Posted on June 17, 2020

Summary

The massive unemployment that has hit the economy has imposed a terrible amount of hardship, and expanded unemployment benefits have cushioned the economic blow for many individuals. The CARES Act, enacted in late March, substantially expanded unemployment benefits, with much of the expansion scheduled to expire at the end of July, and those benefits are taxable under the income tax at the federal level and by most states. Unemployment recipients need to be prepared for the income tax consequences, and information from New York State (see below) suggests that most recipients there are having tax amounts withheld from their unemployment benefits. If most people nationally are having income taxes withheld from their unemployment benefits, we estimate that the effects on withholding could be very significant, perhaps contributing two percentage points to economywide withholding growth. Because of the phenomenon called “real bracket creep” (described below), withholding taxes usually rise by a slightly bigger percentage than wages do in a growing economy, and the opposite occurs in recessions: taxes drop faster than wages. My calculations suggest the tax withholding from unemployment benefits so far in this recession, unlike in the previous recession, may be big enough to be roughly offsetting the real bracket creep effect, meaning that wage and salary declines may be roughly lining up with declines in our measure of constant law withholding—that is, observed withholding growth removing our estimates of the effects on withholding from the recently-enacted payroll tax credits and deferrals that have been reducing withholding but not wages and salaries.

Unemployment benefit amounts were very large in April and May, boosted substantially by the CARES Act provisions that, among other things, added a flat $600 per week federal benefit to any state benefits paid and allowed self-employed workers to qualify for benefits for the first time. Unemployment benefits in April and May together, at about $142 billion, were almost as large as benefits paid in all of 2010 ($157 billion), the year in the Great Recession with the largest amount of benefit payments (see chart above). Clearly we can expect benefits paid in June and July to cause this year’s amount to far exceed the amounts paid in any year, and this year’s amounts are concentrated in a very short period of time. The flat $600 per week amount is large enough that some analysts (for example, see this study from analysts at the University of Chicago) estimate that about two-thirds of recipients, namely those with lower incomes, are receiving as much or more in benefits than they earned when they were employed. That is critical in allowing many families to financially weather the pandemic and to prop up consumer spending and aid the economy. (It does raise questions, though, about the incentives to return to work.) Many families, nonetheless, are still receiving much less in unemployment benefits than what they earned when working, and others don’t qualify for unemployment benefits for various reasons.

Recipients of unemployment benefits need to be prepared for the income tax consequences, given that benefits are fully taxable at the federal level and by most states, and that the amounts received can be significant. States are required to provide unemployment benefit applicants the option to withhold amounts for federal taxes at a 10 percent rate. According to press reports citing the New York State Department of Labor, about 60 percent of recent first-time New York applicants elected to have federal and state taxes withheld from their unemployment benefits. That is more than I would have guessed for several reasons: the cash constraints so many people face; the fact that some people might well earn insufficient amounts, combined with their unemployment benefits, to owe any income taxes, so they shouldn’t have any withholding; people can choose to make quarterly estimated payments instead of withholding; and behavioral economics suggests that many people do not opt in for setting aside funds for future needs (for example, for retirement), but also don’t opt out if they are automatically enrolled. Note that if people don’t have amounts withheld from benefits, or don’t reflect the amount of associated taxes in quarterly estimated payments, then the recipients could have large tax bills next April, or perhaps much smaller refunds than usual.

The magnitude of the unemployment benefits in recent months has a substantial effect on economywide taxes. If 60 percent of unemployment benefit recipients nationwide have taxes withheld at the 10 percent rate, that could imply about 2 percent of withholding in June stems from unemployment benefits alone. (That calculation takes the average of the April and May amounts of unemployment benefits, so about $70 billion, multiplied by the assumed 60 percent who elect to withhold and the 10 percent federal tax withholding rate, which yields a withholding amount of about $4 billion, or about 2 percent of an average withholding month of about $200 billion.) Since we estimated that withholding, adjusted to remove payroll tax credits and deferrals enacted in the CARES Act, fell about 6 percent in the first half of June compared to the same period in June of 2019 (see previous post), the implication is that withholding would have dropped by about 8 percent without the boost from unemployment benefits.

Normally withheld taxes rise a bit faster than wages in a growing economy, and fall a bit faster than wages in recessions, but that may not be happening this time. The reason for withholding normally being more responsive than wages is the phenomenon called “real bracket creep”, where higher wages in a growing economy push taxpayers into higher income tax brackets, boosting their taxes more than their wages; that effect is built into the into the income tax withholding calculations. It also works in reverse in recessions, where the lower wages push taxpayers into lower tax brackets, dropping their taxes by a greater percentage than their income. This time, though, my calculations suggest that the unusually high levels of unemployment benefits and the corresponding effect on tax payments may be roughly offsetting the normal real bracket creep effect in downturns; that may mean that what we observe in withholding declines will line up more closely than usual with changes in wages and salaries. The netting is also complicated by data showing that lower income individuals are being very disproportionately hit by the current downturn, as well as the effect from self-employed workers being covered by unemployment benefits for the first time. We will certainly need more data to better estimate the magnitudes of the various effects.