How Much Revenue Could You Raise With a Poll Tax?

And other reflections on charging people to vote.

Two key dogmas in our secular religion are:

1. Nothing is more important for a citizen than the right to vote.

2. Charging anyone to vote — the so-called “poll tax” — is depraved.

Yet if you remember the adage that “Actions speak louder than words,” there is a deep tension between these two dogmas.

The first dogma, even if exaggerated, implies that people will have a high willingness to pay to vote. After all, if voting is so tremendously important, you will gladly sacrifice piles of other goods to partake.

The second dogma, in contrast, relies heavily on the assumption that a poll tax would sharply reduce turnout, leading to dramatically different — and far less fair — political outcomes.

Most of my research in public choice ultimately relies on the fact that an individual voter’s influence is near-zero. Most obviously, the probability that your vote ever flips any notable political race is rock-bottom. As a result, the expected selfish cost of holding even the most absurd political views is almost nothing, leading to severe, widespread political irrationality.

One common objection to my position is that many, perhaps most, voters don’t know that their probability of decisiveness is approximately zero. My favorite reply is just pointing out that the idea of even a $100 poll tax horrifies most people. Why? Because they instantly realize that turnout would crash. Which doesn’t just show that almost everyone does know that p≈0, but that almost everyone knows that almost everyone knows that p≈0.

The simplest way to save the standard condemnation of the poll tax is that the first dogma uses the word “important” in a purely moral sense. “Nothing is more important for a citizen than the right to vote” doesn’t mean that individual citizens strongly value their own right to vote. It just means that zero-price voting is morally critical for society. On this interpretation, the less individuals selfishly care about their right to vote, the more depraved it is to charge them to vote.

But philosopher Jason Brennan has already thoroughly critiqued this folk morality of voting. Despite its immense popularity, the idea that political participation of ignorant and irrational voters is good is quite crazy.

If someone is ignorant of surgery, the morally appropriate recommendation is not “It doesn’t matter where you cut, just cut.” The morally appropriate recommendation is abstention. Admit you’re unqualified, and exit the operating room.

Furthermore, if someone ignorant of surgery tries to join a surgical team despite his ignorance, the morally appropriate reaction is not, “Participation is his sovereign right.” The morally appropriate reaction is to stop him. Admit they’re unqualified, and shoo them away from the voting booth.

Which demands a full rethinking of the morality of charging people to vote. Sure, a poll tax could be used in a discriminatory manner. But so could licensing people to drive. Historical baggage aside, a traditional poll tax has two obvious effects: reducing turnout of not just the poor, but the apathetic.

There can be little doubt that the politically apathetic are especially politically uniformed; after all, the less you care about anything, the less you know about it. And there is overwhelming evidence that the poor, too, are especially politically uninformed. (In multiple regressions, this turns out to be largely because the well-educated are both richer and more informed. Regardless of the mechanism, however, the conclusion stands: The poll tax causally raises the knowledge of the median voter).

Since objective self-interest has little effect on political preferences, moreover, there’s basically no cognition-motivation trade-off. Rich voters, like virtually all voters, want to advance the best interests of society; they just know more about the best ways to advance those best interests.

Which brings me to a final verboten question: Sticking with a standard flat rate, how much revenue could you raise with a poll tax anyway? Where does the Laffer curve for the poll tax reach its peak?

The Laffer Curve

This paper finds that even token poll taxes reduced historic turnout by 6 percentage-points. Here’s my best guess about demand elasticity for modern U.S. presidential elections, starting from the 2020 benchmark of 66% turnout (roughly 157 million voters) with a $0 tax.

Poll Tax     Expected Turnout     Total Votes     Revenue Raised

$0               66%                            157M              $0

$10             60%                            143M              $1.4B

$50             54%                            128M              $6.4B

$100           48%                            114M              $11.4B

$200           42%                            100M              $20B

$400           36%                            86M                $34.4B

$600           30%                            71M                $42.8B

$800           24%                            57M                $45.7B

$1000         18%                            43M                $42.8B

$1200         12%                            29M                $34.8B

All guesswork, I admit; if you’ve got superior quantitative insight, please share in the comments. But the conclusion that poll tax revenue would peak in the $600-$1000 range is plausible. How high would the poll tax have to be before participation fell to 1%? Maybe $5000?

Compared to the roughly $4.4T in taxes the federal government collected in 2023, a revenue-maximizing poll tax is a rounding error. But from a Pigovian perspective, a poll tax is still a fine idea. While I’m well-aware this won’t happen anytime soon, our quadrennial festival of democratic fundamentalism is in dire need of some blasphemy.