Estimating Trend Inflation with Moving Average and Stochastic Volatility for Nigeria

Abstract

This study was carried out to estimate underlying inflation in Nigeria using Unobserved Component (UC) model. Also, different channels were used to identify the source of inflation persistence and volatility. This was estimated using Bayesian analysis in order to examine the role of priors in improving the precision of the estimates. Following the analysis, a number of observations were made. Firstly, in terms of methodological contribution, we found that the observed inflation in Nigeria can be modelled by assuming constant volatility as the evidence shows that stochastic volatility (SV) is not a source of observed inflation persistence. However, the variance of the trend inflation can be modelled using regime switching models, where both the mean and variance can be subjected to regime changes. This is because we were only able to establish an evidence of instability not volatility in the variance of the trend inflation. Secondly, we found that the inclusion of autoregressive component in the UC specification leads to volatility in the variance of the observed inflation. Therefore, unobserved component model with MA is more appropriate in modeling inflation persistence in Nigeria which is consistent with empirical regularities. Furthermore, the study found that the observed core inflation (which is the headline less food inflation as released by Bureau of National Statistics, NBS) is being understated. This finding has significant implication for the current measure of core inflation and its application in policy decisions. Therefore, we recommend that the NBS should remove energy components from the observed core inflation in order to have a consistent permanent inflation component in Nigeria.