Closing Regional Economic Divides

The Growth Lab Research Seminar series is a weekly seminar that brings together researchers from across the academic spectrum who share an interest in growth and development.

In this talk, Gordon Hanson, Peter Wertheim Professor in Urban Policy at Harvard Kennedy School, discusses his research on how to help lagging regions create better jobs for disadvantaged workers. Traditional industrial regions have fallen behind economically across high-income countries due to globalization, new technology, and now the energy transition. Hanson believes new approaches are needed to diagnose the causes of persistent regional economic distress and the effectiveness of alternative policies in relieving this distress.

Transcript (Part I)

DISCLAIMER: This webinar transcript was loosely edited and there may be inaccuracies. 

Gordon Hanson: And what we're trying to understand is how we improve market mechanisms to develop better quality jobs in the global economy, with a focus in particular on those without a college education for whom the last 40 years have been very tough. Within that broader set of aims. We have a specific interest in places that have fallen on hard times the deindustrialization of high-income countries. Now, the beginnings of the industrialization in emerging economies has created pockets of regional distress, which are surprisingly long-lasting. So I'm going to talk to you about is a bit of our conceptual approach to thinking about how we go about policy design in this context and then tell you about what we know and don't know about the effectiveness of a place-based policies in dealing with these issues. And then give you our plan of action and kind of tell you where we are in the scheme of things. So this is I'm not trying to pitch a set of results to you today. I'm trying to pitch an approach. And as a consequence of that, that feedback I get from you is enormously valuable because we're really in early stages. So there are many ways to motivate what we're doing. My own frame revolves around what's happened to local labor markets in the US over the past three decades. And what we've seen is the entrenchment of joblessness. By which I mean low employment to population ratios that fall as a consequence of some shock. So my own work, I've studied the China trade shock and its impact on fracturing employment and how that has then translated into the absence of lot of employment creation in different places. But there's a growing list of things that are responsible for regional economic distress. Manufacturing decline in general, which began way before the most recent wave of globalization picked ups, picked up steam, was important. If you think about you look around New England and we have cities like Lawrence and Lowell and Holyoke and Springfield, all of which lost manufacturing jobs after 1960 and had a very hard time bouncing back. You go to New Bedford today, which at one point was one of the richest cities. World Center whaling industry then was able to transition into textile and footwear production. Today, outside of a small tourist pocket near the port is a place that is where joblessness and despair have been present for a couple of decades. Technological change matters here. And now we have the ongoing energy transition. Now, this energy transition isn't a new thing. Decarbonization is accelerating. But the movement away from coal began a couple of decades ago. And that wasn't a movement of towards coal, towards non-fossil fuel-based industry, but just a move towards new forms of fossil fuels, natural gas in particular. So if we add up all of these shocks and then look at the regions that have been affected by what we see and hear is I'm showing you in dark blue are places where we saw substantial declines in the employment to population ratio since 1990. Now, why do I focus on this one particular measure? The employment to population ratio is the as close as you can get to a summary statistics on the economic health of a place. So you think about any economic model that tells us about labor supply, labor supply as a function of the real wage that workers can command in the marketplace if that real and real wages are difficult to measure. We can see nominal earnings, but there's all sorts of issues with unobserved characteristics of people. We can see some local prices on a lot of local prices. We can't. So detecting real wages with a great deal of precision is tough. But if real wages are low, people will work. And so that's where the employment-population ratio comes in. A very simple indicator. Now it's all the more powerful in the US context because it's variation across place is enormous and the changes in the employment to population ratio have been large over the past few decades. And so that is a signal of a reduction in labor demand that hasn't been followed by the successful transition of workers into new lines of activity. And that's the core one of the core policy challenges we want to address.

Attendee :Why do you focus or start at 1990?

Gordon Hanson: As I'll talk about we're going to go way back. The story starts in 1910, 1910. We kind of industrialization is not quite complete in the U.S. North. And our focus just isn't just on U.S. health. We've been stories from other countries as well. And there the timing really matters. So industrialization in the U.S., North Qatar really takes off in the late 1800s, keeps going at a steady pace during the first half of the 20th century. You get to 1960, and that's when manufacturing starts to decline. And lots of high-income countries as a share of employment, not not in terms of as a share of GDP, but the share of employment. And so part of the story that we're going to be looking at is how did northern cities adjust to the exodus of manufacturing jobs from 1960 to 1994? The time the shock happened before 1980, before skill, bias, technological change was a thing. We know that places that were specialized in manufacturing in 1960 had a bad couple of decades after that. We actually don't know a lot about the nature of labor market adjustment, those shocks, pre-1990 shocks. One of the things we want to try and understand is are there some useful lessons about successful adjustment to some of those older shocks that we might be able to apply today? So we're going to take as our first set of analyzes, we'll take as a laboratory that us local labor markets going back over the past hundred years and understanding the decline of manufacturing and traditional energy jobs and some other traditional industrial jobs on local economies and how they adjusted and what were the things that made adjustment more successful? So there we just list a few of the shocks that people have studied. And in this work and my past work, I've been very interested in specific types of shock and specific types of shocks for the purposes of identification. You want to know that you've got an exaggerated source of variation in labor demand so that you can then study what happens or kind of we know all of that now. And so if you just use a simple Arctic measure of the projected decline in manufacturing employment for a case, what you see is we and we know this work from on from Erwin Charles Eric Herson and Schwartz that you know the broader declines in manufacturing really mere kind of adjustment to the china shop. So it's about the loss of good jobs for people without a college education. So painful. So the sets of economic ills that come out of this include, one, what could be a spatial misallocation of resources. We've got people stuck in places with low employment rates and other people thriving in places with high employment rates. So Boston versus Holyoke. Holyoke and Boston aren't that far apart. Are there barriers to the movements of capital or labor or technology between those two places, entrepreneurial talent between those places? And if those barriers exist, then we've got a source of inefficiency. Then we've got some endogenous processes at work with agglomeration of forces having been ever present since humans discovered that cities were a good thing. But those agglomeration of forces really seems to have changed in the eighties and nineties and are now really anchored in the concentration of highly educated workers. And that's both due to sets of spillovers on the production side traditional, more salient externalities that seem to be a little more a little more steelworker centric. But also on the consumption side, when you gather lots of bright young people together, they then generate demand for the urban amenities that attract other bright young people to come there. The with if we then get this separation, you get the entrenchment of different of in spatial inequality. And as we have learned, that has real political consequences. The rise of support for populist nationalist politicians in the US, in France, Germany, UK is concentrated in places where we've seen greater of manufacturing jobs lost. This then rolls over into a whole bunch of attendant social consequences. So the work of anarchism echoes even about deaths of despair, manufacturing job loss and job loss and coal and other things isn't the only thing responsible for those deaths of despair, but it contributes to it. And we've got a growing body of work that shows that. So what do we do in the face of all of these issues? One is do kind of what we're doing now. And that is just let market forces work and market forces will do their thing. It takes a long time. So think about Pittsburgh and the loss of upscale manufacturing, which started really began in the 1960s and pickup picked up a pace in the 1970s. Pittsburgh has come back. It's come back at about 60% of its former size. It's come back with a very different composition of people. And it's come back like after 2000. They have like a good 30 years of time in the doldrums. And one of our concerns about that whole generation of kids that are raised in that environment, and that means loss of economic opportunity for them. So then when it comes to policy, the traditional approach is to treat people and not places. And how do we treat people? We can treat people either by conditioning on their labor market status at a given moment in time. Are you unemployed? And then we might condition that nature of that aid on labor market conditions. And the way this works in the U.S. is pretty stingy unemployment insurance that gets generous when the national economy is in recession. Other economies handle this quite differently, and Britain is a little bit more generous than you ask, but kind of the same story. Denmark takes a very different approach in terms of in terms of what we do about the unemployed. The challenge with the way we've dealt with unemployment is that we tune generosity to what's happening to the national labor market. A lot of this manufacturing job loss has happened off cycle. It's happened actually during periods when the national economy is expanding. And so what you have is there is a lack of generosity, of short duration, of benefits in just not doing much on the ground to help folks. We also have means tested entitlement programs. And so that includes Medicaid. So you get some access to subsidized health care, includes negative income taxes. The earned income tax credit in the United States includes food stamps. The there are two challenges with these programs in terms of addressing areas in economic distress in practice. One issue is that we've simply dismantled a lot of these programs, starting with welfare reform in 1986, and we devolve lots of autonomy to states in terms of their design and their generosity. And the states that chose to be less generous happened to be the states where this job loss, the postwar problems, particular reconstitute. The second challenge is that there's a running conflict in U.S. social policy in terms of is it about helping people in poverty, which those in the left advocate? Or is it about helping families, children, which those on the right advocate? The confluence of those two forces creates a system that heavily conditions aid in practice, having kids in the household in an environment where job loss is endemic and male employment rates in particular have fallen. Men become less attractive marriage partners. So we've seen an increase in single headedness over this time period, overwhelmingly dominated by women being at home, women living with their kids. And what that has meant is that men who have suffered negative shocks have access to pretty paltry policies in terms of allowing themselves to adjust. So whatever we think about the effectiveness of treating people in theory, in practice, we haven't done a lot. Now the other approach is to treat regions. And there's and I'll talk a little about kind of what we know about the pitfalls of place based policies. And here though, and whereas the former is trying to help workers move to jobs, the second is helping new jobs to workers. And here, you know, you all know this well in the growth lab. You've got your Wyoming project up and running. There's lots of experimentation happening on the ground. That's what this project is about.

Muhammed Yildirim, Growth Lab Research Director: So there's a question from Zoom. Can you repeat the trend when programing power transfers from federal to states? What happened?

Gordon Hanson: So in 1996, Bill Clinton promised to end welfare as we know it in the United States. And he did. You did a really good job in fulfilling that promise. So what that meant was that the federal government would continue to support certain federal programs, but there were new sets of conditions. One was a five-year lifetime limit on certain sorts of aid and pretty much all the cash based. So direct income support to low-income families was capped at five years. The second thing that happened was saying, we're not going to tell states how to administer the state. We're going to give them a block grant and then they can decide. And the block grants will go for food stamps or the block grants will go for temporary assistance to needy families, or the block grants will go to Medicaid. And then states can decide what they want to do. They can add money or it cannot add money. And so you now had an environment in which states had more autonomy over how much they added on. And there was a sharp variation across states in how much they did. This gets at a long-running tension in the U.S. over social policy with states in the U.S. in order to be more generous, states in the South be less generous. States in the West kind of somewhere between the two parts of the U.S. West, we have rugged individualism at work and little government support. And then we have California, which is taking kind of Bulgaria in 1957 as a model for economic policy. And so after 1996, what happened then was, in effect, a reduction in the overall level of federal support for poor social welfare programs, but also a lot of regional unevenness based on up based on other choices that state states were making. So this project, a big part of this project, is understanding the experimentation that's happening at the local level with the federal government doing much less. And now states and localities left to us to figure stuff out. So how do we justify place-based policies?

Attendee Very good question on this. You seem to take as a given low geographic labor mobility. What's the reason for that? Because you could think of the fourth pillar, which would be let's move people. Is that the next slide?

Gordon Hanson: This is a slide, for example, perfect time. So you know that the reason we kind of missed all of this, the reason we didn't think that globalization and technological change and the energy transition were going to create pockets of distress is that we thought of the U.S. economy particular as being highly geographically. And it was up until 1980. But that wasn't what you could think we should think of as steady state mobility. It was mobility that was occurring during a period in which the West was still completing its transition to a spatial, steady state. It's not that there was still substantial rural to urban migration going on, although there was a bit it was the U.S. is spilling out of the West and the U.S. is filling out California, Texas and Florida. And that meant there's very strong labor demand and that helped grease the wheels for economies that were subject to shocks in that time period. So Kyle Magnum at the Philadelphia Fed has a really good, good work helping us understand how we miss, infer the nature of geographic labor mobility before 1980, 1990, because we look at that as a steady state phenomenon. It was part of a transition. So then you get after 1990 and what we've seen as kind of in the last four years, the steady decline in mobility. So there are two ways to understand that. And we're not really sure what mix of the two gives us a false representation. One is that there are fixed cost to moving. There's uncertainty. There is the absence of social insurance making reliance on family and also for child care. We've got now mom and dad not living together and leaving relatives around to take care of the kids or yourself when you're sick. We've got all of these imperfections which mean that there's stuff that we don't. As your attachment to place get stronger in that environment. What we have is a disequilibrium in some sense in which we have differences in real earnings across place. There's another view that comes along and says, well, if you add this and recall, Moretti has been writing about this on its own and then recently with Rebecca Wyman say, Well, if you actually look at variation and housing crises, prices are across place. That variation is huge. So when you look at actual differences in real earnings, it may not be that big what's happening then? But that doesn't mean that intervention isn't justified. Intervention could still be justified because what we have at work behind that separation is a set of spillovers which lead to overconcentration of innovation and high skilled workers in India and superstar cities. Know one more second. So you then have that which story is matters? Because what do we want to do? Do we want to move workers out of distressed labor markets, or do we want to move skilled workers in? And there's a set of there's a set of theoretical papers, one paper by Cecile Guevarra and Pablo Obama, another paper by Adrian, all of which they kind of have have complementary but distinct frameworks about the nature of the externality at work. And getting it right matters. Because do we treat workers? We treat firms in new places. And we don't we don't really.

Attendee: Although I was also going to ask in the the sixties and seventies one, we assumed that maybe movement was the norm, but maybe it was just kind of part of the process. Also, that's when a lot of the Reagan era immigration kind of flooded certain pockets of places. So how did some of that lead to this inequality entrenchment that you also mentioned before?

Gordon Hanson: Yeah, so that the large scale immigration so immigration kind of reaches its peak as a share of the US population in 1910 at 15% foreign born, childless population and then declines until 1970. And so in 1977, five years then changes in U.S. immigration policy and then shifts in what's happening in Mexico and the incentive to migrate to the United States. That leads to the most recent U.S. immigration waves, which peaked 15 years ago. I know we've got a lot going on in the border right now, but immigration a lot. And so it's that kind of 1965 to 2005 period where that big immigration we passed, much of it happened in the nineties. As you passed in terms of the quantity, the rate of growth was really fast in the eighties, but the big numbers happened in the nineties in 2000 that may have impeded the mobility of that of U.S. foreign workers in terms of moving to places to take advantage of those jobs. As newly arrived immigrants we know this has are a lot more responsive to variation in economic opportunities so that layers on upon.

Muhammed Yildirim: You have two questions from Zoom. So one question is related to the decline of property values. So do you think that this about kind of property values in the declining community? So the property values go down about the fact because of the wealth effect, it might be harder for them to move, but that the declining property values affect a significant portion of the lack of mobility in these communities.

Gordon Hanson: So that the there's something where you look at and just as a methodological point, we think of it as hard to study changes in housing prices over time because there's so much variation. And it's not just what gets sold at any moment in time. In practice, as is often the case in economics and economics, our obsession with selection facts is over overdone. And if you look at changes in housing prices that you get from taking data from the U.S. Census and simply adjusting for number of rooms, age at which something was built and the observable things that changes and property values you get out of that series, out of the high-quality resales. We only have a couple decades super, super high. So going backwards in time, we can then tell you something about what all of this looks. Property prices decline, a number of things happen. One is it helps offset the decline in multiplexes. And so what that means is that disincentivizes moving. The other thing that happens is that it allows for a form of self-insurance. If you're in a high-wage place and wages aren't rising, you can then move to a low-wage place and you can kind of insulate yourself. So all of that means that the reasons for the housing price adjustment dampen labor mobility. Well, we have some other things going on too, which matter for the long run and matter for what we think about kind of the overall efficiency of the outcome. One is that now we're we know that economic opportunity is highly place-based. And so if you think about what happens to the opportunities for the kids who grow up in these places, you know, one world we got access to, you know, it's low wages, low housing prices. That means your tax base isn't that great. Schools aren't that good. That means that funding for regional public universities isn't that good. And you kind of evolve into providing the services that we can or whatever tradeable stuff you can do for bigger, richer places. So that's. So there is even if it were efficient, we might be concerned from an equity perspective about the outcomes that result. But it may not be efficient in the presence of the spillovers. Get overconcentration in superstars or something else that goes on to your home. Equity is collateral that you use to start new firms. We know that most net job growth comes from new firms. And as an aside, not all firms a tiny number of new firms. So we create lots of new firms. Most of them take a few of them, but a few that don't generate most of the job growth. A lot of those new firms rely on home equity for collateral to get loans to do the work by Davis and Haltiwanger. Building on the larger earlier literature shows that that decline of property values reduces investments in new firm creation and that impedes adjustment. And what that means is you can then that reinforces this separation of what's happening in and low wage, low house price, distressed places and other places there. And that is indicative barriers to the mobility of capital, which is another source of inefficiency. One of the things we're doing in the project, so we've got data on the presence of U.S. banks at a very local level going back about 30 years. And I know we've seen some great work in getting a person saying you get this great presentation on concentration of local banking. If you're a couple of weeks ago, the we're looking at something in parallel. You got a bunch of local labor markets in which a single bank accounts for 70, 80% of deposits. Now, they may or may not exercise market power on the lending side, but if those banks get whacked on their balance sheets by what's happening somewhere else, you're then interrupting, interrupting the availability of capital in a place that really needs it to start in firms. And so what we then what we then see here to take all of this is, which is lots of sources of inefficiency due to underappreciated barriers of mobility of labor and problem, age old problems of investments of capital, some perhaps surprising imperfections in the nature of capital markets and our standard well known imperfections that come from the presence of these extra notes. So that creates the potential for place based policies to do something good. The potential, as I'm getting it right, is a whole nother story. Okay. So we know that when it comes to looking at place picks, policies we have, there are certainly examples of failure which we can focus a bit more on. There are also examples of success. So the Tennessee Valley Valley Authority was created in the 1930s in the US in order to improve the economic livelihood of Appalachian. So this is a part of the US that was remote, had limited roads and highways and electrical generating facilities, and you had entrenched, entrenched poverty. So Tennessee Valley Authority actually lasted for over half a century. Most of what it did was in about a 20 year period and built roads and canals and power plants and schools and connected Appalachia to the rest of the world. And that allowed industry to move in. So it allowed the region to transition out of agriculture into manufacturing at a moment in time when agriculture was releasing lots of labor. So that release of labor from agriculture from 1910 to 1960 was just mind blowing. And it's this labor supply bonus. But if you don't have roads or power, that labor supply bonus does happen. So what? Cline and her show in a 2014 paper in the Quarterly Journal of Economics is that this did promote all of this great industrialization in the aggregate. It's not clear it improved national welfare for the U.S. overall, because we kind of just moved from one tradeable activity to another tradeable activity. It didn't hurt, however, that if you if you want to play, if you care about inequality at all, there's no way this wasn't a really good thing. So we have to decide as a society what pereda ways we place on income people according to their well-being. Appalachian was really, really poor, so that's an example of success in my book because even if even though the national efficiency outcomes are unclear on the equity side, this one doesn't just seems really easy to justify. Then we have, you know, the development of practice and workforce development over the last three decades, a story which hasn't been kind of fully told. But what we now have is is kind of well-recognized best practice. And that best practice sometimes called active labor market programs, sometimes federal labor market programs in which you identify workers in need. So this can be youth and underserved communities, could be the long run unemployed. And you do two things. First, you look at local employers in your place and you and you work with them and say, what type of workers do you want to hire? What skills do you want them to have? And sometimes this happens through community colleges. Sometimes this happens through bespoke training programs. Sometimes it happens through private training institutes. And you provide those skills. And what's important here is to actually to really work with employers. So the Industrial Areas Foundation, which was started by Saul Alinsky, achieved fame for being by creating the community organizing framework that President Obama was a part of when he was a young person and he was maligned for being a community organizer. One of the things that the Industrial Areas Foundation has grown into is is building workforce development in communities with lots of low income workers, mostly in the south and the Southwest. And they have been using that developing this program of the past over the course of 30 years. Our Nestor Cortez, one of the co-founders of this operation, was a MacArthur Genius Award winner in the nineties, received a prize from the Kennedy School 1988. I think it's kind of know what was the right thing to do for a long time. Larry Katz has a brand new paper with a team of coauthors here at Harvard, in which he looks at some of the programs implemented by the Industrial Areas Foundation, Project Quest, and also for other programs, all of which were subject to authorities. The other thing that these programs do is to provide wraparound services. That means you, you, you, you encounter workers and find out what's going on. You assess their skills. You have to have some amount of screening. Employers will do. Employers are deeply skeptical of workforce development because they feel like you're peddling these workers and not listening to what might. So you have screening and that means drug tests and pretty low level street drug tests and kind of basic test stuff. And then you provide career counseling. How do workers think about their future career counseling? We know from many different teams, including what I call a colostomy in Italy right now. Super important. And then you provide advice about how do you show up to work? What are the workforce habits? How you think about your career going forward? How do you think about advancing in your career? As problems come up on the job? How do you how do you address those problems? And so the combination of those two things a targeted workforce, a targeted training, and these wraparound services are really effective. Some over the studies that Katz looks at that all look at you have increases in hourly wages of between 15 and 35% over two your time period. And they persist. But some of these programs we follow orders out five years you're not seeing attention effects. David Card and coauthors have just redone a meta analysis of sexual labor market programs. They look at 210 programs across 30 countries because they're looking at so many programs, they have to look at a narrower set of outcomes. They're looking at the changes in employment rates. And what they find is that these sorts of active labor market programs work the things that don't work. The biggest thing that doesn't work are public sector employment programs. And what you just pay people who have a job and those do not work. Here's something I just learned that I had knowing that the U.S. in the 1970s is the U.S. has had workforce development programs through government legislation since 1963. And each decade, you get kind of a new piece of major legislation, which puts a new kind of paradigm in place, current paradigm, which really established under Bill Clinton. And then it was modified by Barack Obama in 2014. But it was kind of consistent with this framework. There are issues with it, which I'll talk about a second. But in the 1970s, the workforce itself in the U.S. was primarily subsidies for workers to have public sector jobs U.S. and that's where 60% of the planning for workforce development went in 1978, or 755,000 U.S. workers on subsidized employment in the public sector pay for personnel. So that part, the practice here, has evolved dramatically over time. The problem today with the way the U.S. does workforce development, it's not like Denmark does it. And Mark, you lose your job, you get generous unemployment insurance and you get into what is, in effect, a national active labor market program. Right. And so Denmark, even though it's much greater generosity, you see no higher you see no increased incidence of not employment associated with that generosity. The U.S. in kind of typical U.S. action, fashion, devolves a lot of autonomy to states. So states get money and then and those go to workforce development boards. 650 of U.S. and they give out vouchers to workers. Some of these workforce development boards are built into colleges. They're responsive to what's going on in the community. The great others are one or two person operations, which just are basically handing out vouchers. In other cases, these workforce development boards get captured by low-wage employers who say free training and sign us up and then that money gets. We do not have a good sense of how much of workforce development money goes into type one kind of best practice that this stuff how much goes into type to just you know being given out randomly and how much goes into type three, which is supporting your local monopsony? That's something we're going to try to address in more. So there's also, you know, other there's lots of anecdotal evidence about other things that might have worked. Let me talk about failures and failures. A fair number of failures on comes to place-based policy. One of the major pitfalls is that when done poorly, what happens is it's just a transfer to landowners, that the policy gets kind of fully capitalized in the land values. And so taxpayers are providing money to landowners, and you aren't really changing the process. So there's a really good paper in AJ Applied Climate Policy that looks at Germany's place-based policies along the border west, Germany's place-based policies along the border with East Germany before reunification. I can get you conditioned on proximity to the border and they do a regression discontinuity design about me. Just inside. Just outside. What's the primary difference between places? Differences in pattern and nothing else? Now, that was also, you know, that was a program which also might have been a dumb program. You weren't conditioning on the right thing. You're just conditioning on your longitude and latitude. You weren't conditioning on local conditions. Another pitfall is that you put money in place and it just gets captured by special interests. And this could be. And so this is different than the first one. First one could be well-intentioned. Everyone's doing the right thing. It's just that market forces lead to that capitalization. The second one is a pernicious outcome in which local business says, Right, I'm going to put money on the table for place-based development. I've got the perfect project in mind for this to work by Kaitlyn Slattery at Columbia Business School, who's going to be visiting here next year and on Star at Princeton. They have looked at using this approach developed by Greenstone Hornbeck and Moretti, where you look at two U.S. counties that were in the running for the plant. One wins, one doesn't. And then what happens to aggregate economic outcomes? And they're they're kind of their canonical results are that when you subsidize investment in a particular industry, you get more that industry get a significant increase in employment in the sector. Subsequent impacts on aggregate employment in the locality are not. So you're just replacing one thing for another. And impacts on land values or earnings appear to be low as well. Now, this doesn't mean space policy doesn't work because they could be dumb politics, they could be we could be investing not in distressed regions. We could be investing in richer labor markets, in poorer states. So there's how you don't take this as evidence that it doesn't work, just that it can be done for. The other very cool result that Caitlyn has is when don't when do tax breaks to firms locating in a place they peak in year two or three of the second term of a governor who's incumbent and up for reelection. So you just see kind of the obvious political cycles. And then folks may know about Kansas City second quarter where the First World War happened during the Civil War. Second World War isn't has as traumatic, but is is worthy of another call this time. So this is Kansas City in the United States, which as a metropolitan area straddles the borders of Kansas and Missouri between 1911 and 1999, 19 2019, the Kansas side of Kansas City said, We want to promote Kansas City, so we're going to go recruit businesses who are the easiest businesses to recruit, the ones already in the city, just across the state border that's going to Missouri. Look at these companies come back and then Missouri said, oh, no, no, we're going to you're going to be able to us we're going to do it back to you. So between 1920 11 and 2019, there were there was about $335 million of tax subsidy spent by the two states just on getting firms to move back and forth across the country. So the definition of wasted spent so place based, we you want to find examples of place based policy gone awry. They're easy to find, but there are examples in which it is done well. Aside from one of our cities with active labor force programs where you're really talking about something that is. A strategy. It's not a policy. It's a strategy. You're looking across different dimensions and saying, What do we want to be doing? Workforce development. What do we want to be doing in terms of helping local firms improve their capabilities? Which firms want to subsidize firms to which firms might need to provide some inducements not to leave? What sort of collective investment in infrastructure we need to make in order for this to happen? And an example that lots of people like to turn to is an organization called The Right Place in Grand Rapids, Michigan, where work at close who was hired to run this in the 1980s, showed up as a you know, as a pretty young professional. And at that moment in time to find Grand Rapids in utter disarray, just and had been suffering from a couple decades of manufacturing job loss and a place that had been a strong furniture manufacturing center for decades was now deep in the doldrums. A set of business leaders in the area came together and said, we need to do something. And what we need to do is find out how we improve the ecosystem here in Grand Rapids. That makes it attracted more business to invest. These what was one of the things that helped is that these 12 businesspeople had significant investments either and traded or traded stock in that locality. So what does that tell you? Well, they're not multinationals or big, big companies. They just not care. They don't you don't have something invested in them. So the nature of the industry was there, which gave you business owners who internalized some of the local consequences of what had gone. And that might be but not a not a that might not be a necessary condition, but in some instances, a special point to get things going. So what did they do? So over the course of a couple of decades, they provided a set of coordinating activities which helped reduce uncertainty to bigger firms and help and help them understand what Grand Rapids was good at doing. They had to figure out what Grand Rapids was good at in the first place. This is something that Rodrik has talked about a lot in his work. How do you know? What do we know about industry? Industry is moving between places all the time. So when your industry leaves, what do you do next? It's hard to figure out. So I don't think picking winners in terms of subsidizing semiconductor firms think of Grand Rapids as being a traditional furniture manufacturer. Like, what do we do next? Furniture is an auto business to be here because it's kind of like the lower end of it is like textiles. Here. You're competing in the thinnest margins, lowest wage exporters globally. So they provided they helped firms figure out where you want to be, helped identify the infrastructure that they needed, help them manage the process of getting permissions. And then on the on the also on the coordination side, working with local public, regional public universities to make sure they were providing the right skills. So they helped expand the engineering offerings. And in local public universities, they helped firms in Grand Rapids identify export markets to which they can profit, in which they can profitably sell. And they convinced some multinational firms that Grand Rapids was a place for the best and because of its tradition, furniture. And then they also invested in some kind of public R&D. You don't think of this as like saying we're going to they were they figured out where innovation should go in in the furniture industry. It's more that you create like it's a form of technical assistance in which you're providing consulting to businesses as they're trying to figure out how they upgrade their operations technologically, digitally and so forth. So in developing country context, we're totally on board with these sorts of the work that McLuhan and David McKenzie and others did with Indian textile firms ten years ago, in which you say you provide you provide McKinsey style management consulting to those manufacturers to stop long run improvements. They just did an update for that study. Ten years out, we had improvements in sales and productivity. At some point the same thing. And you see this gets derisively called industrial policy. It's been asked to aid in developing country contexts and David McKenzie has recent work on a similar on in on interventions with Colombian auto parts manufacturers the ripple there is. How do you provide these things, Bishop? You have to go in on it. I have to give you your bespoke management consultants because they aren't cheap, even when you're using a suit. They are cheap, they say. And so they experimented with a bespoke services, one on one versus group learning. So you think of it as like executive education for business, and that group learning was just as effective as the other. So Dani and I had Bergin here a month or six weeks or so ago, along with some other practitioners and talking about what she was doing and said, you know, you have to just you have this ecosystem and it's not a huge ecosystem. So it's an ecosystem. So an ecosystem can can get off track easily. And so you're not you know, you're not you're not centrally planning this, but you're you're trying to make sure that you're you're providing the things that allow the place to succeed. Now, is this replicable? We don't know. Is there something special about Grand Rapids? There are some special things about. So it was a place where Dutch immigrants settled in the early 1900s and it kind of formed this tight knit community. And they invested in local industry. And you have this kind of place collection of medium sized firms. And so they were sticky, right? They had something going on within the broader region. There was the city of Grand Haven, which was facing similar issues. And they realized what we have. We have firms here. And the problem with that the firms are having is not their factory workers. It's the engineers and the supervisors. And that that's kind of the middle manager type. We got to convince them Grand Haven is a place of employment. So a local business owner who happens to be asking the boss's father, as it turns out, of who is Donald Trump's secretary of education, Edward Prince, he said, okay, downtown, it needs to be a place people want to live. And so how do we make downtown the a place to live? So he said, I'm going to make downtown something that's a 12 month a year attraction. He put heated water pipes under the sidewalks in the downtown and all the stuff this part of Michigan and western Michigan, it's really it's not like Ann Arbor, Detroit. That's. And they had a whole college there, which was right adjoining campus. And they worked with Hope College in terms of trying to develop some programing. And that would just make the city more attractive in terms of cultural activities. And it's now a place that's and you spent $250,000 on these two types. So a great discussion of this in a book called Our Towns by Jeff and Jim and Devin Balanced. So we can find these examples. And what Danny and I want to try and do is understand are the, you know, the Grand Havens and the Grand Rapids, are are they 5% of cases? Are they 20% of cases? Not 30%. Okay. We know what the average treatment effect of jobs is bad. We also know there's heterogeneity and adjustment. Some places successfully transition into other activities. We also want to try and understand what are the conditions in these places that allow, allowed for and either have the social capital or allow that social capital. That's one of the things that we'll be looking at in part of this is how community colleges work. The community colleges in the US, the U.S. needs to centralize. All of these complicated jurisdictions are overlapping and intersecting and conflicting each other. But in many places, community colleges are quite responsive to what happens at local labor markets. Colleges do a couple of things. One of the things they do is they help people who aren't quite ready to go to a four-year university. Get the skills they need to do that. Both my parents in politics and then on the air for your schools. So that's kind of the college track. They also do a huge amount of vocational training and gift certificate programs that could be six, nine months in length in manufacturing and repair and construction and in the classic blue collar trades. We know that in a study done on the state of Michigan, in fact, published in a journal Labor Economics last year, you see increases in enrollment in vocational programs. When local labor markets see upticks in unemployment and people are choosing programs that are distinct from the sectors in which job loss is occurring. So we see young people making the right decisions about training. We also know that community colleges are severely underfunded. Look at the nature of that funding, really various prospects. One of the things we're going to try and understand is the variation in the quality of community colleges in terms of being able to provide that vocational training in response and then try and understand new places with higher quality community college systems in terms of that responsiveness do better in adjusting to job loss occurs. Okay. So I'll close. We'll have plenty of time for you. And after that, with what we're going to do and what we're going to do at just kind of summarizing what we're going to do is three things. One is kind of on the quantitative side, we're going to track the flow of resources that were into the various buckets of play space policy, workforce development, technical assistance, tax subsidies for firms, and then and then specific types of, of, of infrastructure development. And and this just kind of has to be done in force because it's the US, the Department of Labor does the same, Department of Education does the same, housing development, does this thing states do better things? In addition to that, we're working with an organization called Camden, which collects data on what foundations are doing, that this is about 2010. We want to know the volume of resources flowing through private philanthropy because they're very big in this space. We want to know, just in terms of the funding, is funding flowing to places that we that you identify as being distressed? Where can intervention look justified based on various theoretical fragments? We have the books and not super optimistic here can take one example just from the Trump administration. I don't want to pick on the Trump administration. That example can be replicated by Democratic presidents before Trump's opportunity zones created as part of the 2017 tax bill, provide subsidies for investments in low income ZIP codes. There are 8000 odd zip codes in the U.S. to qualify for opportunities. Less than 1% of them accounted for almost all of them. And those seem to be poor zip codes and rich. And that happened with empowerment zones under Obama. So it's not a left/right thing. It's just a poor policy design. So we want to know, are we missed allocating the resources that are on the table? And there's a silver lining to the answer. Going to that mean yes. It means there's money in the system that can be reallocated without having to optimize the spending. We also want to know about the complementarity or substitute ability funding, funding by state, federal and philanthropic sources. Do we see them going to different places, to same places? What's the potential there? So that's the quantitative and qualitative side. We want to understand the conceptual models that that these folks who are overseeing economic development have in their heads. And we want to understand that variation in practice across place. Helpful in this regard is that it's the US and the U.S. as a country and donors told us. So what we have are all sorts of membership associations in which these people belong. So they're all signed up. They have their websites. They're part of bigger umbrella organizations. And that gives us a paper trail in terms of the. And what we have found in our preliminary discussions. And then also in talking to people who just know this space much, much better than we do, is that civic leadership is really important. Like Birgitte Close in Grand Rapids, you don't necessarily need to have she doesn't have a lot of money in herself, but she has a big megaphone and she has developed confidence in people in the community that she can help orchestrate a strategy that makes sense. So we want to be able we want to train great places according to the quality of civic leadership represented by the economic developments that we have. That's really hard to do. So the version of that we actually pull off might be less ambitious, but we're going to see how far we can go. There we are. So this is just kind of what what complicates good practice is this jurisdictional alphabet soup of jurisdictions that exist in the United States. There are commuting zones that we take as our measure of local labor markets. These are collection of counties in which people live and work. You've got lots of work in labor economics that says the right way to think about. That we have the jurisdictions that are defined by policy. And so this is just three alternative jurisdictions at the federal government level wherein in bringing states into their workforce development boards. Some states have plenty of the states have one of their boundaries do not follow the definition of what the commitments are. Here are economic development. So this is Department of Labor's definition. Here are economic development districts defined by the Department of Commerce for you to get access to funding through the Small Business Administration, Economic Development Administration. And these geographies are different and these don't follow the local type. And then we have opportunity zones. These are defined according to zip codes. And and and and their logic is, is distinct skill the one. The one thing about this that is actually useful for empirical research is you're going to have this variation in jurisdictional fragmentation across place, which might be huge and due to just the happenstance or weirdness of how policy at the federal level and state. So one of the things we want to try to understand is, is that jurisdictional fragmentation an impediment to kind of effective local leadership? And finally, as we do all of this, aside from trying to think about estimating treatment effects like space policies or a little bit less interested in that, because so much of that is done, we're more interested in understanding what we think of as the more central policy question. And that is, if you have the right structures in place, do you adjust more successfully when. I mean, literature approaches this as what's that causal impact of policy on output. Why that to me is not the right policy. Policy question is when place why has a negative shock? What? What do you want your values to be for good outcomes over the next five, ten or 15 years? The other question is a fine question, but plenty well-researched. And we want to know that both in terms of these economic development operations, but also the other elements of a place, the quality and in college system and so on. So that's that's where we are now. And we're hiring and expanding and talking and learning, and we're going to be doing this for the next five years. That's what I'll be happy to turn to. Questions, please.

Attendee: Question is about the workforce development, especially in areas which are dominated by workers. So it seems to me that there'll be a limit to where you will take them. They will never move. So my question based on that isn't the solution to try to attract businesses that will. Employ the maximum number of unskilled workers?

Gordon Hanson: So I want you there to think about that. I want to kind of relax the definition of skill a bit and say to think of somebody who has a high school degree and they weren't a superstar in high school. But they weren't you know, they weren't a dropout either. And now we're thinking about what are the career tracks that are available to us. So if I can link that person directly or indirectly to affirm that is seeing expanding sales and rising productivity, whether that firm is headquartered there or somewhere else. Well, we know that being employed by a good the firm really matters for your long term prospects. So if I can get that worker with just a high school education, maybe some extra education and training into a connection to that firm, either as an employee or as an employee of a subcontractor for that firm that then puts that worker on this wage track rather than on this wage. And that what that means is that the accumulation of skill and experience and everything else varies enormously in terms of the possibilities that that are present. So what do we want that training to do? We want to allow that worker to make that job, but that's at the level of the individual. We know that that works on an individual by individual that we're more interested in. Is does it matter in the aggregate if I have better systems in place that do that? Does that make employers more willing to come in and say, okay, I'm confident that I can I can I have access to the workers that I need to grow and develop over time? The thing that the economic development people say again and again and again that the overwhelming constraint and in terms of their location decisions is talent and talent at multiple levels.

Tim O'Brien, Growth Lab Sr. Manager: And I have a lot of questions and thoughts because I'm a part of the Wyoming project. So I'll save most of those. Spread them out over time. But essentially, I think that we want to maximize the ability to apply what you learn over the next two years to try to act on that place. So but one of the early things that that is in the ether is the Infrastructure and Jobs Act and potentially later the Build Back Better stuff. So I just wonder if you could say a few comments about how you think about that. If you have any ideas of what could go wrong or what could go right with it and how does it interact with basically the everything you presented here?

Gordon Hanson: We are at an utterly bizarre moment in the practice of place-based policy in the United States as a consequence of COVID, because the US government has put this incredible largesse on the table relative to what budgets looked like before. And my ask, it's been so long, it's got to get spent. My guess is that the variation and the quality of things that get funded is going to be absolutely enormous. So in the Wyoming context, you have the advantage of Wyoming not being that big and smaller places. Smaller places tend to get a disproportionate share of the funding. That's just how because they have two senators and representatives. You have matters for how much funding you get in addition to your population. So it's an appealing function, and that means they get more than the population would indicate. So that means in terms of thinking about what Wyoming can do if they tap into the right sources of funding, making it possible to support the question of what you do it. So I'll tell you what we're doing in our work, and I don't know if this is going to have relevance for you. So a portion of what we're doing is also looking at places that are going to be disrupted by transition. And that helpful thing here is that energy transition. As I mentioned, it's been quite a long time. So one of the things I'm going to be working on this summer paper for the Aspen Economic Strategy Group is looking at fossil fuel specialized places in 1984 lots of country like and I'm going to be comparing those to observationally similar non-fossil fuel specialized places as of 1980. I'm not going to repeat the exercise for 2000. So I encourage you to for thinking about the Wyoming case is a. They were my comparatives. And the comparatives are not going to be like necessarily in Montana. And you want them to be comparators in terms of kind of the non-fossil fuel, stop it. But you want to look at the fossil fuel places, too, in terms of what they're doing. But you want to know what kind of the time path of these places has been because the accumulated history is super important if you're dealing with local labor markets in West Virginia that have been shedding jobs in coal since 1980, when the coal decline really began in earnest, it's a very different situation than than some of the places in Wyoming, some of the coal places they like 1980, the 1980, 1990 decline or 1989 to 2005 might not a matter. They might have been only whacked in 2010. And so you have less history underfoot. So that's one thing I would look at to understand the comparative economic development in the places you're starting in. In a clear sense, I would then I think the most important thing for workers coming out of fossil fuels is can you get retooled to go do something else? And the worker training and workforce development is essential here. So the quality of the community colleges and other institutions in the workforce is key. We're doing a bunch of stuff on this project and workforce is doing a bunch of stuff on this. We're happy to make our data available for the planning part and at the tell you how we thinking about grading these places in terms of but in terms of quality, I don't know anything about the community college system in Wyoming or whether it's an important part of training there. But it's it has to be the most obvious thing to look at because it's there. It's been doing it for a while and you don't need to create that to know about the institution.

Tim O'Brien: Thankfully, they're heavily interested in expanding the Community College Network and how to help it work better for the struggling and struggling places in the places that they expect to struggle in the future. So it'll be super interesting.

Gordon Hanson: To look at the various projects that Nieves for her foundation have done. They have that. They have they've been doing this in Austin, in a Paso, in the Lower Rio Grande Valley, in Louisiana, in Phenix, and now in Iowa. And you have really different experiences based on what the players have done in those places and what the quality of work training institutions.

(Part II)

Attendee: I want to follow up about the community colleges. I hear that they're very responsive to demand in local labor markets, but could they also act as a vector for new capabilities to spawn new industries in the way that like, you know, larger universities are able to do?

Gordon Hanson: I don't know the answer to that question. My first guess would be not because of that, you know, the hub of the what is their M.O.? Their M.O. is to take care of workers who are not ready for four year schools. This could be they need remedial education. It could be that families are just not willing to let them be that far from home yet. It could be that they've got to work part-time and provide some income before they go away. It could be they just haven't figured it out yet. And so they have to serve such a diverse collection of students and tasking them with being at the frontier is a big ask. Regional public universities are very different in the economic developer types are. The thing you commonly hear is that people who work in economic development, people who work in workforce development don't get along. Workforce development are producing workers and they want you to buy their workers. I don't care who you are. I don't care if you don't want them. That's what is our product by our products. Economic developers say we want to we're trying to close deals with companies. Those companies often don't want more. Workforce development, especially community colleges are different because they're not workforce development, which is has a different objective. They can be more impacted because they have to work with local employers to make stuff work. They don't. On average, they may not. I've no idea about the Wyoming context, but they're different in that respect. Economic, all the types also kind of a little skeptical about the value of our cities because, you know, just think about the incentives. What are the public universities about? They're not about helping local businesses expand. That's pretty far down their list of priorities. Now doesn't mean it can't happen, but you have to deal with the fact that you're engaging institutions whose incentives are not aligned with the incentives of promoting economic growth in Wyoming. What I encourage you to do is to look at what the Economic Development Administration has done in Wyoming over time. If they give out grants or technical assistance to business and they're out there and at least back to the mid-2000. All of it's just all online. They don't get that many their normal funding streams about $4 million a year. Right now they have 5 million. Okay. And we have a TaskRabbit who is working on something. Okay. This makes the difference.

Attendee: It seems to me like in the US context that the retention or attraction of higher-income workers to stay where you are is a really important complement to whatever you do on the business side, whether it's like building more housing, or some of the amenity stuff that you mentioned in Grand Rapids. And I wonder when you talked about earlier the difference between the focusing on moving people to like superstar cities versus focused on places. Is there room to do both or do we really need to pick one or the other? Because, you know, maybe the stuff that that Erik Prince did, doesn't work if we were able to relax this housing supply constraint in Chicago.

Gordon Hanson: Yeah, it's a really important question. And it's and there's even there's ever more murkiness about it. So if you had to ask me this. I would have said. And recruiting skilled workers to small towns is going to be really, really tough on the whole landscape. I don't think it shifted as a bit as much as folks would like to believe, because cities have operated as cities have for 6000 years. But we may have that. But what's happening on average doesn't matter for Wyoming as small enough. That's all it takes. If a piece of this happens to work for Wyoming, that could be a big deal. And the potential is there, as you know. Why why can't Wyoming have it? It's on Boise, Idaho. You know, I see no reason why not. And so there it's the research probably isn't that useful to you, as is understanding the attractiveness of specific places in Wyoming to the work-from-home crowd. Now realize what you're getting here because the folks and the policymakers in Wyoming might say, Oh, we're going to be part of the tech revolution in the silicon range. And that's not what they're going to get. What they're going to get is back-office stuff, folks and human resources folks and or get you onboarding folks who provides our clients services. But that's what Salt Lake City is telling us. That is doing that really well. That's what Boise's like. It's not Austin. They're going to get an awful lot of those are good jobs and the college-educated workers of workers and help create amenities that attract other skilled workers and potentially a very good deal.

Attendee: You mentioned the bad overlap between local labor markets and the administrative units. Do you think the best approach would be to improve that overlap or to centralize administration more?

Gordon Hanson: I don't know. I mean, so let's take a step back and say I think about policy advice we're giving to different policymakers. So if we're talking to the federal government, then obviously coordinating policy across different areas of practice is an important thing and we don't do that. You have Department of Labor, does its workforce development say that economic development, immigration does. It's then Treasury does it stuff through subsidized lending in low-income communities. And then Housing and Urban Development does its thing with empowerment zones and opportunities and so forth. So at the credible level, are you want to think about how you coordinate, knowing that what's going to happen is you even you know, the last thing you want to do is like create an administration to bring all this stuff together, right? Because you know that that will get blown apart. So you have to understand that the process by which this works is that each of these areas of government has enabling legislation that gets updated about once every ten, 15 years. That sets the framework. And you want to think about designing that legislation, working in concert with other important complementary areas of enabling legislation and getting the jurisdictions right. One thing that and I just is way out of my area of expertise, but you need to know that Congress is going to want to pull this stuff apart and make sure that a certain amount of money goes in. That's fine. But there's clear potential there for right now. We don't plan on spending a whole lot of energy talking to people at our level because our government is kind of shut down in terms of policy. So the policy advice that you then give at the state and local level is different because you have to take the federal structure and then the advice is conditional, may be conditional on the extent of fragmentation of it that you encounter.

Attendee: I have a question. All the places that you'll be studying, because it seems based on your example being mainly like urban areas or areas that previously have localized economic activities. But what about rural areas that for like some intrinsic to our geography, geography like disadvantage or they just don't have economy activities or like as a culture or things like that.

Gordon Hanson: So we're not that focused on agriculture because agriculture employs very few people. But keep in mind that in the U.S. after 1968, manufacturing left big cities and went to smaller towns. So a lot of the places we're talking about are places that have these broader communities. And these collection of counties have 20, 40, 60, 80, 100,000 people. So. So we have so where we are, a lot of the places we're focused on are smaller places. So is that rural? Well, it's kind of the rural proximity, and it's a different set of challenges than low-income neighborhoods in big cities. We're not really speaking to that. That's a kind of a different pathology. I know we're over time. I'm happy to take another question or two or maybe two.

Host: Let's thank Gordon again.