Pesky economist won’t let Big Sugar lie

Michael Pascoe
July 25, 2012
The Age

·The debate of Australia’s sugar consumption is heating up.

·Rory Robertson’s bets are getting bigger.

·Having successfully wagered Doomsday forecaster Steve Keen a walk to Mt Kosciuszko over Australian house prices not crashing during the GFC, he’s punting $40,000 that Big Sugar’s favourite academic paper is wrong.

For hounding Peter Costello over being Australia’s biggest taxing Treasurer, Robertson once was described favourably by Ross Gittins as “that pesky Mr Robertson” delving deep into the statistics to prove his case against Costello’s protestations.

Robertson is proving at least as pesky in his passion for questioning Australia’s fondness of sugar.

Taxation or sugar consumption, it’s all a matter of understanding what statistics are credible to an economist, albeit one with a personal belief that sugar is a sweet poison.

What makes a sucrose fixation a business story is the size of the Australian sugar and sugar-dependent packaged food and drink industries and their fight to keep advertising regulations and health warnings at bay, never mind the health industry and the costs of our obesity and diabetes epidemics.

Robertson is putting $40,000 of his own money up for grabs in a wager aimed at settling his fight with what must be Big Sugar’s favourite academic paper. In the process, the argument has been escalated into questions about the academic standards of the University of Sydney in general and of the Nutrients e-journal in particular.

Sweet paradox
The Australian Paradox study by Sydney University’s Professor Jennie Brand-Miller, author of the Low GI Diet book, and Dr Alan Barclay, the Australian Diabetes Council’s head of research, claimed that Australians’ sugar consumption had fallen by 23 per cent over the past three decades while obesity has soared.

Big Sugar has been quick to claim that the study therefore clears sugar of being a cause.

Robertson argues that the Australian Paradox paper is flawed with key statistics proving either unreliable or, when they didn’t support the authors’ thesis, ignored, as previously reported here.

Challenge
Having failed to win any concessions from Brand-Miller and Barclay or the Nutrients journal that published the paper, Robertson took the fight to the university:
“On 7 June 2012 in a letter to University of Sydney Vice-Chancellor Michael Spence, I challenged the University’s scores of fine scientists – indeed, any scientist, nutritionist, medical doctor, economist, journalist or enthusiastic observer anywhere – to prove that my critique of Australian Paradox is mistaken.

“I wrote: “To be clear, I will reward the first successful researcher with $20,000 (cash), if anyone is able show beyond dispute that the available (valid) information really “…indicates a consistent and substantial decline in total refined or added sugar consumption by Australians over the past 30 years”, as concluded in Australian Paradox. Moreover, I will pay a further $20,000 to the charity of choice at the University of Sydney’s low-GI school, and publish a genuine public apology in The Sydney Morning Herald, The Australian and The Australian Financial Review. “

So far, there’s no sign of anyone trying to win the money.

Returning fire
Professor Brand-Miller and Dr Barclay accuse Robertson of factual errors and “misinterpretation of the distinctions between total sugars vs refined sugars, sugar availability vs apparent consumption, sugar-sweetened and diet soft drinks, and other nutrition information. The terminology, strengths and limitations of various nutrition data are readily understood by individuals trained in nutrition.”

Yet in their rebuttal of Robertson’s attack, Brand-Miller and Barclay failed to make much of a case on the central issue of the reliability of sugar consumption statistics and were simply wrong in their “hunch” that led them to ignore another set of statistics that ran counter to the Australian Paradox finding.

In the third key area of dispute, the interpretation of nutrition surveys, each side seems capable of reading data differently.

ABS factor
The lynchpin of the Australian Paradox case rests on the use of United Nations Food and Agriculture Organisation (FAO) statistics which showed a fall in apparent sugar consumption, but Robertson delved further to find that FAO was relying on an Australian Bureau of Statistics survey that had been discontinued a dozen years ago because the ABS thought it was unreliable.

So if it’s not good enough for the ABS, it questionable that it could be good enough for academic nutritionists to use in a matter with important public health implications.
In the second key area of dispute, Robertson says the Australian Paradox did not mention that the only timely official (ABARE) information on Australia-wide “sugar availability” (production less exports) also suggested the trend over the past 22 years had been up, not down.

“The trend in domestic “sugar availability” per capita has been up, from near the bottom of a 40-60kg range to the top of that range in 2009-10,” he wrote.
Ethanol mix-up

After BusinessDay published the original story in March, Brand-Miller sent me a reply to Robertson’s argument. That reply put the “sugar availability” discrepancy substantially down to sugar being used to make fuel ethanol:

“Sugar availability takes no account of food wastage, use in animal food, beer and alcohol fermentation, or in non-food industrial use, and we cannot assume that a steady portion is lost in this way. Globally, raw sugar is an important ingredient for ethanol production. In Australia, ABARE data show that ethanol production as a biofuel for transport rose from 42 million litres to 209 million litres (almost four-fold) from 2005 to 2009.”

A footnote added that the increase in ethanol production would require about 14 kg of sugar per capita per year if 100 per cent raw sugar was used to make it. “Although there are no firm figures for how much raw sugar is presently being used for ethanol production, supplies of C-molasses alone are not adequate, and the absolute amounts are likely to be increasing,” wrote the academics.

There’s a good reason why there are “no firm figures” – sugar is not used for ethanol production in Australia, as the most cursory of Google searches on Australian biofuels would show.

Fuel ethanol here is produced from red sorghum and waste products from sugar and starch production.

I told the Professor I thought she was wrong, she checked and admitted that was the case. Having failed on two of the three key issues with the jury out on the third, I didn’t bother about the reply.

In the Nutrients e-journal, Brand-Miller and Barclay published their reply to Robertson under the title, Australian Paradox Revisited with the ethanol bit deleted.

(It’s only complicating already complicated matters to point the common knowledge Australian per capita beer consumption has been falling for years, so I won’t.)
I passed on the ethanol exchange to Robertson and he did bother about the reply, including it in his own blog on the controversy.

It looks like the Robertson/Keen debate and bet over house prices was a much more straight-forward and measurable battle – but not as potentially rewarding.
Michael Pascoe is a BusinessDay contributing editor.

Read more: