All the rich countries are successful in raising sizable amounts of revenue from taxes with only a rather little tax evasion. Tax avoidance is the use of legal means to reduce taxes, whereas tax evasion uses illegal means. The federal government of the US raises almost 20% of American GDP through taxes on personal and business income, capital gains, estates, and the sale of gasoline and some other goods. The estimates from the 2001 IRS National Research Program indicate that the percent of income not reported is quite low for wages and salaries, but rises to over 50% for farm income, and about 40% for business income. Income tax payments overall are under reported by about 13 percent. What determines the degree of tax evasion?
If taxpayers responded only to the expected cost of evading taxes, evasion would be far more widespread. The reason is that only about 7% of all tax returns are audited (over a 7 year period), and typically the penalty on under reported income is only about 20% of the taxes owed. Virtually no one is sent to jail simply for evading taxes unless that evasion is on a very large scale, or involves massive fraud. If a person were to evade $1,000 in taxes, his expected gain would be 0.93x$1000 -0.07x$200 (=$1000/5) = $916. On these considerations alone, he should not hesitate to evade paying the $1,000, and presumably much more.
To be sure, the expected gain is not the right criterion since most taxpayers would be risk averse regarding audits and punishments, especially if there is some chance of much greater than the average punishment or likelihood of an audit. However, if the expected gain from evading $1,000 were $916, the degree of risk aversion would have to be huge, far higher than the risk aversion that is embodied in pricing of assets, for risk to explain why there is so little tax evasion.
This is not to say that possible punishments have no affect on the amount of tax evasion. Compliance rates are much higher when governments have independent evidence on a person's income since then the probability of audit when he under reports his income is much higher than when they do not have this information. For example, income from independent consulting to companies is better reported than tips on earnings, or than the incomes of farmers and other small business owners because employers report how much they paid to independent consultants, whereas no one reports how much they paid in tips, or how much they bought from a local store. A PhD study in progress at the University of Chicago by Oscar Vela also shows that persons in occupations where integrity is a more important determinant of success, such as law or medicine, are less likely to evade taxes. Presumably, any publicity that an individual in these occupations was convicted of tax evasion would damage his reputation and earnings.
Vela finds that considerations of reputation, along with more traditional variables in the tax evasion literature do help explain how much evasion occurs for different types of income. These variables include the likelihood of audits that varies for different classes of taxpayers, punishments for those audited, marital status (not surprisingly, married persons are less likely to evade taxes), the marginal tax rate, and the ease with which governments can match reported incomes with independent evidence on incomes, such as from 1040 and 1099 tax forms.
Note that tax avoidance as well as tax evasion tends to rise as the marginal tax rate increases. ...
However, audits, punishments, and the other deterrence variables mentioned in the previous paragraphs do not fully explain why there is not much more tax evasion. I believe it is necessary to recognize that most people believe they have a duty, moral or otherwise, to report their taxable income more or less honestly. ...
Becker presents persuasive evidence that the amount of tax evasion varies, as one would expect in a rational-choice model of taxpaying, with variance in the private costs and private benefits of evasion. I am inclined to believe that the private costs are higher than he suggests, which if true would mean that more tax compliance can be attributed to rational fear of punishment than he suggests and less to taxpayers' feeling a moral duty to pay taxes. For example, the civil penalties for tax evasion are quite severe (the fraud penalty is 100 percent of the amount of taxes evaded), and anyone charged with civil or criminal tax evasion will incur heavy legal and accounting expenses in defending against the charge. Although the audit rate is low, it is not random, but rather is higher for those taxpayers who are in the best position to evade taxes without being caught or whose tax returns raise a red flag because of unusually high deductions or other suspicious circumstances. And once one has been caught evading taxes, one can expect the rate of future audits of one's returns to be high. While it is true that underpayment of taxes is rarely prosecuted criminally, even when deliberate, criminal prosecution is likely if the tax evader takes steps to conceal the evasion, as by never filing a tax return, keeping phony books, or forging evidence of deductions. Moreover, the government does occasionally prosecute even small fry.
Thus far I have focused only on punishment costs. But a neglected point in the economics of crime is the information costs of committing a crime. Evading taxes requires more knowledge than stealing a bike. Most taxpayers probably don't have a clue as to how to evade taxes without being caught. It might seem awfully simple--just list your cat as one of your dependents. But to know whether this would work, you would have to know whether the government has any independent source of information about the number of a person's dependents. You can't just go to a lawyer and ask him what the best way of evading taxes is.
Most people comply with most laws most of the time. I believe that in most cases they do this not because they feel any moral duty to comply with law, but because the potential payoff does not seem to exceed the costs, including the information costs that I have emphasized. The reason I doubt that there is much of a felt moral duty to comply with tax law is that there is a vast amount of illegal behavior by normally law-abiding citizens. The flouting of the traffic laws, the theft of employer property, the nonpayment of social security taxes on household help, illegal gambling, and the employment (both personal and commercial) of illegal immigrants are only the most obvious examples. These are cases in which law enforcement is so lax that the expected punishment costs for most violations hover close to zero, and there are distinct benefits from violation.
Still, Becker is unquestionably correct that there is a good deal of tax evasion apart from the social security example. It could be greatly reduced by stiffer penalties and a greater investment of resources in law enforcement. Every dollar spent by the Internal Revenue Service on enforcement brings in several dollars in additional tax revenue, suggesting that an expansion in the IRS’s budget would be necessary to equate the marginal benefits of tax enforcement to its marginal costs. But this suggestion ignores the fact that the benefits are, as a first approximation, merely income transfers, whereas the marginal costs of tax enforcement are social costs. If taxes are evaded, the resulting shortfall in tax revenues is made up by increasing the tax rate, and there is no social loss unless the increase has worse misallocative effects than the evaded taxes would have had, had they not been evaded. One reason, therefore, that tax evasion is widespread is that it may be cheaper from an overall social standpoint to have slightly higher tax rates than to devote additional resources to law enforcement, though the first-best solution might be stiffer penalties, especially monetary penalties. Deliberately lax enforcement would then explain the amount of evasion.
The general question that Becker raises of the moral costs of committing crime is a fascinating one.