Feb 22 2010
The Economics of Snake Oil
One of the common questions (sometimes framed as a claim or justification) about unscientific and implausible treatments that frequently comes up is – if they don’t work, why are they so popular? The assumption (also made by some who oppose regulation of medical products) is that the consumer will perceive the medical value of products and adjust demand accordingly.
I and others who favor more scientific and effective regulation of health products and claims have argued that there are a host of factors distorting the market in favor of health products with appealing claims, even if they lack substance. I am not an economist, however. So it was heartening to read the very same arguments I have been making expressed from an economist’s point of view.
This paper by Werner Troesken, an economist from the University of Pittsburgh, explores the flourishing of the patent medicine industry from 1810 to 1939 in the US (when FDA regulation essentially shut it down).
Troesken explores many factors leading to this success, but what I found most interesting is that when you strip away all the complexity, Troesken shows how ineffective treatments will still flourish in the marketplace.
Inelastic demand with respect to product failures
The core concept of this paper is that the patent medicine industry demonstrated what Troesken calls “inelastic demand with respect to product failures” – in other words, people still wanted to buy patent medicines, even after their previous experience with such products resulted in failure.
He begins with a few reasonable premises. The first is that the patent medicines of the time did not work. We can say with historical hindsight that the products popular at the time were not effective in curing the diseases they were being used to treat – often serious illnesses that were common at the time, like tuberculosis, but which lacked any mainstream medical treatment. Most patent medicines were variations of the same product, in fact – a bit of alcohol, sugar, and some (essentially random) vegetable matter. Often the same products were marketed over and over again with different labels and claims. We can now say with a high degree of confidence that these elixirs did not cure tuberculosis, or seizures, typhoid, or any other ailment that was claimed. Troesken assumes a 100% failure rate for patent medicines.
Troesken also assumes that customers perceived this failure – they knew they were not cured of TB. For a bit more of background on this premise, however, Troesken explains that at the time the empirical model of assessing medicines was dominant. In other words, treatments were assessed purely on whether or not they worked, not on the plausibility of any putative mechanism. This is largely due to the fact that science had not yet progressed sufficiently to think productively about plausibility.
It is a very interesting side note that empiricism resulted in very slow and uneven progress in medicine. Scientific medicine began to improve dramatically, however, when it became increasingly founded upon basic science. I would add, however, that the technology of empiricism also improved – we got better at doing clinical experiments. I would argue that this debate continues today – that between pure empiricism (which functionally is represented by evidence-based medicine) and a more balanced mixture of basic and clinical science, which I and others advocate as science-based medicine.
In any case, Troesken argues that consumers were following an almost pure empirical approach to patent medicines. They did not know enough science, and not enough science was known, to assess the plausibility of the products, so they judged them entirely on whether or not they perceived a benefit.
This is Troesken’s basic model – medicines do not work, consumers judge them solely on whether or not they work, and consumers correctly perceive that they do not.
Interestingly (and I think this is the core insight of this paper) even with this basic model, Troesken argues that the patent medicine market would still flourish – as it in fact did. This is because (here is where the economics perspective comes in) the perceived value of a potential cure is extremely high, compared to the relatively low cost of experimenting with a new treatment. So consumers felt that they had little to lose and the world to gain, leading to repeated experimentation with, and even high demand for, patent medicines.
Troesken acknowledges that repeated failures would decrease the perceived value of patent medicines, but this perceived value would not go to zero – it would decrease asymptotically, and remain substantially high even after multiple failures.
In other words – the allure of a promised cure is so great, people are willing to risk the investment in the treatment even after repeated failures. We therefore cannot rely upon market forces to weed out ineffective treatments from the marketplace.
False positive and other factors
But of course – this is the most basic model. There are many other factors at work, and Troesken hits upon many of the ones I and my colleagues have discussed in the past, but not all of them. For example, he discusses how marketing was designed to produce “false positives” – this led to the testimonial, a fake patient giving a fake testimony about how the patent medicine worked for them. Consumers then incorporate the false positives of testimonials into their calculations, making the product demand even more resistant to personal failures.
Another potential source of false positives was actual active ingredients present in many patent medicines. Products often contained cathartics, purgatives, or even narcotics (and of course alcohol is an active ingredient). These substances may have caused dependence, and in fact some patent medicines may have been little more than a socially acceptable “back door” method of consuming illicit drugs. Or they may have produced physical effects that convinced consumers something biological was happening, such as removing toxins through purgatives.
What Troesken does not discuss much is the false positive of the placebo effect. This is because he is restricting his analysis to claims for cures of real diseases. In his basic model he ignores spontaneous remissions, but he acknowledges this is another source of false positives – sometimes people get better on their own. But the more serious the disease, the more rare this should be.
When we move into the arena of symptomatically treating chronic ailments that are not life-threatening, the placebo effects takes on a larger role. For example, many illnesses are self-limiting – they will get better with only tincture of time, and so any treatment will seem to work. Also, people with a fluctuating illness will tend to seek treatments when their symptoms are at their worst, and so regression to the mean will result in improvement after any treatment they take while symptoms are at their peak. There are also a host of other psychological factors leading people to perceive improvement in the absence of a real biological effect. (See here for more information on placebo effects.)
One factor I think Troesken gets a bit wrong is his conclusion that patent medicines thrived not only despite their lack of effectiveness, but because of it. He argues that if a patent medicine worked, and cured a client, that consumer would be removed from the pool of demand – they would no longer be seeking medicine. This is true (again within the limited model of patent medicines as cures for serious illnesses, rather than symptomatic treatments for chronic illnesses) but he ignores an important factor. If a person survives one illness, they will survive to get another – and when they do, they will likely return to what cured them the first time. Everyone is relentlessly aging, and with age comes increasing illness. Ironically, Troesken does accurately refer to this effect in a different context below.
The rise of science-based medicine
Troesken also argues that improvements in the knowledge base and effectiveness of scientific medicine did not, and would not be expected to, reduce the patent medicine market. The success of snake oil, in other words, should not be interpreted as a failure of or the absence of effective medicine. In fact, Troesken argues, the success of germ theory and increasingly scientific medicine meant that people were surviving the acute illnesses of their youth to become older adults – older adults with chronic ailments. Modern medicine, therefore, increased the pool of demand for patent medicines.
He also points out, and I would amplify, that increasing medical knowledge did not dampen the demand for implausible treatments. This is because the general public, while they may have increased scientific knowledge as scientific knowledge in general increased, would not rise to the point that they would reject patent medicines on the basis of scientific plausibility. They remained empiricists.
This matches our knowledge of belief in pseudoscience even outside the realm of medicine. Acceptance of dubious science actually increases with education, until one gets to the post-graduate level of science education. It takes a very high level of scientific knowledge (and/or, I would argue, an understanding of the principles of scientific skepticism) to feel confident in rejecting a claim based upon plausibility alone. This is true even of health care professionals.
Patent Medicines Today
The classic era of patent medicines may have ended in 1939, but it has had a resurgence in the US since 1994, when DSHEA was passed. DSHEA essentially removed supplements and herbs out from under the yoke of FDA regulation, partly (at least on the right) based upon the claim that market forces were sufficient to regulate such products.
What is amazing to me is how the supplement industry of today is so similar to (and in some cases historically continuous with) the patent medicine industry of the 19th century. All of the same factors are in place – testimonials, cutting supplements with active ingredients (caffeine is now common, and of course some Viagra replacement herbal supplements were found to be cut with actual Viagra), the types of claims that are made for products (combining ancient wisdom with cutting edge science), and more.
I think Troesken makes a compelling argument for the inelasticity of demand for such products, which, in my opinion, undercuts some of the libertarian arguments made against regulation of supplements, and even (by some) pharmaceuticals. Some may still argue that the demands of personal liberty outweigh the goal of protecting the public from worthless products, or that regulations, despite good intentions, either do not work or have negative unintended consequences. I disagree (even while I am sympathetic to these arguments in principle, I disagree with how they are often applied in this context), but these arguments are not addressed by this paper.
What this paper convincingly argues is that we cannot expect market forces to result in better and more effective health products, or even to keep entirely worthless or even harmful health products from the marketplace.