The Trump Tax Plan and National Priorities
Plan Would Shrink Government to Truman-era Levels While Spending $3.2 Trillion on Tax Cuts for Households Making Over $1 Million a Year
The large tax-cut proposal from Donald Trump, in conjunction with his call for balancing the budget, would dictate low levels of government spending not seen since 1948. It also would benefit households making over $1 million to an unprecedented degree, and do so almost certainly at the expense of low- and middle-income people — with programs important to the well-being of many or most Americans shrinking markedly or disappearing altogether. The consequences of the Trump tax plan to core programs, and to low- and middle-income Americans, would be harsh. Even if policymakers didn’t achieve budget balance under the Trump plan and just offset the tax cuts’ costs, the consequences to core programs, and to low- and middle-income Americans, would be harsh.
These findings emerged from our earlier examinations of the Urban-Brookings Tax Policy Center’s (TPC) estimates of the cost of the Trump tax plan and how the plan would affect households at different income levels. The earlier analyses have been updated and consolidated here to reflect more recent discussion of the Trump plan and to drop information about Senator Ted Cruz’s plan, now that he has suspended his presidential campaign.
The Trump plan would reduce revenues by $9.5 trillion over the next ten years, according to TPC. Based on the TPC data, we estimate that the tax cuts going to millionaires would result in a revenue loss over the coming decade of approximately $3.2 trillion. (“Millionaires,” as used here, refers to households with annual incomes over $1 million, rather than to the amount of assets that households may hold.) Other findings include:
Trump’s Written Tax Plan Stands
In a series of interviews in May, Mr. Trump and his representatives made several statements about his tax-cut plan that have led to some confusion about the plan’s design and intent. As part of this discussion, Mr. Trump indicated that while he expected the plan to lead to, and serve as the basis for, negotiations, he regards the proposal he released in September 2015 as the “optimum plan.”* In addition, his campaign eventually announced that the plan would not be changed.** Accordingly, this analysis examines the proposal as announced by Mr. Trump and as described on his campaign’s website. It does not attempt to speculate about future changes that could be made to the plan.
* Nicki Rossoll, “Donald Trump Walks Back Tax Plan, Saying ‘It’s Going to Be Negotiated,’” ABC News, May 8, 2016.
** Jackie Calmes, “Donald Trump Won’t Alter Tax Plan, Spokeswoman Says as Confusion Reigns,” The New York Times, May 12, 2016.
- Measured as a percent of the economy (gross domestic product, or GDP), revenues would fall to their lowest level since 1950 under the Trump tax plan. In 1950, Social Security spending was only 0.3 percent of GDP, and Medicare did not yet exist. Today, Social Security and Medicare expenditures amount to 8.1 percent of GDP and continue to rise, primarily due to the aging of the population. (See Figure 1.)
- Millionaires would receive tax cuts that average about $380,000 in 2025, the last year for which TPC has estimated the distribution of the Trump plan. (All average tax-cut figures in this report are adjusted for inflation and expressed in 2016 dollars.) The after-tax incomes of these high-income individuals would increase by 17.9 percent.
- The tax cuts for the middle class would be far smaller. The middle fifth of households would receive tax cuts in 2025 averaging $2,900, raising these households’ after-tax incomes (before considering the effects of offsetting budget cuts) by 4.9 percent, less than one-third of the percentage gain for households making over $1 million.
- Although households with incomes over $1 million contain less than 1 percent of the population, in 2025 they would receive 38 percent of the Trump tax cuts. In contrast, the bottom 80 percent of the population would receive 32 percent of the cuts. Thus, in aggregate, millionaires would gain more than the bottom 80 percent of households combined.
- People at the very highest income levels would fare best of all. The TPC data indicate that in 2025, the richest 0.1 percent of the population — those in households with incomes of more than $5.2 million — would receive an average tax cut of $1.4 million each. They would secure a larger share of the tax cuts than the bottom 60 percent of the U.S. population.
These figures likely understate the disparate effects of the proposed tax cuts on Americans at different income levels, because the true distributional effect of tax cuts can’t be assessed without also considering how their costs would be offset. The tax cuts would need to be offset sooner or later; otherwise, the government’s annual deficit would exceed 10 percent of the economy in 2026. Further, Mr. Trump has said he would balance the budget in the relatively near future, essentially committing himself to offsetting the cost of his tax plan (in addition to finding additional savings to balance the budget).
Assuming these offsets would occur through reduced government spending rather than tax increases, low- and middle-income households would lose significantly more from the offsets than they would gain from the tax cuts themselves. Only a tiny fraction of spending cuts could come from cutting program spending benefiting millionaires, given how few such households there are and given that federal program spending is dispersed fairly evenly across the population.
Indeed, the only way to offset the tax cuts so they do not result in income being transferred from the rest of the population to people at the top would be to enact offsetting tax increases that reduced the incomes of millionaires by about $3 trillion over the coming decade. The reason for this is simple: offsets both large enough to offset the costs of the Trump tax cuts and progressive enough to avert an income transfer up the income scale simply cannot be accomplished just or primarily by reducing government spending; offsets of this nature can be secured only if much or all of the tax cuts for people at the top are effectively cancelled out.
Thus, it’s all but certain that once its costs were offset, the Trump plan would result in the transfer of several trillion dollars over the next decade from low- and middle-income households to those at the top of the income scale.
The resulting reductions in government programs (from offsetting the tax plan’s costs) would be sharp:
- Under the Trump plan, offsetting the tax cuts through spending reductions while also balancing the budget in 2026 would require cutting program spending (i.e., federal spending for everything except interest payments on the debt) to its lowest level as a percent of the economy since 1948.
- Balancing the budget as well by 2026 under the Trump plan would require cutting all government programs — including Social Security, Medicare, and defense — by about two-fifths if all programs were cut by the same percentage. Balancing the budget without cutting Social Security, Medicare, and defense would require eliminating essentially the rest of government.
- If spending were reduced enough to pay for the tax cuts but not to balance the budget — and Social Security, Medicare, and defense were not touched — all other programs would need to be cut by 53 percent.
- Offsetting the cost of the Trump tax cuts would require cuts in government programs far larger than the deep reductions contained in the 2017 budget plan that the Republican majority on the House Budget Committee approved in March.
As dramatic as these figures are, they understate the pressure that the proposal would place on many government programs. Mr. Trump has also proposed to increase spending in several areas, including increasing spending on veterans’ programs by $500 billion over the next decade. Offsetting the cost of such increases on top of the cost of the tax cuts would require even deeper cuts to other programs.
The price tag of the tax cuts just to millionaires under the Trump plan — more than 1.5 percent of GDP in 2025 — can also be compared to the cost of other potential policy initiatives (whose costs would also have to be offset). Under the Trump plan, the annual cost of the tax cuts for the $1 million-and-above group would be:
- Greater than the cost of ensuring the solvency of Social Security and Medicare for the next 75 years, under those programs’ most widely used actuarial projections, and greater than the cost of ensuring solvency of the Social Security trust fund (but not the Medicare trust fund) even under more pessimistic estimates of the finances of these programs;
- Larger than the entire “poverty gap,” which is the total amount of income needed to raise every individual now living in poverty up to the official poverty line;
- Seven times the total resources that the federal government now devotes to medical research; and
- Significantly larger than the amount needed to return the nation’s crumbling infrastructure to a state of good repair.
Some may question the magnitude of the spending cuts that would be required by claiming that the tax cuts can boost economic growth sufficiently to mostly or fully pay for themselves. This line of argument lacks foundation, as the Appendix explains. The Appendix also discusses other estimates of the costs of the Trump tax plan. Of note, TPC’s estimate of the costs of the Trump plan is lower than other available estimates, including a “dynamic” estimate by the Tax Foundation that reflects its assessment of the plan’s effects on economic growth.
How Low Would Revenue Fall?
In 2026, revenues would amount to 14.1 percent of GDP under the Trump plan — far below projected revenues of 18.2 percent under current law and even further below projected spending that year of 23.1 percent of GDP. In 2026, revenues would be $1.1 trillion lower than under current law.
As a result, revenues would drop to a much smaller share of the economy than has typified the post-World War II era. Going forward, moreover, revenues will need to exceed their historical levels to help pay for the rise in government costs as the baby-boom generation retires.
Under the Trump plan, revenues as a percent of GDP would fall to their lowest level since 1950. In 1950, Medicare had not yet been created, and Social Security was in its infancy, with Social Security spending amounting to just 0.3 percent of GDP. Medicare and Social Security spending together now totals 8.1 percent of GDP and is projected to rise to 9.7 percent of GDP in 2026. In other words, the enhanced national commitment to assisting the elderly and the aging of the population necessitate a considerably larger revenue base than existed more than six decades ago.
Distribution of the Tax Cuts
TPC has analyzed the distribution of the Trump tax cuts in calendar years 2017 and 2025. The latter year’s distribution better exemplifies the tax cuts’ permanent effects as the plan would be fully implemented by then; thus, this analysis places greater emphasis on those figures. As Appendix Table 1 shows, the TPC data for 2025 indicate:
- Households (technically “tax units”) in the bottom fifth of the population would receive average tax cuts of $169 under Trump, increasing their after-tax incomes by 1.1 percent, before considering the effects of offsetting policies to pay for the tax cuts.
- The middle fifth of households would receive tax cuts averaging about $2,900. Their after-tax incomes would increase by 4.9 percent from the tax cuts before considering how the cuts would be financed.
- Under the Trump plan, the average tax cuts for millionaires would equal about $380,000 — approximately 130 times the average tax cut for the middle fifth. Millionaires’ after-tax incomes would rise by an estimated 17.9 percent, or 3.6 times the percentage gain for the middle fifth.
Calculating the share of the tax cuts that various income groups would receive similarly shows the sharply disproportionate nature of the tax cuts. (See Figure 2.) In 2025, a projected 0.8 percent of the population will live in households with incomes exceeding $1 million, but such households would receive 38 percent of the Trump tax cuts. This would be greater than the share of the tax cuts that the bottom 80 percent of the population would receive.
Top 0.1 Percent Would Be Biggest Winners
TPC also examined the tax cuts that would go to the richest 0.1 percent of the population, those with annual incomes exceeding $5.2 million (in today’s dollars). The data show that in 2025, under the Trump plan:
- These households would receive tax cuts averaging $1.4 million. This segment of the population would receive a larger percentage increase in after-tax income (18.3 percent) than any other group.
- The nation’s richest 0.1 percent of people would receive 18 percent of the tax cuts — more than the plan’s combined tax cuts for the bottom 60 percent of the population.
Offsets Would Likely Transfer Several Trillion Dollars From Low- and Middle-Income People to Millionaires
By combining TPC’s estimates of the share of tax cuts going to various income groups with its estimates of the revenues lost due to the tax cuts, we estimate that the tax cuts to millionaires alone would produce a revenue loss of $3.2 trillion over the next decade under the Trump plan. The tax cuts for the top 1 percent, which includes but is not limited to millionaires, would cost $3.4 trillion. (Due to data limitations, these calculations represent reasonable approximations, not precise figures.)
The Trump campaign has provided little information on how the cost of the tax cuts would be offset. If the tax cuts are not offset, the annual budget deficit would increase sharply, to about 10 percent of GDP, and the national debt would jump as well. Mr. Trump has promised both to offset the cost of the tax cuts and to balance the budget.
Quite simply, absent a reversal of the tax cuts themselves, the $3.2 trillion in revenue lost due to the tax cuts for millionaires could not come close to being offset by reductions to spending programs for millionaires. Spending programs are much more evenly distributed across the population than the Trump tax cuts would be, and millionaires constitute less than 1 percent of the nation’s population. Only a tiny fraction of the tax cuts could be offset through spending reductions focused on this group.
Thus, the lion’s share of the spending reductions necessary to pay for the tax cuts would need to come from programs benefiting low- and middle-income households, costing those households substantially more than the tax cuts would help them. This makes it all but certain that once offset, the Trump plan would effectively necessitate the transfer of several trillion dollars over the next decade from low- and middle-income households to millionaire households.
The Magnitude of the Offsetting Spending Cuts
To further illustrate the potential impact of the Trump plan, two hypothetical approaches to reducing spending are examined: one in which all spending programs are cut by the same percentage and one in which combined spending on defense, Social Security, and Medicare is not reduced, requiring deeper cuts in other programs. The second scenario reflects the fact that Mr. Trump has proposed to increase defense spending and that neither he nor congressional Republican budget plans have indicated interest in deep cuts over the coming decade in the two key programs assisting people who are elderly or have serious disabilities.
To pay for the tax cuts themselves, under the Trump plan:
- If all government programs were cut by the same percentage, offsetting the cost of the tax cuts in 2026 would require cutting all programs by 20 percent.
- If spending on defense, Social Security, and Medicare remained unchanged, offsetting the cost of the tax cuts in 2026 would require cutting all other programs by 53 percent.
To also accomplish Mr. Trump’s call to balance the budget, spending cuts would need both to eliminate currently projected deficits and to offset the tax cuts’ costs. If the budget were balanced in 2026, interest payments would amount to about 2.1 percent of GDP that year. With revenues of 14.1 percent of GDP under the Trump plan, that leaves 12.0 percent of GDP available for all federal programs and activities outside of interest payments. To achieve budget balance:
- Total federal spending (exclusive of interest payments) would have to fall, as a percent of GDP, to its lowest level since 1948. (See Figure 3.)
- If all government programs were cut by the same percentage, budget balance would require cutting all programs by 40 percent. Among other things, this means that about $2 of every $5 in basic Social Security benefits would be withdrawn.
- Spending on Social Security, Medicare, and defense is currently projected to equal 12.2 percent of GDP in 2026. This is slightly more than the revenue available for all federal spending under the Trump plan (12 percent of GDP). Thus if Social Security, Medicare, and defense spending were unchanged, all other programs would need to be eliminated entirely to achieve budget balance.
Other spending-reduction scenarios are possible, of course. But under any scenario, balancing the budget over the next decade while enacting the proposed tax cuts would entail hollowing out government in an unprecedented and potentially highly disruptive manner. The essential role that Social Security and Medicare play for tens of millions of Americans could be compromised. Spending on infrastructure, education, the judicial system and domestic security, scientific research, and basic food, housing, and health assistance to the poor could be eviscerated. Such changes could alter the nature of middle-class life and retirement security while also increasing poverty.
Spending Cuts Would Exceed House Budget Committee Plan’s Severe Cuts
The budget that the House Budget Committee adopted March 16 further illustrates the potential nature and magnitude of the spending cuts that would be needed under the Trump plan. The House Budget Committee’s plan would balance the budget over the next decade through $6 trillion of cuts to non-defense programs. The plan would be revenue neutral; that is, its net effect would be to neither reduce nor increase taxes. The plan’s basic structure — it includes no cuts in defense, essentially no cuts to Social Security, and moderate cuts to Medicare over the coming decade — mirrors that of other congressional Republican budgets of recent years.
Our analysis of the House budget found that 62 percent of the plan’s spending cuts would come from programs targeted for low- and moderate-income Americans; those programs would be cut by $3.7 trillion over ten years. By 2026, more than $2 of every $5 (42 percent) now projected to go to those programs would be withdrawn. If implemented, the budget would eliminate health coverage for more than 20 million people and drop millions of people into, or deeper into, poverty. In addition, due to the plan’s steep reductions in non-defense discretionary programs — which would be cut about $1 trillion below the sequestration levels over the coming decade — investments in scientific and medical research, education, infrastructure, law enforcement, environmental protection, and other national priorities could be cut sharply.
While the House GOP plan calls for $6 trillion in spending cuts over ten years, TPC estimates that the Trump plan cut taxes by $9.5 trillion over ten years. The calculations in this paper use slightly lower figures — $9.2 trillion for Mr. Trump’s tax cuts — that reflect the most recent Congressional Budget Office (CBO) economic projections. Thus, as Figure 4 shows, paying for the tax cuts in the Trump plan (without balancing the budget) would require spending cuts 50 percent deeper than what the House GOP budget plan would require. To also balance the budget under the Trump plan would require spending cuts of more than two and one-half times the size of those in the House GOP budget plan.
Proposals to Expand Certain Programs Would Increase Pressure on Others
Mr. Trump has also proposed some significant program expansions. While he has not said explicitly how much he would raise defense spending, he has proposed a fairly specific plan to expand assistance to veterans. The Committee for a Responsible Federal Budget (CRFB) estimates that plan would cost $500 billion over a decade. CRFB also found that the Trump changes to the health care system would cost $550 billion over the decade. These proposals would necessitate larger reductions in other programs than those described above.
If such expansions — or increases in areas such as infrastructure, which Mr. Trump also has mentioned — were enacted in addition to the tax-cut proposals, the squeeze on other programs would become still more intense.
How Else Could $3 Trillion Be Used?
One can compare the magnitude of the $3.2 trillion in Trump tax cuts to millionaires to alternative policies that would provide significant benefit to the broad U.S. population. This is not to suggest there is an extra $3 trillion in the government’s coffers that is waiting to be spent; it is to suggest that if policymakers wanted to devote this level of resources (which ultimately would need to be offset) to new policy initiatives, more beneficial alternatives would warrant consideration.
As Figure 5 shows, the annual revenue loss just from the tax cuts for millionaires under the Trump plan, which would amount to about 1.5 percent of GDP by 2025, would:
- Exceed the cost (about 1.3 percent of GDP) of ensuring Social Security and Medicare’s solvency over the next 75 years, under the 2015 estimates of the Social Security and Medicare trustees. Even under more pessimistic estimates about the size of the trust fund gap, the Trump tax cuts for millionaires would still exceed the cost of ensuring the solvency of Social Security (but not Medicare as well).
- Be significantly larger than the entire “poverty gap,” which is the total amount of income needed to raise every individual living in poverty to the poverty line. In 2014, the latest year for which this information is available, the official poverty gap equaled 1.1 percent of GDP. This should not suggest that poverty could be entirely eliminated with additional annual expenditures of 1.1 percent of GDP; efforts to close the poverty gap necessarily entail increasing assistance to those just above the poverty line as well. But this suggests that an anti-poverty effort of this magnitude could reduce poverty dramatically. For example, the Urban Institute estimated that a policy package that would reduce the number of children in poverty by 60 percent would cost about 0.5 percent of GDP, or one-third of the resources devoted to the tax cuts for millionaires under the Trump plan.
- Be more than seven times the total resources the federal government devotes to medical research (now 0.2 percent of GDP). The large majority of such research occurs through the National Institutes for Health. Substantially increasing the amount of resources for medical research could advance the nation’s capacity to respond to such ailments as heart disease, cancer, and Alzheimer’s disease and might improve health and extend lives for millions of people.
- Be larger than the amount of funding needed to repair the nation’s crumbling infrastructure. There is broad consensus that the nation’s roads, bridges, water systems, and other infrastructure require substantial repair and modernization. Such steps would make the nation healthier and safer, and would boost long-term economic growth. The American Society of Civil Engineers’ latest assessment of the nation’s infrastructure found that to raise its quality from a “D+” to a “B” would cost another $200 billion a year — or an additional 1.1 percent of GDP per year — between 2013 and 2020, above what is spent now.
Trump Plan Inconsistent with Appropriate National Priorities
Revenues at 1950 levels as a share of the economy, as the Trump plan would produce, are simply too low for today. Such a revenue base would necessitate some combination of exploding deficits and extraordinarily severe cuts to programs vital to low- and middle-income Americans. Mr. Trump should provide specifics on how he would address this conundrum. Describing the benefits of his tax cuts without explaining how to pay for them provides a highly incomplete, and hence skewed, picture of their true effects.
Further, the Trump tax-cut proposal would almost certainly shift an extremely large amount of income to the very top of the income spectrum and away from the rest of the population, exacerbating income inequality. The proposal implies that greatly increasing the after-tax incomes of the wealthiest Americans is a higher priority for the nation than various measures that would benefit Americans more broadly — such as repairing the country’s infrastructure, shrinking poverty, restoring long-term Social Security and Medicare solvency, or boosting medical research.
Appendix: Revenue Loss Estimates and Dynamic Scoring
This analysis relies almost exclusively on revenue and distribution estimates from the Tax Policy Center, as CBPP frequently does in its reports on federal tax policies. Citizens for Tax Justice and the Tax Foundation have also released estimates of the revenue losses that the Trump plan would generate.
TPC’s estimate of the Trump plan’s cost is lower than the Citizens for Tax Justice estimate that the plan would reduce revenues by $12 trillion over ten years, as well as the similar $12 trillion estimate that the Tax Foundation has produced. TPC estimates the revenue loss at $9.5 trillion, which we adjusted modestly downward to $9.2 trillion to take more recent GDP estimates from CBO into account.
TPC and Citizens for Tax Justice do not estimate the plan’s potential impact on economic growth. Considering such changes, and their effect on revenues, is now commonly referred to as “dynamic scoring.” The Tax Foundation, in contrast, produces both estimates that do not take macroeconomic effects into account (its $12 trillion estimate) and dynamic estimates that attempt to take such effects into account. Its dynamic score of the Trump plan — $10.1 trillion over the decade — still finds greater revenue losses than TPC estimates.
TPC’s analysis of the Trump plan includes a discussion of assessing the macroeconomic impact of the tax cuts. TPC notes that a wide range of studies find tax reforms in this country, as well as changes in the marginal tax rates that high-income people in a range of other countries face, have had little effect on economic growth:
- “Examination of particular historical examples of tax reform — including shifts between the pre– and post–World War II periods, and the tax changes that occurred in 1981, 1986, 2001, and 2003 — suggest little impact of taxes on growth,” TPC reported.
- TPC also noted that “Cross-country comparisons of changes in output and changes in top marginal tax rates suggest little or no impact of taxes on growth.”
Another reason to exercise caution in using dynamic scores of tax cuts is that there is no settled methodology for doing so. As a 2014 CBPP report observed, different modeling approaches based on different assumptions can produce widely varying economic growth effects. In some cases, one set of assumptions can yield estimates that economic growth will fall while another set of assumptions can yield estimates that growth will rise.
A third reason for caution is that modeling macroeconomic effects is particularly difficult to undertake when critical details concerning any policies to offset the cost of the tax cuts — which themselves could have economic effects — are lacking. Such is the case with the Trump plan.
In short, suggestions that tax cuts will substantially or completely “pay for themselves” by generating a far larger economy should be treated with very deep skepticism. They are a risky basis on which to make major fiscal policy decisions.
|Appendix TABLE 1|
|Distribution of Federal Tax Change Under Trump Tax Plan, 2025|
|Expanded cash income* percentile||Percent change in
|Share of total
federal tax change
|Average federal tax change (in 2016$)**|
|Top 1 percent||17.6||39.8||-329,103|
|Top 0.8% (millionaires)||17.9||38.0||-382,256|
|Top 0.1 percent||18.3||18.0||-1,438,664|
 Isaac Shapiro, “Millionaires Would Gain Trillions Under Trump and Cruz Tax Plans,” Center on Budget and Policy Priorities, April 28, 2016 https://www.cbpp.org/research/federal-tax/millionaires-would-gain-trillions-under-trump-and-cruz-tax-plans and Isaac Shapiro, “Trump and Cruz Tax-Cut Plans Would Shrink Government to Truman-Era Levels,” CBPP, March 29, 2016, https://www.cbpp.org/research/federal-tax/trump-and-cruz-tax-cut-plans-would-shrink-government-to-truman-era-levels. Both of these analyses examined only the Trump and Cruz plans because then-candidate John Kasich did not release enough information about his tax plan for TPC to assess it, and because the proposals of Democratic candidates Hillary Clinton and Bernie Sanders would increase taxes, not reduce them.
 The TPC estimates include the revenues raised by any provisions that increase taxes. As discussed later, this analysis adjusts the TPC revenue loss estimates slightly downward to adjust for new economic growth projections. For TPC’s analysis of the Trump plan, see http://taxpolicycenter.org/publications/url.cfm?ID=2000560.
 This analysis focuses on the effects of the plan in 2026, as recent congressional Republican plans to balance the budget attempt to do so at the end of ten years.
 As noted, all average tax-cut figures in this analysis are expressed in 2016 dollars.
 The TPC’s methodology establishes income percentile classes based on the number of people rather than the number of “tax filing units” or households. That is, each quintile class in the TPC data includes one-fifth of the U.S. population, not one-fifth of households.
 To calculate the revenue loss attributable to millionaires, this analysis first adjusts the total revenue loss estimates made by TPC slightly downward. TPC’s estimates of the ten-year revenue loss associated with the plan are based on economic projections made in 2015. We adjust those figures to take into account the most recent economic data. As a result, we estimate a ten-year revenue loss of $9.2 trillion for the Trump plan compared to the TPC estimate of $9.5 trillion.
Second, TPC only estimates the distribution of the tax cuts for two years, 2017 and 2025. To estimate the distribution of revenue loss over the ten-year period, this analysis averages the share of the tax cuts received by millionaires in those two years and multiplies the result by the ten-year revenue loss that TPC has estimated. TPC found that under the Trump plan, millionaires would receive 31.4 percent of the total tax cuts in 2017 and 38.0 percent in 2025, yielding an average of 34.7 percent.
Finally, the size of the tax cut going to millionaires is likely to differ slightly from what applying the average shares would indicate because of the effects of capital gains tax cuts. The Trump plan proposes lower rates for capital gains, which is likely to induce taxpayers to sell more appreciated assets than they would have under current law. The increase in revenue from the sale of these assets offsets some of the revenue loss from the lower rates. Since millionaires receive a significant majority of all capital gains, this offsetting effect disproportionately reduces the revenue loss attributable to the tax cuts for millionaires. This effect is not fully captured in an average-share calculation. Based on TPC data as well as CBO projections of capital gains taxes under current law, we estimate that the revenue loss for the tax cuts that would go to millionaires would be roughly 1 percent lower under the Trump proposal than the average-share estimates alone indicate.
Taken together, these steps result in the following calculation of revenue loss: $9.2 trillion*.347*.99 = $3.2 trillion.
 Even if all direct program transfers to millionaires were eliminated immediately, a rough CBPP estimate suggests this would offset less than 5 percent of the cost of the Trump tax cuts to millionaires over the next decade. While transfer programs make up a significant majority of federal spending, millionaires, like other households, also benefit from other programs — such as spending on defense, infrastructure, and the environment. In contrast to the tax cuts Trump has proposed, however, the benefits from these other programs are largely public goods and are broadly shared.
 Citizens for Tax Justice (CTJ) examined the net dollar effect of the Trump tax-cut proposal once offsets are considered. CTJ assumed half of the offsets would come in the form of spending reductions and half in the form of tax increases (thus effectively reversing at least some of the tax cuts). CTJ assumed the spending cuts would reduce each person’s income by the same dollar amount and that the tax increases would come in the form of an across-the-board increase in income taxes that would reduce each household’s income by the same percentage. It found that all income groups except for the top 5 percent would lose on balance under the Trump tax plan, once such offsets are assumed. Thus, even under the relatively progressive offset assumptions CTJ used (which include tax increases), low- and middle-income households would lose more from the offsetting measures than they would gain from the tax cuts themselves. Citizens for Tax Justice, “The Net Effect: Paying for GOP Tax Plans Would Wipe Out Income Gains for Most Americans,” March 9, 2016, http://ctj.org/ctjreports/2016/03/the_net_effect_paying_for_gop_tax_plans_would_wipe_out_income_gains_for_most_americans.php#.VuBUWNDYHOA.
 Assuming the offsets come in the form of spending reductions, households with annual incomes of less than $1 million but still in the top fifth of the income spectrum are, on balance, likely to gain significantly as well, though not nearly to the disproportionate degree as millionaires.
 The current baseline projection for spending on all non-defense programs outside Social Security and Medicare equals 7.8 percent of GDP in 2026, or $2.2 trillion.
 Richard Kogan and Isaac Shapiro, “House GOP Budget Gets 62 Percent of Budget Cuts From Low- and Moderate-Income Programs,” Center on Budget and Policy Priorities, March 28, 2016, https://www.cbpp.org/research/federal-budget/house-gop-budget-gets-62-percent-of-budget-cuts-from-low-and-moderate-income.
 This analysis uses GDP projections that CBO issued in March 2016 — which are slightly lower than the CBO projections from 2015 that TPC used when it issued its first analysis of Trump’s tax plan in December 2015. Revenue levels are very sensitive to the size of the economy, so a slightly lower level of GDP would be expected to yield a slightly smaller loss in revenues due to a tax cut.
 Committee for a Responsible Federal Budget, “Adding Up Donald Trump’s Campaign Proposals So Far,” May 9, 2016, http://fiscalfactcheck.crfb.org/how-do-donald-trumps-campaign-proposals-so-far-add-up/. Of further interest, CRFB released a synthesis of the then five presidential candidates’ proposals in which it assessed all specific policies released by that time and concluded that Trump has proposed to spend more, not less, on government overall. See “How Much Will Government Spend Under the Next President,” March 24, 2016, http://fiscalfactcheck.crfb.org/how-much-will-government-spend-under-the-next-president/.
 Haley Sweetland Edwards, “Trump Agrees With Democrats on High-Speed Trains,” Time, March 3, 2016, http://time.com/4247162/donald-trump-trains-infrastructure/.
 The trustees’ estimates of Social Security and Medicare are the most widely used sources of information on the long-term status of the Social Security and Medicare Hospital Insurance (Part A) trust funds. CBO also makes long-term estimates of Social Security and Medicare, using assumptions that produce a larger solvency gap. Under CBO assumptions, while the revenue loss from the Trump tax cuts for millionaires would exceed the cost of ensuring Social Security solvency over the next 75 years, it would exceed most, but not all, of the cost of ensuring the solvency of both Social Security and Medicare.
 Children’s Defense Fund, Ending Child Poverty Now, 2015.
 Emily Feenstra, “Our Nation’s Neglected Infrastructure,” Contingencies, September/October 2015, http://www.contingenciesonline.com/contingenciesonline/september_october_2015?pg=26#pg26.
 James R. Nunns, et al., “Analysis of Donald Trump’s Tax Plan,” Tax Policy Center, December 22, 2015, p. 17, http://www.taxpolicycenter.org/publications/analysis-donald-trumps-tax-plan.
 Paul N. Van de Water and Chye-Ching Huang, “Budget and Tax Plans Should Not Rely on ‘Dynamic Scoring,’” Center on Budget and Policy Priorities, November 17, 2014, https://www.cbpp.org/research/budget-and-tax-plans-should-not-rely-on-dynamic-scoring?fa=view&id=3598.