Stephen Goldsmith explains how technology is about to radically transform how city governments operate

Featuring Stephen Goldsmith
February 21, 2018
23 minutes and 4 seconds

Sometime around 221 BCE, the Qin Dynasty established a new centralized bureaucracy to govern its dominion in China. While the dynasty fell after just 14 years, the bureaucracy survived. Roughly a thousand years later, it had evolved to adopt many of the core concepts of a modern civil service: a professional class of trained government workers selected by merit through exams, and subject to strict rules of professional conduct.

The Chinese civil service would last another thousand years before its abolition in the early 20th century —right around the same time as the idea caught on in the United States. After decades of rampant Tammany Hall-style political patronage, the establishment of civil services in cities around the country radically changed public perception of government, providing stability and accountability, albeit not without tradeoffs.

Bureaucracies became known for fostering cultures where adherence to rules often overshadowed citizen needs, experimentation was discouraged, and civil servants were evaluated based on preset metrics rather than outcomes.

But thanks to recent technological advances, city governments may be poised to undergo another radical transformation, this time shedding their bureaucracies and adopting the role of distributed platforms — hubs through which city workers, not-for-profits, and private enterprises can all provide services.

In this episode of PolicyCast, Professor Stephen Goldsmith explains how cities are now able to radically change the way they operate, to the benefit of citizens and city workers alike.

Goldsmith is the director of Harvard Ash Center’s Innovations in American Government Program and recently co-authored “A New City O/S — The Power of Open, Collaborative, and Distributed Governance” with Neil Kleiman.

Hosted by

Matt Cadwallader

This episode is available on Apple Podcasts, Spotify, and wherever you get your podcasts.

Note: This transcript was automatically generated and contains errors.

Matt: In this book, you propose a complete rethinking of how municipal governments operate, going from a hierarchical system to more of a distributed one. Can you describe it?

Prof. Goldsmith: Alright, well we have a society that’s very distributed. Folks are used to doing things on platforms, whether the platform is Facebook or whatever the case may be, and we have a large number of new technological advancements that take advantage of that platforming. We’ve got a government that was designed over a century ago. I mean, we came out of Tammany Hall. A lot of corruption. The way managed to reduce corruption was control mechanisms and hierarchical controls and lots of restrictions on the way that public employees do work, so now we have this wonderful opportunity to dramatically change the way government works and it’s up against these kind of old procedures and old processes and old approaches which actually reduce the value of the public servant and therefore reduce the value of services to the public.

Matt: You seem to describe kind of three generations of municipal government model over the last, say, 200 years or so, or even really in the last century or so. You just mentioned Tammany Hall. What was that progression over time?

Prof. Goldsmith: Well, we went from a period of unlimited discretion, AKA corruption, right? Mayors hired their buddies and gave contracts to their friends and the like, and then we corrected that with a set of civil service reforms that were designed to reduce corruption and labor contracts that aided in that and a very kind of formalistic approach to government. That worked in the sense that we now were able to produce more routine activities, like the same exact widget every day, manufactured, theoretically, from a competition that picked the best vendor, but that vendor was picked on activities, not on quality. The whole system was based on, “We’re gonna measure your activities whether you’re a public employee, you’re a vendor, and the number of those activities is our definition of success.

Well, of course, many of those activities didn’t lead to outcomes or even outputs that make sense, so then we go to this kind of reform effort that says, “Well, we’re gonna reinvent government. We’re gonna be a little more entrepreneurial. We’re gonna ask for innovation in government.” I was actually part of that kind of wave and pretty excited about it.

Matt: When you were mayor of Indianapolis.

Prof. Goldsmith: When I was mayor of Indianapolis, right? That was reform in an analog world. That was innovation in an analog world. That was innovation that said, “Your job is … Public works managed the trash pickup and you’re gonna get more creative in the way you do that.” Now along comes a digital revolution that says, “We can look at data across agencies. We can look deeply at outliers. We can predict which employees are working best. We can help our employees exercise their discretion better.

We have a number of ways that we can talk about it as we go through the show today, but though all of that excitement now still is layered on top of this old operating system, and we found across the country, we visited over a hundred cities, that the cities that were being innovated were innovated because they were working around their structure of government. They had set up a special office. They had a separate agency, not because they were able to reform, so the purpose of the book is to say, “This is a new operating system of government. This will unleash the technology and it’ll unleash the talents of your employees.”

Matt: You open up the book with an example of New York City with the Mayor De Blasio’s rollout of the pre-K education system. Can you describe what happened there?

Prof. Goldsmith: Yeah, so we use UPK in New York City for the purposes of showing what a distributed platform would look like because in this country today we’re kind of in a fight between, “Well, this is a public activity, this is a private activity. We’re not gonna privatize this,” or like depending on which side of that you are. Well, any complicated, well-executed strategy involves public, private and nonprofit, so in New York when Mayor De Blasio … I worked for Mayor Bloomberg. Mayor De Blasio took over and one of his key campaign promises was universal pre-K, right? He was gonna stand that up very, very quickly.

How do you do that? We have to identify places where those school … pre-K could occur, but they have to be safe. That means they have to be inspected, so there’s a role of government to make sure that, well, maybe they’re in government schools meaning public schools. Maybe they’re in nonprofits, but either place, they have to be safe, so the role of the city, right?

But then you have to reach out to parents, right? How do you get to parents? Well, maybe social media, so then you reach out to parents, but then parents need information on where the schools are, and eventually as the schools get deployed, the quality of those schools, right? Then you begin to think about this is really a distributed system. We have pre-K in public schools. They have pre-K in non-public schools. We have parents making choices among the schools. We have a role for the city, and so we use that example to say, “Think about the delivering of government services as a platform of distributed services,” and UPK preschool in New York City is a great example. In fact, in record time, Mayor De Blasio got that up and running and with pretty good reviews by virtually everybody.

Matt: One thing that I found really interesting about it wasn’t just the distributed model and the technical part that allowed it to happen, but it seemed to have a cultural change element about it. We kind of think about bureaucrats as kind of pencil pushers, that stereotype, but you highlighted a number of civil servants who really too on the project, put it on their shoulders.

Prof. Goldsmith: Right. Your question’s a good one because in this story there are people in jobs that didn’t exist … the jobs didn’t exist before our chief innovation officer, right? Whose job is to set this up. Somebody who thinks about the design, the user design of the services. Those in public employees acted in a more entrepreneurial fashion. However, their entrepreneurial energies would not have been successful had they tried to do it all inside the four walls of government in a traditional sense. There wouldn’t have been enough classrooms, there wouldn’t have been enough teachers, there wouldn’t have been enough information. It would’ve been taken too long for each of the inspections to occur, so this thinking about entrepreneurial activity, but inside a network government, is really pretty exciting.

Matt: One of the reasons that the civil service emerged was to prevent corruption. The whole idea of adherence to rules and regulations, that was all about preventing waste fraud abuse. How in this new distributed model is there a check on that type of corruption?

Prof. Goldsmith: This is a very important question and one that we should watch carefully. The check 10 years ago was lots of rules. Procurement rules and civil service rules. We’re suggesting more discretion. You’re a field worker, maybe child protective service or whatever, and you are now gonna get on your handheld or tablet not only more digital information but organized with decision support systems like does a child have a broken left arm? Did any of her siblings have broken arms? By the way, we have a report from the emergency room of another family member who had a similar injury, so you’re providing information to the field.

Now, if we’re gonna give discretion, then there has been a thought that discretion and accountability were trade-offs. You could get people’s discretion but you would lose accountability because they could exercise discretion different ways, but actually in a new digital world, that’s not really true, because we can see from her GPS stamp, the child protective service worker or the asphalt pothole worker or fill in the blank, how long they were on the site. We can see what decisions they made. We can see whether they’re outliers and taking too many children out of the home or not enough children out of the home. We can see whether the language they’ve used in their comments. We can evaluate the sentiment, actually. Machine evaluate the sentiment in those comments to say were they making one set of decisions for African American kids and another set of decisions for white kids? Do they need more training?

The point of this is we can both concurrently, now, give more discretion to the field and hold people accountable because we can evaluate much more precisely the results of their work and their attitudes.

Matt: Part of this is about encouraging the kind of entrepreneurial spirit of these individual workers without having to get that top down directive. Do you think that kind of having them, their work be tracked that closely, the kind of GPS tracking and tracking what they’re writing in their comments, do you think that might have kind of an anti-effect there?

Prof. Goldsmith: Right, well, it could. I’ve been in local government for 25 years and I have a deeply held view that most people are in local government because they actually want to serve. They’re not there because they want to earn big money. They’re there because they want to serve, and most of them actually get frustrated because they can’t really serve because they’re too busy doing these kind of commodity, mindless activities, so I think to the extent that we give them more discretion, they’ll be more excited.

Now, your question, though, is important in the following sense. We need to use this increased visibility into their work for purposes of training and assistance. Sure, there are gonna be some bad actors that need to be weeded out, but if the 98 percent, 99 or 95 or whatever it is of the public workforce that’s well intentioned can be trained with the results of this information, I think we can create a culture of initiative, but you could very well have an anxiety about kind of a gotcha environment and that will be an important thing to manage.

Matt: Another thing that comes up on this is cost. Of course, any kind of distributed system that is based on technology requires significant IT resources. Do you see this as a situation where there would be private vendors that municipalities would reach out to to create these systems or is it up to kind of individual cities to build their own systems?

Prof. Goldsmith: This is an important question and obstacle, because a fair amount of the money that we spend today in cities and states doesn’t produce the objective it’s supposed to, right? So [inaudible 00:13:07] dollar of input do we get a dollar of output? Not necessarily. Not because of corruption or ineptness, just because that input activity doesn’t produce that, so it’s clear we could save a lot of money and/or repurpose dollars to higher and better use, so that seems clear, but it does take some money and talent to put in these systems.

Now, I think that the barriers to entry have been reduced through cloud services and [SAS 00:13:41] process pricing, right? So that a vendor can say, “I will set this up and run it for you for five years. I’m gonna charge you on a per-user basis so that a medium size city doesn’t have a huge …” When I was mayor of Indianapolis and we were doing a lot of technical stuff a long time ago, those were really big price tags to implement a system. Really big price tags. Here, the price is not insignificant but it also can be calibrated to the size of the organization to the number of users. It can be amortized over a period of time, so I do think there’s a pretty significant role for the vendor community in bringing technology to cities and then reducing the cost of uptake.

Matt: So you do see there is significant savings by implementing these types of [crosstalk 00:14:30]?

Prof. Goldsmith: There’s definitely significant savings. The savings aren’t on day one, but the costs are, right? So you you have to [crosstalk 00:14:36] you’re gonna do that.

Matt: Do you think that this is a … Is this something that is helpful for all city services or are there particular areas where there are benefits and other areas where a hierarchical system suits it just fine?

Prof. Goldsmith: I mean, there’s a number of ways to think about this that we’ve mentioned in the book. We call it “The new operating system for cities” because, A, if cities organize themselves around their resident/user’s experience, they’ll become more responsive, and that responsiveness is important. That goes across every city agency, so your question actually can be answered depending on kind of how one thinks about it. Reorganizing city government around the UX of the citizen, which we mention in the book in a chapter in multiple ways is really important. Personalization and omnichannel choices and the like.

Predictive analytics, which is another form of the digital revolution applied to cities, which says we’re gonna try to figure out where the pothole is before it occurs. We’re gonna try to figure out where the pipe’s gonna break or the street light’s gonna go out. We’re gonna try to figure out where the next child’s gonna be abused rather than the last child. Those predictive analytics are relatively complicated, relatively expensive. They require information sharing agreements and the like, so that conversation, I think, requires city county leadership to look at their most important targets of opportunity.

Some may be fiscal like who’s not paying their taxes. Others may be human services, who’s getting hurt. It could be most helped. Another way to think about that is the purpose here is not just to hold the employees accountable. It’s to hold the third party contracts/vendors accountable. Are those vendors producing the results that they should be producing? So, I think that it really depends what you look at. We talk in the book about “just in time.” That’s a concept to us that seems relevant to use the digital power to change the way we regulate and permit to dramatically reduce the time of doing those things, so that would be a regulatory function. The UX would be across government. The predictive analytics would apply to kind of high value targets.

Matt: It seems like a system like this where things are distributed and we’re talking about the government as a platform provider. It would also allow for more privatization of services and perhaps even competition that’s not several years long contract based. Is that kind of part and parcel of this idea or …

Prof. Goldsmith: Yes, it is, except both because privatization is kind of an ugly word these days and because it’s not exactly applicable, in the sense that we think about privatization as the divestiture of a public responsibility to the private sector. I’m thinking more in terms of how does one blend the talents of nonprofit and for-profit actors under the control and management of the city? They say the city, I’m gonna control pricing, I’m gonna control access. I’m gonna control route structure or whatever, but I’m not necessarily gonna be in the business of running the taxis or running the Ubers, running the software system, right?

I think there will be more competitive opportunities for nonprofit and for-profit high value actors, but inside a network that’s managed pretty carefully by the city or the county or the state, and therefore it looks more like to me a true public private partnership than a pure privatization.

Matt: As I was reading through some of this, I was reminded of after the affordable care act was passed, part of it was a mandate that hospitals and care providers move to digital recordkeeping and knowing some people who worked at local hospitals, they moved to some systems that essentially handled all that, and a lot of them were very unhappy because the user interfaces weren’t great. It made what they were doing a little bit more difficult even though the promise of the system was to make things easier. How do we avoid that in building these types of systems?

Prof. Goldsmith: Well, I’m not sure it’s totally avoidable. I went in for my annual physical checkup. You know, the doctors, instead of looking at me, he’s looking down.

Matt: At the computer?

Prof. Goldsmith: At the computer, and after a while he goes, “I’m really sorry that I’m not talking to you but I’m required to enter these notes as we talk.” Then he complained about how difficult the system was to use, which took a substantial part of our total appointment time for him to tell me about the computer system. He’s a good doctor. I think this will be an issue. In the book, we talk about UX in two ways. One is the ways in which a resident engages with his or her government. Do they walk in? Do they app in? Do they call in? How easy is that? How do we use human-centered design techniques to make that easier? The second way we do that is to say, “Look, this UX needs to also extend to your own employees.” How do they get this information? How do they use this information? What do they put in here? You need to think about the user experience to your employees and then look at how that translates into the user experience, kind of retail user experience to the resident.

Over the years, government’s done a remarkably bad job. If somebody comes up with a good idea … I’ve been actually quite guilty of this, and then you just kind of impose it, right? You say, “Look, look, we just did a great favor for you. You’ve got your new tablet. You’ve got all that information in there. Go use it.” Well, nobody’s really actually kind of worked with a field worker about, “Well, yeah, but this adds time or this makes it more difficult,” so I think the question’s really, really insightful and it will require more care in the design of the systems.

Matt: What do you see as the biggest impediments to this move? Is this something that’s kind of inevitable with technological change or is this something that’s gonna require efforts on the part of mayoral administrations the country over?

Prof. Goldsmith: Well, I think the answer’s “both.” I mean, it’s inevitable eventually. Only so long people are gonna tolerate a bureaucratic experience after they personalize their Amazon order, right? We ought to be [Amazoning 00:21:18] government and we’re not, so that public pressure will increase, and as folks more and more are using mobile tools and other devices for the rest of their lives, they’re gonna want to use them to communicate seamlessly with the city, but these bureaucracies die hard, right? Almost every city government in the US today is set up with a transportation department, a public works department, a parks department, a fill in the blank department.

They’re often populated with really good folks but those folks are kind of captive with the systems in which they work, and if you say, “Yeah, but there may be information in the public works department that will help you explain why the underground pipe is causing the pothole to occur four times a year,” or there may be information in the school files that would help the case worker understand that the child that she’s looking at has been truant from school for three of the last seven days. What’s that mean, right?

We have to figure out a way to kind of accelerate the permeability, the horizontal nature of the way we deliver services and break down those verticals of that. That will be a very slow process.