EXPLAINING THE DETERMINANTS OF THE FREQUENCY OF EXCHANGE RATE INTERVENTIONS IN PERU USING COUNT MODELS

Abstract

The determinants of the frequency of Central Bank interventions (purchases and sales) in the Peruvian exchange rate market are analyzed using weekly data for the period from January 2001 to December 2010 using count data models (Poisson, Negative Binomial and Zero Inflated). Findings show that the deviations of the logarithm of the exchange rate with respect to a long term trend, previous week’s interventions (persistency), the Embig spread, the spread between interbank interest rates, and the spread of prime corporate interest rates are important determinants. In terms of the models used, the Zero Inflated models allow a better fit and performance in predicting the number of interventions (purchases and sales).