Rational Choice Theory, Emotional Decision-Making, Identity Politics, and the Enduring Mystery of the Persuadable Voter
Rational Choice Theory, Emotional Decision-Making, Identity Politics, and the Enduring Mystery of the Persuadable Voter
The Theory: Voters Are Rational Actors Making Informed Choices
The foundational model of democratic theory holds that voters are rational actors who evaluate candidates’ policy positions, compare them to their own preferences, and vote for the candidate whose positions most closely align with their interests. This is a beautiful model. It implies that campaigns should focus on policy substance, that voters should respond to evidence about performance, and that elections should produce governments that reflect the considered preferences of the majority. Political scientists have spent approximately 70 years testing this model against empirical data. The data have not been uniformly cooperative.
The empirical reality of voter behavior is messier, more interesting, and considerably more alarming to democratic theorists than the rational choice model suggests. Most voters have limited knowledge of candidates’ specific policy positions. Most voters cannot accurately describe the policy differences between the parties on more than two or three issues. Most voters make their voting decisions not through careful policy evaluation but through a combination of partisan identity, candidate affect – whether they like the person – economic perceptions – whether they feel financially better or worse off than they did four years ago – and cultural and social identity – whether the candidate seems like someone from their tribe. This is not an insult to voters. It is a description of how humans make decisions under conditions of information overload, limited time, and competing demands on attention. Voters are not irrational. They are using cognitive shortcuts that are reasonable given the constraints, and the shortcuts happen to be largely unrelated to the policy substance that democratic theory assumes drives the process.
Party Identity: The Shortcut That Shortcuts Everything
The most powerful predictor of how an American will vote in any given election is how they have voted in previous elections, which is itself strongly predicted by how their parents voted, which is strongly predicted by the social and cultural community in which they were raised. Party identification is not primarily a policy choice. It is a social identity, as stable as religious affiliation and roughly as amenable to rational persuasion. People who identify as Democrats vote Democratic in approximately 90% of elections. People who identify as Republicans vote Republican in approximately 90% of elections. The campaigns, the advertising, the debates, the scandals, the economic data, the foreign policy crises – all of this activity is targeted at the 10% who might be persuaded, the so-called “”swing voters,”” while the 90% who have already decided are motivated to turn out through messages that confirm their existing beliefs and amplify their existing fears about the other side.
This means that the presidential election, the most expensive and elaborate democratic exercise in the world, is functionally a turnout contest among people who have already decided, with a side competition for the attention of a relatively small group of genuinely persuadable voters in a small number of states. The elaborate machinery of modern campaigns – the billions of dollars, the data analytics, the message testing – is aimed at this relatively small target. Whether the result is worth the expenditure is a question campaigns do not ask because the expenditure is the business model, and questioning the business model is not something the business model funds.
The Economy: The Variable That Overrides Everything
Political scientists have found that the single best predictor of presidential election outcomes – better than candidate quality, better than campaign spending, better than debate performance – is the state of the economy in the months before the election, specifically the direction of economic movement rather than its absolute level. Voters do not evaluate the economy in absolute terms. They evaluate it in relative terms: is it getting better or worse? An economy that is improving from a terrible baseline outperforms a stagnant economy that started from a good one. Voters reward economic improvement and punish economic deterioration, with a lag of approximately six months during which economic signals are translated into political judgment.
This finding, known as “”economic voting,”” is robust across decades of data. It suggests that presidents have substantial incentive to time economic policy for electoral cycles – stimulate the economy before elections, accept contraction after them – which is exactly what they try to do, with mixed success, since the economy is not reliably responsive to political scheduling. It also suggests that presidents are rewarded and punished for economic conditions substantially outside their control – global recessions, pandemic disruptions, commodity price shocks – which raises interesting questions about whether voters are actually evaluating presidential performance or merely correlating incumbency with economic weather. The answer appears to be mostly the latter, which makes presidential electoral accountability a somewhat imprecise instrument for democratic control, though it is the instrument available and the one the system has been using for 235 years with characteristic stubbornness.
The Turnout Question: Who Stays Home and Why It Matters
Elections are not decided solely by how people vote. They are decided by who votes – which is itself determined by who is registered, who is motivated, who faces barriers, who is mobilized by campaigns, and who decides, rationally or not, that their vote does not matter. The decision not to vote is itself a political act, though one whose consequences are distributed differently than the decision to vote. Non-voters tend to be younger, lower-income, less educated, and less partisan than voters. The electorate that shows up is systematically more older, wealthier, more educated, and more partisan than the eligible population. Policy made by governments elected by this electorate predictably differs from policy that would result from full participation. High-turnout elections produce different outcomes than low-turnout elections, not randomly but systematically, in ways that are predictable from the demographic profile of who turns out.
Campaigns understand this perfectly. Get-out-the-vote operations are targeted at specific populations who are identified as likely supporters but unlikely to vote without mobilization. Voter suppression efforts – whether legal or illegal – are targeted at specific populations identified as likely opponents. Both sides are engaging in turnout manipulation; one side does it by increasing participation and the other by decreasing it, and the difference between mobilization and suppression is not always obvious from the outside, particularly when the same legislature that writes voter ID laws is also the legislature that determines which IDs are acceptable, which is a remarkable alignment of interest and authority that has been present throughout American electoral history and has never been successfully described as coincidental by anyone paying attention.
SOURCE: The London Prat