Few people enjoy paying taxes, but as Oliver Wendell Holmes Jr. reminded us, “Taxes are what we pay for civilized society.†On reflection, most of us therefore offer at least implicit support for penalties against tax evasion — penalties that have little meaning unless backed by significant enforcement resources.
Yet prodded mainly by anti-tax Republicans, Congress has cut the Internal Revenue Service budget steadily since 2011. By 2019, the agency was auditing only one in every 222 individual returns, down from one in 90 in 2011. Similar reductions have occurred for corporate returns, and were proportionately larger for the wealthiest individuals and largest corporations.
These cuts have not saved the government money. The former I.R.S. commissioner John A. Koskinen estimated, for example, that every $1 trimmed from the agency’s budget has resulted in $4 in lost revenue. But this estimate refers only to direct, or first-round, losses. Because the extent to which people comply with tax laws depends strongly on the behavior of others around them, the ultimate revenue losses are certain to be much larger.
The influence of peer behavior on tax compliance is nicely illustrated in a thought experiment I call “the waiter’s dilemma.†It assumes that workers face a choice between working in a factory or waiting tables, jobs they consider equally attractive aside from the matter of pay. Factory workers get a weekly salary of $1,000, while waiters get $500 in salary plus another $500 in cash tips. If the income tax rate is 20 percent and waiters declare their tips, both occupations have after-tax weekly pay of $800.
Now suppose that although waiters are legally and morally obligated to declare their tips, the I.R.S. has no way of monitoring them. Waiters who don’t declare tips then get after-tax pay of $900 a week, a $100 premium that would induce at least some factory workers to seek employment as waiters.
If we assume for simplicity that foreign competition prevents factory wages from rising in response to the resulting shortage of factory workers, the increased supply of waiters will cause servers’ wages to fall until this premium for tax evasion disappears. At a waiter’s weekly wage of $375, servers who didn’t declare tips would get $800 after taxes, the same as in factory work.
The moral dilemma confronting honest waiters is immediately apparent. If they declare their tips, their weekly after-tax pay will be only $700, or $100 less than they would have taken home as fully tax-compliant factory workers. Under the circumstances, they might reasonably conclude that paying taxes on their tips would be unfair.
This example illustrates why pressures to cheat in social settings are even stronger than those that would arise in isolation. We see many examples in which people resist temptation even in the face of golden opportunities. But such opportunities become far more difficult to resist when one’s peers are seizing them without penalty.
A creative study by the economists Jörg Paetzold and Hannes Winner sheds light on the extent to which peer behavior influences tax evasion. It employs data from Austria, where the government allows workers to deduct commuting expenses from their income for tax purposes.
Under this allowance, workers report their total commuting distances, and it is then their employers’ responsibility to certify the accuracy of their reports. But because many employers devote few resources to verification, misreporting entails little risk of punishment. Combining detailed tax data with employer and worker location data, the authors found that the claiming of excessive commuting expense deductions was in fact widespread.
By itself, that result is hardly surprising. What was fascinating was the extent to which peer behavior influenced the exaggeration of deduction claims. To estimate that magnitude, the authors focused only on people who had moved to new workplaces during the current tax year.
Those who were now employed by a company where tax evasion was either the same or less common than in their previous workplaces showed no change in their own rates of tax evasion. But the pattern was strikingly different for workers who had moved to a new job in which tax evasion was more common. For these workers, it grew by an even larger proportion than the one by which their new colleagues’ tax evasion exceeded that of their former colleagues.
In addition to direct observation of peer tax evasion, behavioral contagion amplifies incentives to cut corners in other ways. It does so in part by making people less likely to be chosen as partners in transactions that require trust.
Assessments of trustworthiness depend not only on information about you personally — such as your reputation and others’ evaluations of your character — but also on their perceptions of the trustworthiness of people in general. For any given set of personal characteristics, for instance, you would be less likely to be deemed trustworthy in a population in which only 10 percent were scrupulously honest than in one in which that proportion was 90 percent.
The upshot is that quite independently of any change in your own behavior, any change that makes others more likely to cut corners also makes others less likely to judge you to be a trustworthy potential partner. That, in turn, reduces your own payoff from being trustworthy, which increases your incentive to stray, and so on.
What the reductions in I.R.S. funding will continue to unleash, then, is a characteristic feature of all behavioral contagion processes: an explosive chain of feedback loops that greatly amplify any initial change in behavior.
The United States was once firmly a member of the small group of countries whose high levels of tax compliance helped sustain the infrastructure investments needed to support broad economic and social prosperity. In a 2004 study that ranked 30 industrial countries and territories on a six-point tax-compliance scale, for example, it was in seventh place with a score of 4.47. Highest on the list was Singapore at 5.05, followed by New Zealand (5.00), Australia (4.58), Britain (4.67), Hong Kong (4.56) and Switzerland (4.49). Last among the 30 was Italy, with a score of 1.77.
Since that study was published, reduced I.R.S. funding has led to significant reductions in tax compliance in the United States, the Treasury Department reports. But these reductions are only the beginning. It takes time for people realize the extent to which others are evading taxes. And once that happens, compliance will fall much more rapidly.
Robert H. Frank, an economics professor at Cornell University, is the author of “Under the Influence: Putting Peer Pressure to Work†(Princeton, 2020). Follow him on Twitter: @econnaturalist
