Public Policy and the Policy Cycle

Why this matters in public value

Every public leader inherits a decision-making process that looks, from a distance, like a simple pipeline: a problem is noticed, options are weighed, a choice is made, the choice is carried out, and someone checks whether it worked. In practice, the problem that gets noticed is rarely the most serious one, the options on the table are rarely the full set that exists, and the choice that gets made often owes as much to which minister's diary had a gap that week as to the merits of the case. Understanding why the tidy version is a simplification, and where it breaks down, is not an academic nicety; it is what lets a leader recognize a genuine opportunity to act when one appears, rather than waiting for a process that will never arrive in the form the diagram promises.

Public value theory treats this messiness as a design constraint, not a failure to be engineered away. A policy's legitimacy and support depend on it having passed through a process that stakeholders recognize as fair, even when that process was contested and non-linear (see Topic 1.2 — The Strategic Triangle: Legitimacy, Value, and Capacity). A policy's operational capacity to deliver depends on choices made at the formulation and instrument-design stage, long before implementation begins, which means a leader who only engages once a policy reaches their desk for delivery has already lost the chance to shape whether it is deliverable at all. Knowing the cycle, and its limits, is how a leader finds the point of maximum leverage rather than arriving after every consequential decision has already been taken.

The other reason this matters now is that policy problems increasingly do not respect the boundary of a single instrument. A city's approach to reducing traffic congestion might combine a regulation, a charge, a subsidy for alternatives, and a public information campaign, and getting the combination right is a distinct skill from getting any one instrument right in isolation. A leader who thinks only in terms of 'which law do we need' or 'how much should we spend' misses the fact that the same policy objective can often be reached more cheaply, or more legitimately, through a different instrument or a different mix entirely.

Core concepts

The policy cycle is a stylized model of how public decisions are made, typically broken into stages: agenda-setting, formulation, adoption (or legitimation), implementation, and evaluation, often shown as a loop rather than a line because evaluation routinely reopens the agenda (see Public policy). Its value is pedagogical: it gives a shared vocabulary for talking about a process that would otherwise be described in as many ways as there are people describing it. Its danger is that the neat loop invites a reader to assume real decisions move through the stages in order, once, cleanly — which they rarely do.

Agenda-setting is the stage at which a condition in the world becomes a problem that a government or organization is expected to act on, out of the enormous range of conditions that could, in principle, receive that attention. Most social conditions never become agenda items at all; what separates the ones that do is not simply severity but visibility, a compelling narrative, a constituency pressing for action, and often a triggering event. The American political scientist John W. Kingdon built the most widely used account of this stage, the multiple streams framework, which argues that a problem stream (conditions recognized as problems), a policy stream (solutions already developed by specialists, often before any matching problem exists), and a politics stream (the mood of the public, election cycles, and organized interests) run largely independently of one another, and that agenda change happens when a policy window opens and a policy entrepreneur couples the three streams together (see Multiple streams framework; see John W. Kingdon). The counter-intuitive implication is that a solution is frequently ready and waiting before its problem becomes politically salient, rather than being designed in response to it — which is why so many crises are followed by the rapid adoption of a proposal that specialists had been advocating for years.

Policy formulation is the stage at which a recognized problem is translated into a set of candidate responses: draft instruments, their likely costs and effects, and the trade-offs between them, usually developed by officials, analysts, and outside experts rather than by elected decision-makers directly. Adoption, sometimes called legitimation, is the formal stage at which a candidate response is authorized — a vote, a signature, a regulation formally made — and it is at this stage that a policy acquires binding force and, ideally, the legitimacy that comes from a recognized authorizing process, not merely from having been well designed (see Topic 1.2 — The Strategic Triangle: Legitimacy, Value, and Capacity for how legitimacy is built and tested).

Implementation is where a formally adopted policy meets the organizations, budgets, and frontline staff that must actually carry it out, and it is the stage where the cycle's tidiness collapses most visibly. Michael Lipsky's concept of street-level bureaucracy describes how the police officers, teachers, benefits caseworkers, and inspectors who deliver a policy day to day exercise real discretion over how it is applied, meaning the policy citizens actually experience is often substantially reshaped by frontline judgement calls that no legislative text anticipated (see Street-level bureaucracy). A policy can be adopted with full legitimacy and still fail to create public value if its designers never asked whether the organization tasked with delivering it had the capacity, incentives, or discretion-management systems to do so.

Evaluation asks whether the adopted policy produced the intended effects, at what cost, and with what side effects, and it is the stage that formally closes the loop back to agenda-setting: a policy judged to be failing, or a problem judged to be worse than expected, is itself new agenda material (see Topic 2.6 — Evidence Synthesis and What Works for how evidence is gathered, weighed, and synthesized at this stage). Treating evaluation as the end of the cycle rather than its renewal is one of the most common structural mistakes an organization makes, because it treats the loop as a line.

Two influential critiques explain why the cycle so often fails to run as drawn. Incrementalism, developed by the American political scientist Charles E. Lindblom, argues that real policymakers rarely conduct the comprehensive, from-first-principles analysis the cycle model implies; instead they make small, successive adjustments to existing policy — 'muddling through' — because comprehensive analysis of every option is cognitively and politically impossible, and because small steps are easier to reverse if they turn out to be wrong (see Incrementalism; see Charles E. Lindblom). This is not a counsel of timidity: Lindblom argued incrementalism was often the more defensible strategy under genuine uncertainty, not merely the path of least resistance, though critics have long pointed out that pure incrementalism can also entrench a bad status quo by never asking whether the whole approach, not just its margins, is wrong.

The garbage can model, developed by Michael Cohen, James March, and Johan Olsen to describe decision-making in what they called an 'organized anarchy', goes further still: it portrays problems, solutions, participants, and choice opportunities as four largely independent streams sloshing around an organization, with a decision occurring whenever a choice opportunity happens to bring the other three together, rather than a solution being deliberately designed for a given problem (see Garbage can model). Kingdon's multiple streams framework is a direct, policy-specific descendant of this model. Read together, incrementalism and the garbage can model both argue that the neat, rational-sequential cycle is closer to a legitimating story organizations tell about their decisions after the fact than a description of how those decisions were actually reached — which is itself useful for a leader to know, because it changes how much weight to put on 'but that isn't how the flowchart says it should work' as an objection to a real decision.

The second half of this topic's territory is the policy instrument: the specific tool used to translate an adopted policy into effect. The British political scientist Christopher Hood offered the most durable taxonomy, sorting instruments by the government resource each one draws on: nodality (information and the government's position at the centre of networks), authority (the legal power to require, forbid, or permit), treasure (money, whether spent or collected), and organization (the government's own staff, land, and equipment), often abbreviated by the initials of the four resources (see Christopher Hood). This book groups Hood's categories into four families that map onto how most public leaders actually talk about instrument choice: regulation (authority — rules that require, forbid, or license a behaviour, backed by enforcement) (see Regulation); spending, including direct provision, grants, and subsidies that lower the cost of a favoured behaviour or good (see Subsidy); taxation, which raises revenue but can equally be designed to discourage a behaviour, as with a sin tax on tobacco, alcohol, or sugar (see Sin tax); and information or persuasion, from public awareness campaigns to the more targeted behavioural techniques of nudge theory, which redesigns the choice environment itself rather than mandating, paying for, or explaining a preferred outcome (see Nudge theory; see Topic 4.1 — Behavioural Public Policy for the ethics and mechanics of this instrument family in depth). Real policies almost always combine instruments from more than one family — a plastic bag charge is simultaneously a tax and an information signal, and a vaccination mandate is regulation reinforced by public information — and the leader's real instrument-design question is rarely 'which single tool' but 'which combination, in what sequence, at what cost to legitimacy.'

Best practices

1. Name which stage a live issue is actually in before proposing what to do next. Treating a problem still fighting for agenda space as though it needs a formulation workshop, or a fully formulated option as though it still needs agenda-building, wastes the effort spent on the wrong stage's tools.
2. Keep a standing shortlist of ready policy solutions, independent of any current problem. Kingdon's framework shows that windows open suddenly and close quickly; an organization with no developed options waiting in the policy stream will miss most of them.
3. Design the instrument mix before drafting the legal text. Decide whether the objective is better served by regulation, spending, taxation, information, or some combination, and only then write the instrument that delivers that choice, rather than reaching for a law because a law is the default.
4. Ask who will deliver this at street level, and whether they have the discretion, training, and capacity the design assumes. A policy engineered around an idealized implementer fails on contact with real caseloads, real staffing gaps, and real frontline judgement calls.
5. Build the evaluation plan at formulation, not after adoption. Deciding what success looks like, and how it will be measured, before the policy launches protects the evaluation from being shaped retrospectively to fit whatever happened.
6. Treat incremental adjustment as a legitimate strategy, not a failure of ambition, when the evidence is genuinely uncertain. Lindblom's case for 'muddling through' is strongest exactly when a bold, comprehensive redesign cannot be justified by what is actually known yet.
7. Periodically ask whether incrementalism has become a habit rather than a judgement. A policy area that has only ever received marginal adjustments for a decade may need a deliberate, comprehensive review, even though that review is more costly and disruptive than another small step.
8. Expect implementation to generate new problems that reopen the agenda, and plan for that loop explicitly. Build a route by which frontline feedback and early evaluation findings reach the people who can revise the policy, rather than treating implementation as a one-way, silent stage.
9. State the combination of instruments together, in the same document, not as separate initiatives. A charge, a subsidy, and a public information campaign aimed at the same behaviour are easier to evaluate, and to defend, as one coordinated policy than as three loosely related programmes.
10. Check the distributional effect of each instrument in the mix, not only the policy's aggregate effect. A tax that changes average behaviour can still fall disproportionately on low-income households; pairing it with a targeted subsidy or exemption is often what makes the combination legitimate (see Topic 3.5 — Equity, Fairness, and Distributive Justice).
11. Use behavioural instruments as a complement to, not a substitute for, authority and treasure when the stakes are high. A nudge is often the cheapest and least coercive first move, but it is rarely sufficient alone for a problem serious enough to justify regulation or spending (see Topic 4.1 — Behavioural Public Policy).
12. Document the political window that made adoption possible, while it is fresh. Understanding why a policy became adoptable when it did — which triggering event, which entrepreneur, which alignment of streams — helps the organization recognize the next window rather than assuming the same result can be summoned on demand.

Questions to discuss with your team

Do we actually know which stage of the cycle our current major initiatives are in, or are we applying the same playbook to all of them regardless? Walk through your three or four largest live initiatives and place each one on the agenda-setting, formulation, adoption, implementation, or evaluation stage, honestly, rather than by where the org chart says it sits. Discuss whether the tools and meetings you are using match that stage, or whether an initiative still fighting for attention is being treated as though it only needs better delivery planning. Agree what would change about how each initiative is resourced if the stage assessment is correct.

What ready-made solutions are sitting in our policy stream right now, waiting for a window we are not tracking? Ask each team to name one option they believe is well developed but has never found the political moment to be adopted. Discuss what would need to happen — an event, a change in leadership, a shift in public mood — for that window to open, and whether the option would genuinely be ready if it did. Consider whether the organization has any standing practice for keeping such options current, or whether they quietly go stale while waiting.

When we last designed a policy, did we choose the instrument first and the objective second? Trace back a recent decision and ask honestly whether 'we need a law' or 'we need to spend more' was decided before the underlying objective and the alternative instruments were properly weighed. Discuss what a genuine instrument-choice conversation would have looked like, comparing regulation, spending, taxation, and information against the same objective. Agree on a standard step in your policy-design process that forces this comparison before drafting begins.

Where has implementation quietly redesigned a policy we thought we had already decided? Identify a policy whose effect on the ground differs noticeably from what was written and adopted, and discuss what discretion frontline staff exercised to produce that gap. Consider whether that discretion is a problem to close down or a sensible adaptation the formal policy should be updated to reflect. Decide who is responsible for noticing this kind of drift and for deciding, each time, which of those two responses is right.

Are we treating evaluation as feedback into the next agenda, or as a report that gets filed once the programme ends? Review what actually happened to the findings of your last two major evaluations: did they change a live decision, inform the next round of formulation, or sit in a document nobody revisited (see Topic 2.6 — Evidence Synthesis and What Works)? Discuss what would need to change structurally — timing, ownership, a standing forum — for evaluation to genuinely close the loop rather than merely closing the file. Agree one concrete change you will make before your next evaluation is due.

Is our current approach in this policy area a deliberate incremental strategy, or an unexamined habit? Pick a policy area your organization has adjusted marginally, year after year, without a comprehensive review, and discuss honestly which of Lindblom's two readings applies: sensible caution under real uncertainty, or a comfortable default that has quietly outlived its justification. Consider what a genuine, from-first-principles review would cost in time and political capital, and weigh that explicitly against the cost of continuing to muddle through. Decide whether that review belongs on next year's agenda.

In practice: a public value example

A mid-sized coastal city government had spent several years discussing single-use plastic waste on its beaches and waterways without producing a formal policy. Environmental groups had raised the issue at council meetings for years; a voluntary code asking retailers to reduce plastic bag distribution had produced only a modest decline; and a proposal drafted by the city's sustainability office to combine a retail charge with a public awareness campaign had sat, fully formulated, in a drawer for over a year, unable to attract enough political attention to reach a vote.

The situation changed abruptly after an unusually severe storm surge washed large volumes of plastic debris onto the city's main tourist beach two weeks before the start of the summer season, generating extensive local media coverage and a sudden flood of complaints to elected councillors. A newly elected councillor, who had campaigned partly on environmental issues, saw the moment and pushed the sustainability office's existing proposal onto the agenda within days, coupling the dormant policy stream to the newly opened political and problem streams almost exactly as Kingdon's framework would predict. The proposal that had waited a year for attention was adopted within six weeks of the storm, not because it had improved, but because a window had opened.

The adopted policy combined four instruments deliberately, rather than relying on any single one. A regulation prohibited retailers from providing single-use plastic bags at the point of sale (authority). A small charge applied to any paper or compostable bag a customer still wanted (treasure, functioning as a mild sin-tax-style disincentive rather than a revenue measure). A subsidy funded free reusable bags distributed through community centres in the city's lower-income neighbourhoods, to prevent the charge from falling hardest on residents least able to absorb it. A public information campaign, developed with a local university's behavioural science group, redesigned the checkout prompt at participating retailers so that declining a bag became the default option rather than something a customer had to actively request — a nudge layered on top of the regulation and the charge, not a replacement for either.

Implementation surfaced a gap the original design had not anticipated: market inspectors responsible for enforcing the retail ban had discretion over how strictly to apply it to small, family-run stalls in the city's older market district, and early informal enforcement varied widely from one inspector to the next. Rather than treating this as a compliance failure to be punished, the sustainability office worked with the inspectors to formalize a graduated approach — warnings first, fines only after a repeated breach — which the inspectors had effectively already been applying through their own judgement. The formal policy was revised to match the practice its own street-level implementers had already found workable.

An evaluation eighteen months later found a substantial reduction in beach plastic debris, a smaller but real reduction in overall single-use bag distribution, and a distributional effect that was broadly neutral once the subsidized reusable bags were accounted for. The evaluation also flagged that awareness of the paper-bag charge had faded over time, weakening its behavioural effect, and recommended refreshing the information campaign rather than assuming a one-off launch would sustain itself indefinitely — feeding directly back into the agenda for the following budget cycle. This scenario is illustrative and fictional; it does not describe any real city's plastics policy or council decision.

Four sector lenses

Local government

A local authority typically sits closest to implementation and therefore sees the gap between an adopted policy and its street-level reality faster than any other level of government, but it usually has the narrowest instrument set to work with, since major taxation and much regulatory authority sit with national or regional government. The characteristic local skill is combining a limited set of instruments — local regulation, small charges and fees, direct spending, and locally targeted information — creatively enough to make real progress on a problem whose largest levers belong to someone else.

National government

National government typically holds the broadest instrument set, including the capacity to legislate, tax, and spend at scale, which makes instrument-mix design a genuinely national craft rather than an improvisation. The characteristic national risk is the opposite of the local one: with every instrument technically available, national policy formulation can default to the biggest or most visible tool — a new law, a headline tax — when a smaller, cheaper combination would achieve the same public value at lower cost to legitimacy and lower risk of unintended effects.

Social sector and nonprofit

Nonprofits and non-governmental organizations rarely hold formal policy-instrument authority themselves, but they are frequently the policy entrepreneurs who couple a ready solution to an opening political window, and they are often the organizations closest to how a policy actually lands on the people it is meant to help. Their distinctive contribution to the cycle is agenda-setting and implementation feedback: raising a problem's visibility long before government attention arrives, and reporting, credibly and quickly, when a policy's street-level effect diverges from its intended design (see Topic 3.8 — Social Sector and Nonprofit Management).

Multilateral and international

Multilateral and international bodies rarely command binding domestic authority, taxation, or spending in the way a national government does, so their characteristic instruments are nodality and organization in Hood's terms — convening, setting standards, publishing comparative evidence, and coordinating funding that national governments then deploy through their own instrument mix. Because they operate across jurisdictions with very different administrative capacities, the implementation gap between an internationally adopted framework and its effect on the ground is often the widest of any level covered in this topic, and the agenda-setting and evaluation stages of the cycle typically matter more to their work than adoption or direct implementation.

Common failure modes

Mistaking the diagram for the process. A team designs its internal governance as though every decision will move cleanly through agenda-setting, formulation, adoption, implementation, and evaluation in strict order, and then treats every real decision that does not follow that order as an anomaly requiring correction rather than as normal policymaking. This produces governance processes that slow down the genuine, opportunistic windows a policy entrepreneur needs to act on, without actually making the eventual decision any better informed.

Instrument-first design. A policy problem arrives already dressed as a preferred instrument — 'we need a new law' or 'we need to spend more' — and the formulation stage becomes a search for justification rather than a genuine comparison of regulation, spending, taxation, and information against the objective. This costs the organization cheaper, less coercive, or more effective combinations that were never seriously considered because the instrument was chosen before the analysis began.

Designing for an idealized implementer. A policy is formulated around an assumption of full frontline compliance, ample staff time, and no competing local priorities, none of which reflects the street-level bureaucracy that will actually deliver it. The resulting implementation gap is then blamed on frontline staff for 'not following the policy properly', when the deeper cause was a design that never asked what implementation would actually require.

Evaluation as an endpoint rather than a feedback loop. An evaluation is commissioned, completed, and filed, with no standing mechanism connecting its findings back to the next round of agenda-setting or formulation. This is expensive twice over: once in the cost of the evaluation itself, and again in the cost of the next policy cycle rediscovering, from scratch, a lesson the evaluation had already established.

Incrementalism that has calcified into inertia. A policy area receives only marginal adjustments year after year because that is how the organization has always approached it, not because a genuine comparison against a more comprehensive redesign was ever made. The cost is invisible until a crisis or a policy window exposes how far behind a more deliberate approach the incremental path has fallen.

Ignoring the political window that made a decision possible. An organization treats an adopted policy as though it resulted purely from the technical merit of the proposal, forgetting the triggering event, entrepreneur, or political alignment that actually made adoption possible. When the next similar problem arises without the same alignment, the organization is surprised that an equally well-designed proposal fails to be adopted, having mistaken a lucky coupling of streams for a repeatable process.

Maturity model

Checklist

• Current stage of each live policy initiative (agenda-setting, formulation, adoption, implementation, evaluation) identified honestly.
• A shortlist of researched, ready policy options maintained independent of any current live problem.
• Instrument comparison (regulation, spending, taxation, information) documented before drafting begins.
• Distributional effect of each instrument in the proposed mix checked, not only the aggregate effect.
• Frontline delivery capacity and likely discretion assessed before a policy is finalized.
• Evaluation plan and success measures defined at formulation stage, not after adoption.
• Route by which evaluation findings and implementation feedback reach the next agenda-setting round confirmed to exist.
• Political or circumstantial window that enabled a recent adoption documented while it is still clear.
• Areas of long-running incremental adjustment reviewed periodically for whether a comprehensive redesign is now justified.
• Behavioural or information-based instruments checked against Topic 4.1 — Behavioural Public Policy — for ethical limits before deployment.
• Combined instrument package presented and evaluated as one coordinated policy, not as unrelated initiatives.

Key sources

• Charles E. Lindblom, 'The Science of "Muddling Through"', Public Administration Review, 19(2), 1959, pp. 79–88, for the foundational case for incrementalism as a defensible strategy under genuine uncertainty.
• John W. Kingdon, Agendas, Alternatives, and Public Policies, Pearson, for the multiple streams framework's original and fullest statement.
• Michael D. Cohen, James G. March, and Johan P. Olsen, 'A Garbage Can Model of Organizational Choice', Administrative Science Quarterly, 17(1), 1972, pp. 1–25, for the organizational-choice model that Kingdon's framework adapts to policymaking.
• Christopher Hood, The Tools of Government, Macmillan, 1983 (reissued by Palgrave Macmillan), for the nodality, authority, treasure, and organization taxonomy of policy instruments.
• The Organisation for Economic Co-operation and Development (OECD)'s Recommendation of the Council on Regulatory Policy and Governance, for a cross-country account of how regulatory instruments are designed and governed in practice.

References

1. Public policy — Wikipedia — https://en.wikipedia.org/wiki/Public_policy
2. Multiple streams framework — Wikipedia — https://en.wikipedia.org/wiki/Multiple_streams_framework
3. John W. Kingdon — Wikipedia — https://en.wikipedia.org/wiki/John_W._Kingdon
4. Incrementalism — Wikipedia — https://en.wikipedia.org/wiki/Incrementalism
5. Charles E. Lindblom — Wikipedia — https://en.wikipedia.org/wiki/Charles_E._Lindblom
6. Garbage can model — Wikipedia — https://en.wikipedia.org/wiki/Garbage_can_model
7. Street-level bureaucracy — Wikipedia — https://en.wikipedia.org/wiki/Street-level_bureaucracy
8. Christopher Hood — Wikipedia — https://en.wikipedia.org/wiki/Christopher_Hood
9. Regulation — Wikipedia — https://en.wikipedia.org/wiki/Regulation
10. Subsidy — Wikipedia — https://en.wikipedia.org/wiki/Subsidy
11. Sin tax — Wikipedia — https://en.wikipedia.org/wiki/Sin_tax
12. Nudge theory — Wikipedia — https://en.wikipedia.org/wiki/Nudge_theory
13. The Science of "Muddling Through" — Charles E. Lindblom, Public Administration Review — https://doi.org/10.2307/973677
14. Agendas, Alternatives, and Public Policies — John W. Kingdon, Pearson — https://www.pearson.com/en-gb/subject-catalog/p/agendas-alternatives-and-public-policies-update-edition-with-an-epilogue-on-health-care-pearson-new-international-edition/P200000004628/9781292053875
15. A Garbage Can Model of Organizational Choice — Cohen, March, and Olsen, Administrative Science Quarterly — https://doi.org/10.2307/2392088
16. The Tools of Government — Christopher Hood, Springer (Palgrave Macmillan) — https://link.springer.com/book/10.1007/978-1-349-17169-9
17. Recommendation of the Council on Regulatory Policy and Governance — OECD — https://www.oecd.org/en/publications/recommendation-of-the-council-on-regulatory-policy-and-governance_9789264209022-en.html