Testimony of Diane Lim Rogers, Chief Economist, The Concord Coalition Before the Senate Budget Committee, March 1, 2012
Chairman Conrad, Senator Sessions, and Members of the Committee, thank you for giving me this opportunity to testify before you today on the topic of tax reform and the many good things tax reform could accomplish. Your title to this hearing provided a great way to organize my thoughts and my remarks to you today, as we are all contemplating changes to tax policy that will help our economic circumstances in three major ways: (1) to support and expand our economy, (2) to reduce the budget deficit, and (3) to respond to trends in income inequality.
I. Tax Reform to Encourage Growth
Growing the economy through tax policy isn’t as simple as “cutting taxes” to reduce overall tax burdens. Tax cuts all have benefits, but the first thing one learns in an economics class is in a world of scarce resources, we maximize well being by weighing costs against benefits, and at the margin starting from where we are right now. Tax cuts that might benefit particular households and businesses don’t necessarily pass society’s cost-benefit test, even based on a narrower and naïve goal of maximizing GDP.
Our experience with the Bush tax cuts has demonstrated the problem with the simplistic “cut taxes, grow the economy” view. Their major contribution to record-high deficits clearly reduced national saving and economic growth, they were not very effective at growing the supply side of the economy (even according to the Bush Administration’s own Treasury Department), and they are not the kind of tax cuts that provide high “bang per buck” in a recessionary economy. Some of us were questioning the economic wisdom of the Bush tax cuts back in 2001 when our budget projections showed $5.6 trillion in surpluses over ten years (FY 2002-11), worrying that the surpluses should have rather been saved in preparation for the impending retirement of the baby boomers and the associated pressures on the federal budget, as well as for emergencies like wars and natural disasters. Now that we face a forecast of economically unsustainable deficits over the next ten years under a “business as usual” (policy extended) baseline, it’s hard to believe we’ve continued to extend and deficit finance these same costly tax cuts, despite their less than stellar performance.
We now can no longer afford to run suboptimal tax policy that fails to weigh costs against benefits. Constructing smart tax policy within the broader context of fiscal responsibility requires recognizing the connections and tradeoffs between tax rates, tax bases, revenues, public and private saving, and economic growth. The theory behind supply-side tax policy suggests that reducing tax rates encourages taxpayers to work and save and thus is good for the size of the tax base and for revenues. But in practice, tax cuts rarely pay for themselves, as the more extreme Laffer curve version of supply-side economics would suggest. We experienced higher revenues and budget surpluses following the tax rate increases enacted under the Clinton Administration and lower revenues and high deficits following the tax cuts under the George W. Bush Administration. In looking for economically efficient ways to raise revenue, there's room to improve the existing income tax base before we play around with the rate structure or add new tax bases. A tax cut needs to do more than provide just some marginal benefit; there must be enough benefit to make the cut worth its cost, relative to competing demands. If reducing tax rates encourages economic activity but doesn't pay for itself (such as with a rate cut that increases the deficit more than it encourages private saving), it's not necessarily good for the economy.
There is no policy area where conservatives and liberals are further apart than tax policy. Conservatives argue that tax cuts that raise returns to saving and investment, or increase the rewards for work, are always good for the economy, in good times and in bad. Liberals argue that tax cuts primarily raise the incomes of the rich and squeeze out benefits for the poor, and are the least effective type of stimulus when the economy is in a recession. Both sides tend to neglect the adverse long-term economic effects of any type of tax cut that is deficit financed.
The debate is confusing because not all tax cuts are created equal, and the economic effects of those tax cuts differ across three dimensions: (1) the condition of the economy; (2) how the policy affects relative prices (substitution or incentive effects) versus real incomes (income or distributional effects); and (3) how the cost of the policy is paid for. When evaluating the effects of any particular tax cut on the economy, one should ask the following questions.
Where's the Binding Constraint?
In a cyclical downturn, increasing aggregate supply (the productive capacity of our economy) won't do any good, because the problem isn't too little capacity, but too much idle capacity. To increase the level of economic activity, or GDP, we need to increase demand for goods and services so that more of current supply is used. Think of the uses or demand side of the GDP equation — C + I + G + (X-M) — and contemplate what the government's fiscal policy can do to increase consumption (C), investment (I), or net exports (X-M) indirectly via tax cuts and other subsidies, versus increasing direct government purchases of goods and services (G). In terms of the boost to GDP, tax cuts and subsidies are automatically handicapped relative to direct spending, and unless they produce multipliers of greater than 1, they will fall short of the success of dollar-for-dollar direct government purchases.
But in a full-employment economy, fiscal policy is ineffective in increasing demand-side GDP because supply is the limiting factor. GDP can be increased only by encouraging growth in the stock of those productive resources — the supply side of the economy. In this case, we need to ask how we can use fiscal policy to increase incentives to work or to save. How can fiscal policy be reformed to reduce any of the preexisting disincentives and distortions to economic decisions created by current policy?
What Kind of Tax Cut Is It?
Tax cuts typically generate two types of effects on the microeconomic decisions of households and businesses: a substitution effect whereby relative prices are changed to encourage substitution into more lightly taxed activities and away from highly taxed ones, and an income effect whereby the higher cash flow to those receiving the benefits of the tax cut generates a change in their economic activity.
1. Substitution effects and supply-side tax policy. In a full-employment economy, tax policy's effect on relative prices is more important than it is in a recessionary economy. Marginal tax rates are what affect choices concerning the sources and uses of income. Tax cuts that reduce the marginal tax rates on labor or capital income will encourage substitution into greater labor supply or saving, boosting incomes and GDP. Tax cuts without any effect on marginal tax rates, in contrast, do not improve incentives at the margin. An example of a tax cut that reduces average tax rates and boosts average after-tax returns without reducing the marginal tax rate is that of raising the contribution or income limits on tax-preferred savings accounts. Because many higher-income taxpayers are already maxed out on the tax-preferred options, and might continue to be even after the higher limits, increasing the availability of the tax subsidy for those households can cause shifting of existing savings (moving money out of taxable accounts into tax-free ones) without necessarily creating any new savings. The policy would generate positive income effects for these taxpayers even without any substitution effects.
Empirical research on the significance of substitution effects shows that higher-income households are more responsive to changes in marginal tax rates than lower-income households, probably because they can fine-tune their work hours more easily, and because the relative price change itself is usually larger at higher income levels given the progressive, graduated rate structure of the federal income tax. In fact, many lower-income households are entirely exempt from the federal income tax and so are completely unaffected by changes in marginal income tax rates. This has encouraged economists to suggest that flat rate tax systems (with a single marginal tax rate above some exemption level of income) would generate positive and sizable supply-side effects on labor supply and saving. But hold that thought, because how much the tax cut would cost in terms of lost revenue and the deficit would affect the supply side of the economy as well.
Increased supply-side incentives can also be achieved by reducing differences across marginal tax rates on different sources and uses of income. Broadening the income tax base by reducing tax expenditures would raise the overall average tax rate but would do so by raising marginal tax rates only on those sources and uses of income that are currently undertaxed in the definition of taxable income. By reducing the tax advantage to those currently undertaxed forms of income, the substitution effects away from higher-taxed income would actually be reduced and that type of income would be encouraged, even as the economy-wide average tax rate rises.
Must one be a supply-side economist to believe in the existence of these supply-side types of responses? No. Economists of all stripes broadly agree in the theory that households and businesses respond to relative price changes when those agents are given the opportunity and have the capacity to do so. Economists also agree that marginal tax rates matter in terms of their incentive effects. The debate over how valuable to the economy supply-side tax policy can be is largely over how large those substitution/incentive effects are in the real world, relative to the other economic effects of tax policy.
2. Income effects and demand-side tax policy. In a recessionary economy, the income effects of tax policy matter more. The distribution of the dollar benefits of a tax cut will affect how much the demand for goods and services is stimulated. Tax cuts focused on the top marginal tax rates don't deliver anymore dollars to lower-income households who have the highest propensities to consume. The effect on relative prices matters less than the effect on the levels and distribution of after-tax income. In fact, an economy-wide tax cut isn't a prerequisite of a successful demand-side tax cut. Consider a purely hypothetical and purely redistributive (income-effects-only) Robin Hood policy that increases taxes on the rich and gives the proceeds to the poor. This would increase aggregate demand in the economy by simply shifting income away from savers toward non-savers. Note that this is quite contrary to the optimal strategy in a supply-side tax cut designed to increase labor supply and saving.
Perhaps even more curious, fiscal policies that might seem ineffective or unjustified in terms of incentive effects (such as Social Security cost of living adjustment makeups for seniors, tax breaks for new homeowners, and the "Cash for Clunkers" program) might nonetheless have a high bang per buck in terms of stimulating aggregate demand in a recessionary economy. Even if those policies actually do nothing to encourage the economic activity they're ostensibly designed to, as long as they steer dollars to households with high marginal propensities to consume, they can nevertheless turn out to be pretty effective in stimulating demand.
On the flip side, we shouldn't worry much about higher taxes having large dampening effects on demand if those tax increases are mostly on higher-income households with low marginal propensities to consume. We also shouldn't be too concerned about the potential recessionary effects from tax increases that would take effect only after the economy is back to full employment.
The timing of tax cuts matters. At either the business level or household level, temporary tax cuts are likely to have a greater stimulative effect on the demand for goods and services than permanent tax cuts, because the timing of transactions is relatively easy to change — according to University of Michigan economist Joel Slemrod's hierarchy of responses. A temporary tax cut will generate a large effect as the qualified activity is shifted forward whenever a tax cut has a deadline, even if the same tax cut, because it is only temporary, has a much smaller or negligible long-term effect on the components of aggregate supply.
How Is the Tax Cut Being Financed?
1. Deficit financing sometimes helps and sometimes hurts. In a recessionary economy, deficit financing will increase the countercyclical stimulative effect of any particular tax cut on aggregate demand by promoting consumption of goods and services in excess of personal incomes. But that doesn't mean any deficit-financed tax cut (or spending) makes for the best stimulus, because there are longer-term economic costs still associated with the deficit — the debt has to eventually be repaid in higher taxes or reduced spending in the future. That puts limits on the amount of deficit-financed stimulus that's economically justified. We want to maximize the economic bang for the buck in deficit-financed stimulus, so fiscal responsibility requires that we determine a level of deficit spending we deem worth it, put high bang-per-buck spending or tax cuts at the front of the line (ranking fiscal policies from most effective to least), and draw the line at the credit limit we've implicitly established.
In a full-employment economy, however, deficit financing represents a dollar-for-dollar decrease in public saving, making it harder for the tax cut to increase national saving unless private saving is encouraged by more than the cost of the tax cut. This is not quite as high a standard as the tax cut paying for itself (as proposed by the Laffer curve) — which is 1/t times as hard (t being marginal tax rate on private returns to saving). This is why the Bush tax cuts have been evaluated as a net negative for economic growth by William Gale and Peter Orszag within the first few years of the Bush tax cuts, and by Gale more recently.  It also explains why increased tax rates during the Clinton Administration coincided with higher, not lower, economic growth.
The choice to deficit finance now does not permanently avoid a tougher choice. Deficit-financed tax cuts do not pay for themselves, and they imply inevitably higher taxes or lower spending in the future. This intergenerational redistribution is another economic effect of the tax cut.
2. On the other hand, paying for the tax cut sometimes hurts and sometimes helps. In a recessionary economy when the goal is increasing current consumption, offsetting the cost of the tax cut with spending cuts or tax increases will reduce the net stimulative effect on aggregate demand for goods and services. The more the offset affects lower-income households (those most constrained), the larger the negative effect. Paying for a tax cut going to primarily high-income households with a cut in spending that benefits primarily low-income households would likely be contractionary, not stimulative.
In a full-employment economy, however, finding budgetary offsets to the cost of a tax rate reduction is likely to be better for encouraging aggregate supply and boosting GDP than deficit financing. That's because the deficit reduces public saving dollar for dollar, while empirical evidence has shown that the adverse effect of the offsetting policy on private saving is likely to be something less than dollar for dollar.
Tax Cuts Matter, but Aren't One Size Fits All
So are tax cuts good for the economy? It depends. As countercyclical policy during a recession, deficit-financed tax cuts can help stimulate demand, but deficit-financed spending is likely to be even more effective if it is deliberatively targeted toward lower-income households. As supply-side policy during periods of full employment, tax cuts are most effective if they increase incentives at the margin to work and save — that is, by reducing marginal tax rates or leveling rates across different forms of income — but any deficit financing is likely to produce a net negative effect on national saving.
That's why a revenue-raising (relative to current policy, or revenue-neutral relative to current law) tax reform that reduces or levels out effective marginal tax rates and broadens the tax base at the same time is such a win-win-win formula:
Tax cuts are always an attractive option in the political world where budget constraints are often ignored. But in the real world and in real time — where budget constraints bind and opportunity costs matter — policymakers must be mindful of the fact that the effectiveness of any particular tax cut depends on our economic circumstances and goals and how those mesh with the structure of the tax policy.
II. Tax Reform to Reduce the Deficit
When it comes to tax policy, the economy, and the deficit, it is common to hear this simple line of reasoning: “cut taxes, to grow the economy, to reduce the deficit.” In the previous section I explained why the first causal relationship doesn’t usually hold true. And for this section, the bad news is that the second causal relationship doesn’t hold either. We can’t just “grow our way out” of budget deficits; we’ll actually have to make structural changes to our spending and revenue programs.
CBO has shown that our current (still) record-high deficits are largely structural, not just cyclical. The projections show that even as the economy continues to recover, deficits will come down only modestly before they rise again dramatically. The current business cycle has done nothing to change the longer-term path of the federal budget, other than to temporarily create another short-term demand for federal dollars.
Economists agree that the federal budget is on an unsustainable path and that for the continued health of the economy, deficits must eventually come down to levels lower than the growth rate of the economy (allowing the debt/GDP ratio to be stabilized). Even though a sizeable level of deficit spending over the next one or two years can be justified to support the economic recovery, a commitment to bring down deficits to lower, more sustainable levels over the next decade is essential not just for longer-term economic growth but for shorter-term economic stability (via the confidence of global investors in the U.S. economy).
Tax policy has to be part of the solution. It is true that the greatest pressures on the federal budget in the decades to come are in the entitlement programs because of the aging of the population coupled with rising per-capita health costs. But it is hard to see how our society would choose cuts in real, per-capita benefits of the magnitude necessary to both achieve sustainable deficits and keep revenues at the historical average. And even if we would choose to do so, we would never do it very soon; entitlement reforms would have to be phased in much more slowly than tax reforms could take effect.
The historical average level of revenues/GDP has very little bearing on what the right level of revenues is going forward. The right level of revenues is that which is adequate to pay for the government we desire. (And the right size of government is that which we are willing to pay for.) Given the dramatic changes in the structure of our population and the continued growth and evolution of our economy, it is difficult to see how what was right over the past 40 years—and it wasn’t even quite adequate then—could be right over the next 40 years.
The Current-Law Baseline: Why It Matters and Why Budget Rules Should Respect It
Adjusting the CBO current-law baseline to construct the Concord Coalition’s “plausible baseline” (a “business as usual” projection) nearly quadruples the ten-year deficit from $3.1 trillion to $11.8 trillion—with $6.5 trillion of the $8.7 trillion difference due to tax policy and the plethora of expiring, deficit-financed tax cuts in current law. If current tax cuts were extended and deficit financed (as usual), federal revenues would never rise above 17.5 percent of GDP over the ten-year budget window. In contrast, under current law where expiring tax cuts expire as scheduled, revenues rise to 21 percent of GDP by the end of the ten-year window, which is consistent with an economically-sustainable level of deficits over the next 10-20 years according to CBO. So whatever we do on the tax policy front, we should commit to achieving current-law revenue levels.
There are many tax policies that would be consistent with the current-law baseline level of revenues. I have characterized the three main approaches as: “do nothing” (let the Bush tax cuts expire as scheduled at the end of 2012), “do it big” (broaden the tax base by reducing tax expenditures, paying for lower tax rates), and “do it to the rich” (such as via a surtax on millionaires and/or large corporations). Each approach has different relative advantages regarding their economic effects and political attractiveness. The best economic effects would come from increases in revenue accomplished through progressive base broadening/reduced tax expenditures. We could do any combination of the approaches, and, most significantly for this Budget Committee: all would be encouraged in practice with a commitment to strict, no-exceptions, pay-as-you-go rules—on new or extended tax cuts and not just spending increases. Such a commitment would immediately get us back on the path to sustainable deficits.
In addition, to encourage a base-broadening approach to tax reform, the Budget Committee could make further use of their reconciliation instructions to the committees of jurisdiction. On the expiring tax cuts, for example, the budget committees could require the tax-writing committees to come up with a level of revenues at or at least closer to the current-law baseline, effectively requiring that a certain portion of the cost of extending expiring tax cuts be offset with other revenue increases. And in their policy statements or other report language accompanying the budget resolution, the budget committees could recommend that at least x% of those offsets (up to 100%) be comprised of reductions in tax expenditures as listed and scored by the Joint Committee on Taxation.
Is Deficit Reduction Contrary to Economic Growth?
But what about the economy? Can we really pursue deficit-reducing revenue increases without jeopardizing economic growth—goal #1?
For both the short-term and longer-term economic goals, reducing the deficit is not necessarily contrary. The design and the timing of the policies are crucial though. For the short term, specific and credible policy commitments to deficit reduction—even if the policies take effect later—would immediately help to keep U.S. credit-worthiness high and interest rates low. For the longer term, deficit reduction is essential to increase national saving and hence economic growth.
Even if continued deficit spending, in general, is justified in our still-recovering economy, it doesn’t imply that all deficit-financed spending is worth its cost. We should look for opportunities to decrease spending on the least-effective (perhaps even counter-productive) and least-responsive activities, in order to either free up resources for more effective uses or for deficit reduction, considering both short- and longer-term goals. We should be maximizing “bang per buck”—broadly defined—to extend both to the relationship between deficits and national saving, and to longer-term tax policy.
By far the largest and most reliable connection between tax policy and economic growth over the past several decades has been through the beneficial effects of surpluses and the adverse effects of deficits on national saving. The budgetary effects of tax policies have mattered far more than the microeconomic responses to the policies, because the former have been huge and certain, while the latter have been small and uncertain.
The Bush/Obama Tax Cuts
This seems a good place to remind the Committee that the Bush tax cuts continue to be the single most costly policy proposal in President Obama’s budget. This is a choice policymakers have to make and legislate: whether to extend some or all of them, and if and how they will be paid for.
When the CBO completes their analysis of the President’s budget, they will show the effects of the proposals relative to current law. That presentation will make clear that the single most costly proposal in the President’s budget is still the deficit-financed extension of all but the high-end Bush tax cuts, which by the Administration’s own numbers costs $1.3 trillion ($2,173 billion to extend all the tax cuts minus $849 billion in high-end ones that would be allowed to expire)—without counting the cost of associated continued AMT relief (another $1.9 trillion), or associated interest costs, which the OMB tables do not break down but CBO says would be about $790 billion for the full complement of extended tax cuts. Given that only $849 billion out of $4.5 trillion in tax cuts, or less than 19 percent, would be allowed to expire, the associated interest costs from the extension of the remaining cuts are likely to be over $600 billion over ten years. (This is consistent with the OMB showing in Table S-8 that the debt service costs associated with all their adjustments to get from the current-law baseline to their “adjusted baseline” are $640 billion.)
Gathering these numbers from all the different places in the Obama budget tables they appear and adding them up, we find that leaving aside any new tax policies for the moment, the Obama Administration is proposing to extend expiring tax cuts at a total cost (including associated interest costs) of almost $4 trillion over the next ten years alone. In other words, the President is proposing $4 trillion worth of (just) old “Bush” tax cuts in his budget. That figure does not include the costs (or revenue gains) of any new tax policies the President proposes, which speak more in terms of what they target than in how much money they lose or save.
Back in 2001 when we faced $5.6 trillion in surpluses (FY2002-11) there were still at least some of us saying a big tax cut was not a good idea, given the impending retirement of the baby boomers and associated effects on the fiscal outlook. (I was writing the chapter on the fiscal outlook in President Clinton’s final economic report at the time.) If not paying for the tax cuts was not a good idea back then, why would it be a good idea now?
Reducing Tax Expenditures Is Reducing the Size and Scope of Government
Those who oppose deficit reduction on the revenue side of the budget usually assume higher revenues will lead to larger government. But the holes in our income tax base—the special exemptions, deductions, credits, and preferential rates—amount to over $1 trillion/year (about 90 percent of this in the individual income tax and 10 percent in the corporate), nearly as much as all of discretionary spending combined. While not all tax expenditures can be unambiguously labeled “spending in disguise,” most of them can. The Tax Policy Center’s Donald Marron and Eric Toder have estimated that about 65 percent ($600 billion in 2007) of tax expenditures are “replacable by a spending program.” Thus, “filling out” the tax base by reducing these tax expenditures would level out and support lower marginal tax rates (reducing the economic distortions caused by taxes), and reduce both the deficit and the effective size of government.
III. Tax Reform to Promote Fairness
Reducing the deficit by broadening the tax base is also appealing if one is concerned about the inequality in the distribution of income and the “vertical equity” of the tax system. Tax expenditures are very different from other types of (direct) government spending, because they’re spending (“poking holes”) through the progressive income tax system, which affects high-income households far more than others. This makes tax expenditures some of the most regressive government subsidies around, and cutting them one of the most progressive ways we could reduce the size of government and the deficit at the same time.
Income inequality has been increasing, particularly at the very top, and so it seems our desire to ameliorate these changes by trimming any government spending that goes to the rich. But there’s not much on the spending side of the budget that directly benefits higher-income households, and even we were willing to cut Social Security and Medicare benefits on the rich, these would be small changes relative to their incomes. The only place in the federal budget where benefits go disproportionately to the rich (relative to their incomes) is via the tax system. Thus, the only route to substantially reduce income inequality at the top of the distribution is to raise tax burdens and reduce tax subsidies there.
Raising taxes on the rich means the policy should be revenue gaining, helping both of the previous goals: deficit reduction, and hence economic growth (via effects of lower deficits on national saving). The fairness goal doesn’t support revenue-neutral tax reform, unless revenue gained from the rich is used to pay for tax cuts to middle-income households.
For a long time, President Obama has defined “the rich” as those households with annual incomes over $250,000. Limiting tax burden increases to households with over $250,000 in income, however, can be terribly constraining in terms of the dual goals of raising large amounts of revenue in economically efficient ways, especially if changes are limited to statutory tax rate increases. A Tax Policy Center analysis has illustrated the challenge in keeping the current, full-of-holes definition of the tax base, limiting burden increases to these upper-income households, and raising enough revenue to get to economically-sustainable deficit levels. Marginal tax rates on the rich are then forced to climb to levels that would indeed be around the peak of the Laffer curve.
Still, there seems to be a lot of public sentiment to intentionally “target” tax increases on the rich. Warren Buffett’s low effective tax rate (especially relative to that of his secretary) has inspired a near consensus that “the rich” do not pay their fair share of taxes and that any future tax reforms, whatever the levels of revenues produced, should steer more of the tax burden onto those lucky few. For some, the definition of “the rich” has moved into the “millionaires and billionaires” category now. (And that naturally means that more of us will be supportive of the idea of raising taxes on (just) “the rich.”) The Administration’s interpretation of the Buffett rule, as described in their budget (pg. 39) is that “no household making over $1 million annually should pay a smaller share of its income in taxes than middle-class families pay”—even though their specific tax proposals largely honor their long-held $250,000 threshold.
As I wrote in Tax Notes (in a column I titled “Who Wants to Tax a Millionaire”), there are lots of ways to raise tax burdens on millionaires, some economically smarter than others. While a specific surtax on incomes over $1 million seems like the simplest and starkest way to do it, such a policy would raise the distortionary effects of the tax system by raising effective marginal tax rates on income already counted as taxable. Economists would prefer to raise millionaires’ taxes by including more of their total income in what counts as (fully) “taxable income,” or by reducing the size of some of the subsidies given to them through various tax expenditures.
Talk of millionaire tax increases naturally brings proposals for some form of “millionaire surtax” first to mind, but there are many other ways we could increase the taxes that millionaires pay. We could let overall income tax rates increase by, for example, letting the Bush tax cuts expire—either all of them, or just the upper-income provisions. We could tax capital gains and dividend income, and other types of currently tax-preferred income, at the same rate as “ordinary” (labor) income. Finally, we could broaden the tax base to reduce tax expenditures that largely benefit higher-income households, such as limiting itemized deductions to a tax-rate ceiling—a policy the Obama Administration has repeatedly proposed in their budgets.
All these options would raise tax burdens on millionaires—both in absolute and relative terms—but would produce different economic effects in terms of the concentration of the millionaires’ higher tax burden, the level of potential revenue raised, and the incentives to work or save.
The more targeted the tax burden to millionaires, the more limited the total revenue potential, and the more we have to worry about adverse effects on incentives via higher marginal tax rates. The more a proposal‘s tax base is limited to currently-taxable income above a million dollars, the higher the effective marginal tax rate will be at those millionaire income levels, for a given level of revenue. The exception to the targeting versus incentives tradeoff is where we can find ways to broaden the tax base that concentrate higher burdens on millionaires by limiting the benefits of tax expenditures (such as the percentage applied to itemized deductions) or by phasing out benefits completely at higher income levels.
Proposals to let all of the Bush tax cuts expire may seem contradictory to the “Buffett rule” of raising the tax burden on (only) millionaires, and yet it would raise the most revenue from millionaires—and not just a lot more revenue overall. It’s a reminder that even millionaires benefit from lower-income tax cuts, too; in fact, households who are fully above the lower brackets benefit more in dollar terms from reduced lower-bracket rates than do households who fall in the middle of them.
Tailoring Base-Broadening Tax Reform for the Buffett Rule
Although most proposals to reduce tax expenditures in “across the board” ways will naturally burden higher-income households the most, there are ways to explicitly limit the higher burdens to households above a certain tax bracket or income level, to make any base-broadening proposal as progressive as one wants to.
One example of a base-broadening revenue proposal that limits higher burdens to higher-income households is the limit of itemized deductions to 28 percent, which President Obama has now proposed in all four of his budgets. But the President went further in this year’s budget to propose reducing a wider array of tax expenditures to higher-income (still above $250,000 income) households, including the exclusion of employer-provided health benefits. The proposal sticks to the idea of making sure upper-bracket taxpayers benefit no more from these tax expenditures than they would if they were only in the 28 percent bracket. Administratively, limiting exclusions isn’t as easy as limiting deductions, however, because we would need additional information reported on individual tax returns—the addition to taxable income of the previously excluded type of income. But the expanded proposal does improve the bottom line significantly; compared with the limit on itemized deductions only, this version raises about double the amount of revenue, $584 billion over 10 years according to the Administration.
New as well this year in the President’s budget is the proposal to not just let the tax rate on dividends go up to a less-preferred rate but to eliminate the preference entirely for high-income households, treating dividends as ordinary income for them. So even with the Administration’s consistent push to let the upper-end Bush tax cuts expire, they are expanding the impact of the expiration this year, likely encouraged by their new emphasis on the Buffett Rule.
(Even) Broader Tax Increases Still a Preferable Way to Raise Taxes on “the Rich”
Raising taxes on the rich by reducing (just) their tax expenditures is economically preferable to raising taxes on them by raising their rates, but raising taxes on the rich by raising taxes more generally—without discriminating by income level or type of person or business—is better still. Economists would prefer to avoid the awkwardness, complications, and inefficiencies that result from targeting tax increases (or tax cuts, for that matter) to a very specific category of taxpayers. Raising taxes overall, by either letting (even all of) the lower Bush tax rates expire or reducing tax expenditures, would naturally burden higher-income households disproportionately and hence enhance the progressivity of the income tax system, simply because the rich benefit the most from current tax rate cuts (even lower-bracket ones) and tax subsidies.
“Broader” tax reform strategies—which would apply a broader and more uniform definition of the tax base to a broader (larger) number of people and businesses—would allow tax rates to stay low while the deficit is reduced more significantly. And because such deficit reduction would be achieved by efficiency-enhancing yet progressive tax policy changes, the dual goals of adequately supporting demand in a still-recovering economy and encouraging the supply of productive resources over the longer term would be achieved.
Conclusion: A “Trilemma” for Tax Reform?
I have recently heard the three tax reform goals the Committee outlined for this hearing—economic growth, deficit reduction, and fairness—referred to as a “fiscal trilemma,” with the implication that achieving all three goals simultaneously might be difficult. But as I’ve described today, I believe that base-broadening, revenue-raising, tax-expenditure-reducing tax reform can easily be consistent with all three goals.
The Budget Committee could play a very critical role in achieving fiscally-responsible tax reform by enforcing current-law baseline revenue levels through effective budget rules and instructions to the tax-writing committees.
Politically arbitrary labels such as the choice of budget baselines matter a lot, because politicians need these simple metrics to demonstrate their success as policymakers. Republicans will always want to be known as the tax cutters, while Democrats will always push for more progressive taxation. Setting a goal of sticking to the current-law revenue baseline, which is achieved by base broadening rather than higher rates, is a way of honoring the seemingly inconsistent tax policy goals of both parties. It seems reasonable that policymakers should start from a current-law standard, because making changes relative to current law is their legislative responsibility, after all, even if the policy-extended baseline is a more accurate reflection of “business as usual.”
Economically, however, it doesn't matter if we view such tax policy as raising revenue relative to a current-policy baseline or as keeping revenue constant relative to a current-law baseline. All that matters is that the policy raises enough revenue to keep deficits at an economically sustainable level — where the economy's growth has a chance to keep up with the growth of the debt — while minimizing the distortionary effects of taxation.
No matter how one might choose to interpret it — as a policy change consistent with Republican goals of reducing tax rates and government's interference with market decisions (the subsidies given out through the tax code), or as one consistent with Democratic goals of reducing the deficit by progressively raising revenue as a share of our economy — this type of bipartisan tax reform will be crucial to achieving fiscal sustainability. For now this seems the most promising area for significant progress on deficit reduction to happen relatively quickly, while we continue to work on economically smart ways to control spending in the rest of the federal budget.
 See chapter 2 (pp. 79-92) in the Economic Report of the President, January 2001 (http://www.gpoaccess.gov/usbudget/fy02/pdf/2001_erp.pdf).
 Joel Slemrod, “Tax Systems” in NBER Reporter, Summer 2002 (http://www.nber.org/reporter/summer02/slemrod.html) and “Income Creation or Income Shifting? Behavioral Responses to the Tax Reform Act of 1986,” American Economic Review, 85(2), May 1995 (http://www.jstor.org/pss/2117914).
 William G. Gale and Peter R. Orszag, “Bush Administration Tax Policy: Effects on Long-Term Growth,” Tax Notes, October 18, 2004 (http://www.taxpolicycenter.org/UploadedPDF/1000698_Tax_Break_10-18-04.pdf) and William G. Gale, “Five Myths about the Bush Tax Cuts,” Washington Post, August 1, 2010 (http://www.washingtonpost.com/wp-dyn/content/article/2010/07/30/AR2010073002671.html).
 See “The Concord Coalition Plausible Baseline,” The Concord Coalition, updated January 2012: http://www.concordcoalition.org/concord-coalition-plausible-baseline.
 See Economic Report of the President, January 2001, op. cit.
 Donald B. Marron, “How Large Are Tax Expenditures?”, Tax Notes, March 28, 2011 (http://www.urban.org/uploadedpdf/1001526-Expenditure-Estimates.pdf).
 Donald Marron and Eric Toder, “Measuring Leviathan: How Big Is the Federal Government?”, presentation at Loyola Law School, January 14, 2011 (http://events.lls.edu/taxpolicy/documents/PANEL2MarronToderSizeofGovernment-presentationFinal01-06-11.pdf).
 Rosanne Altshuler, Katherine Lim, and Roberton Williams, “Desperately Seeking Revenue,” Tax Policy Center, January 2010 (http://www.taxpolicycenter.org/publications/url.cfm?ID=412018).
 Diane Lim Rogers, “Who Wants to Tax a Millionaire?,” Tax Notes, February 6, 2012, pp. 725-7.
 Stolen from the title of an upcoming tax reform conference at Tulane University (http://www.ssrn.com/update/ern/ernann/ann12048.html).