Last week was our spring break, and I finally took some time to read Gary Taubes’ 2008 book Good Calories, Bad Calories. Taubes is an investigative science journalist who has been writing for years about the science of nutrition and epidemiology, and the book focuses on a long, careful, detailed narrative about how such science has evolved since the mid-19th century. One of the themes that emerges is that some of the most prominent researchers, particularly those advancing the dual hypotheses that fat causes heart disease/overeating causes obesity, did not test their hypotheses for falsification using controlled trials in designing their research, and are also personally invested in doing research that “proves them right”. Thus, Taubes argues, an orthodoxy has formed around these hypotheses when he finds the scientific support for them lacking, and similarly finds support for an alternate hypothesis — refined carbohydrates cause heart disease and obesity. But the orthodoxy resists testing that alternate hypothesis.
I have personal interest in this topic based on my own experience. As a high metabolism athlete for all of my life, I grew up being able to eat almost anything in unrestricted quantities. But when I got my first faculty job out of grad school (at WIlliam & Mary, yay!) in 1992, the combination of teaching and research duties with moving to a swampy climate against which my body rebelled meant a reduction in my activity, bloating because of the humidity, and weight gain. Without really thinking about it (because I hadn’t had to before), I reduced my meat consumption and substituted into (refined and unrefined) carbs. The next two years were right out of Taubes’ book — reduction in calories to manage weight while increasing exercise, but not having enough energy to actually make it meaningful, culminating in what is now known as metabolic syndrome complete with insulin resistance, hormone imbalance, and symptoms of polycystic ovarian syndrome. I then spent two years revamping my diet to reduce refined carbs, include more animal and vegetable protein at every meal, and monitor my hormone and energy levels, and succeeded in reversing all negative symptoms. I returned to the energy levels that have enabled me to do longer and longer distance cycling and triathlon endurance events and the demanding training for them. Even though I don’t eat low-fat, my triglycerides are so low that my doctor marvels at it. Taubes’ argument is consistent with my experience.
Economist Russ Roberts has been experimenting with his diet and exercise for the past six months, following broadly the same principles that I do (including the refined carbs on the weekend), and he reported in on Friday: 20 pounds lost, more energy, feeling of satiation, low triglycerides. Again, consistent with my experience.
You may know Russ for his outstanding EconTalk podcast series, and in November 2011 he interviewed Gary Taubes. The conversation was interesting and informative, and the podcast page lists lots of resources for further reading. One theme that Russ developed in the discussion was that in both nutrition research and economics research, the issues come up of orthodoxy and structuring research questions in ways that generate falsifiable hypotheses when you are studying such a complex, dynamic system as either the human diet/cardio/endocrine system or the human economy. The human traits that incline us toward orthodoxy, whether it’s wanting to prove ourselves right or appeal to authority or some other trait, have led to models and hypotheses that are not supportable or not even meaningfully testable/falsifiable. So for me reading Taubes’ book was a good cautionary tale of the value of humility beyond the analysis of low-carb/low-fat nutrition.
Another insight that comes up in the book that I would add to Russ’ comparison with macroeconomics is heterogeneity. Taubes is careful to point out that individuals have different metabolic experiences and achieve homeostasis with different combinations of fat, carbs, etc., so while low-carb nutrition may allow some people to strike a healthy heart and weight balance, others may be able to eat more carbs and do the same. Heterogeneity means that there’s no one-size-fits-all hypothesis … and as any Austrian macroeconomist will tell you, that’s the argument they put forth about macroeconomic models and aggregation. Heterogeneity in the capital structure in reality means that models abstracting from such heterogeneity are more likely to mislead.