Does New Zealand's inflation target anchor long-term expectations?

ABSTRACT

A claimed benefit of an inflation target is that it provides an anchor for long-term inflation expectations. We evaluate this by using survey-based inflation expectations data for New Zealand to assess whether forecasts for different horizons systematically converge to the RBNZ's target. We find evidence of this from the aggregated expectations data, and for most individual professional economists and financial sector forecasters in our sample. However, the evidence from forecasters in other sectors is mixed. In particular, those from the business community appear more focussed on the RBNZ's inflation track record. This could contribute to episodes of persistently low inflation.

KEYWORDS:

• Inflation expectations
• inflation target
• anchoring
• monetary policy

JEL CODE:

• D84
• E52
• E58

Acknowledgements

The data provided for this study by Aon Hewitt NZ Ltd and the Reserve Bank of New Zealand are gratefully acknowledged, as is the feedback on a previous version of the paper provide by two anonymous reviewers. The usual disclaimer applies.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Notes

1 Other potential benefits include a reduction in the level and volatility of inflation (Bernanke et al., 1999; Johnson, 2002; Levin et al., 2004a), and in the real cost of disinflation (Gonçalves & Salles, 2008; Mishkin & Schmidt-Hebbel, 2007).

2 The United Kingdom's inflation target was reduced from 2.5% to 2% in December 2003, but this coincided with a change in the official inflation measure and the change to the target largely reflected the difference between the two measures. In March 2007, the Czech National Bank (CNB) reduced its medium-term inflation target from 3% to 2% (±1%), but this was prompted by the decision in late 2006 to postpone adoption of the euro to beyond 2010. Hence, the CNB's new target simply maintained the status quo – i.e. that the ECB's 2% inflation target would apply after 2010. A number of other countries (e.g. Canada, Israel and New Zealand) lowered their inflation target in a series of steps during their disinflationary process. However, this ‘phasing in’ of a country's ultimate inflation target is qualitatively different from changing a long-established target.

3 A variation on this theme is to investigate the impact on expectations of changes in the actual rate of inflation (e.g. Clark & Nakata, 2008; Demertzis, Marcellino, & Viegi, 2010; Levin, Natalucci, & Piger, 2004b; Van der Cruijsen & Demertzis, 2011).

4 Two definitions of the ‘medium term’ are needed because, as discussed in Section 5, the expectations data are sourced from two separate surveys. Although both surveys ask for one-year ahead forecasts, their longer forecast horizons do not match.

5 Miyajima and Yetman (2019) report similar findings for South Africa.

6 Karagedikli and McDermott (2018) find evidence that inflation expectations in New Zealand have become more backward-looking in recent years. We shall return to this issue in Section 7.

7 Michael Riddell was the Head of Financial Markets at the RBNZ in 2002, one of three senior managers who reported directly to the Governor.

8 Projected inflation rates as high as 2.6% were described as being comfortably within the target range in later Monetary Policy Statements. Reddell (2016) also recalls that Governor Bollard ‘did not see [increasing the lower bound of the target range] as a route to higher inflation, but rather to cementing in something more like the average inflation outcomes of the previous few years’, which had been in excess of 2%.

9 Data on seven-year-ahead forecasts are not used here because the differences between this series and the four-year-ahead forecast values are typically very small. There are also more gaps in the seven-year-ahead series of several individual forecasters. The results obtained for the forecasters with complete seven-year-ahead data are not qualitatively different from those based on the four-year-ahead data.

10 The RBNZ's Survey of Expectations has an average of 80 respondents, compared with 11 for the Aon Economists’ Survey. Following the cessation of the Aon Hewitt survey, the RBNZ survey began asking respondents for their five- and ten-year-ahead inflation forecasts as well. Both surveys also ask respondents for their forecasts of other key economic variables.

11 A little over one-third of the RBNZ survey's panel are drawn from the financial sector. A further third come from the business sector and about one-fifth represent either labour or the agricultural sector. The remainder (classified as ‘others’) includes economic commentators and advisers.

12 The RBNZ ceased reporting the response rates in 1997.

13 I am grateful to an anonymous reviewer for suggesting this point.