Time-Varying Parameters of Inflation Model in Nepal: State Space Modeling

Abstract

This paper attempts to investigate the stability of time-varying parameters of the random walk model of inflation in Nepal. This study has been motivated with the Lucas Critique (1976) that the monetary/fiscal policy that is exposed to change over time affect the expectations of forward looking economic agents which hence lead to non-constant time-varying parameters of the model. Monthly time series of inflation ranging from August, 1997 to July, 2012 has been utilized for the analysis. Applying the Kalman Filter technique for the estimation of coefficients of random walk model, we found non-constant time varying parameters of both the constant and autoregressive of order one AR(1) coefficient of inflation over the long run. The changes in the expectations of rational economic agents on macroeconomic policies as a result of the problems of policy commitment, credibility and dynamic consistency might have attributed such non-constant timevarying parameters. Therefore, in addition to supply smoothing policies to control inflation in Nepal, consistent and credible policies that are not exposed to change over time may reduce the gap of actual inflation from its targets and hence trigger inflation into desired level.