Economist v nutritionists

March 7, 2012
The Age

Moneyball, the successful book and movie, showed how an economist’s feeling for statistics turned a professional baseball upside down. Now an Australian economist’s examination of the numbers destroys the local sugar lobby’s key defence against linking fructose to obesity and diabetes.

The sugar industry is a big fan of what self-described “economist and former fattie”, Rory Robertson, calls “the low-GI crew” – a high profile group of Sydney University nutritionists who promote the health benefits of food with a low glycemic index and downplay, if not completely dismiss, claims that fructose is a prime suspect in our obesity and diabetes epidemics.

The low-GI crew is about as high profile as academic nutritionists can get: Professor Jennie Brand-Miller, AM, author of the Low GI Diet book; Bill Shrapnel, Sydney University Nutrition Research Foundation deputy chairman; and Dr Alan Barclay, the Australian Diabetes Council’s head of research.

The cornerstone of their defence of sugar is what they have termed “the Australian Paradox” – the claim that Australians’ sugar consumption has fallen by 23 per cent over the past 30 years while obesity and diabetes has soared. Thus, they argue, sugar must be innocent.

There are others who claim sugar is guilty as hell, with none arguing the case against sugar most forcefully than David Gillespie, lawyer and author of Sweet Poison and its sequel Why Sugar Makes Us Fat. Faced with Gillespie’s theories, Professor Brand-Miller cites the Australian Paradox. “That to me blows David Gillespie’s hypothesis out of the window (sic),” she says.

Want a quote attacking those who attack sugar, ring the low-GI crew and you’ll get the Australian Paradox.

But what if there is no Australian Paradox? What if Australians’ sugar consumption has been rising and the low-GI crew’s key statistic is simply wrong?

Enter Rory Robertson, unaware there was an academic debate raging when he came across Gillespie’s book, cut fructose from his diet last May and lost 10 kg without any extra exercise. As a believer then through personal experience, he subsequently found the Australian Paradox more than a little strange, applied his economist’s training to dig into the source of the nutritionists’ statistic and now charges that it is not true.

Taking aim

In a research paper (done on his own time, not that of his bank employer) circulated to friends and colleagues for comment, Robertson takes issue with several aspects of the low-GI crew’s defence of sugar:

“My main concern, however, is the low-GI crew’s unreasonable treatment of the available data on Australian sugar consumption. Its regular claim – “In Australia sugar consumption has dropped 23 per cent since 1980” – is woefully misleading, based as it is on a series that was abandoned by the Australian Bureau of Statistics (ABS) as unreliable a decade ago.

“Last year, Dr Alan Barclay and Professor Jennie Brand-Miller lifted the status of the “it’s not sugar” story a couple of notches, publishing an academic paper that concluded: “This analysis of [i] apparent consumption, [ii] national dietary surveys and [iii] food industry data indicates a consistent and substantial decline in total refined or added sugar consumption by Australians over the past 30 years”.

“The low-GI crew then declared an ‘Australian Paradox’ in the relationship between sugar consumption (down) and obesity (up). Unfortunately, the paper’s conclusion is largely at odds with the available facts on Australian per capita sugar consumption.
“Bizarrely, the low-GI crew seems somewhat unaware that its own charts illustrate clearly that the longer-term trend in measures (i) and (ii) is up not down… the available national nutrition surveys show per capita “total sugars” consumption rose not fell for both adults (between 1983 and 1995) and children (between 1985 and 2007). Second, per-capita soft-drink consumption rose not fell over the available 1994-2006 period.”

Robertson says the paper 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 (Population data in table 4) over the past two decades has been up, from near the bottom of a 40-60kg range to the top of that range in 2009-10.

Apparent consumption

But the big figure in this argument, the cornerstone of the Australian Paradox, is the “apparent consumption” number. What Robertson found after some digging and questioning of the Australian Bureau of Statistics, is that:

“The “apparent consumption” series on which the low-GI crew’s strong conclusion is based (1980-2003) simply was downloaded from the website of the Food and Agriculture Organization of the United Nations (FAO). The low-GI crew may or may not be aware that the downloaded series from the FAO’s website actually was produced by the ABS for decades, until it stopped counting after providing estimates for 1998-99.

“Anyone familiar with the ABS would be aware that it is rather unusual for it to stop producing a dataset that already spans 60 years, particularly when the topic was becoming more rather than less relevant.

“The low-GI crew either remains oblivious to this data dead-end, or simply chooses not to mention it. Either way, it’s hard to say anything useful about “the past 30 years” when the ABS stopped even pretending to measure of sugar consumption after printing an estimate for 1998-99, some 12 years ago…

“You probably guessed that the ABS didn’t give up counting sugar after 1998-99 because it couldn’t find any. The problems began when it came time to add imported sugar to domestic “sugar availability”. Discussions with the ABS confirm that it struggled to know how much sugar was in the rapidly growing imports of things like bakery products, confectionary, soft-drinks, cordial and syrup, processed fruit and vegetables, and “other processed foods”.

“Someone smart might be able to guess how much sugar is in those imports. I have no idea. And neither does the low-GI crew. Most of what we do know is that: (i) all the available refined sugar– about 60kg per capita per year after exports – in 2009-10 had a total worth less than $1 billion at wholesale prices, while (ii) total food imports in 2009-10 amounted to about $10b at retail prices, with imports of sugary items growing pretty strongly. You tell me how much sugar was in those imports? All I’m suggesting is that there’s room for quite a bit…

“Breaking the available 22 years of this dataset into two halves, we find that “sugar availability” was 22% higher in the second half than in the first half. Up 22%, not down 23%!

“Clearly, this chunky rise in “sugar availability” over the 22 years to 2009-10 should have made the low-GI crew question its view that sugar consumption has fallen substantially over “the past 30 years”. That’s especially the case when it almost certainly had seen David Gillespie’s chart for the past half century, showing elevated sugar availability for over much of the period but a particular “soft patch in the mid to late 1980s.

“In summary, and contrary to the inaccurate claims of the low-GI crew, there appear to be no reliable or timely data series showing a significant decline in per capita sugar/fructose consumption over “the past 30 years”.

“The “true” trend in sugar consumption over recent decades remains uncertain but the available evidence – from (i) the two-decade uptrends in sugar availability and sugary imports; (ii) national dietary surveys and (iii) industry data on soft-drink sales – suggests that if anything it’s more likely to be up than down significantly, as claimed.”

Robertson, fresh from winning his high-profile bet against Professor Steve Keen over housing prices, wants to donate $10,000 to a health department or non-conflicted university to help fund a definitive experiment to compare the effect on obese people of a no fructose diet, a low GI diet and a control group eating their normal intake. He is not a scientist and says there is more science to be done – but he does know his way around a set of statistics.

Michael Pascoe is a BusinessDay contributing editor – who has a love of fructose-laden dark chocolate.

Read more