Avaliação dos determinantes macroeconômicos da inadimplência bancária no Brasil

• Ano de defesa: 2008
• Autor(a) principal: Fernando de Menezes Linardi
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: autoregressão vetorial; taxa de inadimplência; Risco de crédito; simulações de Monte Carlo; Método de Monte Carlo; Inadimplência (Finanças) Modelos matemáticos; Bancos Brasil; Risco (Economia)
• Link de acesso: http://hdl.handle.net/1843/AMSA-7FNJU7

This dissertation investigates the relationship between the delinquency rate on Brazilian banks loans and macroeconomic factors for the period 2000 to 2007 using a VAR (Vector Autoregression) model. Loans were divided into State-owned and private financial institutions in order to assess the effect of macroeconomic shocks on the delinquency rate of these institutions. The results show that the delinquency rate of financial institutions is particularly sensitive to shocks on output gap, variation of the index of average income of workers and nominal interest rate. The estimated model produced good out of sample forecasts of delinquency rate and the results indicate that they are not worse than the forecasts of two competing models. The VAR model also allowed us to estimate the correlations of macro variables and to compute the probability that the delinquency rate exceeds a given threshold deemed risky through Monte Carlo simulations. This procedure may be used as an additional tool of credit risk management by the Central Bank and financial institutions.