## Women and Taxes

No. 250
Thursday, February 28, 2002
by Edward J. McCaffery

## The Income Tax, Marginal Rates and Joint Filing

With those qualifications behind us we have a good chance of understanding how the average working wife loses two-thirds or more of her salary to work-related expenses and taxes, and how the Tax Code affects every corner of women's economic lives. We will begin with the best known of all taxes, the U.S. personal income tax.

How Marginal Rates Work. Income tax rate brackets work like the rungs on a ladder. As one climbs higher in the income scale, one enters new brackets on the margin. That is, you do not lose the benefits of the lower rate brackets, or rungs, for the dollars you have already earned. The present rate brackets for an individual look something like Table I.6 It actually isn't all that easy to generate such a table, because the "zero bracket" varies depending on whether one itemizes her deductions or not, whether one has household dependents or not and so forth. To keep matters simple, our Table I is for an unmarried person with no dependents who is taking the so-called standard deduction.

Table I means that a single taxpayer pays no income tax on her first \$7,500; after that, she pays 15 percent of each additional, or "marginal," dollar until her income reaches \$35,000; after that, she pays 28 percent on the margin, and so on.

Suppose, for example, that Sally, as yet unmarried, earns \$40,000. Her income tax will be \$5,525, determined as follows:

• 0 percent of \$7,500, plus
• 15 percent of (\$35,000-\$7,500), or \$4,125, plus
• 28 percent of (\$40,000-\$35,000), or \$1,400.

Note that this is not the same as paying 28 percent of the full \$40,000, which would be \$11,200. Sally's average tax is about 14 percent (\$5,525 of tax divided by \$40,000 of total income), far below her current marginal tax bracket of 28 percent.

Nonetheless, that marginal tax rate is critically important, for it shapes Sally's marginal decisions. Suppose that her boss offers her \$1,000, before tax, to come in and work on Sundays between Thanksgiving and Christmas. On that \$1,000 - which would push her income from \$40,000 to \$41,000 - Sally will pay \$280 in taxes. Looking at the federal income tax alone, Sally should consider whether it's worth \$720 - not \$1,000 - for her to give up these days with her family.7

This is what thinking on the margin means.

How Joint Filing Works. The tax system has to decide what to do for husbands and wives. One option, which the United States largely had before 1948, and which many advanced countries around the world have now gone back to, is called separate filing. Under separate filing, marriage is irrelevant. Husbands and wives fill out their own tax forms, on their own incomes, under something like Table I above. Each spouse has her or his own zero bracket, and so forth.

A second option is called joint filing, where husbands and wives are treated as a single unit for taxpaying purposes. This is what America has had since 1948.

Once a tax system has joint filing, further questions arise as to what to do about the rate brackets for married couples.

"The marriage penalty is not a burden on marriage; it's a burden on two-earner couples."

Should the rate brackets stay the same for married and for unmarried persons? If they did, almost all couples where both spouses worked for pay would suffer a marriage penalty. Suppose, to keep things very simple, that Sally was earning \$7,500 when she got engaged to Bill, who was earning \$60,000. Under Table I, Sally is paying no tax at all; her income falls into her own zero bracket (I chose a low figure, for convenience's sake). If, on marrying, Sally and Bill were to pay taxes under Table I as a couple, then the household income would be \$67,500, and Sally's \$7,500 would have moved from her own zero bracket into the 28 percent joint bracket. Sally and Bill's combined taxes would go up after their wedding day. Hence the "marriage penalty."

Another approach, and what America did in 1948, is to double the unmarried person's rate brackets for married couples. Looking still at Table I, the rate brackets for married couples, filing jointly, would say no taxes on income from \$0 to \$15,000, a 15 percent bracket from \$15,000 to \$70,000, and so on. Under such a system, there are no marriage penalties: taxes cannot increase on marriage. There are, however, marriage bonuses. These arise in couples with one primary or predominant earner. Think again of Sally earning \$7,500 and Bill earning \$60,000. Under Table I, quite a bit of Bill's income as an unmarried person-\$25,000 of it, to be precise-falls into the 28 percent rate bracket. If Sally and Bill got married, however, and if all the rate brackets in Table I doubled, then the household's combined income would all fall in the 0 or 15 percent brackets, the latter now extending up to \$70,000. I'll skip the more detailed math here, but the point is that taxes decrease on marriage for one-earner, or predominantly one-earner couples, under a double rate bracket system of joint filing.

It is perhaps an unfortunate fact of logic that someone's bonuses are someone else's penalties. Under a 1948-style rate structure with its generous marriage bonuses, there were, necessarily, singles penalties. In our example, Bill sees his taxes go down when he marries. If he stays single, he logically can complain of a singles penalty (just as, today, the childless can complain about the tax system when they don't benefit from child credits and the like).8 In part for this reason, Congress tinkered with the rate brackets in 1969, essentially splitting the difference between doing nothing, with its marriage penalties, and doubling the rate brackets, with its marriage bonuses. Specifically, Congress set the "married, filing jointly" rate brackets at 1.6 times the single, unmarried rate brackets. Once again, the precise numbers are complex because of itemized deductions, inflation adjustments, and the like, but the actual 2000 income tax rate brackets for married couples filing jointly look like those set out in Table II.

Now compare Tables I and II. You will notice that the rate brackets in Table II are higher than those in Table I, but they are not doubled amounts; the 15 percent bracket ends at \$59,000, for example, falling between Table I's \$35,000 and its doubled amount, or \$70,000. And so on.

"Slightly more than half of all married taxpaying couples today pay a penalty for being married."

Under this contemporary style of rate structure, there are both marriage bonuses and marriage penalties. Couples with two rather equal earners pay penalties, because they lose the benefit of some of the lower rate brackets (unmarried, they can each earn \$35,000 and never pay tax at a 28 percent rate, for example). Slightly more than half of all married taxpaying couples today pay a penalty for being married, in this sense. Couples with one primary earner, on the other hand, still get a bonus, because the rate brackets that this one earner faces if married are wider and bigger than if he or she had not married. About 40 percent of all married taxpaying couples today get a tax bonus for getting married, again in this technical sense. For the remaining 10 percent of all couples, it comes out about the same.

Why "Married, Filing Separately" Does Not Work. An option under current law, "married, filing separately," does not solve the problem of two-earner married couples on whom the marriage penalty falls.9 To understand why, compare Tables I and II again. Note that there are two effects of being married. One is that you get to combine the income of husband and wife and file a single return; this is known as income-pooling, and it is a benefit - it is in fact what creates marriage bonuses. But a second effect is that the combined income is taxed under a rate structure set less favorably than double the amounts of the individual, unmarried rates; this creates marriage penalties.

"Married, filing separately does not solve the marriage penalty problem."

The "married filing separately" rate structure is set at one-half the rates for married filing jointly; that is, one half of Table II. Couples who choose this option give up the benefits of income pooling while retaining the unfavorable rate brackets; in essence, they have rate brackets set at 0.8 (one half of 1.6) of the unmarried person's ones. Only under rare circumstances not worth mentioning here does "married, filing separately" make sense; well over 95 percent of married couples file jointly, and those that do not tend to be estranged couples unable or unwilling to sign the same tax form.

Sally's Problem. Now we are in a position to see Sally's difficulties more clearly. Suppose that Sally and Bill are married; Bill earns \$60,000; and it's not even an option worth consideration that Bill quit his paid job. Sally is thinking about going to work, going back to work or staying at work. These conditions set the stage for marginal thinking.

If Sally works (or continues to work) and earns \$30,000, the couple's total income will be \$90,000. If she stays home, it will be \$60,000. Under the correct understanding of marginal tax rates, then, all of Sally's \$30,000 will fall in the income tax's 28 percent bracket. The federal government will take \$8,400 of it.

From another - but equivalent - perspective, Sally's \$30,000 will bring home \$21,600 after income taxes, given Bill's work as fixed.

Now \$21,600 is still a significant sum; it's 72 percent of Sally's pretax salary, after all.

But wait. We're just getting started.