Glyphosate and Behavioral Economics

It's hardly possible to turn on the news, to talk to a farmer, or to visit a nursery without hearing something about glyphosate, the active ingredient in many of the world's most popular and effective herbicides like Roundup, Pronto, Eraser, and 750 other products. If you've ever heard of these products, you've probably already heard a thousand times that the science shows no risk associated with glyphosate, despite a single committee from the World Health Organization listing it as a probable carcinogen, and despite prominent court cases awarding billions of dollars to cancer victims with the finding that the cancer was caused by glyphosate. So we're not going to rehash that same "science vs. pop culture" argument again. Instead we're going to look at the science behind how this misinformation spread, to the point that a huge proportion of people in the world believe a harmless product is harmful. The result is very likely to be that we will see more bans of glyphosate, requiring farmers to revert to the older, less effective, and actually-dangerous herbicides. The story of glyphosate misinformation is the perfect example of what I like to call the Endarkenment: the modern rejection of science and facts in favor of misinformation.

Glyphosate was first registered for use in the United States in 1974, and quickly swept the farming industry. It was safe for humans and wildlife, so there were no environmental concerns about its use; it decomposed into the natural compounds carbon dioxide, phosphoric acid, and ammonia; and it killed virtually every weed in a crop field, reducing tilling and erosion. Farm Chemicals magazine named Roundup one of its "Top Ten Products that Changed the Face of Agriculture", and its inventor Dr. John Franz was inducted into the National Inventor's Hall of Fame. Glyphosate's inherent safety to people and animals came from the fact that it includes no ingredients that are harmful. Its mechanism blocks one very specific metabolic process (called the shikimic acid pathway) that is present only in photosynthesizing plants and some microorganisms, and is not present at all in humans and animals. Glyphosate does not pass through your skin; and if ingested, it does not bioaccumulate and it passes right through without being digested or metabolized. Taken all into account, it is not surprising that forty years as one of the most common herbicides on the planet has resulted in not a single observable health consequence.

This is just what we'd expect to see. As glyphosate only interacts with structures not present in humans and animals, the latest report published in Haz-Map, a service of the US National Library of Medicine, states:

..There [is] no association between glyphosate exposure and NHL (non-Hodgkin's lymphoma)... there is little evidence of biologic effects of glyphosate in humans; therefore, no clear biologic mechanism has been proposed.

As far as the agriculture industry is concerned, glyphosate truly is the bee's knees of herbicides. It is tremendously effective at improving crop yields and reducing costs, and not only has no evidence of harm to humans ever been observed, nobody has even been able to propose a plausible mechanism by which it might.

Then, in 2015, the United Nations' IARC (International Agency for Research on Cancer) added glyphosate to its list of compounds in Group 2A of its classifications of carcinogenic agents. Group 2A, though it's titled "Probably carcinogenic", is defined as meaning there is limited or inadequate evidence of carcinogenicity. It's distinct from Group 1 which is for compounds defined as having sufficient evidence of carcinogenicity. Every other major health and science organization worldwide plus over 800 studies disputed this classification, and criticism came fast: notably, that the IARC had excluded the vast majority of those 800 studies and included only a very few small ones; and that theirs had been a "hazard" assessment rather than a "risk" assessment. Now, to you and me, the difference between hazard and risk may seem meaningless, but the terms are quite distinct in the science of toxicity assessment. The moon Io might be a hazardous place to be; but your risk of dying on Io is basically zero.

And although the IARC is only one of scores of such bodies around the world — and still the lone outlier to suggest that glyphosate might be harmful to some — attorneys were fast to react. They went out and searched for everyone who had non-Hodgkin's lymphoma and who had worked on a farm applying Roundup. Soon there were over 4,000 lawsuits filed against Monsanto, the producer of Roundup at the time (the actual number of lawsuits isn't known; some sources report it as over 11,000). Despite the actual data being unassailable — that as Roundup entered the market and its usage increased dramatically over the years, there was absolutely no corresponding increase in the rates of non-Hodgkin's lymphoma, or any other cancer or any other disease, among farm workers or anyone else — eventually one team of lawyers managed to prevail in court and a worker was awarded hundreds of millions of dollars. The precedent thus established, at least two other cases have been won as of this writing, awarding damages exceeding two billion dollars so far. Juries don't assess science, they choose a winner between teams of arguing lawyers.

As no science or statistics can support the finding, it's likely that all these cases will be overturned on the inevitable appeals. However, the fact remains that the belief that glyphosate is harmful has swept the general public. So now, let's dig into how there came to be opposition to glyphosate in the first place, as by any measure, it's safer than virtually every other effective herbicide ever developed.

It is arguable that most people — including those who are either passionately for or against glyphosate — do not know very much about its chemical makeup, its biochemical mechanism within the weed, or the details of the published research on its safety. And yet these people still cling to a polarized position, often strongly. And clearly, glyphosate is not the only such science question where we observe this behavior. We see it in such topics as water fluoridation, BPA in plastic food containers, the sweetener aspartame, vaccine safety, the list goes on and on — even the Flat Earth. What is the societal mechanism by which large numbers of people form shared opinions on things they know very little about?

One method is what behavioral economists call an information cascade. Behavioral economics is a fascinating field. Where classical economic theory models rational decision making, behavioral economics looks at the human weaknesses that compel us to make irrational decisions. We're all susceptible to biases, emotions, cultural influences, cognitive errors, and a whole raft of factors that result in bad decision making. In an information cascade, people make a simple decision based only upon an observation of someone else making that decision, like dominos toppling in a great big expanding fan.

The way it works is this. A person hears about glyphosate, and that it is somehow controversial. The average person has little to no skin in the game for whether glyphosate is harmful or not, so there is no reason for them to take the time and energy needed to learn how to properly research the question. The cost in time and energy is too high for little meaningful benefit, so virtually nobody goes to the trouble of informing themselves. Instead, they observe the choice made by somebody else. When we see someone else make a decision that we don't know how to make ourselves, we assume they know more than we do; so it's natural for us to adopt their decision. Almost all the noise in pop culture about glyphosate is negative, so the natural and expected result — predicted and well explained by behavioral economics — is that many people will decide that glyphosate is unacceptably harmful, and other people will observe that decision and make the same one. Soon, we see exactly what we observe today: a huge proportion of people who do not understand the science believes glyphosate is harmful.

This is, of course, not a comprehensive discussion of information cascades; there are other factors and provisos governing when and how they occur and are most effective. For example, it's important that each member of the cascade has only limited knowledge of the person whose decision they mimic; they know only that that person made a decision, and have little information about why they made it. To learn more about how information cascades play a huge role in so many popular false beliefs, see the references at the bottom of this page.

The result is that we are now where we are. There are countless consumer groups informed by this information cascade who list glyphosate as carcinogenic, and countless lobbying groups for the organic industry who say the same. Armed with white papers from these groups, plus the single, lonely scientific finding by the IARC, glyphosate opponents now march the Earth, seeking out additional misinformation that confirms their preferred conclusion, and adding it to their arsenal.

Bans on glyphosate are already appearing around the world, and will probably continue. It's important to remember that laws and regulations are not informed by science; they are informed by whatever a bureaucrat was persuaded to do, for whatever reason. Bureaucrats and lawmakers are as susceptible to fallacious information cascades as anyone else; in fact, they are a prime example of it. Their aides advise them how to vote, and they're too busy with other matters to open an investigation into how and why that recommendation was given.

The result of these bans is that farmers are required to go back to older generation herbicides, which have been supplanted because they're either not as effective, too expensive, actually harmful to people and animals, or even all three. One measure of the toxicity of a compound is its LD50: higher numbers are safer, lower numbers are more toxic. Two of the main fallback herbicides for farmers who are not allowed to use glyphosate are Diquat and Glufosinate. Glufosinate has an LD50 of 83; a low number meaning it's quite toxic. Diquat has an LD50 of 120; slightly safer. Glyphosate rings in at — wait for it — 6,800. In other words, the chemicals that the anti-glyphosate lobby wants farmers to use instead are about 100 times as toxic.

This is what happens when we allow emotion-driven policy, or anecdote-driven policy, or popular opinion-driven policy: a worse outcome, in almost every case. Data driven-policy — policy informed by classical economics instead of behavioral economics — is, sadly, all too often a unicorn. Bans on glyphosate are equivalent to "personal conviction" exemptions from vaccination; equivalent to more expensive BPA-free water bottles; equivalent to Berkeley's warning labels on cell phones. At best they misinform; at worst they do real harm.

By Brian Dunning
Follow @BrianDunning