Política cambial com dívida indexada em moeda estrangeira no Brasil, 1995-2004

Resumo

This thesis discusses the Brazilian public sector’s currency-linked domestic debt between 1995 and 2004. It argues that the reasons of that debt are not to be found in the goals of optimal debt composition, but rather among the exchange rate policy objectives. The Mexican Tesobonos case offers insights for the Brazilian case, highlighting three major differences, namely the average term of the bonds; the stronger concentration of Mexican debt in international markets; and the importance of non-residents as holders. The Brazilian case was presented in historical perspective, starting with the introduction of the ORTN with dollar-indexation clause by the end of the 60’s. It attempted to follow the development of the Brazilian currency-linked domestic debt from the 60’s onwards. Taking into account that the objectives of that policy instrument are more related to exchange rate policy it was presented a brief survey of the literature on the short-term determinants of the exchange rate, starting with Meese and Rogoff (1983), which had introduced the idea that a random walk is a better forecast to exchange rate than any other model based on macroeconomic fundamentals. It is also presented a review of the literature on foreign exchange intervention, which offers plenty of works devoted on developed countries and scarce contributions on emerging countries. The offering of currency-linked domestic bonds is treated as foreign exchange intervention, in search for its effects on the exchange rate level and volatility. The econometric exercise built a database of the stock of currency-linked domestic on a daily basis, through aggregation of primary auctions of bonds and swaps over the period between 1991 and 2004. The exercise estimated the impact of such debt over foreign currency prices in two periods, namely the administered regime (1995-1999) and the floating period (1999-2004). It was used a Vector Error Correction Model (VECM). Estimated elasticities showed a possible simultaneity bias, which was detected through a Granger causality test. The latter provided evidence of a feedback-type bi-directional causality. Estimations are subject to heteroskedasticity, which was corrected through GARCH modeling. This test showed that the elasticity of the exchange rate with respect to intervention with currency-linked domestic debt was -0.17 in the administered regime and 2.15 in the floating regime.

Ficha do documento

Tipo

Tese

Ano

2005

Instituição

Fundação Getulio Vargas

Fonte

Repositório da FGV

Idioma

Português

Acesso

Não informado

Identificador

oai:repositorio.fgv.br:10438/1768

Temas

EconomiaGestão PúblicaFinanças Públicas

Palavras-chave

Public securities debtOfficial interventionExchange rateForeign currency-linked debtDívida pública mobiliáriaIntervençãoIndexação cambialSéries temporaisGARCHEconomiaPolítica cambial - BrasilDívida pública - BrasilCâmbio