Academic journal article The Reserve Bank of New Zealand Bulletin

Assessing Recent External Forecasts

Academic journal article The Reserve Bank of New Zealand Bulletin

Assessing Recent External Forecasts

Article excerpt

1 Introduction

Due to the lags with which monetary policy affects inflation, macroeconomic forecasting is a critical component of monetary policy (for further discussion, see Drew and Frith 1998). The current Reserve Bank published forecasts are constructed with the assistance of our new dynamic stochastic general equilibrium (DSGE) model 'KITT' (see Lees 2009). However, as part of the decision-making process, the Reserve Bank also considers information from a variety of sources, including forecasts from external agencies. Consequently, the Reserve Bank dedicates significant resource to gathering and assessing external forecasts each quarter. This article seeks to enhance the value of this exercise by establishing the recent forecasting accuracy of the relevant agencies.

Over the past 5-10 years, several pieces of work have analysed the accuracy of Reserve Bank forecasts and compared them to those from external agencies. McCaw and Ranchhod (2002) assessed the Reserve Bank's forecasts between 1997 and 2002 and found that the Reserve Bank consistently under-predicted CPI inflation over that period. While the work concentrated on explaining this bias, it also included a comparison with several external forecasters and found those forecasts were about as biased and accurate as the Reserve Bank. Turner (2006) updated this work by assessing the Reserve Bank's forecasting performance over 2003 to 2005 against New Zealand Consensus forecasts for several macroeconomic variables produced by Consensus Economics Inc. (1) Turner found that Reserve Bank forecasts were at least comparable to this average. In the case of the 90-day interest rate and CPI inflation, Reserve Bank forecasts were more accurate and less biased. (2)

This article updates these assessments of forecast accuracy. However, in contrast to previous work in this area, the majority of our data has been collected two-to-three weeks prior to the Reserve Bank projections being finalised. Consequently, the comparisons are more meaningful, as each forecaster has a similar information set--placing each contender on a level playing field.

The article proceeds as follows. Section 2 describes the forecast data. Section 3 discusses the large changes in economic conditions over our sample period to illustrate the difficulties faced by forecasting agencies during this time. Section 4 details the methodology used to determine forecasting accuracy. Section 5 contains the results of our analysis and notes the limitations of this work. Section 6 concludes.

2 The forecasts

The data for this work comes from Reserve Bank published forecasts and external forecasts collected over the period 2003q1 to 2008q4. In contrast to previous work assessing the accuracy of Reserve Bank and external forecasts, the majority of our forecast data has been collected two-to-three weeks prior to the Reserve Bank projections being finalised. Consequently, the forecast comparisons are more meaningful as most forecasters have had access to a similar information set. The exceptions are New Zealand Institute of Economic Research (NZIER) and NZIER Consensus: (3) their forecasts can be up to three months out of date. However, we include these forecasts in our average measure due to their comparable forecasting performance over our sample period.

As in a horse race, where contenders perform better in some conditions than others, forecasters have their stronger areas of performance. While recent economic conditions have proved difficult for forecasters, comparisons between external and RBNZ forecasts still play an important role in monetary policy formulation.

[ILLUSTRATION OMITTED]

The data includes one-year and two-year ahead forecasts for real GDP growth (March year annual average percentage change), CPI inflation (March year annual percentage change), the 90-day interest rate (March year annual average) and the TWI (March year annual average) from eight external forecasting agencies as well as the Reserve Bank. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.