Consumer Inflation Expectation

The Melbourne Institute Survey of Consumer Inflationary Expectations measures what Australian households expect consumer prices to do over the next 12 months. It is published monthly by the Melbourne Institute of Applied Economic and Social Research at the University of Melbourne. Each month about 1200 households answer by telephone through the Consumer Attitudes, Sentiments & Expectations (CASiE) program.

It does not measure past price rises, which sit in the Consumer Price Index, a price collection by the Australian Bureau of Statistics. It also excludes business, economist and financial-market forecasts.

It is released monthly on a mid-month Thursday at 11 am Sydney/Melbourne time, in AEST around mid-year and AEDT in summer. The exact Thursday varies by month and the next date is announced 1 month ahead in each release. Each release gives the new headline against the prior month plus distribution detail for answers in the 0-5% range and cluster notes. March, June, September and December releases also give 12-month past pay growth, 12-month expected pay growth and rise versus fall shares.

The headline moves with the spread of household answers, not store prices. Five parts below show where each move comes from.

• Headline expected rate, 30% trimmed mean: captures the average householder's expected price change over the next 12 months after dropping extremes.
• Weighted share expecting 0-5%: captures the share of valid answers, excluding don't knows, in the moderate range after survey weighting.
• Weighted mean inside 0-5%: captures the average of moderate answers after survey weighting.
• Response clusters: captures shifts at common answers like 10%, 5%, 2-3% or 0-2% that move the gauges.
• Quarterly pay block: captures past and expected pay growth plus rise versus fall shares, not part of the inflation headline.

In the May 21, 2015 release, the headline rose 0.2 points to 3.6% from 3.4% in April. At the same time the share expecting 0-5% fell 2.2 points to 71.5% from 73.7%. The mean inside 0-5% fell to 2.4% from 2.6% on fewer answers clustered at 5%. That mean had stayed near 2.5% for 12 months, so readers should check all 3 numbers together.

1. Sample: Interview about 1200 Australian householders by telephone each month.
2. Ask expectations: Record each householder's expected rate of price change over the next 12 months.
3. Drop don't knows: Remove all don't know answers before any math.
4. Headline trimmed mean: Order valid answers, trim 30% of extremes from both tails and average the middle 70% with weights. As an illustration, 20 answers trimmed to the middle 14 can give 4.1% while the mean inside 0-5% is 2.7%.
5. Inside 0-5%: Compute the share of valid answers in 0-5% and the mean of answers in 0-5%.
6. Quarterly pay: March, June, September and December releases also give past-12-month and expected-12-month pay growth with rise versus fall shares.
7. Weights and breaks: Results use survey weights, but no weighting scheme is published. No seasonal adjustment step is described. Level comparisons across mid-2014 mix method change with economics. March 2014 read 2.1% with no trimmed-mean label, while June 2014 read 4.0% with the new label.