Banking on the Periphery: Brazil, 1906-1930

Gail D. Triner, Rutgers University

[PLEASE NOTE THAT EQUATIONS MAY NOT BE REPRODUCED ACCURATELY IN THIS VERSION OF THE SUMMARY--THE CLIOMETRIC SOCIETY]

In the interest of brevity, historiographic and citation footnotes are not included. The author will be pleased to supply the complete version of the paper

A substantial body of economic historiography examines and debates the importance of financial institutions in shaping early growth and development. Consensus exists that financial systems play an important role in determining the course of development, although existing work finds nearly endless variety in its specific effects and ambiguous directions of causality. The research from which this consensus is drawn is based almost entirely on the experience of the economies which are currently "developed". We know very little about the historic relationship between finance and development in economies which have not achieved self-sustained growth and industrialization. This lacuna is especially curious given the perceived importance and widespread assumption of the failure of financial systems in contemporary development economics.

The Brazilian experience offers a fruitful opportunity to begin to redress this gap. Recurring financial crises, the existence of a large modern industrial sector and the significant presence of the State in industrial production combine to make the question of historical financial development in Brazil especially interesting. To the extent that it has been considered, Brazilian historiography suggests that during the early years of the 20th century, when agricultural exports were strong and industrial processes became established, the banking system failed to support financial intermediation and economic development within the private sector. These findings stem from analyses of monetary policy and the political economy.

This paper argues that, from the time of its reform in 1906 until the end of the First Republic in 1930, a dynamic banking system in the private sector successfully contributed to development and growth. It finds that banks advanced the reallocation of financial resources from traditional to emerging sectors. Further, they eased commercial flows by significantly improving financial efficiency. Moving away from the usual focus on the public finance and the monetary sector, the paper assesses the role of banking within the private productive sectors of the economy. It explores both the relative importance for banking of growth in the agricultural and industrial sectors, and the responsiveness of different types of banks to the private economy.

Applying classic concepts, articulated by Gerschenkron, Cameron and others, which have been found to be important to understanding banking in other economies, this paper examines the role of banking in Brazilian economic change during the period of early industrialization. Its methodology adds an empirical dimension to the existing historiography of Brazilian banking. The individual balance sheets of the largest and most important banks comprise the major body of data. After adjusting for mergers, new institutions and interstate chartering, the data base includes 21 banks, accounting for 70% to 85% of total banking balances. A simple estimation of the demand and supply of real bank deposits examines the relationship between the demand for banking and economic growth.

During the Brazilian First Republic, from 1889 to 1930, commercial agriculture was the focus of the economy. Coffee and, for a brief intense period, rubber production propelled a dynamic export sector. These commodities received government subsidies when international prices declined. The first formal coffee price support program began in 1906. The Federal treasury or the governments of the major coffee producing states provided various forms of support for most of the remainder of the period. At the same time, industrial processes were firmly established in manufacturing, and a consolidated national economy began to emerge. International markets provided capital to the public sector and to large-scale infrastructure (which often had public sector ownership or subsidies). Manufacturing and industrialization received little direct support from the public sector. Trade tariffs were the main source of government revenue, and indirectly encouraged domestic manufacture. More importantly, growing internal demand and capital accumulation sustained industrial development. By the end of the First Republic, industrial output was still relatively low and represented a small share of the total economy. Nevertheless, the rates of growth of the two sectors differed significantly, and a perceptible shift occurred in the composition of output away from agriculture and toward industrial manufacturing. Although it has not been determined when the process of industrialization became irreversible, there is broad consensus that this was certainly the case by the end of the First Republic, and probably earlier than that.

In 1888, slavery was abolished in Brazil. The economic restructuring that occurred with abolition caused massive disruption of production, demographics, and finance. An uncontrolled expansion in the money supply and widespread formation of incorporated enterprises generated an extremely rapid run-up of money supply and stock markets. After this brief spurt, known as the Encilhamento, markets for money and all financial securities collapsed in 1891. A long period of stagnation, debt rescheduling and recovery followed this frenetic experience. A stronger financial system did not emerge until the banking reform at the end of 1905. The re-chartering of the Banco do Brasil was the predominant feature of this reform. The Federal government's decision to re-introduce the gold standard in order to maintain the value of the local currency (the mil-rŽis) while supporting coffee prices, determined the timing of the reforms. The Treasury instituted and financed the Conversion Office (in current terminology, a currency board) within the newly re-constituted Banco do Brasil to manage the foreign exchange operations that would maintain both the value and convertibility of the mil-rŽis. Simultaneously, the bank was the largest commercial bank serving the private sector.

Their experiences during and after the Encilhamento rendered banks very conservative. After its reorganization, a commercial banking system emerged. Banks took (primarily demand) deposits and they financed short-term commercial transactions. Their charters did not allow them to acquire equity shares or long-term commercial bonds for their own investment portfolios. Banks extended short term credit which was protected by easily identified collateral. Only a few banks extended long term mortgages in order to finance real estate, and state governments subsidized these institutions. With the exception of the Banco do Brasil, banks engaged exclusively in local business. These limitations did not result in insurmountable constraints to their capabilities. Banks easily renewed credit upon expiration, effectively transforming short-term credit into medium- and long- term capital. And, borrowing based on personal creditworthiness and social relations easily funded many business undertakings.

Banks exhibited relative stability during this period. At least as important, the business community perceived banks to be increasingly stable institutions. In 1930 the banking system was about 9 times larger than it had been in 1906 (after adjusting for price inflation), while the total economy and industrial production increased about 2 1/2 and 3 1/2 times, respectively. Although limited in the scope of their activities, the banking system grew rapidly after 1906. Banks increasingly accumulated and re-allocated financial resources, at the expense of either personal or other institutional channels.

The banking system was the conduit for changes in monetary policy; and it was the linchpin between money, public finance and the private productive sectors. Monetary policy fluctuated rapidly and severely in response to the government's financial needs and political goals. International debt obligations, the coffee commerce, and domestic agricultural and industrial production were of primary concern to the government at various times during the First Republic. Monetary policy, more than tax or budget policies, was used to intermediate these competing concerns. The Banco do Brasil was the institution which the federal Treasury used to implement monetary policy and to distribute financial resources.

Three sub-periods usefully categorize economic fluctuation and changes in monetary policy. From 1906 to 1914 the federal treasury attempted to manage the economy within the confines of the gold standard. Money supply, and therefore credit, responded to the determinations and discipline of the foreign exchange markets. Although breaches to the standards of the Conversion Office were routine, credit conditions were restricted during these years. From 1914 through 1923, the Treasury effectively abandoned the gold standard and monetary policy was expansive. Both economic growth and price inflation resulted. A Rediscount Office opened within the Banco do Brasil in 1921, and it orchestrated large scale monetary expansion through 1923. This expansion financed coffee price supports and federal government expenditure and it fueled inflation more than it supported real economic growth. Monetary orthodoxy became the rule once again at the end of 1923. From 1924 to 1926, the Banco do Brasil, acting as a central bank and with the Rediscount Office closed, effected the contraction of money and credit. The money base and prices declined and the level of production remained static. From 1927 to 1930, with all monetary responsibility transferred to the Treasury, the gold standard was re-introduced, coinciding with a strong industrial recession. Whether or not these monetary policies were successful, they defined the environment in which banks operated.

Despite the policy reversals, banks steadily increased their importance to the private sector. The growing share of the money supply in the form of bank deposits was one of the strongest indicators of the strengthening role of the banking system after its reorganization. At the end of 1906, bank deposits comprised 25% of the money supply, increasing to 62% in 1930. The results were larger credit capacity for banks and enhanced efficiency in conducting commercial transactions. The willingness to place deposits with banks, rather than holding currency as a store of resources, implied strengthening confidence in and stability of banking institutions. The most consistent indicator of increased bank stability through the period was the declining ratio of cash relative to the level of deposits (the reserve ratio). The reserve ratio declined from 50% in 1906 to 24% in 1930, and a successively smaller increase in the reserve ratio accompanied each contraction.

The methods of monetary economics study the volume and composition of money. Analyses demonstrate the increasing stability and importance of bank deposits in determining the behavior of the money supply in Brazil, both through the level of deposits and the steadily declining reserve ratio. However, money supply analysis assumes that the factors explaining the demand for money are stable. Substantive changes in the Brazilian economy during First Republic suggest underlying shifts in both the demand for and composition of money. An evaluation of the banking system solely in relation to monetary factors is insufficient. Further examination of the relationship between banking and the productive sector can be accomplished by isolating bank deposits from total money, and focusing on the components of their demand.

Simple supply and demand functions for changes in the real volume of bank deposits can specify the relationship between banking and the real economy (following Sushka, 1976). The bases for this model are the elementary principles of economic theory that supply and demand are (different) functions of quantity and price, and that supply of real deposits equals their demand:

.

. .

D = a1 + b1Y + b2P b1 > 0; b2 < 0 (Equation 1)

.

. .

S = a2 + b3C + b4P b3 > 0; b4 > 0 (Equation 2)

. .

D = S

. where:

D = demand for real bank deposits, annual change

.

S = supply of real bank deposits, annual change

.

Y = output, annual change

.

P = prices, annual change

.

C = real cash balances, annual change

The model estimates the change in deposit volume. Demand is expected to be an increasing function of the change of income (the real level of production) and a decreasing function of the rate of price fluctuation. Supply of deposits is hypothesized to be a positive function of both real cash balances and prices. The simultaneous equations are estimated by two stage least squares procedures, with price level changes determined endogenously in the first stage. The demand side of the model addresses the issues of the relationship between the productive economy and banking.

The nature of the data base allows this model to be applied in a variety of ways. Because the data are the balance sheets of individual banks, rather than the more usual consolidated banking system, balances can be aggregated by various specific characteristics. For the purposes here, regression estimates for privately owned domestic banks, foreign- owned banks and the Banco do Brasil identify their differing responses to economic change. Using income and price series for the industrial and agricultural sectors (from 1908), the model distinguishes sectoral impact. Further, there are sufficient observations to estimate the model for shorter sub-periods within the 24 years. In the interest of brevity, the regressions reported here estimate this model for all of the bank groups for the full 24 year, and the privately owned domestic banks (which were most responsive to the variables of this model) for the three important periods of economic and monetary fluctuation: 1907-13, 1914-23 and 1924-30.

The major limitation of this model is its simplicity. The model incorporates only three independent variables: income, prices and cash balances. Further, in an economy with price fluctuations as rapid and severe as during the First Republic, both experience and theory support the expectation that prices and nominal interest rates moved together. Therefore, the price indices capture a proxy measure for the change in the opportunity cost of holding deposits. Nevertheless, the absence of a time series for domestic interest rates is unfortunate. Another limitation is the assumption that the direction of causation is from the economy to the banking system. This model specifies that growth of banking as a result of change in the productive sectors. It does not take into account the possibility of two- way causation. Granger-Sims tests support the expectation of two-way causation. Notwithstanding these concerns, this model provides a test of the relationship between banking and the economy which is consistent with the available data.

The regression estimates (Table 1) confirm that the demand for real bank deposits was negatively related to prices and positively correlated with production. Changes in demand were more strongly correlated with industrial production than it was with total output. The correlation between demand for banking and agricultural output was weak; and, when it was statistically significant, it was negative in direction. Reflecting its role as the monetary authority, the correlation between the economic variables and the fluctuation of real deposit volume at the Banco do Brasil tended to be the inverse of the results for the private banks and for the banking system overall. The results were stronger, in terms of the strength of relationship (correlation coefficient) and explanatory value (R2) for the private domestic banks than for others.

The growth of private domestic banks demonstrated the strongest relationship both total and industrial output. For the privately owned domestic banks, this simple model estimates a fairly high proportion of the fluctuation in real bank deposits. Consistently, the demand function explains a higher portion of the change in the deposit base of the private domestic banks than any of the other groups identified. The independent variables explain approximately half of the change in private bank real deposits during each of the time periods. The change in the regression results among the sub-periods reflects the shifting role of banking in industrialization. The response of deposits to output fluctuation (correlation coefficient) moderated from 1914-23 to the last period, 1924-30.

These findings hold more strongly when the model is estimated with the industrial production index than the agricultural index. For all of the groupings of banks, and in all time periods (with the exception of 1907-1930, when monetary policy responded to the requirements of agricultural price supports), the absolute value of the correlation coefficients for industrial production were much higher than for agriculture. The importance of industrial production in the explaining fluctuations of bank deposits provides the most succinct demonstration of the empirical relationship between banking and industrialization.

Given the constraints of bank participation in industrial development, what were the mechanisms that generated the apparently paradoxical result of the positive relationship between private sector banking and industrial growth? The available evidence allows an exploration of some of the means by which the banking system encouraged and supported growth. The influence of banking on industrial growth was indirect.

In addition to using the banking system to implement monetary policy, the federal government used it to improve the distribution of financial resources to areas and activities which had previously suffered from their absence. The Federal Treasury successfully expanded the distribution of financial services after the banking system's reform, and it established financial services in some of the areas which had not had access to them previously. Banking services slowly expanded into new activities and geographic areas. Although they provided the basis for consolidating a national banking system from local markets, these policy-directed efforts were not instrumental in determining the evolution of the banking business at this time.

The increasingly effective intermediary function of banks while conducting their own business activities was more important than distribution policies in supporting economic change. Banks continuously renewed short-term credit by discounting commercial paper, and they ensured continuing finance of agricultural commodities (especially coffee) through cyclical fluctuations. Early manufacturing concerns occurred on a small scale and self-financed. As a result of increasingly secure supplies of credit, resources that otherwise would have been engaged in short-term commercial transactions could be securely allocated to capital formation in innovating uses. The improved financial system encouraged economic growth by meeting commercial demand and supporting capital market transactions with improved security and efficiency. Not all impediments to dynamic banking evaporated during the early years of the 20th century. Structural constraints which prevented more aggressive and innovative practices included the slow evolution of property rights and law, and probably, the low level of national income with along with the concentration of its distribution. Nevertheless, the banking system provided passive accommodation to support transition and growth which had been thwarted by its absence during other periods both prior and subsequent to the First Republic.

Brazilian economic historiography, focusing on monetary policy and political economy, assumes a moribund and small banking sector. It may have been small but it clearly was not moribund; banks actively responded to private sector demands. These institutions were more than tools of the federal monetary system. The strong relationship between industry and the (mostly) small scale commercial banks implied a capacity to flexibly allocate resources between sectors. The benefits of economic growth and prosperity in maturing sectors (especially coffee) were transmitted to evolving industrial sectors. Financial institutions actively participated both by intermediating financial resources and by providing financial security. Extreme swings of monetary policy obscured the dynamism of the domestic banking sector.

The evolution of banking practices at this time demonstrated many characteristics in common with other early industrializing economies. The results in the Brazilian case demonstrate the fragility of financial development. Integrated industrial and financial development were not self-sustaining or continuous. During the mid and late 20th century, development and growth have been sporadic. Political exigency, monetary policy and economic cycles re-asserted their strength and influence over financial activities. The banking sector's ability to solidify its integration with and commitment to the private productive sectors has been constrained.

References Available from the Author on Request