Tuesday 30 June 2009

Jim Mora on the BERL report

Afternoons with Jim Mora (Radio NZ) covers our critique of the BERL report. The podcast is available here (link updated). They invited Ganesh Nana on to defend himself. The BERL segment begins at 15:40 or so. My running commentary is below.

At 18:00, the panelists identify the nub of the critique as that BERL tallies all the costs, but includes no benefits. That's not quite how we'd put it. Even leaving all of the benefits issue to one side, BERL overstates costs by almost $2 billion through what we view as errors in method: inappropriate use of multipliers, double counting, and so on. The biggest portion of the difference between our figure and BERL's is their counting of internal costs as social costs; however, even leaving that to one side, there are big problems in the rest of the report.

At 19:05, Nana rightly notes that BERL was not asked to tabulate benefits. But that's somewhat beside the point. There are two basic approaches you can take for a study like this. Either count only externalities, in which case benefits really don't matter, or properly weigh internal benefits against internal costs if you want to weigh internal costs. They chose to count internal costs, but not to weigh against internal benefits.

19:30: Nana complains that his report has been criticized out of context for not having done something it wasn't meant to do: weighing benefits. But we weren't the ones that started down that route. Geoffrey Palmer of the Law Commission mistook the BERL report for a welfare study: not really his fault, since the BERL report conflates welfare with costs in numerous places. But when I called BERL on this very early on and asked that they warn Palmer that their study wasn't suited to that purpose, the principal author of the report said it was none of his concern how their study was used. So it's mighty rich to complain now that it's being taken out of context: they were happy for it to be used out of context by the Law Commission. The second sentence of our executive summary notes that BERL was asked to look at costs; we also note at page 10 the various reasons BERL gives that we should actually ignore benefits. Here's one: "We assume that it is irrational to drink alcohol to a harmful level and that harmful alcohol use has zero private benefit." p. 173 of the BERL report. BERL may have been asked to look only at costs, but they've gone further and explicitly set benefits to zero.

19:40: Nana claims we've complained his study wasn't peer reviewed. Actually, we've claimed that they've pretended to have a peer review: it was reviewed by the authors of its method.

20:49: Nana says adding up benefits is a murky issue. Of course it is; that doesn't mean it can't be done though. As a first cut, get total consumption for each of wine, beer and spirits, get a measure of the price elasticity of demand for each, assume linearity, and calculate the size of the triangle. I'm not saying that that's the best way to do it, but it's a lot better than assuming it to be ZERO.

21:00: Nana goes on again about "rational consumers", etc. I've demolished this argument before. He's also offering BERL's services if anyone wants to measure benefits. I'd be a bit nervous about giving them that contract.

21:51: Internationally accepted methodology. Ahem.

22:23: Nana agrees that it's net harm, not gross harm, that matters. But he didn't see fit to correct Geoffrey Palmer when the Law Commission read his report as measuring net harm. And we can't blame Palmer for that reading: BERL says "net" all over the place; it's just that they're netting out very minor things, and "net" means different things in different places.

22:30: Jim Mora nails the problem: there's no sense in producing a report that measures only costs.

I'll look forward to tomorrow's NZAE meetings; perhaps BERL will there offer some more substantive defense than they have so far.

No comments:

Post a Comment