Americans' favorite tax provisions -- for homes, for charities, for children -- are meant to defend a "middle class welfare state." They're not doing a very good job.
As the presidential election enters its final days, the battle over taxes remains front and center. Both Mitt Romney's plan and prominent deficit reduction proposals rely on "broadening the base" by eliminating tax expenditures, like deductions and credits, that act as spending through the tax code.
The Obama campaign is claiming that Romney will hurt the middle class by cutting popular tax expenditures like the home mortgage interest deduction, while Romney has avoided the politically-delicate question of what he will cut to make his fiscal math add up. Christina Romer, former chair of the Council of Economic Advisors, echoed Obama's argument in the New York Times last month, writing, "Many tax expenditures, like the mortgage interest deduction and the tuition credit, go to middle-class families."
While rhetoric about tax expenditures is bandied about, the actual effects get lost in the fray. Do tax expenditures actually benefit the middle class? Or, put differently, are tax expenditures a "middle-class welfare state", as the Washington Post's Ezra Klein described them last year?
The answer is: not as much as you might think. While there are differences between types of tax expenditures, most of the largest tax expenditures help a few people in the middle class a bit while helping wealthier Americans much more. The tax code may have become a type of middle-class welfare state, but in its current form is one of the most regressive and least efficient middle-class welfare states possible.
However, with median income on the decline, certain tax expenditures do play a role for some members of the middle class even as the highest earners gain an outsized share of the benefits and the system is highly regressive overall.
Think of tax expenditures like serving a very fancy and expensive dinner to Joe (who is very wealthy) and John (who is middle class). We fill Joe's plate with food, including heaping servings of mashed potatoes, ribs, broccoli, and dessert. On John's plate, we put one small half-dollop of mashed potatoes and a few peas. John does benefit from the dinner overall, but barely, while Joe gets massive benefits.
And this fancy dinner costs a lot of money.
In 2011, tax expenditures cost about 1.1 trillion dollars. The Office of Management and Budget projects a similar figure for the current fiscal year (FY 2013), with costs increasing gradually in the upcoming years. The costs mainly include deductions (like the home mortgage interest deduction), exclusions (like the employer-sponsored insurance exclusion), credits (like the earned income tax credit), and lower rates for investment income.
Here is how much middle-income earners have benefited from some of the largest tax expenditures (and estimates of how much they will cost in FY 2013):
Home Mortgage Interest Deduction (100.9 billion dollars in FY 2013): Income filers making 100,000 dollars per year or more got nearly 75% of the benefit from the home mortgage interest deduction in total dollars in 2011. Therefore, filers making less than 100,000 dollars per year, or a vast majority of Americans, got 25% of the benefit.
If we define the middle class broadly as the three middle quintiles (20-80% of earners), averages of Tax Policy Center estimates show that about 76 percent of middle class individuals would not see any increase in their taxes in 2012 if the mortgage interest deduction were eliminated. The average amount that a person in this quintile benefits from the deduction is 312 dollars per year. In other words, for a program that costs on average 559 dollars per person, a person in the middle group would benefit an average of 312 dollars.
Looking at total dollar amounts shows how benefits are distributed overall and is not affected by huge income disparities. To determine how these tax expenditures affect families' budgets, however, we can also look at benefits as a percentage of income. If we use 100,000 dollars per year in cash income as a cutoff for the middle class (which corresponds approximately to the cutoff for the 80th percentile of earners), we can see from the chart below that tax filers in the higher end of that spectrum - and especially right above it - benefited in the realm of 1% of their income on average from the home mortgage interest deduction in 2011.
Despite being regressive, then, the home mortgage interest deduction helps the middle class more than other tax expenditures because it helps a portion of upper-middle class families and because other large expenditures are simply more regressive.
Charitable Contributions Deduction (48.9 billion dollars in FY 2013) Charitable contributions also skew upward: 85.2% of the gains went to those filers making over 100,000 dollars per year in cash income in 2011. As a percentage of income, extremely regressive becomes just regressive: earners under the 100,000 threshold benefited on average between 0% and 0.3% of their income from the deduction, while the highest earners gained 1.2%.
Part of the reason for this is that charitable contributions, like the home mortgage interest deduction are line-item deductions. Therefore, they only apply to those who itemize their tax returns. Since each filer has to choose whether to itemize or take a standard deduction, itemized tax expenditures only benefit those filers whose itemized deductions are big enough to exceed the standard deduction and who have the know-how to itemize. In 2011, less than 30% of all filers itemized their taxes, and more than 80% of the benefits from itemized deductions went to individuals in the highest quintile.
Preferential Rates on Capital Gains (62 billion dollars in FY 2013). The lowered rates on capital gains are the most regressive tax expenditure: in 2011, 96% of the benefits went to the highest quintile of earners. Tax filers with income of more than a million dollars per year or more gained on average nearly 120,000 dollars from these reduced rates (over 6% of income), and only 13% of all filers gained any benefit from the lowered capital gains rates.
In other words, the direct middle class benefits were essentially zero from lowered capital gains rates.
Employer-Sponsored Health Insurance Exclusion (180.6 billion dollars in FY 2013). Determining distributional effects of the largest tax expenditure is difficult because it is deeply interconnected with larger issues of health care provision. This is also why few policymakers have suggested significantly altering the policy unless it is done as part of a larger health care overhaul toward a single-payer system.
Middle-class families do benefit from the exclusion employer health insurance benefits since a majority of all workers receive health insurance through their employers. However, just as with other major tax expenditures, the distribution favors the wealthy. Middle-class families benefit less than the highest earners do, and as employer-sponsored health insurance coverage continues to decline, the middle class will see fewer benefits from this provision.
In a 2009 analysis, Jonathan Gruber estimated the cost and coverage effects of changes to the employer-sponsored insurance exclusion. Leaving the exclusion for payroll taxes aside, the top 20% of earners would bear 50% of the cost of repealing the exclusion for federal income taxes (which means that these earners receive approximately 50% of the benefit of the exclusion). The three middle quintiles combined receive the other half.
Exclusion for Retirement Savings Plans (165.4 billion dollars in FY 2013). Tax exclusions for retirement plans, including 401(k)s and other employer pensions, sound like they should help the middle class, but the distributional effects are similar to other major tax expenditures. Approximately 80% of the benefits from all tax-favored retirement plans will go to the top quintile of earners this year. As a percentage of their income, individuals in the three middle quintiles will benefit about 0.9% on average, while top quintile earners will benefit 3.2%.
With the continued decline in employers offering defined benefit pensions, defined contribution pension values damaged by the stock market, and questions about the long-term solvency of Social Security possibly leading to significant benefit cuts, the tax exclusion for retirement savings plans will not make up for a decrease in middle class retirement benefits.
Taken together, the Tax Policy Center estimates that 2/3 of the benefit from all exclusions (including health insurance, retirement savings, and a number of others) accrued to the highest quintile of earners in 2011.
Yet despite this imbalance, exclusions are far more important to the middle class as a percentage of income than itemized deductions and preferential rates on capital gains. If all exclusions added together were eliminated, the middle three quintiles would see their after-tax income decline by about 4.4% (compared to 0.6% for itemized deductions and less than 0.1% for capital gains).
The differences between total benefits from tax expenditures and benefits as a percentage of income means that eliminating these regressive expenditures without adding further supports will take a toll on middle class budgets. As income inequality and the cost of goods like health care increases, these tax expenditures become more important to the middle class, even if the distribution favors the highest earners.
Earned Income Tax Credit (55.7 billion dollars in FY2013, including money refunded). Any full discussion of tax expenditures also needs to differentiate between deductions/exclusions and credits. Tax deductions (and exclusions) reduce the amount of income that is subject to taxation, while tax credits rebate or offset the direct amount of tax being paid.
About 80 percent of tax expenditures are deductions or exclusions, rather than credits, which is why most of the biggest expenditures fall into this category. But the distributional effects of deductions and credits are very different. Credits - particularly refundable ones, meaning that if the tax burden is reduced to a negative amount the filer receives a positive rebate - are progressive. More than 90% of the benefits of all refundable credits, including the EITC, went to the bottom four quintiles of earners in 2011, and the elimination of all refundable credits would hurt the middle quintiles of earners about 3% of income on average.
This does not mean that credits are exempt from the debate: there is a strong argument to be made that providing benefits through the tax code is less effective and less efficient than crafting more direct social policy. In America, we help offset the cost of raising a child with the child tax credit; in Finland, however, families are given a direct child benefit payment each month. And the use of tax policy (particularly credits) to provide social policy reduces many Americans' federal income tax burdens to zero or less, which has become a point of contention in the public sphere.
Rather than defending current tax expenditures or focusing on reducing tax rates, the debate we should be having is how to create a "middle class welfare state" that actually works. We can continue to use the tax code but make it radically more progressive by replacing deductions with credits or creating mandatory universal tax-favored savings programs, or we can switch from the system run through the tax code to a simpler, more direct slate of policies, such as expanding social insurance programs like Social Security and Medicare and increasing public provision of services like education and childcare.
Both sides agree that the American middle class is "under stress" and facing stagnating incomes, higher costs, and more insecurity. If we are in an economic situation in which any action that would hurt the middle class is wrong, even if it is extremely regressive and inequitable, it should make us think that the overall welfare state for the middle class needs to be stronger. If without the half-dollop of mashed potatoes the middle class is going to starve, we should rethink how we are serving dinner.