Geraldo da Silva e Souza

Evolution of Bank Efficiency in Brazil: A DEA Approach

This paper investigates cost, technical and allocative efficiencies for Brazilian banks in the recent period (2000-2007). We use Data Envelopment Analysis (DEA) to compute efficiency scores. Brazilian banks were found to have low levels of economic (cost) efficiency compared to banks in Europe and in the US. For the period with high macroeconomic volatility (2000-2002) the economic inefficiency in Brazilian banks can be attributed mainly to technical inefficiency rather than allocative inefficiency. State-owned banks are significantly more cost efficient than foreign, private domestic and private with foreign participation. There is no evidence of differences in economic efficiency due to type of activity and bank size. These results may provide some useful guidance for financial regulators and bank managers.

Optimal monetary rules: the case of Brazil

Within a dynamic programming approach, an optimal rule for the central bank to attain its inflation targeting goals is derived. The short-run nominal interest rate is used as an instrument to achieve monetary objectives. The model is tested for the Brazilian economy and compared with results found for other countries. Evidence for the estimated feedback interest rule for the Central Bank suggests that the cost of reducing inflation in an open economy is lower than that of a closed economy.

Two-stage inference using data envelopment analysis efficiency measurements in univariate production models

This article addresses the problem of modeling data envelopment analysis (DEA) inefficiencies as dependent on contextual variables. For this purpose we use a statistical model similar in appearance to inefficiency component specifications in stochastic frontier models. The underlying production response is univariate. The approach is asymptotic and is based on a two-stage statistical inference procedure. In the first stage inefficiencies are estimated using DEA. In the second stage these estimates are modeled as if they were the true inefficiencies by means of a statistical model dependent on the contextual variables. To define this data generating process one could use a flexible family of distributions like the truncated normal. Theoretical inefficiencies are assumed to be independent but not identically distributed. Some of the asymptotic results implied by the two-stage inference procedure are inspected in finite samples by means of Monte Carlo simulations. The procedure is illustrated with an example where a deterministic production model is fitted to research data generated by the major state company responsible for agricultural research in Brazil.

Significância de efeitos técnicos na eficiência de produção da pesquisa agropecuária

This article uses censored and truncated statistical models to assess technical effects potentially affecting the production frontier of Embrapas (Empresa Brasileira de Pesquisa Agropecuaria) research centers. Specifically the article assesses the significance of the factors level of research funds generated externally to the National Treasury, actions of partnership, technical quality of research projects, client satisfaction, change in administration, type and size on the measurement of technical efficiency of the research centers. Efficiency measures are computed both using Data Envelopment Analysis (DEA) and Stochastic Frontier methods. It is concluded that the DEA efficiency measurements provide more information regarding the assessment of technical effects and that revenue generation, actions of partnership, technical quality of projects, and size are significant factors. The statistical analysis leading to these results fits a Tobit model, defined by a distribution in the gamma family, to DEA technical efficiency measurements. This analysis is consistent with the fit of a truncated normal distribution to residuals obtained from the fit of a DEA production function.

Uma metodologia alternativa para mensuração de pressão sobre o mercado de câmbio

This article applies multivariate analysis to measure the exchange market pressures in 20 countries, in the context of the Asian crisis of 1997/98. We develop an index of pressure on the exchange market as a weighted average of exchange rate changes, reserve changes and interest rates changes. Changes in the exchange rate enter with positive weights, changes in reserves have negative weights and changes in interest rates have positive weights. These weights are derived from a measure of the communality of each variable used in the analysis. Using a exchange market pressure index we classify the sample, identifying by Wards method of cluster analysis, the countries most affected by the crisis initiated in mid 1997. The classification is validated with the use of discriminant analysis.

Influência de variáveis contextuais em medidas não-paramétricas de eficiência: uma aplicação com métodos de reamostragem

We evaluate three bootstrap techniques for the statistical analysis of a non parametric production model for which a DEA measure of efficiency is potentially affected by a set of exogenous factors. The application of interest relates to the assessment of the significance of the contextual variables income generation, processes improvement, intensity of partnerships, type and size on the technical efficiencies of Embrapas research centers. It is concluded that the bootstrap of the maximum likelihood estimator provides the best fit from the point of view of Pearson correlation between observed and predicted values and is the most informative in regard to the significance of the variables considered. With the exception of size all contextual variables are statistically significant. Income generation, processes improvement and intensity of partnerships are all positively associated with technical efficiency.

A Probabilistic Approach for Assessing the Significance of Contextual Variables in Nonparametric Frontier Models: An Application to Brazilian Banks

This article presents an empirical application illustrating the use of a nonparametric frontier model relying on a probabilistic definition of the production frontier. The significance of the variable nonperforming loans in productive efficiency is assessed, for a sample of Brazilian banks, using the concepts of condicional and unconditional efficiency measures, in a context where it is not necessary to impose any particular distribution for the production data. The analysis is robust relative to the assumptions of separability.

Credibilidade e crises cambiais: uma aplicação do modelo de Velasco

Este artigo analisa as crises cambiais dos principais paises da America do Sul, no periodo de 1992 a 1998, com base no modelo de Velasco (1996). Este e um modelo que sintetiza dois enfoques: ataques especulativos resultantes de desequilibrios nos fundamentos macroeconomicos e resultantes de profecias auto-rea-lizaveis, mesmo quando as economias apresentam bons fundamentos. Nove paises latino-americanos sao classificados por meio da construcao de um indice derivado de uma funcao perda. Desse modo e possivel agrupar os paises pelo grau de vulnerabilidade as crises, estabelecendo-se zonas de credibilidade. Os resultados indicam que a economia brasileira, por apresentar problemas nos fundamentos macroeconomicos, foi classificada em uma zona de credibilidade nula. A Argentina moveu-se de uma zona de alta credibilidade para uma faixa intermediaria. Isto indica que a Argentina tornou-se suscetivel as crises auto-realizaveis. As demais economias situaram-se numa regiao de alta credibilidade.

Credibilidade e crises cambiais: uma aplicação do modelo de Velasco (Credibility and Currency Crises: Empirical Evidence of Velasco's Model)

This paper analyzes the currency crises of the South American countries during the period from 1992 to 1998, based on the Velasco (1996) model. This is a model that synthesizes two approaches: speculative attacks resulting from unbalanced macroeconomics fundamentals and resulting from self-fulfilling prophecies, even when the economies show good fundamentals. Nine Latin American countries are classified by means of a loss function index. Thus, it is possible to classify the countries by the degree of vulnerability to the crises, establishing zones of credibility. The results indicate that the Brazilian economy, since it shows problems in its macroeconomic fundamentals, was classified in the zone of null credibility. Argentina moves from a zone of high credibility to an intermediate or partial zone. This is an indication that Argentina became susceptible to self-fulfilling crises. All other countries were classified in the zone of high credibility.

Crises Cambiais dos Anos 1990: Uma Análise Sob o Enfoque Custo-Benefício a Barro-Gordon (Currency Crises of the 90's: An Analysis on Barro and Gordon Approach)

Este artigo tem por objetivo verificar se os paises envolvidos nas principais crises cambiais dos anos 1990 seriam anunciados como vulneraveis as respectivas crises, com base em previsoes levadas a cabo com uso de um escore de avaliacao de crise definido pela funcao perda de Barro-Gordon (1983). Este escore de crise depende da inflacao e do crescimento real do produto. E um construto definido com a utilizacao de metodos estatisticos de analise multivariada. Neste contexto os paises da amostra sao avaliados. Identificam-se os mais vulneraveis as crises cambiais por meio da analise de conglomerados. A classificacao obtida e validada com o uso de analise disciiminante. Os resultados sao consistentes com os observados na literatura. This article aims to verify if the countries involved in the main currency crises of the 90s would be announced as vulnerables to the respective crises, basing the predictions on a score of crisis evaluation defined by the loss function of Barro-Gordon (1983). This crisis score depends on inflation and on the GDP growth. It is a construct defined with the use of multivariate statistical methods. In this context the countries in the sample are evaluated. Cluster analysis is employed to identify the more vulnerable to currency crises. This classification is validated via discriminant analysis. The results are consistent with the literature.