The poorly named Inflation Reduction Act has little to do with fighting inflation. It has more to do with tax policy, and more importantly, tax policy enforcement. Indeed, the Internal Revenue Service (IRS) has been given a great deal of extra resources to conduct audits and investigations of potential tax evaders.
The response by many is that more enforcement of tax policy is a desirable thing. Citing numerous studies, notably a famous one in the American Economic Review authored by Gabriel Zucman and two of his colleagues, defenders of greater enforcement point out that most tax evasion is done by wealthy individuals in the top 0.01 percent of the income or wealth distribution. Because of their vast wealth and high incomes, the richest can cheat governments more easily. As a result, the richest in society evade 25 percent of their taxes compared with less than 5 percent for people down the income ladder. Such proportions make it easy to then jump to “enforcement” in order to argue that more audits will allow governments to collect more revenues and close the deficit.
There is, however, one major problem with this: It does not consider that everyone cheats the government in one way or another.
Take the aforementioned famous paper by Zucman and his colleagues. That paper estimates tax avoidance by relying heavily on movements of wealth to foreign countries to avoid tax liabilities. This “offshoring” is costly to do, and only the richest can afford it – ergo the great degree of cheating frequently highlighted in the media.
However, people with lower incomes cheat differently. Because of their modest means, they can more easily cheat governments by working illegally. For example, self-employed workers can “omit” certain cash payments they received in the course of fully legal work. Others can simply work an illegal job and not report anything.
True, random audits ensure that enforcement could eventually catch these individuals. The problem is that these types of cheating are harder to detect because they are small relative to the economy’s size. For example, an undergraduate student who worked during the summer for a moving company for cash payment for three or four days (a classic occurrence in my home city of Montreal), could earn a few hundred dollars that tax authorities could hardly detect. This is why Zucman and his colleagues argue that random audits “are likely to miss some forms of tax evasion in the bottom and the middle of the distribution.”
This is not a minor problem. A few hundred dollars for that undergraduate student represent a sizable share of his annual income. Waiters in restaurants can easily under-report large cash tips that will, at the year’s end, represent a sizable share of their annual income. As such, proportionally to income, tax evasion is greatly underestimated as you move down the income ladder.
This claim of underestimating the evasion at the bottom of the income is consistent with well-known empirical facts regarding the underground economy (i.e., the informal sector, the shadow economy, grey and black markets etc.). For example, we know that poverty is a strong determinant of participation in the underground economy. We also know that the underground economy, which is quite large even in rich countries such as Canada, tends to disproportionately involve people lower down the income ladder.
It is hard to arrive at a number of the extent of the underestimation of cheating by bottom and middle income groups, but there are clear signs of underestimation. With this in mind, one realises that the political quip that the rich just don’t pay what they owe is humbug. More importantly, one realises that greater enforcement by the IRS may lead to more collection of revenues at the top. However, it may also end up taking away income that, dollar for dollar, is more proportionally important for people at the bottom of the income ladder. In other words, the greater enforcement could be regressive. Policymakers and pundits should ponder this possibility before they wrap themselves in the cloth of virtue.
