Oct 29, 2024, 11:01 AM ETTutorial

How much does a vote matter? | Rob Wiblin

The expected value of voting often exceeds the impact of any other activity achievable within the same timeframe, provided the election is competitive and the voter is informed.
The analysis focuses on US presidential elections due to their global priority, large audience reach, and extensive modeling data, though the framework applies to other democracies.
In 2000, a margin of 537 votes in Florida (0.009%) and 366 votes in New Mexico determined the outcome, illustrating that close margins are historically plausible.
Current modeling by Andrew Gelman indicates that as of October 14, 2020, only four states (New Hampshire, Pennsylvania, Wisconsin, Michigan) had odds of a single vote changing the outcome exceeding 1 in 10 million.
The probability of a vote changing an election result is roughly proportional to one over the square root of the number of voters.
In countries with fewer voters like Australia (1/10th US) and the UK (1/5th US), the odds of a single vote flipping a close election are two to three times higher than in the US.
Even in seemingly non-competitive elections, polling errors prevent the probability of a swing vote from dropping below half of its value in a very close contest.
The "court rerun" objection is invalid because a single vote shifts the probability of winning a clean victory, which remains as impactful as shifting the margin in a tied election.

The US federal government allocates approximately $17.6 trillion over a four-year term, averaging $129 per vote cast in 2016.
Multiplying the $17.6 trillion budget by a 1 in 10 million chance of influence yields an expected value of $1.76 million per vote in a competitive district.
If a policy shift involves only 0.5% of GDP (e.g., foreign aid) versus 0.3%, a single vote with 1 in 10 million leverage shifts $17,000 in expectation.
Non-budgetary impacts include foreign policy (trade, war), business cycle stabilization, regulations, immigration policy, social freedoms, and the protection of political liberties.
In a hypothetical nuclear war scenario with a 1 in 1,000 chance under one candidate versus 1 in 500 under another, a single vote would be worth $25,000 per fellow citizen to that outcome.
George W. Bush's administration is cited as an example where divergent policies led to the Iraq War (hundreds of thousands of deaths) while simultaneously funding the PEPFAR program (potentially saving millions of lives via HIV drugs).
Election impact is highest when candidates differ significantly, a condition more likely in systems with low turnout, weak checks on power, plurality voting, and multiple viable candidates.

The argument for "epistemic modesty" suggests that because half the population disagrees with you, your probability of being right is only 50%, which would reduce the expected value of voting to zero.
This objection relies on the assumption that all voters are epistemic peers, which is false given the average voter's low level of information and motivation.
Surveys indicate high levels of political ignorance, such as only 38% of Americans knowing which party controlled the House in 2014 or 20% realizing Social Security spending exceeds foreign aid.
Uninformed voters often rely on heuristics (e.g., "am I better off?") that can cancel out random errors but still provide less reliable data than deliberate research.
Individuals who impartially consider global well-being and research policy positions possess a significant epistemic advantage over the average voter, potentially increasing confidence in their choice to 75% or higher.
Adopting a collective strategy of not voting due to uncertainty would likely worsen overall political outcomes by reinforcing public ignorance.

The direct time cost of voting is low (minutes to hours), but the cost of informed decision-making can be substantial if the voter must research from scratch.
For a highly specialized individual, spending a week researching a candidate could cost $100,000 in forgone earnings, making direct donation to effective charities potentially more impactful.
If a voter already follows politics and can vote wisely with minimal extra effort, voting is generally the most impactful use of their time.
The effort required to research is most justifiable in elections where the difference between candidates is stark and the election is competitive.
Voters may defer decision-making to trusted individuals if they lack the time or expertise to research independently.

Personal appeals from friends and family increase voter turnout by 10 percentage points or more, making them 10 times more effective than stranger appeals.
The historical cost to persuade one voter to support a candidate via door knocking or phone banking is estimated at $30 to $100.
2020 data suggests the cost per vote for TV ads in US presidential elections has risen to $100 to $1,000 due to high campaign saturation.
In 2020, the Trump and Biden campaigns allocated approximately $30 per potential voter in tipping-point states, compared to only $3 per voter in the Democratic primaries.
Campaign spending per voter is significantly lower in the UK and Australia (approx. $10 per voter in marginal seats) than in the US.
Research suggests social science efficacy findings may weaken over time due to replication failures and increased campaign saturation.
Donating to underfunded campaigns or volunteering to contact friends in competitive districts can offer a high return on investment if the social value of a vote exceeds $10,000.

Voting is altruistic and high-impact if: (1) the election concerns major issues, (2) one candidate is substantially better and identifiable, and (3) the voter is in a competitive seat.
The expected impact of a single vote in a competitive district is sufficient to justify the time investment for most informed individuals.
Friends and family should be encouraged to vote, as personal appeals are the most cost-effective method for generating additional votes.
Individuals in high-impact careers may find it more effective to specialize in solving global problems and donating earnings rather than dedicating significant time to political research.
Joining or donating to political campaigns is a viable high-impact strategy, particularly when campaigns are underfunded and the candidate gap is large.
The author explicitly states they vote whenever possible, viewing the risk of error as outweighed by the massive potential upside of influencing government allocation.