Econometric Models for Forecasting Regional Economic Growth and Their Improvement

This study develops and tests a panel autoregressive framework for forecasting regional economic growth in Uzbekistan's 14 territorial units using official SIAT data for 2010-2025 (224 region-year observations; 210 after introducing a one-year lag). Pooled OLS and two-way fixed-effects specifications are compared using region-clustered standard errors. The pooled OLS estimate indicates positive growth persistence (coefficient = 0.288, p < 0.001), while the lag coefficient becomes insignificant after controlling for region and year effects (0.207, p = 0.135). An expanding-window one-step validation for 2023-2025 yields an MAE of 1.22 and an RMSE of 1.62 percentage points. The baseline forecast implies mean regional growth of 7.12% in 2026, converging to 6.80% by 2030. The results show the importance of controlling for territorial heterogeneity and common shocks and indicate that a parsimonious autoregressive model should be treated as a benchmark rather than a complete structural model of regional growth.