Chapter 33 — The Policy Process: How Problems Become (or Don't Become) Laws

"Policy is what governments choose to do or not to do." — Thomas Dye, Understanding Public Policy

"Solutions float around looking for problems to attach themselves to." — John Kingdon, Agendas, Alternatives, and Public Policies

33.1 The integrative chapter

You have now read about thirty-two chapters of this book. You have read about the founding (Part I), the institutions (Part II), the politics that flow through them (Part III), and the substantive policy domains where institutions and politics collide (Chapters 27–32). This chapter does something different. It does not introduce a new domain. It builds the analytical framework that ties all the prior chapters together.

The question is simple enough to state and complicated enough to fill the rest of the book: how does a problem become a law, and then operational reality? Or, just as often: why doesn't it?

That question matters because most of what your government does — and most of what it fails to do — is the cumulative result of policy decisions. The Affordable Care Act covers more than twenty million Americans because Congress passed a law in 2010, the Supreme Court mostly upheld it in 2012, the Department of Health and Human Services wrote thousands of pages of regulations, fifty governors made decisions about Medicaid expansion, and the federal exchange enrolled people each year for more than a decade. Each step of that chain is a policy decision. If any of them had broken, the policy would not exist as you experience it.

The opposite is also true. Comprehensive immigration reform does not exist as a policy because, on at least four occasions between 2006 and 2024, the chain broke. Each time, the streams partially aligned — there was a recognized problem, a proposed solution, and at least the appearance of political opportunity — and each time, somewhere along the chain, a critical link failed. To understand why is to understand the policy process.

This chapter offers a vocabulary for that understanding. The vocabulary is borrowed from political science: the stages model, the streams model (John Kingdon), the garbage-can model (Cohen, March, and Olsen), the veto-players framework (George Tsebelis), the implementation gap (Pressman and Wildavsky), and the policy feedback tradition (Suzanne Mettler, Paul Pierson). Each of these traditions captures something the others miss. None of them, on its own, explains the whole picture.

Three commitments shape what follows.

Process is power, again. This phrase has appeared throughout the book, especially in Chapters 7, 8, 16, and 25. It applies once more here. The procedural rules of policy-making — committee jurisdiction, filibuster cloture, reconciliation eligibility, regulatory comment periods, judicial standards of review, federalism's allocation of authority — are not neutral. They allocate advantage. Whoever wrote the rules, or whoever inherited them, holds power that does not always show up in vote totals.

Policy success and policy failure are both real categories. It is fashionable, on both ends of the political spectrum, to narrate American government as comprehensive failure. That is not what the data shows. The Clean Air Act of 1970, as amended in 1977 and 1990, has substantially reduced ambient sulfur dioxide and nitrogen oxide. The Children's Health Insurance Program, enacted in 1997, covers more than seven million children who would otherwise be uninsured. Medicare Part D, enacted in 2003, gives nearly fifty million seniors prescription-drug coverage. The Bipartisan Infrastructure Law of 2021 is funding tens of thousands of infrastructure projects across all fifty states. These programs exist. They were enacted by ordinary congressional process. They are doing what their drafters said they would do. A theory of the policy process that cannot explain success is incomplete.

The veto-points framework is structural, not partisan. The American system has more institutional veto players than any other advanced democracy. That is a fact about the constitutional architecture. It is not a partisan complaint. It produces deliberation and stability when those are needed; it produces gridlock and slow response when those are costly. Whether the trade-off is worth it depends on what you think democracy is for, and on which problems you most want government to solve.

By the end of the chapter you should be able to:

• explain the classic stages model and its limits;
• apply Kingdon's streams model to a real policy moment;
• identify the major veto points in the American system and explain how they bias outcomes toward the status quo;
• distinguish the major adoption pathways (statutory, regulatory, judicial, state-level, treaty);
• identify why some major reforms succeed and others fail repeatedly; and
• evaluate proposed process reforms on their own terms, without prejudging.

33.2 The classic stages model

The stages model is the oldest analytical framework for understanding policy. It is the framework that shows up in textbooks from the 1970s, in instructor handouts, and in lobbying-firm onboarding decks. It pictures policy moving through six stages, in order:

1. Problem definition / agenda-setting. A condition becomes a problem when actors collectively decide it is one — when they argue it must be solved.
2. Formulation. Solutions are proposed. Bills are drafted. Regulations are sketched. Court briefs are written.
3. Adoption. Some authorized body — Congress, the president, an agency, a court, a state legislature — chooses one of the proposed solutions.
4. Implementation. The chosen solution is translated into operational reality. Regulations are written. Forms are designed. Field offices process applications. Money is disbursed.
5. Evaluation. Did it work? Empirical evaluation by analysts; political evaluation by voters and interest groups; judicial evaluation by courts.
6. Feedback. The results of the policy reshape the politics that produced it. Successful programs create constituencies that defend them; failed programs invite calls for reform or repeal.

The stages model is useful. It gives you a checklist. It tells you the components without which no policy actually exists.

But the stages model is also incomplete, and the way it is incomplete matters. Three problems.

Problem one: the stages do not happen in order. Implementation often begins before adoption, as agencies anticipate likely outcomes and begin planning. Evaluation often begins before implementation is complete, as opponents prepare critiques. Feedback happens continuously, not at the end. Formulation can occur years or decades before agenda-setting — the Affordable Care Act's individual mandate, for example, was formulated by the Heritage Foundation in 1989 and implemented by Massachusetts in 2006, well before it became part of a national agenda in 2009.

Problem two: many policy stages run simultaneously, on different schedules. A policy may be in implementation (ACA exchanges launching in 2014), evaluation (CBO scoring updates in 2015, 2016, 2017), feedback (the constituency formed by the exchanges making repeal politically costly in 2017), and adoption (Inflation Reduction Act premium-subsidy enhancements in 2022) all at once. Treating these as sequential underestimates how dynamic the process is.

Problem three: solutions chase problems as much as problems generate solutions. The "garbage-can" model (Michael Cohen, James March, and Johan Olsen, 1972) describes organizations as "garbage cans" into which actors throw problems, solutions, and decision opportunities. Coupling among the three is loose, intermittent, and often opportunistic. A solution that has been sitting on the shelf for a decade — a tax credit for clean energy, for instance — gets attached to a problem (climate change, energy security, manufacturing competitiveness) when a decision opportunity (a reconciliation bill) opens up. The policy makes sense not because the problem produced the solution but because the available solution found a problem to attach to.

A more accurate picture of the policy process recognizes the stages but does not enforce them as a sequence. The stages are components, not steps.

33.3 Kingdon's streams: when does a window open?

The most influential single framework in agenda-setting is John Kingdon's Agendas, Alternatives, and Public Policies (1984; revised 1995, 2003). Kingdon studied federal agenda-setting in health and transportation policy and proposed a model with three streams flowing in parallel:

• The problem stream. Conditions become recognized as problems through several channels: changes in indicators (homicide rates rising, child-poverty rates falling), focusing events (a school shooting, a financial crisis), and feedback from existing programs (Medicare costs growing faster than projected).
• The policy stream. Specialists — academics, think tanks, agency staff, congressional staff, interest-group analysts — generate, refine, and circulate policy ideas. Most ideas die. A few survive in what Kingdon called the "policy primeval soup," waiting for an opportunity.
• The political stream. Public mood, electoral results, partisan composition of Congress, presidential agendas, interest-group pressure. The political stream defines what is currently politically possible.

A policy window opens when the three streams couple — when a recognized problem, an available policy solution, and a political moment line up. The window may open because of a focusing event, an electoral change, or an indicator update. It does not stay open long. Kingdon argues that successful policy entrepreneurs are those who keep their preferred solution on the shelf, ready to be coupled to whatever problem and political moment arrives.

Three implications follow.

First, the same problem can wait years for a window. The 1965 Medicare and Medicaid laws were the result of policy ideas that had been in circulation since the 1940s. The 1986 tax-reform act drew on academic critiques of the existing tax code that dated to the 1960s and 1970s. The Affordable Care Act drew on a Heritage Foundation framework from 1989 that became a Massachusetts law in 2006 that became a federal proposal in 2009. The window opens, the solution is at hand, and adoption happens. Without the prior decades of policy work, the window opens onto an empty room.

Second, the same solution can wait years for a problem. Cap-and-trade emissions trading was developed by economists in the 1960s and 1970s. It got attached to the acid rain problem in the late 1980s, became the centerpiece of the 1990 Clean Air Act amendments, and substantially solved the problem. It was later attached to the climate problem, formed the basis of the EU emissions-trading scheme, the Regional Greenhouse Gas Initiative in the Northeast, and California's economy-wide system. The solution shaped how the problem got framed.

Third, focusing events matter, and they cut both directions. The 9/11 attacks opened windows for the PATRIOT Act, the Department of Homeland Security, the Authorizations for Use of Military Force, and ultimately the wars in Afghanistan and Iraq. The 2008 financial crisis opened windows for TARP, Dodd-Frank, the auto-industry bailout, and (eventually) the ACA. The COVID-19 pandemic opened windows for the CARES Act, the American Rescue Plan, expanded Child Tax Credit, paid leave provisions, and (more durably) the Inflation Reduction Act and CHIPS Act. The January 6 attack opened a window for the Electoral Count Reform Act of 2022. The murder of George Floyd opened windows for state and local police-reform measures, though the federal George Floyd Justice in Policing Act stalled in the Senate. The Russian invasion of Ukraine in 2022 opened a window for an aid package that drew on Cold War alliance-management instincts. The Israel–Hamas war beginning in October 2023 opened windows for both supplemental aid and unprecedented intra-Democratic-Party conflict.

Focusing events do not guarantee policy. They open windows. What flows through depends on whether there is a usable solution and a workable political coalition.

33.4 Where do solutions come from? The formulation stage

Many readers, when they think of policy-making, think of Congress. But Congress mostly adopts policy. It rarely formulates it. The drafting work is done elsewhere.

Think tanks. A major shift in American policy formulation occurred in the 1970s, when conservative donors built a deliberate institutional infrastructure: the American Enterprise Institute (founded 1938 but expanded in the 1970s), the Heritage Foundation (1973), the Cato Institute (1977), and later the Mercatus Center (1980), the State Policy Network (1992), and a dense network of state-level affiliates. The infrastructure was designed to do three things: produce policy ideas, train staff who would carry the ideas into government, and pre-position those ideas for adoption when political windows opened. Heritage's Mandate for Leadership (1981, with successor volumes through 2024) was an explicit playbook for incoming administrations. The progressive equivalent — the Center for American Progress (founded 2003), Demos (2000), the Roosevelt Institute (with deeper roots but reorganized in the 2000s) — was built later and is less well-resourced. Other think tanks (Brookings, founded 1916; the Urban Institute, 1968; the National Bureau of Economic Research, 1920) cultivate a more academic, less explicitly partisan posture, though their staffs and outputs are not without political valence. The Niskanen Center (2014) tries to occupy a center-libertarian space; the American Compass (2020) and the Manhattan Institute (1978) work on a populist-conservative agenda; the Roosevelt Institute and Economic Policy Institute work on labor-progressive economics.

The point is not which think tank you trust. The point is that policy ideas have to come from somewhere, and most of them come from these institutions before they reach Congress.

Agency expertise. The career civil service (Chapter 11) is a major source of formulation. Agency staff with subject-matter expertise draft proposed regulations, evaluate options, and brief political appointees. The EPA's career staff drafted what became the 2009 endangerment finding for greenhouse gases. CMS career staff designed the operational architecture of the ACA exchanges. The Federal Reserve's research staff produces the analytical framework that shapes monetary policy. When a new administration arrives with a policy agenda but few specifics, agency career staff fills in the specifics.

Congressional staff. Each member of Congress has a personal-office staff (typically 15–20 people for a House member, larger for a senator). Each committee has a majority and minority staff (anywhere from a dozen to over a hundred). The Senate Legislative Counsel and House Legislative Counsel offices employ career attorneys who draft bill text. When a bill emerges from committee, it has been touched, in nearly all cases, by these career staff far more than by the members whose names are on it.

Academic research. Long-term influence, often indirect. Economists' work on optimal tax theory, criminologists' work on incarceration, public-health researchers' work on smoking and obesity, epidemiologists' work on infectious-disease modeling — these influence policy through diffusion into think tanks, agencies, and congressional staff over decades. Direct academic influence on a specific bill is rarer but does occur (Jonathan Gruber's models of the ACA's individual mandate, for instance, or Raj Chetty's work on tax credits).

Interest groups. Chapter 24 covered interest groups in detail. For formulation, the relevant point is that they often arrive with bill text in hand. The American Legislative Exchange Council (ALEC) provides model legislation to conservative state legislators; the State Innovation Exchange (SiX) does the same for progressives. The Chamber of Commerce drafts business-favored regulatory language; environmental groups do the same for environmental rules. Drafting is itself a policy-influence mechanism, because whoever writes the first draft shapes what is debated.

33.5 Adoption pathways: more than one route to law

Most introductory accounts depict policy adoption as a bill-becoming-a-law story. That is one path. It is not the only one, and increasingly it is not the dominant one for major policy. Six pathways deserve attention.

33.5.1 The classic congressional bill

Bill introduced; referred to committee; committee markup; committee report; floor consideration in chamber of origin; passage; sent to the other chamber; same process; differences resolved in conference or by amendment exchange; final passage; presentment; signed or vetoed. This is the path Chapter 8 described in detail. It is also surprisingly rare for major policy. The 1990 Americans with Disabilities Act, the 1996 welfare reform, the 1997 Children's Health Insurance Program, the 2001 No Child Left Behind Act, the 2003 Medicare Modernization Act, the 2008 Higher Education Opportunity Act, the 2014 Veterans Choice Act, the 2018 First Step Act, the 2018 Farm Bill, the 2020 CARES Act, the 2021 Bipartisan Infrastructure Law, the 2022 CHIPS Act, the 2022 Electoral Count Reform Act, and the 2024 Lankford-Sinema border bill (which failed) followed something close to the classic path. Most of these required at least sixty Senate votes — the cloture threshold. Notice how few there are.

33.5.2 The reconciliation bill

The Congressional Budget Act of 1974 created a budget-reconciliation process for fiscal legislation. Reconciliation is non-filibusterable. A reconciliation bill needs only fifty-one votes (or fifty plus the vice president) to pass the Senate. It is constrained by the Byrd Rule, which excludes provisions that do not have a budgetary effect, that increase deficits beyond the budget window, or that touch Social Security. But within those constraints, reconciliation is the dominant pathway for major fiscal legislation since 2001.

The 2001 Bush tax cuts, the 2003 Bush tax cuts, parts of the 2010 Affordable Care Act (the reconciliation amendments to the Senate-passed bill), the 2017 Tax Cuts and Jobs Act, the 2021 American Rescue Plan, and the 2022 Inflation Reduction Act all passed via reconciliation. None could have cleared a sixty-vote cloture threshold under the partisan composition of the Senate at the time. Both parties have used reconciliation. Both parties' major legislative wins this century have substantially come through it.

The implication is structural. The filibuster has shifted major legislation toward fiscal frameworks because fiscal frameworks are non-filibusterable. This shapes what can be policy. Climate policy in 2022 became, in substantial part, tax-credit policy because tax credits fit reconciliation. Social policy in 2010 became insurance-market regulation tied to subsidies because that was reconciliation-eligible. The procedural rule shaped the substantive form of the policy.

33.5.3 Executive order and regulation

When Congress is gridlocked, presidents act. Chapters 9 and 11 covered the relevant authorities — executive orders, presidential memoranda, agency regulations, guidance documents, and enforcement priorities. The Obama administration used executive action for DACA (2012), the Clean Power Plan (2014), and overtime-rule changes (2016). The Trump administration used executive action for the travel ban (2017), regulatory rollbacks across multiple agencies, and immigration enforcement priorities (2017–2021 and again from 2025). The Biden administration used executive action for student-loan forgiveness (2022, struck down in Biden v. Nebraska), climate regulations, and asylum reforms.

The strength of executive action: it is fast. The weakness: it is reversible by the next administration, it is constrained by statutory authority, and since 2022 it is constrained by the major-questions doctrine (West Virginia v. EPA, 2022) and since 2024 by the end of Chevron deference (Loper Bright Enterprises v. Raimondo, 2024). A policy that exists only by executive action exists provisionally. The Clean Power Plan never took effect; DACA has survived court challenges but is constantly contested; student-loan forgiveness was struck down. Permanent policy still requires statute.

33.5.4 Judicial doctrine

Some of the most consequential policy changes in American history were not enacted by Congress but announced by the Supreme Court interpreting the Constitution. Brown v. Board of Education (1954) ended de jure school segregation. Roe v. Wade (1973) recognized a constitutional right to abortion. Citizens United v. FEC (2010) reshaped campaign finance. Obergefell v. Hodges (2015) recognized a constitutional right to same-sex marriage. Dobbs v. Jackson Women's Health (2022) reversed Roe. Students for Fair Admissions v. Harvard (2023) ended race-conscious college admissions in their then-current form.

Each of these is policy change, in the sense that ordinary Americans' lives are different because of them. None passed through Congress. Each generated, on enactment, a political reaction that reshaped subsequent legislative agendas. Policy by judicial doctrine has the strength of finality (only constitutional amendment or judicial reversal undoes it) and the weakness of legitimacy contestation (was the court the right body to make this decision?).

33.5.5 State-level innovation

The "laboratories of democracy" phrase comes from Justice Brandeis's dissent in New State Ice Co. v. Liebmann (1932). The substantive idea is older. States can adopt policies that the federal government has not, and the federal government can later learn from, scale up, or harmonize them.

Examples cluster.

Massachusetts healthcare (2006) → Affordable Care Act (2010). The individual mandate, exchange-based purchasing, subsidies up to a defined income level, and insurance-market regulation: all in Massachusetts first, all federalized four years later.

California vehicle-emissions standards → federal harmonization (and contestation). California's preexisting authority under the Clean Air Act (Section 209's waiver provision) to set stricter vehicle-emissions standards has, since the 1970s, effectively driven national auto manufacturing. The Trump administration revoked the waiver in 2019; the Biden administration restored it in 2022; the second Trump administration is contesting it again.

Colorado and Washington recreational marijuana (2012) → state spread (24 states by 2026) → federal pressure for descheduling. The federal government has not legalized marijuana. The states have driven national policy by sequential adoption.

State-level paid family leave → federal proposals. California (2002), New Jersey (2008), Rhode Island (2013), New York (2016), Washington (2017), Massachusetts (2018), and a growing list of others. None of this is yet federal law, but the policy infrastructure is being built bottom-up.

State-level minimum-wage increases → wage-floor diffusion. The federal minimum wage has been $7.25 since 2009. Thirty states and many more cities have set higher state and local floors. Most American workers earning above $7.25 now do so because of state policy.

Heartbeat bills, abortion bans, and abortion-protection laws → post-Dobbs state policy patchwork. When Dobbs returned abortion to the states (2022), states moved in opposite directions, producing a patchwork of state policy that is now functionally the policy environment for most Americans.

33.5.6 International and treaty pathways

Article II of the Constitution requires Senate ratification of treaties by two-thirds. That threshold has not been routinely met for major agreements since the late twentieth century. The Comprehensive Test Ban Treaty was rejected by the Senate in 1999. The Convention on the Rights of Persons with Disabilities was rejected in 2012. Major climate agreements (Kyoto, Paris) have been entered as executive agreements rather than as treaties.

Executive agreements — which do not require Senate ratification — have largely replaced treaties as the operational form of international commitment. They are correspondingly easier to enter and easier to exit, with all the consequences for credibility and durability that Chapter 32 discussed.

33.6 Implementation: the link that often breaks

Jeffrey Pressman and Aaron Wildavsky's Implementation (1973, with a famously long subtitle) studied a 1966 Economic Development Administration program intended to create jobs for unemployed minorities in Oakland, California. Federal money was appropriated. Ground was broken. Years later, almost nothing had been built and few jobs had been created. Pressman and Wildavsky asked why.

Their answer, briefly: every non-trivial policy depends on a chain of institutional acts. Congress appropriates. The agency promulgates regulations. State or local government applies. Sub-agencies coordinate. Contractors bid. Inspectors approve. Each link in the chain has decision points. At each decision point, some probability the link breaks. If the chain has thirty links and each link has a 90 percent probability of holding, the overall success probability is 0.9^30, or about 4 percent.

The implication is that implementation gaps are not anomalous; they are structurally expected. Successful implementation requires either fewer links or a higher probability of each link holding. Fewer links means more centralized policy (federal direct-administration rather than federal-state-local handoffs). Higher per-link probability means more capacity, clearer authority, and less veto-point density.

Two implementation cases are worth dwelling on.

The Affordable Care Act, 2010–2014. The ACA's implementation was extraordinarily ambitious. The federal government had to design and operate insurance exchanges in states that declined to build their own; states had to decide whether to expand Medicaid; HHS had to write thousands of pages of regulations; navigators had to be trained; the IRS had to design systems for premium tax credits. The HealthCare.gov launch in October 2013 was a famous near-disaster: the federal exchange website crashed under load and barely functioned for the first six weeks. The cause, retrospectively analyzed, was a failure of project management — too many subcontractors, no integrated systems testing, no single accountable executive — combined with the political constraint that the launch could not be delayed without enormous cost. The site was eventually fixed, enrollment recovered, and by 2015 the exchange was functioning. The implementation gap was substantial but not, in the end, fatal.

The Inflation Reduction Act, 2022–present. The IRA's tax-credit-driven approach to climate policy is, by design, less implementation-heavy than a regulatory approach would have been. Tax credits are administered through the existing tax system. Direct-pay provisions allow non-tax-paying entities (including municipalities and rural cooperatives) to receive payment for clean-energy investments. Even so, the IRA's electric-vehicle credit provisions, manufacturing-domestic-content requirements, and grant programs have generated thousands of pages of Treasury and IRS guidance. The implementation is going better than the ACA's launch — faster, with fewer crises — but it is still an enormous administrative undertaking, and the second Trump administration has begun reversing components of it through executive order and new guidance.

State Medicaid expansion as an implementation case. Under the original ACA, Medicaid expansion to all adults below 138 percent of the federal poverty line was mandatory. The Supreme Court in NFIB v. Sebelius (2012) made expansion optional for states. Forty states and the District of Columbia have, as of 2026, expanded; ten have not. The empirical effect is large: people in non-expansion states are less likely to be insured, to receive preventive care, and to receive timely treatment for serious conditions. The implementation gap, in this case, was created by the Court's ruling and is sustained by state-level political decisions.

The Pressman-Wildavsky lesson generalizes: a policy is not what is enacted. It is what is implemented. And implementation depends on a chain of acts that can break at many points.

33.7 The veto-points framework

George Tsebelis's Veto Players (2002) offers a comparative framework for thinking about why some political systems produce more policy change than others. A veto player is an actor whose agreement is required for a policy change. The more veto players, and the more ideologically distant they are from each other, the more difficult policy change is. The smaller the winset — the set of policy points that all veto players prefer to the status quo — the more biased the system is toward stability.

The American system has many veto players, by design.

• The House of Representatives must pass a bill.
• The Senate must pass a bill — and, on most legislation, must clear a sixty-vote cloture threshold.
• The President must sign (or, with two-thirds of both chambers, be overridden).
• The Supreme Court, on constitutional questions, can strike down a statute.
• States, on policies that depend on state implementation (Medicaid, education, much of environmental regulation), can refuse to play.
• Within Congress, committee chairs historically held strong agenda-setting power; today party leadership has largely centralized that, but committees still matter.
• Within the executive, agency rulemaking can be slow-walked or fast-tracked depending on political appointees, career staff, and judicial review.

The comparative consequence is that American policy change is structurally harder than policy change in, for example, the United Kingdom (where a parliamentary majority can do nearly anything) or Germany (where the federal structure adds veto points but party discipline is much higher).

This is not a partisan claim. It applies whether the party in power is Democratic or Republican. The 2009–2010 Democratic Senate had sixty senators and passed the ACA; it lost the sixtieth seat (Scott Brown's election after Edward Kennedy's death) and could finish the law only via reconciliation. The 2017 Republican Senate had only fifty-two seats and could not pass an ACA repeal even via reconciliation, because three senators (McCain, Collins, Murkowski) defected. The 2021–2023 Democratic Senate had only fifty seats and could pass only what reconciliation allowed.

Veto players are also why narrow majorities cannot govern in the American system the way they can in parliamentary systems. The system was, in the framers' design, intended to require broader coalitions before policy could change. That intention has produced the system we have.

The normative debate is whether the design is well-suited to current conditions. Defenders of the structure argue that deliberation and stability are precisely what the framers wanted, that the current frustration with gridlock reflects unrealistic expectations of what national politics can produce, and that loosening the constraints would enable policy whiplash. Critics argue that the structure was calibrated for an eighteenth-century population of three million in thirteen states, that contemporary problems (climate change, pandemic preparedness, immigration) move faster than the system can respond, and that the result is a permanent mismatch between policy needs and policy capacity. Both positions are serious. Neither is silly.

33.7.1 Veto players in operation: a 2017 case

A concrete example clarifies how veto players operate together. The 2017 attempt to repeal and replace the Affordable Care Act provides the clearest worked case of recent years.

Republicans entered 2017 with unified control: a Republican president (Trump), a Republican House majority (241 seats), and fifty-two Republican senators. They had committed for seven years to repealing the ACA. The political stream's alignment seemed total. The policy stream had ready alternatives (the American Health Care Act in the House; the Better Care Reconciliation Act in the Senate; the "skinny repeal" amendment as a fallback). The problem stream had been characterized for years as ACA-driven (premium increases, narrow networks, tax penalty for the uninsured). All three streams looked coupled.

What broke? The Senate's veto density. Reconciliation rules limited what the bill could include. The Byrd Rule's "extraneous matter" provisions excluded large parts of the proposed reform. Three Republican senators — John McCain, Susan Collins, and Lisa Murkowski — defected for different reasons (process objections from McCain; coverage objections from Collins; Medicaid objections from Murkowski reflecting Alaska-specific concerns). With fifty-two seats, three defections were enough to fail. The skinny-repeal amendment failed 49–51 in the early-morning hours of July 28, 2017.

The lesson is structural. Even unified party control does not guarantee policy change in the American system. Three veto players (three senators) can block what fifty-five legislators (the entire House majority plus a presidential signature) want. Whether that is appropriate depends on whether you are the fifty-five or the three. Both perspectives have constitutional grounding. Neither side is entitled to dismiss the other.

33.8 Worked examples: when streams couple

The clearest way to understand the streams model is to apply it to recent cases.

33.8.1 The Affordable Care Act, 2008–2010

Problem stream. Forty-six million uninsured Americans (Census, 2008). Health-care cost growth outpacing wage growth. Employer-based insurance fraying. Pre-existing-condition exclusions producing horror stories that became fixtures in political coverage.

Policy stream. A solution architecture had been on the shelf for two decades. The Heritage Foundation had developed an individual-mandate framework in 1989. Mitt Romney's Massachusetts had implemented it in 2006. Academic economists (notably Jonathan Gruber at MIT) had modeled scaled-up versions. The Obama campaign and the Senate Finance Committee had refined a draft.

Political stream. Barack Obama's 2008 election with a mandate that included healthcare. Sixty Democratic senators (briefly). A speaker (Pelosi) committed to passing major legislation. A unified party leadership.

The streams coupled. The window opened. The Senate passed the bill on Christmas Eve 2009. The window then partly closed (Massachusetts elected Scott Brown, ending the sixty-vote majority). The House passed the Senate bill in March 2010 unchanged, and reconciliation was used for amendments. The window had stayed open just long enough.

33.8.2 The Tax Cuts and Jobs Act, 2017

Problem stream. Frustration with the existing tax code's complexity; perceived global tax-rate competition; corporate tax-rate gaps with peer nations. (Note: the empirical evidence on whether U.S. corporate effective rates were globally uncompetitive was contested. The political perception of a problem was widespread.)

Policy stream. Rate-cut-and-base-broaden frameworks had been on the shelf since the 1986 reform. Conservative think tanks had refined them throughout the 2000s and 2010s. The Camp draft (2014, House Ways and Means Chair) was ready and in detailed form.

Political stream. Republican unified government with fifty-two Senate seats, a willing president, and a unified party position on tax cuts.

The streams coupled. Reconciliation was the vehicle (the corporate-rate cut was the centerpiece, and reconciliation could pass it). The bill passed in December 2017. The Byrd Rule required that the individual-side cuts sunset after 2025 to avoid post-window deficit increases — which is why much of the 2017 individual tax code is currently scheduled to expire on December 31, 2025, generating the agenda-setting moment for the 2025 tax debate.

33.8.3 The Inflation Reduction Act, 2022

Problem stream. Climate change with worsening indicators (record temperatures in 2021–2022, increasing climate-related insurance losses). Inflation in 2022, providing the bill's name. Manufacturing competitiveness, especially with China.

Policy stream. Tax-credit-driven climate policy had been developed and refined over multiple decades. The "Energy Innovation and Carbon Dividend Act" framework had circulated. The "Clean Energy for America" framework had circulated. The American Jobs Plan (initial Biden proposal, 2021) had laid groundwork. The Build Back Better bill (Fall 2021) had developed details.

Political stream. Democratic unified government with a fifty-fifty Senate, requiring all fifty Democrats. Joe Manchin's months of negotiation with Senator Schumer culminated in the surprise announcement of a deal in late July 2022. Senator Sinema joined after carve-outs. The bill passed via reconciliation in August 2022.

The window in this case stayed open through extraordinary persistence and was nearly closed at multiple points. Without Manchin's last-minute reversal, the policy stream would have produced no policy at all in 2022 — and a Republican House majority in 2023 would have closed the window for the rest of the term.

33.8.4 The CHIPS and Science Act, 2022

Problem stream. Semiconductor manufacturing concentrated in Taiwan and South Korea, creating supply-chain vulnerability dramatized by COVID-era chip shortages. Strategic competition with China, including export controls.

Policy stream. Industrial-policy frameworks for semiconductors had been developed by think tanks (CSIS, CNAS) and by academic economists. The "Endless Frontier Act" had circulated since 2020.

Political stream. Bipartisan consensus that something needed to be done about chip manufacturing. Sufficient Republican support to clear cloture (a rarity in this period), making it possible to pass via regular order rather than reconciliation.

The streams coupled. The bill passed by 64–33 in the Senate and 243–187 in the House, signed in August 2022. Implementation began immediately, with the first major awards (Intel, TSMC, Samsung) in 2024. As of 2026, the law's outcomes are still being measured.

33.9 Why some policy succeeds

If the policy process is as constrained as Section 33.7 suggested, why does any policy ever pass? Eric Patashnik's Reforms at Risk (2008) and his subsequent work studied this question. A policy passes when:

1. A focusing event aligns the political stream with a specific problem. The policy-stream solution must be ready, but the political moment is the binding constraint.
2. The policy can be designed to fit available procedural channels. If sixty Senate votes are not available, the policy must be reconcilable. If reconciliation is unavailable, the policy must be regulatory.
3. The coalition contains members who hold otherwise irreconcilable preferences. This is often achieved by side payments, geographic targeting (highway projects, military bases), or bundling that creates personal stakes for marginal members.
4. There is sufficient elite consensus that the existing framework is failing. This is the precondition for reform. When the existing framework still serves enough actors, no amount of agenda pressure produces change.

Worked examples of policy success.

The Clean Air Act sulfur-dioxide trading program (1990 amendments). The cap-and-trade program for SO₂ emissions reduced acid rain by approximately 88 percent below 1980 levels by 2010, at a cost (according to retrospective EPA studies) far below ex ante projections. Why? The problem was tractable (clear pollutant, well-understood chemistry). The solution was tested (economists had argued for cap-and-trade for decades). The political stream aligned (President Bush sought a major environmental accomplishment, and the framework had bipartisan support). The implementation was straightforward (emission monitoring is technically feasible, allowance trading happens through markets).

The Children's Health Insurance Program (1997). CHIP expanded health coverage to children in low-income families above Medicaid thresholds. Why did it pass when the broader 1993–1994 Clinton health-reform effort had failed? Smaller scope. Bipartisan sponsors (Hatch, Kennedy). Sub-Medicare-and-Medicaid budget impact, fitting reconciliation-style rules. State-level flexibility that reduced federalism objections. The window was narrower; the policy fit through it.

Medicare Part D (2003). A Republican-led expansion of Medicare to cover prescription drugs. Why did Republicans expand a New Deal social program? Combined motives: senior voters' interest in drug coverage, partisan competition with Democrats on a popular-coverage issue, and a structure (private-insurer administration) that fit conservative preferences for market-based delivery. The bill was large, contentious, and required arm-twisting on the House floor (the famous three-hour vote). It passed.

The Bipartisan Infrastructure Law (2021). A $1.2 trillion infrastructure package, passed by 69–30 in the Senate and 228–206 in the House. Why did this pass with bipartisan support when so much else does not? Geographically distributed benefits (every state and most districts get visible projects). Long-running policy stream (infrastructure proposals had been refined for years). Political stream alignment (the Biden administration wanted a bipartisan win; some Senate Republicans wanted to demonstrate they could legislate). The bill passed regular order. As of 2026, projects are funded across all fifty states.

33.10 Why some policy fails

The complement is just as instructive. Some major reforms fail repeatedly, and the failures are illuminating.

Comprehensive immigration reform. In 2006, in 2007, in 2013, and in 2024, comprehensive reform packages — combining border-security investment, legalization of a long-resident unauthorized population, employment-verification, and visa-system modernization — got close. None passed. The 2007 effort cleared key Senate procedural votes before a coalition of conservative Republicans and labor-aligned Democrats voted it down. The 2013 Gang of Eight bill passed the Senate 68–32 but was never brought up in the House. The 2024 Lankford-Sinema-Murphy compromise — perhaps the most concession-heavy package on enforcement in recent memory — was killed on the Republican side after presidential-candidate Trump opposed it. The pattern suggests a structural problem: as long as the issue remains politically valuable as a divider, parties have an electoral incentive against passage. Case Study 02 develops this in detail.

Comprehensive entitlement reform. Multiple commissions (Simpson-Bowles 2010, the Domenici-Rivlin task force 2010, the Gang of Six 2011) produced detailed proposals. None became law. The structural reason is that meaningful reform requires either tax increases (anathema to one party's median voter), benefit cuts (anathema to the other party's median voter), or both. Without a budget crisis large enough to break the political stream's status-quo bias, nothing passes.

Major climate legislation pre-2022. Cap-and-trade for greenhouse gases passed the House in 2009 (Waxman-Markey) but died in the Senate. The politics of climate had become more partisan than the politics of acid rain in 1990 had been. The 2010 election then produced a Republican House. The window closed. It would not reopen for federal climate policy until the 2022 IRA, and by then the policy had been redesigned (tax credits, not cap-and-trade) to fit reconciliation.

Healthcare reform pre-ACA (1993–1994). The Clinton Health Security Act, drafted by a task force led by Hillary Clinton, never received a floor vote. It died of a combination of forces: its complexity was used against it (the "Harry and Louise" advertising); its administrative architecture was attacked as bureaucratic; the small-business lobby mobilized; and the Senate Democratic leadership could not hold sixty votes. The lesson was internalized in 2009: the next major effort would be sold differently and would use existing (private-insurance-based) infrastructure rather than build new bureaucracy.

The voting-rights legislation of 2021–2022. The For the People Act and the John Lewis Voting Rights Advancement Act passed the House but could not clear cloture in the Senate. The filibuster was the binding constraint. Senate Democrats could not assemble a majority for filibuster reform on this issue (Manchin and Sinema declined).

The pattern across failures: either the political stream collapses (immigration repeatedly), the procedural channel does not fit (cap-and-trade and reconciliation), or the policy itself proves too complex for the political coalition to hold (Clinton healthcare).

33.10.1 Policy feedback: programs make politics

A development that the stages model misses entirely is what political scientists call policy feedback. Suzanne Mettler's The Submerged State (2011) and Paul Pierson's earlier work (most influentially "When Effect Becomes Cause," 1993) developed the theoretical framework: enacted policies do not merely respond to politics; they create politics. They generate constituencies, shape interests, structure expectations, and reorganize who shows up to defend or attack them.

The clearest cases are entitlement programs. Social Security, enacted in 1935, immediately created a constituency of beneficiaries and prospective beneficiaries who organized to defend it. The American Association of Retired Persons (AARP) is in part a political consequence of Social Security's existence: there were not millions of organized seniors mobilizing for retirement security before 1935 because retirement security as a concept was not embedded in American policy. Once it was, the constituency followed. Subsequent reform efforts have repeatedly run into AARP's organized opposition. President George W. Bush's 2005 Social Security partial-privatization proposal failed in substantial part because AARP mobilized against it.

Medicare, enacted in 1965, did the same thing in a different form. Hospital systems, physician groups, and private Medicare-Advantage insurers all developed economic stakes in Medicare's continuation. Repeal became, after a decade or two, almost unimaginable. The program's existence reorganized the politics around it.

The Affordable Care Act provides a more recent and more contested example. By 2017, the exchange constituency (those receiving premium subsidies and the insurers selling on exchanges), the Medicaid-expansion constituency (states that had expanded and their now-insured residents), and the pre-existing-condition-protection constituency (essentially all Americans, in different probabilities, who might at some future point need insurance after a diagnosis) had become organized political forces. The repeal effort failed in part because of policy feedback: the ACA had created constituencies that the 2009 ACA-design debates had not yet produced.

Policy feedback runs the other way too. Failed implementations create constituencies for reform. Welfare-reform politics in the 1990s drew on a constituency of recipients, advocates, and conservative critics shaped by AFDC's earlier failures. The 1996 Personal Responsibility and Work Opportunity Reconciliation Act would not have been politically possible without three decades of accumulated dissatisfaction. The constituency for change is itself a feedback effect.

The implication for analyzing the policy process: the relationship between politics and policy is bi-directional. Politics shapes policy at the adoption stage. Policy then shapes politics through implementation and feedback. The next round of adoption operates in a politics that the prior round of policy partly created.

33.11 The "permanent campaign" implication

Frances Lee's Insecure Majorities (2016) studied a structural change in American politics: as control of Congress has become more genuinely competitive (no longer the durable Democratic majorities of 1955–1994), members have shifted toward what Hugh Heclo called the "permanent campaign." Both parties expect the next election to potentially flip control. The dominant question for any roll-call vote becomes: does this help us in the next election?

The implication for policy-making is corrosive. Bipartisan deals require both parties to give up something, but a deal that hands a partisan accomplishment to the other side is electorally costly. Compromise, which requires both sides to share credit, becomes electorally suspect. The space for substantive policy negotiation between cycles narrows.

This is not a normative claim about which party is responsible. Lee's data show that both parties exhibit the pattern. It is a structural observation: when majorities are insecure, the policy process becomes more campaign-shaped and less negotiation-shaped.

The reform implication, which Frances Lee and Yuval Levin have separately developed, is that some procedural changes might restore deliberative space — committee markup that actually amends bills, regular-order budgeting, reduced top-down party leadership control, longer debate windows. Whether these reforms would actually work, or whether the underlying electoral incentives would defeat them, is contested.

33.12 The administrative-state question

Chapter 11 covered the administrative state in detail. Two recent developments shape the policy-process picture this chapter is drawing.

The major-questions doctrine. In West Virginia v. EPA (2022), the Supreme Court held that on questions of "vast economic and political significance," an agency must point to clear congressional authorization for its action. The doctrine's contours are still being worked out. Its initial impact has been to constrain agency rulemaking on issues like climate (the Clean Power Plan was effectively foreclosed), student loans (the Biden forgiveness program was struck down in Biden v. Nebraska, 2023), and regulatory ambition more generally.

The end of Chevron deference. Chevron U.S.A. v. Natural Resources Defense Council (1984) held that courts should defer to reasonable agency interpretations of ambiguous statutes. Loper Bright Enterprises v. Raimondo (2024) overruled Chevron. Courts now decide statutory meaning de novo. The shift moves interpretive authority from agencies (with subject-matter expertise) to courts (without it). Defenders argue the move restores judicial responsibility for legal interpretation; critics argue it transfers technical questions to the wrong institution and increases policy uncertainty.

The combined implication is that policy that survives must increasingly be statutory, with clear authorizing language, rather than regulatory and discretionary. This raises the bar for major policy change. Reconciliation-eligible fiscal frameworks become even more important. Executive action becomes even more reversible.

33.12.1 Implications for the rest of the policy story

The administrative-state shift since 2022 has consequences that ripple through the rest of this book. Three deserve mention.

First, climate policy is now harder to do regulatorily than it was in the early 2010s. The Clean Power Plan strategy — using EPA authority under existing Clean Air Act provisions — is largely foreclosed by West Virginia v. EPA. Climate policy that survives must run through statute (the IRA's tax-credit framework), through state action (California's vehicle-emissions waivers, RGGI in the Northeast), or through narrowly-focused agency action that does not trigger major-questions concerns. This is not a partisan observation; it is a structural one. Regulatory ambition on climate now faces a higher legal hurdle than it did fifteen years ago.

Second, executive action on immigration faces similar constraints. The Obama administration's DACA survives, in modified form, but its legal foundation is repeatedly contested. Broader executive-action approaches — large-scale changes in enforcement priorities, sweeping regulatory expansions of asylum eligibility — are vulnerable to litigation under both major-questions doctrine and broader administrative-procedure-act challenges.

Third, financial regulation is constrained. The CFPB's rulemaking authority is contested in multiple Supreme Court cases. The SEC's climate-disclosure rule (2024) was challenged on major-questions grounds. Banking regulation, securities regulation, and consumer-protection regulation have all become more legally fragile.

These developments do not mean agencies cannot act. They mean agencies must act more narrowly, with clearer statutory hooks, and with greater attention to litigation risk. Policy ambition in these areas now requires either congressional statutory authorization (which raises the bar) or careful regulatory design within smaller authorities (which limits scope).

33.13 Reform proposals: steel-manning each side

The American policy process is widely criticized. Reform proposals come from across the political spectrum. The following five are the most prominent. Each is steel-manned.

Filibuster reform. Proposals range from elimination of the filibuster on legislation, to a return to the "talking filibuster" (requiring continuous debate to sustain), to exemption of specific issue areas (voting rights, judicial nominations). Defenders argue the filibuster forces consensus, prevents whipsaw policy, and protects minority voices in a chamber that already malapportions toward small states. Critics argue the filibuster has shifted from a rare instrument used in extraordinary cases to a routine sixty-vote requirement that effectively prevents most legislation, that the Founders did not intend it (the original Senate operated by majority rule), and that its current form blocks deliberation rather than enabling it. Both positions have empirical support.

Anti-gerrymandering reform. Independent redistricting commissions (now operating in California, Arizona, Michigan, Colorado), mathematical fairness standards (the efficiency gap, the partisan-symmetry test), and constitutional amendments are proposed. Chapter 35 develops these in detail. Defenders of independent commissions argue they reduce safe seats, restore competitive elections, and require members to represent broader coalitions. Critics argue commissions are not actually neutral, that constitutional amendment is unrealistic, and that mathematical fairness standards smuggle in partisan assumptions.

Permitting reform. The 2022 IRA's clean-energy investments require massive infrastructure build-out (transmission lines, geothermal, wind, solar, EV charging, hydrogen). The current permitting environment — overlapping federal, state, and local approvals; environmental review under NEPA; opportunities for litigation — produces multi-year delays. Senator Manchin's 2022 permitting-reform proposal would have streamlined approvals for both fossil and clean energy. It died of a combination of left-environmental and right-fiscal opposition. Defenders argue current permitting prevents the country from building the things current policy claims to want. Critics argue NEPA and related processes protect communities (especially low-income and minority communities) from environmental harms and that streamlining would shift costs to the most vulnerable.

Bipartisan-mandate budget reforms. Proposals to require supermajorities for new entitlement programs, automatic stabilizers tied to economic indicators, and statutory PAYGO with teeth. Defenders argue these would force sustained fiscal trade-offs rather than indefinite deficit financing. Critics argue they would lock in existing distributions, prevent counter-cyclical action, and impose status-quo bias on top of an already status-quo-biased system.

Congressional process reforms (Frances Lee, Yuval Levin tradition). Strengthen committee markup; restore amendment rights on the floor; reduce omnibus packaging in favor of single-subject bills; require regular order budgeting; increase committee staff capacity. Defenders argue these reforms would restore deliberation and broaden coalitions. Critics argue the underlying electoral incentives drive members away from bipartisan deals regardless of procedure, and that procedural reforms alone cannot fix incentive problems.

None of these reforms is guaranteed to work. None is obviously misguided. The point of steel-manning is that the policy-process is itself a contested policy question, and intelligent disagreement about how to fix it is a signal of intellectual seriousness, not a failure of analysis.

33.13.6 The under-discussed reforms

Five additional reforms attract less attention but matter for completeness.

Sunset provisions and reauthorization requirements. Many programs have indefinite authorization. They run on autopilot until specifically repealed. Sunset clauses force periodic reconsideration. Defenders argue sunsets compel programmatic evaluation and adjustment to current conditions. Critics argue sunset provisions create policy uncertainty that itself harms outcomes (recipients cannot plan; agencies cannot recruit; long-term investments are deterred), and that the sunset window often becomes a hostage moment when partisan coalitions extract concessions unrelated to the policy under review.

Single-subject bill rules. Most state constitutions require legislative bills to address a single subject. Federal Congress has no such rule, producing omnibus bills running thousands of pages and combining dozens of unrelated subjects. Defenders of a federal single-subject rule argue it would force transparency and prevent log-rolling that obscures accountability. Critics argue that omnibus bills are how complex compromises actually get assembled, that single-subject rules at the state level are unevenly enforced, and that the rule would simply shift complexity to multi-bill coordination.

Lobbying and revolving-door restrictions. Tighter rules on the movement of staff between Congress, agencies, and the firms they regulate. Longer cooling-off periods. Disclosure requirements on lobbying contacts. Defenders argue these reforms reduce capture and the appearance of capture. Critics argue restrictions on movement reduce expertise (the people who understand a regulatory area best are often those who have worked across both sides) and that strict enforcement creates First Amendment problems.

Independent fiscal scoring authority. Proposals to expand the Congressional Budget Office's mandate to include longer-run scoring, dynamic scoring (with explicit modeling of behavioral responses), and post-enactment evaluation. Defenders argue better scoring discipline would reduce optimistic-projection abuse. Critics argue dynamic scoring opens scoring to ideological judgment about behavioral responses and that retrospective evaluation should remain agency-specific rather than centralized.

Deliberative democracy mechanisms. Citizen assemblies, deliberative polls, and structured public-input processes. The Irish citizen assemblies on abortion (2016–2017) and constitutional reform are often cited. Defenders argue these mechanisms produce considered judgment that is less subject to short-term media cycles than legislative process is. Critics argue they are bypasses of constitutional process, that their representativeness depends on participant selection, and that their non-binding nature limits their value.

These reforms get less air time because they are technical. They are not less important. A policy-process textbook that names only the high-profile reforms (filibuster, gerrymandering) understates the range of available levers.

33.13.7 What reforms cannot fix

It is honest to note the limits of any reform package.

The deepest constraints on the American policy process are not procedural. They are electoral (the permanent campaign, geographic sorting, primaries) and cultural (declining trust in institutions, fragmentation of media, polarization of mass opinion). A more efficient legislative process operating in a more polarized electorate may not produce better policy; it may simply produce faster swings between policies that each partisan coalition prefers.

This is why so many reform discussions pivot, eventually, to electoral reforms — campaign finance, voting rights, redistricting, ranked-choice voting, multi-member districts, term limits — rather than legislative process reforms. The legislative process is downstream of the elections that produce it. Fixing the legislature without fixing the elections may be limited.

Conversely, fixing the elections without fixing the legislature may also be limited. A more representative Congress is still constrained by veto-point density, by the filibuster, by judicial review. The two reform agendas are complementary; neither replaces the other.

This chapter does not recommend a particular reform package. It maps the available options. The choice is yours. The textbook's view is that the choice is genuinely contested, that reasonable people disagree, and that the worst posture is the one that announces the answer is obvious.

33.13.8 The international comparison, briefly

Other constitutional democracies have made different trade-offs.

The United Kingdom: parliamentary supremacy, single-chamber effective veto, no judicial review of statutes (until the Human Rights Act of 1998 and even then in restrained form). Result: rapid policy change when government changes (the 2010 austerity package, Brexit's implementation), and high vulnerability to whiplash when government changes again.

Germany: federal structure with a Bundesrat (state-government chamber) that has limited veto powers; constitutional review by the Bundesverfassungsgericht (Federal Constitutional Court) which is highly active. Coalition government forces ongoing negotiation; basic-law amendments require a two-thirds vote in both chambers. Result: more stability than the UK, less stability than the US; substantial constitutional discipline.

France: semi-presidential system with strong executive authority during cohabitation periods of unified government; weaker during periods of legislative-executive division. Fifth Republic constitution gives the executive substantial decree authority within statutory frameworks. Result: rapid policy change in unified government, gridlock in cohabitation.

Canada: Westminster parliamentary system at the federal level, but a federal structure with strong provincial powers and a constitutional Charter of Rights and Freedoms (1982) enforced by the Supreme Court of Canada. Strong party discipline produces decisive parliamentary majorities; provincial governments operate independently in their constitutional spheres. Result: more responsive than the US to mass electoral mood, but with substantial provincial-federal coordination requirements.

The American system is at one extreme of the comparative space: many veto players, weak party discipline, federalism with substantial state authority, a constitutionally entrenched bill of rights actively enforced by courts, a presidential election entirely separate from legislative elections. No other major democracy combines all of these features. The United States is, in this sense, an outlier. Whether the outlier status is a strength or a weakness depends on which performance dimension you prioritize: stability, responsiveness, deliberation, redistribution, civil-liberties protection, or fiscal discipline.

The point of the comparison is not to recommend imitation. It is to remind readers that the American configuration is one choice among many, and that the trade-offs are visible in the comparative record.

33.14 What you should take from the chapter

A short summary, before you move on. The next chapter (Chapter 34) will examine money in politics — the dimension of political life that runs through this chapter at every stage. Chapter 35 will examine elections and voting; Chapter 36, voting rights; Chapter 37, the events surrounding the 2020 election and January 6. The remaining chapters of Part V build, on the foundation this chapter provides, the politics of the contemporary period.

Five things to carry forward:

The stages model is a vocabulary, not a sequence. Use the stages to identify the components of a policy event, not to predict their order.

Policy windows open briefly. Whoever is ready when they open shapes what passes. The deep work — formulation, coalition-building, framing — has to happen before the window opens, not during.

The American system has many veto players. This is structural, by design, and produces a status-quo bias that is feature for some purposes and bug for others.

Adoption pathways matter. Statutory, regulatory, judicial, state-level, and treaty pathways differ in speed, durability, and constraint. The choice of pathway shapes the substantive policy.

Reform proposals are themselves policy questions. The procedural rules of policy-making are not neutral. Steel-man each reform on its own terms before deciding which you support.

Process is power. So are policy outcomes. They are connected, but they are not the same thing. A textbook that does justice to both is doing the work of political science.

33.15 A note on what comes next

The chapters that follow build on this one. Chapter 34 examines money in politics, the dimension that runs through every adoption stage and shapes which policy ideas ever reach the formulation stream at all. Chapter 35 examines elections and voting, the political-stream input that determines which coalitions are available for adoption in any given session. Chapter 36 examines voting rights, the procedural rules that determine who counts in those elections. Chapter 37 examines the events surrounding the 2020 election and January 6, the most acute recent test of whether the policy process can sustain a peaceful transfer of power.

Each of those chapters can be read as a deeper investigation into one input or constraint of the policy process this chapter has mapped. The framework here is meant to provide vocabulary; the chapters that follow provide substantive examination of particular conditions.

A final thought. The chapter has been honest about gridlock, veto-density, and the structural status-quo bias of the American system. It has also been honest about successful policies — the Clean Air Act amendments, CHIP, Medicare Part D, the Bipartisan Infrastructure Law, the IRA, CHIPS. The picture is mixed. It is not a story of comprehensive failure; it is also not a story of comprehensive success. It is the story of a system designed to require broad coalitions, sometimes operating in a politics that can produce them, and often operating in a politics that cannot. Whether the design has aged well is genuinely contested. Whether the design can be reformed without losing its protective features is genuinely contested. Whether the deeper electoral and cultural conditions can be reformed at all is genuinely contested.

Take this chapter's vocabulary as a starting point, not an answer. The hardest questions about the American policy process are not vocabulary questions. They are judgment questions. Make your own judgments — but make them on the strongest version of each side.