FORECASTING STATE BUDGET REVENUE IN UZBEKISTAN: A BACKTESTED TREND-MODEL APPROACH FOR 2025–2027
Keywords:
Tax revenue forecasting; state budget; trend model; Holt exponential smoothing; backtesting; Uzbekistan. JEL: H68, C53, E62.Abstract
Uzbekistan's post-2017 reforms left a short, structurally shifted revenue series that rules out ARIMA, so this paper builds transparent trend forecasts of state budget revenue for 2025–2027. Annual execution data for 2018–2024 (Ministry of Economy and Finance) are modelled with a log-linear constant-growth trend and additive-trend models (linear and Holt), ranked by an out-of-sample backtest (train 2018–2022, test 2023–2024) and checked against the approved 2025 budget. Revenue rose from 79.1 to 274.3 trln UZS (23.0% CAGR), with growth decelerating to 18.4% by 2024. The additive trend is far more accurate out of sample (MAPE 4.0% vs 14.9%) and its 2025 forecast (298.6 trln UZS) is within 3.2% of the plan (308.5); it projects about 299, 331 and 362 trln UZS for 2025–2027
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
