Monday, 18 November 2013


I've pretty weak priors about AgResearch's restructuring. The government's looking to move some of the work currently done at Invermay up to Lincoln. Dunedin Council and the Otago Regional Council don't like the plan, as expected: whether or not the proposal does net overall good, it's not good for Otago.

Dunedin and Otago commissioned BERL to analyse the business case. I've searched around online for it and haven't been able to find it. It's not in the Dunedin Council press release. I can't see it on the BERL site. A Google site search of BERL for "AgResearch" doesn't turn it up. Similarly, searches for BERL and AgResearch only turn up newspaper reporting on it. So I've not been able to read the piece or evaluate it.

Stephen Joyce dismissed it without reading it:
A spokesman for Mr Joyce responded to questions about the report by hitting out at the quality and independence of Berl.
''The minister stresses he hasn't had the opportunity to read the report but was interested in comments about robustness because in his experience in the five years he has been a minister he has had concerns about the robustness of some of Berl's analysis.
''It's also difficult to describe Berl as 'independent' because again, in the minister's experience, they tend to write reports which uncritically support the views of those who commission them,'' the spokesman said.
The way the report had been presented and comments from Mr Cull suggested the report was ''another part of an ongoing campaign to influence the outcome of AgResearch's deliberations''.
BERL protests their independence:
Mr Joyce last night stood by his comments, adding that in his experience Berl was an advocate for ''preserving the way things were, regardless of the real economics of changing circumstances''.

Mr Joyce's original criticisms came after Berl - which was contracted by the Dunedin City Council and the Otago Regional Council - questioned AgResearch's case for restructuring and slashing about 80 jobs at Invermay.

Mr Joyce through a spokesman told the Otago Daily Times on Friday - despite not having by then having read the report - that he had previously had concerns about the robustness of some of Berl's analysis work and that it was difficult to describe Berl as independent because it tended ''to write reports which uncritically support the views of those who commission them''.

Berl executive director and chief economist Ganesh Nana on Saturday said he was ''disappointed and saddened'' by Mr Joyce's comments.

''We ... totally dispute the assertion that we are not independent or that we write reports to clients' order,'' Dr Nana said.

''That's something that we don't do, but if we did do that we wouldn't have lasted 55 years,'' Dr Nana said.

He also took issue with the minister questioning the quality of Berl's work.

''I'd like to know which [analysis] he is referring to, because I know quite a few that may not have come across his desk, but have gone across some other ministers' desks and they have been quite comfortable with them.''

Berl had a good reputation and he stood by the findings in the report.
I don't discount that BERL could be right this time. There's been some reporting that prior internal AgResearch documents supported strengthening Invermay.

But I'm not sure that Joyce is wrong; instead he seems to be Bayesian. Consider the prior history here.

  • Auckland and Bay of Plenty wanted somebody else to pay for fast broadband rollout. BERL reported that dairy production in the Bay of Plenty would receive a boost equivalent to seven percent of the export value of the region's dairy product. They also reckoned the Auckland region's GDP in 2025 to be about $1000 per capita higher if broadband rolled out in 2012 instead of 2015. Neither seems at all plausible. 
  • The Rail & Maritime Transport Union and City of Dunedin wanted to show how great it would be for the country if KiwiRail were directed to ignore profitability and do all its rolling stock construction in Dunedin. BERL told them that the government would wind up being $65-70 million net to the good. Again, pretty implausible. Or, if you buy that case, why don't we pay Holden to run its car manufacturing in New Zealand? The subsidies they're getting in Oz are starting to seem a bit unsustainable. Maybe it'd work here.
  • Labour and the Greens wanted to show that their NZPower plan would be great. BERL told them that the economy would get a $450m annual boost, so long as we assume an economy with deficient demand such that there are unemployed resources that can be drawn into production without opportunity cost. If you find the right set of starting assumptions, you can come up with the right answer. Question is whether that's the set of starting assumptions that best models reality over the requisite period of analysis, or the set that tells the folks commissioning the study what they want to hear. 
  • BERL told NZ First that the biggest cost of dropping inflation targeting would be the loss of face for "mainstream" economists. Maybe there's a case for NGDP targeting over inflation targeting, though I'm not convinced of it in a small open economy. But that wasn't what NZ First was going for. 
  • The Canterbury DHB wanted to know just how terrible alcohol is. So BERL told them, after assuming, at footnote 14, that there were zero savings from moderate alcohol use: they specifically set equal to zero the coefficients in their source aetiological table for conditions where alcohol reduced costs rather than increasing them. They also did this in their earlier report on the social costs of alcohol for MoH and ACC, more fully critiqued here.
  • The Greens wanted to know just how awful asset sales are. So BERL told them just how costly it would be. Assume that revenues from sales go towards building other assets that will provide an equivalent dividend yield but that time-to-build means a few years' delay in getting the new stream. Surprise surprise, selling the assets is terrible. See comment above about choosing one's assumptions. 
It isn't hard to look over the track record and come to a position like Minister Joyce's. 

And where substantial parts of the current government are expected to hold this opinion of BERL's work, one wonders who the ultimate customer of their reports is meant to be. Even if you believed their reports to be of the highest possible quality, if the Minister can dismiss the reports without reading them, and if you have reasonable reason to expect that he would, well....


  1. Once upon a time in NZ, we had a severe rabbit infestation problem. The country wanted to get rid of rabbits and paid people a sum of money every time they brought in a rabbit tail (or pair of ears) to those in charge of the scheme. This did not work, hunters did not want to remove their income entirely and some even, illegally, breed rabbits. Only when the bounty scheme ended and people were paid to hunt rabbits did the problem diminish. If this little history has any relevance it is probably that a market in elephants will probably help keep them from extinction. But equally probably a good thing if illegal hunters and importers can be prevented from profiting from what they do.

  2. This doesn't show anything. You're never going to see BERL's reports which go against their clients' interests because their clients never release such reports. So you can't establish that they're biased by pointing at all their reports which go in favour of their clients, as they're not a representative sample. All you're establishing here is that they do a lot of reports.

  3. Entirely possible that there are tons of reports they've done that have never surfaced. But do note the method critiques in those listed above.

  4. I completely agree that to save us we have to hunt it and that’s what I do in trading. I find been aggressive as the best policy to attack. If I stay defensive and try not to do anything then sooner or later I will end up with losing. I use OctaFX broker with which I can use their rebate service and earn up to 15 USD profits per lot size trade so even with saving I get plenty of rewards with them.