Un análisis del ciclo económico de la República Dominicana bajo cambios de régimen[Analysis of business cycle of the Dominican Republic using Markov Switching model]

This paper presents a univariate model that analyzes systematic changes in the behavior of the business cycle in the Dominican Republic, capturing changes in average growth and identifying differences between contractions and expansions with respect to their persistence and duration. To do so, it uses the classic algorithm described by Hamilton (1990, 1991) that consists of two parts. In the first part, population parameters, including joint probability density of unobserved states, are estimated. In the second part, using a nonlinear filter and smoothed probabilities, probabilistic inferences are made about unobserved states. Our results suggest that the characteristics of the distribution functions estimated for each scheme differ, both in mean and standard deviation. Thus, for a recessive event or contraction, average quarterly growth was around -0.33% with a standard deviation of 0.45%, whereas for an expanding statistical event, estimates were 0.23% and 0.27%, respectively.