Inflation Dynamics and Forecasting Performance in Developing Economies: A Cross-Country MIDAS Analysis

Abstract

This study examines the inflation dynamics and its forecasting performance for 26 developing economies for the year 2000-2023 using Mixed Data Sampling (MIDAS) regression. The methodological farmwork includes an Augmented Phillips curve approach incorporating asset prices, financial indicators, and macroeconomic indicators. The MIDAS regression technique is used to incorporate the shortcomings of data availability in different frequencies. It incorporates the high frequency indicators, financial indicators, and low frequency macroeconomic indicators in one regression. Hence, the variables are not needed to be aggregated or disaggregated to match the frequencies in same regression. Several MIDAS specifications are estimated such as Beta, PDL/Almon, Exponential Almon, step and UMIDAS to get the RMSE and Theil index for in-sample and out-of-sample forecasts. The results showed the effectiveness of MIDAS models specifying the predominance of Beta method over other specifications for many countries. Inflation occurred as information-rich and multidimensional process which is caused by real economic activity, monetary policy indicators, and other external factors while monetary aggregates are shown as weak predictors of inflation. The relationship between the Phillips curve remained the same in the case of almost all countries but it is more of a hybrid nature and the context (geographical nature of economies) matters. There are cross-country heterogeneity and a need for flexible, data-intensive modeling techniques to effectively predict inflation and design policies in developing economies. This analysis also re-evaluates the Phillips curve relationship and highlights that it is still relevant, but it has now become hybrid and context dependent. The research makes a significant contribution to literature as it gives a more in-depth comparison of the cross-country inflation processes and dynamics in the developing economies.