A Puzzle in Voter Behavior

Link post

Everyone knows old republicans are against immigration. This demographic fact has persisted across decades so it may seem normal and obvious, but it actually presents a challenge to many economic theories of voter behavior. In particular, the anti-immigration preference among old republican voters isn’t well explained by theories of “rational” or self-interested voting.

These theories suggest that people vote for the candidate whose policies would benefit them the most. This is intuitive and has some empirical support. Democrats, who are highly educated, favor student loan forgiveness and Republicans, who are older, favor expanding social security. But there are just as many counterexamples to this theory. High-income Democrats vote for increased taxes and farm-and-fracking Republicans vote for tariffs which reduce the value of their export industries.

Anti-immigration sentiment among old republicans is another counterexample to the predictions of straightforward self-interest in the voting booth. This group of voters face less of the purported negative impacts of immigrants than almost any other group. They don’t compete with immigrants in the labor market since they’re retired or close to it. They’re insulated from any increases in crime which happen mostly in dense city centers (if at all).

Old republicans also capture lots of the upside from extra immigration. Young, working age immigrants buoy social security and Medicaid. Homeowning older voters benefit from upward pressures on land prices due to migration and population growth. Low-skilled immigrants also ease staffing pressures at nursing homes, significantly improving outcomes for residents. This is especially beneficial to older Republicans who are more likely to smoke, be obese, and rely on social insurance.

Most of these voters would personally benefit from greater immigration, so why do they vote against it? It isn’t just ignorance and irrationality. It’s rational ignorance and irrationality.

Consider this model of an election: We get to choose between candidate D and candidate R. The straightforward self interest model simply says that we choose whichever candidate benefits us the most. If we’re older, we might choose candidate D since expanded immigration would increase the quality and decrease the cost of our medical care later in life, perhaps worth tens of thousands of dollars in health and wealth.

The rational irrationality model notices that we only get these benefits if candidate D actually wins. If candidate R is thousands of votes ahead, then no matter who we choose to vote for, we’ll get R’s policies. Similarly, if candidate D was going to win with our without our vote then no matter what, we get D’s policies. So unless our single vote decides the election, the policy differences between D and R have no direct impact on the self-interested welfare impact of our voting decision.

Instead of considering the difference between the value of each candidate’s policy proposals, we consider the difference between each candidate’s policy proposals multiplied by our chance of changing the outcome of the election. We’re one vote out of tens of millions so this chance is miniscule. Even when one candidate’s policies would benefit us on the order of hundreds of thousands of dollars, the expected value of a vote for that candidate is tiny since it will almost certainly not change the outcome.

So why do people vote at all? Well, lots of them don’t, suggesting that the small costs of voting are often enough to outweigh the expected value of a vote for one candidate over another. But what voter turnout there is can be explained by noticing that there are rewards to voting which don’t rely on affecting the outcome of the election: The feeling of fulfilling civic duty, indulging in nationalism or xenophobia, cheering for your team colors, or seeing yourself as helping those in need.

These emotional benefits are modest, and are outweighed when the stakes are high. Democrat billionaires still use tax advantaged savings accounts and write-offs because doing otherwise would cost them too much. But they don’t face these costs at the voting booth. Even though a higher tax rate might cost them dearly, their single vote has almost no chance of causing or preventing this policy change. So they are free to indulge in the emotionally rewarding rhetoric of providing for the needy.

This theory of voter turnout also makes predictions for voter behavior. First, rational ignorance: Voters won’t do much research on the similarities and differences between candidate’s platforms since any gains they might get from a clear-eyed evaluation of the candidates are trivialized by their tiny chance of changing the winner. Second, rational irrationality: Voters won’t work to correct biases that muddy their decision making since picking the right candidate doesn’t matter much and indulging in biases can be emotionally rewarding.

So why do old republicans support cutting immigration even though that might land them in an understaffed nursing home for the rest of their lives? It’s a collective action problem. As a group of millions, their votes do have massive influence on immigration policy. But as an individual voter in the booth, it makes sense to consider everyone else’s vote as already cast. In that case, the election is almost certainly decided one way, with or without your vote. So it’s more rewarding to defect and indulge in an emotionally rewarding cheer for your team, your colors, and your country.