Posted on May 31, 2022
The amount of federal income and payroll taxes withheld from workers’ paychecks and remitted to the U.S. Treasury Department–what we call tax withholding–grew much more slowly in May than in the prior recent months, according to our estimates. The amount of tax withholding, which moves with overall wages and salaries in the economy, was about 5 percent higher in May compared to May of a year ago (so-called year-over-year growth). That is much slower than the growth of between 9 percent and 11 percent recorded in each of the five previous months (see chart below). We construct that measure of withholding growth from daily reports released by the U.S. Treasury, but adjusted by us to remove the estimated effects of tax law changes (currently a relatively small adjustment) and to standardize the makeup of business days across months (see our methodology). Other, more direct measures of wages and salaries in the economy–from the Bureau of Economic Analysis (BEA) in its GDP Accounts and from the Bureau of Labor Statistics (BLS) in its monthly establishment survey–have been growing in recent months at a similar clip to tax withholding, although those other measures are not yet available for May. We will see if the BLS establishment report for May, to be released on Friday of this week, shows a slowdown in growth in employment and/or average hourly earnings. That could certainly have effects on financial markets.
The big drop in withholding growth in May has data anomaly written all over it, but we can find no measurement-related reason for such a rapid drop-off. Although the timing of the Memorial Day holiday was different this year than last year–when it occurred on the very last day of the month instead of the second to last day of the month this year–that should not significantly affect our measurement of withholding growth adjusted for calendar differences. (Note that without the calendar adjustments, withholding in May this year would be well ahead of last year’s amounts, because no amounts are due and recorded on the Monday holiday, but rather on the following day, which was in June last year and in May this year.) The calendar adjustments didn’t seem particularly unusual, as we had a number of daily withholding growth measurements in the second half of May that were all in a very tight band of 5 percent to 5.5 percent year-over-year growth. Also, withholding growth in May of last year, as we measure it, was very similar to that in the months surrounding it, suggesting that nothing unusual occurred last year that is distorting this year’s measurement.
A change in withholding growth such as we see for May would normally be indicative of a slowdown in growth of either employment or average hourly earnings. There can, however, be months in which noisy data or special circumstances cause withholding to deviate from wages paid, so we will see what further data show–mainly upcoming data on wages from BLS and BEA, as well as data from the Treasury Department on withholding in June.