FORECASTING NPL OF COMMERCIAL BANKS IN UZBEKISTAN USING VAR MODEL
Abstract
This study examines the forecasting of non-performing loans (NPL) in Uzbekistan's banking sector by focusing on key macroeconomic indicators — GDP growth, interest rate, exchange rate, and inflation rate. I analyze monthly data from March 2020 to October 2024 by employing a Vector Autoregressive (VAR) model to capture the dynamic relationships between endogenous variables. The findings suggest that lagged GDP growth and interest rate have a significant positive association with the NPL rate, reflecting delayed effects of economic growth and borrowing costs on loan performance. The VAR model's predictive accuracy was validated against historical data, and future forecasts for the next six months were generated with a 60% prediction interval.