In this paper, we use the Markov switching model, the state-space model with Markov switching heteroskedasticity and the local level model with GARCH(1,1) disturbances to investigate the link between the level of Egyptian inflation and its uncertainty. We use different ways to measure inflation uncertainty. First, we consider it as the variance of unanticipated inflation. Second, we measure it as the unconditional variance of unanticipated changes in inflation. Finally, we measure it by the conditional variance modeled as GARCH effect. We find evidence of a positive effect of inflation level on inflation uncertainty for the three models. By making the distinction between the long run and the short run, we conclude that inflation has a significant positive effect on uncertainty in the short run but no effect in the long run. We state that the cost of inflation is mainly due to the association between higher inflation and higher short-run uncertainty. Reversing the causality link between inflation and its uncertainty, we find that inflation uncertainty has a positive effect on inflation level in the short run but this effect dies out in the long run, which is indicative of a stabilization monetary policy in Egypt. According to the likelihood-based information criteria (Akaike and Schwarz Bayesian criteria), the state-space model with Markov switching heteroskedasticity and the two-state Markov switching model outperform all others.
Read full abstract