Thoughts on Baselines and the Toomey Proposal

by Reihan Salam

Recently, the New York Times published an article on the debate over the Social Security payroll tax. The headline read as follows:

G.O.P. and Democrats Differ on How to Prevent Social Security Payroll Tax Increase

Is this headline flatly and flagrantly untrue, or is it an example of choosing a particular baseline? The article, by Robert Pear and Jennifer Steinhauer, both of whom strike me as honest, intelligent reporters, provides context:

Republicans in both houses of Congress said Wednesday that they wanted to prevent an increase in Social Security payroll taxes that would otherwise occur in January, but they remained far apart from Democrats on the specifics of how to do it. …

Senate Republican leaders introduced a bill that would keep the payroll tax rate at its current level for another year. The cost is roughly $120 billion. Senate Republicans would offset most of the cost by freezing the pay of federal employees through 2015 and gradually reducing the federal work force by 10 percent.” [Emphasis added]

Several paragraphs down, we find the following:

Last December, Congress temporarily reduced the employee’s share of the Social Security payroll tax by 2 percentage points, to 4.2 percent of wages. If Congress does nothing, the rate will revert to 6.2 percent in January.

Here is another headline that would have been, by my lights, equally accurate:

G.O.P. and Democrats Differ on How to Extend a Social Security Payroll Tax Cut

The baselines underlying the two headlines are different. In the first case, Pear and Steinhauer’s editors chose to emphasize increases relative to the current level of Social Security payroll tax. In my second theoretical headline, they would instead emphasize the fact that the Social Security payroll tax was temporarily reduced. It is not obvious to me that Pear and Steinhauer’s editors are being dishonest as such by choosing one baseline and not the other, though of course others — particularly those who oppose extending the Social Security payroll tax cut — might disagree.

I mention this because Kevin Drum believes that I am being dishonest. First, he quotes a recent post of mine (thanks for the link, incidentally):

[Brownstein] suggests that the GOP is aiming for “a deficit plan that relies solely on spending reductions (particularly in entitlements) while preserving tax cuts for the affluent.” As Keith Hennessey has explained, congressional Republicans have made a strategic shift on taxes. Opposition to any net tax increase was a way to secure leverage in negotiations over the long-run fiscal trajectory. But leading Republicans have demonstrated an openness to net tax increases, and in particular to an increase in the average tax rates paid by the most affluent households, provided it is part of a package that secures structural, architectural Medicare reform.

Then he adds:

This is just flatly untrue, and I’m a little surprised at how flagrantly it’s been bandied about lately. The Toomey proposal in the supercommittee did indeed include about $300 billion in new taxes, some of which would have fallen on high-income earners. But Toomey’s proposal explicitly tied this to a demand for permanent extension of the Bush tax cuts for the rich (those are the red bars on the right, the ones that fall exclusively on those earning more than $200,000). Unlike the other Bush tax cuts, these cuts are very much a matter of contention between the parties, so Toomey’s proposal, in plain English, was this:

*Republicans will agree to raise taxes on the rich by $300 billion

*If and only if Democrats agree to permanently extend $700 billion in tax cuts for the rich.

Conservatives try to handwave this away by claiming it’s “just a baseline game” or some such, but if it were just a game they wouldn’t care about it. In fact, as they know quite well, we’re talking about actual tax dollars paid by actual people that affect the actual budget deficit. Compared to the law as it stands now, Toomey’s deal explicitly demanded that net taxes on the rich go down by about $400 billion after 2013. That was the only deal ever on the table. If you want to claim that it’s a breakthrough merely for Republicans to propose a reduction in the net size of the Bush tax cuts, that’s fine. But a net increase? They’ve never even come close to offering that. [Emphasis added]

This is actually a very interesting document for a number of reasons, and I’m grateful to Kevin for illustrating a number of habits of mind and argumentative strategies that are not uncommon on the left and the right. I am also grateful to him for giving me the opportunity to raise a number of related issues. So let’s take this piece by piece:

(1) “This is just flatly untrue, and I’m a little surprised at how flagrantly it’s been bandied about lately.“ 

Another way to approach this issue is to put ourselves in the shoes of Pear and Steinhauer’s editors. Using the current level or current policy as a baseline seems like a reasonably coherent approach, though of course it creates a bias towards the current level or the current policy. In the context of the Pear and Steinhauer article, this implicit bias is in favor of maintaining a lower level of Social Security payroll tax that had prevailed before the temporary cut. 

It is certainly true that I didn’t make explicit reference to a current policy baseline, in part because I made a passive assumption that in my community of readers, my implicit baseline was understood. Blogging often involves shorthand. Rather than make every assumption explicit, we assume that understanding will emerge in the course of an ongoing narrative formed over many blog posts over a period of time. Indeed, this is one of the pleasures of reading blogs — I have a sense of how Tyler Cowen thinks about a wide range of issues, and the kind of stories to which he gravitates. 

One ongoing storyline has been that congressional Republicans have opposed net tax increases relative to the current policy baseline. For months, congressional Republicans have signaled an openness to base-broadening measures, provided the revenue would be applied to reducing marginal tax rates and not to deficit reduction.  

To characterize others as dishonest requires that we first rule out the possibility of good-faith disagreement. That is usually a last resort for me. We’ll return to this theme in a moment. 

(2) “The Toomey proposal in the supercommittee did indeed include about $300 billion in new taxes, some of which would have fallen on high-income earners.

As I understand it, the Toomey proposal, which has not been entirely fleshed out, is inspired the Feldstein-Feenberg-MacGuineas cap on individual tax expenditures, which the CRFB has described as follows:

Both parties are studying cuts in tax expenditures (TEs) as a means to reduce the federal deficit. Super Committee Republicans have offered a proposal to limit the amount of tax expenditures to help generate about $250 billion in new revenues, while Super Committee Democrats have proposed an automatic trigger to limit TEs (a la the Feldstein-Feenberg-MacGuineas approach). So, this is a good time to revisit “Capping Individual Tax Expenditure Benefits,” a study on an approach to limit TEs by Martin Feldstein, chairman of Reagan’s Council of Economic Advisors, NBER’s Daniel Feenberg, and CRFB President Maya MacGuineas.

They show that a TE benefits cap of 2 percent of household income (AGI) would lower the federal deficit by $278 billion in a single year. They slice and dice that sum (their Table 3) into forty-eight segments, six different TE categories across eight income groups. They also indicate annual sum savings of $208 billion if the cap were 3 percent or $110 billion with a cap of 5 percent.

The political appeal of the TE benefits cap is that it doesn’t require lawmakers to go after a particular tax expenditure, e.g., the mortgage interest deduction. Rather, it attacks all tax expenditures in an at least notionally political neutral way. 

So the Toomey proposal posits that Congress can impose something like a Feldstein-Feenberg-MacGuineas cap to raise, say, $2.78 trillion over the decade-long budget window. (Multiplying by ten makes the insane assumption that the economy will look much the same over the next ten years. Assuming there is a sustained recovery, we can call this a conservative estimate.) Sen. Toomey wants to use the bulk of this revenue to cut marginal tax rates and retain a relatively small portion of it to raise revenue relative to current policy. 

The individual cap on tax expenditure benefits can be deployed in a wide range of contexts. For example, under the current law baseline, taxes are scheduled to increase by $3.8 trillion on January 1, 2013. Adding a cap on individual tax expenditures would further raise revenue by a dramatic margin.

Or we could allow current policy to remain in place except for the addition of a Feldstein-Feenberg-MacGuineas cap, which would represent a (for our purposes) $2.78 trillion net increase relative to current policy. To be sure, this would represent a steep tax decrease relative to current law, i.e., the universe in which the tax trigger is pulled. Sen. Toomey’s proposal goes much further towards the current policy baseline than this by devoting most of this new revenue to cutting marginal tax rates. 

(3) “But Toomey’s proposal explicitly tied this to a demand for permanent extension of the Bush tax cuts for the rich (those are the red bars on the right, the ones that fall exclusively on those earning more than $200,000).

As short as this sentence is, it offers a really rich vein for exploration. Toomey’s proposal called for jettisoning the Bush-era tax code and creating a new schedule of marginal tax rates and a new structure for tax expenditures. It does, however, build on current policy rather than current law to determine what constitutes an appropriate revenue target.  

Now for the more interesting part of the discussion. The 2001 and 2003 tax cuts, or the Bush-era tax cuts, are described in many different ways. Critics tend to separate them into two buckets: (a) the “Bush tax cuts for the rich” and (b) the “middle-class tax cuts.” Alan Viard of AEI describes the cuts as follows:

The 2001 tax cut lowered the top four brackets, moved some income previously taxed in the 15 percent bracket to a new 10 percent bracket, increased the child tax credit, gave tax relief to married couples, and expanded other tax breaks.[1] The 2003 tax cut accelerated some of the provisions of the 2001 tax cut and lowered capital-gains and dividend tax rates. Without further action, most of these provisions will expire at the end of 2010.

Notice that if Viard’s description is correct, the aforementioned framing immediately leaps out at us as misleading. All of the provisions of the 2001 tax cut, including those characterized as (b), benefit high-earners. As Viard writes:

Although marginal-tax-rate increases are distortionary at any income level, rate increases at the top income levels generally create the largest distortions per dollar of revenue. That partly reflects the fact that the rates at the top income levels are already high, so further increases are more damaging. But it also reflects the fact that rate increases at the top raise revenue from only some of the affected taxpayers’ income. For example, consider a proposal to increase taxes by 5 percent of the income above $250,000 (which approximates the expiration of the high-income rate reductions for a married couple with ordinary income). The resulting revenue is less than 5 percent of the affected taxpayers’ incomes because the tax applies only to the income above the $250,000 threshold; for a $400,000 couple, for example, the 5-percentage-point tax increase applies only to the last $150,000 of income, and the revenue is only $7,500. Because the disincentive effects depend upon the marginal tax rate applied to the last dollar, though, they are as severe as if the couple had to pay an extra 5 percent on their entire income (except that the disincentive will not prompt the couple to reduce their income below $250,000).

Critics of the high-income rate reductions often use this as an argument against the anti-tax sentiments of high-earners. Most HENRY (high-earner, not rich yet) households will only experience a slight tax increase under proposals to roll back the high-income rate reductions, as only part of their income is above the threshold. Yet as Viard explains, the incentive effect is significant. This is why capping individual tax expenditure benefits represents a very different and in many respects much smarter strategy. It more closely matches the revenue gain to the scale of the disincentive effect. Unfortunately, this is a difficult concept to understand, particularly for those to whom high marginal tax rates have a totemic significance. Viard continues:

The opposite pattern holds for tax hikes in the bottom bracket, which actually leave marginal rates unchanged for most of the taxpayers from whom additional revenue is collected. For example, suppose that the 10 percent bracket, which will apply to the first $17,000 of a couple’s taxable income in 2011 if extended, reverted to 15 percent. Taxpayers in that bracket would face a 5-percentage-point increase in ¬disincentives and would pay an additional 5 percent of taxable income. At the same time, all couples in higher brackets would also pay an additional $850 in tax because the first $17,000 of their income would be taxed at 15 rather than 10 percent. These higher-bracket taxpayers would not, however, face any additional disincentives because there would be no change in their marginal last-dollar tax rates. The size of the distortions depends upon the extent to which taxpayers can reduce their taxable income in response to higher marginal tax rates. [Emphasis added]

Another way of putting this is that tax cuts in the brackets aimed at low- and middle-earners will do nothing to encourage high-earners to work and earn more. They just represent a really big revenue loss from high-earners. That might be a perfectly good thing if we want to limit the amount of money the federal government takes in, but not if we want to get the most “bang for the buck,” i.e., the most growth-enhancing tax code.

The cuts to capital-gains and dividend tax rates represent a somewhat more complicated issue. High-earners are overwhelmingly the beneficiaries of capital gains and dividends, yet there are potential spillover benefits for the broader economy, the extent of which are a subject of considerable debate and speculation. Let’s just say that some people believe that all U.S. workers benefit when investment increases and others believe that relatively few people capture the benefits. 

(4) “Unlike the other Bush tax cuts, these cuts are very much a matter of contention between the parties, so Toomey’s proposal, in plain English, was this: (a) Republicans will agree to raise taxes on the rich by $300 billion (b) If and only if Democrats agree to permanently extend $700 billion in tax cuts for the rich.

Drum and I have different definitions of plain English. To start, he begins with the assumption that some of the Bush-era tax cuts are “a matter of contention between the parties.” This could mean many things. It suggests — to me, at least — that he is introducing a third baseline. There is current policy — the current level of taxes — there is current law — building a $3.8 trillion tax increase relative to current policy come January 1, 2013 — and then there is a baseline based on what is not “a matter of contention between the two parties.”

This actually has a funny impact on the political conversation. The “middle-class tax cuts,” which, as we’ve established, benefit the rich as well as the non-rich, are sacrosanct because congressional Democrats embrace them. This is despite the fact that the revenue loss from the “middle-class tax cuts,” including the revenue loss from the cuts that go to high-earners, is actually much larger than the revenue loss from the high-income rate reductions. Last fall, Josh Barro wrote a post on some of the other implications of this approach, drawing on testimony from CBO chief Douglas Elmendorf: 

In the CBO’s strong labor response model, a full extension of the tax cuts knocks 0.6 percentage points off of GNP, whereas a partial extension as proposed by President Obama (which would raise earned and unearned income tax rates on most filers making over $250,000 per year) would cost 1.2 points of GNP. That is, a partial tax cut extension is worse for the economy than a full extension, even though the full extension grows the national debt by more.

The implication is that the extending the portion of the tax cuts that principally benefits high earners would grow GNP by 0.6 points, despite such an extension’s effects on the deficit. It is the rest of the tax cuts (lower marginal rates in the bottom four brackets, marriage penalty relief, the higher per child tax credit, etc.) that do not stimulate enough economic activity to offset the added debt burden.

Even if you prefer CBO’s weak labor response model (which assumes workers have low sensitivity to tax treatment), the presentation shows equal economic damage from a partial or full extension, despite the fact that a full extension grows the deficit by about a third more than a partial extension. With this assumption, tax cuts for the rich aren’t an economic boon, but they are at least a free lunch. [Emphasis added]

I mention Josh’s post because I think there is a case to be made for low marginal tax rates for high-earners and for making the overall tax-and-transfer system more progressive by curbing tax expenditures and making public expenditures more progressive, e.g., instituting a higher minimum Social Security benefit while reducing the benefit for those with higher lifetime earnings, etc. That is, I actually don’t think the model of hiving off the contentious “Bush tax cuts for the rich” from the apparently uncontentious “middle-class tax cuts” makes sense. 

If anything, I actually think the “middle-class tax cuts” should be a subject of more contention, because they do such a poor job of meeting the needs of low- and middle-earners and the broader economy. That might be why I have a different sense of what is and is not contentious. Instead of trying to adjudicate contentiousness, I started from current policy, not because it is or is not contentious but rather because that is where we happen to be at the moment. (That’s why we call it “current policy.”) 

So in plain English, Drum seems to be baking in his political assumptions while claiming that those who bake in their political assumptions are being flagrantly dishonest. This actually doesn’t strike me as dishonest. It does strike me as a failure of empathy. Is it possible for others to see the world differently and not be stupid or dishonest? The answer to that question is usually yes. 

To turn to Drum’s (a), the $300 billion isn’t actually an increase on the rich, as I understand the proposal. Some of it will be borne by high-earners while some of it will be borne by middle-earners, which is a reason why the Toomey proposal has been criticized by some observers on the left. (There is also the concern, from Jed Graham, that is will raise expenditures on the PPACA health exchanges.) From Drum’s perspective, this presumably makes the Toomey proposal “worse.” (My own view, again, is that the overall distribution of taxes and transfers is what matters.) As for Drum’s (b), this goes back to our earlier question of the appropriate baseline. Do we roll with Pear and Steinhauer’s editors in the case of the Social Security payroll tax and take current policy as the appropriate baseline or not? 

(5)”Conservatives try to handwave this away by claiming it’s “just a baseline game” or some such, but if it were just a game they wouldn’t care about it. In fact, as they know quite well, we’re talking about actual tax dollars paid by actual people that affect the actual budget deficit. Compared to the law as it stands now, Toomey’s deal explicitly demanded that net taxes on the rich go down by about $400 billion after 2013. That was the only deal ever on the table.

This is a particularly important passage. The entire question of what is and is not our implicit baseline is dismissed as conservative hand-waving. There’s little more to say about that.

We are indeed “talking about actual tax dollars paid by actual people,” etc. And the Toomey proposal implicitly suggests that current policy is the place to start, and that we can and should raise revenue on a net basis relative to current policy. To translate this into slightly different language, Toomey’s proposal wanted to be ensure that net taxes on the rich don’t increase by X amount after 2013 relative to current policy, but rather that they increase by X-n amount. 

(6) A parting thought: President Obama has insisted that he wants to leave current policy tax rates in place for 95% or 98% of U.S. households. Because of that, I suppose we could say that current policy tax rates for said households are “not contentious.” But of course many critics, particularly but not exclusively on the left and the center, suggest that we must increases taxes for a much wider swathe of U.S. households. Do this critics make the president’s implicit baseline more contentious? Or is a baseline only uncontentious if, say, congressional Democrats or Republicans are involved? 

It is easy to see why these ideas might make sense inside of someone’s head. But one of the challenges of our work is that we’re trying to communicate across pretty deep divides in how we understand and asses the world around us.

To be very clear, I was using a current policy baseline. I will try to let you know the baseline I am using in the future, but I might forget. I encourage you to nudge me when I do. Rest assured, I am not trying to mislead you. (Nor were Pear and Steinhauer’s editors, I’m guessing.) 

P.S. Ramesh Ponnuru has characteristically astute related thoughts on his disagreement with David Weigel. 

The Agenda

NRO’s domestic-policy blog, by Reihan Salam.