How to Be a Policy Entrepreneur in the American Vetocracy

A review of "Recoding America: Why Government Is Failing in the Digital Age and How We Can Do Better," by Jennifer Pahlka
April 2nd 2024

This article originally appeared in American Affairs Volume VII, Number 4 (Winter 2023): 20–28.

To paraphrase economist Robert Solow, once you start thinking about state capacity, it’s hard to think about anything else. Or, at least, that’s been the case for Jennifer Pahlka, the founder of Code for America, a deputy CTO during the Obama administration, and the au­thor of Recoding America, a scintillating new book about how government services are actually implemented.

Pahlka’s camp of policy wonks (of which I am a member) is broadly concerned with improving state capacity, the ability of the government to accomplish its policy goals. Pahlka is most interested in what happens when policies get implemented, and how kludgy government services can be delivered more effectively. Recoding America focuses on the many cases where policymakers unknowingly undermine their own objectives when they ignore how policy implementers actually turn legislative text into government-services reality.

Pahlka’s book lays out a tremendous framework for policy entrepreneurs—individuals inside or outside of government working to change policies—who want to engage the administrative state to improve implementation. But America’s state capacity is hobbled by more than just poor implementation. We can build on Pahlka’s framework by examining the deeper forces in American political institutions that prevent new policies, such as permitting reform, from ever making it to the implementation stage.

In our highly polarized political environment, five structural factors originally intended as checks and balances have metastasized into what Francis Fukuyama calls a “vetocracy”: bicameralism, the Senate fili­buster, presidentialism, adversarial legalism, and federalism. To be sure, these institutions also have benefits, such as guarding against the tyranny of the majority. But to actually change laws, policy entrepreneurs must design their solutions to run this gauntlet of veto points. It may be difficult, but it is possible to build state capacity while operating within the constraints of U.S. institutions: the federal government can use carrots and sticks to align state and local policy with its goals; Congress can use its authority to limit judicial review; and legislators can add riders to must-pass legislation.

Case studies on the importance of policy implementation

For a book about government websites and interagency meetings, Recoding America is remarkably enthralling. Pahlka vividly conveys how treating software development like a box-checking exercise can completely stymie well-meaning policy. Take this story about veterans filling out an online form to apply for health care benefits from the VA:

“The health care application wasn’t a web form like the kind you use when you’re booking a rental car or taking a survey. It was a paper form that had been put online as a PDF. If the computer you were using had a particular (and outdated, even then) version of Internet Explorer, and a particular (also outdated) version of Acrobat Reader, and it was set up to open PDFs in Reader by default, you could click on the link for the form on the VA web­site and it would appear in your browser. You could then type your information into each field on the form, save it, and send the completed document to the VA electronically. The computers inside the VA ran that exact combination of Internet Explorer and Acrobat Reader and were set up to open Reader by default. Whenever anyone inside the building clicked on that link on the website, the form would load. It worked fine.

“But outside the building, very few computers were set up that way. Some of the older computers . . . tested were running even earlier versions of the software. Others were running more recent versions than those inside the VA. The team tested the application on all kinds of computers running all kinds of common combinations of software. In none of the tests did the application open in the browser.”

In Pahlka’s vignettes, implementation is often the difference between success and failure. And she has no shortage of case studies to cite: from the Obamacare rollout disaster, which managed on its first day to serve a total of eight people, to the California unemployment insurance backlog during the pandemic, which became so large that only about one in a thousand support calls were answered. In the most stark example, she contrasts two criminal justice reforms in California, one of which was trivially easy to implement at scale and the other virtually impossible.

In 2014, California passed Prop 47, a referendum that reclassified all burglaries of property under $950 from felonies to misdemeanors, and applied retroactively. But rap sheets do not include a code for whether the value of the stolen property was under $950, meaning people trying to change their criminal records would have to dig through paper files and hope they could find enough evidence for a judge to conclude the theft qualified. As Pahlka says, “the way the law was written makes automated expungement impossible.”

Two years later, California passed Prop 64, which expunged marijuana-related felonies and was relatively simple to implement in an automated fashion. The Code for America team developed an algorithm to scan electronic court files and identify eligible records. This tool has already found 144,000 convictions eligible for expungement. “Instead of decades of work,” Pahlka writes, “the software program had taken about thirty seconds to run.”

Implementation is necessary but not sufficient

The remedy to these kinds of problems is relatively straightforward: policymakers should, even in the early stages of crafting a policy solu­tion, proactively look ahead to the implementation phase to understand what is technically feasible. Yet while U.S. policy would surely be im­proved in many dimensions by implementation-forward thinking, it is not a panacea. For resolving a thornier set of issues, paying attention to the implementation phase is necessary but not sufficient. One particularly salient example is reforming the permitting process to accelerate the buildout of clean energy infrastructure. As Pahlka notes at the end of her book, passing federal subsidies was just step one in decarbonizing the economy:

“The work is certainly teed up for that administrative state. Despite never-ending headlines about gridlock in our nation’s capital, the 117th Congress (2021–22) is on track to pass well over two hundred laws, including such whoppers as the Infrastructure Investment and Jobs Act, the chips and Science Act, and the Inflation Reduction Act (IRA). The implementation challenges of these and other legislative outputs nationally cannot be overstated. . . . [We’ll need to] dramatically streamline solar permitting in three thousand counties and almost twenty thousand towns and cities, [and] navigate stakeholder input processes in the countless jurisdictions new power lines must travel through. . . .

“‘Fifty years from now, I’m going to explain to young people what happened, when we’re all huddled around a tire fire for warmth after the climate collapse,’ Lucas Merrill Brown [a Biden administration environmental policy staffer] told me. . . . ‘A lot of people worked really hard to build political consensus to fight the climate crisis. We passed half a trillion dollars in funding. And then we kept using fossil fuels because we couldn’t figure out the permitting.’”

But permitting is not merely an implementation challenge for the administrative state. According to the repeat Project at Princeton’s ZERO Lab, if the United States doesn’t more than double the rate at which it expands electricity transmission capacity, then 80 percent of the emissions reductions from the Inflation Reduction Act could be lost. The administrative state cannot solve this problem merely by processing developer proposals more quickly, or launching a digital platform for submitting applications. To put it bluntly, more efficient implementation won’t solve climate change—we will need to fundamentally over­haul the way the permitting process works.

Major electricity transmission projects involve complex coordination between federal, state, local (and sometimes hyper-local) officials, as well as securing rights of way from private landowners. To streamline this process, Congress will need to pass new laws. But that legislation will be dead on arrival if proponents don’t understand how to navigate the American vetocracy (even if they take Pahlka’s strong advice about implementation). Policy entrepreneurs have to understand the structural barriers to American state capacity, and develop a strategy to overcome them.

The super-trifecta problem

Collectively, three institutions—bicameralism, the Senate filibuster, and presidentialism—make it extraordinarily difficult to pass or update laws compared to parliamentary democracies. In the modern era, it is extremely rare for one party to have a “super-trifecta,” controlling the presidency, a majority of the House, and a sixty-vote supermajority in the Senate. Since 1981, there has only been one period—between July 2009 and February 2010—when this occurred. The Affordable Care Act, arguably the most significant piece of legislation since the Great Society programs, passed during this brief seven-month window, in a party-line Senate voteReconciliation bills, which require only a simple majority in the Senate, are strictly limited to “budgetary matters.”

Senator Joe Manchin wanted permitting reform to be included in the Inflation Reduction Act, but permitting is not primarily a budgetary issue and the IRA was a reconciliation bill (passed on a 51–50 party line vote in the Senate). He later tried and failed to include it in a continuing resolution and as an amendment to the National Defense Authorization Act (NDAA). Absent a vetocracy, comprehensive permitting reform probably would have passed by now, likely as a direct trade for Manchin’s vote on the climate bill.

While reforms to make legislation easier to pass are a long-term state capacity goal, what are policy entrepreneurs to do in the meantime? Given how hard it is for legislation to clear the super-trifecta threshold, Congress has developed a pattern of grouping policy priorities in a small number of “must-pass” omnibus bills. It’s possible for crafty policy entrepreneurs to tailor their proposals to fit within one of these legisla­tive vehicles.

For example, continuing resolutions (CRs) and appropriations bills allow federal agencies to utilize designated funds for their operations and programs. Failure to pass either a CR or a full appropriations bill causes a government shutdown. Despite the prevalence of shutdown threats during divided government, the budget almost always passes, making it a reliable vehicle for riders.

As a key national security bill, the NDAA has been passed annually for sixty-one consecutive years. The farm bill, usually reauthorized every five years, is another must-pass vehicle. Debt ceiling adjustments, which permit the U.S. Treasury to borrow up to a specified limit, can also become “Christmas tree bills,” as failure to pay debts on time could be catastrophic for U.S. creditworthiness. Congress has approved 103 separate debt-limit modifications between the end of World War II and the present—never breaching the ceiling (though coming close at times).

While adding the entire permitting reform package to must-pass legislation last year didn’t work, the same tactic has already worked twice this year in a piecemeal fashion. The most recent debt-ceiling deal in June included incremental reforms to the permitting process. Even more recently, a narrow permitting-reform amendment was added to the NDAA to exempt from environmental review the construction of semi­conductor facilities that receive chips Act subsidies. It seems plau­sible that reforms to permitting interregional transmission and altering the NEPA judicial review process could be added to the next appropriations bill, the next NDAA, or the five-year FAA reauthorization.

Adversarial legalism

Pahlka says that poorly designed government services often come about because an agency’s “overriding goal was not to design a service but to avoid being sued.” This threat of litigation hangs over every government action, and merits a more thorough explanation beyond the scope of Recoding America. The U.S. legal system is based on common law, which is predicated on the principle that similar cases should be decided similarly. Judicial decisions are considered a crucial source of law, and courts often rely on precedents to guide future case rulings. While a common law system can evolve over time as judges hear new cases, it can also lead to greater uncertainty regarding government actions.

Moreover, the specific characteristics of the American common law system amount to what Robert A. Kagan calls “adversarial legalism.” Under adversarial legalism, disputing parties (rather than judges or gov­ernment officials) bring claims, identify controlling legal arguments, and gather evidence. Kagan concludes that adversarial legalism has two char­acteristic consequences: costliness and legal uncertainty—which form a lethal combination for state capacity. If federal actions are costly, then the government will shy away from acting frequently. And if there is legal uncertainty, then private sector actors will not be able to rely on government actions.

Pahlka touches on this briefly when discussing how the environmental movement has leveraged the court system against government agencies:

“We may think of the 1960s and 1970s campaigns to clean the air, preserve wildlife habitats, and stop chemical pollutants from being dumped in waterways as being anticorporate, and indeed it was corporations committing these atrocities. But overwhelmingly, it was not companies that the activists sued but the government agencies that gave the green light for the companies’ projects. . . . Laws that made new construction projects much harder, for instance, are today huge impediments to mitigating both the housing crisis and the climate crisis, blocking not only affordable housing but also critical renewable energy infrastructure like solar and wind farms. The environmental movement got really good at stopping bad things from happening but less adept at enabling, and speeding up, good things. Now the habit of suing has itself become a great threat to our environment.”

The environmental movement has become accustomed to challenging NEPA reviews under the Administrative Procedure Act. A recent study surveying major transportation and energy infrastructure between 2010 and 2018 found “predevelopment litigation on 28% of the projects requiring an environmental impact statement, 89% of which involve a claim of a NEPA violation.” According to a 2014 GAO survey, the risk of lawsuits significantly increases the cost and time to prepare reviews in an attempt to make them “litigation-proof.” This incentive to head off lawsuits has driven the growth of environmental reviews from dozens of pages in the 1970s to thousands of pages today. Even with extremely long environmental reviews, bad faith actors can still find something that hasn’t been comprehensively addressed, as in the ongoing fight over congestion pricing in New York City.

Policy entrepreneurs who encounter problems downstream of adversarial legalism should prioritize reforms that reduce litigation risk. Given the high costs and uncertainty created by adversarial legalism, Congress already routinely uses its constitutional authority to constrain the scope of judicial review. For example, under the Federal Arbitration Act of 1925, Congress can mandate arbitration for certain disputes, removing them from the judicial system. Alternatively, Congress can establish specialized courts to handle specific types of cases more effi­ciently; for instance, the United States Tax Court specializes in federal income tax disputes. Congress can also make it harder for plaintiffs to establish the legal standing needed to sue.

In the context of permitting, there are multiple ways policymakers could create more legal certainty for project developers. They could impose a time limit on judicial injunctionslimit standing to those who participated in a public comment period, and create a shorter statute of limitations on filing lawsuits. Other options include skipping the district court level and sending NEPA cases directly to the D.C. Circuit Court of Appeals.

Ideally, reforms would protect good-faith lawsuits intended to im­prove a project or prevent significant public harm, while filtering out bad-faith lawsuits intended to protect some narrow private interest. Under adversarial legalism, permitting reform ideas need to include some limitation on judicial review, or lawsuits will continue to delay projects and lengthen environmental reviews.


The federal system of government can lead to complex, overlapping policies and legal systems, as Pahlka briefly notes when talking about her experience working on the California unemployment insurance backlog: “federalism also makes the job of the Department of Labor, which has to regulate each state labor agency to ensure compliance with federal law, exceptionally difficult.”

Federalism has been a significant issue for permitting, as projects often need to secure approval from local, state, and federal officials. One of the key challenges of large linear infrastructure projects, like trans­mission lines, is that they span multiple jurisdictions (and usually need to acquire the rights of way from property owners). This leads to an obvious holdout problem for major projects. Any actor with the ability to veto the project can withhold its approval until it has extracted the maximum amount of value from the project.

To overcome this hurdle, policy entrepreneurs can design federal legislation that includes both carrots and sticks to guide state and local governments toward certain policy decisions. By leveraging federal funds and regulatory authority, the national government can nudge states to adopt policies that align with federal objectives, even in areas traditionally under state jurisdiction.

A few examples from other areas of policy: federal highway funds can be one effective tool for corralling states. Under the Clean Air Act, states must develop and implement State Implementation Plans to achieve or maintain the National Ambient Air Quality Standards. If states fail to comply, the EPA can withhold up to 100 percent of federal highway funds or impose a Federal Implementation Plan. The National Minimum Drinking Age Act of 1984 threatened states with the with­holding of 10 percent of federal highway construction funds if they didn’t raise the minimum legal drinking age to twenty-one. Faced with the potential loss of substantial funding, every state eventually complied. The “Race to the Top” program, launched in 2009, had states compete for a share of more than $4 billion in federal education grants by adopting reforms favored by the federal government. The promise of substantial funding incentivized many states to adjust their educational policies.

In each of these areas, policy entrepreneurs used a mix of carrots and sticks to try to overcome inherent coordination problems in our federal system. In the case of siting, planning, and cost allocation for interregional transmission lines, Congress could empower FERC to take over the process, just as it did for pipelines under the Natural Gas Act of 1938. Alternatively, the federal government could set a minimum trans­fer standard between regions, and let states and public utilities decide how best to fulfill that requirement. Or Congress could tie IRA, Infra­structure Investment and Jobs Act, and chips subsidies to states and localities that streamline their permitting processes. Lastly, Congress could direct federal agencies to grant immediate approvals for decarbonization projects where environmental impacts are well understood, either due to the nature or location of the project.

Policy is upstream of implementation

Across numerous stories, Pahlka shows the importance of the implementation stage of policy. She persuasively argues that government needs more product managers focused on user experience and fewer project managers focused on completing a checklist of requirements. She makes a compelling case for civil service reform, and a general shift to bring more technical capacity in house instead of relying on external contractors (the development was outsourced to thirty-four different vendors across sixty separate contracts). Perhaps the lowest-hanging fruit in software development: government engineers should adopt an agile workflow rather than the waterfall methodology.

Pahlka is also right to note that implementation has traditionally been neglected:

“The bill passed. We celebrated. What comes next is merely ‘operational in nature.’ If we want to escape that fate, on climate or any other existential issue we face, implementation can no longer be policy’s poor cousin. It can’t be beneath the attention of our most powerful institutions, and it can’t be beneath our attention as a public.”

But as the permitting issue shows, fixing administrative implementation alone won’t cure what ails American state capacity. Policy entrepreneurs working to advance reforms still need to navigate the system in Washington or their ideas will never make it to the implementation phase.

In an era of polarized politics, the combination of bicameralism, the Senate filibuster, presidentialism, adversarial legalism, and federalism has resulted in a vetocracy. While this institutional arrangement makes poli­cy change in the United States difficult, progress is possible. Federal policies can include carrots and sticks for state and local governments to get them onboard. Legislation can include limitations on the scope of judicial review to provide more certainty regarding federal actions. And advocates can modify their proposals to be included in must-pass legislation. Only policy entrepreneurs who have a clear-eyed view of the bottlenecks to U.S. state capacity can work to unlock them.