The Impact of Inflation and Unemployment on the Real Growth of Household Incomes: Dynamic Panel Evidence and Forecasts for the Regions of Uzbekistan

This study estimates how inflation and unemployment affect the growth of real per capita household income across the regions of Uzbekistan and develops a forecasting model whose accuracy is verified out of sample. Official data from the SIAT database of the National Statistics Committee cover 14 regions over 2010–2025 (224 observations). We combine Pesaran CD and CIPS tests, a dynamic panel ARDL model with Driscoll–Kraay standard errors, Mean Group and two-way fixed-effects estimators, and compare forecasting models in pseudo-real time over 2019–2025 using RMSE, Theil’s U and Diebold–Mariano tests. A one-percentage-point rise in last year’s inflation lowers real income growth by 0.78–1.11 percentage points, and in 2019–2025 a one-off inflation shock reduced the real income level cumulatively by 1.02%. The strongly negative unemployment effect produced by a simple national OLS regression (−2.45; p = 0.027) is an artefact of the 2018 change in the unemployment methodology: once the regime shift and common shocks are controlled for, the effect is insignificant (−0.06 to −0.70). A parsimonious dynamic panel model with a regime shift reduces the root mean squared forecast error by 27% relative to a naive forecast (Theil’s U = 0.73; DM test p = 0.012), and its 80% prediction intervals cover 88.8% of realised outcomes. The model projects real income growth of 6.4–6.5% in 2026–2028; if inflation exceeds the central bank’s path by 3 percentage points in 2027, growth slows by 3.6 percentage points in 2028. Disinflation towards the 5% target and forward-looking indexation of transfers are therefore central to protecting real incomes.