Brazil: Recent Economic Developments
Author: International Monetary Fund
Publication Date:
15 Apr 1998
eISBN:
9781451805888
Language:
English
Keywords:
ISCR; CR; state; federal government; export; derivatives activity; Banco da Amazônia; bank owner; replacement cost; average equity; Loans; Nonperforming loans; Commercial banks; Foreign banks; Global
Download PDF (2.8 MB)

This paper describes major economic developments in Brazil in 1997. A number of issues were analyzed in the paper, including the slow progress being made in the negotiation of the fiscal adjustment programs with the states, the sustainability of the growing current account deficit, as well as the strength of the banking system following macroeconomic stabilization. The paper discusses the post-Real crisis in the states and the state adjustment programs being negotiated with the federal government. Privatization and the associated foreign direct investment flows are also described.

Abstract

This paper describes major economic developments in Brazil in 1997. A number of issues were analyzed in the paper, including the slow progress being made in the negotiation of the fiscal adjustment programs with the states, the sustainability of the growing current account deficit, as well as the strength of the banking system following macroeconomic stabilization. The paper discusses the post-Real crisis in the states and the state adjustment programs being negotiated with the federal government. Privatization and the associated foreign direct investment flows are also described.

V. An Assessment of Risks in the Brazilian Financial System65

A. Introduction

115. This chapter analyses the Brazilian financial system which is the largest of the emerging market economies (with the exception of China) and by far the largest in Latin America.66 Section B reviews the structure and recent evolution of the Brazilian financial system, with particular emphasis on the banking system. Section C presents a preliminary analysis of the performance of banks with majority domestic shareholders based on detailed financial statements since December 1994. In order to assess the resilience of the private banks to the sharp increase in interest rates in October 1997, the impact of a doubling of nonper-forming loans on the financial conditions of individual banks in that group is simulated in Section D. Section E discusses the public sector banks (state banks and federal banks). Section F reviews valuation and disclosure practices, as well as the ways that market discipline takes place in Brazil. Section G looks at the supervisory framework and the role of the central bank, and Section H at the regulatory framework for banks’ derivatives and foreign exchange activities.

B. Structure and Recent Evolution of the Financial System

116. The Brazilian financial system is primarily composed of a large banking sector, mutual investment funds that are mainly managed and administered by banks, pension funds, and insurance companies, most of which are owned by or associated with bank holding companies. In 1997, total assets of these main components of the financial system are estimated to have amounted to about R$800 billion, of which about 70 percent were accounted for by the banking system, 20 percent by mutual investment funds, and 10 percent by pension funds and insurance companies.67 There are also investment banks, credit, financing and investment companies, and numerous smaller financial institutions such as securities dealers, brokerage firms, leasing companies, investment funds, credit cooperatives, mortgage companies, and real estate finance companies (Table 7). Most of these institutions also tend to be part of bank holding companies, and apart from the pension funds, insurance companies, and stock exchange houses, most fall under the supervision of the central bank, which is responsible for the supervision of over 3,000 financial institutions.

Table 7.

Brazil: The Financial System and Summary Supervisory Responsibilities

article image
Sources: Central bank; Securities and Exchange Commission (CVM); Superintendency of Private Insurance Industry (SUSEP); and Secretariat of Complementary Pensions (SPC); and Fund staff estimates.

CVM supervises banks that are listed on Brazil’s stock exchange, jointly with the central bank.

117. The investment funds and insurance industries have experienced strong growth since 1994. The insurance industry grew by about two and a half times since the inception of the Real Plan, with total premiums of R$18 billion in 1997. Most of the growth is accounted for by the development of term life insurance and capitalization products. Several leading international insurance companies have entered the Brazilian market in 1997.

118. The number of investment funds has grown from 1,000 in 1994 to almost 3,000 in 1997, following reforms that were enacted in 1995 with a view to increasing the transparency and supervision of the funds’ activities. Assets have grown from less than 2 percent of GDP in 1991 to almost 19 percent in 1997, or R$160 billion. At present, there exist about 30 different categories of investment funds, of which 25 are for residents and 5 for nonresidents. The central bank is in charge of the supervision of 17 types of fixed-income funds (funds with more than 51 percent of assets invested in fixed-income securities); Brazil’s securities and exchange commission is in charge of the supervision of 13 types of equity funds.68 Private pension funds had assets of R$70 billion in 1996.

119. The banking system is composed of a large private banking sector, several federal government-owned banks, including the two largest in the country, and the state banks. The structure of the Brazilian banking system as of mid-1997 is summarized in Table 8. Since the beginning of the Real Plan, the Brazilian banking system has undergone profound changes. As noted in Chapter IV, during the years of high inflation, the banks earned substantial profits from treasury operations (mainly arbitrage on interest rates and currencies) and from float on basic banking services (bill and tax collections).69 Since the stabilization of the economy in the second half of 1994, the banks’ profitability and capitalization have come to depend more on lending operations and fee income. The adjustment of the banking industry to the new low inflation environment was set back in 1995, when the central bank raised interest rates to cool domestic demand and protect the capital account, partly in response to the Mexican crisis. Higher interest rates in turn affected the debtors’ ability to repay bank loans and financial intermediation costs increased sharply reflecting increased provisions for nonperforming loans. Table 9 shows the evolution of nonperforming loans in the system, by category of bank, for the period June 1994 to June 1997.

Table 8.

Brazil: Summary Structure of the Brazilian Banking System

article image
Sources: Central bank; and Fund staff estimates.

Latest comprehensive data available from the central bank.

In addition, the federal government development bank, BNDES, is primarily engaged in medium to long-term lending, either directly or through agent banks, and implements the government’s privatization program by providing technical assistance on this area and through BNDESpar, which provides financing for privatization operations. The funding of BNDES is obtained mainly from Brazil’s mandatory unemployment and disability insurance fund and from borrowing abroad. It does not take deposits from the public. There are also five state-owned development banks.

Including the assets of Banespa as of the last available balance sheet before central bank intervention in December 1994. In January 1998, following the general agreement between the federal government and the state government of São Paulo concerning the restructuring of the latter’s debts, Banespa published its financial statements for 1995, 1996 and 1997. These indicate assets of R$49 billion at end-1995, R$63 billion at end-1996 and R$71 billion at end June 1997.

Table 9.

Brazil: Nonperforming Loans as a Proportion of Total Loans 1/

article image
Sources: Central bank; and Fund staff estimates.

Nonperforming loans are defined as loans in arrears plus loans in liquidation, and may be taken as an approximate equivalent of loans that are substandard, doubtful, or loss.

Estimated by Fund staff based on the balance sheet data of private banks with majority domestic ownership.

Special classification rules apply to loans of state banks to state governments; only installments that have fallen due and have not been paid are classified as non-performing. Mortgage loans may also be classified in this manner; this affects mainly the federal government owned Caixa Econômica Federal.

Central bank data, except estimated by Fund staff for December 1994 and June 1995.

120. Following a peak in the reported delinquency rate of about 9 percent in August 1996, the banking system has generally been able to reduce the incidence of nonperforming loans to more manageable levels. Most banks were able to absorb the associated costs, because low levels of lending resulted in generally high capitalization ratios. However, there remains a fair amount of disparity in the situation of individual banks, in particular among the private banks with a majority domestic shareholding.

121. The recent evolution of the structure of the Brazilian banking system is summarized in Table 10. Three trends are apparent from the data that cover the period since 1994: the share of the private sector banks in total assets has increased from 36 to 38 percent; there is a movement toward greater concentration among private sector banks (the share of the five largest private banks in total private banking assets has increased from 58 to 63 percent); the share of foreign ownership has grown from 13 to 18 percent, up to June 1997 and to 21 percent in January 1998. Correspondingly, the share of publicly owned banking assets has declined from 51 to 44 percent, reflecting mainly the exit of two state-owned banks from the system and the operational limitations placed on a few other state-owned banks by the central bank.70

Table 10.

Brazil: Banking System Assets

(In billions of reais, unless otherwise indicated)

article image
Sources: Central bank; and Fund staff estimates.

Private banks with more than 75 percent domestic ownership.

Share of five largest private banks in total assets of private banking system.

122. These trends, combined with the need to further reduce the incidence of nonperforming loans and with the supervisory pressures for adequate provisioning and capitalization, underscore the intensification of competitive pressures in the system. These pressures intensified during 1997 as may be seen in the sharp acceleration of mergers and acquisitions in the banking industry (Table 11). During the period June 1994 to December 1996, an estimated 7 percent of private banking assets changed ownership; in 1997 a further 20 percent of private banking assets changed ownership, most of which as a result of foreign entry. Overall, about 30 banks have changed ownership since the beginning of the Real Plan, of which 15 were taken over by foreign banks. The mergers and acquisitions should help reduce costs through the rationalization of operations.

Table 11.

Brazil: Major Mergers and Acquisitions Among Brazil’s Private Banks

(Number of banks)

article image
Sources: Central bank; rating agencies; press reports; and Fund staff estimates.Note: Including in 1997 an acquisition by a foreign bank that was announced on January 5, 1998. Excluding the privatization of state-owned banking assets, which in 1997 amounted to about R$6 billion, or 3 percent of private banking assets.

123. Two other developments that took place in the second half of 1997 have been the entry of major American, British, and Swiss institutions into the investment banking business, and the entry of two leading Brazilian investment banks into the retail banking business. The larger foreign presence as well as the larger size of the top private domestic banks imply that competition will intensify for the second-tier retail banks.71 It is expected that more second-tier banks will exit the market, look for strategic partners in order to achieve the critical mass needed to compete, or be forced into increasingly regional/specific business niches, such as consumer lending or lease financing.

124. The second-tier banks are also subject to increasing competitive pressures that have grown as a result of the greater integration of the Brazilian economy into the global financial system. Prime Brazilian corporate borrowers have been increasingly able to bypass the domestic banks and deal directly with the large multinational banks that offer more attractive rates through better access to funding.72 In addition, borrowing by nonfinancial corporations in international capital markets has increased sharply in the past two years. Corporations are also beginning to issue local currency debentures and commercial paper. Therefore, most domestic banks have been forced to turn to financing middle-market borrowers (and more recently consumers). The financial statements of the middle-market borrowers are typically succinct and frequently not audited, making credit assessment expensive and often incomplete, especially on the part of the large number of banks that were established in 1988–90 and that have not yet had to build risk-minimizing loan underwriting and monitoring systems. As a result, some banks tend to base their lending decisions on the quality and size of collateral, rather than on projections of borrowers’ repayment capacity.

125. Some factors tend to mitigate this picture of increasing lending risks, at least for some segments of the banking system. In particular, it is estimated that the average debt of the Brazilian corporate sector is on the order of 30 percent of its equity, indicating that Brazilian corporations can, in general, offer attractive cofinancing terms. Moreover, it is estimated that for private sector banks one-fourth of the value of loans in default is recovered within the year, a further third within the following two years, and whatever else is recovered thereafter.73 However, these figures represent averages that can mask sharp disparities across industries or regions and may inordinately affect certain banks whose lending portfolio is not reasonably diversified.74

C. Private Domestic Banks

126. In order to analyze the soundness of the private domestic banks and their resilience to macroeconomic shocks, this section presents a detailed analysis of the financial conditions of a large sample of these banks. It also includes a sensitivity analysis to quantify the impact on bank capital adequacy of a doubling of nonper-forming loans that could result from a macroeconomic shock, such as the recent doubling of interest rates.

127. The analysis is based primarily on financial statements submitted by banks to the central bank covering about two-thirds of private bank assets. Although data are submitted monthly, the analysis uses mainly the half yearly returns, because it is a requirement that these be audited by external auditors. Whenever needed, these data are supplemented by other sources from the central bank or from the private sector. In general, results need to be interpreted with caution, since ultimately they depend on the quality of the underlying practices concerning loan classification that are being followed by individual banks. In particular, the simulation will be sensitive to the quality of the data on nonperforming loans in June 1997. Moreover, a number of domestic banks that exited the system prior to June 1996 (some of which were large) were not included in the sample provided by the central bank. However, in the data that were provided, the assets of such banks appear to have been added to the assets of the acquiring banks that were included in the sample. Without detailed data of the banks that exited the system during the sample period, it was generally not possible to make the comprehensive adjustments that would be required to ensure consistency of the historical, backward-looking analysis. However, the data do provide a good basis to conduct a forward-looking simulation exercise, as is done here, to measure the impact of a doubling of nonper-forming loans on the capital adequacy of the private domestic banks.

128. Banks were grouped into four categories depending on the value of their capital adequacy ratio. Group A contains banks with capital adequacy ratios of less than 10 percent in any given period (the regulatory minimum in Brazil as of December 1997); group B contains banks with ratios between 10 and 12 percent; group C is for ratios between 12 and 16 percent; and group D, for ratios above 16 percent. The results, consisting of a set of standard indicators of bank performance, are presented in Table 12.

Table 12.

Brazil: Private Domestic Banks: Analysis of Bank Soundness 1/ December 1994-June 1997

(In percent, unless otherwise indicated)

article image
article image
Sources: Central bank; and Fund staff estimates.

Banks with less than 50 percent foreign participation and with demand deposits in June 1997.

Capital adequacy for December 1994 is estimated (before January 1995, banks were required not to exceed a 15/1 ratio of assets to supervisory capital). The standard was increased from 8 to 10 percent in May 1997, and to 11 percent in November 1997. The definition of supervisory capital consists of tier 1, plus revaluation reserves.

Demand, savings, and time deposits of non-financial entities.

Supervisory capital minus non-provisioned non-performing loans as a proportion of risk weighted assets.

Defined as personnel plus administrative expenses as a proportion of banking income.

Excluding banks that exited the system before December 1996. The estimated assets (deposits) of such banks are R$13.7 billion (R$5.4 billion). Banks that exited the system between June and December 1997 had estimated assets of R$30 billion and deposits of R$11 billion.

Weighted average.

129. Based on the data, it is possible to highlight the main characteristics of the Brazilian private banks, mainly by comparing the performance of the reportedly highest capitalized groups (C and D) with the least capitalized groups (A and B).75

  • Brazilian private domestic banks are generally well capitalized by international standards. About 75 percent of deposits throughout the period December 1994 to June 1997 were in banks with capital adequacy ratios in excess of 12 percent (group C and D). In June 1997, the weighted average capital adequacy ratio of those banks was 17.5 percent. However, this is partly the result of the high capitalization of the largest private banks, three of which are in group D. In general, the high level of capital in the Brazilian private banks may be explained by the relatively low level of their assets that carry 100 percent risk weight, with loans (which generally carry a 100 percent risk weight) still representing less than 50 percent of assets. Banks accumulated capital by means of profits generated during the years of high inflation. The relatively high levels of capital gives banks the possibility of expanding lending operations, especially to market segments with higher profit margins such as consumer lending. Indeed, the drop in the capital adequacy ratio over the past year for banks in groups C and D is largely explained by the recent sharp increase in such lending.

  • Brazilian private banks are generally profitable. The returns on average equity for banks in groups C and D were on the order of 13 to 15 percent throughout the period. The efficiency ratio was between 50 and 75 percent for those banks. Both measures compare well with values for major international banks. Group B banks were also able to maintain fairly high levels of return on average equity, on the order of 12 to 15 percent.76 As of January 1998, only a few banks had released preliminary earnings data for the second half of 1997, and these showed earnings that were above those of previous periods.

  • All groups have maintained high levels of provisions, in excess of 100 percent of non-performing loans, through most of the period and for most of the four groups. The phenomenon of overprovisioning is largely explained by the tax treatment of specific loan loss provisions. Only groups A and B banks sometimes had provisioning levels below 100 percent as the incidence of non-performing loans increased in 1995–96 and overwhelmed earnings.

  • Excess provisions constitute hidden reserves that should be taken into account in the assessment of capital adequacy. To this effect, an indicator labeled “capital cushion” in Table 12 has been calculated by adjusting the ratio as computed by the central bank for the extent of overprovisioning. Overall, this adjustment improves reported capital adequacy by adding about 1 percentage point to the average capital adequacy ratio of the sample. As expected, the upward adjustment reflects the overprovisioning of banks in groups C and D. In general, weaker banks (which here are in groups A and B) will be reluctant to provision in order to improve the return on capital, and indeed, for group A banks, the adjustment is slightly downward. For group B banks, the difference between actual and adjusted ratio is, however, minimal. It is known that some of the banks which are exposed to a large construction and agrobusiness corporation, that began to restructure R$550 million of commercial bank loans in mid-1996, have not fully provisioned these risks (although loans have been appropriately reclassified). The impact of this reluctance to provision for capital adequacy will become fully apparent in the simulation exercise conducted below.77

  • Under the assumption that banks which exited the system between December 1995 and March 1997 (the last one being Bamerindus) were in groups A and B, the level of assets (deposits) in weak banks declined from 47 percent (42 percent) of the sample’s assets (deposits) in December 1995 (the beginning of the previous period of banking problems) to 32 percent (28 percent) in June 1997.78

130. A recent development that could be a source of concern is the net migration of about a dozen medium-sized banks representing about 24 percent of the sample’s assets in June 1997 from groups C and D to groups A and B since December 1996.79 Moreover, the average profitability of group A banks (about 11 percent of sample assets in June 1997) turned sharply negative in December 1996 and stayed negative in June 1997, reflecting recent large investment in systems and personnel, dependence on short-term interbank funding, and persistently high levels of nonperforming loans whose income cannot be recognized. The migration of banks to lower capital adequacy groups is partly explained by the mechanical impact of the growth of lending, on the capital adequacy ratio and indeed similar movements can be seen in group C and D banks, as mentioned already. Figure 26 shows the pattern of this migration between December 1996 and June 1997 for the largest 20 banks in the sample and the resulting clustering of medium-sized banks at lower levels of capital adequacy in June 1997.80

Figure 26.
Figure 26.

Brazil: Private Domestic Banks: Total Assets and Capital Adequacy Ratio, December 1996-June 1997

(For the 20 largest private domestic banks)

Citation: IMF Staff Country Reports 1998, 024; 10.5089/9781451805888.002.A005

Source: Central Bank.

D. Simulation of the Impact of a Macroeconomic Shock on the Private Banks

131. In order to evaluate the resilience of the Brazilian financial system to a macroeconomic shock, it is appropriate to quantify the impact of a sharp deterioration of the quality of the loan portfolio of the private domestic banks. For this purpose, a simulation was conducted to measure the deterioration of capital adequacy that would result from a doubling of nonper-forming loans (from the level existing in June 1997), such as could be brought about by a macroeconomic shock. Such an event would result in levels of loan delinquencies equivalent to about 150 percent of previous peak levels, except for group D banks, and would imply a significant shock to the system (Table 13). Previous peak levels were recorded in the first half of 1996 after the central bank increased interest rates sharply in March 1995.

Table 13.

Brazil: Incidence of Nonperforming Loans

(Share of non-performing loans in total loans, in percent)

article image
Source: Table 12.

132. The simulation work was based on three scenarios that are reported in Table 14. In scenario A, the actual capital adequacy ratio as calculated by the central bank is adjusted downward by the level of nonprovisioned nonperforming loans that existed in June 1997, with the result that 15 banks (12 percent of sample’s assets) would have a capital adequacy ratio of less than 10 percent, the minimum required by the central bank by the end of 1997. In scenario B, the capital adequacy ratio is recalculated under the assumption of a doubling of nonper-forming loans from the level of June 1997. Undercapitalized banks would then increase to 30. Under scenario C, a conservative estimate of profits for 1997 is factored into the calculation of capital adequacy: profits for the whole of 1997 are estimated to be half of the value obtained by extrapolating the trend of profits for the last semester of 1996 and the first seven months of 1997, and this value is added to supervisory capital for the purpose of calculating capital adequacy. Under scenario C, 23 banks (16 percent of assets) would be formally undercapitalized.81

Table 14.

Brazil: Potentially Undercapitalized Banks: Simulations Based on June 1997 Data

(In millions of reais, unless otherwise specified)

article image
Sources: Central bank; and Fund staff estimates.

Capital adequacy ratio recalculated as supervisory capital, minus nonprovisioned nonperforming loans, as a proportion of risk weighted assets.

Excluding banks that have exited the system since June 1997 and one bank with a reputable international bank as large minority shareholder.

Capital adequacy ratio as under Scenario A, with the added assumption that the June 1997 level of nonperforming loans doubles and is fully subtracted from supervisory capital.

Capital adequacy ratio as under Scenario B, with the added assumption that 1997 profits are fully used toward provisioning nonperforming loans. Profits for 1997 are estimated to be half of the value obtained by extrapolating the trend of profits for the last semester of 1996 and the first seven months of 1997.

Weighted average capital adequacy of banks excluded between scenarios B and C: 8.2 percent.

Weighted average capital adequacy of banks excluded between scenarios B and C: 10.5 percent.

133. However, each scenario must be further adjusted to take into account the exit of banks from the system since June 1997 (banks that became majority foreign owned since June 1997) and thus have the backing of head offices abroad. This is reported in the columns “Adjusted Scenario A,” “Adjusted Scenario B,” and “Adjusted Scenario C.” Under adjusted Scenario C, the outcome is that a group of 18 banks representing 7 percent of the assets of the private domestic banks would be undercapitalized and would be candidates for corrective action by the supervisory authority. The three largest banks in this group had aggregate assets of R$8.8 billion (55 percent of that group or 4 percent of private banking assets), in June 1997 and were the fifteenth, eighteenth and nineteenth private banks in the sample in terms of asset size.

134. It is useful to compare the results of the simulation of such a shock to the financial system with the previous period of banking problems in 1995–96. This comparison, made in Table 15, shows that the value of assets (or deposits) in banks that may be considered under-capitalized following the shock is roughly one-third of what it was during the previous period of banking problems. There is an upward bias in this comparison, however, as the data for 1995–96 only include those banks that actually exited the system, and exclude banks whose owners decided to recapitalize, whereas the data for 1997–98 taken from the simulation assume that all undercapitalized banks exit the system. If adjustments could be made for this bias, less than 7 percent of private banking assets would be considered at risk under this scenario and the R$16 billion of assets and R$4 billion of deposits should be interpreted as upper limits of a range of outcomes following the shock.

Table 15.

Brazil: Extent of Banking Sector Weaknesses: Comparison of 1995-96 and 1997-98

(In billions of reais, unless otherwise specified)

article image
Sources: Central bank; and Fund staff estimates.

Including all private domestic banks that exited the system between December 1995 and March 1997.

Data from the last column of Table 14, assuming that none of the concerned bank owners would recapitalize their bank.

135. The extent of banking sector weaknesses also can be assessed by referring to the use of public funds for restructuring private banks during the previous period. Disbursements under the central bank’s special facility for restructuring the private banking sector, PROER, amounted to R$21 billion for seven operations of financial assistance, for the period December 1995 to March 1997. If the proportion of PROER disbursements to assets that exited the system in 1995–96 is maintained in 1997–98 and assuming that no bank owners decide to recapitalize their bank and that the central bank decides to provide financial assistance to all 18 banks, R$10 billion may be taken as an upper limit of further PROER type operations, should the shock materialize to the extent assumed. The absorption of such liquidity would represent about 5 percent of outstanding federal government securities, compared to about 13 percent for the period 1995–96.

136. Disbursements under the deposit insurance fund that was created in 1996 amount so far to about R$3 billion, of which R$2.5 billion were used in the large restructuring operation of Bamerindus. The net present value of the fund is estimated by the central bank at between R$3.5 and R$5.5 billion, depending on the assumed long-term discount rate. Although only part of the deposits are covered by the deposit insurance fund, it is likely that the fund would be insufficient to cover further operations related to the exit from the system of R$4 billion worth of deposits, should a shock of this magnitude materialize.

137. In conclusion, the extent of weaknesses in the Brazilian private banking system appears to be appreciably smaller than two years ago. However, should the level of nonperforming loans increase significantly following a large macroeconomic shock as simulated here there would continue to exist up to 7 percent of assets that are in banks that may be considered weak. But if the order of magnitude of the problem is similar to the simulated one, the monetary impact of helping banks exit the system with the use of public funds should be relatively small.

E. Public Sector Banks

138. The segment of the banking industry consisting of the state banks (banks majority owned by state governments) is undergoing a process of reorganization, through privatization, conversion into development agencies, recapitalization or liquidation, as part of an explicit policy to reduce the role of the states in the banking sector.82 As discussed in Chapter II, losses of the state banks will be covered by the state governments. Initially, losses will be financed by federal bonds and the states will compensate the federal government as part of the restructuring of the state debt currently being finalized. Most of the state debt is owed to the state banks.

139. In the case of state banks being liquidated or converted into development agencies, remaining deposits in those banks will be transferred to the federal banks with the liquidity of the federal banks ensured by the issuance of the above-mentioned bonds by the federal government. In February 1997, a facility (PROES) was created at the central bank to provide bridge loans to the federal banks in this process; up to October 1997, PROES had disbursed R$854 million on account of the restructuring of Bemge, one of the two state banks of Minas Gerais.83 Bemge, with assets of R$3 billion, is to be privatized in the first half of 1998 to be followed in the second half by the privatization of Banespa, the state bank of São Paulo; market participants in Brazil expect considerable interest in this privatization.84 Banerj, the state bank of Rio de Janeiro with assets of R$8 billion, was privatized in July 1997. Of the remaining 23 state owned banks, the authorities have so far liquidated 2 (Alagoas and Amapá), and expect to be able to privatize another 6. When the exit policy is completed, the most important remaining state bank will be the savings and loans institution Nossa Caixa Nosso Banco of São Paulo. In June 1997, its assets of R$13 billion made it the country’s eighth largest bank; as part of the restructuring agreement for the debt of the state government of São Paulo with the federal government that was approved by the senate in December 1997, Nossa Caixa will benefit from a swap of claims on the state government for negotiable federal securities worth R$7 billion.

140. The two large federally owned banks, Banco do Brasil (a commercial bank) and Caixa Econômica Federal (a savings and loans institution), have made good progress in restructuring their operations and restoring capital adequacy. These banks represent about 35 percent of Brazilian banking assets.

141. The new management in place at Banco do Brasil since 1995 has embarked on an ambitious program of investment in systems, skills, and new lines of business (insurance and capital markets), that will allow it, among other things, to publish its financial statements according to United States Generally Accepted Accounting Principles (GAAP) by 1998 and to strengthen its internal controls and procedures for risk management, especially in loan approval and monitoring.85 The associated implementation of new management principles, as well as the central bank’s requirement to comply with a new capital adequacy standard by January 1995, forced the bank to write off losses in its lending portfolio (R$5 billion out of a portfolio of R$47 billion, with a further R$6 billion of rural sector loans being restructured under a securitization program that was financed by the federal government) and expense foreign exchange valuation losses (R$2.7 billion). The recapitalization of Banco do Brasil took place in June 1996 for an amount of R$8.3 billion and was underwritten mostly by the federal government through R$6.4 billion in federal securities that pay floating market rates and have maturities of up to 15 years.86

142. Banco do Brasil has a large stock of directed loans to the agricultural sector. At end 1996, the bank’s rural lending represented 54 percent of its lending to the private sector (R$22 billion), 80 percent of which were concentrated to large farmers and agrobusinesses, and 6.4 percent of which were nonperforming, even though 27 percent of rural loans (mostly small value loans), had been restructured. Negotiations for restructuring the rest of the rural loan portfolio are continuing. The reported level of nonperforming loans in the remainder of the private sector credit portfolio stood at 19 percent.87 However, provisions covered 113 percent of total nonperforming loans and were further increased to 138 percent at end June 1997.88 After decreasing to 5.7 percent at end 1995, the bank’s capital adequacy ratio stood at 11.1 percent at end-1996 following the recapitalization and declined somewhat to 9.6 percent in June 1997 as a result of a recent expansion in consumer lending.

143. Banco do Brasil further improved its net worth in March 1997 by R$2.6 billion, when the central bank allowed banks that sustained one time restructuring losses to gradually deduct such losses from future profits in order to meet well-defined future obligations, such as pension outlays or provisions on loan losses that result from differences between tax laws and accounting rules on the deductibility of specific provisions. Banco do Brasil, which sustained large losses in 1994 and 1995 (as a result of the more conservative provisioning of its rural lending and the recognition of the foreign exchange valuation losses), used a tax credit of R$11.3 billion in order to meet the large obligations of its pension fund (R$8.7 billion) and to constitute new provisions on rural lending (R$2.6 billion). Banco do Brasil’s liquidity appears to be well ensured by its large branch network.

144. The Caixa Econômica Federal (CEF) is Brazil’s main savings and loan institution, with total assets roughly equivalent to those of Banco do Brasil. CEF had cash-flow problems in 1992, because of nonperforming loans to the states that were made under the Fundo de Garantia de Tempo de Serviço (FGTS).89 As a result, CEF stopped making new housing or FGTS-funded loans until the end of 1996, while the state debt was rolled over by the federal government in 1994 under law No. 8727/94. Subsequently, CEF has been acquiring claims of state banks on those states that have signed debt rescheduling agreements with the federal government, the liquidity of which may be ensured by the central bank’s PROES facility pending the formal exchange of those claims for federal negotiable securities. The total of state debt on CEF’s books represented R$1.1 billion (16 percent of CEF’s credits) at end June 1997; it is being serviced by the federal government. Since then, CEF has had a good liquidity position as a result of the stability of its funding from deposit accounts (56 percent of its liabilities) and from funding through FGTS (21 percent).

145. CEF’s liquid assets (interbank placements and negotiable securities holdings) represented about 15 percent of assets in June 1997. CEF’s nonliquid assets are mainly comprised of housing credits (36 percent of total assets), claims on state governments (11 percent), mortgages acquired under PROER (the program for restructuring private banks, 7 percent),90 commercial loans (7 percent), and FGTS urban infrastructure loans (4 percent). About 15 percent of the housing credits and PROER mortgages are claims on the government on account of the mortgage guarantee fund (FCVS), a program to subsidize individual housing.91

146. The reported incidence of nonperforming loans is low (less than 1 percent), reflecting the large share of individual housing loans in the portfolio and the more lenient classification requirements that apply to such loans (as noted earlier, only installments in arrears are classified). CEF management estimates that, if the principal of such loans had to be classified, nonperforming loans would only increase by R$1.2 billion (less than 2 percent of the loan portfolio). Because CEF is often the only bank for such borrowers, the cost of defaulting on a CEF mortgage is prohibitive in terms of access to banking services. Since 1996, CEF has been required to make general provisions against FGTS loans and is fully provisioning nonperforming loans in its commercial lending. As a result, provisions for credit risks represented 5.4 percent of the loan portfolio in June 1997 compared to 4.3 percent a year earlier. These figures reflect a conservative policy of overprovisioning compared to the level of nonperforming loans recorded on the books, which provides a cushion sufficient for absorbing a five fold increase in loan defaults.

147. The CEF’s capital adequacy, under Brazilian standards, is estimated to have been 7.8 percent in June 1996, 8.2 percent in December 1996, and 7.5 percent in June 1997.92 The recent deterioration may be attributed to the resumption of housing and urban development loans. In the case of CEF, the capital adequacy calculation is somewhat distorted by the high proportion of FCVS assets, which although they are claims on the federal government, carry a 100 percent weight under Brazilian standards of capital adequacy. If those assets carried a weight of zero percent, it is estimated that CEF’s reported capital adequacy would have been 9 percent in June 1997.

148. The CEF plans to securitize its housing loans, as a means to originate new lending. CEF’s intention is to focus on its comparative advantage in the origination of residential mortgages. The process of securitization will involve the standardization of the existing loan contracts, particularly stripping the FCVS component. CEF continues to buy back the housing loan portfolio that the private banks have originated under FCVS. This buy-back, combined with the securitization, enables the federal government to progressively recognize FCVS liabilities and correspondingly contain FCVS contingencies. At the end of the process, CEF will be the residual holder of FCVS claims, which it can afford to hold to maturity, because of its otherwise good cash flow.

149. The third largest federally owned bank, Banco Meridional, with assets of R$2.5 billion at end 1996, was privatized in November 1997. The other federally owned banks, Banco da Amazônia and Banco do Nordeste, had estimated assets of R$4 billion at end-1996.

F. Valuation, Disclosure Practices, and Market Discipline93

150. The valuation and disclosure practices followed by the leading private Brazilian banks compare favorably with practices followed in the mature financial markets, such as the United States. The leading private domestic banks publish statements that are prepared both according to Brazilian GAAP and U.S. GAAP; the large publicly owned Banco do Brasil intends to do so in 1998, and rating agencies have noted the high quality of its recent disclosure practices.

151. With respect to the consolidation of operations in subsidiaries and affiliates, Brazilian accounting and consolidation rules for banks are more stringent, and Brazilian financial statements have a higher level of disclosure in that area than that required under U.S. GAAP, especially for companies whose shares are traded in stock exchanges. In general, shareholdings exceeding 25 percent in an affiliate have to be consolidated line by line. Although financial instruments are not necessarily marked to market, the leading banks in Brazil use mark to market valuation methods and departures from that principle by the second-tier banks is taken into account by banking counterparties and rating agencies in the form of higher funding costs, lower credit lines, and lower ratings. In addition, recent regulations require banks to disclose more information on securities holdings.94

152. The central bank gives banks some discretion in recognizing impaired loans. Loan classification and provisioning are generally triggered by payment arrears lasting more than 59 days, rather than by analysis of borrowers’ ability to repay; in the case of export/import loans, classification is triggered after 29 days of the loan being overdue and 100 percent provisioning is required. The level of mandated provisioning depends on the existence and quality of collateral. In the case that classification as bad loans is mandated, all claims out-standing to the particular borrower must be classified as such (as indicated above, housing loans and loans to state governments are an exception to this rule).95 The rules on income recognition appear appropriately strict.

153. In practice, the leading banks interpret the supervisor’s rules for loan classification and provisioning and for income recognition in the most conservative sense: loans that are classified as nonperforming are immediately provisioned at 100 percent.96 Moreover, as indicated before, private Brazilian banks tend, with a few exceptions, to overprovision. Accordingly, it is not clear that changes in the loan classification system toward one based on the borrower’s ability to repay would necessarily result in higher levels of provisions compared with the present system, as levels of provisions in such a new system tend to be more progressive (less stringent) than the norms followed by the leading Brazilian banks.97

154. Operations with related parties and the concentration of lending to borrowers under the same ultimate control are both regulated, but the level of disclosure and the scope of definition of related parties are not as comprehensive as in the United States. However, the leading banks disclose information in a manner broadly comparable to United States practice, as do the second-tier banks that have submitted to due diligence of the rating agencies. Aggregate information on the geographic concentration of risk and on meaningful sectoral risk is not readily available, a factor which tends to complicate peer group analysis. The full introduction of the central de riscos (see below) should correct this anomaly.

155. The leading banks disclose comprehensive information on interest rate and foreign exchange mismatches, off-balance sheet contingencies and operations in the derivatives market (see below).

156. Market discipline in Brazil may be seen in the extent of peer pressure, the development of rating agencies, and the strengthened role of external auditors after the well-known problems that emerged in the case of Banco Econômico and Banco Nacional in 1995. The intensification of competition with large scale entry of financial institutions of international reputation have brought about a considerable strengthening of peer pressure over the past two years, especially on the second-tier banks. Ongoing assessment by peers has become easier as the end of inflation has tended to force improvements in the transparency of accounts. The failures of the large banks mentioned above and the widespread problems of the state banks have pushed the banking system to strengthen the credit analysis of banking counterparties. Since then, most interbank operations, and these can be a sizable source of funding for second-tier banks, take the form of repos, and only prime banks have access to uncollateralized credit.

157. The international rating agencies are present in Brazil and financial groups with significant operations are routinely scrutinized. All financial institutions are required by the central bank to publish semi-annual financial statements that are audited by independent auditors, as well as submit to the central bank monthly unaudited financial statements. Listed institutions also have to publish the first and third quarterly statements. Many financial groups are increasingly tapping cheaper foreign funding (as permitted under the resolution 63 family for on-lending operations, and since 1992 under Annex V of Resolution 1289 for foreign listed depository receipts) and in this context have been exposed to the broader scrutiny of international capital markets.

158. Most auditing firms in Brazil are affiliated with the leading international auditing firms. Following episodes in which external auditors of international reputation issued unqualified opinions of impaired banks, the central bank strengthened the role of the external auditors in the appraisal of banks.98 Audit work is now organized to complement the inspection rounds of the central bank; auditors must immediately report to the central bank any fact that would materially affect the financial condition of a bank, including any refusal to disclose needed information; external auditors may be held liable for not reporting such facts; a bank cannot retain the same external auditors for more than four consecutive audits (two years); the bank has to designate a member of its board of directors as the person directly responsible for ensuring the proper disclosure of financial information to the auditors and to the central bank.99 However, in a recent case of a merger involving a second-tier bank, the external auditors were pressured by the central bank into giving unqualified opinions, in order to afford the central bank more time to find a buyer for that bank. This episode has adversely affected the work of some of the rating agencies, who rely on published information supplemented by due diligence.

G. Supervisory Framework

159. There are four bodies that supervise the Brazilian financial system: the central bank, the securities and exchange commission (CVM), the insurance industry superintendency (SUSEP), and the secretariat of complementary pensions (SPC). All report to the national monetary council (CMN), comprised of the president of the central bank, the minister of finance and the minister of budget and planning (see Table 7 for a summary of the respective supervisory responsibilities).

160. The central bank’s supervisory activities are discussed in detail below. It should be noted that, in addition to supervising the financial institutions listed in Table 7, the central bank regulates derivatives markets (exchanges and over the counter) that trade nonequity contracts, as well as the participation of foreign investors in all derivatives markets.

161. The CVM supervises all corporations that are listed on one of Brazil’s stock exchanges, including the 16 banks (13 private and 3 majority owned by the public sector) that are listed. It regulates the issuance, trade and custody of all equity instruments in Brazil (stocks, debentures, subscription bonds, commercial paper, depository receipts, warrants, stock indices, and shares in equity mutual investment funds) and their derivatives. CVM is member of the International Organization of Securities Commissions (IOSCO) and of the Council of Securities Regulators of the Americas (COSRA).

162. The central bank’s supervision of banks and financial groups has been considerably tightened over the past two years.100 Between July 1994 and December 1997, the central bank intervened a total of 43 banks, of which 32 were private (31 of which were liquidated or are in liquidation proceedings), 7 were state banks, 2 were banks with foreign participation, and 1 was an investment bank. As a result, about 16 percent of the 271 banks that existed in the system in July 1994 were intervened. The central bank also intervened (mostly through liquidation) a further 64 nonbank financial institutions during this period.

163. A central bank report from 1996 indicates that since the creation of the central bank in 1964 about 80 percent of its interventions in financial institutions took place during the period following the implementation of Real Plan. The constitution of 1988 had allowed freer entry into the banking system and introduced the concept of universal banking, with the result that the number of banks almost doubled. Central bank supervision therefore had to deal with a much larger number of institutions, and financial data that were distorted by inflation. As float income came to an end with macroeconomic stabilization in 1994, the continued unsound management of certain banks (some of which had previously benefited from regulatory forbearance) resulted in their capital becoming inadequate after the central bank raised interest rates in March 1995. The legislation on bank intervention as it existed until 1995 (mainly the regime of special temporary administration of banks by the central bank) effectively did not allow for preventive intervention or corrective action to preserve the bank as a going concern, except through the nomination by the central bank of a temporary administrator whose decisions on restructuring or liquidation (such as selling assets) could be blocked in court by bank shareholders/directors.101 As a result, preventive intervention was in practice only feasible in the case of state-owned banks; in such cases an understanding could generally be worked out by which the state governments would not challenge the decisions of the temporary administrator.102

164. In order to ensure the orderly exit of weak banks, the central bank created PROER in November 1995, a legal and financial mechanism to change the ownership of private sector banks.103 Under the legal provisions of PROER, the central bank can formally ask the owners of weak institutions to recapitalize and restructure operations under a prompt corrective action procedure, or, if the owners’ plan is rejected as inadequate, the central bank can intervene and formally dispossess the owners with a view to selling the bank or liquidating it. In addition, should the central bank be forced to intervene, the personal assets of any of the owners, directors or auditors that were involved with the bank during the previous 12 months are escrowed, pending a due diligence investigation of the financial condition and operations of the bank. Under the financial provisions of PROER, the central bank can quickly sell an intervened bank by finding new owners, as the take-over is facilitated by the use of public funds to separate the “bad bank” from the “good bank.”104 In practice, the central bank intervenes once the new owners have been identified and the financing deal finalized.

165. The privately funded and mandatory deposit insurance scheme (Fundo Garantidor de Créditos) was extended to all deposits in 1996; in the last operation to date under PROER (the take-over of Bamerindus by HSBC), the deposit insurance fund provided guarantees for central bank lending to the bad bank part of the intervened bank. The main impact of PROER and the deposit insurance fund has been to convince the owners of private banks that the central bank will not hesitate to dispossess them if warranted and that it is willing to pay the price to do so (so far some R$21 billion or 2.8 percent of 1996 GDP). As a result, moral hazard appears to have been significantly reduced in the system. Furthermore, the authorities have allowed since 1996 100 percent foreign ownership of banks and one foreign bank has benefitted from large PROER assistance in its take-over of a Brazilian bank.105 The policy of allowing large scale foreign entry, as well as the momentum generated by PROER may be viewed as turning points in Brazilian banking history and have resulted in the unprecedented number of mergers and acquisitions that were documented previously.

166. With these staff-intensive restructuring operations largely behind it, the central bank was able to devote increased resources to the thorough inspection of the small- and medium-sized private banks in 1997. The top five or six domestic private retail banks are viewed by the market as being on a sounder financial footing and are subject to the market discipline implicit in these banks’ international exposure. The data on the reported capital adequacy of the large banks that were discussed in the section dealing with the private domestic banks support the central bank’s strategy of monitoring them less intensively for the time being. Staffing constraints dictated that the inspections of small- and medium-sized banks be conducted in two rounds and these will be concluded by the beginning of 1998; some of these inspections involve examinations of foreign operations and, where allowed by host supervisors, these are conducted in joint teams with the host supervisors. Similarly, Brazilian bank inspectors participate as host supervisors in the on-site examination of foreign banking operations conducted by home supervisors.

167. Since January 1995, the central bank has assessed bank capital adequacy on the basis of a Basle-type risk-weighted capital adequacy ratio.106 In May 1997, the regulatory threshold for corrective action was increased from 8 percent capital adequacy to 10 percent, effective at the end of 1997; in November 1997, the ratio was further raised to 11 percent, effective by the end of 1998.107 In Brazil, only tier 1 shareholders’ equity, plus revaluation reserves, count toward meeting minimum supervisory capital requirement.108 As a result, Brazil’s current capital adequacy standard is equivalent to 13–15 percent under Basle rules, depending on the capital profile of the particular bank.109 Brazil has also had in place capital requirements for the coverage of counterparty credit risks arising from the banks’ derivatives activities; these have also been more stringent than recommended by Basle. However, Brazil does not yet have capital charges for market risks, including foreign exchange risk. In their monthly submissions of information to the central bank, banks are required to disclose the details of their derivative positions and submit their own calculation of the capital charge against the associated risks; this information has to be certified by a designated member of the board of directors of the institution.110 The parameters on which these calculations are based, as well as the prices used to mark the positions to market, are set uniformly for all banks by the central bank (based on the data provided by Brazil’s main commodities and futures exchange), instead of being left to the discretion of banks as is allowed under Basle.

168. The calculation of supervisory capital is adjusted for consolidation, in order to account for the effects of shareholdings in other financial or nonfinancial entities and cross-border ownership structures.111 For shareholdings abroad that are consolidated in a bank’s financial statements, if the central bank does not receive the authorization to inspect any of the entities in which these investments were made, 75 percent of the value of the assets that are consolidated were deducted from supervisory capital as of January 1, 1998, and 100 percent will be deducted from June 1, 1998.

169. Should a bank fail to meet the minimum capital requirement, the August 1994 regulation that established Brazil’s capital adequacy standard also spells out the procedure for prompt corrective action. The bank’s board has five days to comply with the central bank’s notification to meet with the supervisory authority, at which point, the board is given two weeks to present a restructuring plan. If accepted by the central bank, the bank is given six months to meet the objectives of the restructuring plan, including the restoration of its capital. The implementation of the plan is monitored both directly by the central bank and by means of monthly reports that external auditors have to submit to the central bank. In particular, the bank is prohibited from distributing profits in excess of the 25 percent level mandated by Brazilian corporate laws; it is also prohibited from engaging in operations that would further weaken its capital adequacy. In the meantime, the capital deficiency must be covered for no more than 90 days by a blocked deposit of the bank owners at the central bank, as the bank owners make the necessary arrangements to pay in more capital.

170. The central bank started to put in place in 1997 a centralized credit risk assessment system (central de riscos (CR)).112 As of December 31, 1997, the system is used to measure the overall exposure of the banking system to any one borrower, with differences in the details that are required for the reporting of large exposures. Every bank financing (loan, leasing operation, advance, guarantees given, or losses on such operations) that exceeds a value of R$50,000 and any collateral attached to such financing have to be reported monthly to the CR, loan by loan, classified by entity exercising ultimate economic control over the borrower; before December 31 financing operations were reported by borrower. In case a bank for the first time reports a loan of more than R$50,000 to an economic entity, then that bank’s loans of less than R$50,000 to that entity also must be reported loan by loan. For bank financing operations of less than R$50,000, only the aggregate of such operations classified by entity exercising ultimate economic control must be reported. The central bank may disclose this information to other lenders in the system with the agreement of the borrower. The central bank intends to broaden the reporting requirements to include credit scoring (the leading banks already use scoring methods as a tool of credit risk analysis), and the separate identification of restructured loans. In this way, the central bank will be able to recognize discrepancies in the risk evaluation that different banks make of the same borrowers. Once the credit scoring information is available and streamlined, it will serve as the tool for reforming Brazil’s loan classification system, permit the effective enforcement of such a system, and the better monitoring of banks’ classification and provisioning practices.

171. The central bank formally launched on December 11, 1997 a program for the modernization of its supervision practices (programa de aperfeiçoamento dos instrumentos de atuação do Banco Central do Brasil ao sistema financeiro nacional—Proat). The broad objectives of the program are to strengthen the central bank’s supervisory practices, upgrade the skills of its human resources, and evolve a strategic vision for the future of the financial system. The central bank expects that the program will permit a deeper understanding of the banks’ lines of business, especially in the nonbank financial sector, and generally enhance its evaluation of the banks’ internal procedures for the management of risks.113 The program benefits from a technical assistance loan of US$20 million over three years that the Board of the World Bank approved in November 1997.

H. Derivatives and Foreign Exchange Activities

172. The Bolsa de Mercadorias e Futuros (BM&F) located in São Paulo has grown from being the seventh largest commodities and futures exchange in the world in 1993 to the fourth largest, after Chicago’s two markets and London. BM&F’s turnover of notional contract value increased 46 percent in 1996 to reach US$7 trillion, with about 500,000 contracts being negotiated daily. BM&F is by far the largest such exchange in Brazil. As is the case of exchanges in other countries, BM&F turnover has been concentrated in interest rate and currency derivatives (mainly futures and options on interest rate swaps). As explained below, BM&F maintains a policy of encouraging simple contracts, for which the risks borne by the counterparties can be easily monitored.

173. There are three interesting features of the Brazilian derivatives market (both exchange-traded and over the counter) that suggest a market characterized by reasonably low levels of risks compared to other developed markets around the world. First, over the counter (OTC) contracts in which a financial institution is either a counterparty or a broker have to be registered in one of two custodial systems (CETIP or BM&F); as a result, only transactions directly contracted between two nonfinancial entities are not registered. The practice in other countries is that OTC contracts, for which risks are typically larger than for exchange traded contracts, are not centrally registered, so that surveys of OTC markets are the only means of measuring their size and composition. Second, in the registration of the contracts, the financial institutions have to disclose the identity of the final clients, so that the custodial systems maintain the accounts of those parties (about 6,000 accounts at BM&F), rather than the accounts of the financial intermediaries, as is the practice in the custodial systems used by exchanges in most developed economies. Finally, if the parties to an OTC transaction register their contract with the custodial system of the BM&F, they can ask to benefit from the same guarantees that apply to BM&F’s exchange-traded contracts, and BM&F will directly set and monitor limits on the final counterparties, rather than relying on the intermediaries to perform this analysis and enforce limits on their clients.

174. Data reported to the central bank by banks that operate in Brazil on their derivatives activities for September and October 1997 are presented in Table 16. These show that OTC contracts in which banks are a counterparty were only two times larger than exchange-traded contracts, compared to six times larger in an April 1995 BIS survey of derivatives markets.114 More importantly, contracts with guarantees accounted for some three-fourths of total outstanding notional value. As a result, the replacement cost (for contracts without guarantees), which provides a better perspective of the actual credit exposure, was less than 0.5 percent of the notional value, compared to 5 percent in the 1995 BIS survey.

Table 16.

Brazil: Derivatives Activities of Private Banks, as of End-September and End-October 1997

(In billions of reais)

article image
Sources: Central bank; Commodities and Futures Exchange (BM&F); and Fund staff estimates.

OTC contracts in which financial institutions are either counterparty or broker have to be registered in one of two custodial systems, CETIP or BM&F. Parties who register contracts in BM&F have the option of purchasing the same guarantee that applies to exchange-traded contracts.

Contracts in which a financial institution acted as a broker.

Only for OTC contracts without guarantees. All other contracts are excluded from capital requirements.

Calculated as 10 percent of replacement cost plus 16 percent of potential future exposure (as of December 1997).

Calculated as 11 percent of replacement cost plus 20 percent of potential future exposure (as of December 1998).

Estimated as 50 percent of 8 percent of (replacement cost plus potential future exposure).

175. In June 1997, the central bank introduced capital requirements for certain derivatives activities of banks. These requirements are to cover counterparty credit risks, but not yet market risks, related to derivatives. Counterparty credit risks arise from the probability that a counterparty, from which the bank expects payment in settlement of derivative contracts, defaults. Market risks related to derivatives arise from the probability that changes in market prices will result in losses to banks that have to make payments in settlement of derivatives contracts whose value is sensitive to those changes.115 The Basle Committee on Banking Supervision introduced capital requirements on counterparty credit risks related to derivatives in 1995, and on market risks in 1996.116

176. The methodology adopted by Brazil results in higher capital requirements than under Basle. Table 16 presents estimates of the differences between the capital requirements under Basle and under Brazilian regulations. The replacement cost and the potential future exposure of the contracts (without guarantees) provide the basis for the capital charge to cover counterparty credit risk. Data for September and October 1997 suggest that the capital charge for counterparty credit risks in derivatives is about four times as large as required under the Basle capital accord, and that foreign banks’ credit risks measured in this way are about four times larger than for domestic banks.

177. The central bank relies on the banks to set internal limits on foreign exchange risk and does not require banks to meet formal end-of-day limits on their foreign currency exposure or to set aside capital to cover such risks.117 The verification of the adequacy of the banks’ internal limits is part of the on-site examination process. There are a number of regulations on bank foreign exchange activities, which by and large do not prevent banks from taking or holding speculative positions against the domestic currency.

  • the net of all foreign exchange operations involving the actual conversion of domestic currency into foreign currency or vice-versa cannot result in a position of more than US$5 million at the end of the day (long position) or a maximum of R$7.5 million (short position for large banks) without incurring penalties. These limits are very small compared to the capital of the majority of the banks. During normal times, the effect of these regulations is to create a high turnover on the foreign exchange market during the day and to ensure that the foreign exchange liquid reserves of the banking system are on the books of the central bank at the end of the day. Banks are allowed to continuously adjust their spot exposure by holding or taking positions in excess of the above limits, but they must deposit the excess at the central bank at a zero remuneration.

  • off-balance sheet derivatives operations that give rise to foreign exchange exposure are subjected to a capital charge for credit risk that is stricter than Basle, as indicated in the above discussion of derivative risks. As concerns market risk, Brazilian banks are very active in providing forward cover (either through exchange-traded contracts or over the counter) to their clients that wish to hedge foreign currency exposures that arise from foreign-currency denominated loans given by Brazilian banks or from foreign-currency denominated notes issued on the international capital market. To hedge their own resulting short dollar position, Brazilian banks sell offsetting contracts in the market or purchase securities with foreign exchange guarantees issued by the government (NTN-D series) or by the central bank (NBC-E series). The extent of the demand for hedging operations on the part of the Brazilian corporate sector has grown sharply following heightened pressures on the exchange rate in July 1997 during the crisis in Thailand. Outstanding federal securities with foreign exchange guarantee increased in value from R$1.5 billion in July 1997 to R$38 billion in December 1997, a figure which may be compared to an estimate of foreign-currency denominated lending by Brazilian private domestic banks of R$15 billion in June 1997.118

  • Off-balance sheet guarantees denominated in foreign exchange, regardless of whether they have been called or are likely to be recoverable, and regardless of the counterparty, are taken into account at 100 percent in the calculation of risk-weighted assets that are used to assess capital adequacy.

178. Information disclosed by the large banks in their investment prospectuses, as well as the inspections of small- and medium-sized private domestic banks that the central bank had completed by October 1997 suggest that the extent of currency mismatch in Brazil is minimal during normal times.

65

Prepared by Gabriel Sensenbrenner.

66

The assets of Brazil’s five largest banks were US$260 billion, compared to US$230 billion for Korea, US$220 billion for Taiwan Province of China, US$140 billion for Thailand, US$100 billion for Malaysia, and US$100 billion for Mexico (data as of end-1996). The unweighted capital asset ratios were as follows: 8 percent for Brazil (12 percent for the five largest private banks); 5 percent for Korea; 4 percent for Taiwan Province of China; 7 percent for Thailand; 6 percent for Malaysia; and 3 percent for Mexico.

67

Staff estimates exclude the assets of investment banks and most categories of nonbank financial institutions, for which comprehensive data were not available.

68

Nine types of these investment funds currently exist. The 60-day fixed income funds represent about 70 percent of assets (R$110 billion) and equity funds represent a further 10 percent (R$18 billion). Most of the assets invested in equity funds are invested in so-called “carteira livre”, which are aggressive funds for which the managers are explicitly allowed to take substantial exposure to derivatives.

69

Float income of the private banking system for the year ending in June 1994 was R$4.7 billion, and only R$0.1 billion for the year ending in June 1995.

70

The size of the state banking system is somewhat under reported in the current data, because of the special treatment given to Banespa data until December 1997. Indeed, the most recent data communicated by the central bank on deposit shares indicate a somewhat opposite evolution. Private domestic banks have lost 5 percentage points of market share to foreign banks, federal banks, and state banks, reflecting the effect of the larger branch network of the federal and state banks during the recent period of remonetization of the economy.

71

In the second half of 1997 alone, the central bank granted banking licenses to another eight foreign banks.

72

A recent survey of IBGE indicates that 53 percent of corporate revenue is generated by 5 percent of diversified companies operating mainly in the mining, oil, construction, and manufacturing sectors. One may assume that these are typically prime borrowers.

73

These rates apply to the book value of loans as of the initial contracts. When a loan becomes delinquent, Brazilian private banks impose stiff penalties and late charges which, because they snowball, quickly force borrowers to negotiate, and are a means to avoid tortuous litigations.

74

The level of loan sectorization in Brazil prevents an assessment of the risks inherent in sectoral loan concentration. Banks report sectorization according to seven categories: public sector, industry, commerce, other services, households, agriculture, other. Information on geographic concentration is also not readily available.

75

It should be noted that the differences in capital adequacy between groups A and B, on one hand, and groups C and D on the other hand that are shown in Table 12 tend to be underestimated by the fact that, as noted above, some banks in difficulties (which would have been classified in groups A or B) were excluded from the sample.

76

Earnings of some of the leading banks have been bolstered in 1996 by the realization of government-subsidized mortgage assets (FCVS), at some 40 percent of book value, that these banks had generally written off. These windfall profits are estimated to have amounted to R$1.5 billion, or about 35 percent of the earnings of the largest 50 banks in 1996. In addition to the realization of these hidden reserves, further pressure on bank returns comes from the tightening of the taxation of banking income since 1996. The tax rate on banks is 43 percent compared to 35 percent for nonfinancial entities, and income from affiliates abroad has been taxed at 25 percent since 1996.

77

The reluctance to classify loans as nonperforming, by “restructuring” them, cannot be assessed based on available data. Market participants are of the view that the practice has diminished in importance over the past two years. Evidence from medium-sized banks in groups A and B that have submitted to due diligence of one rating agency indicates that restructured loans represented on average 5 percent of the loan portfolio at the end of 1996.

78

This adjustment is possible because the central bank provided information on the assets and deposits of banks that exited the system before June 1996.

79

Data for December 1996 and June 1997 are not contaminated by sampling problems to the same extent as are data before December 1996. Of the group of medium-sized banks that moved (banks with more than R$1.4 billion in assets), two banks moved from group B to group A; two banks moved from group C to group A; six banks moved from group C to group B; and one bank moved from group D to group B.

80

As a point of comparison, the aggregate capital adequacy ratio of the Mexican banking system (excluding 11 banks under intervention), declined from 10.2 percent in December 1993 to 9.8 percent in December 1994 according to the Comissão Nacional Bancária e de Valores.

81

No adjustment was made for loan recovery in any of the scenarios.

82

This policy was formalized in Medida Provisória No. 1514-2 of October 2, 1996, reissued as 1590-15 on September 24, 1997. These established the mechanisms to effect the reduction of the importance of state banks in the banking system.

83

The other, Credireal with assets of R$8 billion, was privatized in August 1997.

84

Banespa reported assets of R$27 billion when it was taken over by the central bank in December 1994. In January 1998, following the general agreement between the federal government and the state government of São Paulo concerning the restructuring of the latter’s debts, Banespa published its financial statements for 1995, 1996 and 1997. These indicate assets of R$49 billion at end-1995, R$63 billion at end-1996 and R$71 billion at end June 1997. Only about 6 percent of assets are claims on the private sector. As part of the debt restructuring agreement, Banespa will benefit from a swap of claims on the state government for federal negotiable securities worth R$63 billion. This swap will enable Banespa to restore its liquidity; up to October 1997 (latest available data), it had monthly outstanding TBAN financing (the central bank’s last resort facility) of about R$2 billion.

85

Most of the information on Banco do Brasil is taken from the June 9, 1997 investment prospectus for the placement of up to US$1 billion in global medium-term notes that will be listed in Luxembourg.

86

In the operation, the government increased its voting and nonvoting stake from 29 percent to 73 percent. Other shareholders (Banco do Brasil’s pension fund, BNDES, and private entities) subscribed the balance.

87

The overall level of nonperforming loans disclosed in the investment prospectus was 14 percent at end 1996.

88

Loan recovery has so far been minimal, less than 4 percent of loans in default, reflecting difficulties in collecting loans from the larger rural borrowers, since most small loans have been restructured in the context of the securitization program.

89

For a full description of FGTS, see SM/97/44. FGTS is Brazil’s unemployment and disability insurance fund. The fund is managed by CEF on behalf of FGTS’ board and therefore its assets are not on the books of CEF. The investment policy of FGTS is to match the cash flow from its placements with expected redemptions. Long-term placements take the form of loans for urban renovation and for moderate income housing.

90

The part of PROER assets which is nonperforming (83 percent) is being funded by a long-term line of credit from the central bank of R$5.7 billion.

91

For a full description of FCVS, see SM/97/44. FCVS subsidies are at a floating rate and are a function of the difference between the lenders’ actual cost of funds and the repayment capacity of the borrowers indexed on their income. Although the program instituted in 1967 was discontinued in 1990, the long tenor of mortgages (15 years) creates a potentially large future liability for the federal government. So far, FCVS obligations that have been verified (because the underlying mortgages have matured) amount to R$18 billion, of which CEF holds R$1 1 billion. The government’s contingent liability could be as high as R$3 billion per year until 2008. However, the government has taken steps to reduce this exposure by: (i) offering incentives to homeowners to prepay FCVS mortgages; (ii) auditing FCVS claims to ensure compliance with requirements; (iii) acquiring FCVS claims at market value; (iv) swapping FCVS claims for negotiable federal securities, thereby reducing the interest cost to the government; and (v) allowing the use of FCVS claims at face value as privatization money. CEF is in charge of implementing (i), (ii), and (iii) and its liquidity stands to benefit from (iv).

92

Staff estimates based on data provided by CEF.

93

Information on valuation and disclosure practices is based on investment prospectuses that have been prepared by the banks that issue notes on the international capital market.

94

The secondary market for certain securities that were issued by state governments has not been as liquid as the market for federal securities, raising concerns as to the proper valuation of such holdings on bank books. Such concerns should be addressed in the context of the recent agreements between the state and the federal governments as these ensure that such securities have the same rating as federal securities.

95

Comparative information on loan classification and provisioning requirements for the major Latin American countries, including Brazil, is contained in International Capital Markets, SM/97/190, pp. 206-207. The information shows that Brazil is the only country in the sample with a loan classification system that is still strictly based on payment incidents.

96

Overprovisioning is largely explained by the tax regime that applies to specific loan loss provisions. Because banks are not allowed to fully deduct such provisions from earnings for tax purposes, overprovisioning is a way of smoothing earnings across periods, thereby minimizing tax outlays, in some cases considerably.

97

A loan classification system with typical provisioning ranges would be as follows: normal (general loan loss provision); watch (5 to 10 percent provisions); substandard (25 to 40 percent); doubtful (50 to 80 percent); loss (100 percent). Classification is generally determined by a determination on the borrower’s probability of repaying the loan, augmented by data on payment arrears.

98

In one well-publicized episode of bank failure that involved large-scale fraud, a number of third parties, including the auditing firms, the rating agencies, a prominent U.S. investment bank, and the leading Brazilian banks, were each unable to detect the extent of the fraud.

99

Medida Provisória No. 1334, March 13, 1996 and Resolution No. 2267 of the National Monetary Council, March 29, 1996, incorporated as Articles 3 and 9 of Law No. 9447/97, March 14, 1997.

100

The central bank participated in the elaboration of the Core Principles of the Basle Committee on Banking Supervision. It is also participating in the survey of compliance with the principles that the Basle Committee has initiated in January 1998.

101

The regime of temporary special administration (administração especial temporária) was created in 1987, following a banking crisis. The central bank may also use two other regimes, which were created in 1974: the regime of liquidation (liquidação extrajudicial), in which extrajudicial proceedings are used to liquidate banking assets immediately following the nomination of a liquidator; or the regime of formal intervention (intervenção), in which the bank is closed for restructuring by a central bank appointed administrator. Formal intervention can be used (and has been used in the cases of Nacional and Bamerindus) as an effective means of changing bank ownership, if the central bank administrator is replaced by a new management team mandated by new owners immediately following formal central bank intervention. However, the central bank does not consider it a credible enforcement tool, as long as new owners have not been identified, because protracted closure runs the risk of destroying much of a bank’s franchise value.

102

In the case of Banespa which was intervened in December 1994, the state governor of São Paulo nevertheless blocked through legal action the publication of financial statements until December 1997. The appointment of a new management of the bank by the central bank was, however, not effectively challenged.

103

PROER was created by Medida Provisória No. 1179, November 3, 1995, as a policy framework for the resolution of private banks. The legal means for implementing PROER were enacted as Medida Provisória No. 1182, November 17, 1995. No. 1182 has since been incorporated into Law No. 9447, of March 15, 1997. This law, in conjunction with the law of the central bank (1964), the law that established the regimes of extrajudicial liquidation and of intervention (1974), and the law that established the regime of temporary special administration (1987), may be used to assess Brazil’s degree of compliance with Principles No. 1 and 22 of the Basle Core Principles, dealing with the preconditions for effective banking supervision and formal powers of supervisors. Other Brazilian legislation and regulations address the areas covered by the other Core Principles. However, the staff has not had the opportunity to conduct a thorough assessment of Brazil’s compliance with the Core Principles.

104

In essence, the financial assistance that the central bank provides to the new owners of the good bank part (in the form of a liability on its books) is offset by an asset representing a claim on the bad bank part. By law, this claim must be collateralized, and this is achieved by discounting enough remaining (good) assets that are left in the bad bank as the financing deal is being structured.

105

In November 1995, under Resolution No. 2212, the central bank also removed the requirement that the minimum capital of new foreign banks be double that of new domestic banks; instituted a minimum capital adequacy ratio of 32 percent for new banks; required that individuals or corporations filing to invest into banks have a minimum net worth of 220 percent of their intended investments; and further strengthened the licensing process by evaluating the net worth of the ultimate shareholders of new banks, instead of the net worth of proxy shareholders.

106

Resolution No. 2099, August 17, 1994, as amended until December 1997.

107

Resolution No. 2399, June 25, 1997; and Circular No. 2784, November 27, 1997.

108

According to Annex 1 of Resolution No. 2099, universal banks are prohibited from issuing debentures, which would typically be counted in tier 2 supervisory capital under Basle.

109

This order of magnitude is taken from information disclosures of a sample of leading banks.

110

Circular No. 2583, June 21, 1995.

111

Resolution No. 2302, July 25, 1996.

112

Resolution No. 2390, May 22, 1997.

113

Almost all of Brazil’s 3,000 mutual investment funds are managed and marketed by commercial and investment banks, and these activities are supervised by the central bank and the securities and exchange commission. On November 27,1997, the national monetary council tightened the regulations of these funds. By Resolution No. 2451, banks have until March 31, 1998 to sell control of their asset management units. Furthermore, upper limits have been introduced on the value of the performance bonds and margin payments that equity funds and fixed-income funds of foreign investors post to exchanges as guarantees for the settlement of obligations arising from derivatives positions. For the most aggressive funds (defined as those with notional outstanding contracts in excess of three times net asset value), such values cannot exceed 50 percent of the net asset value of the fund, compared to 15 percent before (Circular No. 2785). Finally, fund managers have to inform investors that the funds are not covered by any deposit insurance guarantee (Circular No. 2786).

114

See International Capital Markets Report—Background Material (SM/97/190), pp. 120–26, for the results of the BIS survey and other data on world derivatives markets.

115

The main sources of market risks in banking are interest rate risk (resulting from the effects on earnings of maturity mismatch) and foreign exchange risk (resulting from the effects on earnings of currency mismatch).

116

See Basle Committee: Amendments to Annex 3 of the Capital Accord (concerning certain off-balance sheet items), April 1995; Amendment to the Capital Accord to Incorporate Market Risks, January 1996.

117

By end-1997, under the amendment issued in January 1996 to the Basle capital accord of 1988 to incorporate market risks, the internationally active banks of the G-10 countries have to provide capital (including a new category of tier 3 capital) to cover market risks, in particular the risk of losses arising from movements in exchange rates.

118

The level of lending denominated in foreign exchange was not separately reported in the financial statements provided by the central bank. An indirect estimate based on the available data on assets and liabilities explicitly identified as foreign exchange denominated in the financial statements indicates that about R$15 billion or 21 percent of lending (combining all maturities) is denominated in foreign exchange for the banks in the sample. Other banks, especially foreign-owned banks, also extend foreign-currency denominated loans. Foreign exchange funding of private domestic banks were in the order of 19 percent of total liabilities in June 1997, and on the order of 50 percent for foreign-owned banks.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Brazil: Recent Economic Developments
Author: International Monetary Fund
Volume/Issue:
Volume 1998: Issue 024
Publisher:
International Monetary Fund
ISBN:
9781451805888
ISSN:
1934-7685
Pages:
219
DOI:
https://doi.org/10.5089/9781451805888.002
  • View in gallery

    Brazil: Private Domestic Banks: Total Assets and Capital Adequacy Ratio, December 1996-June 1997

    (For the 20 largest private domestic banks)

Brazil: Private Domestic Banks: Total Assets and Capital Adequacy Ratio, December 1996-June 1997

(For the 20 largest private domestic banks)