Want to Fix the Government? Fix the Code

BIANNA GOLODRYGA, HOST: Next, a technology and how it is not being leverage well by the U.S. government. The former U.S. deputy chief of technology says that stacks of policy are holding officials back. In her new book, “Recoding America,” Jennifer Pahlka shares the struggle and how it is hurting the American public. She joins Walter Isaacson to offer potential solutions.

(BEGIN VIDEO CLIP)

WALTER ISAACSON, HOST: Jennifer Pahlka, welcome to the show.

JENNIFER PAHLKA, AUTHOR, “RECODING AMERICAN”: Thanks for having me.

ISAACSON: You’ve got this great book, “Recoding America.” And I would like to walk through how you got there. I remember back when I was — “Teach for America,” you started a “Code for America” and that put you on this path. Tell me about that path.

PAHLKA: I was working in the tech media world and it was around 2008 or 2009 when we were getting this new president elected based on his ability to use the internet, and it really started —

ISAACSON: We’re talking about Obama there

PAHLKA: Yes. I’m talking to President Obama. And I was working on event called Web 2.0, which was about this new kind of web, the participatory web that everybody got to be involved with and things moved quickly with lightweight applications that were easy for people to use. And I started saying, hey, you know, if it can help this guy get elected, can help him govern better? And that’s how we got involved with this whole idea of Gov 2.0, which is what inspired me to start “Code for America” and get people from the tech industry into government and helping it work that way, where people could participate and it would be easy to use.

ISAACSON: And that helped get people into government, sort of as a year or so of service you’d for a community or a city in government in coding. But then, you went on Todd Park, the chief technology officer under President Obama, and you created the U.S. Digital Service. That was a lot more than just a year of volunteer.

PAHLKA: Yes. I think my journey into government has been one of learning quite a bit. I loved what we did and I believe the original ideas of “Code for America” didn’t quite account for how much, you know, the tech industry could help government, but we really needed to learn about government, and I think government now is in a place to help the tech industry quite a bit. And part of that to learning was when Todd Park recruited me to come to the White House. And I had seen what was going on over in the U.K. where they had created something called the Government Digital Service, which, yes, was much more than a one-year service program. It was the great technologists and designers and content strategists working at the center of government, really changing how the U.K. operated and how the government could communicate with its citizens. And so, that was the inspiration for the U.S. Digital Service. And also, I think, out of that we got another wonderful group in the General Services Administration called 18F, and both of those groups are now groups and agencies around federal government. And also, many states have these groups that do digital service work, which is just a little bit distinct from sort of the legacy wave that government has done I.T. operations.

ISAACSON: One of your other stints in government, you and I were both on the Defense Innovation Board with Eric Schmidt as the chair, the former CEO of Google. And I think we were all kind of shocked that the Defense Department itself, where you’d think it would be the best, they were just having trouble implementing things. Tell me what you learned there.

PAHLKA: The dynamics of these very long procurement processes that focus on gathering oftentimes thousands and thousands of requirements, and results and software for people who work in government and people have to touch governments, that really doesn’t work well, that’s just so over scoped. Those dynamics, and various other human dynamics, were present to exactly the same, really, in the Defense Department as they were in the USDA or states or cities. And so, it was really a great learning experience to see how these dynamics come about in places that look very different from each other. And yet, so many parts of government are suffering from this dynamic of, as I call it in the book, building concrete boats.

ISAACSON: That’s because it’s like — in your book, that’s a great scene because it’s sort of the requirement, or somebody has a request and they say, we’re never going to question requirements. So, they tell us to build a concrete boat, we’ll build a concrete boat. So, let’s start there, which is requirement. That seems to be at the heart of the problem of implementation. And so many people pile requirements on. Give me some examples of that.

PAHLKA: I work this summer — I’m sorry, two summers ago when — in COVID, on the unemployment insurance crisis here in California. And when we came in, they were in the middle of doing a procurement for a new unemployment insurance system. They had been about to bid it out, I think, when the pandemic hit and sort of took them off course. That request for proposal had, I believe, 6,700 requirements in it. So, they’re really just trying to recreate all of the complexity of the system that they have but on new technology.

ISAACSON: Wait, wait, wait, wait. 6,000 requirements? I mean, who puts these in and what are they talking about?

PAHLKA: Well, if you — I think the way to explain why there’s 6,700 requirements, and it may not be the exact number, but it was in that ballpark, is to you understand the complexity of the systems that they are trying to build, and I don’t just mean the technology systems, the legacy of policy and process and procedure that’s accumulated over many, many years. In fact, in unemployment insurance, if you think about it, that program was created at the 1935 Social Security Act. It’s been quite a few years since then, and that’s, you know, almost nine decades of people at the federal and state level, from the executive, judicial and legislative branches, throwing down new policy and process requirements and having them sort of land in this way on a program that then they start to really buckle under the weight of 90 years of new memos and new changes without ever going back and rationalizing it or simplifying it. And the way that really hit home, for me, when I was working at the Employment Development Department in California in August of 2020, when they had this backlog of what turned out to be 1.2 million claims that hadn’t been paid since March and April, was a colleague of mine was working with a claims processor to try to understand, you know, where were leverage points, why we were doing it this way, what we change?

And one of the guys she was talking to kept saying, I’m not quite sure about the answer to that question. I’m the new guy. I’m the new guy. And eventually, she said, well, how long have you worked here? And, he said, well, I’ve only worked here 17 years. The people who really know this process of these — processing these claims have worked here for 25 years or more.

So, certainly, the technology gets very complex and you get these, you know, wild numbers of requirements but — and it’s in part because we’re not doing the hard work of making choices about what to actually prioritize, but it’s also because we are dealing with 90 years of policy cruft and clutter that nobody has bothered to simplify.

So, I don’t think we’re going to get an unemployment insurance system that’s going to scale to meet the need in the next downturn unless we go look at the policy and process, not just keep, you know, blaming all these technology failures. The technology is just buckling under the weight of all that policy and process clutter.

ISAACSON: One of the things that stood out was the Food Assistance Program in California that you tried to help untangle.

PAHLKA: So, when we first started working, this was at “Code for America” in 2013, on food assistance. One of the forms that was used by — I think it was 28 of the California counties, had 212 questions on it. It did not work on a mobile phone. So, most low-income people access the internet through a mobile phone, but they would have to go find a computer at the library to use it, and it really was — the reason why so many people who were eligible for a SNAP in California weren’t getting it, they couldn’t get through this form. Even at the library, it would take too long. The library computer would timeout after half an hour and they hadn’t finish filling it out. And because of all of the processes and procedures that they had to jump through after filled out that form. And working with those counties to find a better way to do this, with fewer questions that worked on a mobile phone, involved really understanding that they — the way they made decisions about this was that all the counties got to vote on what features were going into this. And when you can all add but no one subtract, that’s when you get, you know, classic concrete boat with 212 questions.

ISAACSON: I’m writing a book on Elon Musk, and his rule number one is question every requirement no matter who gave it to you, whether it was the government, the military or whatever. And so, he’s become obsessed with that, unlike at Boeing where they’re trying to do the same type of rockets, but they’re following every requirement. There seems to be a danger in both approaches. In other words, ignoring requirements as well as following them to build a concrete boat, as you say. How do you balance that and give me some examples of where people could challenge a requirement and actually make it work better?

PAHLKA: Yes. I think that the idea around challenge requirement needs to connect to what was the intent of the law, what are we actually trying to achieve here, does this requirement get us to that goal? And I think the people on the ground have a better sense of whether that’s going to be the thing that gets us there or something else is. And really saying, OK, do we understand how the user is going to use this? So, being able to bring in the voice of the person, whether it’s a claims processor, you know, inside a government agency or the person making the claim, the person unemployed, those voices are very rarely represented when those 6,700 requirements are being put together. So, I give an example, not from unemployment insurance, but for Medicare, really amazing public servant named Yadira Sanchez who was around for healthcare.gov. She got thrown a whole bunch of really difficult work that helped healthcare.gov go from failing to succeeding. And in that learned about user needs, user research, more agile development, less relying on these requirements. And then you see her sort of evolve in her leadership to the point where she is looking at the requirements that are given her and saying, you know what, that’s not the best way to fulfill the intent this law or this program. So, for example, she’s asked to do these data extracts of pharmaceutical data to give to the ecosystem. It’s part of what Medicare does is they provide a lot of data to the provider world to help, so that they can, you know, evaluate their outcomes. And she says, well, those data extracts, with all the policies and processes we have, are going to take us nine months to get out the door. We have all this packaging to do, all this testing. It’s very expensive. And then, the people who want to use that data to make health care better aren’t going to get it for nine months, and they only get a certain moment in time.

There is a way to do this technically called an Application Programming Interface, an API, that will let anybody plug into that data in real-time in a safe way. And she says, let’s do that instead. Everyone is happier. They have better access to better data, it cost less in the long run and it works better. And then, they get much faster. But she is literally looking at a law that says to her very specifically, do these data extracts, and she’s doing something a little different. That’s the kind of boldness that perhaps Elon Musk might, you know, be very proud of.

ISAACSON: But it’s also the kind of boldness that can get you in trouble.

PAHLKA: Yes.

ISAACSON: In other words, if somebody messes up a product in government, they don’t get in that much trouble. But if they violate a step or a procedure or a process, they fear they can be, you know, indicted or something.

PAHLKA: Yes. There’s a lot of fear. And I call that in the book, the accountability trap. And it’s something, I think, we should all really understand and uses a basis to have some empathy for public servants. Because you are held accountable to both outcomes, like does the site work, and processes and procedures, and those two are very frequently at odds. And we’re asking our public servants on a day-to-day basis to make a choice between the two. But it’s the fidelity to process and procedure that their careers very often rely on.

ISAACSON: You know, one of the other sentences in your book I liked was, government leadership has typically seen implementation, in other words, getting something implemented, as a second-class job compared to the people who do the big policy stuff. And especially. I might say, Democrats when it comes to environmental things, they want so many processes and procedures before something can be done that that policy gets shaped by the people who can implement it.

PAHLKA: Absolutely. I think it’s important to realize that when I’m — all these things I’m writing about in the book where we end up with not the outcome that we intended, almost all of the things that are piled on to make it hard are really well-intentioned. I agree with the idea that we should check various environmental things before we do a building, but at some point, you have to say, for instance, well, we need to build infrastructure to electrify our country. And if we are not able to do that in 10 years because of all of the requirements and the process, that is actually going to get us to the wrong outcome.

ISAACSON: Is that sort of one of the problems with the housing crisis?

PAHLKA: Very much, yes. I live in California where our housing crisis is quite acute, and there’s become a realization that many of the procedures that were put in, for very good reasons, are now hurting us. They make everything take too long and get too expensive. And that is an excellent corollary to how we build government software.

ISAACSON: Tell me, and you do so at the end of the book, how do we get out of this mess?

PAHLKA: It’s not going to be easy. There’s no silver bullet. I do think that there are very specific things that each of us can do, and I think it starts with just thinking about implementation as the important part of our government’s functioning and what we hold our government leaders accountable to instead of just passing a law or policy.

I mean, we tend to celebrate when elected leaders get something passed. I was, for instance, really happy when we got the Inflation Reduction Act passed because it’s our shot at avoiding a climate collapse. But if we then don’t go and implement that, why would we — why did we pass in the first place? Why did we do all that hard work to get it passed?

So, I think we, as the public, need to think about implementation more than we think about what’s getting past or certainly more than we do today. And our elected leaders have to shift their focus as well. They are so focused on what policy can I say I got through Congress or a state legislature, and less on, what is the health of the bureaucracy that needs to implement this? And what can I do to increase that health, such that every law and policy that gets passed can actually be implemented well?

Because what really matters isn’t the words on a page, what matters is the outcome. Do we really feel it? Do we get this new electrical infrastructure? Do we have a medical — you know, a health care system that people really understand and can work? Do people who are supposed to get benefits actually get them? That’s where we need to turn our attention and that’s going to be, you know, the responsibility of people in administrative agencies, certainly the people who oversee them in legislative branches but also, the American people, because guess what, the elected leaders do what we hold on countable to, at least to some degree.

ISAACSON: Jennifer Pahlka, thank you so much for joining us.

PAHLKA: Thank you for having me, Walter. It was a pleasure.