Microlevel Analyses of DSGE Model Parameters: Evidence from Kazakhstan

Dynamic stochastic general equilibrium (DSGE) models are widely used by central banks, government agencies and financial organizations to conduct simulation and forecast relevant macroeconomic indicators in the economy. The most important inputs into all DSGE models are structural parameters which are either calibrated from other sources or estimated via Bayesian methods. Using non-public microlevel data, we estimate ten structural parameters for Kazakhstan: the elasticity of substitution between exports and imports, constant relative risk aversion, intertemporal elasticity of substitution in consumption, Frisch elasticity of labor supply, the depreciation rate of physical capital, capital and labor shares, and the elasticity of substitution between tradable and nontradable goods. Various econometric techniques such as fixed-effects, generalized method of moments (GMM), Arellano-Bond, and non-linear iterative maximum likelihood estimation are used to obtain consistent estimates of the models' coefficients. The structural parameters can be used in calibrated DSGE models as fixed parameters or as prior information in Bayesian estimation of the models.