Agenda capture

Agenda capture is the persistent control, by a narrow set of actors, over which issues and questions reach public attention, official decision makers, and the ballot. Political scientists Peter Bachrach and Morton Baratz argued in 1962 that power has “two faces,” one of which works by keeping certain issues “out of the public domain” rather than by winning votes on them. Because what is never asked is never decided, agenda power can shape outcomes before any public debate begins.

In plain words: before people vote or give an opinion, someone decides what the choices are. A news editor picks the stories. A legislative leader picks which bills get a hearing. A pollster picks the list of answers. A petition needs a certain number of signatures before it reaches the ballot. Each of those gates is needed, because no one can consider everything. The problem starts when the same few hands control the gates and the public cannot see how. This page explains how agenda power works, what has been tried to open it, and why every fix creates new gatekeepers.

The problem, defined precisely

Agenda capture sits on top of several older concepts, and clear terms keep it from becoming a catch-all complaint.

Agenda. The set of issues receiving serious attention from the public, the media, or a decision-making body at a given time.

Agenda setting. In communication research, the influence of news coverage on which issues the public considers important. Maxwell McCombs and Donald Shaw (1972) wrote that in choosing and displaying news, “editors, newsroom staff, and broadcasters play an important part,” and that “the media may set the ‘agenda’” of a campaign.⁠²

Gatekeeping. Any rule or role that decides whether an issue moves forward: a committee chair, a signature threshold, an editor, a survey designer.

Second face of power (nondecision). Bachrach and Baratz argued that “there are two faces of power, neither of which the sociologists see and only one of which the political scientists see.” The less visible face involves sustaining values and procedures “that help him keep certain issues out of the public domain,” and “neither involves nor can be identified with decisions arrived at on specific issues”.⁠¹ The article also refers to the “mobilization of bias”.⁠¹

Question framing. The wording, options, and order of a question, which can change the answers. See Question design.

Regulatory capture. A related but narrower idea. Daniel Carpenter and David Moss define it as “the result or process by which regulation, in law or application, is consistently or repeatedly directed away from the public interest and toward the interests of the regulated industry, by the intent and action of the industry itself”.⁠⁷

Agenda capture, as used on this page, means control over what reaches the agenda, not over the final decision. It is a matter of degree, and it includes both deliberate exclusion and structural bias that no one intends.

Scale and evidence

Agenda capture is difficult to measure, because its clearest form, an issue that never appears, leaves no record. The evidence below therefore comes from specific gates where the effect can be observed.

Question lists shape stated priorities. In a 2008 Pew Research Center experiment on the most important issue in the presidential vote, 58% chose “the economy” when it was offered in a closed list, compared with 35% who volunteered it in an open-ended version, and 43% of open-ended respondents gave an answer not on the list.⁠³ Gallup’s long-running “most important problem” question is asked in open-ended form.⁠⁴ Whoever writes the list holds real power over the result. See Priorities.

Lobbying and the status quo. Frank Baumgartner, Jeffrey Berry, Marie Hojnacki, David Kimball, and Beth Leech studied 98 issues and found that lobbying efforts face “an entrenched Washington system with a tremendous bias in favor of the status quo,” and that “resources explain less than five percent of the difference between successful and unsuccessful efforts”.⁠⁶ This cuts two ways: organized money does not reliably win, but the starting agenda is hard to move.

Scarce attention. Joshua Kalla and David Broockman note that officials’ time is finite, “so when they decide to spend time hearing the concerns of some individuals, they have less time to hear others’,” and their experiment found donors gained meetings with senior staff three to four times more often.⁠¹⁶ See Money in politics.

Fabricated participation in agency agendas. Under the Administrative Procedure Act, federal agencies must publish notice of proposed rules in the Federal Register and “give interested persons an opportunity to participate” through written comments, then consider “the relevant matter presented”.⁠⁸ In 2021 the New York Attorney General reported that “nearly 18 million of the more than 22 million comments” the Federal Communications Commission received in its 2017 net neutrality proceeding were fake. The report found that “the nation’s largest broadband companies funded a secret campaign to generate millions of comments,” and that most of 9.3 million fake comments on the other side, supporting net neutrality, were submitted by a single 19-year-old college student using automated software.⁠¹⁰ The U.S. Government Accountability Office found in 2019 that “agencies aren’t required to collect or verify commenters’ identity information” and that unclear posting practices could lead the public to wrong conclusions about who commented and how many.⁠⁹

Ballot access. Twenty-four states, the District of Columbia, and the U.S. Virgin Islands have a citizen initiative process (NCSL’s count notes that Mississippi’s was found void in 2021); 23 states have the popular referendum; legislative referrals can appear in all 50 states. Initiative signature thresholds are typically a percentage of votes cast for a statewide office in the preceding general election, and states often restrict content, for example through single-subject rules.⁠¹¹ Mississippi’s requirement of signatures from five congressional districts became impossible to meet after the state dropped to four districts.⁠¹¹ The rules of the gate, not only the voters, decide what reaches the ballot.

Causes and competing explanations

Several mechanisms produce agenda capture, and scholars weigh them differently.

Scarcity. Legislatures, newsrooms, and voters can attend to only a few issues at once. Some gatekeeping is therefore unavoidable, and the question is who controls it.⁠¹⁶

Organization. Interests that are concentrated, organized, and funded can monitor and act at every gate; diffuse interests often cannot. Bachrach and Baratz’s point is that this advantage operates partly through rules and values that keep issues off the table, not only through visible lobbying.⁠¹

Status quo bias. Baumgartner and colleagues’ finding that the system favors the status quo suggests that much agenda power is defensive: blocking change is easier than causing it.⁠⁶

Media selection. McCombs and Shaw’s agenda-setting hypothesis locates part of the power in news selection.⁠² Where local reporting declines, fewer outlets make those choices locally. See Local news collapse.

Capture of agencies. Carpenter and Moss argue that capture exists “by degrees rather than by its presence or absence,” that much of it is “weak” capture compatible with regulation still serving the public, and that the picture is “quite different from the one George Stigler envisioned, in which capture by industry was virtually inevitable and complete”.⁠⁷ They also note that groups other than the regulated industry, such as “labor unions or certain activists,” can capture a regulation.⁠⁷

Manufactured signals. Where participation is cheap and identity is unverified, the volume of input can be faked, as the FCC proceeding showed on both sides of the issue.⁠^10,9

Who is harmed

Agenda capture harms people whose concerns never become questions, and that harm is real even though it is hard to count.

Diffuse and unorganized publics lose most, because they lack the resources to watch every gate. See Unequal participation and Representation gap.

Decision makers are harmed when the signals they receive are distorted: an agency that receives millions of fabricated comments cannot easily read genuine public views.⁠¹⁰

Voters are harmed when ballot questions reflect who could afford to qualify them rather than what most people wanted to decide, though whether initiatives on balance favor organized or popular interests is disputed.

Organized interests can also be harmed by opaque gatekeeping, when a committee chair or an agency staff quietly blocks an issue. Agenda capture is not only a problem of the powerful against the weak; it is a problem of unaccountable control.

What has been tried elsewhere

Many democracies have built formal routes for the public to raise issues, and each has a gate of its own.

The citizen initiative (United States). States with initiatives let citizens bypass the legislature if they meet signature thresholds.⁠¹¹ In Meyer v. Grant (1988), a unanimous Supreme Court struck down Colorado’s ban on paying petition circulators, holding that circulating a petition is “core political speech”.⁠¹² That protects political expression; critics argue it also lets well-funded sponsors buy ballot access. Both readings can be true.

The European Citizens’ Initiative. One million EU citizens, with minimum numbers in at least seven member states, can ask the European Commission to propose legislation. The Commission must state “within 6 months” what action it will take, but “the Commission is not obliged to propose legislation”.⁠¹³ The initiative opens the agenda without transferring the decision.

UK Parliament petitions. An e-petition with 10,000 signatures receives a government response; with 100,000 “it may be debated in Parliament.” The Petitions Committee decides whether to hold a debate and can decline, for example if the issue was recently debated.⁠¹⁴ A threshold produces attention, not a vote.

Citizens’ assemblies (Ireland). Ireland’s assemblies use stratified random selection: letters go to randomly chosen households, and members are selected to reflect the population by gender, age, geography, and other characteristics (99 members in the assembly on drug use). Each assembly is established by the Oireachtas, which defines its terms of reference.⁠¹⁵ Random selection broadens who deliberates, but parliament still chooses the topic.

Comment-process reform. After its 2019 review, GAO made eight recommendations to agencies to clarify how they post and treat comments and identity information; the GAO page lists all eight as closed and implemented.⁠⁹ Clearer posting rules do not, by themselves, verify who is commenting.

Measuring the agenda. The Comparative Agendas Project classifies government policy activities “into a single, universal, and consistent coding scheme” and lists 20 active projects across countries and subnational units.⁠⁵ Measurement does not open the agenda, but it makes shifts in attention visible and comparable.

The strongest arguments across the political spectrum

Agenda power is criticized from every direction, and defended from several.

The institutionalist argument for gatekeepers. Gatekeeping filters proposals for feasibility, legality, and cost, and protects deliberation from being swamped. Status quo bias provides stability and protects minorities from sudden change.⁠⁶ The weak point is that filters chosen by insiders can quietly exclude legitimate concerns, which is Bachrach and Baratz’s warning.⁠¹

The populist and direct-democracy argument. Voices on the left and right argue that elites, parties, and donors control what is debated, and that initiatives, petitions, and referendums return agenda power to citizens.⁠¹¹ The weak point is that every open route has thresholds and costs, and paid circulation and organized mobilization may let the best-resourced groups use the open door most.⁠¹²

The public choice and libertarian argument. If government holds power to grant favors, interests will compete to capture it; the surest remedy is to reduce the favors available. Carpenter and Moss summarize this view as Stigler’s, in which capture was “virtually inevitable and complete” and the response is to weaken or remove regulation.⁠⁷ The weak point, in their account, is that much capture is weak and limited, so removal can sacrifice real public benefits.⁠⁷

The reform-by-design argument. Carpenter and Moss argue that because capture is usually partial, the better response is to reduce its severity, or prevent it, rather than to purge regulation wholesale.⁠⁷ The weak point is that safeguards themselves need maintenance and can be captured.

The deliberative argument. Advocates of randomly selected bodies argue that sortition gives agenda influence to ordinary people who are not self-selected activists.⁠¹⁵ The weak point is that topic selection and expert briefing remain in official hands, so a deliberative body can deepen an agenda without setting it.

The constitutional-conservative argument. Some argue high thresholds and single-subject rules for ballot measures protect constitutions and laws from transient or poorly drafted proposals.⁠¹¹ Critics reply that thresholds can be raised to block popular measures; whether a given threshold protects or obstructs depends on the case.

Unresolved questions

• How can nondecisions, issues that never reach the agenda, be measured at all?⁠¹
• Do open agenda routes (initiatives, petitions) on balance favor broad publics or organized sponsors, and under which rules?⁠^11,12
• Can participation processes verify that inputs come from real, distinct people without creating identity databases that themselves become risks?⁠^9,10 See Identity and registration.
• Should randomly selected bodies be allowed to choose their own topics, and with what limits?⁠¹⁵
• Who should write and revise a participation framework’s subject list, and what review should that face? See Governance.
• How much does the decline of local news change who sets local agendas? See Local news collapse.