Quantcast

Bayesian Analysis of Structural Changes in a Linear Regression Model: An Application to Rupee-Dollar Exchange Rate

Research paper by Anoop Chaturvedi, Arvind Shrivastava

Indexed on: 20 Nov '15Published on: 20 Nov '15Published in: Journal of quantitative economics : journal of the Indian Econometric Society



Abstract

The objective of this paper is to find the points of structural shift in the exchange rate model under a Bayesian framework which incorporates the possibility of shift or no shift in both disturbances precision and regression parameter. The Bayesian analysis of a linear regression model has been carried out under the mixture of prior distributions for the parameters. In order to find the structural shift points, a test based on posterior odds ratio for testing the hypothesis of no structural shift against the alternative hypothesis of shift due to change in disturbances precision and regression parameters has been developed. Further, to illustrate the theoretical findings, the effect of interest rate, growth rate of trade balance and GDP growth rate (market price) on Indian rupee-US dollar exchange rate has been investigated. The application of proposed model indicates that the strongest structural shift of exchange rate captured in the years 1997–1998 and 2007–2008. The stability in the foreign exchange market was disrupted during 1997–1998 and 2007–2008 due to intensification of East Asian crises and global financial crisis respectively.