A small scale forecasting and simulation model for Azerbaijan (FORSAZ)

In our study, we model both steady state and short-run dynamics of the important aspects of the national economy using quarterly data for the period 1999Q1-2016Q2. We explicitly model government, money market and external sector, but omit household sector, labor market, wage dynamics and volume of the physical capital specifications due to serious data quality issues. Using Fully Modified OLS (FMOLS) co-integration methodology we explore co-integration relations among the variables. Coefficient estimates of short-run dynamics are in compliance with our ex-ante expectations. Stability tests indicate that the system seems to exhibit stability around its steady state values and model variables converges to their steady state values approximately within 140 periods (2016Q3-2050Q4). Impulse-response analysis also show stable convergence of the model and predict economically consistent results. The results of in-sample and out-of-sample simulation exercises for the inflation, the government consumption and the imports are satisfactory. However, it seems that the model cannot adequately capture ex-post dynamics of NFA and reserve money. In general the results indicate that model can be used for the specific policy analysis and forecasting of main macroeconomic variables of Azerbaijan.