Jack D. Guenther
During the past seven years, inflation in Brazil has averaged over 20 per cent a year, but this has not impeded economic growth in real terms of about 10 per cent a year. This success in “living with inflation” frequently has been attributed—in part, at least—to the system of indexing or “monetary correction” adopted in 1964 which provided for wages, financial obligations, taxes, the exchange rate, and other economic variables to increase pari passu with the general level of prices, thereby minimizing the distortions usually associated with inflation.
However, indexing in Brazil is not as all-embracing as is sometimes suggested and almost nothing is indexed through a simple 100 per cent link to prices. In the case of wages, for example, prices are only one element in the wage formula, which itself has undergone several revisions during the past decade to keep it in line with the Government’s general wage policy. In the case of financial obligations, only about one fifth of the liabilities of the Brazilian Treasury and financial system outstanding at the end of 1973 was indexed—mainly government bonds and mortgage paper—and the procedure for linking these obligations to prices also has been changed several times when the system was not providing results consistent with the interest rate and other policies being pursued by the Government. Finally, in the areas of the exchange rate and charges for public sector services, there is no formal system of indexing; these important prices are adjusted upward frequently, but at the discretion of the authorities after taking account not only of movements in prices but also of other factors. A survey of the system shows, therefore, that prompt discretionary action by the Brazilian authorities has probably been more important than indexing in adapting the economy to inflation. In fact, with the resurgence of inflation to over 30 per cent in 1974, the problems produced in Brazil by indexing became more acute and led to its abandonment even in some of the limited areas where it was previously applied.
Review of past trends
The main features of Brazil’s economic and financial crisis prior to 1964, which led to the introduction of indexing, are well known. Price increases, which had averaged about 20 per cent a year during the 1950s, began to accelerate rapidly in 1959. Strong trade unions negotiated successively larger wage adjustments, credit was granted liberally to accommodate higher costs, and the resulting wage-price spiral pushed inflation to over 80 per cent in 1963. As the Government tried to suppress inflation, the distortions in relative prices also became more severe. Exports stagnated because of delays in adjusting the exchange rate; apartments were left empty as rent adjustments lagged; and electricity, telephone, and other public services deteriorated as the charges for these services were kept artificially low. Interest rates, although nominally high, were negative in real terms, and savers sought protection in real estate or the transfer of funds abroad. Finally, Brazil’s public finances were seriously weakened, because all charges for public sector services and certain minor taxes were specific rather than ad valorem; moreover, taxpayers delayed payments to take advantage of the loss in purchasing power of the cruzeiro.
Faced with this combination of high inflation and serious distortions in relative prices, the new Brazilian Government which took power in April 1964 adopted a medium-term plan to: (1) eliminate inflation gradually over a period of four or five years; and (2) minimize inflationary distortions during the transition period. While the reduction of inflation was slower than expected, the program to minimize the distortions arising from inflation has been more successful: most real interest rates have been positive; the exchange rate has been adjusted regularly to compensate for increases in domestic costs relative to those abroad; and rents, tax revenue, charges for public services, and other key economic variables all have roughly kept pace with inflation. Saving and investment in the economy have been high, output has risen at an unprecedented rate, and the balance of payments has been strong.
Readjustable treasury bonds
Readjustable treasury bonds, which were indexed for inflation in addition to paying a small nominal interest rate, were issued in 1964: these bonds have set the pattern for most other forms of “monetary correction” of financial assets, although such indexing has not spread very widely in the private sector. The 1964 law did not prescribe what index would serve for adjusting the bonds, but in practice the bonds were linked to the wholesale price index with a lag of several months. Initially the treasury bonds were not well received, and most Brazilian savers preferred to continue converting their idle funds into foreign exchange. To meet this problem, the Government offered a new series of bonds in May 1965 which gave purchasers the option of monetary correction based either on the wholesale price index or on the exchange rate with respect to the U. S. dollar, whichever was more favorable. This option could be chosen ex post, but after the crawling peg exchange system was adopted in 1968, the rate of depreciation of the cruzeiro in relation to the U. S. dollar was always less than the rate of increase in wholesale prices, and monetary correction based on wholesale prices was consistently the more favorable option.
One of the most troublesome problems encountered in Brazil’s indexing has been the lag between price increases and the subsequent adjustment of the indexed paper. For various operational reasons, the average lag in adjusting treasury bonds to inflation was about five months until 1972. In 1973, with inflation declining, this lag was considered to be interfering with the efforts to control inflation. The Government, therefore, temporarily eliminated the lag by using “projected” inflation for the current months where statistics were lacking, thus introducing a departure from the practice of linking to actual prices.
In the late 1960s and early 1970s, sales of these indexed treasury bonds to the Brazilian private sector were relatively small, and their role in Brazil’s capital market remained quite limited. At the end of 1973, only about 15 per cent of the outstanding bonds was held by the general public. Most of the bonds were in the hands of the commercial banks (which are permitted to use them to fulfill part of their minimum legal reserve requirements) or the National Housing Bank (which receives indexed deposits and therefore needs indexed liquid assets until the funds are lent for housing). With the acceleration of inflation in 1974, however, indexed bonds for the first time became more attractive to the private saver than nonindexed paper. This caused considerable problems in domestic capital markets, as private savers drew funds out of nonindexed accounts to purchase the bonds. In response, the Government temporarily discontinued the sale of indexed bonds and switched exclusively to treasury bills. These treasury bills, which circulate in Brazil much more widely than the bonds, have an average maturity of six months and have never been indexed on the grounds that linking such a short-term asset to a price index would not be administratively feasible.
“indexing of private sector financial obligations other than those related to housing finance remains rather limited”
Apart from treasury bonds, the only other major financial transactions that have used indexing extensively are those pertaining to housing finance. Since 1964 both assets and liabilities in the housing sector have been linked to either a price or wage index and adjusted monthly, quarterly, or annually in accordance with movements in the index. There have been frequent changes in the system of indexing, however, and, as in the case of treasury bonds, the Government has enjoyed wide freedom to change the index in order to meet the needs of its housing policy or its overall economic policy.
As a result of the system of indexing, the Brazilian mortgage market, which was almost nonexistent in 1963, has grown substantially in the past decade. The system has had a series of problems, however, including at various times a high level of defaults, and this has led to a gradual movement away from full indexing. The most frequent (although incomplete) explanation for these defaults was that the family incomes of some mortgagees did not grow pari passu with the mortgage obligations. The problem became so severe that in 1972 the system was revised to ensure that mortgage payments declined in real terms over time. All obligations continued to be indexed, but the initial service schedule on mortgages, which previously involved equal total payments (amortization plus interest) over the life of the contract, was switched to one of equal amortization payments. Under this system, because real interest payments became smaller over the life of the loan, total monthly payments declined in real terms despite the indexing.
When inflation rose to over 30 per cent in 1974, the Government again revised the system of mortgage financing to reduce the “burden” produced by indexing. Under a new system of tax rebates, the effective adjustment of mortgages in 1975 will be equal to inflation minus 10 percentage points. The rebate is at least Cr$240 per year, and the relief from indexing, therefore, is greater than 10 percentage points on those mortgages in which annual service payments are less than Cr$2,400 (about US$300). Indeed, on many small mortgages, even the nominal service payments will decline in 1975 under the new system. Because the savings instruments that finance housing continue to be fully indexed, the Treasury in effect will have to subsidize the Housing Bank by paying the rebates from the budget.
Other financial assets
Contrary to popular belief, indexing of private sector financial obligations other than those related to housing finance remains rather limited in Brazil. At the end of 1973 only about 19 per cent of the total financial obligations issued by the Brazilian Treasury and the Brazilian financial system was indexed (Table 1). The Brazilian treasury bonds accounted for half of the indexed paper, and the remainder was mortgage bonds and savings deposits, which are used almost exclusively to finance housing. On several occasions during the past decade, the Government, in its efforts to develop a medium- and long-term capital market in Brazil, has tried to encourage the spread of indexing into other parts of the private sector, but the movement appears to have been resisted by borrowers. Financial institutions, of course, are unable to accept indexed deposits unless they can find borrowers willing to contract loans with indexing. Private firms, however, generally have been reluctant to assume such “open-ended” obligations, apparently preferring the certainty of rates fixed in nominal terms, even if these rates are high, rather than rates fixed in real terms where the nominal interest might turn out to greatly exceed their projections.
The erroneous impression that indexing is very extensive in Brazil’s private sector is partly due to the ex ante monetary correction advertised by Brazilian commercial banks on their time deposits. This ex ante correction for inflation merely means that a certain projected rate of inflation presumably has been allowed for when setting the total interest rate. The practice of ex ante monetary correction originated partly to circumvent Brazil’s usury law, which limited interest to 12 per cent; the borrower would agree to pay, for example, 12 per cent interest and a further 15 per cent allowance for inflation. At times the portion corresponding to inflation also received a different tax treatment from the remainder of the interest—on the grounds that “interest” which merely permitted the saver to keep pace with inflation should not be considered taxable income. The fact is, however, that the total return paid on instruments with ex ante monetary correction is agreed in advance, and the instrument is not indexed in the sense that the return varies with the actual behavior of prices.
|Bills of exchange||37||17|
Index of business loans
The National Development Bank is at present the only major source of indexed loans to business. This bank uses mainly its own capital and part of the resources of the social security system to make fully indexed loans. Late in 1974, however, borrowers from the bank became alarmed as their obligations began to be adjusted upward in line with the acceleration of inflation. Consequently, the demand for loans began to lag. The bank authorities were concerned that this might deter investment and announced, therefore, that adjustment of annual service payments on its loans will in future be limited to a maximum of 20 per cent regardless of the rate of inflation. The loan principal will continue to be fully indexed, however, which will require the borrower to make additional service payments beyond the original maturity of the loan.
The coexistence of indexed and non-indexed financial assets has led to considerable problems, particularly when the rate of inflation varies significantly. The recent upsurge in Brazil’s inflation caused indexed treasury bonds to become more attractive than fixed-interest securities for the first time in over five years. During the course of 1974, the total return on indexed treasury bonds rose from 17 per cent (13 per cent monetary correction and 4 per cent interest) to almost 40 per cent, while interest on nonindexed time deposits increased only from 21 per cent to 27 per cent. This shift in favor of indexed paper led to an outflow of funds from many financial institutions, which the Brazilian authorities first tried to offset by fiscal measures reducing the taxation of interest on nonindexed assets. Finally, however, they made the decision to suspend temporarily the sale of treasury bonds to the public between August 1974 and February 1975, when sales were resumed on a limited basis. Also, to partially offset the changed flow of funds, the Central Bank early in 1975 diverted resources from the housing system, which was capturing indexed funds greatly in excess of its needs, to the investment banks, whose nonindexed offerings were unable to attract private savings. The Central Bank reportedly will pay the difference to the Housing Bank between the indexed return required on their investments and the nonindexed rate that will be paid by the investment banks.
After 1964 the new Brazilian Government eliminated strikes and collective wage bargaining and ordered that future wage adjustments in the industrial sector be made in accordance with an official formula. Total employment in the unionized sector, for which the wage formula was designed, was less than 2 million workers out of a work force of 30 million. This sector included, however, the strongly organized groups, and the purpose of using the standard formula for these groups was to ensure relatively uniform wage adjustments rather than adjustments dependent on the relative strength of each union. Other wages in the economy continued to be set by the Government independently of the formula (as in the case of the minimum wage and the wages for national and local government employees), or remained completely free of control (as in the areas of services and agriculture).
“restraint on wages undoubtedly was a major reason for Brazil’s success in reducing inflation rapidly”
The official wage formula used in Brazil since 1965 is quite different from a cost of living escalator clause operating during the life of a wage agreement. Each contract is signed for a 12-month period, and contract renewals for the various industries are spread throughout the year; once the wage is determined there is no link to prices and no adjustment in nominal wages is made until the contract expires. The rise in the cost of living is taken into consideration only when a new annual contract is signed. Furthermore, in the mid-1960s even the new contract did not include full compensation for past inflation. The formula was specifically designed, in fact, to break the previous pattern of fully compensating for past inflation which, together with escalator clauses during the contract period, had been largely responsible for the wage-cost spiral under the previous system of collective bargaining. The formula was based on the recognition that the maintenance of average real wages did not require full compensation for past inflation when inflation was declining.
Specifically, wage increases granted under the 1965 formula reflected three components, which were designed to (1) raise the real wage up to the average of the previous 24 months; (2) provide for the maintenance of this average wage by granting an additional amount equal to one half of the “projected” inflation during the year covered by the wage agreement; and (3) allow a small additional amount for higher productivity. The allowances for future inflation and productivity were announced each year by the Ministry of Planning. Not surprisingly, the estimate of future inflation always was less than the actual inflation turned out to be. For example, in the year beginning August 1966, the projected inflation used in the wage formula was only 10 per cent, whereas the actual increase in the cost of living was 30 per cent. As a result, average real wages declined during 1966 and 1967, despite the professed goal of the formula to maintain them (Table 2). This restraint on wages undoubtedly was a major reason for Brazil’s success in reducing inflation rapidly in these years—from 90 per cent in 1964 to about 25 per cent in 1967—while at the same time increasing investment and strengthening the balance of payments.
|12 months||In per cent|
|1966||41.9||30.0 3||30.0 3|
Refers to previous month; figures for December represent the increase In prices during the 12 months ended in November
Under the official wage formula.
Estimate based on partial data.
Refers to previous month; figures for December represent the increase In prices during the 12 months ended in November
Under the official wage formula.
Estimate based on partial data.
Wage formula revised
Brazil’s wage formula was revised in 1968, and since that date real wages granted under the formula have risen gradually. The major change was the introduction of a new component in the formula to compensate for the excess of actual inflation over projected inflation during the previous contract period: in calculating the real wage base for the coming year, the revised formula used not the actual real wage received in the past year but the intended (higher) real wage that would have been received if inflation had been as projected. Also, the allowance for the growth in productivity was raised gradually from zero to 3 per cent. Because of these 1968 adjustments to the formula, real wages granted under the formula rose by an average of about 2 per cent a year from 1968 through 1973.
Because wage contracts are signed for 12 months, a temporary decline in real wages still can occur with an upsurge of inflation during the life of the contract. To deal with this problem, in December 1974 the authorities departed from the formula and granted a special advance wage boost of 10 per cent to all workers whose contracts had been signed in the period January-June 1974 and whose wage adjustments, therefore, had not yet fully reflected the mid-year upsurge of inflation. At the same time, they further liberalized the formula to provide in future a straightforward increase in real wages of 4 per cent a year. These latest discretionary wage boosts were taken at a time when consumer demand was showing signs of weakness and output in some sectors of manufacturing was declining for the first time in almost ten years.
After 1968 Brazil’s revised wage formula probably tended to perpetuate past price rises without any significant effect in either accelerating or decelerating the rate of inflation. There were occasions during that period, however, when exogenous forces tended to change the rate of inflation. For example, in 1972, relatively good harvests led to a decline in the rate of price increases; and more recently, the sharp increase in world prices of petroleum and other commodities, together with excess domestic demand, produced an acceleration in Brazil’s inflation. Once these price trends were set in motion, either increasing or decreasing the previous rate of inflation, the wage formula tended to reinforce the trend. However, this reinforcement of price trends probably was less pronounced than under the system of collective bargaining and escalator clauses that existed before the 1964 reform. This is because—except for the advance in December 1974—the Brazilian formula adjusts wages to compensate for actual price trends only when the 12-month contract expires, which means that wage payments do not fully reflect new price trends until a full year has elapsed.
Other areas of indexing
Apart from wages, treasury bonds, and mortgages, the other principal areas where some form of indexing is used in Brazil are house rents and certain fiscal transactions.
During the late 1960s and early 1970s, three schemes existed for adjusting house rents, depending on when the contracts were signed: (1) contracts signed between 1964 and 1967 were adjusted each year by two thirds of the percentage increase in the minimum wage; (2) contracts signed prior to 1964 were adjusted by the full increase in the minimum wage plus an additional 10 percentage points; and (3) contracts signed after 1967 were free of controls. By 1974, however, most rents payable on pre-1964 contracts had been brought up to the market level, and almost all rent contracts are now freely negotiated. In practice, however, many of these freely negotiated contracts have clauses which link payments to the readjustable treasury bonds.
In the area of public finances, several types of indexing have been introduced. One of the first acts of the new Government, in July 1964, was to index unpaid taxes and related liabilities. Taxpayers had used many devices to delay paying taxes, including testing the assessments in the courts in order to benefit from inflation while the case was under consideration. The 1964 legislation provided that all tax liabilities not paid in the calendar quarter when they became due would be subject to monetary correction. The post-1964 tax laws also provided for the annual revaluation of fixed assets and working capital; the revaluation of the fixed assets was made mandatory, while the revaluation of working capital was optional. At present, after several changes, the index for readjustable treasury bonds is used for all these purposes.
Indexation has been used also—but rather sporadically—for income tax brackets in Brazil, in order to ensure that the brackets are adjusted upward roughly in line with inflation. This policy antedates the 1964 reform, for already in 1961 the income tax brackets were defined as multiples of the minimum wage and were, therefore, adjusted automatically each time the minimum wage was changed. This link to wages proved unsatisfactory after 1964, when minimum Wages for two or three years were adjusted upward by a lower percentage than the increase in prices, which would have resulted in increased income tax burdens at all levels of real income. This increase in tax rates was not desired by the Government, and in November 1964 the law was modified to provide that the income tax brackets be adjusted in accordance with movements in prices rather than the minimum wage. In 1973 and 1974, the procedure again was modified, and the Government departed further from indexing. In 1973, as part of the policy to improve income distribution, the upper brackets were increased by 15 per cent and the lower brackets by 26 per cent; and again in 1974 the upper brackets were raised by 12 per cent and the lower ones by as much as 41 per cent. Thus, the authorities have exercised a considerable amount of discretion in adjusting the tax brackets in the past decade. Although the brackets have been changed regularly, the size of the increase has been dictated by general policy considerations rather than by rigid adherence to an index.
The system of minidevaluations
There are two other areas—the exchange rate and charges for public services—where regular adjustments also have been made to offset the effect of rising prices, but without any rigid system of indexing.
“one of the most important achievements in Brazil since 1964 has been the strengthening of public sector finances”
In August 1968 the cruzeiro was depreciated by 12 per cent in terms of the U. S. dollar, and the authorities announced that in the future it would be depreciated frequently by small amounts. The criteria used for adjusting the rate would be price changes in Brazil compared with those abroad, the level of international reserves, and the behavior of exports.
Although the cruzeiro has been devalued roughly in line with price differentials in recent years, this has not been done through a formal system of indexing, as is sometimes suggested. The economic authorities have been free to devalue by somewhat more or less than indicated by price differentials if that course seems advisable in the light of overall policy considerations. In December 1971, for example, the cruzeiro was depreciated fully in line with the U. S. dollar; but in February 1973, at the time of the second U. S. dollar devaluation, with Brazil rapidly gaining reserves and struggling to control inflation, the cruzeiro was appreciated by 3 per cent with respect to the U. S. dollar. The Brazilian decisions on these occasions were made after considering all the advantages and disadvantages of following the U. S. devaluation (particularly the effects on the balance of payments and domestic prices) rather than by a predetermined link of the exchange rate to price indices.
Any attempt to rigidly index the exchange rate would encounter formidable difficulties. For example, there would be the problem of selecting which price series to use in Brazil and in other countries, particularly in a world of floating currencies. The Central Bank has stated that the value of the cruzeiro now is being set mainly in relation to a trade-weighted package of currencies rather than in relation to the U. S. dollar alone, but the composition of the basket and the weighting of the currencies in the basket have not been announced.
Although the exchange rate is not formally indexed, the authorities have been able through the system of mini-devaluations at frequent, irregular intervals to convince exporters, borrowers, and foreign lenders of their firm intention to maintain a realistic exchange rate. Producers are prepared to undertake long-term investments in the export sector, with the confidence that the exchange rate will be adjusted roughly in line with relative costs and that export prospects will continue to be attractive when the investment is completed. This new policy toward exports, which traditionally was lacking in the Brazilian economy, undoubtedly contributed to the tripling in the U. S. dollar value of Brazil’s exports between 1968 and 1973 and particularly to the rapid development of manufactured exports.
Charges for public sector goods and services
One of the most important achievements in Brazil since 1964 has been the strengthening of public sector finances, which had seriously deteriorated during the years of high inflation. This improvement was aided by indexing tax arrears to discourage taxpayers from delaying payments. Probably the most important innovation in this area, however, has been the policy of frequently adjusting the charges for public sector goods and services—such as electricity, transportation, water, wheat, and petroleum—in line with rising costs.
After 1964, charges for all public sector services were adjusted upward by amounts greatly exceeding current inflation. It was, in fact, the necessity for a large amount of such corrective inflation that caused the Government to opt for a gradual rather than an abrupt halting of inflation after 1964. By the early 1970s the charges for public services in Brazil were high even by international standards, and the large public sector investment programs being carried out in electricity, transportation, steel, and other areas are now being financed to a substantial extent by user charges.
It is important to point out that while these charges for public goods and services have risen roughly in line with inflation in recent years, this is not a result of any system of indexing, but rather of discretionary action. As in the case of the exchange rate and most interest rates, the adjustment for inflation in this area is accomplished by administrative decision rather than by an automatic link to prices.
Brazil’s experience gives strong support to the view that the removal of distortions caused by inflation is essential if a high rate of economic growth is to be achieved. In particular, where economic development has high priority, those economic variables that are important for encouraging saving, investment, and exports cannot be permitted to lag behind general price increases. The Brazilian case does not, however, provide much evidence that widespread indexing is the most appropriate way to achieve this goal. Rather, it seems to show that discretionary action by economic managers was the main element in Brazil’s success, and that the indexing was highly selective, flexible, and subordinated to general policy considerations.
Finance & Development is now available in an Arabic edition. Readers who wish to receive the Arabic language edition should send their current address label and include their address in Arabic. New readers may also send their requests to:
Finance & Development
International Monetary Fund Building
Washington, D.C. 20431