Parameter drifts, misspecification and the real exchange rate in emerging countries

Abstract

This paper reviews the baseline framework for the analysis of emerging economies. Using Argentinean data, I estimate a small open economy model with stochastic trend, working capital constraint and augmented with time-varying parameters. I find that "structural" technological and financial parameters of one-sector model are time-varying during 1936-2006. Time-varying parameters correlate with the real exchange rate, suggesting potential misspecification of the one-sector model. Therefore, I propose a two-sector model that endogenously accounts for the real exchange rate. In this model, stationary productivity shocks and the country premium together explain a large share of the variability observed in the data.

Approximating Time Varying Structural Models with Time Invariant Structures

“…In applications, procedures like the one described in section 3 or the one of Seoane (2014) can be used to select the interpretation of the additional shocks.…”

Detecting and Analyzing the Effects of Time‐varying Parameters in Dsge Models

“…In fact, models with n 1 structural shocks and n 2 time‐varying parameters, models with structural shocks and models with n 1 structural shocks and n 2 measurement errors are observationally equivalent: where ; and . In applications, procedures like the one by Seoane (), can be used to select the interpretation of the additional shocks.…”

“…and v t = Ru t . In applications, procedures like the one by Seoane (2016), can be used to select the interpretation of the additional shocks.…”

“…Such a requirement is crucial, for example, to distinguish structural from reduced-form models and to conduct correctly designed policy counterfactuals. Dueker et al (2007), Fernandez-Villaverde and Rubio-Ramirez (2007), Canova (2009), Liu et al (2011), Vavra (2008), Dew-Backer (2014), Meier and Sprengler (2015), Seoane (2016), Castelnuovo and Pellegrino (2018), among others, have shown that the parameters of dynamic stochastic general equilibrium (DSGE) models are not time invariant and that variations are small but persistent. Parameter variations do not necessarily imply that DSGE models are not structural (see, e.g., Cogley and Yagihashi, 2010, Chang et al, 2013, Schmitt-Grohe and Uribe, 2003, Hansen and Sargent, 2010, and Cogley et al, 2015, but they create concerns about the economic interpretation of the results.…”

“…Earlier work by Parkin () studied whether the parameters of a Real Business Cycle (RBC) model are a function of one particular omitted variable. Seoane () finds endogenous parameter variations, interprets it as evidence of misspecification, and proposes a more complex constant parameter model that fits the data better. Kulish and Pagan () characterize the decision rules and the likelihood function of a DSGE model when predictable structural breaks occur.…”