I can address that question.
You're quite correct, we are already making distinctions in the GST rules as they are. With regard to your side point, I guess, about the nanny state, in this particular context we're talking about children. Children have nannies, and that's probably a good thing.
But on the general question of the GST, there aren't any studies done. There are some data available that can be used to forecast the effect of the GST on shipments of soft drinks, and those data are actually quite interesting. When the GST came into effect in 1992, it actually lowered the tax on domestically consumed soft drinks because they had previously been subject to the manufacturer's sales tax, which was 13.5%. Domestic consumption went up a little bit, but exports went up quite considerably because they were no longer subject to tax whatsoever.
Although the GST revenue data are only collected on an aggregate level, the officials at the Department of Finance could do a fairly good estimate of the impact by looking at historic A.C. Neilson sales data, which can be broken down by product category. They could just look at the rules and see which ones would be subject to tax.
With regard to the question about how we could develop nutrition criteria to make the distinctions, there's a professor at the University of Oxford in the United Kingdom who was contracted by the U.K. Food Standards Agency to come up with what they call a food scoring system. I think it's a very promising technique. What he did was assign foods to three levels of nutritional quality: very helpful, mediocre, and low.