Pod Insights The Ezra Klein Show The Book I Wish Every Policymaker Would Read About this episode My pitch for this episode is simple: Jennifer Pahlka has written one of the best policy books I’ve ever read.
Pahlka served as deputy chief technology officer in the Obama White House, and she’s the founder and a former executive director of Code for America, a nonprofit that works to enhance government digital services. Over the course of her career, Pahlka has become obsessed with an area of policy that is too often ignored by policymakers: implementation. She was part of the effort to rescue HealthCare.gov in 2013 and was tapped by Gov. Gavin Newsom in 2020 to help fix California’s unemployment insurance system as it buckled under the weight of the Covid response.
It has become a common refrain that the U. S. government is often terrible at delivering even basic services. But Pahlka’s new book — “Recoding America: Why Government Is Failing in the Digital Age and How We Can Do Better” — puts forward a deeper theory of why government services are so awful, how policy implementation so often goes awry and what it would take to fix those systems so that government could better live up to its promises. It’s an argument that anyone who cares about government in the 21st century needs to take seriously.
Book Recommendations:
Implementation by Jeffrey L. Pressman and Aaron Wildavsky
Radical Help by Hilary Cottam
“Mandate for Leadership” (chapter 3), edited by Paul Dans and Steven Groves
Listen to this podcast in New York Times Audio, our new iOS app for news subscribers. Download now at nytimes.com/audioapp
Thoughts? Guest suggestions? Email us at [email protected] .
You can find transcripts (posted midday) and more episodes of “The Ezra Klein Show” at nytimes.com/ezra-klein-podcast, and you can find Ezra on Twitter @ezraklein. Book recommendations from all our guests are listed at https://www.nytimes.com/article/ezra-klein-show-book-recs.
This episode of “The Ezra Klein Show” was produced by Emefa Agawu. Fact-checking by Michelle Harris, Mary Marge Locker and Kate Sinclair. Mixing by Jeff Geld. Our production team is Annie Galvin, Jeff Geld, Roge Karma and Kristin Lin. Original music by Isaac Jones. Audience strategy by Shannon Busta. The executive producer of New York Times Opinion Audio is Annie-Rose Strasser. And special thanks to Isaac Jones and Kristina Samulewski.
Transcript 12,951 words · 2 speakers
Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.
0:22 Ezra Klein? In 2009, Jen Polkow founded Code for America. And the idea behind Code for America was to get the best technologists working on the public's problems, not just private websites and e-commerce platforms and social networks. Even then, it was clear that public digital infrastructure was going to be really important, and it was lagging far behind the private sector. And Code for America was built to bridge that gap. Then in 2013, Polkow became Deputy Chief Technology Officer in the Obama administration. She was very deeply involved in the effort to rescue healthcare.gov and helping create the United States Digital Service. Then she went back to Code for America, and then in 2020, was tapped by California Governor Gavin Newsom to help untangle the mess of California's unemployment insurance program
1:06 Ezra Klein? as it buckled under the weight of the COVID response. I mention all that because Polkow has worked directly on and then helped oversee and advise teams working on the digital delivery of government services, really at every level of government across much of the US. So she's someone who badly wants government to do good in the lives of people and has spent years now confronting the reasons it so often falls so short of its goals. One of her big points, one I've come to appreciate a lot more in recent years, is that even liberals who care a lot about government don't care enough or track closely enough how implementation actually happens. Delivery often happens out of sight, except for the people who need that delivery.
1:49 Ezra Klein? In our media, I mean, we're part of this, I'm part of this problem too, there's a ton of focus on politics, on elections, on big policy questions and fights and theories. But then it all passes, and the nitty-gritty of how that policy actually shows up in people's lives is left up to someone, somewhere. And when it doesn't... Show up in people's lives or even makes people's lives worse because of how it is implemented. There's often no outcry because there's no attention and so there are no fixes. Now, Polka has written this book, Recoding America. And I want to say it clearly. This is one of the best policy books I've read. It's a book I hope future governments and current governments will absorb. We get at a lot of it in this conversation, but not enough.
2:34 Ezra Klein? But one thing she really does is tell clearly the stories and tell clearly the lessons she has seen and learned about why things go wrong, even when the people involved are trying to make them go right. So people who care about these issues really should pick it up and read it in full. But this conversation is a really rich place, I think, to start. As always, my email, [email protected] .
3:01 Ezra Klein? Jen Polka, welcome to the show.
3:03 Speaker 01 Thanks for having me.
3:04 Ezra Klein? So I want to begin with a quote that feels in a way like the organizing thesis of the book, which is, "When systems or organizations don't work the way you think they should, it is generally not because the people in them are stupid or evil. It is because they are operating according to structures and incentives that aren't obvious from the outside." And I love that because for me, it's a bit of a creed on reporting on government. But you've been inside government. So tell me a bit about how you came to that view.
3:34 Speaker 01 I've seen so many people frustrated with government services. And they make a lot of assumptions about how that government service got to be the way it is. One example I spent a lot of time on was SNAP in California, Supplemental Nutrition Assistance Program, where the application form was over 212 questions. It took a long time to fill out. And so people naturally assume that this is done on purpose, that the people making the form must not want people to get SNAP. And for example, there, that's not really what was going on. California is a very poor welfare state. There's a huge number of incentives for local communities to sign people up for SNAP. It's great for their economy.
4:17 Speaker 01 It's great for getting people out of emergency rooms, for instance. But there are just so many stakeholders in this that they all got to pile on their questions and you ended up with something that was really, really burdensome. And it's hard to see those incentives from the outside. That's just one example of them. But our public servants are incented to do things that if you look at them rationally, don't make sense.
4:40 Ezra Klein? You have a great line in that section, which I'll probably get wrong in the paraphrase, but you say something like, "People often think it was designed this way, when the truth is it wasn't designed at all."
4:51 Speaker 01 Yeah. I think that's true of so many things that we think are systems. I also think it's true of the policy. Take unemployment insurance, for example. We've had layers and layers of changes on it over time. Those changes come from federal government, they come from state government, they come from executive, legislative, and judicial branches. And so what you end up with 90 years later, almost, is something that's just sort of a mess of changes, but not something like thoughtfully crafted by policymakers in DC that is, you know, now operating in some rational way at the state level.
5:27 Ezra Klein? So I thought a lot about whether or not to do this because there is probably a best practice of podcasting where you start in the examples that are going to be attainable and people already know. And we have a bunch of those, healthcare.gov and California Employment Insurance. But I want to start actually with weedsier one, which is, tell me about the Enterprise Services Bus.
5:50 Speaker 01 Okay, you want to hear about the ESP?
5:52 Ezra Klein? Yeah, because I think it gets at this point of systems where clearly nobody was trying to do something wrong here, and yet.
6:00 Speaker 01 Yeah. I think everyone was trying very hard to do it right. So a friend of mine, Matthew Weaver, was assigned to go help a team at Raytheon, who had a contract with the Air Force, to do essentially the next generation of software for the satellites that enable Global Positioning System, GPS. So something that's pretty important to all of us, especially people like me who need it all the time. And when he got there, he found that one of the big things that was hanging them up was that they had to get the data from the satellites to the ground stations. And there's a very industry-standard way of doing this. It's the obvious way you would do it. But instead, the team had inserted this thing called an enterprise service bus,
6:44 Speaker 01 or ESB, which technical people will know as something that was like really popular in the 80s and 90s, but it's a big clunky piece of software. And it was kind of creating this Rube Goldberg machine that made the data go like from the normal, the easy protocol into this crazy mess of code and then back out again. And it would timeout. You couldn't get the data from one place to the other in time. They'd spent months on this and they couldn't figure it out. And he's like, why are we using this ESB when this other protocol would work fine? It's a requirement in the RFP that the Air Force had put out.
7:25 Ezra Klein? An RFP being a request for a proposal. So the RFP is how the Air Force gets, ends up with Raytheon building this thing.
7:31 Speaker 01 Exactly.
7:32 Ezra Klein? And we're going to talk about the contracting in a bit.
7:33 Speaker 01 Yeah, we'll talk about the contracting. So Raytheon can't take it out because they say, no, the Air Force is requiring us to do this. Why is the Air Force requiring it? Because the Department of Defense requires it. Why does the Department of Defense require it? Because there's something called the federal enterprise architecture that requires it. The story of the federal enterprise architecture goes back to this act called Klinger-Cohen that was, I think, 1996, where Congress said, like, we need to get a better handle on digital. We're not doing technology well in government. Let's make every agency create a plan and then let's have those plans be coordinated. And the federal enterprise architecture was the attempt to coordinate those plans.
8:15 Speaker 01 So one of the things they put in there was that things needed to be interoperable. And they offered in there as an example, this very common type of software at the time, enterprise service buses, as an example of how you might get things to be interoperable. They weren't actually a great idea at that time. They were already kind of becoming sort of obsolete, like essentially, application programming interfaces, APIs that sort of run the internet now kind of took them over. The problem really wasn't that they were specifying something outdated. It's that because it was in this document as a suggested way of doing what the federal enterprise architecture said should happen.
8:59 Speaker 01 When it got translated down to the Department of Defense enterprise architecture and the Air Force enterprise architecture, each layer in that hierarchy interpreted it more and more rigidly until at the Air Force enterprise architectural level, it's assumed to be required in all software forever. And so the team is saying, I can't take this out. The law requires it. In fact, the law does not require an enterprise service bus. But the people who have made all of the documents that govern the creation of the software for these satellites believe that it is.
9:39 Ezra Klein? When I was reading that story, it reminded me of a piece I just spent a couple of months working on about what I called everything bagel liberalism. And that was a piece that was in part about affordable housing and part about semiconductor contracting. And it was a weird piece of work, to be honest, in its structure. But when I was reporting on the basically request for proposal the Biden administration put out for its semiconductor grants, one of the things that came up as I was talking to members of the administration about all of the additional layers of requirements and asks and, you know, having a childcare center and this and that that ended up in the bill, I would hear about people say, look, we think those things are good. You don't realize we actually fought a bunch of these battles
10:24 Ezra Klein? and knocked a bunch of stuff out that otherwise would have been there. And then I would read the text and I would see the thing they told me was knocked out in the text. But it was a suggestion. It was a, we're not requiring this. But, you know, one good example of how you might achieve this goal we are requiring is this. And it was really striking to me that people were telling me on background that they had fought and won this fight to not have X in the notice of funding opportunity. And then I had read the notice of funding opportunity and X was there. And it was true that it wasn't required, but it seemed very likely that the semiconductor firms were going to think it was required, or at least think, and this was a plain reading to me too,
11:07 Ezra Klein? that they would have a better chance of winning the grant if they put it in there. And so there is this way in which even when people seem to think they've created flexibility in the actual language.
11:20 Ezra Klein? Once the lawyers get hold of it and the contractors get hold of it, it's really not that flexible.
11:26 Speaker 01 There's so much better safe than sorry thinking going on. And I would say with contractors and with people who do the compliance, right? Those bids that those companies give back, they know they'll be judged by people who are really just checking boxes. Like, does this meet this? Does this meet this? And they're going to check that box. I think it's a good example of the sort of disjuncture between the thinking of policymakers and the reality of how that policy plays out. You know, another good example of it is FISMA. It's the Federal Information Security Management Act.
12:03 Ezra Klein? Yeah. Which I only know because I have your quote on this sitting here in my next paragraph notes.
12:08 Speaker 01 You know, I've gotten pretty good at remembering like what all the acronyms stand for, but there are a lot of them. It can be hard. So FISMA is a law that covers cybersecurity and it has these 300 controls in it. So 300 different things that you, according to law, could do to secure an application.
12:25 Ezra Klein? So it's like a menu, not 300 mandates?
12:27 Speaker 01 But in practice, if you ask pretty much anybody who does software in government, they have to do all 300. And it's because I think it's evolved such that in order for people to sort of cover their butts, they have to explain why you didn't do numbers, you know, 138 through 192 or something. And it's really hard to do that. In fact, it's sometimes easier just to do the controls than it is to explain them. But the people in the middle who are doing the compliance have no idea if for your application, numbers 35 and 95 are the two really, really important ones. And you should spend the rest of your time on the thing that's most important to security, which is testing.
13:10 Speaker 01 All they know is that they will be responsible if anything bad happens. And if it was, you know, control number 96 that got you, it's their butt on the line. So they will insist on all of them. And if you go talk to the people who wrote FISMA, they will say, we never intended that.
13:28 Ezra Klein? Well, this gets at a, I think, a really important point you make about. That incentive system, how it is actually designed versus how it might look from the outside, you write that public servants are trapped between two distinct systems of accountability. So what are the two?
13:42 Speaker 01 Well, you know, when things go wrong, we're very upset with public servants. I mean, healthcare.com is a famous example. There were like 10 hearings in one month about this and they're televised and we're upset because the thing doesn't work. That's an outcome. But what public servants are actually hired for, promoted for, what their careers depend on, is fidelity to process. It's a compliance framework. And so, yeah, if you didn't do number 96 and something goes wrong, that's what the sort of administrative state will focus on. The people who do oversight focus on whether you checked all the boxes, not whether it works.
14:22 Speaker 01 And in fact, it's quite true in people's careers that they are not punished for something not working, as long as they can say, I did what the procedure prescribed.
14:34 Ezra Klein? Something this gets at is that we have civil service protections. That politicians are mad at the outcome of something civil servants did. Might allow them to embarrass the agency. They might even cut funding for the agency in the future, although often they don't. But they can't fire anybody.
14:52 Speaker 01 No.
14:53 Ezra Klein? They can't really create that many consequences for the agency. Whereas failing to follow process, you really can get fired. You really can lose your promotion track. There really are consequences. And so, in a weird way, the people who made the policy in the first place, and in theory are exposed to the voters who are supposed to be the judges of the policy, actually do not have power, really, to hold the civil servants accountable. And there are good reasons for that. I mean, there are good reasons for civil service protections, but there is also consequences to it, which I think frankly, on the left are pretty underplayed.
15:27 Speaker 01 Yeah, I mean, I heard so many times when I worked in government, you could get fired for that. In fact, you hear frequently, you could go to jail for that. It doesn't really match the reality. Like, people don't get fired that much, and they certainly don't go to jail. We used to have a joke about, really? I'm gonna go to procurement jail? Where exactly is procurement jail? Right? It's like such a nerdy piece of law and policy that certain people understand, but there's this perception that like a small violation of it is gonna end you in jail. It doesn't really match the reality, but the culture has definitely evolved. To highlight any deviance from process and make it very, very risky. It can be risky on a like, just a practical level.
16:11 Speaker 01 Like, you might not get demoted. You might not lose your job, but suddenly people are not wanting to work with you. There can be sort of cultural consequences to it and just your ability to get the job done.
16:24 Ezra Klein? Tell me about the Paperwork Reduction Act.
16:27 Speaker 01 So this is a great example of a law that sounds great. And it sounds something like something that's really important to me because it's supposed to reduce the paperwork burden on the American public. And I think the paperwork burden is a real problem, in part because, as they say in the book, paperwork favors the powerful. Like, the more paperwork that's required, people with resources will get through it and people without them will not. So this law said essentially something we all believe, what, you know, if you measure it, you'll be able to manage it. So I always call it the comically misnamed Paperwork Reduction Act because it's designed to measure the burden of any, say, form that you have to fill out.
17:10 Speaker 01 Like, your tax form, for instance, the 1040, that has an O M B number, which is required by the Paperwork Reduction Act. And it means that the Office of Information and Regulatory Affairs, known as O I R A, reviewed the time it would take to fill out this tax form and stamped its approval. And so it's got this measurement on it. The problem is that it got interpreted very broadly to say, any time you speak to a member of the public about a form like this, that has to get O I R A approval. So in the tech world and consumer product world, we have something that's really important, the usability of our services, called user research.
17:54 Speaker 01 The way we make a form easier, like that tax form, is that we watch people filling it out, and when they get stuck at a question, like, what were your last year's state taxes? We go, okay, we've got to make that easier on that person, and we redesign it. So we have this very important discipline called user research that is how we streamline forms and applications and make them a lot easier to use. So it involves something like Watching you fill out that tax form and seeing where you get stuck, seeing where you have questions, asking questions like, does this question actually need to be on this form? How do we make it easier for the user? But when you assume that every single interaction with a member of the public needs to go
18:39 Speaker 01 through OIRA, then your user research needs to go through an OIRA approval process. And because that process can easily take nine months and sometimes longer, you essentially can't do user research.
18:53 Ezra Klein? But this is weird because you point this out in the book. It would seem from a plain reading of the law that user research is exempted.
19:01 Speaker 01 Yes.
19:02 Ezra Klein? And yet?
19:03 Speaker 01 And yet, there are examples of this going on every day in federal government now. There's also examples of people getting through it and not being required to get OIRA approval for user research. But essentially what happens is, in fact, a friend of mine whom you know, Erie Meyer, who's at the Consumer Finance Protection Bureau, had this happen to her the other day. She was already doing user research and they said, oh, no, we're going to have to send this through OIRA approval. I mean, they were already doing it, which is the crazy thing. And essentially, the person in the agency decides that better is safe than sorry. It's better to be safe than sorry. And says, I understand your interpretation, but it's my butt on the line if someone else
19:47 Speaker 01 calls a foul on this, so we're sending it up to OIRA for their review. Now, OIRA could say, obviously, this is wrong and silly and there's no need to get our approval, but they also take a better safe than sorry stance, again, because of the incentives to them. And the result of Erie's request was that they said, this will have to go up on the federal register and we will have to get comments from the public about whether these people can get comments from the public. Oh, and in the meantime, there'll be no user research or no comments from the public happening until this process is over.
20:22 Ezra Klein? I just want to note that everybody in this story is a liberal who wants government to work well.
20:27 Speaker 01 Yeah.
20:27 Ezra Klein? I just think that's important to say here because it's something you say towards the end of your book, that working in government is a really valuable thing to do, in part because it will change your view on what sometimes makes government work poorly. And one of the almost spiritually difficult things about reading your book, if you're somebody who does care about government, is knowing that this is a lot of people who are doing their best, who have, in many cases, much more difficult jobs, who require much more endurance and creativity to serve the public than, you know, anybody in the private sector does. And yet, we're getting this outcome somehow, right? I've known a lot of people at the top of government, the people who work in, you know,
21:10 Ezra Klein? offices in the West Wing, and they're all very, in some conceptual way, aware. The government is inefficient, they don't want it to be, right? Bill Clinton had Al Gore do reinventing government, Barack Obama hired the first CIO to look over the government, chief information officer, right? Joe Biden is somebody who, when he was a senator, often talked about government inefficiency. Their chiefs of staff are impatient, management-oriented people.
21:34 Speaker 01 Yeah.
21:34 Ezra Klein? And yet, somehow, from the top, even though they talk about this all the time on the campaign trail, like, no Democrat wants to be a pro-inefficient government Democrat, this does not filter down somehow. Why?
21:49 Speaker 01 I think you have to look not just at the executive branch and the administrative agencies, though you shouldn't not look at them. I mean, I think it essentially comes down to you come in with a passion for cutting the red tape and making this easier, and two things happen: you get worn down, and you have other priorities. It's really, really hard to spend all your time on this, and you're not in control of it. When I worked in the White House trying to stand up what became the United States Digital Service, we were originally going to do it in GSA.
22:23 Ezra Klein? Which is the...
22:24 Speaker 01 General Services Administration.
22:27 Speaker 01 And the head of the GSA is a fantastic public servant named Dan Tagarini. And he really, really wanted this unit in his agency. And he was actually blocked by a guy below him who kept calling out that sort of better safe than sorry on some very obscure interpretation of how the money that we were going to use to stand up the unit could be used legally. I mean, it was wrong, but he could construe the language that way. So you have to actually have some empathy for people at the top who are fighting as hard as they can to get stuff down and are blocked by people not just above them, but around them and below them.
23:08 Ezra Klein? There's another dimension here that you point out, and it's a sociological observation about government that as soon as I read it, I realized it's completely true, that knowing a lot about politics is very high status in government. If you're considered like a tactical political genius, that's great. Knowing a lot about policy, having very big policy ideas, that's also very high status. Implementation isn't.
23:33 Speaker 01 No.
23:33 Ezra Klein? Why?
23:34 Speaker 01 I think it goes back very deep in our culture. I mean, I think that's not just true in government, let's be honest, right? Like, big ideas people get a lot of status in our society.
23:43 Ezra Klein? Thank you for listening to my TED Talk.
23:46 Speaker 01 Well, yes, big ideas guy. A friend of mine who worked in the government digital service in the UK pointed me to this idea that's written about the British civil service, that it is literally divided between the intellectuals and the mechanicals. Now, I think that the digital age has complicated this in a lot of ways. If you look at, say, metaphysical Silicon Valley, if coding is a mechanical task, it's the way you implement something, like, all those companies were founded by coders. Like, mechanicals in that framework are sort of at the top. And lots of interesting, good and bad things have come out of that culture, which is really distinct from the culture of government, where the intellectuals are at the top and have the power.
24:31 Speaker 01 And the people who write the code or do the designs that implement these programs are really, really, really far down. Like, they're at the very bottom.
24:40 Ezra Klein? In a way, they're not even at the bottom, they're outside.
24:42 Speaker 01 They're outside, this is true.
24:43 Ezra Klein? Tell me about Circular A76.
24:47 Speaker 01 So a couple decades ago, the federal government said, this is actually coming out of OMB, the Office of Management and Budget, but it really derives from a long history of trying to streamline government, said, if someone else can do your job in government, we're going to outsource it. It's the reason we have things like the concession stands in national parks are run by third parties, like they're not run by the government. That's all A76. And there's a real valid justification for some of this. There's a famous case in which, I think during World War II, government decided to make its own steel because there wasn't enough. Well, by the time they finished that, it was wildly expensive and very bad. And it's a commodity. The market can provide us good steel. That
25:31 Speaker 01 is the best way to do that. We should not get in the business of steel. And so on that basis, they sort of roped a lot of other things into that, like concession stands in federal parks. There's no reason that's inherently governmental, which is an important phrase that very few people talk about. Because the distinction between something that is inherently governmental, like a government person must make this decision, i.e. most policy decisions, and something that is inherently commercial is like, okay, well, we'll make the policy decision and then we'll contract it out to somebody who just delivers on it. And it creates this real separation between policy and implementation, which has a lot
26:13 Speaker 01 of complications to it. And I think over time has sort of snowballed into some really bad outcomes because we're trying to do policy over here, implementation over here. I have my hands very far apart, since your listeners can't see, and so much gets lost in the middle.
26:28 Ezra Klein? So that's in a way the guiding principle here. You say it's gone now really far. I mean, you go quoting someone else as far as really saying that civil servants manage, they don't implement. So particularly in the digital sector, tell me about what the government does and what it contracts out. Like, what functions in your experience are done by somebody on a public payroll? And what functions are done by somebody who is being paid out of public taxes but is on a private payroll?
26:58 Speaker 01 So let me speak to what's still pretty standard but changing, but like when I was in government in 2013 and 2014 was really, really standard. If you're trying to get, say, healthcare.gov built, and you're the digital person responsible for it, what you will actually do is work with contracting officers, people to gather a bunch of requirements, lawyers, and you will have to be master of an incredibly complex apparatus of procurement rules and operations and forms that need to be filled out and processes that need to be followed that actually is a very robust skill set and knowledge base in its own
27:42 Speaker 01 in order to hire people who will actually do the digital implementation of whatever it is you're doing. So I talk about somebody in the book who's in fact a fantastic digital leader and gets really good outcomes because he is a good digital leader. But like 90% of his time is spent on procurement and legal.
28:02 Ezra Klein? So I want to talk about how procurement actually works because just like it is different to hire and fire in the public sector because of civil service rules, it is different to procure, to pay an outside vendor to do something in the public sector because we've a lot of rules meant to clamp down on cronyism, on patronage. We're trying to make the process fair and we make it cumbersome. One of the things you talk about a bit is the way the system can get slowed down by vendor protest. That if you don't win a contract, you can actually protest the entire process, holding it up, sort of like if you are upset about people building something, you can bring an environmental quality act objection. Tell me about these protests.
28:47 Ezra Klein? Tell me about the ways in which the vendors actually have quite a lot of power in the process.
28:52 Speaker 01 Yeah, these rules were designed to make the process fair. So for instance, an agency will announce a winning bid. And if other companies who bid on this think that either something genuinely went wrong or more likely that they can simply take some advantage out of this, they can protest the bid. And because these bids are so wildly complex, I mean, the paperwork required to bid on a major government project, whether it's technology or anything else, is massive. There's a lot to review. And There's very little evidence of actual competency that are required in this paperwork. It's all evidence of compliance. And so it's very easy to find something that's not compliant.
29:37 Speaker 01 But the result is easily six months, often a year or two delays in getting something done.
30:10 Ezra Klein? I think actually, the first time we started talking, I became more aware of your work, was in the healthcare.gov debacle.
30:15 Speaker 01 Mm-hmm.
30:16 Ezra Klein? And at the core of that debacle was procurement. And who got chosen to build that? So how did these rules end up affecting who got that contract?
30:27 Speaker 01 If you look at the timeline in the ACA, and I'm going to forget the exact times, but essentially, they only had sort of three years. The Centers for Medicare and Medicaid Services, CMS, had about three years to get healthcare.gov up and running. They needed to launch it by October 1, 2013, if I recall. And as I just described, the process of getting a bid together that's incredibly complex, getting all your requirements, getting that contract out to bid, reviewing the bids, and then awarding it, and then potentially protests, that's easily three years right there. So they had to do something else. So CMS actually had some vendors on a very obscure vehicle, we call it, called IDIQ.
31:12 Speaker 01 It's stands for indefinite delivery, indefinite quantity, which is essentially a way of saying, "I've hired you to be around and do what I need."
31:22 Speaker 01 Actually, an IDIQ can be a very good thing to have someone just onboard.
31:25 Ezra Klein? They're basically on retainer.
31:26 Speaker 01 They're on retainer, but like, yeah, for a wide variety of things. And so CGI was one of the companies that was on an IDIQ already at CMS. So they didn't have to bid certain pieces of it out. Now, let me be clear, there were 60 different contracts to vendors to build healthcare.gov. Some number, like 12 of them, went to CGI, which is considered the main contractor, and others went to others. So it's not like they had the whole thing. But they really couldn't have done that full process where they bid it out. Because if anybody had protested, they would, had no hope of making their congressionally mandated deadline.
32:07 Ezra Klein? I think it's worth spending a minute here because I covered the drafting of the Food and Care Act really closely. I covered how they were thinking about building the insurance.
32:16 Speaker 01 Oh, and you missed this.
32:16 Ezra Klein? And I'm, not only did I miss this, but, but I can say with certainty that there was a lot of excitement in Congress about, we're going to create these online digital marketplaces and it's going to be a next generation website and be able to compare everything. And I'm sure in Congress's mind, having talked to people, the way they thought about this is the government would go for this unbelievably marquee project and get a really great website design player. I mean, Google doesn't do a ton of government contracting in this particular way. They do other things. But someone like a Google, right? Like just whoever's the best. And this would be a great project. And instead, you end up with a pretty obscure contractor, and to me, this is important, who, I'm not saying this is their only competency, but across a bunch of these contractors,
33:01 Ezra Klein? what they have is a competency in contracting. They know how to work with the government, which is a huge pain. And this feels like another place where the procurement rules and the way different regulations, et cetera, have been interpreted, it's really not what the legislature was trying to do. They wanted a great website. What they got was a total disaster. But they didn't think go back and change procurement rules either, really, to my knowledge.
33:28 Speaker 01 Well, the legislators would say, we have given agencies, and I'm not sure about this particular case at that particular time, but this happens in the Department of Defense all the time, we have given you lots of exceptions that you can use. There's something called an other transaction authority, which basically gets rid of a lot of those rules and lets you sort of choose who you want to choose in a streamlined kind of way, but they just don't get used. And that, again, goes back to, people say, better safe than sorry. If something goes wrong, someone is going to come back and say, what justification did you have for using OTA, other transaction authority, instead of the more robust way that we know will ensure, quote unquote, fairness? So it's really that culture that comes back to it.
34:13 Speaker 01 I mean, it's a reasonable example, I think, of what I talk about in the book, is culture eats policy, right? The policy that is designed by policymakers wants to get the outcome of flexibility for the agencies because they know that the rigidity is hurting the outcomes. But the culture just goes back to, I see that you've given us that, I'm choosing not to take it because I feel like it's going to make me safer.
34:37 Ezra Klein? So healthcare.gov was this huge project. It had to be contracted from top to bottom. But I also want to talk about the ways in which this kind of contracting makes it hard to update systems as policy is changing. And I want to use a pretty horrific example that you talk about in the book, which is the child separation policy at the border. So obviously, the actual policy there was terrible, but one thing that made it much worse was the rigidity of the system being used. Can you go through that?
35:08 Speaker 01 When the Trump administration ordered that kids and parents be separated at the border, the systems that tracked families could only assign one, what's called a number, alien number, to a family. That's how it was set up and had worked for years.
35:26 Ezra Klein? So just horrible. It's called an alien number.
35:28 Speaker 01 It's called an alien number, though they call it a number, which is also really confusing. So if a family of four comes through, they get one a number because in the past, they were treated as a unit.
35:41 Speaker 01 This horrific and cruel policy comes down. They separate them. They now give you one A number, your partner one A number, and your kids get A numbers separately. And there is no affordance for connecting your A number with your children's. They were pinning sticky notes with A numbers on infants who had been taken from their parents. And there were wonderful public servants who were writing down, keeping spreadsheets, trying to track the connections between parents and kids. It wasn't like there was no effort made, but there wasn't a systematic way to do it. There was a lot of chaos. You know, the person who's trying to reunite them is not necessarily the person who separated them in the first place.
36:24 Speaker 01 And so when, thankfully, that order was reversed, there was very little capacity within the Customs and Border Patrol to fulfill the demand to reverse it and reunite these kids because the software didn't have any affordance for making this connection.
36:41 Ezra Klein? So something you then talk about is it would have been a fairly obvious move, again, without either of us endorsing a terrible policy. But it would have been a fairly obvious move if you're doing that policy to change the software so that the A numbers are tracked together. Why didn't they? What would that have taken?
37:01 Speaker 01 So I should caveat by saying I don't know the specifics of the CBP contracts that govern this software. And I've never seen this software in action. But I do know that the way this generally works in government is if you want an update to your software, you take it out to bid, and that takes a really long time. And yes, the company that may have originally written it will bid on it, but you're not guaranteed to give it to them. Maybe that's the right thing, maybe that's the wrong thing. But if you have a new vendor, they have to get up to speed on the code base. But essentially, the way we think about software in government, it's a project. Like, you start it, you finish it, and then there's a little bit of money trailing along for you to keep it up and running.
37:44 Speaker 01 So when you have a change like this, that's frankly incredibly urgent, you don't have people who work on that software day to day, week to week, that are there, ready to change it and know how the codebase works. So in contrast, Google didn't create search and then lay off 95% of the staff and keep just a skeleton crew around it to keep it up. The way things that really work in the real world are staffed is on an ongoing basis, not only to make sure that changes in some connecting database somewhere can be updated so the thing doesn't crash, but also because we live in a dynamic world. Like, we can't expect there not to be policy changes that have to be incorporated.
38:27 Speaker 01 And frankly, we should be constantly looking at the software and saying, is it doing what it could be doing as well as it could? There's constant improvements in the consumer products that we use. So the curve of staffing in a consumer product is sort of like, it starts really slow, and then as the product becomes more and more clearly valuable and popular, the staffing goes up and up and up. It's just this like constant curve up. That's how the software is constantly better and more usable and more valuable to us. In government, it's this weird sort of like, there's nothing, then you hire all these people, and then there's this couple-year period where lots of people are working on it, and then it drops to almost zero.
39:09 Speaker 01 So after almost zero, you have no capability to change that software. You have acquired software, but you have not acquired the capability of doing what that software is supposed to do, of performing the function on an ongoing basis that that agency is responsible for.
39:27 Ezra Klein? I think this brings us back to healthcare.gov, which is a real, I think, before and after moment for how digital is viewed inside the government, because it is built and broken by this contractor, a set of contractors. And the rescue effort is a much more internal process. It is built on now the idea that you actually need internal capacity and flexibility. And because the president is focused, there's a lot more ability to move quickly and make decisions. So tell me a bit about the principles and staffing that the rescue effort is built on.
39:59 Speaker 01 The rescue effort came at a time when we were trying to prove the point that you needed internal capacity. So as disastrous and horrible as it was, it was actually really useful for me and others who were saying, hey, we can't just have procurement competence in government. We have to have some digital competence in government. In fact, we need it in the halls of power in the White House, in OMB. And I think one of the reasons it worked is that they could see through the rescue effort that we could bring in people whose competence wasn't how do we contract with company, but how does the software actually work and what's wrong with it. So the people who came in were software engineers. Many
40:41 Speaker 01 of them happened to have this title at Google and other companies called site reliability engineers, SREs. You've never heard of an SRE in government. There literally never was one. That was always the responsibility of an outside company. So you had people who could actually look at the code. And in fact, we found that oftentimes the people at the vendor who were supposed to have done the coding didn't really even know how to code. So you just had all these people who knew how to work with government, but didn't actually know how to make the site work. And bringing in folks who could actually get in there in the weeds and was the thing that changed it. The process up to that point
41:25 Speaker 01 had relied a lot on planning. There's this real focus on planning in government for lots of jobs. If you want to get promoted, you have to have this thing called a PMP, a project management professional. And it's all about upfront planning. But it's really hard to plan in a situation like healthcare.gov where they are still finalizing the regulations and rules that govern it, right up to the launch of the site. So the skill that was needed was not this very traditional plan everything, then execute to it. You needed people who were basically good at sort of what's going on with users? Is it working? Just go find problems and solve problems at a very detailed technical level, instead
42:09 Speaker 01 of looking at project plans that were not actually connected to the experience that users have.
42:15 Ezra Klein? One thing you say about the healthcare.gov experience is that There were a lot of reasons it went wrong. And, you know, there are GAO reports on it.
42:25 Speaker 01 Very good ones.
42:26 Ezra Klein? I wrote a piece for Business Week about it back at that time. And, you know, I mean, there was a huge amount of attention on this. But you say if you could boil it down to one, it's that there was a lot of process management, program management, but not a lot of product management, not a lot of people willing to say, no, we're not going to do that, empowered to say, no, we're not going to do that. Can you talk about that distinction?
42:49 Speaker 01 So, project management is the art of getting things done. It is hugely important. I value good project managers very highly, and everyone should. Product management is the art of deciding what to do in the first place. And if we don't do that, then the project managers just are overwhelmed with a sort of undifferentiated, unprioritized mess of requirements that they have to fulfill. And, yes, healthcare.gov, there was a lot written in the law about what needed to happen. It was probably overspecified. But there was also not a willingness to sort of even imagine that you might do less. In fact, people were told, you can't do less. You can't not serve every single edge case from day one.
43:34 Speaker 01 For instance, that's illegal. There's reasons why they think it's illegal. I do not think it's illegal, and I think we've proven since then that it's not. But that willingness to say, it will launch serving the most common set of users because that way we can handle the people who can just use a website to get their health insurance. And then the people who have crazy exceptions, or rarer exceptions, are going to go through the call center, they're going to go through those in-person service centers that they've set up around the country, which makes sense. Those people actually probably need a person walking them through it because of their exceptions. It doesn't mean that the website couldn't eventually serve all of those people.
44:16 Speaker 01 It means that to get a site that launches and works on day one, you constrain what you're trying to do, get those first people through the door, test the site that way, and then you add the functionality later. That's good product management. And until recently, it literally hasn't existed as a discipline within government.
44:37 Ezra Klein? There's both product management as a discipline, and I think there's also product management, or what in this case I would call prioritization,
44:42 Speaker 01 Yeah.
44:43 Ezra Klein? as a political value, as something the leadership values. And you tell a story in there of a conference call while the site is in crisis with Marilyn Taverner, who is head of CMS, the Centers for Medicare and Medicaid Services. Could you tell that?
45:00 Speaker 01 Mina Shung was somebody who joined the, what we'll call rescue effort, though I do try to stay away from that term because there are so many great public servants who were part of the rescue effort, and they often feel unseen by that. But Mina Shung came in to help. And one of her responsibilities was to staff a daily call with the administrator, Marilyn Taverner, and others, including the main contractors. And the site was in total disarray. And I don't remember exactly how long this was in, but this was before we were headed towards, OK, it's starting to work again. And one of the things that she brought up on this call was that in the original spec, there was going to be a separate site in Spanish.
45:45 Speaker 01 Makes sense. I think government services should absolutely be in as many languages as we can to serve people as well as we can. But at that point, even the English-language site was very far from being able to work. So Taverner brings this up on the call and says, where's my Spanish site? Now, there's all these vendors on the call, so Mina knows that they're going to try to placate her and say, we're going to get to that, and then divert resources to the Spanish site. So she gets ahead of them and just jumps in and says, I'm sorry, administrator, but you can either have one site that works or two that don't. And Taverner hangs up on her. She doesn't want to hear it. And God love Mina, she just kept persisting. But her job at that point
46:30 Speaker 01 was to keep the team's eye on the ball, which is, we have to get this site to the point where it's operational, and then we can use those resources to make a Spanish version. But it would be pointless to make a Spanish version when this has something that didn't work.
46:46 Ezra Klein? One thing that is a worrying dimension of that story is that, and you see this in government a lot, that when things are in crisis, people realize they have to bend rules, they have to prioritize, you know, they can't do everything at once, they have to make the first thing work first. But a lot of things aren't in crisis, or, and I think this is actually more often the case, they are, but they're not in crisis for the right people. Which is to say, I think there's a lot of attention to delivery and implementation when the people being delivered to and implemented upon are middle class, are upper middle class, are rich. So healthcare.gov was affecting so many people simultaneously. If you think
47:28 Ezra Klein? about the unemployment insurance failures during the expanded unemployment insurance during COVID, that was affecting a lot of middle class people who had political power all at once. There are obviously a lot of programs where they work very poorly for poor people in an ongoing way. But they're not a crisis. And so the ability to say the administrator, or to the secretary, or to the president, or whomever, if they even care about this, hey, we're just not going to do that because this crisis is so big, we have to do the first thing first, isn't there. And so I worry a bit about that. I'm curious to hear your reflections on that, how things work sort of in what you might call implementation peacetime versus implementation wartime.
48:11 Speaker 01 Yeah, we don't get as much done in peacetime. And I think one of the tactics of people who I respect in this community is to make it a crisis, right, to make the crisis visible. I'd argue that it was a crisis that very few people were getting enrolled in SNAP, but it was not hitting the headlines, for sure, for all the reasons that you mentioned. And some great people at Code for America sort of made that a lot more visible and change got to happen. I talk in the book about what was certainly a crisis at the Department of Veterans Affairs, where veterans couldn't sign up for healthcare, but it was completely invisible to the people inside the building because the form that was broken, if you tried to use it from any computer at a library or at your home, wasn't set
48:54 Speaker 01 up with the same combination of Internet Explorer and Adobe Reader that the computers inside the building were set up with. And so the form worked fine for them. And the tactic that the team used to finally get this fixed was to record a veteran trying to use the service. And show the people in power how humiliating and frustrating the experience was for veterans. But until they were able to bring this veteran named Dominic's experience to life inside the building at the Department of Veterans Affairs, they were literally told there was no work to be done because the requirements in the contract had all been fulfilled and all the boxes had been checked.
49:38 Speaker 01 So you've got to sometimes make the crisis visible when it just won't be because the people who it affects don't have a voice.
50:09 Ezra Klein? One thing that comes out of the Veterans Administration story you tell is the way that what a crisis is to the people using government services and what a crisis is to the people making them can be very different. And I'd say throughout the book, you're very sympathetic, even to people you're at cross purposes with. You're trying to get something done and they're kind of standing in your way and you tend to see where they're coming from. But in this particular story, in the veteran's story, you talk about somebody you call Kevin. And so the attitude that he and his team have developed towards development, towards keeping themselves out of hot water. Tell me about Kevin.
50:46 Speaker 01 Kevin was also at the Department of Veteran Affairs, but this is actually a separate interaction. I had started working with a small team to understand problems with the Veterans Benefit Management System, which is sort of the internal way that you would process people's requests for benefits. We had been told that there was terrible latency in the system. Essentially, you can click on a link to start processing something and you'd wait for a really long time before the page would load. It's very frustrating. We met with the guy the first day and he said, I'm so glad the White House decided to send people to verify that everything is okay because latency has been solved. Now, we found out later, he solved latency by defining it as over two minutes.
51:28 Speaker 01 So if you clicked and had to wait two minutes or one minute and 59 seconds for that page to load, you were not to report it as latency. So that was the solution. That was my first clue that this man I called Kevin was more interested in how things looked than how they actually operated. But as I asked him questions about this system, why had they decided to do X instead of Y? What role did the claims processors have in this particular part of it? He kept saying, that's not my call. I don't know that. And I asked him why. And he said, look, I've spent my entire career training my people not to have an opinion on the business requirements. If they ask us to build a
52:10 Speaker 01 concrete boat, we'll build a concrete boat. And I said, good Lord, why? And he said, because that way, it's not our fault. And honestly, that moment for me, I was seven months into working in government then. I'd worked at Code for America before. I'd worked with government. I was not new to the sort of dysfunctions of government, and I shouldn't have been shocked. But I really did take it as a gut punch. I didn't blame him in a certain sense, because I knew what he meant about the way we structure responsibility in government. But at the time, I think 18 veterans a day were committing suicide.
52:51 Speaker 01 Many of them were still waiting for their benefits. And it just seemed like such a callous thing to say, when if he would push back a little and say, this concrete boat is never going to float, we could have gotten more veterans their benefits.
53:09 Ezra Klein? When I finished the book, I had this note to myself that this book is about personnel, capacity, and autonomy. That it's about who works in government, what kinds of things government does versus doesn't do. And then the autonomy, the agency, the power we give these people to do the jobs we ask them to do.
53:30 Speaker 01 That's your TED Talk.
53:30 Ezra Klein? Yeah, there you go. My TED Talk, a summary of Jen Balka's book. But one thing that I have always worried about, being close to a lot of people who work in agencies, and I think that you see in that story.
53:43 Ezra Klein? Is when you make it really hard and annoying to work in government. It's not that it pays a little. You have a good section on this. It's not the government pays so little, but it is often very frustrating work. You have to deal with all these weird procedures and the lawyers tell you can't do it. I mean, working on code in government, and this is true in a lot of your stories, including the stories of your heroes, often requires an almost superhuman level of persistence. You're trying to get around a lot of very annoying things, and sometimes you get a hero out of that. But it also, over time, drives out a lot of good people who just don't want their days to be annoying. And it also selects for people who can make the internal compromise in that story Kevin made, where you sort of shut off the part of you that says, that doesn't make sense, and I'm going to argue.
54:28 Ezra Klein? You lose enough fights like that. The lawyers tell you no enough times, you get yelled at by your superiors enough times, and either you leave, or you say, I'm going to stop fighting. And that's where things can go, I think, really badly. That it's not just that we need good people to work in government. You hear that all the time. But we need government to be a good place for people to work. And for a lot of creative and patient people, particularly those at mid-levels, not those who are, you know, appointed to run an agency or run an office in the White House or something, it gets very frustrating. I've seen a lot of really good people go in, and particularly when they're in older agencies that already know how they do things, I've watched them go right back out. And I don't think they weren't committed. I just think people, it's hard to have a really irritating job.
55:13 Speaker 01 I really think that we're so focused on getting these outcomes from a policy sense. And to solve the problem you're talking about, we've got to really focus on just basically the overall health of the civil service. Yes, it's frustrating to fight these battles. I think there's increasingly, certainly in my experience, like, a community of people who support each other around it that really makes it tolerable. When people stay, it's because they've got someone's shoulder to cry on when they need it and people to tell them, keep going. I have found that incredibly valuable and incredibly meaningful to me to be part of that. But like, for instance, one challenge they face is they also really can't hire
55:56 Speaker 01 other people quickly. And there's a lot that we-
55:59 Ezra Klein? Or fire people.
55:59 Speaker 01 Or fire people. We could do so much, and by we, I mean more our elected leaders and how we hold them accountable, to making that environment less destructive, less wearing you down, and giving public servants, like those who are high-profile, who are just persistent and heroic, the tools they need to succeed. But we just don't think of the tools of legislators or the legislative branch as being good at that, right? They have money, they have rules, and they have oversight. And those things aren't working very well. But if they targeted them at sort of the health of the civil service and really focused on it, I actually think we could make that a place that doesn't drive people away as much.
56:43 Ezra Klein? Well, tell me what that would look like.
56:45 Speaker 01 So, for example, there are bottlenecks in the system that are just sort of vestigial. Like, OPM requires that there be industrial organizational psychologists that approve a bunch of things, the same way that we have OIRA as a bottleneck to getting user research out the door. Remove it. But nobody's on that. Nobody's looking at these little things like that, that one by one could actually speed hiring and make it easier for people. But there's probably 7,000 little things like that. And one of my complaints, since you started on the left, right, is that the right actually has a lot of plans around this that are focused on firing people. The left really doesn't pay much attention to
57:28 Speaker 01 this. They're not obsessed with the details of making the civil service have greater autonomy and freedom and hiring the right people, and yes, probably firing the right people. I actually want the right and the left to say, this is actually more important than this, that, or the other policy outcome that we're trying to get, because it'll enable all of those policy outcomes. But it's real work that, like, somebody has to fund and somebody has to do. I know probably half a dozen things that would go in that, but somebody has to go find the other 36.
58:03 Ezra Klein? My impression of this area of politics is that it's become, as many things do, very dysfunctional through a kind of crude polarization, which is to say that civil service reform, by which I mean making it easier to hire and fire, giving the civil service more autonomy, but also making sure there's more accountability, including accountability from the political system itself, because it is often, parts of that have often been a feature of the right wing attacks, and often they are attacks that also want to outsource a lot of functions of the civil service and starve it of money and so on. They've been coded as right. So the way to defend and show you care about government is to protect it from these attacks. But if you then cover government, or you actually
58:46 Ezra Klein? know people who work in it, even people in the civil service don't like how this works. I mean, the difficulty of listing a job to get anyone in, we're so worried about patronage, we're so worried about corruption, they've made the whole thing work really poorly, and we drive people out because they just cannot stand it over time. And many of the people can, we're either asking, again, superhuman work out of them, or we're accepting people who maybe are not the people we wish were in charge of some of these things. And that's not to take away from a lot of people doing hard work, but it's also a kind of fallacy, if you've ever worked in any kind of organization, to suggest that everybody in an organization is the right person at the right place. I think the left has a lot of trouble with
59:30 Ezra Klein? this kind of statement, but it is important to be able to fire the right people to make organizations work. It's not pleasant, it's not what anybody should want to do, but to not be able to do it is a way to guarantee organizational dysfunction.
59:42 Speaker 01 I agree with that. I think the order of operations, though, should be first, we make it easier to hire.
59:47 Ezra Klein? Yes, no, totally. I think the reason this all gets stopped on the left is they're so afraid of the firing conversation. That's why I bring this up.
59:54 Speaker 01 They are, and there's some good reasons for that, right? So if Schedule F comes back in a second Trump administration or...
1:00:00 Ezra Klein? Say what Schedule F is.
1:00:02 Speaker 01 Yeah, so Schedule F was something that Trump instituted, I think, in the last couple of weeks of his presidency, that would allow him to arbitrarily reassign someone from a civil service position into a political position. What's the difference? Obviously, you can fire a political position, a political appointee, with very little paperwork. It's just, it's quite easy. Whereas a civil service person, you're supposed to not be able to fire unless there's, you go through, jump through a whole lot of hoops. I guess what I would say first is both people on the left and the right who work in government, and of course, civil servants are meant to be non-political, are frustrated with the same thing. And I think that polarization is really sad because we're actually responding to the
1:00:45 Speaker 01 same frustration of this sort of demo sclerosis, things moving too slowly. But on a very practical basis, like people I see working in government are now trying to solve problems. Yes, they have these obstructionists around them saying, sorry, this is how it's always been done. The ESP has to stay in the software. We're going to build a concrete boat. But the first way to combat that is to get other people in to balance them out. There's a concept always in government and much theory about government, which is you're trying to balance between agility and stability. And bringing in more people with an agility mindset is a good way to balance that, whereas
1:01:26 Speaker 01 just firing everybody is a great way to potentially, in another Trump administration, enable authoritarianism. So there's a reason they're upset about it. But I think that if the left doesn't start acknowledging that we have the same frustrations with the civil service and the administrative state as the right, we're never going to get anywhere. And I actually spend a lot of my time with people on the right with whom I agree a lot more than I thought I would.
1:01:54 Ezra Klein? I want to disentangle some thoughts here, at least for me, which is one, I think people go too quickly to the idea that if you created a lot more autonomy and flexibility within the civil service, made it easier to hire, easier to fire, that has to mean it's easier for the political system to hire and fire, which I do think is something we should be at least a bit careful with. Whereas I'm talking really about the ability of the civil service to hire and fire internally unto itself. Which is a different question already because they're just different rules. That's why in Schedule F, Trump was trying to move people from civil to political. But you could change the rules for what the civil servants can do internally and still keep, you know, I think maybe you do want less wall than you currently have, but a
1:02:38 Ezra Klein? pretty good wall between politics and the civil service. And then the other thing is, I really am skeptical that the right and the left have the same concerns here. And there are places where procedures might converge. There are places where some of the frustrations might converge. I, for instance, see this a lot in a lot of the work I'm doing around permitting reform and kind of supply side issues, that often something that is important to me now is deregulating the government.
1:03:08 Speaker 01 We're not supposed to call it deregulating.
1:03:10 Ezra Klein? I know.
1:03:10 Speaker 01 But that is what it is.
1:03:11 Ezra Klein? But that is what it is. I want the government to deregulate itself oftentimes. And deregulation is coded as a thing you do to make the government less powerful on the right. But oftentimes, I think regulation is one reason the government is not able to achieve its ends. And so it kind of sounds to some people like I'm saying the same thing that the right is saying, but I'm not. I actually want a more active and capable government. And there are things you might want to do from a deregulatory standpoint that would get you there. There are other things you would want to do from a deregulatory standpoint that would not. So many people on the right do not want the EPA to be able to regulate carbon. That is a kind of deregulatory decision, but it isn't the one I would make. I would like it to be easier when the government wants to build a clean energy for them to get the permits done.
1:03:56 Speaker 01 Yeah.
1:03:57 Ezra Klein? That is a deregulatory decision that makes the government stronger. So I just think that's also true sometimes within the civil service debates, that I would like to see a more capable, stronger, I'd like to see more people employed in many cases by the government. We've outsourced a lot of functions, which is not oftentimes what the right means when they say they want to do civil service reform. I don't, they often don't mean they want to bring a lot of these functions back into government purview.
1:04:21 Speaker 01 Let me say quick, three quick things to that because it's a really important point, and it does get. Gnarly.
1:04:26 Ezra Klein? Yeah, totally.
1:04:27 Speaker 01 One is, yes, but I also think change happens when people with different views do have something in common, even if they're going to use it for different ends, right? And I think we probably have some fertile soil in there to get change done, even if people on the right want to do it, want that change for a different reason. Obviously, we need to be careful about that. I agree, we should definitely be regulating pollution.
1:04:50 Ezra Klein? Oh, it's sort of the last thing I was going to say on that, which is that there's the right and there's Trump.
1:04:53 Speaker 01 Yes, exactly.
1:04:54 Ezra Klein? And there are people on the right you really can have this conversation with, it's just not Trump.
1:04:58 Speaker 01 So that was my third point. My second point was, it isn't regulation or not regulation, it's how we do it. And when we talk about competencies and capacities in government, I think one of the main points I'm trying to make is, that has to be a core competency in government, and we don't even recognize it.
1:05:18 Ezra Klein? I really take the big point of the book is saying, there is politics, there's policy, and there's delivery. And right now, delivery is so subordinate to the people who run politics and the people who run policy, that in addition to delivery being done poorly, the people delivering often don't think they have any power. They don't have power to interpret the language they're given, they don't have power to say no to the lawyers, there's nobody empowered over a process to be flexible as it is ongoing. They often don't have the power to change contracts quickly. And that shift of power from policy or politics to delivery, that feels to me like the great
1:06:03 Ezra Klein? revolution you're trying to kick off here. And I'm curious, if you had two or three places you could start kicking it off, right, aside from everybody reading your book, and they should read your book, what would you do? Because, yeah, culture can be policy, but policy can also create culture.
1:06:18 Speaker 01 I think there's a lot of things that can happen that fall under the banner of where we pay attention and where our elected leaders pay attention. So, one really practical thing, and I have an example of it at the end of the book, is that when we're writing law and policy, who's at the table? And I take very much your point about everything big old liberalism, where you already have too many people at the table, and that sounds like just adding someone else on. But if we want to make law and policy that are implementable, then we have to have implementers at the table when that law and policy is being written. And those people are likely to be, and hopefully are, much closer to the users of the system,
1:07:03 Speaker 01 and they're going to provide insights that make your policy better. But they're also going to start to shift that power balance that you talked about. So like, an example I talk about in the book is clearing of criminal records, where the way a particular law was written made it kind of impossible to implement at scale. But in the aftermath of that, you've got people with technology skills who spend time with the people who have these convictions and understand the constraints in their lives and how they're going to interact with it, at the table when decriminalization laws are being written. Those decriminalization laws are usually the ones that contain records expungement provisions.
1:07:47 Speaker 01 That's a huge difference. Like, people can barely believe that this is happening, that you've got coders and designers and user researchers at the table. This is at the state level, when laws are being written. So that's one thing. You know, there's a saying, "What we pay attention to grows." So, our elected leaders have this oversight responsibility as part of their toolkit. And we also have the whole apparatus of oversight in the executive branch, offices of inspectors general, the Government Accountability Office, GAO. And what do they focus on? Failures.
1:08:25 Speaker 01 What if they lifted up the public servants who were having success by empowering themselves to have that flexibility and made half of their work or more shining a light on them and saying, this is what we want public servants to be like, and we're going to reward them? I think it would make a huge change in the culture. And I do have several sort of policy prescriptions in the book about things like how money is allocated that I think are important to pursue. But in the end, I keep coming back to those policies fail if the culture doesn't change. So the ways we pay attention, what we talk about, who we elevate, I think those are all kind of reasonable levers and good places to start.
1:09:11 Ezra Klein? So I think that's a good place to end, too. So what are three books that have influenced you that you would recommend to the audience?
1:09:17 Speaker 01 The first is a book from 1973 with the longest subtitle I have ever seen. It is called Implementation: How Great Expectations in Washington Are Dashed in Oakland, or Why It's Amazing That Federal Programs Work at All, This Being a Saga of the Economic Development Administration, as Told to Two Sympathetic Observers Who Seek to Build Morals on a Foundation of Ruined Hopes. What do you think of that?
1:09:43 Ezra Klein? I love that because that's exactly how pamphlets in early American history and kind of like 17th and 18th century Europe sound. And I love those title formats.
1:09:53 Speaker 01 Well, you're going to love this book. It covers a project that started in, I think, 1967, where politicians and bureaucrats went and made like a big announcement about investment that was going to happen in Oakland, which is my hometown and somewhere I think you have lived. And everybody celebrated and then goes into excruciating detail in the same way some of my book goes into excruciating detail about why several years later, essentially, nothing had happened. And they really conclude that this separation between policy and implementation is the only reasonable explanation for the failure. And they pull up all the other obvious reasons you would think that it failed and really debunk them. And it's just, it was
1:10:35 Speaker 01 sort of like they were reading my mail back in 1973. So I really love that. The next one is Radical Help by Hillary Cottem, who is a Brit, looking at the social safety net, the workforce training system, the healthcare system. And she takes that concept I have in the book of user needs over government needs and just pushes them so much further in a way that is really inspiring and I think very critical of all of these government systems that we have today. I mean, it's in Britain, but it's very similar to the US. And I just think she moves the overturn window about what it would mean to truly meet user needs in government-funded systems.
1:11:18 Speaker 01 And I hope everybody who works in social services and healthcare would read it. The last one is a bit controversial, and it's not one I'm recommending in its content, but I do think people of all political stripes should read.
1:11:33 Ezra Klein? Retreats are not endorsements here.
1:11:34 Speaker 01 Exactly. Definitely not an endorsement. But Chapter Three of this new report called Mandate for Leadership: The Conservative Promise is essentially the blueprint for what the Heritage Foundation thinks the next Republican president should do in all areas. But Chapter Three covers civil service rules. It's kind of hidden in there, but it does say that they would bring back Schedule F, which we can debate, but I think is a bad idea. But it's ultimately the way I want to call attention to-- they are paying a lot of attention to how to reform the civil service. And I would like there to be counter proposals on the table, since theirs is so thorough and detailed.
1:12:16 Ezra Klein? Jen Palka, your book is called Recoding America. I really, really recommend it strongly to people. They really should read it. We did not cover a lot that is important in it. Thank you very much.
1:12:27 Speaker 01 Thank you, Ezra. This was great.
1:12:39 Ezra Klein? This episode was produced by Emma Fagau, fact checking by Michelle Harris and Kate Sinclair, mixing by Jeff Geld. The show's production team is Annie Galvin, Jeff Geld, Roche Karma, and Kristin Lynn. Original music by Isaac Jones. Audience strategy by Shannon Busta. The executive producer of New York Times Opinion Audio is Annie Rose Strasser. And special thanks to Sonia Herrero and Kristina Samulewski.
Previous Beyond the ‘Matrix’ Theory of the Mind Next What Communes and Other Radical Experiments in Living Together Reveal Page updated October 26, 2026 · transcript and summary are machine-generated
