Jennifer Pahlka is founder of Code for America, served as U.S. Deputy Chief Technology Officer, and is author of the newly published Recoding America: Why Government Is Failing in the Digital Age and How We Can Do Better. She spoke with City Journal associate editor Daniel Kennelly.
Your book opens with a discussion of several failures of government policy implementation, from Medicare’s unwieldy bureaucracy driving doctors out of the profession to the difficulties surrounding the launch of healthcare.gov to Covid-era economic-relief programs. What unites all these failures?
In each of these cases, you have well-meaning, hardworking civil servants operating within a system that is collapsing under its own weight. Often, what the public sees is a website, like healthcare.gov, or a DMV site, and they conclude that government is bad at technology. But behind that clunky technology are decades of accumulated policy and process cruft. It is constantly added to and never removed; policy frameworks become increasingly complex, never simpler. When I was working on unemployment insurance benefits during the pandemic, one claims processor who had been at the department for 17 years explained that he was still learning all the ins and outs. He routinely deferred to colleagues with 25 years of experience or more. If it takes that long to master the rules that govern giving people money when they’re out of work, there’s something wrong.
Could you give a specific example of problematic program implementation and its effects on people who relied on it?
Value-based care in Medicare is a good example. The Medicare Access and CHIP Reauthorization Act passed in 2015 with the intent of improving the care that patients receive by paying doctors more for better-quality care. But the way that doctors had to submit their quality data was already maddening to begin with, so many sole practitioners were planning to stop taking Medicare patients altogether rather than invest in the new software and the expensive retraining that it would require. This, of course, would degrade the quality of care. The burdensome, confusing implementation of past programs had left doctors with little trust in the Centers for Medicare and Medicaid Services. Without careful attention to doctors’ needs, the new law would have the opposite effect from what Congress intended.
Fortunately, in this case, the team doing the implementation fought to simplify the policy complexity that had made past systems so confusing and burdensome. Their mantra became “I get that it’s complicated. But it has to make sense to a person.”
Where do we find the solutions to these implementation problems? Better policy? More regulation, money, or oversight? And whose job should this be?
It starts with the public holding elected leaders accountable for implementation. When candidates ask for your dollar or your vote, do you ever ask them what they’ve done to make the bureaucracy more efficient or less risk-averse? Or to ensure that existing laws have been well implemented? Mostly we just want to know if they share some of our values, and we expect that this affinity will translate into the kind of country, state, or city we want to live in. It rarely does.
Leaders can also make specific changes to support better implementation. Today, funding happens in giant chunks that breed mega-projects; it could start small so that teams can learn. Today, oversight focuses on adherence to process, increasing risk aversion. And it focuses on failure; public servants only know what gets you in trouble. They need to know what “good” looks like—not just staying away from “bad.” The stories in my book provide some of these models.
Tell us about the United States Digital Service (USDS), its creation, and the impact it has had on government-service delivery.
Several of us were trying to stand up what became USDS in 2013, when healthcare.gov failed. It marked the first time that the White House’s top priority was the functioning of a website. The White House is generally very focused on policy and politics, but it had to face the reality that its signature policy initiative was threatened by poor implementation.
USDS doesn’t really do tech for agencies; it helps agencies get over humps and adopt more effective, modern practices. Some of its biggest successes aren’t about tech at all, like SME–QA [Subject Matter Expert Qualification Assessments], a path for hiring in the federal government that ensures that candidates are actually assessed for their skills, which the standard process very rarely does. USDS is best when it helps agencies do their jobs better.