What makes people think the tax system is fair? In a fascinating new study, my Tax Policy Center colleague Vanessa Williamson found that people’s attitudes about tax fairness changes when they get new information about how much other people pay. This is especially true when they learn about the taxes low-income people pay. 

In February, Vanessa ran an experiment in cooperation with the news website Vox. Readers were given a chance to participate in a voluntary tax quiz, and more than 7,000 did. They were divided into five groups. One was asked questions that encouraged them to think that high-income people pay relatively low taxes, another was asked questions about the high taxes the rich pay. The third was asked questions about parts of the law that reduce taxes for low-income people, and the fourth was nudged into thinking poor people pay relatively high taxes. 

Then, after taking the quiz, the participants in these four groups were asked their opinions about the tax system. Did they think high-income people paid too much, too little, or their fair share? What about low-income people?  What is the highest rate anyone should pay? Should the government redistribute income by raising taxes on the rich? And, finally, did they think their own taxes were fair?

Finally, people in the fifth group (the control group) were asked their opinion of the tax code before taking the quiz. Vanessa compared their answers with those from the people who gave their opinions on tax fairness after taking the quiz.   

What happened?

First, Vanessa found that facts matter. When people learned about certain aspects of the tax code, their views on fairness changed, sometimes dramatically. This was especially true when they were asked to think about taxes paid by the poor.

When people were asked before taking the quiz (the control group) about taxes paid by the poor, 56 percent thought that low-income people paid too much. But among those who were first asked questions that suggested taxes fall heavily on people with low incomes, 68 percent thought poor people paid too much. And among those who were asked questions that suggested tax breaks benefit the poor, only 47 percent thought they pay too much.

When it came to their views about the rich, Vanessa found much more modest effects. Those who answered questions suggesting that the rich take advantage of special tax breaks and overseas shelters were a bit more likely to say that the rich don’t pay their fair share, but their views moved by only 4 percentage points, much less than among those who were asked similar questions about low-income people.  People barely changed their views at all when asked questions that suggested the rich pay high taxes.

Vanessa also found an interesting see-saw effect. If people thought that taxpayers at one end of the income spectrum paid too much, they also believed that those at the other end paid too little. For example, if they learned that the poor were paying relatively high taxes, they were more likely to think the rich were not paying enough.

How did taking the quiz affect the respondents’ views about their own taxes? It turns out that the control group (those who were asked their views before learning new information about the tax code) were most likely to think their own taxes were fair. But once people were asked about taxes paid by the rich and the poor, they were more likely to think their own were less fair. And for the most part it didn’t matter much which quiz they took or what their responses were.

A few caveats. This was not a representative sample. Overall, Vox readers are more politically engaged, more liberal, younger, and wealthier than the public as a whole. And interestingly, while Vox readers overall are divided evenly among men and women, two-thirds of those who chose to participate in the quiz were men.  As a result, Vanessa couldn’t say that all Americans would respond this way, but she could measure how getting information affects some people’s opinions about taxes.

Still, the results are intriguing. They raise a number of important questions for further academic study. And they may suggest to politicians that people’s views about the tax code can be manipulated by presenting them with “facts,” whether real or alternative.