Trend Inflation and Inflation Persistence in Australia: A New Keynesian Perspective

Abstract

The true nature of inflation dynamics is an ongoing matter of debate and investigation in modern macroeconomics. That such attention is devoted to the dynamics of inflation is due to its importance, not only for understanding the nature of business cycles, but also for determining the appropriate path for monetary policy. Modern models of inflation are typically derived from the seminal contributions of Calvo (1983) and Taylor (1980) which imply a purely forward-looking New Keynesian Phillips curve (NKPC) where inflation depends on its future expectation and the level of real marginal costs. Despite its theoretical elegance, the purely forward-looking incarnation of the NKPC has been shown to perform poorly against the data. The empirical shortcomings of the NKPC are generally attributed to its inability to replicate the innate persistence which is present in inflation (see, for example: Fuhrer and Moore, 1995). In order to enhance the degree of persistence within the model several authors have proposed somewhat ad-hoc rationales for the inclusion of lagged inflation terms in the NKPC (see, for example: Gali and Gertler, 1999 and Christiano, Eichenbaum and Evans, 2005). While these `hybrid Phillips curves' do indeed improve the fit of the model, their questionable microfoundations are an obvious source of criticism. This thesis attempts to ascertain the extent to which inflation dynamics in Australia can be explained by the NKPC without having to rely on arbitrary backward-looking terms that have limited structural meaning. In particular, this analysis considers whether an adapted version of Cogley and Sbordone's (2008) extended model of timevarying trend inflation is suffcient to explain inflation persistence in Australia without the need for a backward-looking term in the NKPC.