An economic perspective of obesity and seed oils
Since I tend to write quite a bit about health topics, I thought some of my readers might be interested in hearing about my latest research paper, which I posted to my website yesterday.
In it, I combine economics with several ideas from biochemists and medical doctors about what is driving obesity and other metabolic diseases like heart disease, diabetes, dementia, and cancer. The main finding is that vegetable seed oils seem to play a big role, and that the incentive structure of the US economy reinforces consumption of seed oils.1
What the data shows
The main story in the data can be told in a few graphs:
Over a 60-year horizon, the rise in obesity doesn’t correlate so much with increased calories in the food supply or declining rates of physical exercise.
Obesity also doesn’t correlate much with sugar supply (particularly since 2000) but does correlate strongly with the supply of vegetable seed oils.
Obesity rates do correlate strongly with the supply of seed oils in US food
How did seed oils become prominent?
Seed oils in food didn’t really exist before 1900. The original seed oil was cottonseed, which began being extracted for various industrial uses starting in the mid-1800s. However, it wasn’t suitable for human consumption.
Industrial innovation
There were two main innovations that happened to make it possible:
Deodorization
Hydrogenation
With these two processes, cottonseed oil could be turned into a solid fat which could be substituted for pork lard or beef tallow or butter at a fraction of the price. In 1911, Crisco—“crystalized cottonseed oil”—was born.
The diet-heart hypothesis
Seed oils really started to take hold in the 1960s, after several influential papers published by Ancel Keys and his collaborators implicated saturated fat as a cause of heart disease through elevated levels of cholesterol in the blood. The American Heart Association in 1961 recommended replacing saturated fats with unsaturated fats because unsaturated fats reduce blood cholesterol levels. Shortly after, soybean oil consumption shot up (see Figure 1 of this open-access academic paper).
The race to eliminate saturated fat and the rise of trans fats
With American culture focused on heart disease, food scientists began perfecting ways to use hydrogenation to create solid-like fats out of unsaturated fats (seed oils). Partial hydrogenation, in particular, resulted in a fat that was not fully saturated, but which mimicked the properties of saturated fats. Unbeknownst to everyone, partially hydrogenated seed oils were even more dangerous than saturated fats were thought to be. Unfortunately, this wasn’t discovered until decades later.
The fall of trans fats and the rise of new applications of seed oils
The heyday of trans fats was roughly 1975-2000. By the mid-2000s, the damage of trans fats was well recognized, and companies started shifting away from partial hydrogenation to come up with other ways of getting solid texture out of vegetable oils. The result was a continued rise of seed oils in the 2010s.
How do seed oils promote obesity?
To answer this question, we need to know a bit more about how metabolism works, as well as a little bit about a fatty acid known as linoleic acid.
How metabolism works
Growing up in the 90s, I was taught that weight gain was caused by eating more calories than one burned. One problem with this idea is that calories are a measure of how much energy it takes to literally consume (oxidize) a food. But the human body doesn’t do this to the food we eat. Instead, the body has to convert food calories to adenosine triphosphate (ATP), which is the energy currency of each cell in the body.
ATP is produced in the mitochondria of the cell and can be generated from the breakdown of fat, sugar, or amino acids (components of protein).
Thus, if we want to know why the body isn’t burning the excess fat it’s carrying, we should be asking why the mitochondria are choosing to generate ATP from sugar or amino acids instead of from the fat that is sitting right there.
Polyunsaturated fats and oxidation
There are three types of fats: saturated, monounsaturated, and polyunsaturated. And within polyunsaturated, there are omega-3 and omega-6.2 Polyunsaturated fats have two or more double-bonded Carbon atoms in their molecules. This makes them more prone to oxidation, which can create free radicals.
When we eat polyunsaturated fats (PUFAs), some of the fat gets used to create ATP, some of it gets put into cell membranes, and some of it gets stored in the extra layer of chub that we wish we didn’t have.
One particularly disruptive PUFA is the omega-6, linoleic acid. It is easily oxidized and is plentifully found in all of the seed oils. Linoleic acid causes problems in mitochondria because of its propensity to oxidize.
Dr. Cate Shanahan in her book Dark Calories hypothesizes that linoleic acid causes insulin resistance and diabetes because it forces the body to use sugar instead of fat to produce ATP. This results in reduced blood sugar levels (hypoglycemia) which kicks off an “arms race” between the liver and pancreas, resulting in chronically elevated blood sugar levels. See the figure below for more details:
Omega-6 PUFAs and appetite disruption
Anthony Hulbert in his book Omega Balance also points out that omega-6 PUFAs like linoleic acid also disrupt the “endocannabinoid system” in the body, which is how appetite is regulated, among other functions.
The incentive web reinforcing seed oil consumption
This is where economics comes into the picture. If seed oils are indeed bad for our health (this is still a contentious topic), it will be difficult to eradicate them due to how the incentive system is currently set up. Here’s a picture showing why:
Information
There are several interconnected incentives that all tilt toward increasing levels of seed oil consumption. Foremost is the problem of information. Scientists publish research papers, but they need funding for their research. So they have to get research grants from the government, or secure funding from industry sources. This means that the research that scientists do will be influenced by what the government wants researched, or what food and drug companies want researched.
Government
Government nutrition recommendations are also influenced by lobbying activities of food and drug companies, as well as the research findings that scientists publish.
Food companies
For food companies, seed oils present substantial cost savings over traditional (pre-1900) forms of fats, since they can be scaled up in production and can be produced in more uniform batches. Oilcrop seeds also have more uses than just seed oils—some of the byproducts of seed oil production can be fed to animals, used as emulsifiers to improve food textures, or even turned into biodiesel.
Consumers
For consumers, seed-oil-based products are cheaper and more convenient. Food companies are able to make products with seed oils that have improved shelf life and improved taste and texture. Moreover, seed oils are vegan/vegetarian/Kosher/Halal and so appeal to the least common dietary denominator. All of these factors add up to consumers preferring seed oils to traditional fats.
Drug companies
Drug companies play an indirect role in the incentive web depicted above. They fund research and lobby the government, and they also have influence over medical education. All of this adds up to a distortion in medical advice that favors pharmaceutical remedies to traditional remedies. Moreover, medical billing is structured such that doctors are highly incentivized to perform procedures over counseling with patients about how to best implement lifestyle changes.
How do we reduce seed oil consumption?
Given the web of incentives above, it will take a multifaceted effort to curb seed oil consumption. Here are a few potential ideas:
Require PUFAs to be on every food label, just as trans fats and saturated fats currently are. And make sure to require separate line-items for omega-3 and omega-6.
Define what “healthy” food is. Right now, the USDA’s “Healthy Eating Index” penalizes saturated fat consumption and rewards unsaturated fat consumption, but doesn’t distinguish between monounsaturated, omega-6 PUFA or omega-3 PUFA.
Require more rigorous testing of food formulations before they are widely introduced into the food supply. We don’t want a repeat of trans fats with some new food industry innovation that we don’t know the full ramifications of.3
Adjust the current levels of taxes and subsidies on seed oil crops and foods using seed oils. However, food taxes are always tough because everyone needs to eat, and you don’t want to end up taxing the poor at higher rates than you tax the rich.
Treat seed oils like alcohol or tobacco and put age/location restrictions on their use. There’s lots of evidence that seed oils are especially bad for children.
Difficult tradeoffs
In the end, there seem to be some difficult tradeoffs between cost-effective food production and human health. Moreover, returning to traditional fats means consuming more animals, which brings up questions about environmental sustainability and animal welfare. However, Casey & Calley Means in their recent book Good Energy argue that “regenerative farming” is more cost-effective than factor farming once one takes into account all of the subsidies and negative externalities of the status quo.
Hope in technology?
I am hopeful that the age of artificial intelligence can enable greater health sensing, which will give consumers more information about how their behaviors are affecting their health. If seed oils are indeed a major problem, this influx of information would reduce demand for seed oils and put pressure on food companies to “give the people what they want.”
What if it isn’t seed oils that are causing obesity?
The evidence I’ve presented is circumstantial, but won’t convince harsh critics. There are plenty of other theories out there for why obesity has exploded; see Slime Mold Time Mold’s blog series on the subject.
Seed oils certainly don’t explain everything, and I’m not claiming that they are the sole cause. Rather, I’m analyzing the economic incentives driving seed oils, noting that they seem to be playing an outsized role in obesity and that they also seem to be able to explain the rise in diabetes, dementia, cancer, etc. I discuss this in section 5 of the paper for those interested.
Vegetable seed oils refer to industrially processed vegetable oils that require chemical extraction. These include soybean (#1 most common in US), canola, corn, cottonseed, sunflowerseed, safflower, grapeseed, and ricebran oils. Olive oil and avocado oil are notably not considered to be seed oils. Peanut oil may or may not be considered a seed oil, depending on who you ask.
You might have heard of omega-3 fats being in fish oils and good for brain health. Omega-3 fats are associated with antioxidants, while omega-6’s have been shown to be pro-oxidative. See Anthony Hulbert’s book Omega Balance for an in-depth review of the nuances between the two.
I don’t feel great about Bill Gates’ recent blog post where he explains “why I’m making big bets on novel fats and oils.” But at least he understands the importance of fats and oils.