SALT Substitutes: A Replacement Payroll Tax?

This week I’m looking at three major proposals for states to respond to partial repeal of the federal tax deduction for state & local taxes. Yesterday we covered some background and evaluated a plan to use the charitable contribution deduction in place of the SALT deduction. Today, it’s…

Repeal and Replace

Our second option has been associated most prominently with the economist and CEPR blogger Dean Baker. Baker’s idea (previewed by Daniel Hemel and David Kamin here at WSD in November) is that states could swap out their individual income tax for a payroll tax. What makes this work is the fact that TWNN does not repeal SALT deductions for businesses. In Baker’s design, the statutory burden of the new payroll tax would be imposed on employers, at the same rate and with roughly the same base as the former income tax. Employers could then claim this new tax as a federal deduction. Wages should adjust to hold workers roughly harmless after all the dust settles.

To see this, imagine that Dean earns $1m in Maryland, where in 2017 he pays a 10% tax. His for-profit employer (FPE, Inc.) takes a $1m business expense deduction for the cost of his salary, reducing its taxable profits from, say, $10 to $9m. At a 35% federal rate, this deduction saves FPE $350k. Assuming a 30% federal individual rate, Dean’s 2017 takehome is $1m — 100k — (30% * $1m — $100k) = $630k.

Now TWNN is enacted and Maryland repeals & replaces its income tax. Rates are set such that FPE pays $100k in payroll taxes to Maryland. Dean’s compensation is reduced to $900k to offset the new tax. FPE gains a new deduction of $100k for the payroll tax, but it also loses a $100k salary deduction. It still reports $9m in net profits. Dean pays 30% of $900k, or $270k in federal taxes, taking home $630k.

One sticky point for this plan is that wages are not usually this flexible, especially not in the downward direction. For instance, we might think that workers will not recognize that their after-tax takehome pay is constant under both options, and resist the offsetting pay cuts needed to make the scheme work. Over a longer time period, wage growth will stagnate (to an even greater degree than at present) or employment will slow so that the average expected wage equilibrates, but this implies that Dean’s plan might contribute to a downtick in employment rates.

If it works, though, it turns out that this mechanism does not depend at all on the employer’s tax rate. For instance, in effect repeal and replace shifts the employee’s deduction to her employer, but what if the employer does not claim deductions? Governments, nonprofits, and money-losing startups don’t have any federal taxable income to offset with deductions for payroll taxes. But this shouldn’t matter. Let’s say Dean works at CEPR, a nonprofit, instead of FPE. Presumably, the equilibrium wage for Dean at CEPR reflects his employer’s total cost of employing him. If this was $1 million before repeal and replace, it will still be $1m after. All that repeal & replace does is swap out $100k of salary and rename it payroll tax, but this renaming has the fortuity that payroll taxes paid as a result of employing a worker are not considered taxable income for that worker.

Here are the numbers. In 2017, it costs CEPR $1m in salary to employ Dean, who takes home $630k, as above. In 2018, under TWNN and repeal and replace, CEPR must pay $100k to Maryland. To hold its costs constant, it reduces Dean’s salary to $900k. Since 30% of $900k is $270k, Dean will still take home $630k.

In fact, things are even better than these examples suggest, because repeal and replace will also reduce federal payroll taxes. By reducing a worker’s salary by the amount the worker formerly paid in state income tax, repeal and replace also reduces the federal wage base for FICA taxes — as much as a 15% savings for employees earning under around $120k. These savings can then be divided between employer and employee. Too bad for the Social Security Trust Fund (don’t worry, it’s not a big deal).

It’s not all roses. Adjusting an employer-administered payroll tax to reflect the individual circumstances of individual workers will be difficult. The biggest challenge is probably dual-earner households. What if Dean’s spouse earns $50k? In theory, the rate at which the replacement payroll tax is imposed on Dean’s spouse should represent the marginal rate Maryland imposed on households earning $1.05 million. If the rate were set based only on spouse’s income, it would probably fail to replace the income taxes formerly collected from the household.

There are solutions here, but they are of uncertain effect. For example, the state could allow payroll tax rates to vary by household, though doing so accurately would probably require workers to disclose to their employer (or at least its payroll administrator) their spouse’s earnings. In addition, because employers rarely individualize compensation packages for line workers, such a system would tend to shift some of the tax burden from those formerly in higher brackets to those who were in lower. That is, employers would tend to reduce the salary of every Senior Clerk by the average increase in payroll tax for a senior clerk, instead of tailoring salary changes to the payroll costs of each employee. Fortunately, many states have a flat enough rate schedule that these issues do not even arise or do not represent a meaningful amount of money.

Investment earnings are also a complication, as Baker acknowledges. There is probably no way to shift a tax on capital gains to another state taxpayer, especially in the case of cross-border investment. Possibly states could reduce individual taxes on capital and increase taxes on business, but that exacerbates the familiar problems that small, open economies face in taxing mobile capital. One small partial fix would be to include investment earnings when determining the marginal rate for a variable-rate payroll tax, ensuring that households with substantial investments would be subject to payroll tax rates that approximate the rate they would have faced under an income tax. Again, though, a variable-rate system might not work in practice.

Another implication of the loss of individualization would be that it would be much more difficult for states to pursue policy through their tax system. Without an income tax, states could not offer charitable contribution deductions, tax credits for organ donation or renewable energy, and so on. My view is that this would mostly not be much of a loss. Nearly all of these programs could, and in many cases should, be replaced with direct spending programs.

For instance, the state charitable contribution deduction represents poor tax planning. As we’ve seen, when state taxes are federally deductible, state tax deductions increase federal tax. If states replaced deductions dollar for dollar with a matching grant to a donee’s charity of choice, it could potentially deliver just as much value to the nonprofit sector without also hiking the generous donor’s federal taxes (though of course the net result would depend on the extent to which donors respond to matching grants rather than deductions, a question on which we currently have little data). This same problem plagues state organ-donor credits, though in that special case it is arguable that federal law would prohibit the state from directly compensating the donor (see Bilgel & Galle 2015 for more discussion).

Still, reengineering all these systems will incur startup costs both for government and for those individuals who employ them. There can be advantages to leveraging the administrative capacity of a tax system for other policy goals. And political economy constraints might prevent some existing and laudable ideas from securing support if state income tax systems were scrapped. What, for instance, would be the political fate of state EITC’s?

Finally, repeal and replace has cross-border wrinkles that will need ironing. Suppose a California resident who works in Nevada, where there is no income tax. California may well lack “nexus,” or the constitutional right to exert its legal authority, over the Nevada employer. Some federal statutes prohibit states from imposing income taxes on certain workers (notably railroad employees and military service personnel) domiciled in another state, and there will certainly be lawsuits over whether a state payroll tax can properly be applied to such a person’s wages.