3 Liberalization, Crisis, Intervention: The Chilean Financial System, 1975–85
  • 1 https://isni.org/isni/0000000404811396, International Monetary Fund


In the mid-1970s, the Government of Chile undertook a comprehensive program of economic liberalization. A central component of this program, particularly after 1977, was a drastic overhaul and deregulation of the country’s financial system. This reform was carried out in the spirit of the well-known McKinnon-Shaw prescription that abolishing “financial repression” is essential for sustained economic development (McKinnon, 1973; Shaw, 1973): banks were privatized, regulations were relaxed, interest rate ceilings were abolished, and integration with world capital markets was increased. The financial reform was accompanied by stabilization policies aimed at reducing the persistent macroeconomic disequilibrium, which was reflected particularly in hyperinflation during the early 1970s.

In the mid-1970s, the Government of Chile undertook a comprehensive program of economic liberalization. A central component of this program, particularly after 1977, was a drastic overhaul and deregulation of the country’s financial system. This reform was carried out in the spirit of the well-known McKinnon-Shaw prescription that abolishing “financial repression” is essential for sustained economic development (McKinnon, 1973; Shaw, 1973): banks were privatized, regulations were relaxed, interest rate ceilings were abolished, and integration with world capital markets was increased. The financial reform was accompanied by stabilization policies aimed at reducing the persistent macroeconomic disequilibrium, which was reflected particularly in hyperinflation during the early 1970s.

During the stabilization cum reform program, both the financial and the real sectors of the economy appeared to be making progress for several years before a major financial crisis developed. Between 1977 and 1981, inflation fell (albeit slowly), and output boomed while domestic capital markets vigorously expanded (Table 1). But, by late 1981 the financial system was indisputably submerged in a major crisis. By the end of 1981, the nonperforming assets of banks had reached an estimated 22 percent of capital and reserves and were to rise to 47 percent at the end of 1982 and 113 percent in May of 1983 (Arellano, 1983a). Two successive waves of government “interventions” (one in November 1981 and one in January 1983) were necessary to rescue or liquidate the troubled financial intermediaries. By early 1983, 11 commercial banks (including the nation’s 2 largest private banks) and 5 financial companies (financieras) were under government intervention or had been liquidated. Many other financial institutions survived only because of a generous infusion of government subsidies. The severity of the crisis contributed to a substantial weakening of macroeconomic performance and major readjustments in policies. The fixed exchange rate regime, maintained for almost three years, collapsed in June 1982. During 1982 and 1983, inflation rebounded somewhat, and real output fell by more than 14 percent in 1982 alone.

Table 1.

Main Macroeconomic Indicators, 1975–86

article image

December to December percentage charge; Central Bank of Chile. CPI base: 1978 = 100.

Figures in parentheses are from Cortázar and Marshall (1980) CPI.

December to December percentage changes, International Monetary Fund, International Financial Statistics.

Measured in December of each year Nominal exchange rate adjusted by a trade-weighted index of the exchange rates and CPIs of Chile’s 16 major trading partners. Base: December 1978 = 100.

Cortázar (1983). Uses Cortázar-Marshall price index as deflator.

United Nations (1983). Arithmetic average of quarterly survey results. Does not include Minimum Employment Plan.

Overall deficit of the nonfinancial public sector, as a percent of GDP.

IFS. 1980 prices.

Central Bank of Chile, Boletín Mensual, December Indice General de Acciones. Base: December 1978 = 100.

IFS, in millions of U.S. dollars.

This chapter focuses on the possible causes of the crisis in the Chilean financial system and analyzes the measures adopted to deal with the crisis and their effects. The Chilean process resembles the “classical” boom-and-bust cycle of financial crises, as described by Minsky (1977) and kindle-berger (1978).1 The study therefore examines the upswing in financial intermediation that occurred in the late 1970s and analyzes the weaknesses that eventually led to systemic difficulties. The analysis focuses on the interaction among financial and macro variables and policies.

The macroeconomic policies and outcomes up to the 1982 crisis are well known and have attracted ample academic attention (Corbo (1985); Dorn-busch (1984); Edwards (1985); Foxley (1983); Harberger (1985); Sjaastad (1983); Zahler (1983)).2 it suffices here to outline some of the main changes in the macroeconomic environment that accompanied the financial reform. A major aim of the stabilization policy during this period was to lower the rate of inflation, which had approached hyperinflationary levels in 1973. To a large extent, the inflation reflected the need to finance huge fiscal deficits, which in 1973 had reached 22 percent of GDP.

A tighter fiscal policy was an important component of the policy package. Although the growth of fiscal expenditures was reduced, the Government took mainly strong measures to increase revenue, including tax reform, price increases for goods and services produced by the public sector, and the sale of many enterprises that the previous administration had nationalized. As a result, the public sector achieved a significant overall surplus by 1979.

Monetary and exchange rate policies also were modified. The Government followed a policy of moderate expansion in domestic credit combined with exchange rate adjustments that followed a crawling peg. The strategy was to lower the rate of inflation gradually while improving the reserve position of the Central Bank and preserving the competitiveness of Chilean exports.

The exchange rate policy was changed in December 1977, when the Government introduced a daily schedule for the exchange rate for the ensuing 2 months. In February 1978, this schedule was extended for 11 months, and at the end of that year, it was extended through December 1979. These schedules implied a declining rate of devaluation, with the aim of driving inflation closer to international levels. In addition, it was hoped that preannouncement of the devaluation rates would help remove uncertainty and align domestic interest rates with those abroad. In June 1979, the exchange rate was pegged at Ch$39 per U.S. dollar—the level that, according to the schedule, would have been reached only in December. Moreover, the Government announced that there would be no further devaluations.

Perhaps the most dramatic policy changes occurred in the foreign sector. The authorities moved rapidly to eliminate trade restrictions and sharply reduced both the dispersion and the levels of tariffs, but they moved more slowly in freeing capital movements. Most capital outflows remained restricted through the 1970s, and all new foreign borrowing or refinancing of existing credits by commercial banks, except for short-term lines of credit, remained subject to prior approval of the Central Bank.

These policies succeeded in lowering inflation to less than 10 percent per year by 1981, while achieving a rate of growth in GDP of 7 percent per year over the 1976-81 period. But in 1982, and 1983, the country plunged into a major recession; GDP fell 14 percent in 1982, and unemployment increased from a low of 11 percent in 1981 to more than 22 percent in 1982. This recession had both foreign and domestic causes. The most important foreign causes were the collapse in the prices of Chile’s main export commodities (especially copper), the appreciation of the U.S. dollar, and higher interest rates abroad. The main domestic cause was a deterioration of competitiveness, caused by the combination of a fixed exchange rate and backward-looking wage indexation. The cumulative effect of high domestic interest rates since 1975, following the liberalization of the financial sector, weakened the financial health of enterprises and served to deepen the recession. Furthermore, the Chilean Government addressed the declining competitiveness and deteriorating external balance by restraining domestic demand and lowering nominal wages. When those measures proved insufficient, the Government abandoned the fixed exchange rate of the peso; on June 15, 1982, the peso was devalued by 15 percent. Following further sharp depreciation, the peso declined to Ch$66 per U.S. dollar by September 1982, when it was announced that exchange rate adjustments would follow a crawling peg.

Although inflation accelerated from 10 percent in 1982 to 23 percent in 1983, the economy then began to recover. In 1984, real GDP grew by almost 6.3 percent, while inflation continued at about 23 percent. Recovery continued in 1985, when GDP grew by about 2.4 percent, although inflation accelerated to about 26.5 percent. The rate of unemployment fell sharply to about 12 percent by 1985; real average wages declined steadily between 1982 and 1985.

Against this background, this chapter systematically analyzes the connection between macro and purely financial developments and separates the micro from the macro causes of the financial crisis. It is usually taken for granted that undesirable macro outcomes such as high real interest rates and overvalued exchange rates adversely affect the position of borrowing firms, and thus the financial system as a whole. However, the more subtle question of how the financial sector disequilibrium affects the health of the macroeconomy is seldom asked. This chapter discusses the possibility that the troubles in the financial sector may have contributed substantially to the creation of instability elsewhere in the Chilean economy. In turn, the macro disturbances that those troubles helped create would come back to haunt the banks, via a weakened pool of borrowers and large quantities of nonperforming assets.

This chapter also examines the interrelationships among several widely accepted hypotheses concerning Chile’s financial crisis. The financial reform was quite controversial when implemented, but ex post there is remarkable agreement as to “what went wrong.” Many such explanations, however, are partial equilibrium analyses that focus only on a particular feature of the financial problem (high interest rates, for instance). This paper juxtaposes these explanations and explores their mutual compatibility. Finally, this paper benefits from recently published work (Gálvez and Tybout (1985); Corbo and Sanchez (1985); Arriagada (1985)) on the microeconomic performance of Chilean firms during the relevant period. This information permits a better discussion of some common assumptions about the link between firm bankruptcies and bank problems.

Distributional and welfare issues are dealt with only superficially. The process of financial reform and the measures to deal with the financial crisis have probably had major distributional and welfare consequences, but their measurement and evaluation should be the subject of a separate study.

I. Financial Reform

The process of financial reform began shortly after a new administration came to power in 1973. Broadly speaking, the reform included institutional, regulatory, and international aspects, which will be considered in this section. By 1980, the process of liberalization was essentially completed. From 1981 onward, the changes would be mostly in response to the growing imbalances.

Institutional Changes

The financial system was almost entirely under state control in late 1973. After a long period of mixed ownership, virtually all financial institutions had been nationalized during the period of 1970 through 1973. In addition, a number of entities linked to ministries and state agencies directly supplied credit to sundry activities.

In 1974, Chile had 20 domestic nationalized commercial banks. The following year all but the Banco del Estado were sold to the private sector. The system thus became a predominantly private one, although the Banco del Estado has maintained a large and active presence.

The process and terms by which the banks were transferred to private groups have been criticized. Differential access to credit and weak (and eventually abolished) legislation against concentration of ownership meant that the bulk of commercial banks ended up in the hands of a few large conglomerates, some of which had also recently acquired scores of reprivatized manufacturing and service enterprises (Arellano (1983b)).3 Interlocking ownership and management patterns were eventually blamed for sonic of the mistakes made by these institutions.

Even before the commercial banks were privatized in May 1974, short-term transactions at free interest rates had been allowed. As a result, a new kind of financial entity arose, the financieras, which initially enjoyed great flexibility and a competitive advantage (interest rates on commercial bank deposits were not freed until a year later). In December 1976, problems in some of the financieras prompted an increase in their supervision.4

The development of financieras was only one manifestation of a general increase in new financial intermediaries during the 1970s. The authorities’ efforts to lower barriers to entry brought in foreign banks. The evolution of the system is described in Table 2, which shows that by 1981, Chile had 23 domestically owned banks (1 of which was state owned), 18 foreign banks, and 13 finance companies (financieras).

During this process, the Sistema National de Abormsy Préstamos (SINAP) (National Savings and Loan System) faced serious difficulties. Founded in 1960 to provide housing credit, it captured a large portion of domestic financial savings during the 1960s. The SINAP had made long-term commitments (housing loans) financed with short-term indexed deposits, which were inexpensive as long as interest rate ceilings existed. When the ceilings were abolished in 1975, the SINAP was caught in a squeeze. It effectively went bankrupt in 1976, and the state guaranteed its deposits.

Table 2.

Structure of Financial System, 1974–84

article image
Source: Larrain (1985).

Only one domestic bank has been state owned since 1975.

Regulatory Reform

Chilean bank regulations changed often during the reform period, so space limitations permit only a general description here. The general aim of these changes was twofold: to promote the rapid growth of the financial system and to increase competition.

As has been noted already, controls on interest rates were lifted in 1975. During the period of so-called financial repression (1930-75), the combination of interest rate ceilings and high inflation had consistently yielded negative real rates, but this situation changed after liberalization. It was hoped that positive real yields on domestic financial assets would reduce preferences for foreign or nonproductive assets. Indexed savings instruments and domestic dollar assets were to help in this task. The attractiveness of domestic financial investments was also to be enhanced by changes in tax legislation, which took account of inflation and ensured that thereafter only real interest earned counted as taxable income.

Another much-criticized feature of the previous system—an array of quantitative controls on credit—also was eliminated. Selective credit controls had been a widely used tool of economic policy until 1975, but after that date, it was expected that the market mechanism alone would allocate credit and improve the allocation of resources.

To reduce the cost to banks of holding required reserves, which could become quite substantial under inflationary conditions, the Central Bank began paying competitive interest on such reserves in May 1976. Subsequently, reserve requirements were gradually lowered, to 10 percent for sight deposits and 4 percent for time deposits, and interest payments on reserves were phased out.

At the same time, the system moved toward multipurpose banking. To increase competition and lower costs, distinctions among commercial, investment, mortgage, and development activities were abolished. In addition, foreign banks were also allowed to open branches in Chile and to purchase Chilean banks. The result was rapid financial widening. According to Luders (1986):

Savings operations (before a monopoly of the Banco del Estado) were permitted for all banks, and housing mortgage transactions (previously a virtual monopoly of the SINAP) were also expanded to all banks. The volume and diversification of government and Central Bank papers in the market increased noticeably; the range and number of mutual funds increased manifold; businesses began to issue significant amounts of commercial paper which were intermediated by depository institutions, stock exchanges and mutual funds; the insurance business expanded its list of products; consumer credit offered by financial institutions expanded noticeably, etc.

New regulations also increased the powers of the bank supervisory-agency (Superintendency of Banks) and changed various aspects of financial activity, such as capital requirements and credit limits. Those that are most relevant for the purpose of this paper are as follows:

  • The jurisdiction of the Superintendency of Banks was broadened to include all financial institutions and its name was changed to Superintendency of Banks and Financial Institutions.

  • The Superintendency was authorized to provide general information about the quality of the assets and liabilities of financial institutions in order to increase market transparency.

  • In November 1974, the capital requirements for banks were increased to take account of past inflation; thereafter, these requirements were automatically adjusted once a year in accordance with the Consumer Price Index (CPI). The maximum debt/capital ratio was maintained at 20, but noncompliance was penalized more strictly, at the rate of 2 percent daily on the excess debt.

  • In 1974, the law limited the maximum individual holdings of bank shares to 1.5 percent (for individuals) and to 3 percent (for firms and organizations). However, these limits were abolished in 1978 because they had proved very difficult to enforce.

  • In 1980, the treatment of bank borrowers was made more uniform, with regard to their credit limits in relation to bank capital and reserves. Previously, the legislation favored corporations over other organizations and individuals. The new regulation set a uniform limit at 5 percent for unsecured credits and at 25 percent for secured credits. These limits were halved in the case of borrowers linked to the bank.

  • In 1980, to facilitate underwriting operations, the limit on bank investment in a given enterprise’s shares was raised from 10 percent to 20 percent of the bank’s paid capital and reserves, but the limit of 10 percent of the firm’s capital was maintained.

  • Limits on bank borrowing abroad were modified several times, and finally abolished in 1980.

These changes gave more freedom to the financial markets and placed more emphasis on rules than on the authorities’ discretion in supervising the system. As the financial crisis unfolded in late 1981, some of these regulatory provisions were modified, as discussed in Sections IV and VII.

Capital Flows

The liberalization of international capital flows has been one of the most widely discussed features of the Chilean experience. The process is discussed in detail in Arellano and Ffrench-Davis (1981), Ffrench-Davis (1983), Mathieson (1979), McKinnon (1982), and Edwards (1984).

After many years of controls, the capital account was liberalized gradually. Starting in August 1976, capital inflows were subjected to a minimum maturity requirement of two years (between 1974 and 1976 the minimum requirement was only six months), and to a varying but small deposit requirement (non-interest-bearing) with the Central Bank.5 This minimum maturity remained in force until the crisis in 1982, when it was abolished.

Banks were allowed to borrow abroad in dollars but not to assume the exchange risk. The principal of the counterpart domestic loans had to be indexed to the exchange rate (i.e., de facto denominated in foreign currency). There were also some limits (relaxed over time) on banks’ foreign indebtedness as a percentage of capital and reserves.

The chief purpose of such restrictions was to ensure some control over domestic monetary policy. It was feared that, given very high domestic interest rates, total liberalization would induce huge and destabilizing capital inflows.6 Despite the restrictions on capital inflows, the inflows proved massive, as seen from Table 3, reflecting in part the policy of preannouncing the exchange rate until mid-1979, which made investing in Chile very appealing.7 Particularly massive inflows of capital occurred after the change in the exchange rate regime to fixed rates. The monetization of the ensuing reserve buildup was blamed for the slow progress in reducing inflation (Harberger (1985); Corbo (1985)).

As Table 4 suggests, this process led to the rapid accumulation of external debt. In contrast to the situation in Brazil, Mexico, and other Latin American countries, the bulk of the borrowing in Chile was done by the private sector: by 1982, 73.2 percent of net foreign debt had been incurred by the private sector. Most important for the discussion here, financial institutions accounted for the lion’s share of private external borrowing (72.9 percent of total inflows, private and public, in 1981).

A drastic and thoroughgoing trade liberalization also was carried out. Quantitative controls were abolished and tariffs swiftly rationalized and lowered; by June 1979 a uniform tariff of 10 percent (excepting automobile imports) was in place. The fact that liberalization of the capital account lagged substantially behind that of the current account in Chile has been much discussed. Both ex post and ex ante it has been argued that this sequence maximized the chances for adequate macroeconomic control (McKinnon (1982); Edwards (1984); and Calvo (1986)). Indeed, McKinnon (1982) has gone as far as to say that “Chile is to be treated as a norm or standard of reference” in this regard. But experts disagree on whether the speed and magnitude of the liberalization process were adequate.

Table 3.

Net Capital Inflows, 1977–82

(In millions of U.S. dollars)

article image
Source: Le Fort(1985).

II. Effects of the Financial Reform: Some Macroaggregates

A central objective of financial reform is the improvement of the economy’s savings and investment performance. This section reviews some of the available evidence on the course of domestic savings and real investment. The discussion, which is far from exhaustive, is intended only as background for subsequent sections.

Domestic Savings

The impact of financial liberalization on the economy’s overall rate of savings is difficult to determine a priori. First, liberalization affects only the institutionalized part of the financial market: little can be said about its effects on noninstitutionalized sources of finance, which tend to be important in repressed financial systems. Second, microeconomic theory provides no unambiguous propositions on the effect of higher interest rates in the institutionalized market—a typical first-round effect of financial liberalization—because the income and substitution effects point in opposite directions. Moreover, higher interest rates may affect the distribution of income among sectors whose propensity to save may be different (e.g., enterprises and households). Empirically, the matter is also far from being settled.8 In his recent empirical paper Giovannini (1985) concludes that in developing countries, the interest elasticity of savings is likely to be small. The experience of Chile also seems to point in this direction.

Table 4.

Outstanding External Debt, 1975–83

article image
Source: Central Bank of Chile.

Outstanding medium - and long-term public and publicly guaranteed debt repayable in foreign currency.

Includes suppliers’ credits to private sector; lines of credit for imports of capital goods to commercial banks. Banco del Estado, and development banks; credits to private sector under Articles 14, 15, and 16 of International Exchange Law and DL 600; and short-term lines of credit to commercial banks. Banco del Estado, and Central Bank of Chile.

GDP in current Chilean pesos converted to U.S. dollars at the following exchange rates: 1975, 4.911; 1976, 13.054; 1977, 21.529; 1978, 31.656; 1979, 36.80; 1960, 39.0, 1981, 39.0; 1962, 50.91; and 1983, 78.60.

Sum of medium- and long-term public debt and other external liabilities.

Table 5 presents some evidence on the evolution of savings in Chile. The data show significant year-to-year variation in the ratio of gross national savings to GDP, with a peak of almost 20 percent in 1974 and a trough of about 2 percent in 1982. Moreover, the average saving ratio for the liberalization period (1974-83) did not differ significantly from previous periods: for that period the saving ratio averaged 10.7 percent, compared with about 12 percent between 1966 and 1973; however, excluding the recession years of 1982 and 1983, the ratio increases to just 12.6 percent (only 11.6 percent, if the 1974 peak is excluded). The paucity of data makes it even harder to draw conclusions about the distribution of savings between the private and the public sector. Nevertheless, available evidence suggests that the private sector was almost always in deficit and the public sector in surplus, and that the sectoral saving ratios also show substantial variation during the period. The privatization of the social security system completed in 1981 may have had a negative impact on total savings, insofar as part of the contributions previously paid into the state social security were spent instead of saved.


The data in Table 6 indicate a significant increase (as a percentage of GDP) in private investment over the 1975-81 period followed by a sharp decline between 1982 and 1985, reflecting the effects of the recession and the financial crisis. The sharp increase until 1981 is particularly significant in light of the high real interest rates on loans charged during this period. To a large extent, this increase in private investment was offset by a decline in public investment, which was part of the program to reduce the fiscal deficit and the size of the Government. As a consequence, average gross fixed investment as a percentage of GDP was only marginally higher between 1975 and 1982 than it had been between 1966 and 1974.

Therefore, the effect on GDP growth of the shift in the composition of investment would depend on whether the marginal productivity of the projects undertaken by the private sector was higher or lower than the marginal productivity of alternative public sector projects. A definite answer to this question would require further study, which is beyond the scope of this paper.

Table 5.

National and Domestic Savings, 1960–85

(As percentage of GDP)

article image
Sources: United Nations, National Accounts Statistics: Analysis of Main Aggregates, 1983/84, New York (1987); and Central Bank of Chile.Note: Figures in parentheses relate to net savings of the public sector.
Table 6.

Gross Domestic Capital Formation, 1960—85

(In percentage of GDP)

article image
Source: Central Bank of Chile, Dirección de Politica Financiera, Cuentas Nacionales de Chile.Note: Figures in parentheses relate to private sector fixed investment.

III. Effects of Financial Reform: Growth of Financial Assets

One of the most remarkable features of the Chilean experience after 1975 is the tremendous expansion of financial intermediation. As Table 7 shows, total financial assets rose from 19.7 percent of GNP in 1975 to 48.1 percent in 1982. The share of the organized financial system in overall financial intermediation rose as well: papers issued by commercial and development banks and financieras rose from 16.2 percent to 70.7 percent of total financial assets in the same period. Between 1976 and 1981, value added from financial services expanded at a real annual average of 18.4 percent, more than two and a half times the rate of expansion of the economy as a whole.9

Table 7.

Financial Assets1

article image
Source: Central Bank of Chile, Boletín Mensual.Note: Definition of total financial assets excludes equity and time and demand deposits held by the public sector.

These are financial assets held by the public (i.e., liabilities of the financial system).

Includes currency and demand deposits and time deposits held by the private sector.

Paper issued by commercial banks, development banks, and financieras.

At the same time, there were significant shifts in the composition of financial assets. A crucial shift was the move away from money and toward short-term, highly liquid, interest-bearing assets. Quasi-money accounted for 8.16 percent of GDP in 1975 and 22.32 percent in 1986.10 This tendency follows naturally the upward trend of short-term interest rates paid on quasi-money after liberalization and the richer menu of financial assets available to the public.

The boom in financial intermediation, however, did not substantially lengthen the average maturity of financial instruments. As shown in Table 8, much of the expansion in financial instruments came from the shift from money to interest-earning assets with maturities of less than 90 days. Longer-term assets (with maturities over 90 days) languished. They accounted for 54 percent of total assets in 1975, and for only 44.4 and 37.2 percent in 1980 and 1981, respectively.11

These figures do not include equity markets, which experienced a boom of their own. The (real) stock price index rose by 2,685 percent between 1976 and the peak year of 1980 (Table 1). Prices of real estate, land, and other comparable stores of value rose dramatically as well.

This boom in financial intermediation seems inconsistent with the poor performance of domestic savings. An influential view, formulated for the general Southern Cone case, offers an explanation that may be worth quoting at some length:

In a financially repressed economy with a history of persistent inflation, wealth is held as money, land and capital. . . . Money is held because of its property as a means of payment; capital, because of its expected yield in use; and land, as a shelter against inflation. Expected land yields may be low, but they are strongly correlated with inflation rates. . . . In this context, financial reform-mongers typically propose introducing an indexed government bond as an instrument of financial liberalization.12 In the presence of such an attractive asset with a strong back-up market, saving propensities should increase and a higher proportion of Wealth should be held as productive capital .... (However), indexed bonds tend to replace capital (and money) rather than land in private portfolio holdings. True-market-oriented financial reforms are accompanied by a general liberalization of interest rates, in the context of a demand-contractionary package of policies. . . . An excess supply of money may also obtain, in spite of contractionary policies, if the demand for money is sufficiently lowered by the introduction of the indexed bond (Díaz-Alejandro and Bacha (1982)).

This critical view of the consequences of financial liberalization captures some of the tendencies described so far: the extent to which the growth in some financial assets was the consequence of substitution away from money; the boom in interest-bearing financial assets; and the sluggishness in the demand for real capital. But, it fails to account for the boom in overall financial intermediation, not just in close money substitutes. It also neglects the upward dash of the price of the existing capital stock, which took place even as the demand for investment goods remained low. An explanation of these phenomena requires a closer look at the details of the Chilean situation.

Table 8.

Maturities of financial Assets, 1973–82

(In percentage of total)

article image
Source: Arellano (1983b).Monetary assets = Private money holdings: Currency + demand deposits.Less than 90 days = Pagarés Descontables Banco Central (PDBC), Depósitos y Captaciones Bancos y Financieras, Ventas PDBC con pacto retrocompra, Ahorro a la vista, Emisión Pagarés Descontabies de Tesorería (PDT), PDBC Reserva Técnica, CEPAC, Venta Carters Bancos, Operaciones con Cuentas de Ahorro Sistematico (CAS).90 days to 1 year = Cuotas Ahorro Corvi, Depósitos y Captaciones Bancos y Financieras, Ahorro a Plazo, Depósitos de Ahorro e Inversión (DAI), Cuentas de Ahorro SINAP, Valores Hipotecarios Reajustables (VHR), Ahorro Sistemático Cooperative, Depósitos Plazo Banco Estado.More than 1 year = Cuentas Ahorro Reajustables (CAR), Pagarés Reajustables de Tesorería (PRT), Bonos y Letras Hipotecarias, Bonos Hipotecarios SINAP, Debentures, Pagarés Banco Central Sistemático Previsión, Bonos de Reconstrucción, Pagares Reajustables Caja Central.Note: Until 1975 all time deposits in banking system are classified in column (2).

A commonly mentioned cause for the spectacular growth of financial assets is the increase in private wealth. It is conjectured that increases in expected permanent income or in the yield of assets held by the public created an upward wealth effect, some of which spilled into higher demand for financial assets. As Barandiarán (1983) puts it:

The change in expectations toward higher wealth was the decisive factor in the extraordinary expansion of private demand for goods and services. . . . The financial system grew alongside private sector wealth: capital gains realized by the owners of real assets turned in part toward the financial system, producing a monetization of real private assets (in the sense that highly liquid paper was issued on the basis of these assets).

How solid a foundation this process constituted for the growth in bank and financiera liabilities is hard to assess. Whether increases in expected permanent income were effective or only perceived must be a moot question, but some of the optimism displayed at the time seems justified in the wake of a massive removal of distortions, a change that had the potential for increasing real income. High interest rates produced attractive short-term yields on financial assets, which also created a perception of increased wealth. Finally, booming stock, land, and real estate prices certainly increased wealth and hence the accumulation of financial assets. However, it seems very likely that such exploding asset prices were unrelated to market fundamentals (like the conceivable real yield on capital and land) and hence unsustainable. In fact, econometric evidence (Meller and Solimano (1983)) suggests that indeed there was a speculative bubble in stock prices, of the sort described by Blanchard and Watson (1982).

Another key element in the expansion of financial assets was the reorientation of savings toward the financial system.13 During the period of “financial repression,” the Government served as a channel for an important portion of national savings. These were in turn directly invested by the Government or lent to state enterprises and the public. As these uses of funds declined after 1973, more and more of the resources obtained by the Government were channeled through the financial system (the liabilities of the consolidated financial system vis-à-vis the Government rose). Another important change was privatization of the social security system, beginning in May 1981. The bulk of social security contributions were thereafter deposited in the financial system. By year-end 1982 such funds accounted for 7.7 percent of total deposits.

Finally, the single largest source of the expansion of domestic bank liabilities was foreign borrowing. Particularly between 1978 and 1981, capital inflows exceeded current account deficits, leading to a buildup of international reserves (Table 1). Whether the borrowers were domestic banks or firms, such funds would eventually find their way into higher deposits in the financial system.14 At the end of 1981, reserves amounted to almost 10 percent of GDP and to 35 percent of all paper issued domestically by the financial system.

Of course, capital inflows are not an exogenous variable. In Chile, the bulk of foreign borrowing was contracted by domestic banks, which actively sought foreign loans. The same can be said of domestic deposits: banks and flnancieras have ample means at their disposal to regulate their acquisition of deposit liabilities and are willing to attract funds only insofar as they expect to lend them out again, at a profit. The expansion in bank liabilities described must have had a counterpart in a corresponding increase in bank loans. Indeed, it is likely that vigorous growth in the demand for bank credit may have acted as a “pull” factor, prompting banks to attract additional resources to meet this demand. To explore such a conjecture we must examine the course and nature of bank lending, a subject to which we now turn.

IV. Effects of Financial Reform: Expansion of Domestic Debt

Between 1974 and 1982, the accumulated stock of debt of the nongovernment sector went from 5.0 percent of GDP to 61.7 percent (Table 9). The inability of private economic agents to pay interest and principal on such a staggering stock of debt was eventually to become the key weakness of the Chilean experiment.

Much of the debt was denominated in dollars, a factor that increased the service burden after the 1982 maxi-devaluation. Table 10 offers a breakdown of total loans of the financial system by currency and maturity. In 1981, peso loans accounted for almost 62 percent of the total, and dollar loans for the rest. Moreover, such debt was predominantly short term: 67 percent of loans denominated in pesos in 1981 had maturities of less than one year. Dollar loans were somewhat longer term, given the two-year minimum maturity required of capital inflows, but cheaper dollar credit tended to be rationed and was inaccessible to many.15 The short-term nature of credit available to firms was a cause of concern throughout the period. As early as 1976 Corbo warned, “Domestic currency debt is made up almost entirely of 30-day loans. These credits are renewable according to different mechanisms, but firms are subjected to a rate of interest they cannot predict beyond thirty days and are faced with the problem of renegotiating each month the total of their credits.”

Table 9.

Loans of Financial System to Nongovernment Sector, 1969–82

article image
Sources: 1969–78, Central Bank of Chile, Series Monetarias; 1979–82, Central Bank of Chile, Boletín Mensuai; and 1979—82 (total loans only), Superintendency of Banks and Financial Institutions, Información Financiera.

Starting in 1979 includes development banks and financieras.


Table 10.

Currency and Maturity Structure of Loans, 1978–82

(In billions of pesos)

article image
Source: Arellano (1983b).Note: Overdue loans are excluded.

The first factor to explain the growth in debt ratios (Table 9) is the financial euphoria that seems to have overtaken Chilean economic agents starting sometime in 1977 and peaking in 1980 and 1981.16 Barandiarán (1983) has maintained that “the generalization of optimistic expectations about the prospects for the national economy was the main cause of the increased indebtedness of firms and households in 1980 and 1981.” But as Barandiarán himself emphasizes, such euphoric expectations are hard to justify, even after taking into account hopes about increased national permanent income: these hopes cannot explain a tenfold increase in real domestic indebtedness.

A substantial amount of bank credit directly or indirectly financed an increased demand for consumer durables. As already mentioned, years of import restrictions had created pent-up demand for imported consumer durables, whose low relative price during the period of peso overvaluation made them even more attractive. At the same time, expectations of a real devaluation or of a return to higher tariffs or both made overstocking a rational course of action (see Calvo (1986)). As a result, there was a flood of consumer imports, many of which were purchased with bank credit at variable interest rates. This last feature made repayments particularly difficult in 1981 and 1982, as real peso rates rose. Similarly, dollar credit was contracted for cheaply, but it became almost prohibitively expensive after the devaluation. Repossession of items purchased on credit became common in 1982.

Other practices added a great deal to this boom in credit demand. An important share of credit was extended to grupos (financial and manufacturing conglomerates) seeking to purchase firms or other existing assets. An unusual sort of competition took place among these conglomerates, in which market share or sheer size seems to have mattered more than profitability or efficiency. In a few years, these groups had acquired practically all the nation’s largest manufacturing and banking firms, and in the process had sent stock prices soaring. Whether these groups acted speculatively (expecting to sell later at a higher price) or whether they thought the exorbitant purchase prices were justified by high present value of expected real return is irrelevant. The point is that asset prices behaved “as if a speculative bubble were taking place (Meller and Solimano (1983)), and the process probably absorbed a significant share of the nation’s available credit.

Stock prices also were inflated by the practice of grupo firms of trading stock among themselves, thus boosting the price of their shares above market value. These stocks were then used as collateral for bank credit. This was but one of many practices intended to facilitate the use of bank credit by grupo firms.17 Often such credit was used to purchase banks themselves, in an operation that Luders (1986), following the terminology of the time, terms a “bicycle”: “The bank would grant a loan to a corporation controlled by the new owners of the same bank; and the corporation would use the proceeds of the loan to pay for the shares it was acquiring.”

The consequences were twofold. First, a good portion of bank credit went to finance the consolidation or expansion of grupos. Second, loans to firms belonging to the same conglomerate as the lending bank came to account for a substantial portion of the banks’ portfolios; this problem became known as the problem of the cartera relacionada. By June 1982, the cartera relacionada accounted for 21.1 percent of the loans of the five largest private banks. In the ease of the Banco de Santiago, the nation’s largest private bank, this figure was 45.8 percent (Arellano (1983b)).

Finally, perhaps the single most important factor behind the growth of domestic indebtedness was the rolling over of credits and the capitalization of interest. As Arnold Harberger (1985) has put it, in Chile there existed “substantial’false demand.’ The false demand for credit consists of the rolling over of what are essentially bad loans.” By 1981, many borrowers were in a tenuous situation. As Gálvez and Tybout (1985) have documented, real peso overvaluation and sustained high real interest rates were major sources of difficulties for manufacturing firms, particularly those competing in the export and import sectors. Tables 11 and 12 show estimates of firm bankruptcies and of loans in default, respectively. By 1981, the figures for loan defaults amounted to 2.3 percent of the portfolio of the total financial system. This share rose to 8.2 percent in 1982 and 18.5 percent in 1983 (see Table 12). The implicit dangers, however, were greater than these figures would suggest. Chilean banks as a whole were undercapitalized with respect to their historical averages (Behrens (1985)), so that the ratio of nonperforming assets to total capital and reserves was extremely high: Arellano (1983a) estimates the ratio at 11 percent in 1980, 22 percent in 1981, 47 percent in 1982, and 113 percent in 1983.

Table 11.

Number of Total Bankruptcies, 1974–82

article image
Source: Fiscalía Nacional de Quiebras, taken from Luders (1986).
Table 12.

Loan Defaults, 1974–83

(As percentage of total loan portfolio)

article image
Source: Behrens (1985) , taken from Luders (1986).

Including “risky portfolio” loans sold to Central Bank.

Furthermore, the line between a performing and a nonperforming asset becomes fuzzy when rollovers and capitalization of interest are widely used to keep many problem loans on the books. There is substantial consensus (Arellano (1983b); Zahler (1985); Harberger (1985); and Luders (1986)) that such practices accounted for a large share of the expansion in bank credit. Under the extremely high interest rates, interest accumulated on domestic currency loans between 1977 and 1982 would add up to 72 percent of outstanding peso loans by year-end 1982.

Real interest rates (corrected by the CPI) on loans averaged 77 percent per year between 1975 and 1982 (see Table 13). Why would firms and households continue to borrow in pesos at rates so far above any conceivable return on real investment? One possibility is that, locked into short-term credits, firms continued to borrow with the expectation that “interest rates would soon decline.” Another possibility is that the real rates we can compute are ex post, while ex ante expected rates were much lower. In 1981 and 1982, for instance, real rates rose largely because of a swift decline in inflation associated with the appreciation of the dollar in international markets.18 Such mistakes in expectations can perhaps explain a one-time increase in indebtedness, but not a sustained increase such as that witnessed by Chile. In the latter part of the period it is likely that many firms were willing to borrow at any ex ante rate simply to stave off bankruptcy. On the basis of data obtained by Gálvez and Tybout (1985), net earnings dropped sharply after 1980, especially for firms producing tradables, and the gearing ratio of firms began to rise. This situation is consistent with the hypothesis that capitalization of interest and other related lending can be considered “distress borrowing” for a significant number of firms.

Table 13.

Alternative Real Loan Rates, 1975–83

(In percent)

article image
Sources: Central Bank of Chile, Boletín Mensual, Cuentas Nacionales, 1960–83.

Nominal loan rate in pesos corrected by CPI.

Nominal loan rate in pesos corrected by the implicit deflator of fixed capital formation statistics. See Cuentas Nacionales.

Nominal loan rate in pesos corrected by IGPA (Indice General Precios de Acciones).

Arithmetic average.

It has been argued that firms acted this way because continued postponement of bankruptcy made a government bailout more likely. According to Arellano (1983b).

As borrowers become aware that this is a generalized phenomenon that has no micro but only macroeconomic solutions, they become indifferent to the interest rates they are charged. Debts become mere entries in the books—banks cannot collect them. This must have been the perception of large debtors—certainly the conglomerates—during much of the period, and has certainly been the perception of the immense majority of debtors since 1982.

What borrowers expected ex ante we may never know; but ex post those who expected that mounting debts would bring some kind of official help turned out to be correct.

Although many institutional features of the post-reform Chilean financial system facilitated excessive risk taking and unsound lending patterns, the legislation aimed at curtailing such patterns was weak or nonexistent until 1980, and the Superintendency of Banks allowed those practices insofar as they did not violate the letter of the law. In 1980, the Superintendency began to set up a system for the classification of loans according to risk—irrespective of whether they were overdue—and the corresponding rules were established for “individual” provisions for nonperforming loans to supplement the “overall” provisions.19 In late 1981, the authorities adopted measures that limited the amount of bank exposure to a single enterprise and to a bank’s own subsidiaries. But it was not until 1982 that a set of comprehensive measures that tightened bank supervision were approved. The regulations included a more precise definition of the limit on loans to a single enterprise, which took into account the interlocking ownership of firms.20 In the case of banks classified as unstable or poorly managed, the Superintendency was empowered to regulate the bank’s new lending, rolling over of existing credit, and collateral requirements. The Superintendency also began to develop a formal system of rating financial institutions based on “CAMEL” (capital, asset quality, management, earnings, and liquidity) indicators, and commercial banks were prohibited from investing in equity capital, agricultural land, merchandise, or livestock and from accepting stock as loan collateral (so as to limit the use of inflated stocks as collateral).

Minimum capital asset ratios in force in Chile at the time also made it expensive for banks to write off loans. When a loan is declared to be bad, capital and surplus must be reduced by the amount of bad debt, but the loan portfolio must be reduced by an even greater amount because of regulations on the minimum capital asset ratio. Banks were quite unwilling to do this at a time of high loan rates and booming credit demand.

Furthermore, banks were able to engage in this sort of risky lending because they were not subject to the discipline of depositor or state supervision. There were no explicit peso deposit guarantees in Chile until January 1983, but apparently there was a widespread perception that the Government would rescue depositors in the event of a bank crunch. This perception was reinforced at the time of the collapse of Banco Osorno in 1976, when the Government granted a 100 percent bailout of depositors and other creditors.

The issue of implicit guarantees in Chile has been widely discussed. That peso deposits were perceived to be guaranteed seems clear: there were no major bank runs even as practically all the nation’s major banks teetered on the verge of collapse, and this fact was amply discussed in policy circles and even in the press.21 Whether bank owners or managers perceived their investments or reputations to be guaranteed is clearly a different matter.22 If this was the case, it has been argued, bankers had incentives to undertake excessively risky investment. Again, ex ante expectations are unclear, but some ex post observations can be made. Starting in late 1981 many bank managers lost their jobs, and some even went to jail. When the Government intervened in a bank, the owners temporarily lost control over the bank, and eventually saw their ownership diluted by the issuance of new stock. Many banks, whether subjected to intervention or not, received generous subsidies.23 The alternative outcome for comparison is not the entirely unrealistic one in which bank owners or managers lose nothing, but the one in which banks are immediately liquidated, asset prices plummet, and bankers are left with little or nothing in their hands. Compared with the alternative outcome, Chilean bankers appear to have enjoyed a substantial government “guarantee” ex post.

At any rate, if an implicit deposit guarantee had not existed, the logic of the market would have necessarily disciplined bankers. Realizing that many banks were in difficulties, depositors would have transferred their funds toward the more solid institutions. If accounting were pristine and information flows efficient, the collapse of unsound banks would not have affected the rest. If, conversely, a panic had been generated, the system might have failed anyway, but much earlier.24,25

The issue of guarantees to foreign creditors warrants special mention. As Díaz-Alejandro (1985) has emphasized, the Chilean foreign borrowing was largely private and was contracted under explicit government assurances that it would remain so, but the story turned out to be very different. In the words of Arnold Harberger (1985).

The major international banks seem to have acted in concert, leaving the Chilean government no serious alternative but to assume the position of a reluctant guarantor. There can be little doubt that if each foreign creditor bank and each Chilean debtor bank had been left to work out its financial affairs under the applicable laws, a fair share of the foreign debt of the (failed or failing) Chilean banks would have been written off, and Chile’s current debt service problems would consequently have been less.

V. Effects of Financial Reform: Interest Rates

One of the most puzzling features of the Chilean experience is the behavior of interest rates. Accordingly, it is one of the most thoroughly studied (Arellano (1983a and 1983b); Sjaastad (1983); Cortés (1983); Ffrench-Davis (1983); Meller and Solimano (1983); Corbo and Matte (1984); Rosende and Tosso (1984); Zahler (1985); Harberger (1985); Edwards (1986)) with a variety of methods. Zahler (1985) examined events chronologically and attempted to ascertain the changing weight of different factors in interest rate determination. But it is also possible to put forth a few hypotheses that, with some caveats, could apply to the period as a whole. We have taken the latter approach here, drawing generously from the published work on the subject.

Three main features characterize the behavior of Chilean interest rates in this period:

  • High nominal peso rates, not easily explainable by international interest parity considerations;

  • Extraordinarily high real interest rates on peso loans, exceeding an annual average of 76 percent between 1975 and 1982; and

  • A high spread between loan and deposit rates denominated in pesos.

The subsections that follow discuss these features in some detail.

Table 14.

Dollar Interest Rates, 1975–83

(Annual percentages)

article image
Source: Zahler (1985).

Nominal peso rates on short-term operations (30-90 days) minus actual exchange rate devaluation.

Annual average of monthly data on six-month dollar rates.

Loan rate minus LIBOR.

Simple arithmetic average.

Nominal Interest Rates

As Table 14 reveals, nominal peso deposit rates (corrected for ex post devaluation) exceeded LIBOR for much of the period, except for the quarters after the maxi-devaluation of June 1982. A similar phenomenon emerges for peso loan rates. The differential is also substantial in this latter case, and highly positive for all years except for 1982. Only a small fraction of this differential can be attributed to the premium over LIBOR charged by international lenders to Chile. The average premium charged on loans to Chile was 1.55 percent in 1978, 0.99 percent in 1979 and 1980, 0.89 percent in 1981, and 0.97 percent in 1982 (Edwards (1986)). In short, nominal interest rates were “abnormally” high, given that accepted theory would predict a tendency toward interest rate equalization.

Of course, such a prediction assumes substantial capital mobility. It is useful to recall, however, that there were restrictions on capital inflows (albeit in declining magnitude) throughout the reform period.26 These might explain a small portion of the remaining differential, but as these restrictions were relaxed over time, the spread fell only slightly.

Restrictions were applied not only on the length and type of capital inflows but also on the uses to which they could be put by domestic banks. As already mentioned, domestic banks were not permitted to take positions in foreign currency, and hence to arbitrage directly.27 Sjaastad (1983) has argued that these restrictions, by requiring several transactions, made arbitrage very costly and therefore may be held responsible for the stubborn spread. Under such circumstances, “a Chilean banker may be indifferent between paying 3 percent per month for domestic funds as opposed to 1.5 percent for dollars.” But as Cortés (1983) has pointed out, Chilean banks at no time seemed bent on arbitraging, and they did not oppose the restrictions. The two principal uses of foreign credit were for grupo loans and foreign trade financing. Hence, Cortes concludes that the Sjaastad hypothesis, although theoretically correct, is probably quantitatively unimportant.

Another commonly held explanation of deviation from interest parity (Arellano (1983b), and Zahler (1985)) alludes to market segmentation: borrowers who were privileged by their contacts or collateral (presumably the grupos) could borrow cheaply abroad; other borrowers had to resort to expensive credit in pesos. The idea is appealing in its simplicity and informal observation suggests that it probably has some validity, but it is difficult to test rigorously. Further doubts are cast by the Galvez-Tybout observation (1985) that grupo firms do not seem to have enjoyed abnormally low borrowing costs.

Other authors (Corbo and Matte (1984)) have emphasized that Chilean and foreign assets were imperfect substitutes, and hence there was little reason to expect a convergence to interest parity. They have provided evidence to this effect by estimating a simple monetary model with three assets: money, domestic interest-earning assets, and foreign assets. In this model, under perfect substitutability and no sterilization, a decrease in domestic credit by the Central Bank should be matched by an equal capital inflow (reflected in higher reserves), which would leave the monetary base constant. For Chile, the estimated offset coefficient was —0.34, suggesting that asset substitutability was less than perfect (Corbo and Matte (1984)). The logic of this argument is indisputable, but once again its applicability to the specifics of the Chilean interest rate situation is unclear. Whatever the reason (and despite this apparent limitation in asset substitutability), capital inflows to Chile were massive and unprecedented. That they had only a second-order impact on interest rates is still puzzling, even in the context of imperfect asset equivalence.

What about devaluation expectations? Even in a partially open economy, expectations of devaluation must have had some effect on peso interest rates. The problem, of course, is how to compute such expectations. Ex post devaluation was zero in 1980 and 1981 and very large in 1982, but fears of an abandonment of parity may have already been alive in 1981. Using Bayesian methods, in fact, Le Fort (1985) has estimated that the expected rate of devaluation rose from 2 percent in July 1979 (a month after the nominal rate was fixed) to more than 26 percent in May 1982, just prior to the maxi-devaluation. Assuming away such expectational problems, Edwards (1985) estimated an equation for the nominal interest rate as a function of ex post realized devaluation and other variables. The coefficient on the devaluation variable was significant and had the right sign, which suggests that devaluation expectations did indeed play a role.

Real balances (lagged one period) also were included in the Edwards equation, and also proved to be significant. This result fits the framework postulated by Edwards and Khan (1985): in an economy that is only partially integrated with world capital markets, domestic monetary policy will normally play a role in determining interest rates.

Real Interest Rates

Real interest rates in pesos (for both loans and deposits) were extraordinarily high between 1975 and 1983 (Table 15). In the words of Meller and Solimano (1983), “This sole fact inevitably had to lead to an economic and financial collapse.” Certain features of the evolution of real rates over time can be explained with ease: After the capital account was partially liberalized in 1979, rates tended to fall; in 1981 and 1982, when inflation was lower than expected,28 rates surged upward. Rosende and Tosso (1984) have suggested such a sequence. Using a simple Fisher-type consumption-savings model, they argue as follows: In the early years of the period, increases (real or perceived) in permanent wealth tended to reduce savings and to push up the real interest rate. This pressure was relieved somewhat by the entry of large amounts of foreign savings after 1979. As the crisis approached, capital inflows dried up. In the midst of a recession, domestic savings dried up as well, and the real interest rate shot up once again. The dynamics of such a process are highly plausible. However, even at their lowest in 1979 and 1980, real interest rates on peso loans averaged 19.4 percent (LIBOR in those two years averaged 13.05 percent)- Additional explanations still seem to be warranted.

Table 15.

Real Interest Rates, 1975–83

article image
Sources: Central Bank of Chile; monthly rates taken from Boletín Mensual (several issues) annualized and corrected by official CPI. Rates correspond to short-term (30-90 days) bank transactions.

Arithmetic average.

As Edwards and Khan (1985) have emphasized, in a semi-open economy both “external” and domestic factors should help determine interest rates. The obvious external factors in this case are the international real interest rate and the expected real devaluation. Edwards (1985) reports that both factors seem to have been significant.29 By contrast, Meller and Solimano (1983) attempted to assess the impact of the expected nominal devaluation (estimated with parallel exchange market data on ex post devaluation) on the real interest rate. Estimation of a reduced form of the Dornbusch (1980) model yielded a coefficient with a “wrong” sign for expected devaluation, suggesting no effect. Hence, the evidence is mixed concerning the effect of devaluation expectations on the real rate of interest. Casual empiricism, however, suggests that in 1979 and 1980 (at least) both the policymakers’ commitment to the fixed exchange rate and the level of international reserves were high. At that time, expectations of devaluation must have been very low, yet nominal deposit interest rates were still in the 40 to 50 percent range (Table 16). As far as demand factors are concerned, demand for real liquidity was fueled, as already mentioned, by the “bad loan problem” faced by banks. Harberger (1985) identifies this “artificial demand for credit” as one of the two key reasons why real interest rates in Chile were so high (the other being the “decapitalized” state of Chilean firms).

Table 16.

Nominal Interest Rates, 1975–83

article image
Sources: Central Bank of Chile; monthly rates taken from Boletín Mensual (several issues) and annualized. Rates correspond to short-term (30–90 days) bank transactions.

Arithmetic average.

On the supply side, various authors have singled out a “credit crunch” as a cause of high real rates. Edwards (1985) expresses the view that the absence of sufficient real liquidity stands out as a significant factor. Meller and Solimano (1983) conclude that “monetary policy plays an important role in the determination of the real interest rate.”

Care must be exercised, however, in ascertaining the source of this alleged credit crunch. Domestic credit grew 47.2, 30.4, and 97.1 percent in 1980, 1981, and 1982, respectively. (The corresponding inflation rates were 31.2, 9.5, and 20.7 percent.) Hence, booming demand, not tight supply, seems to have been at the root of this excess demand for credit, as Harberger hypothesized.

Spread Between Loan and Deposit Rates

Table 16 also suggests that there was a substantial spread between lending and deposit rates over the period, which accounts for part of the high absolute level of real loan rates. Although real interest paid on deposits was high (particularly in 1981 and 1982), it cannot fully explain the high rates charged for loans. Meller and Solimano (1983) maintain, as a result of their econometric work, that the spread is totally independent of the variables that determine the real interest rate. Hence, microeconomic factors must have been at work in maintaining the wedge between loan and deposit rates.

Early in the financial reform period the combination of high reserve requirements and substantial inflation was blamed for the spread. But as has been discussed in Section I, interest began to be paid on bank reserves in May 1976, and required reserve ratios were gradually brought down to very low levels. Hence, the effect of reserve requirements on bank costs was minimized over time. Another cost-based argument is that made by Sjaastad, and mentioned earlier, about the existence of arbitrage restrictions, but this argument is probably unimportant. Finally, it is commonly conjectured that Chilean banks’ unit operating costs were much higher than those for comparable institutions in more developed countries. This feature has been variously attributed to inefficiency, inexperience, setup costs, and economies of scale, among other things, but assessment of its importance must await a detailed microanalysis of the banks’ books. It is also possible that bank managers attempted to offset the high risk of their portfolios with a higher spread, or that bad loan reserve provisions forced the spread upward—such provisions constitute an “operating cost,” after all.

Frequent mention is also made in the literature (Arellano (1983b); Meller and Solimano (1983); Zahler (1985)) of the oligopolistic characteristics of the Chilean banking system, but the dynamics of the process remain unclear.30 Barriers to entry were reduced, and monopoly rents should have fallen over time. Vigorous competition seems to have existed among banks and financieras, but for some reason it focused on quantity and product differentiation and seems to have had little effect on financial prices.

A plausible though imperfect hypothesis to explain this puzzle goes like this: for a given market-determined deposit rate, the spread will depend on the cost of holding reserves, on operational costs, and on dividends paid to stockholders. Other things being equal, the higher the dividends paid, the larger is the spread. In Chile, banks found themselves with a large portion of nonperforming loans, so that their effective cash inflow was small despite the high loan rates charged. Had nonperforming loans been of manageable size, banks could have lowered dividends or management salaries or other flexible costs to cushion the blow. Given the magnitude of accumulating bad debt, however, a crisis was inevitable.31 Bankers then faced a dilemma: the higher the bank earnings and dividends are, the worse the cash flow of the firm and the sooner the crash would come. At one end, if dividends were infinity, the crisis would have occurred immediately; at the other extreme, if dividends were zero, the crisis would be postponed but the sum of dividends paid over the period also would be zero. Somewhere in between there had to be a rate of earnings/dividends—and hence the spread—that maximized the present value of bank owners’ earnings. This was presumably the stable state the Chilean bankers chose.

VI. Four Theories of the Financial Crisis: Separating Micro from Macro Causes

The macroeconomic environment began to deteriorate dramatically in 1981. According to the indices computed by Corbo (1985), the real exchange rate appreciated about 45 percent between the second quarter of 1979 (when the nominal parity was fixed) and the end of 1981.32 The trade balance worsened accordingly, and, for 1981, the current account deficit reached a staggering 14.3 percent of GDP (Table 1). Between October 1981 and June 1982, the Central Bank of Chile lost US$555 million in international reserves. The need for a real devaluation was by then beyond doubt. Policymakers considered a mandatory cut in money wages as a possible alternative to nominal devaluation, but the idea was soon dropped. The devaluation (18 percent in nominal terms) came in June, and it was accompanied by a relaxation of the indexation of minimum wages, which had put a floor to private sector wage adjustments, to ensure that a real devaluation would take place as well.33 The public deemed the devaluation insufficient and speculation continued. An additional US$890 million in reserves was lost in the rest of the year, as floats and crawls were tried to stabilize the exchange market. An almost 90 percent nominal devaluation was to occur before the exchange market was stabilized. To cushion the impact, the authorities swiftly set up preferential exchange rates for dollar debtors.

The Chilean financial crisis had begun to unfold in late 1981. By then the difficult macroeconomic conditions were taking their toll on the profit statements of firms, and business bankruptcies surged—most prominently, CRAV, a large sugar refining company, failed in May 1981. During the year, real interest rates, which had fallen in 1980, began climbing again, placing further strains on a weakened financial structure characterized by high debt/equity ratios among enterprises and a significant amount of nonperforming assets in the portfolios of banks.

The first public shock came in November 1981, when the Government decreed the “intervention” of three banks, a development bank, and four financieras that together accounted for more than one third of the loan portfolio of the financial system. Two of the banks were among the six largest private commercial banks in the country (Banco Español de Chile and Banco de Talca); the other was a small regional bank. The banks that were subjected to intervention began losing deposits in September 1981; the decline in deposits from August to December amounted to Ch$24 billion, of which almost Ch$16 billion took place in November. Throughout 1982, Chile was plagued by rumors and by evidence of further firm and bank problems, made more urgent by the rapidly worsening macroeconomic situation.34 Fears were confirmed in January 1983, when the Government placed seven banks and one financiera under temporary government management.35 These institutions held 45 percent of the outstanding loans and 41 percent of the total deposits in Chile. They included the nation’s two largest private commercial banks (Banco de Chile and Banco de Santiago), which served as standard-bearers for the two most powerful grupos. Thereafter, a wide array of programs would be put into effect to stabilize the remaining institutions and to restructure and restore to health the ones in which the Government had intervened.36

What triggered a crisis of this magnitude? The financial system had long been engaged in a rather unstable process of credit expansion, but until sometime in 1981 international reserves and the real economy seemed to be booming anyway. There were two immediate activators of the collapse: First, capital inflows declined sharply (Table 3). In the first half of 1982, net capital inflow totaled US$889.2 million—only 36 percent of the inflow in the previous semester. Of course, much of this reduction had exogenous causes—the conflict between Britain and Argentina over the Falkland/Malvinas Islands, the Mexican problems of August 1982, and the generalization of the Latin American debt crisis. Second, asset prices dropped as a result of bankruptcies and the reversal of previously optimistic expectations. This drop had an unusually large effect on the financial health of firms, because a good portion of profits until then had consisted of (unrealized) capital gains.37 After the fall in asset prices, many firms that had apparently suffered only from short-term liquidity problems were revealed to face long-term insolvency.

The hypotheses about the deeper or structural causes of the financial sector problem can be classified into four groups:

  • External shock explanations: The negative impact from abroad (higher interest rates, deterioration of the terms of trade, credit rationing) was so large that no domestic financial system could have withstood it.

  • Macro-shock explanations: The macroeconomic environment was sufficiently hostile, and variables (like the real interest rate and exchange rate) sufficiently out of line, to ensure the collapse of the financial system.

  • Inherent market instability explanations: Unregulated markets can be destabilized by random events. This is especially true of financial markets, where elements such as expectations and trust play a key role. What happened in Chile was nothing but the unfortunate disequilibration of an otherwise sound market.

  • Inadequate financial reform explanations: Reforms in Chile were carried out with little or no regard for possible market imperfections. Competition and the free flow of information were not sufficiently encouraged. The outcome was a predictable combination of several market failures (oligopoly pricing, moral hazard, adverse selection, etc.).

Of course, these four views are but caricatures that require appropriate qualifications. They serve, however, as focal points to structure our discussion.

From the point of view of the interaction of a troubled financial system with the economy at large, the first two explanations above are equivalent: both maintain that the problems came from outside the financial system. Two counterexamples are available to evaluate such an argument. First, consider what happened to the Chilean financial system the last time it suffered an external shock of this magnitude—in the 1930s. Despite all the caveats that such a comparison requires, it can be noted that during the Great Depression no Chilean bank went bankrupt; nor did any bank from the “major countries” in Latin America: Argentina, Brazil, and Mexico (Díaz-Alejandro (1983)). A second counterexample is the recent fate of Latin American countries that did not liberalize their financial systems but suffered great external stress as Chile did. Brazil, Peru, Colombia, and Venezuela experienced many macro and micro problems over the past decade, but not financial problems of the magnitude experienced by Chile. Of course, maintaining financial repression probably entails substantial costs, and so does keeping banks alive through more or less open subsidization (as some Latin American governments seem to have done in the 1930s). Such counterexamples, however, do suggest that governments have tools at their disposal that can, to some extent at least, reduce the vulnerability of a financial system to the vagaries of the macro environment. It seems undeniable that the Chilean Government did not use such tools in time during the crisis of the early 1980s.

It is also important to note that not all banks in Chile experienced the same difficulties in the recent period. The conservatively managed Banco del Estado (the nation’s second largest) emerged from the crisis practically unscathed. So did the various foreign banks operating in Chile. Of course, all banks, private and state-owned, domestic and foreign, were subjected to the same macroeconomic storms. The key to their different fates seems to be in the differences in micro management, to which we now turn.

In contrast to the first two hypotheses, the last two hypotheses focus on the financial system itself in searching for the causes of the crisis. The third hypothesis stresses that such a crisis could have happened to any market system, given the inherent instability of financial markets and the severity of exogenous shocks during the period; the last hypothesis is that a crisis was bound to happen to this particular market system because of the nature of the institutions with which it had been endowed at the time of the financial reform. Obviously, severe shocks can disequilibrate even the soundest of markets. But whereas a sound market tends to cushion the effects of perturbations, an unsound one tends to magnify them. Unfortunately, the Chilean market seems to belong in the latter category. Regardless of the role played by other factors, it is difficult to escape the conclusion that the Chilean financial problem had perverse micro dynamics of its own, which in turn was made possible only by the peculiar pattern of ownership and regulation with which the system emerged from the process of reform. Ex post at least, the market failures visible in the Chilean financial markets belong to the textbook variety (moral hazard, adverse selection, oligopolistic pricing), which were fostered by certain government policies (implicit guarantees, lack of portfolio supervision, allowance of interlocking ownership, and lending patterns).

Furthermore, a strong case can be made that financial sector imbalances contributed to create the macro disequilibrium, and not just the other way around. Theorists of stabilization increasingly recognize the crucial role that credibility plays in disinflation-cum-reform programs,38 Announced exchange rate policy clearly enjoyed substantial credibility until sometime in 1981, and such confidence was buttressed by healthy international reserves. But the stabilization program coexisted with a mounting private financial deficit that was clearly unsustainable in the long run and that, under the (apparently justified) assumption of implicit guarantee, would eventually become the responsibility of the Government. What measures (domestic credit creation, inflation, etc.) the Government would choose to deal with this problem could not be anticipated, but they were unlikely to be compatible with the maintenance of a fixed exchange rate, a cautious monetary policy, and careful management of expectations as the backbone of the anti-inflation policy.39

The point can be put slightly differently. It is often argued, as in the Argentine case, that the combination of a slow crawl of the exchange rate with a large fiscal deficit would make the policy package inconsistent and hence unworkable. What, then, was the inconsistency in the Chilean program, in which fiscal deficits had been contained? What variable played the role that the budget deficit played in Argentina? One possibility is the system of backward full wage indexation (Edwards (1985) and Corbo (1985)); combining it with a fixed exchange rate was tantamount, it is argued, to imposing two numeraires on the economy. Eventually, one had to give. The financial deficit can be suggested as an additional source of “incredibility” in the Chilean case. Díaz-Alejandro (1985) has hinted at this possibility:

The massive use of Central Bank credit to “bail out” private agents raises doubts about the validity of pre-1982 analyses of the fiscal position and debt of the Chilean public sector. . . . Ex post, it turned out that the public sector, including the Central Bank, had been accumulating an explosive amount of contingent liabilities to both foreign and domestic agents, who held deposits in, or made loans to, the rickety financial system.

In other words, it is conceivable that the financial sector deficit played the role usually reserved for the fiscal deficit. As is the case with any deficit, it can be financed with bonds or with money creation. The Chilean deficit was financed at first with “bonds”: banks borrowed abroad and central bank reserves swelled but domestic credit creation was cautious. Eventually, as the limit on foreign borrowing was reached and domestic bankruptcies increased, many banks were revealed to be significantly overextended. The Central Bank had to intervene or heavily subsidize them; in addition, it had to assume responsibility for servicing the foreign debt contracted by banks. The prudent fiscal and monetary stance had to be abandoned (at least temporarily), and the end of the exchange rate regime could not be far away.40 In the last quarter of 1981, the Central Bank extended almost US$ 1 billion of credit to the financial sector and lost about US$230 million in net foreign reserves; in the first six months of 1982 the Central Bank extended a further US$535 million of credit to the financial system, and lost an additional US$325 million of foreign reserves. There can be little doubt that the financial crisis contributed significantly to the loss of reserves and eventually to the collapse of the stabilization plan. First, the intervention of financial institutions led to a rapid expansion in net domestic credit of the Central Bank to support those institutions. Second, those interventions undermined confidence in the financial system, which contributed to the decline in the demand for domestic financial assets observed in 1982 and 1983. These two factors played a major role in generating expectations of higher inflation and of a depreciation of the peso, as international reserves were drawn down.

This is of course a highly stylized account, but it captures one crucial feature of the Chilean problem. The stabilization plan relied heavily on the moderation of inflationary expectations. This aim was initially achieved, but public confidence could not be sustained as long as a good share of the nation’s banks and firms were perceived to be insolvent. After a period of confidence the public eventually adopted precautionary measures (such as beginning to accumulate foreign exchange) to protect itself from future policy changes. As a result, the Government’s stabilization policy became even more difficult and costly to apply.

VII. Reacting to the Crisis

The Government’s reaction to the macroeconomic troubles that became evident in 1981 appeared at first to be guided by the expectation that an “automatic adjustment” would occur in response to such disequilibrium. In particular, high real interest rates would reduce the pressure of domestic demand. The problem of an overvalued exchange rate also would correct itself: in the event of a balance of payments deficit, reserves would flow out and the resulting lower money base would push down the price of nontradables. But because such adjustments appeared to take too long, the Government decided to take a more activist policy stance beginning in May 1982. The Government aimed at easing the adjustment in relative prices by devaluing the peso. Moreover, it began to use monetary policy actively to moderate increases in interest rates, especially as external financing increasingly dried out. These efforts to reduce domestic rates were later helped by the decline in international interest rates; in this regard, beginning in 1984, the Central Bank strove to keep domestic rates competitive with international rates. In 1985, repayment of domestic debt of the nonfinancial public sector and increased reliance on foreign borrowing eased the need for public sector domestic borrowing. The Central Bank provided large credit flows at predetermined rates to financial institutions and “suggested” deposit rates to banks.

At the micro level, the Government’s reasoning before the devaluation was similar: If any firms were near bankruptcy, they would voluntarily declare themselves so. The judicial system would carry out the necessary settlements and liquidations. If asset prices fell as a result, that would only reflect equilibrium conditions in the corresponding markets. The main exception to this hands-off position was a tightening of bank supervision by the Superintendency of Banks. As mentioned, in 1982, the Government attempted to encourage grupos to reduce their debt to the public and particularly to their own banks. The problem of the cartera relacionada was on its way to becoming a public issue.

Barandiaran (1983) conjectures that in mid-1981 firms’ problems were still of a manageable magnitude so that a wave of liquidations would have been feasible both economically (asset prices would not have plummeted) and logistically (the judicial system could have handled the resulting workload). By mid-1982, a drastic solution of this sort was no longer feasible. A macro relief package was necessary, and it had to come from the Government.

The first round of intervention of financial institutions involving eight of them, came in November 1981. At this time, substantial amounts of credit began to flow from the Central Bank to the banking system. The June 1982 devaluation put further strains on the ability of banks and firms to service dollar debts, and more vigorous palliative policies began to be adopted. Further rescue measures would be put into effect as part of the early 1983 bank interventions and until early 1985.

Measures to Aid Domestic Borrowers

Reprogramming of Loans to Firms

In 1983 and 1984, the Central Bank established schemes to enable banks to reschedule a portion of their loans to firms. Lenders granted firms lower interest rates and longer maturities. Because these measures would worsen banks’ liquidity and reduce their profits, the Central Bank introduced a scheme of subsidization that is described in the next subsection.

Under the 1983 program, rescheduled debts of firms in pesos carried a real interest rate of 7 percent. The terms for dollar loans were identical except that the 7 percent interest applied to each payment in U.S. dollars (equivalent) on the basis of the current exchange rate. All borrowers were to repay the rescheduled credits in ten years, with a grace period of five years for principal and one year for interest. The second debt rescheduling, implemented in 1984, expanded the amounts eligible for rescheduling, lowered interest rates (5 percent real for the first two years, rising in steps to 7 percent from the sixth year), lengthened the maturities to up to 15 years, and granted more favorable treatment for smaller debtors.

From the point of view of debtors, these were not easy terms. In the first round (1983) 51 percent of the financial system’s loans were reprogrammed at subsidized interest rates (7 percent in real terms). The remainder was left at market rates (15 percent, real). The resulting average effective real rate—more than 11 percent—was still quite onerous, especially given the recessive economy.

This program was substantial: by the end of 1984 rescheduled loans accounted for 21.4 percent of domestic credit.41 There were some restrictions on eligibility: foreign trade credit and loans to legally bankrupt firms, to other financial institutions, and to holding companies could not be reprogrammed. Nevertheless, this “blanket” program made no effort to separate viable borrowers from those that were not. Whereas, under the 1983 program, the Central Bank sought to reschedule a mandatory minimum 30 percent of the debts of all eligible debtors, the 1984 program varied the rescheduled percentage according to the size of loans. Since late 1985 the possibility of a “case by case” rescheduling exercise had been discussed, but only a program for small debtors was implemented—in 1986—and without any subsidy or credit from the Central Bank. This program included supervisory incentives and penalties aimed at encouraging banks to regularize the situation of small borrowers.

Reprogramming of Housing Loans (Mortgages)

A similar program was established for mortgage debts. Amounts to be rescheduled included installments unpaid since 1981 and a decreasing percentage of installments payable between 1983 and 1987. The rescheduled amounts became loans denominated in UFs (for Unidad de Fomento, a unit of account indexed to the CPI) and carrying 8 percent annual interest. The Central Bank purchased part of the rescheduled loans from the financial institutions with a note in UFs also at 8 percent, and refinanced the remainder of the rescheduled payments with a central bank line of credit carrying 7 percent real annual interest.

Under this scheme, more than 36,000 loans were reprogrammed. Rescheduled amounts were not very large—about US$120 million outstanding as of August 1985.

Sectoral Lines of Credit

As part of efforts to cushion the effects of the crisis on some sectors and to revive other sectors, the Central Bank established some special lines of credit. Some of these have been targeted to working capital needs, payroll financing, labor hiring incentives, construction and public works, and reforestation. They all provide credit to firms at below-market rates. They also carry an implicit subsidy for the financial intermediary involved, because the cost of central bank credit for these lines of credit is lower than banks are allowed to charge the user.

Measures to Aid Financial Intermediaries

Emergency Loans to Banks

At the start of the crisis, emergency loans were granted to banks in which the Central Bank had intervened and others. When more comprehensive aid packages were put in place, loans were repaid and converted to equity.

Subsidies to Facilitate Reprogramming of Loans

As noted earlier, the loan reprogramming was facilitated by the provision of subsidies to the banks and financieras involved: The Central Bank extended to financial institutions lines of credit to be repaid within ten years at a real interest rate of 5 percent, with a grace period of five years for principal and one year for interest; in turn, financial institutions were to use the funds to buy six-year central bank notes on which payments were made quarterly at an annual rate of 12 percent.42,43 In short (and abstracting from the different maturities), banks exchanged their own paper yielding 5 percent for central bank paper yielding 12 percent, thereby receiving a 7 percent spread (with no risk) as subsidy.

Purchase of Risky Loans by Central Bank

Beginning in 1982, the Central Bank decided to facilitate the recapitalization of the banking system. The first scheme, announced in July 1982, involved a central bank purchase of the substandard loan portfolio of banks up to 100 percent of the capital and reserves of each bank. The Central Bank would pay for this purchase by issuing non-interest-bearing central bank bonds. Banks would have to repurchase 5 percent of these loans every six months for ten years. Both the loans and the bonds would be adjusted for inflation. This scheme improved the balance sheet position of the banks involved by temporarily removing substandard assets from their portfolios and by effectively giving them the possibility of writing off bad loans over a ten-year period.

A more comprehensive recapitalization scheme was introduced in February 1984.44 In essence, the new scheme provided for the Central Bank to purchase with cash substandard loans at par for up to 150 percent of capital and reserves of each bank. To reduce the monetary effect of this cash purchase, the selling bank had to use the proceeds to repay any outstanding emergency loans from the Central Bank, and had to use any remaining balances to purchase central bank pagarés (IOUs). Once again, the implicit subsidy was substantial: the central bank pagares carried a real rate of return of 7 percent (UF plus 7 percent) and a four-year maturity. Banks could also exchange an additional amount of substandard loans (up to 100 percent of their capital and reserves) for a non-interest-bearing, nontransferable central bank note. Loans exchanged for central bank notes had to be repurchased over ten years at their initial real value. Loans sold for cash had to be repurchased at their initial value plus a 5 percent real interest rate over ten years. Bank shareholders were required, at the time of sale, to devote all dividends on their shareholdings to this repurchase until the repurchase was completed. New capital contributions also faced a partial limitation on dividends—only up to 30 percent of earnings could be distributed as dividends—until the loan repurchase was completed. By the end of August 1985, the amount of bad loans sold by the commercial banks had reached US$2.36 billion (domestic banks accounted for 96 percent; and “intervened” banks for 75 percent). Although initially only banks that had not been subjected to intervention were allowed to sell substandard loans to the Central Bank, from 1985 the intervened banks also were allowed to participate. As a result, the purchases of substandard loans continued through 1987.45

In addition, the Government decided to recapitalize intervened banks directly. A law passed in February 1985 allowed the Superintendency of Banks to require intervened institutions to increase their equity to make them financially viable.46 The new stock would be offered first to existing stockholders and then to third parties. Any portion of the required capital stock that was not subscribed would be purchased by CORFO, Chile’s state development agency. CORFO would pay for these shares by assuming the emergency credits that intervened banks had earlier received from the Central Bank. In other words, the Central Bank emergency credits were converted to equity. CORFO would have to sell its bank shares within five years, at a rate of 20 percent per year as a minimum. Any unsold stock would be automatically transferred free of cost to the shareholders that had already contributed to the equity increase. Moreover, the law established that at no point could CORFO shares of a bank exceed 49 percent of the bank’s capital. Any losses under this program would accrue to the Central Bank, although the law provides for reimbursement from the Treasury up to a certain amount.

To attract interest in the shares of effectively bankrupt institutions, CORFO made the terms very attractive. Only a small down payment was necessary; the balance was payable in 10 years at 5 percent real interest.47 With some limits, individuals with no tax arrears enjoyed even more favorable conditions: 15 years and no real interest. Buyers of such stock also enjoyed substantial fiscal advantages: not only tax-exempt dividends but also a tax credit equivalent to 20 percent of the value of shares purchased.

Through these mechanisms, the five intervened banks were returned to the private sector during 1986. Four were fully capitalized and sold, including the largest two in the nation, Banco de Santiago and Banco de Chile. The shares of these two banks were sold to more than 57,000 shareholders by December 1986, and their management was subsequently fully privatized. The fifth bank, Colocadora Nacional de Valores, was merged with the Banco de Santiago.

Other Relief Measures

Preferential Exchange Rate

In mid-1982, the Central Bank established a preferential exchange rate for debt-service payments on certain dollar-denominated loans.48 Since December 1982, the resulting subsidy has been paid with negotiable central bank notes that have a maturity of six years.49 In 1984, the Central Bank introduced some restrictions on access to this program. In 1985, further restrictions were enacted, and it was decreed that the program would be gradually phased out.50 As a result, although the devaluation of the peso in February and June of 1985 increased the total cost of the program for that year, the cost subsequently dropped.51

Interest Subsidies on Swap Operations

Since 1983, swap operations have provided a hedge to domestic agents possessing dollar liabilities. Outstanding swaps of dollars derived from the repayment of certain foreign currency loans by domestic residents to local banks were converted into dollar-denominated accounts at the Central Bank, carrying a yield of six-month LIBOR plus a spread.52 Banks can only draw funds from these accounts in order to purchase foreign exchange to service their external debt. Against the remaining balance the Central Bank issues peso-denominated lines of credit to banks at a predetermined real interest rate. In the course of 1985, measures were taken to moderate the effect of the subsidy. The premium over LIBOR that the Central Bank paid on these deposits was gradually reduced, and eliminated by February 1987, and the Government made efforts to decrease the immediate monetary effect of the program. The Central Bank also established a schedule for the elimination of the subsidy on swaps; for new operations, this subsidy was eliminated by the end of April 1987.

The most important result of all these relief programs is that they kept the banking system functioning, however tenuously. The importance of this development for an economy trying to climb out of a deep recession cannot be overstated. As the overall economy recovered somewhat (real GDP grew by 6.3 percent in 1984 and 2.4 percent in 1985), the business bankruptcy problem was less dramatic than it was in 1982, and banks’ fortunes rose accordingly. In 1985, the banking system as a whole showed after-tax profits amounting to 6.2 percent of capital, the first such profits since 1981. In 1986, the system registered profits of about 4 percent of capital.53

The bank rescue operation, together with the decline in international interest rates, may also have facilitated a moderation of domestic interest rates (Table 17). This overall encouraging trend obscures some short-term cyclical fluctuations. In the course of 1985, for instance, as banks tried to shield themselves from the losses from certain problem loans, real loan rates rose slightly (even as deposit rates edged down). But real rates declined again in 1986, presumably aided by the fall in international rates.

The counterpart to the improved asset position of many banks (fewer problem loans, less cartera relacionada) has been a dramatic increase in central bank financing of commercial banks. For instance, this financing became more important than deposits as a source of funding (Table 18). Moreover, as of June 1985, outstanding borrowing in pesos from the Central Bank amounted to four times the capital and reserves of the banking system and was equivalent to almost 50 percent of the sum of loans and foreign assets of the system. A similar point can be made for the income figures, dependent as they are on central bank support. As a result of the support extended to the rest of the financial system and of the preferential exchange rate subsidy, central bank credit (in terms of liabilities to the private sector) rose by 94 percent in 1982 and by a further 433 percent in 1983 (Table 19). The bulk of central bank credit in 1983 consisted of credit to financial intermediaries and transfers to banks and the corporate sector as part of the preferential exchange rate and other programs involving subsidies. During 1984, the credit expansion of the Central Bank slowed down to 174 percent, even though the Central Bank continued to be the principal channel of foreign loans to Chile.

During 1985, central bank credit increased by about 200 percent, largely reflecting the losses arising from the preferential exchange rate and swap subsidies discussed earlier, which are registered as “other assets” in Table 19. Credit expansion was financed to a much larger extent from domestic sources in 1985 than in 1984: liabilities to the private sector increased by about 90 percent, especially reflecting the issuance of medium-term notes to finance the preferential exchange rate subsidy. The issuance of central bank paper limited the immediate liquidity effect of the subsidies, which nevertheless has had to be offset as this paper matured and its interest was paid.54

Table 17.

Interest Rates on 30–89–Day Operations of Commercial Banks, 1981–86

(In percent per month)1

article image
Source: Central Bank of Chile, Síntesis Monetaria y Financiera.

Weighted average of the rates on all operations during the month.

The nominal rate less variation in the CPI during the month.

Table 18.

Central Bank Financing of Commercial Banks, 1981 and 1985

(In percent)

article image
Source: International Monetary Fund, International Financial Statistics.

Similarly, despite the progress on the interest rate front, some problems proved to be more difficult. Despite the fall in deposit and lending rates, spreads continued to be substantial, as banks attempted to offset losses from bad loans not sold to the Central Bank (Table 17). The reduction in interest rates owed a great deal as well to government aid. Without these transfers from the Central Bank, interest rates charged to new borrowers would probably have been much higher.

A reasonable case can be made that the measures taken since 1982 served to postpone and dilute over time the effects of the crisis. Progress was made in many areas, but key issues remained to be settled.

One such issue was the determination of conditions that would enable individuals and enterprises to service their rescheduled debts without incurring further unpayable obligations. As mentioned earlier, the conditions faced by borrowers (unlike those for lenders) were quite stringent from the start, creating great pressures for further reprograminings and concessions. According to Arellano (1984).

The granting of very strict reprogramming conditions—which are not sustained later on—has been typical since 1982. . . . In the end, experience has demonstrated that unrealistic conditions such as the ones that prevailed both before and after the reprogrammings cannot be applied to debtors, and in the long run they end up raising costs and further eroding credibility.

Even after loans had been reprogrammed, the bad debt problem persisted: in 1985, overdue loans as a percentage of all loans started to rise again. Improved economic conditions beginning in 1986 have gradually diminished the importance of this problem.55

Table 19.

Central Bank Operations, 1982–85

(Percentage change with respect to liabilities to private sector at beginning of period)

article image
Source: Central Bank of Chile.

Excludes holdings of treasury notes on account of the 1983-85 capitalization of the Central Bank. These notes are included in other assets.

A negative sign implies an increase in capital and reserves.

Includes foreign liabilities on account of deposits placed by the corporate sector in the Central Bank in the context of the 1983-85 rescheduling agreements with foreign commercial banks.

Includes medium-term notes issued by the Central Bank to finance the preferential exchange rate subsidy.

Table 20.

Holdings of Financial Assets by Private Sector, 1982–85

(Rate of growth, in percent)

article image
Source: Central Bank of Chile.

Foreign currency deposits are valued at the end-of-period exchange rate.

Includes time and savings deposits, mortgage bonds, and foreign currency deposits.

Nominal changes deflated by changes in CPI.

Reform of Bank Regulations and Legislation

As part of the response to the financial crisis, the prudential regulations were strengthened (as discussed in Section IV) and a new legislative framework was developed to govern the restructured financial system. But, in the immediate aftermath of the crisis, it was necessary to allow sufficient flexibility in the application of various legal and accounting regulations to permit time for adjustment. In the period following the 1981 and 1983 interventions, this flexibility encompassed the time allowed for an overdue loan to be transferred to the nonperforming portfolio, the term for constituting individual provisions, the rules governing disposal of physical assets acquired by banks from debtors, the accruals of interest, and the definition of capital and reserves for the purpose of capital adequacy regulations. Only in late 1984, new regulations were issued to return to more orthodox accounting practices. Also, since January 1983, the general deposit guarantee had been extended from time to time, and in June 1986, deposit insurance was granted only to domestic Chilean banks that met specified minimum capital requirements and only upon application by each bank.56

In November 1986, a new banking law was enacted that modified several key aspects of existing legislation in order to give formal status to existing supervisory practices, to further strengthen prudential regulations, and to streamline the deposit guarantee scheme. The main aim of the reform was to minimize the need for state intervention in the financial system by facilitating market self-regulation.57 The law also tried to eliminate some of the practices that had led to problems in the past, such as lending to related groups.

To increase information to the public, the new law directed the Superintendency to publish detailed information on the type and quality of the assets of each of the financial institutions it supervises, at least three times a year. The law also sets different levels of confidentiality. Strict confidentiality applies only to deposits and other sources of funds to the financial system. Banks are allowed to release other information to firms specializing in the analysis of financial institutions and, more generally, to anyone with a legitimate interest—provided such disclosure cannot be seen to harm a bank customer.

The law also strengthened bank supervision by converting previous superintendency regulations into law and making some superintendency powers more explicit. It also defined more precisely the concept of bank client, in order to take into account interrelated ownership.

The foregoing measures were intended to give depositors better information about the soundness of financial institutions. As a consequence, the authorities thought that state deposit insurance should be reserved only for small savers, for whom 90 percent coverage of their time deposits with a maximum of 120 UFs remained. The elimination of the state insurance on other deposits has been taking place gradually, and was to have been completed in 1989. Sight deposits, however, will still be fully protected. Banks have to maintain in liquid central bank or treasury assets the equivalent of sight deposit balances exceeding two and a half times the bank’s capital. Moreover, should a bank be unable to meet its sight deposit obligations by selling off those liquid assets, the Central Bank would advance the necessary funds. In other words, sight deposits remain fully protected, but banks’ freedom to invest such deposits has been circumscribed. A bank with insolvency problems can, subject to approval by the Superintendency of Banks and Financial Institutions, enter into agreements with its creditors proposing ways to meet its obligations. All bank creditors are eligible for these agreements, except holders of sight deposits and time deposits—to the extent that the latter are covered by the state deposit insurance. These agreements may include partial or total capitalization of eligible bank liabilities, extension of the maturity of the liability, and partial write-off. Lack of agreement of an insolvent bank with its creditors is one of the grounds for the Superintendency to liquidate the bank.

The new law has tightened capital requirements. A bank has to increase its capital not only when it no longer meets the minimum capital requirement, as previous law mandated, but also when it does not comply with the required ratio between its capital and reserves, and deposits and other liabilities. The law establishes the period within which recapitalization must take place; in case of noncompliance, the bank is forbidden from increasing its loans or any other investments, except central bank instruments. The law also facilitates the rescue of one troubled bank by another bank. The rescuing bank can lend up to the equivalent of 25 percent of its capital to the troubled bank for two years. The loan can only be repaid if the latter bank has been sufficiently recapitalized, without including the loan for this computation. If the loan is not repaid within two years, it can be used as a capital contribution, in case of merger of the two institutions, or to capitalize the borrowing bank, but with the shares remaining in the hands of the borrowing bank.

Although it is too soon to assess the effects of the new legal framework on the functioning of the banking system, the foregoing reforms constitute an interesting attempt to balance the desire to minimize the state’s interference with private banks against the state’s perceived responsibilities in monetary policy. The law appears to assign a much greater degree of responsibility to the state to ensure the smooth functioning of the payments system than to protect the private sector’s time deposits and other liabilities.58

VIII. Avoiding and Managing Financial Crises: Some Lessons from the Chilean Experience

This section presents nine lessons about managing (and avoiding) financial crisis from the Chilean experience:

  • Financial liberalization is likely to be characterized by rapid growth in quasi-money and the creation of new forms of financial intermediation. But conventional indicators of financial “deepening” or “widening” may be misleading in two senses.

First, ratios like M2/GDP may rise precipitously as a result of substitution away from other assets or sudden inflows of foreign savings. The growth of short-term financial assets does not imply (or require) a better national savings performance; nor does it necessarily entail an efficiency gain, as, for instance, in the channeling of resources away from stores of value with low marginal product toward more productive assets.

Second, financial “deepening” and “widening” are not necessarily tantamount to the emergence of well-developed financial markets, if “well-developed financial markets” means a situation in which key markets exist and relative prices are reasonably close to any conceivable range of social optima. Markets for long-term transactions, in particular, develop only slowly, even in the context of cautious fiscal and monetary policies and declining inflation. Real interest rates can remain extremely high and erratic, far from the course of variables like the marginal product of real capital. For these and other reasons, investment can stagnate even as financial intermediation booms.

  • Conceptually, two internally consistent ways of organizing a financial system can be distinguished: (1) laissez-faire, without supervision but also without government insurance and a binding and credible pre-commitment ruling out future bailouts; and (2) government-provided deposit insurance, which is then sold to banks at actuarially fair prices (to reduce the adverse selection and moral hazard problems typical of insurance schemes, the Government exercises stringent supervision of bank activities).

The choice between the two systems (or combinations of them) should ultimately be guided by empirical considerations, but one warning should be kept in mind: Laissez-faire will work in financial markets only if information flows are efficient and inexpensive, so that customers can easily assess the soundness of a bank’s operations; and if the behavior of depositors and creditors is rational, so that problems in one bank will not lead to runs at other otherwise sound institutions. If these two rather demanding conditions are not met, then government-provided deposit insurance ought to be the choice of the cautious policymaker.

  • The Chilean experience suggests that effective bank supervision should not be limited to the enforcement of certain key required minimum ratios. Interlocking ownership patterns cause such huge problems that they breathe new life into the old-fashioned principle that cautions against banks’ ownership of firms, or vice versa. It is highly doubtful that Chilean banks would have extended so many unsound loans if many of them had not gone to banks’ own related enterprises. Without this sizable cartera relacionada, a financial crisis of Chile’s proportions could never have occurred.

  • When financial liberalization is carried out in a high-inflation setting, a special problem arises. Highly negative (controlled) real interest rates usually are a serious problem under financial repression, and reformers clamor for their elimination. The Chilean experience from 1975 to 1985 suggests that sustained highly positive real rates can constitute at least as serious a problem. When real rates stubbornly remain above any conceivable rate of return on investment, problems among businesses will become widespread. A business’s borrowing decision is then distorted: rather than borrow to invest or to finance working capital, many businesses will borrow to pay interest or simply to stave off bankruptcy. Under institutional arrangements such as Chile’s, the problem can quickly snowball and will inevitably lead to a systemwide crisis.

It is also important to notice that even in an economy as open as Chile’s, policy can have an important effect on domestic interest rates, both nominal and real. A central conclusion of this paper is that domestic supply and demand factors were the most important causes of high interest rates. This suggests, among other things, that it may have been a mistake to assume, without further verification, that only fully open economy models were to be used in designing and evaluating policies in Chile.

  • An “artificial demand” for credit, arising from continuous bank lending to troubled enterprises, seems to have been a major force behind high domestic interest rates. The rolling over of scheduled principal payments and the capitalization of interest tended to increase the demand for credit. In turn, banks tended to raise deposit rates in order to attract domestic resources to finance these practices. Loan rates went up correspondingly, bringing further difficulties to the immense majority of firms whose peso borrowing was short term.

It is noteworthy that the upward pressure on interest rates remained even at times when monetary policy was somewhat expansionary. This suggests that, when setting monetary targets, architects of stabilization programs should closely monitor institutional developments in credit markets and the composition of credit recipients. What may seem like a generous rate of credit expansion to the private sector may turn out to be unduly restrictive if the lion’s share is taken up by the needs of troubled firms.59

  • Although macroeconomic shocks can aggravate economywide financial problems, the ways in which financial crises can aggravate macroeconomic disequilibrium are less well understood. As already mentioned, business problems, credit demand, and interest rates will probably be connected. In the presence of high real interest rates, private investment will predictably suffer. But most important, the accumulation of “excess” domestic debt can introduce serious distortions into the formulation and effects of macroeconomic policies. If the perception exists that depositors, borrowers, or financial intermediaries are under the Government’s protective umbrella, agents in the economy can come to expect that future bailouts will require the abandonment of fiscal and monetary prudence. Even if explicit or implicit insurance does not exist, a wave of bankruptcies may induce policymakers to adopt expansionary policies. Alternatively, the authorities may choose to create inflation in order to wipe off part of the debt and shift some of the adjustment burden from debtors to creditors. In short, the presence of a serious financial imbalance calls into question the credibility of any stabilization program that may be put into practice. Moreover, this loss of credibility can become a self-fulfilling expectation.

  • Voluntarily declared bankruptcies (of firms or banks) will not act as a brake to the unstable financial dynamics that this paper has described. Managers face great incentives to postpone bankruptcy almost at all costs, realizing that the bigger the problem, the more likely it is that some sort of relief will be forthcoming. In the early stages of the problem, the authorities would be well advised to use all measures within their power to counter this logic. Practices like distress borrowing should be monitored and discouraged. The price of financial stability may well be eternal vigilance. This maxim holds even for governments with a preference for laissez-faire: when the bad debt problem becomes sufficiently large, prior commitments against bailouts become unenforceable and hence less credible.

  • If the problem becomes so widespread that government intervention is inevitable, a central aim of policy should be to uproot the problem as decisively as possible. Decisive action begins with taking stock of the situation at once and avoiding the temptation publicly to minimize the importance of the problem—or worse, implying that with a bit of patience it will go away. It is important to make a realistic estimate of the present value of the transfers necessary to restore troubled banks and firms to health. If subsidies are to be granted, debts rescheduled, and so on, the terms should be sufficiently generous—given a cautious assessment of the market prospects of the enterprises involved—to make them definitive. Otherwise, firms will soon come to realize that another round of government assistance is necessary. But uncertain of the terms, managers are likely to postpone important investment decisions. The propagation of a wait-and-see attitude among business groups unnecessarily delays macroeconomic recovery.

  • In designing aid packages, policymakers face a trade-off between the across-the-board and the case-by-case approaches. The former fails to distinguish between firms suffering different degrees of difficulties and may turn out to be unnecessarily expensive; the latter system not only is cumbersome and administratively costlier, but also can run into sticky political complications. In most cases, some combination of these two approaches is likely to be optimal. Whatever the combination, at some point, the decision has to be made to stop subsidizing nonviable enterprises, to the extent that they can be identified. Moreover, any case-by-case reschedulings should be unanticipated (not announced in advance) and definitive: unanticipated, so that moral hazard problems will be avoided (managers who expect that troubled firms will get better deals in the future may have an incentive to slacken or to manipulate the company books); definitive, so that the expectation of additional case-by-case reschedulings will not create moral hazard problems in the future.60


  • Arellano, José Pablo (1983a), “El Financiamiento del Desarrollo,” in Reconstrucción Económica para la Democracia (Santiago: Editorial Aconcagua, 1983).

    • Search Google Scholar
    • Export Citation
  • Arellano, Jos;é Pablo(1983b),” De la Liberalización a la Intervención: El Mercado de Capitales en Chile 1974–1983,” Colección Estudios CIEPLAN,No. 11 (December 1983).

    • Search Google Scholar
    • Export Citation
  • Arellano, Jos;é Pablo, “La Difícil Salida al Problema del Endeudamiento Interno,” Colección Estudios CIEPIAN,No. 13 (June 1984).

  • Arellano, Jos;é Pablo, and R. Ffrench-Davis, “Apertura Financiera Externa: La Experiencia Chilena en 1973–80” (unpublished; CIEPLN, 1981).

    • Search Google Scholar
    • Export Citation
  • Arriagada, Pedro, “Adjustments by Agricultural Exporters in Chile during 1974–82,” in Scrambling for Survival,ed. byV. Corbo and J. de Melo, Staff Working Paper No. 764 (Washington: World Bank, 1985).

    • Search Google Scholar
    • Export Citation
  • Barandiarén, Edgardo, “La Crisis Financiera Chilena,” Documento de Trabajo, Centro de Estudios Públicos, No. 6 (October 1983).

  • Behrens, R., “Los Bancos e Instituciones Financieras en la Historia Economícs de Chile” (unpublished thesis,Catholic University of Chile, 1985).

    • Search Google Scholar
    • Export Citation
  • Blanchard, O., and J. Watson, “Bubbles, Rational Expectations and Financial Markets,” in Crises in the Economic and Financial Structure,ed. by Paul Wachtel (Lexington, Massachusetts: Lexington Books, 1982).

    • Search Google Scholar
    • Export Citation
  • Calvo, Guillermo, “Incredible Reforms,” paper prepared for the Conference on Debt, Stabilization, and Development, Helsinki, August 1986.

    • Search Google Scholar
    • Export Citation
  • Central Bank of Chile, Boletín Mensual.

  • Calvo, Guillermo, Síntesis Monetaria y Financiera.

  • Calvo, Guillermo, Dirección de Política Financiera, Cuentas Nacionales de Chile.

  • Calvo, Guillermo, Series Monetarias.

  • Corbo, Vittorio, “Necesidades Financieras de las Empresas: La Función de las Instituciones Financieras,” Estudios Monetarios V,Banco Central de Chile (1976).

    • Search Google Scholar
    • Export Citation
  • Corbo, Vittorio, “Reforms and Macroeconomic Adjustment in Chile During 1974–84.” World Development,Vol. 13, No. 8(August 1985).

  • Corbo, Vittorio, and Ricardo Matte, “Capital Flows and the Role of Monetary Ploicy: The Case of Chile,” Documento de Trabajo, Economics Institute, Catholic University of Chile, No. 92 (1984).

    • Search Google Scholar
    • Export Citation
  • Corbo, Vittorio, and José Miguel Sánchez, “Adjustments by Industrial Firms in Chile during 1974–82,” in Scrambling for Survival,ed. byV. Corbo and J. de Melo, Staff Working Paper, No. 764 (Washington: World Bank, 1985).

    • Search Google Scholar
    • Export Citation
  • Cortázar, Rene, “Wages in the Short-Run: Chile, 1964–1981,” Corporación de Investigaciones Económicas para Lartinoamerica, Sotas Tecnicas, No. 56 (April 1983).

    • Search Google Scholar
    • Export Citation
  • Cortázar, Rene, and Marshall Jorge, “Indice de Precios del Consumidor en Chile: 1970–1978,” Colección Estudios, CIEPLAN,No. 4.

  • Cortés, Hernán, “Políticas de Estabilización en Chile: Inflación, Desempleo y Depresión 1975–1982,” Cuadernos de Economía,No. 6O,Catholic University of Chile (1983).

    • Search Google Scholar
    • Export Citation
  • Díaz-Alejandro, Carlos F., “Good-bye Financial Repression, Hello Financial Crash,” Journal of Development Economics, Vol.18 (September/October 1985).

    • Search Google Scholar
    • Export Citation
  • Díaz-Alejandro, Carlos F., and “Stories of the 30’s for the 1980’s,” in Financial Policies and World Capital Markets: The Problems of Latin American Countries,ed. byR. Dornbusch, M. Obstfeld, and P. Aspe-Arnella (Chicago: University of Chicago Press, 1983).