Some criticize tax expenditures by noting that high-income earners get a much larger share of this type of spending than those with lower incomes. These policies therefore offset somewhat the progressive tax structure in the United States, under which higher-income Americans pay a larger proportion of their income in taxes that those who make less. If Americans knew that tax expenditures mainly go to the well-off, would they be less likely to support them?
No. In a recent article in Political Behavior, we find that citizens favor expanding benefits via the tax code — even when it’s clear that higher-income households will get more of those benefits.
The home mortgage interest deduction as a case study
The tax expenditure most familiar to many Americans is the home mortgage interest deduction. Under this long-standing policy, families can deduct the interest they pay on their mortgages from their federal income taxes. In 2014, this deduction cost the federal government $72 billion, with $66 billion going to households earning more than $75,000; $30 billion of that went to those earning over $200,000. We conducted an experiment to see whether, when Americans understood that high-income households get most of that benefit, they still supported the policy.
We recruited an online sample of approximately 3,300 respondents in June 2018 with the help of Lucid, a survey sampling firm; respondents were nationally representative with respect to age, education, household income, gender and minority identification. We randomly assigned respondents to see one of two ways of explaining the home mortgage interest deduction: one that was regressive and another that was progressive. In the regressive explanation, we told respondents that eligible households earning between $75,000 and $200,000 get benefits from the deduction that are about twice as large (in dollars) as those for households earning less than $75,000. In the progressive explanation, we told respondents that the benefit for households earning less than $75,000 are about twice as large (as a percentage of their incomes) as the benefit for the wealthier households.
How we explained the facts didn’t matter, to our surprise. After seeing an explanation of the policy, respondents were asked whether they wanted to increase or decrease home mortgage interest deduction benefits on a five-point scale. No matter which explanation they read, respondents were, on average, equally in favor of increasing the benefits. In fact, respondents shown the regressive frame — in which well-off households got twice as much money as those less well-off — still favored expanding the deduction, despite believing that the policy increased economic inequality.
However, we did find that respondents shown the regressive framing were more supportive of changing the home mortgage interest deduction to provide greater benefits to lower-income households, relative to respondents shown the progressive framing, though this finding had greater statistical uncertainty than we would usually accept in order to draw a strong conclusion.
Americans believe tax expenditures cost the government less money than direct alternatives
Why do Americans support tax breaks even when they benefit the wealthy? We attempted to answer this question in a second set of experiments, where we proposed two hypothetical policies: nutritional assistance to all newborns and their mothers and helping employed workers get additional training. We made sure that these policies were as universal as possible and didn’t target specific groups — like the unemployed — that might be perceived to be more or less deserving of the benefits.
For each of these policies we offered different ways to deliver the benefits. The nutrition policy would be delivered to households either as a monthly $100 debit card or as an equivalent tax credit. Similarly, the job training policy would be delivered to workers as a $7,000 direct deposit into their checking accounts or as a tax credit for the same amount.
Using the same survey-participants as before, we found that respondents were much more likely to support both policies, on average, when the tax code delivered the benefits than when beneficiaries received cash assistance or its equivalent.
Here’s why: When considering both policies, respondents told us they believed a tax expenditure would cost the government less money than the cash transfer.
People believed their own taxes would go up more if the job training policy deposited money in beneficiaries’ accounts than if beneficiaries took it as a tax credit. And for the nutrition assistance, they believed that beneficiaries would have to work harder to take the tax credit than they would to use a debit card.
Republicans have a stronger preference for tax expenditures than do Democrats
We find that Republicans have a much stronger preference for delivering both policies through the tax code, when compared to Democrats. Democrats are actually indifferent, on average, between the direct deposit and tax credit versions of the job training policy.
Republicans express a strong preference for the job training tax credit, believing that citizens are less likely to misuse it and less likely to grow dependent on government assistance than they would with the direct deposit.
Lessons for policymakers
With the 2020 elections coming, politicians may try to mobilize voters against tax expenditures that benefit particular citizens more than others. Our research suggests that might not succeed.
The key to changing tax expenditures might be to make changes around the edges instead of launching overhauls — as Republicans appear to have done in their recent tweaks to the home mortgage interest deduction, which lowered the limit on the value of mortgages that qualify for the benefit.
Vivekinan L. Ashok is a Presidential Postdoctoral Fellow in the Department of Government at Cornell University and a faculty affiliate of the Roper Center for Public Opinion Research.
Gregory A. Huber is the Forst Family Professor of Political Science at Yale University and a resident fellow of the Institution for Social and Policy Studies.