Banking and Economic Development: Brazil, 1889-1930 (review)

"This book," writes Triner, "finds that the emergence of Brazil's modern banking system during the first three decades of the twentieth century produced a system that worked efficiently, rationally, and productively" (p. 9). Moreover, "the influence of banking was most heavily felt in emerging sectors of the economy, such as industry and urban real estate development" (p. 13). Contrast this view with Steven Topik's: "[T]he banking system remained lamentably small, concentrated, and conservative" (The Political Economy of the Brazilian State, 1889-1930, 1987, p. 52). Topik states further, "[T]he Brazilian banking community was not composed of dynamic entrepreneurs striving to forge industrialization."

Nevertheless, the views of Triner and Topik overlap more than the quotations would suggest. Triner recognizes that the Brazilian banking system in 1930 was still small, just as Topik knows that it grew rapidly from a tiny base after 1906. Even allowing for nuances, however, Triner's appraisal is more positive than most, an assessment she backs up with references to theories of financial intermediation and development, extensive use of primary sources, and incomplete forays into econometric analyses. She brings our knowledge of banking under the Republic a long way, but her journey starts from a low base. Her contribution, though impressive, falls short of her goal, implicit in the book's title, of delineating how banking contributed to Brazil's economic development from 1889 to 1930. Even the wealth of information she contributes to our knowledge of the history of banking does not justify her conclusion that "banking strongly supported" structural changes that boosted manufacturing (p. 117).

Her econometric evidence linking banking and industry comes from a 1907-30 time-series model of the market for deposits, in which the growth of demand depends on inflation and output growth. Finding a negative coefficient on inflation, she concludes that inflationary monetary policy "spurred the construction of manufacturing capacity" because the resulting currency depreciation made imports costly (p. 105). But all the negative coefficient shows is that inflation increased the velocity of money (the speed at which money travels over a given period of time). Exchange rates are not included in the equation, nor is industrial capacity (as opposed to output). It is likely that inflation did cause a temporary real depreciation of national currency that stimulated domestic substitutes for imports, but that same depreciation made machinery, nearly all imported, more costly, which restrained the expansion of capacity. Her deposit demand equation does yield a positive coefficient on the growth of industrial output, but that may merely reflect the elasticity of both deposits and industrial output with respect to urban income. Contradicting the weak time-series evidence is the disparity between São Paulo's rise to industrial dominance (from 1907 to 1919 the state's share of Brazil's industrial output rose from 16 to 31 percent) and its stable share in banking (from 30 to 28 percent) (p. 210). Triner's statement that São Paulo had one-half of total deposits in 1906 (p. 159) is an overestimate based on the improbable implicit assumption that the share of Rio Grande do Sul, 16 percent in 1907, was zero in 1906 ( p. 210). In contrast, greater Rio's share of industrial output fell from 38 to 28 percent, while deposits rose from 26 to 30 percent. And in Rio Grande do Sul, the output share fell from 13 to 11 percent, while deposits rose from 16 to 28 percent.

Just as the positive time-series correlation does not prove that banking boosted industry, a negative correlation across states does not disprove it. Perhaps the banking system channeled funds from agricultural to industrializing areas. Triner says that the banking system encouraged a flow of funds among agriculture, industry, and government. But it may have been that agriculture, through coffee price supports...