Modelling New Zealand Inflation in a Phillips Curve

Article excerpt

This article presents some simple estimates of Phillips curves for New Zealand inflation and outlines a recent reorganisation of the inflation process in the Reserve Bank's Forecasting and Policy System (FPS). While modern economic theory suggests the traditional Phillips curve should be used only with care, empirical estimates for New Zealand suggest it continues to have some value. We find that estimates of the impact of resource pressure (the "output gap") on inflation are easiest to obtain from an equation on the non-tradables component of CPI inflation, and show that this relationship can be improved statistically by introducing a (fairly smooth) measure of inflation expectations on the right hand side. We present some evidence that the relationship between resource pressures and non-tradables inflation is stronger in New Zealand than some comparable countries, and further evidence that this could be related to the cyclicality of housing construction costs in New Zealand. In the latest version of the Reserve Bank's macro-model, FPS, the inflation process has been written with a tradables/non-tradables split and an explicit empirical measure of inflation expectations. This does not greatly change model properties but allows the model's congruence with the data to be assessed more directly.

1 Introduction

A central bank charged with controlling inflation needs to monitor and understand the inflation process within its economy. Building up a view of the determinants of inflation will typically involve a range of statistical techniques and economic theories.

While the Reserve Bank takes an eclectic approach to thinking about the possible causes of inflation, the Bank's core model of the economy (the Forecasting and Policy System or FPS) focuses on a fairly simple story of inflation determination, the so called "Phillips curve" which relates inflation to measures of resource strain in the economy.

A recent article in the Bulletin (Hodgetts, 2006) described some changes in New Zealand's inflation process over the past 20 years. This article builds on that discussion by providing more details on the Phillips curve structure in use within FPS, which has recently been altered significantly to allow better monitoring of the inflationary process. We begin by discussing the ideas and history behind the Phillips curve and some recent related theoretical developments in section 2. In section 3, we describe some representative Phillips curves estimated on New Zealand non-tradables and tradables inflation data, showing a significant relationship between non-tradables inflation and resource pressure. We show that this relationship looks more reliable in New Zealand than in some other countries, and consider further why this might be the case. In section 4 we describe how this evidence has been calibrated into the most recent version of FPS.

2 Some history of the Phillips curve and recent developments

The simplest Phillips curve, which Bill Phillips, a New Zealand economist, sketched around 1958, showed an empirical inverse relationship between inflation and unemployment "in the previous 90 years or so of United Kingdom data" (Laidler, 2001). The curve has attracted a lot of fame and notoriety ever since.

The notoriety of the Phillips curve came about because of the risk that it might be interpreted as a potentially exploitable relationship for policy purposes. A possible interpretation of the curve was that lower unemployment could be permanently achieved at the cost of a higher (but stable) rate of inflation. However, as shown by Milton Friedman and others, this depended on the assumption that inflation expectations would remain stable as inflation rose. If inflation expectations instead responded gradually (adaptively) to rising inflation, there would only be a temporary reduction in unemployment unless the central bank was prepared to tolerate steadily increasing rates of inflation. …