Forecasting Inflation Rate of Bangladesh Using Exponential Smoothing (ETS)

Abstract

The focus of the study is forecasting performance of Exponential Smoothing (ETS) models for Bangladesh’s inflation rate. We have used officially published monthly data covering 2000-2024. Inflation in Bangladesh has not followed a smooth or stable path over this period, with episodes of food price spikes, exchange- rate pressures, and policy shifts, which makes forecasting particularly challenging. Instead of imposing a single specification, we estimate and compare a range of ETS models, since different structures capture trend, seasonality, and short-run movements in different ways. Based on out-of-sample forecast accuracy, one ETS specification consistently performs better than the others. When its forecasts are compared with those from standard ARIMA models, ETS appears to track turning points and short-term movements more closely, although the differences are not uniform across all horizons. These results suggest that ETS provides a useful and relatively flexible framework for forecasting inflation in Bangladesh, and may be especially well suited to emerging economies where inflation dynamics tend to shift over time.

Social Science Review [The Dhaka University Studies, Part-D], Vol. 42, No.2, December 2025, pp. 211-230