The Consumer Price Index: Updated and Revised

In case you missed it, last month’s CPI numbers from DOS came with some changes.

First, the basket weights and components were updated to reflect 2014’s Household Expenditure Survey. This included adding 32 items, and removing 12. Weight changes were also made – as I suspected might happen, the drop in crude oil prices resulted in a reduced weighting for transport costs. Utilities costs (rent, water, electricity and gas) saw the biggest increase in weighting.

Second, the approach that DOS had taken previously was to rebase the CPI series to the new base year (2015 in this case). Instead, this time they’ve added to their methodological arsenal by chain-linking the new series to the old 2010 series. Essentially, they’ve saved researchers a lot of time and grief by splicing the new and old series together, so we don’t have to.

Third, they’ve added a core inflation measure. This had some cleverness embedded as well, as instead of leaving out whole categories, they’ve removed specific items instead – those with highly volatile prices, and those with administered prices:

1. Meat
2. Fish
3. Seafood
4. Eggs
5. Coconuts and nuts
6. Vegetables
7. Potatoes and other tubers
8. Spices
9. Fresh fruit
10. Cooking oils
11. Flour and other cereal grains
12. Sugar
13. Alcoholic beverages and tobacco
14. Water
15. Electricity
16. Gas
17. Fuels and lubricants for transport
18. Transport services

That directly addresses not just price volatility, but also those criticisms that the headline CPI didn’t reflect “true” inflation because it included prices of goods that were fixed by the government. Personally, I’ve never thought the latter was a valid argument, but this is a good response.

Does the base year and other revisions change the CPI significantly? Nope – the difference between the old and new weights amount to about half a point, or about 0.03%. Still and all, it satisfies some of the itch that the statistics reflect changes in consumer preferences and prices more accurately – to be precise, 0.03% more accurately.

As for how inflation is doing, it’s zooming at the moment (log annual and monthly changes; 2000=100):

Headline inflation hit 3.5% in January, up from 2.6% in December 2015. The Pain Index rose 2.7%, after just 0.8% in December, while (my) Core inflation measure rose a blistering 4.5% in January and 4.6% in December.

Some people (thankfully very few in the industry), are seeing this as proof that inflation is high and rising. Actually, what’s really telling here is the month on month growth rates – they’re all negative. January’s elevated growth rates are all a product of the base effect, from very low petrol prices last year, as well as the impact of GST on consumer prices.

It’s easy to see from looking at the raw level data (index numbers; 2000=100):

See that dip at the beginning of 2015? That’s the base from which the January growth rates were calculated. As we get further along in this year, inflation numbers will drop as the denominator returns to its “normal” path. Once we get past April (and the anniversary of GST implementation), CPI inflation will return to its “normal” trend growth rate. Much of the price increases of the past year have been due to petrol price volatility, and one-off changes to prices (like GST and the hike in alcohol and tobacco duties).

In point of fact, if petrol prices stay anywhere around today’s level, I wouldn't be surprised to see sub 2.0% inflation for the last three quarters of the year.

Technical Notes:

January 2016 Consumer Price Index report from the Department of Statistics