The Policy Window

In brief

John Kingdon, a political scientist at the University of Michigan, published Agendas, Alternatives, and Public Policies in 1984 after interviewing officials in health and transportation policy across several years. He wanted to answer a question that rational models of government could not: why does an idea that has been available for years suddenly become something the government acts on, while equally sensible ideas sit ignored? His answer was that agenda change is not a tidy sequence in which a problem is spotted, solutions are weighed, and the best one is chosen. Instead three largely independent streams run through the system at once: problems, policies, and politics. Most of the time they run apart. Change happens in the rare moments when they are coupled, and a "policy window," a brief opening, lets a joined package onto the agenda. The person who does the joining is the policy entrepreneur.

The full treatment

The problem it answers

Textbook accounts of policymaking describe a rational cycle: identify a problem, generate options, evaluate them against goals, select the best, implement, review. Kingdon found that this described almost nothing he observed. Solutions often existed long before the problems they were attached to. Officials could not say why a given issue rose to prominence in a given year. Timing seemed to matter more than merit. Kingdon needed a model that treated government less as a calculating machine and more as an arena of ambiguity, where the crucial event is not choosing among options but getting an issue taken seriously in the first place. His focus is therefore narrow and deliberate: not the whole of policymaking, but agenda setting, the stage at which some subjects command attention and most do not.

How it works

The three streams flow on their own tracks. The problem stream is about how conditions come to be seen as problems that government should fix. A condition becomes a problem through indicators (a rising fatality rate), through a focusing event (a crash, a crisis, a disaster), or through feedback from existing programs that are visibly failing. Nothing is a problem in itself until it is framed as one.

The policy stream is Kingdon's most distinctive idea. He borrowed a metaphor: proposals float in a "policy primeval soup," generated by specialists, think tanks, and researchers who advocate their pet solutions for years. Ideas survive in the soup through a process of "softening up," repeated exposure that gradually makes a proposal technically feasible and acceptable to the community. Solutions, in this picture, exist before problems and go looking for them.

The politics stream runs separately again: the national mood, organized interests, turnover in the legislature, a change of administration. A new government or a shift in public sentiment can make the same proposal viable this year that was dead last year.

Change requires coupling. When a window opens, an entrepreneur attaches a ready solution from the policy stream to a recognized problem, and presses while the political mood allows. Kingdon's phrase for the person who does this is the policy entrepreneur: an official, a lobbyist, an academic, anyone willing to invest reputation, time, and resources to join the streams at the right instant.

The window and the focusing event

The window is the opening through which a coupled issue reaches the agenda. Kingdon stressed that windows are brief and often unpredictable. Some are predictable (a budget cycle, a scheduled reauthorization); many are not. A focusing event, a plane crash, an oil spill, a mass shooting, a financial collapse, can force a problem onto the agenda overnight. But the event alone does nothing. Unless a worked-out solution already exists in the soup, and the politics allow, the window closes with nothing passing through it. This is Kingdon's sharpest claim: the disaster does not generate the policy. It creates a moment in which a policy that was already sitting there can be attached.

What it draws on and what it rejects

The framework is built on Herbert Simon's bounded rationality: decision makers have limited attention and cannot process everything at once, so what gets attention is itself the scarce resource being fought over. Kingdon adapted the "garbage can model" of Cohen, March, and Olsen (1972), which described organizations as collections of problems, solutions, and choice opportunities dumped together and coupled almost by accident. Kingdon's move was to loosen the garbage can further and apply it to national government, replacing the idea of rational selection with the messier idea of streams that occasionally align.

Lineage

The direct ancestor is the garbage can model of organizational choice (Cohen, March, and Olsen, 1972), itself a reaction against rational-actor theories of the firm. Behind that stands Herbert Simon's work on bounded rationality from the 1940s and 1950s, and Charles Lindblom's 1959 account of policymaking as "muddling through" by small increments rather than comprehensive analysis. Kingdon synthesized these into a model aimed squarely at agenda setting. The framework was later formalized and extended by Nikolaos Zahariadis, whose Ambiguity and Choice in Public Policy (2003) tightened the concepts and applied them comparatively beyond the United States, turning Kingdon's rich but loose description into something closer to a testable framework.

The strongest case for it

Its great strength is fidelity to how policy change actually feels to the people inside it. Practitioners recognize the picture at once: the years a proposal spends going nowhere, then the sudden window after an election or a crisis when everything moves. The model captures timing, contingency, and luck, which rational-choice accounts cannot. It explains a genuine puzzle, why solutions so often precede problems, and why the same idea fails then succeeds with no change in its content. And it is generative: it has produced a large body of empirical studies across health, environment, education, and foreign policy, and it travels, having been applied well beyond the American federal setting for which it was built.

The strongest case against it

The most persistent charge is that the framework is too flexible to be wrong. Because "coupling" and "windows" can be identified only after a change has occurred, critics argue the model risks becoming an after-the-fact narrative that fits any outcome and predicts none. It describes vividly but forecasts poorly.

A second line of attack comes from the leading rival model. Frank Baumgartner and Bryan Jones, in Agendas and Instability in American Politics (1993), advanced punctuated equilibrium theory: policy is marked by long periods of stability broken by sudden large shifts, driven by changes in how issues are framed ("policy images") and by which institutional "venue" gains control of an issue. Their account is more structural and has generated large datasets showing the characteristic pattern of stasis and lurch, which they argue explains the same phenomena with more rigor than the streams metaphor.

Third, critics question the independence of the three streams. In practice, the political mood shapes which conditions get seen as problems, and problems shape which solutions get funded, so treating the streams as separate may be a convenient fiction rather than a fact. Zahariadis himself acknowledged the streams are not fully independent. Finally, some object that "policy entrepreneur" is close to a residual category: when a window is exploited we call someone an entrepreneur, and when it is not we say none was present, which explains everything and therefore little.

Where it stands now

Multiple-streams remains one of the two or three dominant frameworks in policy-process research, taught in nearly every public-policy program and used in a large and continuing body of empirical work. Reviews of the field, including work by Michael Howlett, Allan McConnell, and Anthony Perl, place it among the most applied lenses, alongside punctuated equilibrium and the advocacy coalition framework. The main scholarly effort now is to make it more testable: to specify when a window is likely to open, what makes an entrepreneur effective, and how the streams interact rather than merely coincide. It has not displaced its rivals, and it is best understood not as a law but as a way of seeing, one that reliably points attention to timing, framing, and the people who exploit the moment. That is why it endures.

Test yourself

Think of a policy that passed in your country after a disaster or a scandal. Ask whether the event created the solution, or merely opened a window for a solution that specialists had been pushing for years. If the answer is the second, you have found a coupling, and the person who made it happen was the entrepreneur.