Chilean transitional process was arduous and frustrated by severe institutional imbalances. 17 Difficulties resulted largely from a very unfavorable set of initial conditions, namely, a rapidly escalating rate of hyperinflation that severely repressed the interest rates paid by the entire financial system, and a pronounced institutional imbalance, aggravated by the exposure of SINAP to large-scale programs of office and residential construction, which made their liquidity quite vulnerable to deposit withdrawals. The process of liberalization underwent several stages. In May 1974, all previous interest rate ceilings on nonadjustable instruments with a maturity of less than one year were replaced by an overall ceiling of 200 per cent for any type of commercial bank operation, and investment banks were allowed to pay up to 50 per cent more than the normal banking rate, thereby ending the quasi-monopolistic position held by SINAP in the attraction of funds. This caused a redirection of resources from SINAP, especially toward investment banks; endangered the liquidity position of SINAP, particularly after a modification of the formula to readjust SINAP’s readjustable mortgage deposits, Valores Hipotecarios Reajustables (VHRs) in the second half of the year; and forced the Central Bank to step in on behalf of the Government to finance withdrawals of deposits from SINAP. In November 1974, the interest rates that commercial banks could pay on any time deposits (and earn on commercial paper) were freed. As a result of massive withdrawals from SINAP, confidence in the system dwindled, and despite the financial support of the Central Bank, it verged on bankruptcy. Finally, despite these difficulties, the remaining 200 per cent ceiling on the savings deposits rate was eliminated, and bank lending rates were freed. As a result, all interest rates rose substantially and became consistently positive in real terms for the first time in many years. This eventual accomplishment was, however, achieved at a high cost during the transitional period because the financing of withdrawals from SINAP had become a major cause of the continuously rapid increase in the money supply, and of inflation.
II. The Extent of Financial Repression
It would be difficult to determine whether the changes in policy described earlier were generally the result of greater enlightenment or constituted a rather belated adjustment to the new inflationary environment that characterized the 1970s. On the one hand, as inflation worsened in all the countries during the first half of the 1970s, efforts to implement equilibrium policies were, in some instances, thwarted by new and sometimes unexpected developments. On the other hand, in some instances, new policies had been considered or initiated before becoming absolutely necessary, but events simply overtook the authorities’ desire for better adjustment and made it more difficult to achieve. In any case, as is shown later, the new policy attitudes did not, for most countries, reduce the extent of financial repression nor did they improve the structure of interest rates.
The statistical series on liability interest rates used in this study, to aid country comparisons, must be taken with some caution, because these are not necessarily representative of the interest rate structure for countries that have sophisticated financial systems and are composed mostly of legal figures, since the effective rates were never recorded. 18 Unfortunately, casual evidence suggests that effective rates were, in some cases, significantly different from the legal figures, not only within the legal constraints but also because there were instances of noncompliance with the law. Finally, there are statistical gaps, particularly in regard to interest rates that were free.
The series on the savings deposit rates (shown in parentheses in Table 6) represent the lowest end of the interest rate structure on financial assets, with the exception of currency and demand deposits, on which no interest was paid. In 1967, savings deposit rates in the banking system ranged from about 2.5 per cent in Haiti to 8 per cent in Argentina, with the exception of Brazil where it might have been more than 20 per cent. Rates paid by other financial institutions that were allowed to receive savings deposits, especially the savings and loan associations, were slightly above or below rates paid by commercial banks, except in Chile, where savings and loan associations paid substantially higher rates than banks. In most cases, rates were in line with the moderate rate of inflation that occurred in 15 of the 19 countries.
Throughout the decade there was a general tendency toward increasing the savings deposit rate in most countries, presumably in response to the accelerating rate of inflation. But this adjustment lagged significantly in many countries and was generally incomplete. By 1976, the savings deposit rate was in the range of 50 per cent per annum in Argentina and Uruguay, about 30 per cent in Brazil, and somewhat below 200 per cent in Chile. However, in all other countries, except Colombia and Bolivia, it was below 10 per cent.
A sample of liability rates representing the highest level existing in each country during most or all of the decade, subject to availability of data, was selected. (See Table 3.) 19 Despite the partial unavailability of data in some countries—which can, however, to some extent be supplemented by an educated guess—some impressions of trends in maximum interest rates during the decade could be formed. First, interest rates on the high side of the structure generally tended to move upward by more than the rate on savings deposits. Second, this was especially true in those countries in which inflation accelerated rapidly toward hyperinflation, that is, Argentina and Chile. Third, some policies (e.g., partial indexation in Chile, Colombia, and, to a lesser extent, Paraguay) created a disparate structure of interest rates on instruments that were essentially similar.
|Colombia (Savings and loan)5||—||17.6||23.8||24.6||20.9||…|
|Costa Rica (Time deposits)7||7.0||8.7||13.0||14.0|
|(Finance company certificates)7||…||12.0||13.0||14.0|
|(Savings and loan)18||—||14.0||18.0||12.0||12.0|
As a result, the interest rate structure on liabilities of financial institutions tended to be more divergent as the general level of interest rates rose during the decade. This is reflected in the increasing difference between the savings deposit rate (minimum) and the high rate (maximum)—most other rates being within these two boundaries. One of the underlying factors was that, except for Uruguay, the economies of the more inflationary countries were relatively large and advanced and, therefore, their financial structure tended to be more diversified. The diversification of institutions, operations, and maturities created the vehicle for a more diversified interest rate structure, but this was not necessarily a sign of a greater tendency toward equilibrium.
The divergent structure of interest rates might have reflected to some extent a basic problem inherent in existing monetary systems posed by the lack of payment of interest on currency and demand deposits. This institutional constraint on the rates earned on two of the major financial assets held by the public apparently tended to limit the increase in other interest rates at the low end of the maturity structure. This consideration suggests that perhaps in economies with high rates of inflation no system of corrective measures is likely to succeed fully in eliminating the undesirable effects of inflation. (See the subsection, Financial repression, interest rate structure, and income distribution.)
From a statistical point of view, the sharp divergence in some of the rates on liabilities poses the issue of how representative each of the series was within the structure of interest rates, and leaves open the question of what was the effective degree of overall financial repression. One answer to this question could be obtained by weighting each interest rate by the relative share of the financial instrument in question with respect to the entire structure of liabilities. That task is beyond the scope of this paper.
Indices of financial repression
As noted earlier, in judging the appropriateness of an interest rate it is important to determine whether it approaches the expected rate of inflation. The prevalence of negative expected real interest rates thus provides empirical content to the notion of financial repression. Financial repression is a complex concept, however, even when limited to the interest rate sphere because of the multitude of interest rates that form the interest rate structure and the difficulty in quantifying the expected rate of inflation, and also because, in practice, short-run adjustment problems as well as external considerations might require a modification of this basic position.
Nevertheless, to make an empirical approximation of financial repression, it may be useful to examine whether liability rates ranging from those on savings deposits to those on high-yielding assets were positive or negative in real terms with reference to both the actual and our illustrative expected rates of inflation, which represent the boundaries of maximum and minimum adjustment that might be required under most normal circumstances. The existence of a negative real rate on savings deposits based on both inflation measures may be an indicator of at least some degree of financial repression. Negative real rates on liabilities with high nominal yields may indicate the prevalence of substantial financial repression. These indices are displayed in Tables 4 and 5, respectively.
|Costa Rica2 (I)||5.8||2.9||4.1||2.3||3.8||2.3||–7.1||–16.4||–3.7||10.1|
Effect of the high inflation of the period 1973–76 on real interest rates
In 1967 the real savings deposit rate was negative in only 6 of the 19 countries. Negative rates arose in Argentina, Chile (except in savings and loan associations), and Uruguay, and to a lesser extent in Colombia and Peru, because they were typically inflationary countries, whereas in Bolivia this resulted from an atypi-cally high rate of inflation compared with adjacent years. The remaining countries achieved positive real rates as a result of moderate inflation, except for Brazil, which achieved them as a by-product of indexation of other instruments.
That relatively pleasant picture changed dramatically during the high-inflation period 1973–76. During 1973 and 1974, all countries had a negative real savings deposit rate (except for Brazil in 1973). Then some improvement took place throughout 1975 and 1976. In 1975, the substantial reduction in the rate of inflation in Bolivia, Nicaragua, and Paraguay resulted in a return to a positive rate, while Chile and Brazil achieved positive real rates despite the continuing high rate of inflation. In 1976, four additional countries moved to a positive real rate. In Costa Rica, Honduras, and Panama, this was due mainly to the reduction in the rate of inflation, while in Uruguay it was due largely to a substantial increase in the rate of interest.
The sample of high liability rates also reveals a striking contrast between 1967 and the period 1973–76. In 1967 this real rate was negative only in Bolivia (as a result of atypically high inflation in that year) and in Argentina and Uruguay, two countries with overwhelming financial repression. By 1973 there were only six countries in which this high real rate remained positive: Brazil, Colombia, El Salvador, Honduras, Peru, and Venezuela. In 1974 the number was reduced to two countries: Argentina, as a result of a temporary reduction in the rate of inflation; and Colombia owing to full monetary correction on deposits in savings and loan associations. The situation improved slightly in 1975 when the rate became positive in Bolivia, Chile, Honduras, Nicaragua, Panama, and Paraguay; but it turned sharply negative in Argentina owing to hyperinflation and somewhat negative in Colombia (except for the unorganized money market) as a result of the ceiling imposed on indexed deposits of the savings and loan associations. A more substantive improvement was registered in 1976, when the rate remained negative in only five countries: Argentina, Ecuador, Guatemala, Mexico, and Peru.
The indices of financial repression based on the illustrative expected rate of increase in prices present a picture that is similar, in general terms, to the raw indices. Owing to the smoothing -out effect of the autoregressive method used to compute the expected indices, however, it is not surprising to find that, by these indices, fewer countries can be classified as financially repressed during the inflationary outbreak of 1973–74, nor that the situation was reversed by 1976 when the rate of inflation dipped in a number of countries.
In 1967, eight countries were financially repressed according to the index of the expected real savings deposit rate. Brazil and Haiti were in this category, in addition to Argentina, Bolivia, Chile, Colombia, Peru, and Uruguay, which were also identified as repressed by the raw index. The sample of expected high real liability rates shows Brazil, Chile, and Uruguay as having been financially repressed in that year, whereas the actual rate index pinpointed Bolivia and Argentina in addition to Uruguay, but not Brazil and Chile. The discrepancy for Brazil is due to the influence of the pre-1964 high inflation rates on the expected indices, which one would be inclined to disqualify because the 1964 reform probably lowered inflationary expectations. On the other hand, the expected high real rate indices, which show that Argentina and Bolivia were not financially repressed in 1967, provide a more accurate picture than do the actual indices, since these countries had a year of atypically high inflation compared with adjacent years.
Because inflation had just begun to escalate rapidly in 1973, the expected indices for that year do not indicate widespread repression (in contrast to the indices based on actual inflation). Indeed, by the expected indices, even at the peak of inflation in 1974 and in the aftermath in 1976, several countries escaped financial repression. The expected real savings deposit rate remained positive in Brazil and Guatemala through 1976 (Chilean data do not allow an accurate assessment). The index of the expected high real rate shows 6 repressed countries in 1974 compared with 17 countries in terms of the actual rate, but in 1976 the expected rate index shows 10 countries with financial repression compared with 5 in terms of the actual rate.
Financial repression, interest rate structure, and income distribution
The significant disparity between liability interest rates that existed in some countries, particularly during the high-inflation period 1973–76, created a situation in which, in some of the countries, only part of the interest rate structure was negative in real terms, thereby raising difficult questions concerning the overall degree of financial repression and the distribution of income. While more research on differences in the financial structure among countries would be needed to answer such questions, perhaps a generalization can be made that with the exception of the relatively large and more developed countries, such as Argentina, Brazil, Chile, and Mexico, savings and time deposits were a dominant component in the structure of total financial sector liabilities to the public. If so, given the relatively small differentials that existed in most countries between the savings deposit rate and rates on time deposits, negative real savings rates constitute a presumptive indicator of the financial repression experienced by the average and below-average saver, while high-income groups fared much better, since they had greater access to higher-yielding instruments that sprang up in many countries throughout the decade. These developments were often associated with the growth of specialized mortgage banks and financieras in a number of countries.
A more favorable development from the point of view of the distribution of income was the rapid expansion of the savings and loan associations in a number of countries. In some cases, this also helped to reduce the overall financial repression.
Finally, the expansion of parallel or noninstitutional markets with attractive interest rates in some countries—notably Colombia—as a result of the repression of the institutional financial system might have helped to reduce the overall degree of financial repression, but probably aggravated the perverse effects of institutional repression on the distribution of income.
The development of foreign currency obligations in local institutions and the availability of foreign financial outlets for local investors presented new challenges to the monetary authorities of the area.
Domestic rates in domestic currency and in foreign currency; the dollarization problem
In many of the Latin American countries surveyed, it was possible for domestic residents to open deposit accounts and to acquire other local financial assets denominated in foreign currencies. In some countries, the markets for instruments in foreign currency sprang up in a legislative vacuum, while in others it was legally permitted within an explicit regulatory framework. The U. S. dollar was by far the most frequently used foreign currency in Latin American countries.
Instead of attempting a comprehensive survey of interest rates on foreign currency instruments—a difficult task given the lack of data in this area—this paper outlines three examples of this dollarization problem and the ingredients for a practical policy of containment.
In Uruguay, during the prolonged period of widespread financial repression that preceded the de facto interest rate liberalization in 1976–77, the policy of depreciating the Uruguayan peso encouraged private asset holders to acquire foreign currency assets issued domestically, the effective interest rates of which in peso terms considerably exceeded the maximum rates applicable on any domestic currency instruments. The dollarization was further aided by public offers of Uruguayan Treasury bonds denominated in U. S. dollars.
In Bolivia, the dollarization problem received the first serious impetus in the aftermath of the 1972 devaluation of the Bolivian peso, which raised the attractiveness of foreign currency instruments relative to those in pesos. To cope with this growing problem, interest rate regulations on domestic currency deposits were changed from a maximum to a minimum of 10–11.75 per cent, and foreign currency rates were subject to a 9 per cent ceiling. Thus, the policy of simultaneously fixing a floor on domestic currency rates and a ceiling on foreign currency rates could be used to contain the dollarization problem by increasing the effective interest rate differential in favor of domestic currency instruments up to a desired level, after allowing for the expected rate of change in the external value of the domestic currency.
In Mexico, until 1976 the authorities contained the dollarization problem by limiting permission to constitute foreign currency deposits in private banks to residents in the U. S. border areas and to foreign residents, and by keeping foreign currency rates in line with domestic currency rates and a stable exchange rate. This situation changed dramatically on March 22, 1976 when—in the wake of diminishing confidence in the external value of the Mexican peso—the authorities almost completely liberalized the previous constraints in order to allow private banks to accept foreign currency deposits from all residents. This provoked a massive conversion of domestic currency into foreign currency deposits. 20
Domestic rates and foreign rates; Eurodollar rates and private capital movements
It is well known that the relationship between foreign interest rates and domestic rates (whether in domestic or in foreign currency) is an important determinant of private international capital flows. However, interest rates in foreign and domestic currency are not comparable unless the rate of change in the exchange rate between the two currencies is expected to be zero. The problem of measuring exchange rate expectations is a very difficult one. Instead of attempting to derive any measure of exchange rate expectations, this paper uses the actual changes in the exchange rates with respect to the dollar as an adjustment factor for comparison of domestic currency and foreign interest rates. (See Table 6.) The comparison is between the adjusted one-month Eurodollar rate and the local rates on savings deposits, which are instruments roughly comparable in maturity.
The adjusted Eurodollar rate reflects the enormous exchange rate adjustments that took place in some of the countries, particularly in the more inflationary ones. All adjustments were in the direction of enhancing the relative value of the U. S. dollar except for a minor appreciation of the Venezuelan bolívar in 1973. In Brazil, Chile, and Colombia, adjustment took place in every year, and in Argentina and Uruguay, depreciation occurred in most years. Bolivia, Costa Rica, Ecuador, and Mexico each suffered a more or less sudden and pronounced depreciation of their currencies, and the Peruvian sol was depreciated in 1967 and again during 1975 and 1976. The remaining eight countries—less inflationary on average—did not change the value of their currencies in relation to the dollar (except for the noted appreciation of the Venezuelan bolívar).
Data show the widespread prevalence of an interest rate differential in favor of the Eurodollar rate, not only during the depreciation of domestic currencies (in which case the differential usually tended to become enormous) but also under exchange rate stability. The main exceptions were Brazil, owing to its especially high nominal rate, and Bolivia, except for the year of the depreciation (1972). In terms of periods, 1976 was the exception because the Eurodollar rate dipped to 5 per cent while domestic rates reached, or stayed at, an all-time peak in every country except Mexico.
Although it would be premature to draw any firm inferences from the preceding comparison of rates for the behavior of capital flows, the evidence is at least consistent with data showing that Latin American countries registered private short-term capital outflows. This is not to say that some high-yielding domestic assets could not have rates above those of comparable instruments in international markets, as happened in Mexico, thereby causing some longer-term inflows. Nevertheless, the bulk of foreign long-term capital inflows in Latin American countries is known to be associated with borrowing by the public sector.
This paper has analyzed the evolution of policies on the structure of interest rates on financial sector liabilities to the public in 19 Latin American countries during 1967–76, and has examined the degree of financial repression and some of its associated problems.
Whereas at the beginning of the decade virtually all countries still applied relatively low interest rate ceilings (the best-known exception being that of Brazilian indexation), a number of them moved throughout the decade to the use of interest rate floors and of various forms of indexation and toward interest rate liberalization. Nevertheless, this new experimental attitude did not always reflect greater enlightenment in the formulation of interest rate policies, but rather constituted a belated adjustment to the severe inflation that occurred in the region, particularly in the period 1973–76.
On the whole, indexation techniques produced disappointing results, owing, at least in part, to the unwarranted use made of them. Even in Brazil—the most successful experience—indexation proved to be difficult to apply on a fully comprehensive basis, and it required continuous retouching through discretionary action by the authorities. Problems were severely magnified when indexation was applied on a partial basis, and it was accompanied by severe inflation, as in the savings and loan associations in Chile and Colombia, because this tended to require credit-supporting measures that tended to aggravate inflation. Indexation with respect to a foreign currency, either directly or through a value-guarantee clause, could also prove to be potentially dangerous.
The trend toward liberalizing interest rates in Latin America was more recent and more successful than indexation. In Brazil, the 1975 transition to interest rate freedom was orderly and successful from the start. Similarly successful liberalizations were achieved in Uruguay during the period 1976–77, and in Argentina during 1975–77. On the other hand, the liberalization process in Chile in 1974–75 was accompanied by severe institutional imbalances (resulting from an initial institutional disequilibrium and from the piecemeal fashion in which liberalization was applied) that became a factor in the Central Bank’s inability to control credit and in the ensuing hyperinflation. The long experience with nonregulation in El Salvador was frustrated by an understanding among the banks to fix maximum interest rates on deposits. These varied experiences demonstrate that two conditions are necessary to liberalize interest rates successfully: a competitive financial system and an interest rate structure that approaches equilibrium, thereby suggesting the need, in some countries, for positive discretionary action to achieve those conditions.
The two unadjusted indices of financial repression used in this paper—the real savings deposit rate and a sample of high interest rates—indicate that while in 1967 financial repression was limited to a minority of inflationary countries, it spread to the entire region throughout the inflationary period 1973–76. However, in terms of the illustrative expected real rates, several countries that were less affected by a prolonged bout of inflation escaped financial repression. In the majority of repressed countries, financial repression was incomplete in the sense that it was not indicated by all indices.
The upward adjustment that took place in interest rates was not equiproportional for all rates. Rather, it appears that the structure of interest rates tended to diverge in most countries, especially the more inflationary ones. This divergence in the interest rate structure probably produced perverse effects on the distribution of income.
Both the lack of a sufficient adjustment in the level of interest rates to inflation and the divergent structure of interest rates indicate that inflation can present serious problems even when corrective measures are undertaken. Any corrective measures are bound to have undesirable side effects. In addition, the divergence in the structure of interest rates might be the reflection of the existence of a floor constraint on the structure of interest rates imposed by the lack of payment of interest on such an important instrument as currency.
Another area in which the problems raised by inflation are revealed is that of adjusting the structure of interest rates to external influences. Theoretically, developing countries need to ensure that domestic deposits, in domestic and foreign currency, do not fall below comparable foreign rates in order to avoid capital outflows, and also that domestic deposits in domestic currency earn higher yields than those in foreign currency, in order to avoid the dollarization problem. In practice, the relationship between domestic inflation and depreciation of the currency requires adjustment for expected changes in the exchange rate, but it is difficult to measure those expectations.
Imperfections in interest rate data further impede a full evaluation of the problems discussed earlier. At best, the legal interest rate series that this paper has reconstructed can offer a rough approximation. There is a need to give greater attention to the task of measuring effective market interest rates to register interest rate movements that take place within the established legal norms, particularly under interest rate freedom. This is a task that central banks could carry out more easily than any other institution and from which they would benefit the most. In addition to the immediate operational needs, more accurate data than are currently available will be necessary to test the numerous hypotheses that have been proposed in recent years on the effects of interest rate policies in developing countries.
Mr. Galbis, Senior Economist in the Central Banking Service, is a graduate of the University of Barcelona, Spain, and the University of Wisconsin.
An earlier, more comprehensive, version of this paper was presented at the XV Meeting of Central Bank Technicians of the American Continent, held at Port-of-Spain, Trinidad and Tobago, November 19–24, 1978. In addition to colleagues in the Fund, the author is indebted to some delegates to that meeting for their comments. He is, however, solely responsible for any remaining errors.
There are some well-known exceptions, particularly in reference to developing countries in Asia. See Anand G. Chandavarkar, “Some Aspects of Interest Rate Policies in Less Developed Economies: The Experience of Selected Asian Countries,” Staff Papers, Vol. 18 (March 1971), pp. 48–112; U Tun Wai, “Interest Rates in the Organized Money Markets of Underdeveloped Countries,” Staff Papers, Vol. 5 (August 1956), pp. 249–78; U Tun Wai, “Interest Rates Outside the Organized Money Markets of Underdeveloped Countries,” Staff Papers, Vol. 6 (November 1957), pp. 80–142; and U Tun Wai, “A Revisit to Interest Rates Outside the Organized Money Markets of Underdeveloped Countries,” Banca Nazionale del Lavoro, Quarterly Review, No. 122 (September 1977), pp. 291–312.
As recently as 1975, none of the Latin American central bank bulletins presented substantive information on relevant interest rate series. See Vicente Galbis, “Money, Investment, and Growth in Latin America, 1961–1973,” a paper presented at the XII Meeting of Central Bank Technicians of the American Continent, in Punta del Este, Uruguay, November 2–7, 1975. (It was published in Economic Development and Cultural Change, Vol. 27, April 1979, pp. 423–43.)
See Ronald I. McKinnon, Money and Capital in Economic Development, The Brookings Institution (Washington, 1973); Ronald I. McKinnon, ed., Money and Finance in Economic Growth and Development: Essays in Honor of Edward S. Shaw (New York, 1976); Edward S. Shaw, Financial Deepening in Economic Development (Oxford University Press, 1973); Vicente Galbis, “Financial Intermediation and Economic Growth in Less Developed Countries: A Theoretical Approach,” Journal of Development Studies, Vol. 13 (January 1977), pp. 58–72.
See Vicente Galbis, “Theoretical Aspects of Interest Rate Policies in Less Developed Countries” (unpublished, International Monetary Fund, February 9, 1978). A preliminary version of this paper was presented at the XIV Meeting of Central Bank Technicians of the American Continent, in San Carlos de Bariloche, Argentina, November 6–11, 1977.
Correlation analysis between the rates of change in the consumer price index, the wholesale price index, and the gross domestic product (GDP) deflator for each country (subject to availability of data) shows that, except in a few cases, the correlation coefficients were above 0.9.
Under normal circumstances, the relevant period and the weights attributed to each observation within the period may be estimated empirically by econometric methods, in order to derive more precise illustrative figures of the expected rate of inflation. See Vicente Galbis, “Interest Rate Policy Framework for a Developing Economy: The Case of Peru” (unpublished, International Monetary Fund, November 16, 1976).
Country details are examined in the unpublished version of this paper, available upon request from the author, whose address is Central Banking Service, International Monetary Fund, Washington, D.C. 20431.
“Dollarization” is defined here as the increasing holdings of assets denominated in dollars in the portfolios of the public.
At the same time, the liberalization of the right to accept U. S. dollar-denominated deposits accentuated the dollarization of the Mexican economy. (See the subsection, Domestic rates in domestic currency and in foreign currency; the dollarization problem.)
See Vicente Galbis, “Saving and Financial Intermediation in Peru: Analysis and Policy” (unpublished study prepared for the International Bank for Reconstruction and Development, January 1977); Galbis, “Interest Rate Policy Framework for a Developing Economy” (cited in footnote 6).
For a discussion of the experiences of these three countries from the point of view of partial indexation, see the subsection, indexing of financial savings instruments.
See Banco Central de Chile, “Políticas y Métodos de Determinación de Niveles y Estructura de Tasas de Interés en Chile,” Boletín (July 1976), pp. 1043–59.
By contrast, Guatemala freed the minimum rates on savings deposits in state banks in 1972, as stated earlier.
Partial indexation is defined as indexation of financial instruments in only a subset of financial institutions.
See Jack D. Guenther, “’Indexing’ versus Discretionary Action—Brazil’s Fight Against Inflation,” Finance and Development, Vol. 12 (September 1975), pp. 24–29; Alexandre Kafka, “Indexation for Inflation in Brazil,” in Essays on Inflation and Indexation, Domestic Affairs Studies, No. 24, American Enterprise Institute for Public Policy Research (Washington, 1974), pp. 87–98.
See Vicente Galbis, “Colombia: Evolution of the Financial Sector and the Monetary Policy Instruments, 1967–74” (unpublished, International Monetary Fund, October 1975).
See Banco Central de Chile (cited in footnote 12); Donald J. Mathieson, “Financial Reform and Stabilization Policy in a Developing Economy” (unpublished, International Monetary Fund, March 6, 1978), pp. 22–27.
The series used here are only a small subset of those contained in the original version of this paper, which is available upon request from the author, whose address is Central Banking Service, International Monetary Fund, Washington, D.C. 20431.
In Colombia and Paraguay, two alternative series are considered in order to adequately reflect the significant change in the financial structure caused by the creation of savings and loan association systems. In Costa Rica, the alternative series shows a rate that is slightly higher than the main series, but it has the drawback of having been unavailable for the earlier years of the period.
See Banco de Mexico, Informe Anual (Mexico, D. F., 1976), pp. 39–48.