Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.
Love it or hate it, regulation is one of the most common activities federal agencies undertake. Rule-making itself is among the most rule-bound processes, both to keep it under control and to ensure everyone’s voice is taken into account. Now a study, under the auspices of the National Academy of Public Administration, offers ways the regulatory process...
Love it or hate it, regulation is one of the most common activities federal agencies undertake. Rule-making itself is among the most rule-bound processes, both to keep it under control and to ensure everyone’s voice is taken into account. Now a study, under the auspices of the National Academy of Public Administration, offers ways the regulatory process can become more agile. NAPA fellow Michael Fitzpatrick joins The Federal Drive with Tom Temin to discuss this.
Tom Temin We should qualify you a little bit. You have worked in two administrations for the Office of Information and Regulatory Affairs, OIRA. So this is something kind of near and dear to your heart.
Michael Fitzpatrick That’s right I’ve spent six proud and busy years helping to coordinate regulatory policy in the Office of Management and Budget (OMB) and then have spent over a decade since in private sector, dealing with regulations from the other side. So I can bring both of those perspectives to the conversation.
Insight by Google: Granting access to, sharing and securing data in the cloud continues to challenge agencies across government. During this exclusive webinar, moderator Jason Miller will discuss cloud strategy, security and applications with agency and industry leaders.
Tom Temin And let’s begin with a definition of terms the study by NAPA, and we’ll get into a little bit of how the study was developed, but what is agile rulemaking? What’s the vision for agile rulemaking relative to the way it’s done now?
Michael Fitzpatrick Sure Tom, well as we get started, let me just say that all of my thoughts and opinions on our conversation here are mine alone, and not those of my current employer at Google. But my employer is very concerned with agile regulation, as are many companies. And agile regulation is, I think, an evolution of a concept that in past years has been called better regulation or smarter regulation. And it builds upon those concepts, which is how can regulators carry out their important duties and obligations to protect the public in ways that also balance against maintaining economic growth and job creation and innovation. And this is really the central challenge for regulators, is to strike that balance on complex, evolving issues. And agile regulation is a set of concepts and practices that we hope will allow regulators to become more nimble, more flexible, more innovative in their own regulatory designs and processes so that they can deal with what is becoming a more and more complex world that’s coming at them faster and faster.
Tom Temin So then it’s important to underscore then that the study looks not just at the process of creating regulations, which itself has issues, but also with the resulting regulations and how effective they are.
Michael Fitzpatrick Absolutely. In regulation, which you rightfully noted at the top is really the principal means by which the federal government makes policy on a daily, monthly and yearly basis. As we see Congress can be episodic at best in passing policies, but regulations come every day. And there are two fundamental components to making regulation. One is the process by which the agency engages with itself, with other members of the federal family — so White House offices and other federal agencies — and most importantly with the public. And by the public, I mean everybody: regulated entities and businesses, but academia and civil society, individual citizens. They all have to be part of the process. And so you have important process functions that need to be carried out under law under the Administrative Procedures Act and the various executive orders and then you’ve got, what’s the design of the regulation itself? What is this process yielding? What’s the substantive form of the rules that are being written, and both of those important components can be involved by agile tenants or practices.
Tom Temin And before we get into the tenants tell us how the study was done. Who was involved and how long it took?
Michael Fitzpatrick Sure it took almost a year. It was a combination of staff very informed and expert staff at NAPA, the National Academy of Public Administration, but also an expert advisory group — on which I was proud to sit — that included other experts from government, from academia and from the private sector, all of whom have extensive experience with the regulatory system, but also who are really dedicated as affiliates of NAPA to trying to earnestly provide recommendations and new designs and suggestions on how to make the system work as best it can.
Tom Temin We’re speaking with Michael Fitzpatrick, he’s Director of Global Strategy and Innovation at Google and a fellow with the National Academy of Public Administration. And let’s get into some of the top tenants. There are nine of them. What do you feel people should know first about what constitutes agile rulemaking?
Michael Fitzpatrick So Tom, the nine tenants are spread through four categories. One is the public need, understanding what the public needs with respect to regulatory protections. The second is regulatory design, which we’ve just discussed, which is one of the significant components of regulatory policy. The third is internal processes, which again, is that process of making the regulation and engaging with the public. And then fourth is continuous learning. Let me highlight a few at the front end here. First, let me take a quick step back and say the world is becoming more complex as we all know. We’ve got global pandemics, we have wars, we have supply chain disruptions because we have a thoroughly integrated global economy. We have climate change that we are grappling with. And in particular for regulators, we have the challenge of rapidly evolving and disruptive — and I mean that in all the senses of the word not just the negative sense of the word — just impactful technology, that is shaping society in many good ways, but also presenting challenges, and what we in the regulatory business called negative externalities, or risks and costs to society. And for regulators, this is a huge challenge because these technologies, and this technology development, is as we’ve said, rapid and impactful. It can be inscrutable and complex technically, and regulators face a real challenge in trying to both understand the technologies themselves, but also importantly, the impacts on society good and bad, and then how to design solutions that are still going to be relevant and appropriate in 5, 10 or 15 years when that technology will have evolved, and when its relationship with society will have evolved. So I wanted to get that out there as one of the fundamental reasons why we think it’s so important that regulators consider some of these agile tenants.
Tom Temin And getting to that one about public need understanding changing external conditions and evolving societal economic and environmental needs, that means that agility should also look at when regulations can be retired, for example. It seems like every other administration looks for retirements and sunsetting. And in fact, I think even the Obama administration on the Democratic side, also had a gambit to reduce regulations that were no longer applicable or were obsolete and it seems like agility would build that into the agency process without needing an executive order every four or eight years.
Michael Fitzpatrick Absolutely, absolutely. And you’ve identified one of the two I was going to speak about, I was also going to speak about the regulatory toolkit, which I’ll do in just a moment. But that’s actually relevant to the fourth category, which is continuous learning. And by that we mean, regulators should be working and designing processes to allow them to observe how their policy is playing out in the real world, and also how the technology is continuing to evolve and engage with society. At the front end, continuous learning can be incorporated in flexible modes of regulation, like pilot projects, and waivers, for instance, with automated transportation, where you’re allowing controlled experiments to occur in certain places around the country before you regulate across the entire sweep of society. But it can also play a critical role at the back end, once regulations are developed and are in effect, to continuously observe and learn from how they are actually interacting with society. Much of the work we do in designing regulations, and this is what OIRA managed for the White House under all presidents since Ronald Reagan, is ex-ante or before the fact analysis of how we think the regulation will act in the real world. It’s informed, it’s empirical, but it’s still a guess. The way you really know how a regulation is engaging with the world is 5, 10, 15 and 20 years after the fact. And so, we believe, and many presidents have tried to develop policies to encourage agencies to look back or retrospectively review regulations already on the books to assess. Are they doing better than we expected? Are they doing worse than we expected? Is it a mixed bag? And should this regulation be retired? Should it be modified or should it be strengthened? All of those are possible answers. Now, let me just put a footnote in here for the challenges of this. You are right, in the Obama administration, and I worked on this with with Cass Sunstein while at the White House, we did engage in a retrospective review process that was pretty robust. Every president has had some variation on it over the last 30 years. It’s a big challenge for regulators, for agencies, to do in a consistent sustained way across the full corpus or body of regulations. And that’s because of resources and time. It is expensive and time consuming to go back and look at lots of regulations and really assess how they’re doing and then, of course, to engage in the new rulemaking that’s necessary to adjust that.
Tom Temin Sure.
Michael Fitzpatrick So I just want to point out that if folks want agencies to really engage in this in a consistent and robust way, Congress is gonna have to step up and provide the resources necessary.
Tom Temin You know it always comes around to that too. Just for a moment I want to talk about the process of rulemaking because that has been kind of on the back burner, but steadily boiling away, if you will, for some time now. And that is to automate some of the assessment of comments that are submitted by the tens, or thousands, or millions that are identical, this kind of thing.
Michael Fitzpatrick Right.
Tom Temin And also just putting more automation and more machine learning in a way that doesn’t distort the intended outcome, but also speeds up the whole process for the agency. Is that part of the agility picture here?
Michael Fitzpatrick Absolutely. In our agility tenants that go to internal process, we talk about making workflow more visible by using plain language and regular transparent updates to the public about regulatory plans and constructing small internal but inclusive teams that can collaborate. But one of the things we do focus on as well is regulators themselves harnessing this new technology that I’ve been speaking about to make the regulatory process more inclusive, more transparent, more responsive and more efficient. So you can imagine, first of all, that harnessing the internet, you can open up the process and make it accessible to individual citizens and small businesses and nonprofits, because they can now engage with agencies online. They can get access to vast amounts of information about past regulations that are already on the books, but also about regulations that are under development. They can engage with the agency virtually, with one on one meetings or to attend public hearings. So the transaction costs of engagement are lowered and the opportunities are raised using technology. In addition, as you noted, good regulation relies on good data. Empiricism. And the process is intended to actually, it was sort of the original crowdsourcing mechanism, the Administrative Procedures Act passed in 1946. It said, put your proposal out to the world and solicit feedback from anyone and everyone, assess it, respond to it and then produce your final work product, which by the way, you’re going to have to defend in court. So this is a serious exercise. Using artificial intelligence (AI) and machine learning (ML) techniques, agencies are going to be able to ingest that vast amount of public input to bundle it and categorize it and organize it in ways that will allow both themselves to use it more usefully, but allow the public to access it in more useful ways. And let me mention one other concept. And that is for regulations that are already on the books, imagine, if you will, if there were AI and ML based tools that would allow a small business person for instance — or an entrepreneurial who was interested in expanding their business, or getting into a new line of work or introducing a new product or technology — they could access the federal body of regulation through this mechanism on say regulations.gov, and describe the activity or the product that they were going to engage in or introduce and have the machine learning aggregate all of the rules and regulations that might be relevant to them that they need to consider in engaging with that. That could be extremely cost effective and useful for entrepreneurs and small businesses.
Tom Temin Sure.
Michael Fitzpatrick And one hopes that over time, [the] General Services Administration (GSA) will get the resources to institute these kinds of tools.
Tom Temin We’re speaking with Michael Fitzpatrick, director of Global Strategy and Innovation at Google and a fellow at the National Academy of Public Administration. And that aggregation could also help agencies understand what they can retire and trim back and so forth by the same token.
Michael Fitzpatrick I think that’s right. I think AI and machine learning tools, always, of course, sort of being supervised in some way by the human beings that they’re meant to assist, could absolutely carry out data aggregation, data crunching and data assessment projects that would allow agencies to better assess the impact of their regulations.
Tom Temin And do you think this agile approach can exist coincident with a societal debate about federal regulation. There was a Supreme Court ruling that was controversial and will be debated for a while long time yet about how much agencies can interpret what is in the written law as passed by Congress. And we’re not going to resolve that one today, but it seems like that’s an issue that somehow needs to be inculcated into the federal workforce so that it can regulate effectively, but yet, within the law.
Michael Fitzpatrick Yes, I mean, I think the West Virginia decision deals with a complex and controversial concept of statutory interpretation and delegation of authorities from Congress to the regulators. So I’m not sure how much impact it has on the ability for agile regulation tenants to take hold. But I do think that, and regulators know this, in order for the regulatory state, if you will, in order for regulations to continue to demonstrate their efficacy, their effectiveness, in order for there to be opportunities to continue to regulate in ways that protect the public but also don’t hinder innovation and economic growth, they are going to need to adapt to this very new world, and are going to need to adapt their processes and their thinking, so that they can grow and mature along with it. And I think that has impacts on the ultimate sort of validity and power and agency that they have to continue to make policy. So the better the regulatory state does its job in protecting the public, as it’s been commanded to do by Congress through statute, in ways that are flexible and nimble and future proof, and they can evolve so that they don’t unnecessarily hinder job growth and economic growth and innovation, that is the sweet spot for regulators to continue to do their jobs effectively. And to have the mandate to do so.
Tom Temin Yes. And for this agile future, then it sounds like there are some technical investments that need to be made, some legal statutory work that needs to be done and some workforce development.
Michael Fitzpatrick Absolutely. And we come back to the question of resources once again, which is if you want regulators to test test test, if you want them to look back at their regulations, if you want them to hire technical experts in house who give them the capacity to really understand these new technologies and to build smart policies around them, if you want them to be able to obtain the technologies that are necessary in order to do the things we just talked about — the AI and ML based technologies — to make the regulatory process run better, that all takes money. And so it’s got to be a partnership between the executive branch and the legislative branch to make that happen.
Tom Temin Michael Fitzpatrick is director of Global Strategy and Innovation at Google and a fellow at the National Academy of Public Administration. Hey, thanks so much for joining me.
Michael Fitzpatrick It’s my pleasure, Tom. Always happy to get up and talk about regulation.