Financial Frictions in a DSGE Model for Latvia

This paper builds a dynamic stochastic general equilibrium (DSGE) model for Latvia that would be suitable for policy analysis and forecasting purposes at Latvijas Banka. For that purpose, the DSGE model with financial frictions of Christiano, Trabandt and Walentin (2011) is adapted to Latvia's data, estimated, and studied as to whether adding the financial frictions block to an otherwise identical (baseline) model is an improvement with respect to several dimensions. The main findings are: 1) adding of financial frictions block provides a more appealing interpretation for the drivers of economic activity and allows reinterpreting their role; 2) financial frictions played an important part in Latvia's 2008 recession; 3) the financial frictions model beats both the baseline model and the random walk model in forecasting CPI inflation and GDP, and performs roughly the same as a Bayesian structural vector autoregression.