Why I, too, Am Skeptical of Market Failure Corrections

Jon Murphy recently posted an explanation for why he is skeptical about the use of taxes to offset market failures. His reasoning was that the use of tax policy will inevitably be distorted by political incentives, and such incentives may not be at all aligned with what is socially beneficial.

I agree this is a major issue. One of my favorite recent explanations of this problem came from Scott Alexander. Alexander used the example of how in theory, taxes and subsidies could be used to nudge people into eating a healthier diet. But Alexander then goes on to note:

After providing numerous examples of the kinds of subsidies and restrictions that result from the political process as it actually exists, Alexander concludes “Given our existing government, it shouldn’t be let within a light-year of getting to determine anybody’s diet. Speculating that maybe the people who administer the program will be virtuous competent individuals who act for the good of the public, is saying that the thing which has already happened won’t happen.”

But there’s another reason why I am skeptical of this approach, one that holds even if we assume away all problems about political incentives. But first, here’s a (seemingly) random digression – what is the impact of time-restricted feeding on how much people weigh?

Time-restricted feeding (also known as intermittent fasting) is a somewhat popular method people use to help lose weight. Time windows vary, but the most common method is called the 16:8 method, where one goes 16 hours between eating, and consumes all their food during the remaining 8 hour window. A person who does this might skip breakfast, wait until noon before they consume anything with any calories, and then eat between noon and 8pm. Then, they’d wait until noon the next day to start eating again. 

Diet and nutritional studies are notoriously difficult to carry out and often have very divergent findings. But there was a really interesting meta-analysis that looked at the effect of time-restricted feeding among Muslims who observe Ramadan. This is the practice of fasting between sunrise and sunset which, as the study notes, could be a fasting window of between 9 and 22 hours depending on how far one lives from the equator. It’s also a practice observed by hundreds of millions of people, which gives a much better sample size than most nutritional studies. 

So, what effect does this have on people’s weight? The answer is “all of them.” It has every possible effect on people’s weight. Some people who observe Ramadan fasting lose weight, others maintain their weight, and others actually gain weight. Some people lose weight because restricting the time they have available to eat leads to them consuming fewer calories than they otherwise would. On the other side, some people approaching the end of their fasting window find themselves in a physical state known, to use a technical term, as being “insanely fricking hungry” and will gorge themselves when their eating window begins, ultimately consuming more overall calories than they would have if they had just eaten throughout the day. And for others, these two effects basically balance out and their total caloric intake remains unchanged. As the meta-analysis put it, “Effects of Ramadan fasting on weight vary between individuals, ranging from weight loss to weight gain, depending on whether or not energy intake in the non-fasting period under- or over-compensates for the lack of energy intake during the fasting period.”

So what does this have to do with the use of taxes and subsidies to offset market failure? Well, the use of such taxes and subsidies implicitly assumes that people will respond to taxes or subsidies in a specifically predictable and desired way – and people can in fact react in all kinds of different ways to taxes and subsidies, just as people’s total calorie intake can respond in every kind of way to time-restricted feeding. 

One famous example of this is the cobra effect. As I have described it before:

Culling the cobra population was judged to produce positive externalities, and was thus judged to be underprovided on the market. Policymakers subsidized the killing of cobras because they anticipated it would lead to an increased amount of cobra hunting, thus offsetting the market failure by increasing cobra culling to a socially optimal level. But people reacted differently than policymakers anticipated. Instead of cobra hunting, people began cobra breeding. So the attempt to use subsidies to decrease the cobra population had the opposite effect. 

But is this just some isolated case? Or is there reason to believe that the inability to predict the specific ways people will respond to taxes and subsidies is the rule rather than the exception? In my extended review of Jeffrey Friedman’s book Power Without Knowledge, I outlined an argument Friedman made that this issue is the rule rather than the exception, and why this undermines the arguments made by economists with technocratic aspirations, who imagine they can skillfully guide behavior across society by just using taxes and subsidies to create the “right” incentives. Friedman argued that “incentives alone cannot actually produce behavioral predictions or, therefore, policy advice.”

This, Friedman argues, is because “knowing that the perceived incentive will affect these agents’ behavior is useless—for predictive purposes—if the economist does not also know exactly how it will affect it. But this requires knowing exactly how agents will interpret their situations in light of the perceived incentive. Only if they interpret their situations the way the economist does will the incentive ‘matter’ in a way the economist will be able to predict.” But, as Friedman goes through great pains to argue, different people perceive things in different ways and think in different ways, which means the way people will respond to any given incentive will be variable and unpredictable. As a result, economists (and policy makers more generally) lack “the ability to predict future agents’ subjective interpretations of how to behave under future circumstances as the agents themselves will perceive and interpret them.” 

A book length demonstration of this very issue is Scott Hodge’s recent book Taxocracy: What You Don’t Know about Taxes and How They Rule Your Daily Life. It’s a fairly fun and breezy read. Hodge outlines all kinds of examples over the course of centuries where people’s responses to taxes – responses that were not anticipated by the policymakers levying the tax – have created all kinds of unanticipated outcomes. Some of them are merely amusing, like how some old houses in France are built rather like mushrooms – relatively small and narrow first floors with wider floors above. Homes were built this way because “property taxes were based on the square footage of the land a house occupied. So, people cheated on the tax collector by designing a small ground-floor level and wider stories above it.” 

But other times the results are less amusing and more disastrous. King William III instigated a tax on windows, on the assumption that dwellings and buildings with lots of windows were likely to be owned by the wealthy, and thus this would serve as a way to tax the rich. However, 

Granted, in neither of those cases were the taxes passed as a means of correcting a market failure. But the fundamental problem – that people will react to taxes (or subsides) in all kinds of ways that you can’t predict – is just as true whether the taxes (or subsidies) are meant to correct a market failure or are simply for the more generic purpose of raising revenue.

Friedman argues that this undercuts the arguments in favor of technocratic policy – including the use of taxes and subsidies to alter behavior in a way that corrects market failure. Friedman wrote “if we have reason to think that we cannot accurately know the results of a certain action (such as a specific technocratic action), then our knowledge of the beneficial outcome of taking that type of action cannot serve as the rationale for it, as technocracy demands, since we lack such knowledge. Likewise, if the defender of technocracy concedes that it is likely to produce unintended consequences but allows, too, that she does not know what they are likely to be, then her putative knowledge of the beneficial results of technocracy (the prevention, alleviation, and solution of social problems) cannot serve as the rationale for it, for she lacks knowledge of what lies on the cost side of the ledger.” 

So even without politically misaligned incentives (a very real problem in its own right) there is another problem with attempting to use taxes and subsidies to correct market failure. Because, paraphrasing Friedman, if we have reason to think we cannot accurately know the specific ways people will change their behavior in response to Pigouvian taxes or subsidies, and I think we do in fact have good reason to think this, then the claim that the taxes or subsidies will alleviate a market failure cannot serve as the rationale for that policy.