Prospective voting is the act of casting a ballot based on what you expect candidates or parties will do in the future, rather than rewarding or punishing them for what they have already done. It is one of the two dominant frameworks political scientists use to explain voter behavior, and recent research suggests it plays a more central role than many traditional election models assumed. A forecasting model built entirely on prospective voting theory, focusing on which candidate voters trusted more on key issues, predicted the 2024 U.S. presidential popular vote with an average error of just 0.65 percentage points across the campaign season, outperforming many polling-based forecasts that relied on retrospective economic indicators.1SAGE Journals (Research & Politics). Prospective voting and the issues and leaders model: Forecasting the 2024 U.S. presidential election
What Makes Prospective Voting Different
The opposite of prospective voting is retrospective voting, where you evaluate the incumbent’s track record and decide whether to keep them or throw them out. Classic retrospective models lean heavily on measurable outcomes: GDP growth, unemployment, inflation, whether the country is at war. The logic is simple and intuitive. If things are going well, reward the party in power. If not, vote for the other side.
Prospective voting flips that orientation. Instead of asking “how have things been?” you ask “how will things be?” That means evaluating campaign promises, assessing which candidate seems more competent on the issues you care about, and forming expectations about future economic conditions or policy directions. In practice, most voters blend both approaches. But the balance between backward-looking and forward-looking reasoning varies enormously depending on the voter, the election, and the circumstances.
The distinction matters because the two approaches can lead to opposite conclusions. A voter using retrospective logic during an economic downturn will almost always punish the incumbent. A voter thinking prospectively might still support the incumbent if they believe that candidate has a better plan for recovery than the challenger. This is one reason why incumbents sometimes survive bad economies and challengers sometimes lose despite favorable conditions.
Who Actually Votes Prospectively
Not everyone engages in forward-looking evaluation equally. Research consistently finds that political knowledge is a key divider. Voters with higher levels of political expertise tend to engage in more sociotropic, ideological, and policy-based voting, meaning they weigh broader national considerations and specific policy positions rather than just personal economic circumstances.2Political Research Quarterly. Political Expertise and Issue Voting in Presidential Elections That kind of issue-centered evaluation is inherently prospective: you are comparing what each candidate says they will do and judging which agenda you prefer.
A study using incentivized forecasting tasks found a strong connection between political interest and the accuracy of people’s prospective beliefs. In other words, people who pay close attention to politics are better at predicting what candidates and parties will actually do once in office. But the same study uncovered a complicating wrinkle: strong partisan identification was negatively related to accuracy. The more intensely someone identified with a party, the worse their prospective predictions tended to be.3Quarterly Journal of Political Science. An Assessment of Citizens’ Capacity for Prospective Issue Voting using Incentivized Forecasting Interest helps, but tribalism hurts.
Economic expectations follow a similar pattern. Prospective economic evaluations carry the most weight among the most informed voters. These voters rely on national economic forecasts when they believe the incumbent party is likely to win, and they shift toward personal financial expectations when the election outcome is uncertain or they think the challenger has the edge.4Political Behavior. Who Votes for the Future? Information, Expectations, and Endogeneity in Economic Voting Less informed voters, by contrast, tend to lean more on retrospective judgments, essentially using recent experience as their main guide.
How Partisanship Warps Future Expectations
One of the biggest challenges with prospective voting is that people’s expectations about the future are not formed in a vacuum. They are deeply colored by existing partisan loyalties. Research across multiple countries shows that supporters of the current government tend to expect the economy to improve, while opponents tend to expect it to decline, regardless of what objective economic indicators suggest.5Comparative Political Studies. Partisan Preferences, Electoral Prospects, and Economic Expectations
This creates a chicken-and-egg problem for researchers. When a voter says “I’m voting for Candidate X because I think the economy will improve under them,” is that a genuine forward-looking assessment, or is it their partisan loyalty generating a convenient economic forecast? The answer, in most cases, is probably both. People genuinely do form economic expectations, but those expectations are filtered through the lens of who they already support. The practical effect is that prospective economic voting is partially a mechanism for expressing partisanship in economic clothing.
Emotions compound the problem. A recent study in the United Kingdom distinguished between two psychological states that shape expectations about the political future: optimism and hope. Optimism, the researchers found, drives positive expectations about societal outcomes in general. Hope, on the other hand, drives partisan expectations about election outcomes specifically. In fact, partisan bias in predictions about who would win an election showed up only among people who scored higher on measures of hope. Those who were less hopeful were more evenhanded in their electoral predictions. Interestingly, providing positive information from polls and expert commentary dampened hope’s distorting effect by raising expectations among those who were otherwise unhopeful.6SpringerLink. Hope, Optimism, and Expectations for the Political Future
Do Campaign Promises Actually Get Fulfilled
Prospective voting only works as a democratic tool if the promises candidates make have some relationship to what they actually do in office. There is a persistent cynicism about this, a sense that politicians say whatever gets them elected and then do as they please. The evidence, while messy, is more encouraging than the cynical take would suggest.
An analysis comparing Donald Trump’s and Joe Biden’s presidencies found that Biden achieved a higher rate of promise fulfillment at roughly 55%, compared to about 42% for Trump. A meaningful share of Biden’s fulfilled promises, about a third, came through traditional legislative channels, compared to less than a fifth for Trump, whose fulfilled promises relied more heavily on executive action. The study also found a positive correlation between how important the public considered an issue and whether the president followed through on related promises.7Emory University Digital Repository. Campaign Promises to Public Policy: An Analysis of Promise Fulfillment during the Presidencies of Donald Trump and Joe Biden
But raw fulfillment rates can be misleading. Not all promises are created equal, and treating a minor administrative pledge the same as a signature campaign commitment distorts the picture. Research on UK elections found that weighting promises by how central they were to the campaign dramatically changed the assessment. For one party, applying a centrality weighting reduced the apparent promise-keeping rate by 21 percentage points. Promises about the European Union and immigration, which were far more important to voters, were weighted six to seven times more heavily than miscellaneous administrative commitments that padded the raw numbers.8Political Studies. Which Promises Actually Matter? Election Pledge Centrality and Promissory Representation In short, whether politicians “keep their promises” depends heavily on which promises you count and how much each one matters.
For prospective voters, this means the strategy is only as good as the voter’s ability to distinguish core commitments from filler. A candidate might deliver on dozens of minor pledges while abandoning the two or three promises that actually mattered to their base. Evaluating that track record requires paying attention not just to what was promised, but to which promises were central to the campaign.
Why Voters Prefer Sooner Over Later
Even when voters are thinking about the future, they tend to discount it. Policies whose benefits arrive quickly are preferred over those that take years or decades to pay off. This temporal discounting is a well-known feature of human psychology, and it creates a structural bias against long-term policy investments like infrastructure spending, climate action, or pension reform. Politicians who promise immediate results have a built-in advantage over those whose best proposals would take a generation to bear fruit.
The good news is that this discounting effect appears to have limits, and it does not seem to be driven by pure selfishness. A large pre-registered study of over 2,400 people found that while citizens do prefer nearer-term benefits, they show no special preference for policies that pay off within their own lifetime as opposed to after it. People discount the future in general, but they are not drawing a line at their own mortality. The study also found that the timing effect was substantially smaller than other features voters cared about, like the overall size of the benefit. People were clearly willing to accept later payoffs if those payoffs were larger.9European Journal of Political Research. Policy discounting across and beyond the lifespan
Personal circumstances can also shift how much weight people place on the future. Becoming a parent, for instance, appears to lengthen people’s time horizons. A study using a matched panel design found that new parents showed increased support for stopping climate change compared to otherwise similar individuals who did not become parents. Additional experiments suggested that longer time horizons explained at least part of this shift.10Political Behavior. Valuing the Future: Changing Time Horizons and Policy Preferences This finding fits a broader pattern: when something makes the distant future feel more personally relevant, people become more willing to weigh it in their political decisions.
Prospective Voting During Crises
Crisis moments put the prospective-versus-retrospective tension in sharp relief. When a disaster hits, some voters focus on who is to blame (retrospective), but others pivot hard toward who they think can best handle the situation going forward (prospective). The pandemic offered a natural experiment in this dynamic.
A study of voting behavior at the onset of the COVID-19 pandemic found that initial local outbreaks favored whichever political party was governing at the state level and hurt the far right. The researchers attributed this to a “strategic alignment” mechanism: voters chose the party or candidate they believed was best positioned to help them through the crisis, regardless of whether that party had caused or worsened the situation.11SpringerLink. Strategic Alignment in Times of Crisis: Voting at the Dawn of a Global Pandemic This is prospective voting in its most urgent form. In a moment of genuine fear and uncertainty, a significant number of voters reached for whoever seemed most capable of navigating the problem, not whoever deserved credit or blame for the status quo.
The crisis context also connects to a broader finding about how voters handle political ambiguity. You might expect that voters would always prefer candidates who are specific and precise about their plans. But research on context-dependent voting shows that voters can actually develop a taste for ambiguity, even when they are generally risk-averse in other domains.12Journal of Public Economics. Context-dependent voting and political ambiguity During uncertain times, a candidate who keeps their options open may seem more adaptable than one locked into a specific plan. This helps explain why crisis-era campaigns sometimes reward vagueness over detail, a counterintuitive outcome for the prospective voting framework, which generally assumes voters want clear information about future plans.
Forecasting Elections with Prospective Voting Models
For most of modern political science, election forecasting has been dominated by retrospective models. These “fundamentals” models use variables like economic growth, presidential approval ratings, and incumbency to predict outcomes. They work reasonably well in normal elections, but they struggle when voters are making strong forward-looking judgments that break from recent trends.
The “Issues and Leaders” model represents a direct attempt to forecast elections using prospective voting theory. Rather than relying on economic fundamentals or incumbent approval, it measures voter perceptions of which candidate is more competent on the key issues and which one demonstrates stronger leadership qualities. For the 2024 U.S. presidential election, this model produced an Election Eve forecast of a near tie: 50.2% for Donald Trump and 49.8% for Kamala Harris in the two-party vote. That was more cautious than most polling-based forecasts, many of which showed a Harris lead. Trump ultimately won the popular vote by about 1.5 points, meaning the prospective model’s election-eve prediction was off by roughly a point and a half, while its average error across the entire campaign season was just 0.65 points.13SAGE Journals (Research & Politics). Prospective voting and the issues and leaders model: Forecasting the 2024 U.S. presidential election
One election does not prove a model, and the Issues and Leaders approach is relatively new compared to established fundamentals models that have decades of track records. But its performance in 2024 is a meaningful data point. It suggests that asking voters “who do you trust to handle this issue?” and “who seems like a stronger leader?” captures something real about how elections are decided, something that backward-looking economic data alone can miss.
Local Economic Conditions and Prospective Choices
Prospective voting is not limited to national elections or Western democracies. Research on Russian parliamentary elections in the Ural Federal District found that local unemployment rates were the most significant factor driving voter choices, outweighing monetary income levels. Voters in regions with higher unemployment tilted toward opposition parties, though which opposition party benefited varied by region. In one autonomous area, unemployment drove support toward one opposition party, while in a neighboring region the same economic anxiety fueled support for an entirely different one.14Общество: политика, экономика, право. Prospective Voting in the Study of Electoral Behavior in Parliamentary Elections (Using the Example of the Ural Federal District)
This pattern illustrates something important about how prospective voting interacts with local conditions. The same economic grievance, high unemployment, can push voters toward completely different political options depending on which party is seen locally as the most credible alternative. Prospective voting is not a monolithic force pointing in one direction. It is filtered through local political landscapes, the credibility of specific candidates, and even which party has the strongest ground presence in a given area.
AI Microtargeting and the Manipulation of Forward-Looking Voters
If prospective voting depends on expectations about the future, then anything that shapes those expectations has the power to shape votes. This is where artificial intelligence enters the picture, and not in a reassuring way.
AI-driven microtargeting in political campaigns works by building detailed profiles of individual voters and inferring their specific psychological vulnerabilities from their data. This allows campaigns to deliver highly tailored political messages designed to influence each person’s expectations and decisions. The concern is not just about persuasion in the traditional sense. It is about the capacity to construct a personalized information environment that steers individual voters toward particular conclusions about the future without those voters realizing they are being steered.15PubMed Central. The Complexities of AI for Social Good: An Explorative Study on Adversarial Machine Learning to Facilitate Deliberative Decision-Making in Elections
For prospective voters specifically, microtargeting is a potent tool because forward-looking decisions rest on beliefs about uncertain outcomes. If you are deciding based on what you think a candidate will do about the economy, health care, or immigration, then whoever controls the information you see about those topics controls the raw material your expectations are built from. Retrospective voters are somewhat harder to manipulate on their core judgment, because personal economic experience is hard to override with messaging. But expectations about the future are inherently soft. They are constructed from information, trust, and imagination, all of which are susceptible to manipulation at scale.
This does not mean prospective voting is inherently more vulnerable to abuse than other forms of political reasoning. Every mode of voter decision-making can be exploited. But the forward-looking nature of prospective voting creates a specific surface area for influence operations: the voter’s model of what will happen next. As AI tools grow more sophisticated in their ability to predict and shape individual beliefs, the integrity of prospective voting as a democratic mechanism depends increasingly on the quality and diversity of the information voters encounter.