Estimation of a small open economy DSGE model for Kazakhstan

Abstract

This paper adapts the DSGE (dynamic stochastic general equilibrium) model of Medina and Soto (2007) in the context of Kazakhstani economy, and fully estimates it using Bayesian methods. The main goal of the paper is to contribute to the scarce macroeconomic modeling literature on Kazakhstan.Overall, we find that the oil price shock is key in explaining the variance of virtually all the variables of interest -in particular, it accounts for more than 40% of variance in real exchange rate over the longterm horizon. Furthermore, while the oil price and commodity (oil) production shocks contributed positively to the country's GDP growth in real terms before the Great Recession, their effects have been primarily negative during the two major economic crises of 2007 and 2015, and the fiscal policy has had mixed success in counteracting them. The model produces relatively accurate in-sample predictions for most of the variables compared to VAR and random walk, and outperforms them in forecasting monetary variables. Lastly, the counterfactual exercises show that the choice to adopt the floating exchange rate policy in 2015 has prevented a larger output slump in the short-term at the cost of stronger currency depreciation, and that countercyclical fiscal rules would have greatly mitigated the immediate negative impact of the 2007-08 and 2015 crises, in addition to making real output more stable overall.