Statisticians in bit of a twist over revised consumer price index

PUBLISHED : Tuesday, 03 May, 2011, 12:00am
UPDATED : Tuesday, 03 May, 2011, 12:00am


Compared with the old CPI series, the new CPI series generally shows smaller year-on-year rates of increase. This is because when the prices of various goods and services change, households tend to buy more of the goods and services with relatively smaller price increases ... to substitute those with larger price increases ...

Government news release, January 28

It happens every five years, right on schedule. Our statisticians conduct a survey of household expenditure patterns, adjust the weights of the consumer price index for it and, well, stripe me pink, inflation is lower than we thought it was.

We're there again. The change has just been made and whereas the previous calculations yielded a year-on-year rate of inflation in March of 4.57 per cent, we now get only 4.38 per cent from the new and improved way of doing things. Could have gone either way, you know, but it just happened to be lower this time. Funny how that happens.

If you say so, the statisticians will come out and tell you that it makes perfect sense. You naturally buy less of things that have gone up most in price. We're all bargain hunters, aren't we? The new weighting arrangement only reflects this.

Except that they seem have got themselves into a bit of a twist about it. In the very same news release in which this justification is made we have the acting commissioner for census and statistics, Lily Ou-yang, saying that a higher expenditure weighting for housing is 'mainly attributable to the general rise in rentals for private housing over the past five years'.

So in this case people buy more, not less, of things that have gone up most in price. Hmmm ... interesting.

And then she does it for food, too. A higher weighting for food in the total consumer basket, she says, is 'due to the increases in both prices of various basic food items and the costs of meals bought away from home'.

Very well, we can understand this. A roof over your head and food on your plate are necessities of life. If their prices go up you can't just say that you'll do something else with your money. You're stuck. You have to pay.

But then it's rather odd that our statisticians should invoke the argument that you buy less of what costs more when it does not apply to food and housing, which together now constitute more than 59 per cent of the total consumer basket. Want to have a rethink on this one, fellas?

It becomes even odder when you see that they have taken the weighting for durable goods down to 5.27 per cent of the basket from 5.5 per cent before.

Unlike food and housing, it's probably true that if the price of a television goes way up you would wait a little longer to buy a new one. Conversely, if that price goes way down you are more likely to rush to the shops for one of those big, flat screen jobbies.

And prices of durable goods have gone way down since the last household expenditure survey, almost 30 per cent down relative to the rest of the consumer basket. That should push its weighting up, not down. What happened, Mrs Ou-yang?

It brings to mind the way that five years ago our statisticians revised the weighting of jewellery down to 0.28 per cent of the basket from 1.53 per cent before.

There was nothing in the retail sales figures to suggest such an enormous contraction of spending on geegaws but inflation in jewellery prices at the time was running at more than 20 per cent. How convenient that this suddenly did not matter so much anymore.

It seems a little unusual that they could not have used some of the same magic this time to give durable goods a higher weighting and thus push down the official inflation rate even further. It would have been legitimate this time too.

Very puzzling indeed, but perhaps the first thing our statistics department needs to do is learn to get its story straight. The weightings go up because the price goes up, no, because the price goes down except when the price goes up and then the weightings go down and then ...

No, let's try that one again. When the price goes up, the weighting goes ... goes ... up except when it goes down ... and ... hmmm ...