Sunday, June 01, 2008
You might be about to discover that you were wrong on both counts.
The Competition and Consumer Commission is likely to recommend a compulsory national system of “unit pricing” as a result of its grocery inquiry. Family First has introduced a bill for such a scheme into the Senate, and Coles has told the inquiry that even without legislation it’ll go ahead on its own such scheme within the twelve months. Aldi already does it.
What will you find when every item for sale has a card in front of it saying how much it costs per litre or kilo? Surpises.
Here are a few uncovered by Family First:
The milk in a 3-litre carton of Pauls Milk is actually more expensive per litre than the milk in a 2-litre carton.
The fruit salad in a Coles 1kg can of is actually more expensive per kilo than the fruit salad in a 825g tin (an astounding 32 per cent more expensive in fact)...
And the best value Kleenex toilet paper is often the 6-pack rather than the 8-pack or the 12-pack.
We think we’ve got prices sussed, but many of us take a lazy mental shortcut. We assume that the big volumes are always cheaper just because they usually are.
When a team led by Gordon Mills of the University of Sydney examined supermarket prices in 1995 it found that in a full 9 per cent of cases, the largest volumes carried the largest rather than the smallest unit prices.
It happened the most for toothpaste, canned meat, flour, snack foods and tissues.
Those products were often packed in odd sizes, perhaps in order to make the unit prices hard to calculate. Toothpaste was packed in tubes of 70g, 110g, 140g and 175g.
For most brands almost all of the surcharge occurred on the biggest toothpaste tube, rather than the second-biggest. The biggest tube was in fact only 25 per cent larger than the next biggest, but it was hard to tell that without getting out a calculator.
In his resulting book Retail Pricing Strategies and Market Power, Gordon Mills made an intriguing suggestion about why the most expensive flour, tissues and so on were to be found in the biggest rather than the smallest packets.
It was that the people who could afford to buy the biggest packets were well-off enough not to care about properly calculating the price. They only thought they did.
Joshua Gans writes:
Peter Martin thinks that the one thing we might see from the ACCC’s grocery inquiry is unit pricing for groceries. He points to the fact that often larger volumes of food products actually sell for more per unit than smaller ones. Buying a larger one would be a bad idea as you could just buy a requisite number of smaller ones and be better off.
OK, so consumers might prefer that information. Even if you can calculate per unit volumes not having to do so would be a time saver. So why don’t supermarkets generally offer them?
When you try to think of reasonable explanations — integer constraints, packing costs or what have you — it only increases the rationale for some transparency so that consumers feel the firm’s cost pain and accommodate it.
Thus, we are left with the strategic: put simply, it looks like an innumeracy tax. Those who can’t or can’t be bothered calculating may end up paying more than those who can. But even that doesn’t get us the whole way there as an explanation: those very same people who can’t also have to be the people who are relatively price insensitive. Now, the notion that the rich can’t be bothered is attractive but given how widespread this practice is, I suspect those who can’t, can’t also easily check between supermarkets as to whether they are paying more for their whole basket of groceries.
So what happens if we mandate unit pricing? Let’s assume that eliminates the innumeracy tax. The problem is that supermarkets will no long have a means of targetting price sensitive customers and sorting them from the insensitive ones. In a relatively competitive market, prices on some products might rise. And sorting out the average seems tricky (although not impossible with a formal model — something I don’t have time to do here).
In any case, it is not entirely clear whether a policy to mandate unit pricing would improve matters; although ideologically, I am in favour of more transparency if it can be cheaply done; i.e., assuming supermarkets themselves can calculate unit prices! Of course, after FuelWatch, no such policy could surely be introduced without a clear policy trial or experiment and some careful econometric analysis. After a week of Parliamentary debate on statistical significance, our political standards on these things appear to have shifted.
Also: Alan Kallir, Public submission to ACCC Grocery Prices Inquiry, March 11, 2008