Income tax burdens on the average family of four

Yesterday being Tax Day (the due date for on-time calendar year income tax returns), there were naturally various stories in the media about people's tax burdens and such. One that came up at least twice concerned that old chestnut of U.S. family structure, the good old "family of four."

As it happens, I was a member of such a family as a child, and I am again a member of such a family as a parent. But meanwhile I suppose a lot of things have changed.

In yesterday's NYT, Ross Douthat has an op-ed mentioning "families of four" that I didn't get to read until today (I've been really busy lately). I didn't have it in mind, therefore, when speaking yesterday to Natasha Lennart of Salon about the taxation of families of four, so it's lucky for me that her interest didn't lie in the Douthat column (although I suppose she would have told me if it did).

As per Lennart's Salon column that was posted yesterday evening, her interest lay in a report posted by the Tax Policy Center finding that, "for the second year in a row, a family of four earning the median income is paying less in federal income taxes than at any time since at least 1955."

I posited that this presumably reflected not only continuation of the Bush tax cuts, but more particularly the continuing economic slowdown. For example, suppose a lot of two-earner families are down to one earner. While obviously this is bad for them if they want or even need the two incomes, it would certainly tend to reduce their tax liability. If this is a major factor in the data, then presumably family-of-four tax liabilities will start rising again if employment levels ever get back to where they ought to be.

In fact, however, I gather that the Tax Policy Center data is based on a representative two-parents, two-kids, one-earner family of four. So it's lucky for me that I was only quoted saying things that actually are true:

"In 1955 'family of four' described a single-earner nuclear unit: a married man and woman with two children.

"'We're certainly more heterogeneous these days,' New York University's Wayne [Perry] Professor of Taxation, Daniel Shaviro, told Salon."

Lennart then notes that, while the Tax Policy Center data that she was citing assumes a traditional 1950s-style family, "in 2005, according to census data, 42 percent of families had two income earners, and around 34 percent of children lived in single-parent families. Of course, this does not change the fact that taxes are historically low. But the 'four person family' statistic does not pay attention to the fact that a dual-earner family is dually burdened with Social Security taxes and cannot claim the same spousal benefits. As Shaviro noted, 'Secondary earners are taxed fairly highly.'

"There is nothing misleading about analyzing the income tax burden of a married couple with two children with just one earning spouse. It is mistaken, however, to assume that is what's understood by the term 'family of four' in 2011."

Anyway, back to parts of the interview that luckily didn't appear in the column. Not having yet read the Douthat column, I suggested that, while household heterogeneity is an important issue in certain contexts - e.g., distributional effects of the fiscal system as between one-earner couples, two-earner couples, and singles, as well as the system's excessive discouragement of work by "secondary earners" in a couple - it wasn't necessarily crucial (in the sense that ignoring it would be misleading) when one is talking about historically low (since the 1950s) U.S. income tax burdens today, which presumably are a function of the rates plus the down economy.

Douthart raises the issue as follows:

"Today ... a family of four making the median income — $94,900 — pays 15 percent in federal taxes. By 2035, under the C.B.O. projection, payroll and income taxes would claim 25 percent of that family’s paycheck. The marginal tax rate on labor income would rise from 29 percent to 38 percent. Federal tax revenue, which has averaged 18 percent of G.D.P. since World War II, would hit 23 percent by the 2030s and climb even higher after that.

"Such unprecedented levels of taxation would throw up hurdles to entrepreneurship, family formation and upward mobility. (Or as the C.B.O. puts it, in its understated way, they would 'tend to discourage some economic activity,' and 'harm the economy through the impact on people’s decisions about how much to work and save.')"

He makes this point to challenge the Obama side of what many see as the "Obama versus Ryan" debate over our fiscal future, on the view that entitlement growth (in particular on healthcare) will definitely have to slow, whether not as much or in the same manner as the Ryan plan posits.

Douthat then got criticized (such as here) for using an overstated income number. Apparently the number he should have used, for taxable income purposes, is $75,700. He has acknowledged that he misread a chart that includes employment-based health insurance and the employer’s share of payroll taxes, neither of which is in taxable income. So we have a rare happy ending, in terms of on-line snark exchanges, as the above-linked critic accepts in an addendum that it was an honest mistake.

But Douthat's choice of a representative family could be questioned as potentially misleading (albeit still, I trust, honestly so) in a sense that I would have mentioned, in my Salon interview with Lennart, if only, ahem, I had gotten to read the Douthat column in the morning (re-queue whining about how busy I am lately). All else equal, Douthat would have had to cite lower numbers had he chosen a one-earner family or a single individual, given how joint returns amalgamate spousal incomes. The relative burden on two-earner families, from how we treat such households via joint returns with rate brackets being significantly less than double those on single returns, is important, but makes such a family not entirely representative. (So re-queue here the heterogeneity point.)

Another important aspect that Douthat may not have recognized was the following. The tax rate increase for two-earner families that he shows presumably reflects a key fact about income tax brackets under present law. They are indexed for inflation, but not for real GDP growth. Accordingly, there is a tendency over time for movement towards having everyone, not just high-earners, end up in the top bracket.

E.g., suppose that, with the gracious assistance of Wikipedia, we compare median incomes, in constant (i.e., inflation-adjusted) dollars for past eras as compared to today. Between 1950 and 2000, the U.S. median income for men grew 80 percent in real terms, while that for women almost tripled. (The latter, of course, also reflects women's greater entry into the workforce.)

For this reason, when economists try to project future tax revenues under a "current policy" approach rather than a "current law" approach, they will typically assume that, over time, tax rate brackets are roughly indexed for real GDP growth, as well as for inflation, thus causing the brackets to kick in at the same points in the contemporary income distribution in, say, the 2030s as today. After all, permitting the U.S. income tax to evolve into a system in which more and more people hit the top rate bracket would be a genuine policy change, albeit the natural consequence of simple legal inertia.

Obviously, a long-term U.S. fiscal gap and what I would call a reasonable (and indeed indispensable) willingness to address it on the tax side, as well as on the spending side, would naturally raise the question of whether we should continue current rate bracket policy by raising the tax rate brackets for real GDP growth as well as for inflation. But my default would certainly be to assume current policy in this regard.

While I'm sure it was not deliberate, I do think that Douthat's column ends up being potentially misleading in substance insofar as it treats real bracket creep as illuminating with regard to whether we should raise tax rates (at least for the top bracket) in keeping with Obama's but not Ryan's plan. The question of how high tax rates should be is analytically distinct from where we should set the break points between brackets. And the real bracket creep tendency of present law does not by itself carry any implication that the right solution is to lower the top rate a la the Ryan plan.