1. Why is Brazil different and how?
2. What effect did the inflation have on people's behavior?
3. Why was there little or no lending at any interest rate fixed for more than one day?
4. What was wrong with Brazil's capital markets?
5. How were banks performing at that time?
6. Why had the previous efforts failed?
7. What in fact was the Real Plan?
8. Was this plan any more likely to succeed than the previous plans?
9. What effect would the Real Plan have on the economy and on financial markets?
10. What were the debt and the foreign exchange markets suggesting?
11. Evaluate Banco Itau's strategy in the wake of increasing competition?

Banco Itau

ABSTRACT

In 1994 Roberto Egydio Setubal, Banco Itau's president and CEO, was charged with the task of overseeing the adoption of the newly implemented Real Plan, a strategy designed to reduce Brazil's growing inflation problem. This case discusses how Setubal prepared Banco Itau for this national economic transition.

Introduction

Banco Itau's president and CEO, Roberto Egydio Setubal, sat in his Sao Paulo office in early September 1994, fielding phone calls from his top management team as he helped orchestrate one of his most important tasks to date: responding to the implementation of the Real Plan, a stabilization program aimed at halting Brazil's skyrocketing inflation. If everything went as planned, hyperinflationwhich Brazil had endured for more than a decadecould be substantially reduced, leaving the financial community to operate in entirely new circumstances. Setubal's first task had been to make sure that Banco Itau's more than 1,000 branches had the proper amount of the new Brazilian currency, the real (plural: reais), which had gone into circulation on July 1, 1994. To convert to the new currency, banks had to divide the old currency, the cruzeiro, by 2,750 and then perform the tedious chore of converting all accounts and currency holdings. But the biggest challenge would be implementing a strategy that would make money in this new economic environment. From his position overseeing Brazil's second-largest private bank, Setubal had watched many changes in Brazil's economic environment, including the implementation of five other economic stability plans over the previous eight years. All had failed, and it was not yet clear whether the new Real Plan would succeed. Banco Itau was therefore treading carefully in this uncharted territory, preparing for both the best-case and the worst-case scenarios. While all of the other plans had been implemented with an element of surprise, the Real Plan was the first plan to be fully disclosed prior to its adoption; this had given banks about three months to prepare for the transition. First and foremost, the plan's goal was to end Brazil's spiraling inflation, which had reached an astounding 2,500% for 1993. Most banks, including Banco Itau, generated most of their revenue purely from floattaking in deposited checks, collecting them quickly, earning the overnight interest rate (which could be as high as 2.5% per day) for a day or two, and then making the cash available to the customer. The unstable economic environment did not permit much normal bank lending; the government's claim on bank assets crowded out and significantly depressed the private sector, while the government counted on banks buying their bonds to finance the public sector deficit. Setubal had to consider how Banco Itau's strategy and balance sheet would change in the new economic environment. With this in mind, he scheduled a meeting with his top management team to discuss what options Banco Itau should pursue. Brazil in 1994 Brazil is the fifth-largest country in the world and the largest country in Latin America, occupying approximately 3.2 million square miles, or 60% of South America's land mass. In late 1994 Brazil had a population of about 150 million, the sixth-largest in the world. It was subdivided into 26 states, which enjoyed considerable political and economic independence. Brazil's two largest cities, Sao Paulo and Rio de Janeiro, had populations of approximately 16 million and 11 million, respectively. The population was growing at a rate of nearly 2% a year and was expected to reach 170 million by the year 2000 (see Exhibit 1). In 1988 Brazilians voted on a new constitution that established a presidential form of government comprised of executive, legislative, and judicial branches. Executive power rested with the president, who was elected by popular vote every four years. The legislative branch was comprised of 81 senators elected for eight-year terms to the National Congress and 513 deputies elected for four-year terms to the Chamber of Deputies. The judicial branch was headed by the Federal Supreme Court and the Superior Court of Justice. On December 17, 1989, Fernando Collor de Mello, having campaigned on an anticorruption platform, became Brazil's first president elected by popular vote since 1960. However, President Collor himself was later charged with corruption and resigned his post in December 1992 in the midst of impeachment proceedings established by the Congress. Vice President Itamar Franco became acting president until Fernando Henrique Cardoso, finance minister and author of the Real Plan, was elected president in the autumn of 1994.

Brazil's Economy

Over the previous 30 years, Brazil had moved from an agricultural economy to a more diversified industrial economy. In the 1960s Brazil had adopted a development strategy of high growth and import substitution. This strategy produced deepening deficits in the national and state budgets, which were financed in the 1970s by heavy foreign borrowing that international banks made readily available. The public sector's role in the economy also expanded, and significant structural distortions such as subsidies and tax incentives were introduced.

Brazil was a net importer of oil throughout that 30-year period. New and more severe pressures were placed on the economy following the sharp rise in oil prices and international interest rates in 1979. However, Brazil did not slow its growth rate and continued to borrow until 1982, when it followed Mexico and many other developing countries in defaulting on its external debt. During the ensuing global debt crisis, which lasted throughout the 1980s, Brazil and Mexico were the two largest debtor countries, owing more than $100 billion each to hundreds of international banks. Brazil was also the last major country to reach a settlement with the international banks under the Brady Plan.1

After defaulting on its external debt, Brazil needed to finance its continuing budget deficits internally, and did so largely through selling local-currency treasury notes and bonds to the central bank as well as to the commercial banks. But hyperinflation sharply limited the possible terms on government paper. Some treasuries were indexed, paying 6% real interest plus the floating inflation rate. But most treasuries were of very short termof anywhere from one day to several weeksand paid high fixed rates of interest. Even with short maturities, the bonds were often sold with a one-day guaranteed repurchase agreement, effectively shortening the term to one day. At the peak of hyperinflation in 1989-1990, the entire internal national debt had to be rolled over every day. Brazil resisted the growing trend toward privatization and tried to avoid economic contraction by supporting money-losing state enterprises and subsidizing numerous private enterprises. The resulting inflation was managed by indexation: wages, prices, interest, and other payments were formally linked to one of several inflation indices. The institutionalization of inflation through indexation in turn made halting inflation more difficult. High inflation spiraled into hyperinflation by the late 1980s. Brazil's first attempt at stabilization came in 1986 with the Cruzado Plan, which was initiated with the dual intentions of ending high inflation and implementing fiscal reform. The key element was a price freeze that lasted 10 months before failing, after which inflation rose faster than ever. The Cruzado Plan was followed by four other plans. The Bresser and Summer Plans, in 1987 and 1989 respectively, were characterized by price freezes. The Collor Plan I was launched in 1990, followed by the Collor Plan II in 1991. Under the Collor Plan II, the indexation system was revised to include only certain types of contracts as well as financial obligations with a term of more than 90 days. It was no more successful than its predecessors. See Exhibit 2 for a summary of these plans, and Exhibit 3 for a graph of monthly inflation from 1979 to the end of 1994.

The Real Plan

On July 1, 1994, Finance Minister Fernando Henrique Cardoso introduced the Real Plan, the main objective of which was to combat inflation through pegging the Brazilian currency generally to the US dollar while pursuing tight fiscal and monetary policies. The government would push for fiscal reforms in Congress that would, it hoped, end decades of the chronic budget deficits that fueled inflation. The Real Plan included structural reforms, such as revamping the tax and social security systems as well as creating incentives to stimulate private sector participation in government monopolies, such as oil, telecommunications, and infrastructure. To stimulate competition, the government also introduced measures to deregulate certain industries, such as steel and transportation, which had previously been controlled by government enterprises.

The Real Plan was based loosely on Argentina's successful currency stabilization, which featured a currency board with large dollar reserves and a firm one-to-one linkage with the US dollar. But the Real Plan involved no currency board and no system for managing the exchange rate. It was simply announced that the exchange rate would be 1.000 real/$ for an undetermined period of time, and in principal Brazil's accumulation of $38 billion in foreign exchange reserves were available to support this rate. The initial market reaction was strongly positive. The foreign exchange rate rose quickly to 0.915 reais/$ in the first week of July, and the government said that it would try to establish one-to-one parity in 30-40 days. By August 31, the real had strengthened further to 0.889 reais/$. Actual inflation had fallen from nearly 50% per month in June to 4.33% in July, 3.94% in August, and 1.75% in September. But a survey had found that 52% of Brazilians thought that the plan would not lower inflation permanently, while only 35% thought it would.

The Brazilian Banking System

In mid-1994, the Brazilian banking industry could be characterized as very large and profitable. There were over 250 banks operating in Brazil with approximately $400 billion in assets, the largest number in Latin America. The 10 largest banks held approximately two-thirds of the industry's total assets. See Exhibit 4 for a listing of Brazil's largest banks. Federal and state banks, which were created primarily to promote economic and social development in specific regions, accounted for approximately 60% of industry assets in 1994. Two of the three federal banks, Banco do Brasil and Caixa Economica Federal, were by far the two largest banks in the country. Most state banks were used as political instruments: the great majority of their loans were to states that owned them to finance either political projects or the state budget deficit, with little or no regard to ability to repay. Indeed, the difficulty encountered by states in repaying loans to their state banks would soon lead several of those banks to collapse. Among the private banks were 30 commercial banks, 20 investment banks engaged in specialized lending, underwriting, and asset management, and 204 "multiple-service banks" that were authorized to engage in both commercial and investment banking. The two largest private banks were Banco Bradesco and Banco Itau, both multiple-service banks with many branches. As the country's highest decision-making body on monetary, credit, and financial affairs, the National Monetary Council (CMN) oversaw monetary and foreign exchange policies. It also administered the inflation indexation system. The Central Bank of Brazil was the banking sector's main regulatory body. It was responsible for implementing policies directed by the CMN that related to monetary policy and exchange control, as well as for regulation of financial institutions and foreign investments. The Central Bank's president was appointed by the Brazilian president, served an open term, and could be replaced at any time.

In 1988 Brazil's banking laws were liberalized to greatly facilitate entry into the banking business and to enable all domestic banks to engage in a wide variety of businesses, such as investment banking and insurance, leasing, and credit card companies. Most banks took immediate advantage of the new laws and aggressively entered the new businesses. The main beneficiaries of the new laws were the small banks, and dozens of new banks were chartered, often owned by industrial groups. Foreign banks, which had previously been excluded from Brazil's banking system, were now also readily admitted to it. Yet lending remained somewhat riskier than in other countries, partly because the legal system favored borrowers over lenders. In fact, each of the stabilization plans listed in Exhibit 2 entailed changes in the laws governing loan contracts to further benefit borrowers in a number of ways. Furthermore, lending skills had somewhat atrophied, since in the 1980s the only projects funded were those that were favored by the federal government or by state governments and which benefited from their guarantee. Even the large banks made little effort to study various industries or the national or international outlook. On the face of it, the private banking industry seemed enormously successful. Its average return on equity was higher than all other industries in Brazil, and its revenues represented 14.5% of GDP in 1993, compared to 3% in the United States and 4% in Germany. Closer examination, however, revealed that this profitability resulted almost entirely from market transactions in a high-inflation environment, as opposed to traditional deposit and loan activities. The inflation drove interest rates up as high as 2.5% per day. Such interest rates meant that people kept minimum balances in checking accounts, yet could not keep zero balances. Banks had to redeposit 50% of these balances without interest as reserves at the Central Bank. The net amount (deposits less reserves) was invested in overnight interbank loans or government bonds and earned the full nominal interest rate, while the customer suffered the loss in purchasing power. The government paper often matured in 30 days but was frequently subject to one-day repurchase agreements, so that banks earned income with little or no risk. The government tolerated this, and indeed encouraged it, because it created a steady market for government bonds. This game was exceptionally profitable for the banks. Banks earned further income on float. Checks, which were widely accepted in Brazil as a means of exchange, would be deposited daily in banks. The banks could convert these checks into cash either on the same day or on the following day, but would not make funds available to the customer until at least one dayand sometimes even two or three daysafter collection. This float was also subject to 50% reserve requirements, but the remaining balance was again invested in overnight interbank loans or government bonds. The banks' concentration on short-term assets, government bonds, and inflation-induced monetary gains caused private sector loans to fall from 50% of GDP in the late 1970s to 25% by 1993. If the Real Plan succeeded and inflation subsided, the "floating income" (income from demand deposits and checks in collection) would dwindle to a small figure. Indeed, Exhibit 5 shows that this is exactly what happened. Setubal wondered what the new banking environment would be.

Banco Itau

Banco Itau S.A. originated as the Banco Central de Credito, which was founded in 1944 and became Banco Federal Itau in 1964 after the merger with the former Banco Itau. Through a series of mergers, acquisitions, takeovers, and investments during the 1960s and 1970s, the bank continued to grow, becoming Brazil's second-largest private bank in 1974. In 1988, Banco Itau became a multiple- service bank, consolidating its commercial and investment portfolios. It was the financial arm of Itausa-Investimentos Itau S.A., the holding company of a large array of businesses positioned in finance, insurance, processed wood and finishing materials for the building trades, consumer electronics, chemicals, real estate, planning, and engineering. Banco Itau was a public company. Its common and preferred stocks were listed on the Sao Paulo and Rio de Janeiro stock exchanges and traded daily. The bank's executive board consisted of the president, three senior vice presidents, six executive vice presidents, six executive directors, and the general counsel. The administrative council elected members of the executive board annually. Information about the educational degrees of its senior officers is provided in Exhibit 6. The Villela and Setubal families were the main stockholders of the holding company; Setubal's father, Olavo Egydio Setubal, was the holding company's largest shareholder and Banco Itau's chairman of the board, while Setubal himself had become president and CEO in April 1994. Banco Itau operated in commercial banking, investment banking, export and import financing, foreign exchange operations, leasing, consumer credit, real estate financing, fund management, brokerage, and underwriting. Its primary focus was retail banking, with 85% of its 1994 net interest margin originating from this segment. The large corporate market and the middle market represented 7% and 8%, respectively, of net interest margin in 1994. Banco Itau's strategy was to grow through acquisition in Brazil and also to move into other markets. In May 1994 the bank had established Banco Itau Europa in Portugal; it was also planning to open Banco Itau Argentinawhich would be the first Argentinean bank created with Brazilian equity early in 1995. In 1986 Banco Itau an and other banks experienced the effects of the Cruzado Plan. They had concluded that if the Cruzado Plan succeeded, they would need to shift their focus from arbitrage and float- related activities to lending and similar conventional banking services. They began to offer such services in 1986, but found a number of surprises. First, the amount of money that could be made on spread lending was not nearly sufficient to cover the banks' cost structures in the absence of float income. Second, when the Cruzado Plan collapsed, inflation returned with a vengeance, wiping out the monetary value of cruzado loans. Bank Itau and other Brazilian banks that had cooperated with the Cruzado Plan suffered major losses. Setubal considered 1986 an expensive but valuable education. Banco Itau then began looking at its cost structure to determine how it could become more efficient. It examined its branches for profitability and cut the number of its employees dramatically, from nearly 90,000 in 1986 to just over 37,000 in 1994. Even with these reductions, however, Banco Itau's capital-to-employee ratio of $43,000 compared unfavorably to that of international banks; for example, Banco di Roma's capital-to-employee ratio was $290,000, and Sumitomo Bank's was $1 million. Banco Itau invested in technology and created a credit-analysis decision model to analyze creditworthiness, although the bank made only modest use of these tools during hyperinflation. At the same time, the bank continued to invest in the treasury market. The combination of trading, float- related profits, and low operating costs made Banco Itau one of the most profitable banks in the private sector. A Baring Securities report estimated that Banco Itau earned 41% of its 1993 revenues from float and only 28% from credit operations. Banco Itau's balance sheet at June 30, 1994, is shown in Exhibit 7. Its deposits were subject to Central Bank reserve requirements, which were raised when the real was introduced at the end of June, so as to contain credit growth. Demand deposits paid no interest but were subject to non- interest-bearing reserve requirements of 50% until June 30, 1994; after that, the reserve went to 60% for float balances (demand deposits in the process of collection and payment) and to 100% of all other demand deposits in excess of their balance on June 30, 1994. Savings accounts paid a variable rate of interest equal to an inflation index plus 6.17% per annuma high real rate that was attractive to depositers and expensive for banksand were also subject to reserve requirements. Interbank deposits consisted largely of overnight money with no reserve requirement, and Banco Itau's one-day Interbank Certificate of Deposit (CDI) interest rate was a benchmark for many other transactions. Time deposits were issued to individuals and corporate customers, generally for 30-day terms at negotiated rates, and were subject to reserve requirements. Borrowings consisted largely of foreign currency borrowings for trade transactions, and so were matched against foreign currency assets. Strong Brazilian banks could sell dollar deposits to international banks, but had to pay a spread of 1.5%-2.5% above LIBOR to do so. See Exhibit 8 for Banco Itau's consolidated income statement.

Strategic Possibilities

With the implementation of the Real Plan, Banco Itau faced some immediate decisions. It could switch its strategy completely by substantially reducing its treasury holdings and seeking more loans based on the infrastructure it had built. It could also begin to charge fees for certain services that many consumers had taken for granted, such as checking accounts and ATM withdrawals. However, customers were likely to resist such changes. A precipitous drop in inflation sharply lowers interest rates. However, for the Real Plan to succeed, monetary policy had to be extraordinarily tight, especially because fiscal reforms would take time to legislate and might not be feasible at all in the short-term. A severe monetary policy would drive up the real rate of interest, creating interesting opportunities for lenders but likely imposing hardships on borrowers. In this environment, it became important to understand the real rate of interest. In July 1994 a typical 30-day working capital loan commanded an interest rate of up to 6% per month, while the inflation rate from June 20 to July 21 was 4.33% per month.2 Treasuries actually yielded more than this; Banco Itau's holdings of treasuries and the related yields in July and August are shown in Exhibit 9. The terms of these bonds were usually three to six months, but many were sold under one-day repurchase agreements and were therefore classified as overnight bonds. Some idea of inflationary expectations and real interest rates in this transitional period can be obtained from interest rate futures and foreign exchange futures. There was no proper yield curve in Brazil, either for treasuries or for bank deposits. But the Commodities and Futures Exchange (BM&F) traded one- to four-month futures on the overnight CDI interest rate. For example, on August 31, 1994, a set of R$100,000 contracts on CDI futures settled at the prices shown in Exhibit 10. From these, one could estimate where market participants thought interest rates would be during the coming four months. Furthermore, exchange rate futures were also traded on the BM&F. The foreign exchange futures prices, such as those for August 31, 1994 (see Exhibit 10), were based not on interest ratessince an appropriate set of real deposit rates did not existbut rather on true expected future values of the exchange rate, given inflation expectations. These numbers surely contained important clues about inflationary expectations and therefore about real interest rates and the likely future of the Real Plan. Setubal wondered about the market for private borrowing. To be sure, there were many individuals and businesses starved for credit under the old regime, and most of them would be eager to borrow. On the other hand, private borrowers would have great difficulty paying high real rates of interest.3 Furthermore, very tight monetary and fiscal policies would almost certainly depress the economy; as a result, borrowers might face high real rates at a time of recession. Setubal wondered if this was really the time to expand lending. Treasuries could be funded with either domestic real deposits or international dollar deposits. Dollar LIBOR was in the vicinity of 5% per annum (see Exhibit 11), and Banco Itau could borrow from international banks at about 2% above LIBOR. If the bank held real treasury bonds funded with dollar borrowing, it could earn an immense spread. Of course, this would mean taking foreign exchange risk: if the real devalued versus the dollar, the bank would absorb the full loss. But the real was now so strong and the political will to maintain its credibility so clear that the risk might be worth running. Even if the bank decided that foreign exchange risk should not be incurred, the foreign exchange market might still provide an indirect source of real funding that could be superior to local real funding. There was no market in Brazil for (real) deposits of terms of different lengths; unlike the dollar market, the real market did not have an orderly scale of deposit rates for one month, two months, three months, etc. But the existence of a futures market in the real/$ exchange rate made it possible to fund in dollars and then purchase forward the dollars needed to repay the deposits, thereby hedging the foreign exchange risk and providing "synthetic" real funding of various terms. Setubal wondered if this might not provide a creative way of funding treasuries

and he estimated what the bank would earn if the August portfolio of treasuries were funded in this way.