By their nature, public policies have always contained an element of paternalism—the notion that policymakers more than citizens know what is best for the society and, perhaps, the individual. Although some paternalistic approaches are motivated by religious or moral beliefs, and progressive thought itself has always had a strong foundation in paternalism, the growing popularity of behavioral economics has sparked new, seemingly objective, rationales for policies that aim to reduce the incidence of activities that policymakers have deemed “wrong” for society. Taxes on tobacco, alcohol, sugar, and marijuana are only four of innumerable examples of the state’s efforts to “nudge” citizens toward “better” choices.

But vice taxes illustrate the pitfalls and failures of that approach to public policy and to taxes in particular. One common trait of efforts to tax tobacco, alcohol, sugar, and marijuana is that policymakers generally failed to account for the likelihood of both avoidant behaviors and the use of substitute goods. When taxes on tobacco increase, smokers become more efficient users, purchase cigarettes in lower-tax jurisdictions, buy products on the black market, or reduce spending on other goods. In response to alcohol taxes, consumers produce their own, purchase in lower-tax jurisdictions, or switch to lower-cost brands with equivalent alcohol content. In response to taxes on sugar-sweetened beverages, consumers substitute with other high calorie but untaxed drinks or with food high in fat and carbohydrates. Efforts to legalize and tax marijuana with high excise taxes have had the effect of keeping the market underground. Making it harder to consume alcohol can increase marijuana consumption and perhaps vice versa.

Policymakers’ response to these well-supported patterns is rarely to undo the taxes and other policies that caused them; instead, they tend to favor higher taxes and more regulations, which in turn create even more unintended consequences. Ironically, the externalities policymakers seek to reduce are often the result of their own actions—i.e., other public policies.

What all efforts to tax vices have in common, especially from the progressive movement to today, is the widespread inclination of policymakers consigning too much faith in their own ability and that of “experts” to improve civil society over and above what would have occurred in the absence of their interventions. Like any attempt to engage in social engineering, using the tax code to shape behavior betrays four critical mistakes.

First, there is no guarantee that policymakers and the experts on which they depend possess complete, accurate information from which to draw inferences about human behaviors, much less to model the underlying data completely and appropriately, and much less to draw implications for tax policy. Economists Rizzo and Whitman argue that to make better policy, policymakers must have six different types of knowledge about consumer preferences at both the individual and population level, a requirement that is impossible to meet. Merely collecting more data does not necessarily generate knowledge, nor is it a useful approach. Underlying measurements may lack both accuracy and precision, and social scientific laboratory experiments and econometric modeling cannot hope to duplicate the complexity of the real world. It should thus come as no surprise that academic research on the efficacy of vice taxes is best described as noncommittal. Some studies suggest the taxes “work,” others find the opposite, and others conclude that it depends. In short, so-called experts producing the associated research have covered all the possible outcome angles. It should also not come as a surprise that replication problems and a spike in retractions have recently plagued the social and physical sciences, which should—but likely will not—increase skepticism of their use in policymaking.

Second, policymakers and experts are not immune from the very cognitive biases possessed by the ordinary citizens they imply are inferior. Common sense dictates that this simple fact should compel policymakers and experts to question their own abilities to direct tax policy toward socially optimal ends. But surprisingly, 96% of studies in top behavioral economics journals that recommend paternalistic policy interventions do not question policymakers’ cognitive abilities or biases. This comes despite mounting evidence to the contrary. State policymakers in particular have a strong tendency to imitate actions taken in other states, the so-called bandwagon effect. Such decisions may be motivated by competitive concerns, but psychology also plays a role. Policymakers interpret the popularity of a policy in other jurisdictions as a proxy for success and seek to emulate that success, an overreaction that creates policy bubbles. Policymakers also respond to emotional manipulation. On the other side of the coin, theory and research suggest that policymakers have an intellectual reluctance to oppose policies they previously supported and may support inefficient policies out of fear that changing positions will significantly harm their likelihood of reelection.

Third, scholarly biases impose their own problems. Analyses of vice taxes frequently ignore substitution. Once substitution is factored in, researchers often find that the efficacy of vice taxes declines and that the taxes may actually do more harm than good. Still another analytical bias is that of ignoring the full costs and benefits of policy changes. Perhaps the best example is ignoring the “benefit” of having smokers die sooner, and hence saving health care and social security costs, or the “cost” of having a healthier population—i.e., health and entitlement costs are higher.

Fourth, policymakers and the experts on which they depend are economic creatures that respond to incentives, and the incentives they face are not tied to improved social welfare. Elected officials have one primary goal—reelection—and in a political system heavily influenced by money, reelection comes not by improving social welfare, but by creating the illusion that you have. Academic researchers are beholden to a hyper-competitive grant system; insufficient grant support could mean a loss of prestige or, worse, job loss. The incentive is to do whatever is necessary to win grants, and that often means tailoring research findings to the clients’ desires. Competitive pressures also drive faster research and data collection, which can have the ripple effect of poor measurements and insufficient modeling of econometric data.

Given all of the above, why do vice taxes persist? Taxes on tobacco and alcohol have historical momentum: tobacco and alcohol. Both taxes benefit from lingering moral sentiments against consumption and, having been on the books for over a century, the likelihood of their termination or reduction is low. Over that timeframe, taxes on both have become self-sustaining. The taxes support large bureaucracies tasked with collecting them, and the members of those bureaucracies, as economic creatures, generally resist proposals that may reduce their prestige or jeopardize their job security.

But vice taxes are entangled in countless other bureaucracies, where they provide the funding to maintain programs for favored constituencies. Perhaps the best example is the use of tobacco taxes to subsidize the Children’s Health Insurance Program or other policies that mandate vice tax revenues be allocated toward public education. Because these programs have broad, emotionally favorable constituencies, and the burden is believed to fall on a much narrower segment of the population engaging in unfavorable behaviors, it is not surprising that vice taxes have lasted as long as they have. Consider the difficulty of trying to balance “the children” against “big tobacco,” “big sugar,” or “big alcohol.”

Finally, vice taxes invite rent-seeking from multiple parties. Proposals to increase or expand vice taxes invite heightened lobbying activity, just as they do for any other consumption tax. Disparate parties may seek or lobby in favor of vice taxes not for any public health or moral reason, but because the taxes are a convenient source of revenue that avoids having to reduce budgets—and thereby reduce the state’s role in individual affairs and civil society.

This commentary is an excerpt from the recently published Tax Politics and Policy by Michael Thom, an Assistant Professor at the USC Price School of Public Policy. Text reprinted with permission by the author.