Your browser is not supported

Our website does not support the browser you are using. For a better browsing experience update to a compatible browser like the latest browsers from Chrome, Firefox and Safari.

Great Expectations: performance of survey inflation expectations at improving model-based inflation forecasts

This analytical note considers how much survey inflation expectations improve headline inflation forecasts.

Meltem Chadwick and Tyler Smith

Great Expectations: performance of survey inflation expectations at improving model-based inflation forecasts
Great Expectations: performance of survey inflation expectations at improving model-based inflation forecasts

This analytical note considers how much survey inflation expectations improve headline inflation forecasts.

What this paper is about

Meltem and Tyler talk about their research and share their key findings.

Meltem: Hi I'm Meltem

Tyler: and I'm Tyler, researchers here at Te Pūtea Matua.

Tyler: Recently we looked into inflation expectations and how we can better predict future inflation, based on how much people expect prices to change over time.

Meltem: We found that knowing what people's expectations of inflation will be in the future, is essential for effective monetary policy. This is because if people expect high future inflation, prices will generally rise to what people are willing to pay.

Tyler: But if people feel future inflation will be low, they probably won't want to pay more for goods and services. We found that including inflation expectations and inflation forecasts significantly improves their results.

Meltem: We also learned that 1 and 2 year ahead household inflation expectations add the most value to forecasting headline inflation and are more accurate than longer term expectations or models.

Tyler: You can read the full analytical note on the Reserve Bank's website.

Key findings

  • This Note presents empirical evidence suggesting that New Zealand households tend to pay more attention to inflation when it is high than when it is low. In the academic literature, this is known as rational inattention.
  • Rational inattention can result in a non-linear relationship between actual inflation and households’ inflation expectations. Once inflation rate rises above a certain threshold, for example, 2%, household inflation expectations line up closely with actual inflation. Thus, inflation expectations may be slow to respond to monetary policy announcements, and this may make it more difficult for the central bank to rein in high inflation by raising interest rates.
  • It is important for monetary policymakers to monitor this potential non-linearity in how households perceive and internalise inflation data.

The Analytical Notes series encompasses a range of background papers prepared by Reserve Bank staff. Unless otherwise stated, views expressed are those of the authors, and do not necessarily represent the views of the Reserve Bank.

 

We considered 14 survey-based inflation expectation series, ranging from expectations of 1 to 10 years ahead:

  • The ANZ business outlook (ANZBO) is a monthly survey of 300 to 400 firms which we averaged into a quarterly frequency to match the frequency of inflation outturns. Sectoral sub-splits are also available (grouped into retail, manufacturing, agriculture, construction and services sectors), allowing us to examine the relevance of sectoral expectations of inflation.
  • The survey of household inflation expectations (SHIE) and the survey of expectations (SoE) are both RBNZ surveys. The household survey collects responses from around 1000 individuals, from which we take the published mean result for the relevant quarter.
  • In comparison, the survey of expectations collects responses from professional forecasters and businesses in New Zealand. Response rates for this survey have declined, with current releases only including around 30 to 40 responses each quarter.
  • We also utilize Consensus forecasts for inflation expectations for New Zealand. These are published semi-annually in April and October from expectations by a group of international and domestic economists.