Using a Restricted Kalman Filtering Approach for the Estimation of a Dynamic Exchange-Rate Pass-Through pp. 255-268
Authors: (Rafael Martins de Souza, Luiz Felipe Pires Maciel, Adrian Pizzinga, National School of Statistical Sciences (ENCE), Brazilian Institute of Geography and Statistics (IBGE), and others)
Abstract: In this paper we propose linear state space models to estimate the time-varying passthrough of Brazilian price indexes against the US Dollar/Real exchange rate from 1999 to 2007. The methodological framework encompasses the restricted Kalman filtering under a reduced modeling approach, under which it becomes possible to check whether some economic hypotheses are supported by the data. The paper has three main targets. The first is to decide whether models of null (or of full) pass-through are acceptable to the price indexes investigated here. The second is to carry out likelihood ratio tests for the significance of some economic exogenous variables, which are termed determinants in this paper and are theoretically associated with the passthrough. The third is to analyze the behavior of the Kalman filter estimates of the pass-through from the best models.