24 OCTOBER 2011CONTRIBUTOR
Shane LeeSenior Rates Strategist
+61 2 9226 4632Shane.Lee@anz.com
STRATEGY AROUND THE Q3 CPI RELEASEINTEREST RATE STRATEGY
In this note we look at the risks around the CPI using data published by Bloombergon economists’ expectations for the quarterly increase in the trimmed mean of theCPI basket.At the end of each quarter, economists build bottom-up forecasts of the CPI usingdata from private sector data vendors on such items as automotive fuel, rents, fruitand vegetables, utilities etc. Using this approach it’s possible to get a good handle onprice movements of around 50%of the basket. However, it leaves the unknownprice movements (the remaining 50%) to be determined by an econometric modeland this is probably where there is most room for error.
We find that if automotive fuel prices rise and rise by more than 1.5% thenthere is a 77% probability that the trimmed mean outcome will exceedexpectations. Conversely, if fuel prices fall and fall by more than 1.5% thenthere is a 43% chance that the outcome will be below expectations.Data for Q3 suggests that fuel prices fell by 1% in the quarter. Thehistorical relationship between actual and expected trimmed mean outcomesand fuel prices implies that it is more probable that economists’ expectationswill be met.
The Bloomberg survey indicates that economists are expecting a 0.6% rise in thetrimmed mean that would put the YoY rate at 2.7%, up from a revised 2.5% in Q2.Indeed, this should be sufficient for the RBA to consider a modest 25bps rate cut atits 1 November Board Meeting. We expect the trimmed mean measure of consumerprices to rise by 0.5% in the quarter. The range of forecasts is from 0.5-0.8%.
OUTCOMES VS ECONOMISTS’ EXPECTATIONS
Bloomberg have published data on actual underlying CPI results against economistestimates since Q3 2006. On average, economic forecasters under-estimatedinflation in both 2007 and 2008 as it continued to climb well beyond 3% (Figure 1).This trend continued throughout 2009 and in Q1 2010. In the remainder of 2010,they over-estimated inflation, while in the first half of 2011 it was under-estimated.Of course the ABS revisions have changed the actual outcomes over thisperiod and most notably the Q2 2011 result. However, these revisions weremade using a new seasonal adjustment process that was not used prior toQ3 2011 and should not impact our analysis of actual verses expectedinflation outcomes.