The popular view is that campaign contributions buy legislative votes — that donors write checks in exchange for favorable roll-call decisions, and that the connection between money and policy outcomes is both obvious and pervasive. Political scientists have spent decades trying to establish whether this is true, and their findings are more complicated than either the populist indictment or the industry-funded counterargument suggests.
The research does not support a simple story of vote-buying. It does support a more nuanced and arguably more troubling account of how money shapes policy: not primarily through direct exchanges but through access, agenda-setting, attention, and the systematic amplification of some interests over others in every stage of the legislative process.
The Methodological Challenge
The fundamental problem in this research is reverse causation combined with selection effects. Industries and organizations contribute money to candidates who already support their policy positions. A pharmaceutical company does not give to a senator who opposes drug company interests; it gives to senators who are friendly to its concerns — either because of shared ideology, constituency ties, or prior relationships. Any observed correlation between pharmaceutical contributions and favorable pharmaceutical votes might reflect the contribution causing the vote, or it might reflect both the contribution and the vote flowing from an underlying alignment that existed before any money changed hands.
Formally, this is an omitted variable problem: the variable that drives both giving and voting — underlying sympathies and ideological alignment — is not directly measurable, and its omission from statistical models produces estimates of contribution effects that are difficult to interpret causally. This problem has led to the strange circumstance in which a meta-analysis of 36 studies on contributions and roll-call votes concluded that only one in four found evidence that contributions affected votes — not because researchers found contributions clearly harmless, but because the methodological bar for establishing causality is high and most research designs cannot meet it.
The causality problem goes both directions. Studies that find no effect of contributions on roll-call votes may be accurately capturing the absence of direct vote-buying, or they may be missing real effects because their research designs cannot isolate them. Studies that do find effects may be finding genuine causal relationships, or they may be picking up the underlying alignment that drives both giving and voting.
The Roll-Call Vote Literature
The most comprehensive meta-analysis of the direct contributions-to-votes research was Ansolabehere, De Figueiredo, and Snyder’s 2003 paper “Why Is There So Little Money in Politics?”, which examined 36 studies of PAC contributions and congressional floor voting. Their conclusion — that the evidence for a systematic money-to-votes relationship is weak — was striking given the widespread belief that such a relationship exists and has been documented.
Their interpretation was that political contributions function more like consumption by donors than investment — that donors give because they support candidates and causes, not primarily to extract policy benefits. This view is consistent with the observation that contribution amounts are small relative to the potential policy benefits they might purchase, and that individual donors rarely seem to calibrate their giving to the specific legislative value of the candidates they support.
But this meta-analysis finding has been contested. Thomas Stratmann’s 2005 meta-analysis of the same literature found evidence of significant money-to-vote effects that Ansolabehere et al. had downplayed. Other researchers have argued that the failure to find roll-call effects in broad cross-sectional studies may reflect the research design rather than the absence of effects: studies that aggregate contributions across many industries and many bills may dilute real effects that would be visible within specific industries and on specific votes.
The most methodologically sophisticated recent work on this question illustrates both the difficulty and what careful design can reveal. A 2022 study of the sugar industry published in the American Political Science Review took advantage of a natural experiment created by two nearly identical votes on anti-sugar-subsidy amendments — one in 2013 and one in 2018. Between the votes, the sugar industry dramatically increased its campaign contributions to House members, with sugar contributions rising by more than 50 percent in real terms between the two votes. Using district and incumbent fixed effects to control for underlying support levels, the authors found that increased contributions significantly increased the probability that a member would vote against sugar subsidy reform. The effect was causal, not merely correlational, because the research design controlled for a district’s inherent supportiveness of the issue.
The study’s authors concluded that the broad literature’s failure to find systematic roll-call effects may reflect the difficulty of isolating effects in wide-sample studies, while targeted research on specific industries with identifiable economic interests does tend to find positive relationships — “even controlling for district fixed effects.”
Beyond Roll-Call Votes: Access and Agenda-Setting
Even if direct vote-buying is rare and difficult to document, the influence of money on the legislative process operates through channels that roll-call vote studies are not designed to detect.
Richard Hall and Frank Wayman’s foundational 1990 study on committee participation found that PAC contributions affected not primarily how members voted on final passage, but how actively they participated in committee deliberations on behalf of donor interests — whether they sponsored or opposed amendments, how much floor time they spent on an issue, whether they mobilized colleagues. They characterized contributions as being allocated “in order to mobilize legislative support and demobilize opposition” and found that contributions “did buy the marginal time, energy, and legislative resources that committee participation requires.” This is a subtler form of influence than direct vote-buying, but it shapes legislative outcomes nonetheless — particularly in the committee stage, where most bills are written and amended before floor votes occur.
The access research discussed earlier reinforces this picture. The Kalla-Broockman field experiment established that campaign donors receive disproportionate access to legislators and their senior staff. Access matters for policy outcomes not because it directly determines votes but because it provides the opportunity to shape how legislators understand issues, what options they consider, and which concerns they hear repeatedly. An industry that can obtain meetings with relevant committee members and their staff before legislation is drafted has advantages that an industry without such access cannot easily replicate.
The Gilens-Page Analysis
The most widely cited recent study on the relationship between money, preferences, and policy outcomes operates at a higher level of aggregation than the roll-call vote literature. Martin Gilens and Benjamin Page’s 2014 paper “Testing Theories of American Politics”, published in Perspectives on Politics, analyzed roughly 1,800 policy proposals across several decades, coding the preferences of economic elites, business-oriented interest groups, mass-based interest groups, and average citizens on each proposal, then examining which preferences were associated with actual policy outcomes.
Their findings were unambiguous on one point: “the preferences of the average American appear to have only a minuscule, near-zero, statistically non-significant impact upon public policy” when the preferences of economic elites and organized business interests are controlled for. Economic elites and business-oriented interest groups have substantial independent impacts on policy outcomes; average citizens and mass-based interest groups have little.
The Gilens-Page study has generated significant debate, including about what “economic elites” means in their dataset (they proxy elite preferences with the views of the top ten percent of earners, which is a rough measure), whether their model correctly isolates independent effects, and how to interpret cases where elite and mass preferences happen to align. Some critics have noted that ordinary citizens often get the policies they favor, but largely in cases where those policies are also favored by elites. The study does not show that elites always win; it shows that when elite and mass preferences diverge, elite preferences predict outcomes and mass preferences do not.
This does not establish a direct causal role for campaign contributions specifically. Wealthy individuals and business interests have many channels of political influence beyond contributions — lobbying expenditures, revolving door relationships, information provision, and the basic political weight that comes from economic scale and social proximity to governing elites. But the Gilens-Page findings are consistent with a system in which campaign contributions are one part of a broader structure of influence that systematically amplifies some voices over others.
Lobbying and Policy Outcomes
The research on lobbying expenditures and policy outcomes is separately documented and reaches more consistent conclusions. Lobbying has been found to affect whether bills move through committees, whether legislation passes, and whether specific provisions survive the amendment process. A Journalist’s Resource review of the lobbying literature found specific empirical evidence of lobbying effects on tax rates, visa allocation, climate policy, and various regulatory outcomes.
Research on lobbying status quo effects suggests that it takes roughly 3.5 lobbyists advocating for policy change to overcome one lobbyist defending the status quo — meaning industries with established regulatory arrangements and ongoing lobbying presences have a structural advantage in maintaining those arrangements against reform. The policy implications are directional: industries that have been heavily regulated often lobby successfully to prevent tightening of those regulations, while industries seeking new favorable arrangements lobby to create them.
Specific studies have found that a 1 percent increase in lobbying expenditures is associated with a reduction in a corporation’s effective tax rate of between 0.5 and 1.6 percentage points, and that corporate campaign contributions at the state level are associated with lower state corporate tax rates. These are observational findings and face the same endogeneity challenges as the contributions-votes literature, but they are consistent with a picture in which organized political spending shapes policy in measurable ways at the margins.
What “Influence” Means in Practice
The concept of influence that most of the research actually documents is not vote-buying — direct exchanges of money for specific legislative decisions. Direct vote-buying of this kind would be illegal, and the evidence does not suggest it is widespread. The influence documented in the research is more structural and more diffuse.
It operates through access: organizations that contribute obtain meetings, briefings, and informal relationships with legislators and staff that non-contributing organizations do not get. It operates through agenda-setting: legislators who hear repeatedly from contributors about the consequences of proposed regulations develop a familiarity with those perspectives that shapes how they define problems and what solutions they consider. It operates through mobilization: contributions buy the legislative effort — the amendments sponsored, the colleagues lobbied, the floor time spent — that shapes how a bill is written before any final vote occurs. And it operates through representation: as the Gilens-Page study documents, the policy preferences of economically advantaged groups are more likely to be reflected in policy outcomes than the preferences of average citizens, suggesting that something about the political system consistently amplifies some voices over others.
Where the research is genuinely contested is in quantifying these effects with precision. The methodological challenges are real, not merely convenient objections by interested parties. Establishing that contributions causally change votes — as opposed to correlating with votes because contributors and legislators share underlying sympathies — is genuinely hard. Researchers who are skeptical of strong causal claims have legitimate methodological grounds for that skepticism.
Where the research is less contested is in the directional story it tells: that organized interests with sustained political spending relationships obtain better access, more legislative attention, and outcomes more consistent with their preferences than unorganized interests without such relationships. That pattern is documented across multiple research designs, multiple policy domains, and multiple time periods. The magnitude of the effect is debated; the existence of a systematic advantage is not.
America’s Plan covers campaign finance and governance reform. Related articles address campaign money and lobbying, the FEC’s enforcement record, and who actually funds American campaigns.
This article was researched and drafted with AI assistance under human review. See our full AI and editorial practices.