Dynamism and Stagnation: An Outlook

The ability to be flexible, dynamic, and responsive is particularly important during moments of shock, when outside forces are roiling the economy and society. Unprecedented external shocks, from the pandemic to wars to whatever other enormous event is happening geopolitically, define our current moment. At the same time, reacting to shocks has become increasingly difficult, because we have hamstrung ourselves. Our world is constantly changing but we have not positioned ourselves to respond flexibly to that change.

For twenty years after the Second World War, Detroit chugged along reasonably well despite have a relatively stagnant car industry. This stagnation became much more catastrophic when the industry faced global competition and radically increased gas prices in the 1970s. America's regions have been experiencing shocks, including from China, long before the shocks of the last five years. Yet some regions have responded to those shocks far better than others, partially based on local flexibility and entrepreneurship. In our largest cities today, the office markets face particularly large shocks due to the shift to working from home. Finding an effective response to that is going to be enormously difficult: our zoning codes have created islands of skyscrapers dedicated to offices with no other interspersing functions. Regulated single purpose streetscapes make it particularly hard for cities to be flexible and reshape.

We need dynamism and the risk of stagnation is severe. This talk’s patron saint is Mancur Olson, who forty years ago wrote The Rise and Decline of Nations. Here are two quotations:

Stable societies with unchanged boundaries tend to accumulate more collusions and organizations for collective action over time.

Distributional coalitions slow down a society’s capacity to adopt new technologies and to reallocate resources in response to changing conditions and thereby reduce the rate of economic growth.

In 1988, when I first read this book in graduate school, Olson’s prediction sounded a bit like New York City or parts of California, but we had Texas and Ronald Reagan. American as a whole did not seem so sclerotic. Thirty-five years later, Olson’s view increasingly describes America, not just pre-Thatcher England. This is the world that surrounds us, a world in which there are rules that constrain whether you can open a new business or build a new structure.

Responding to Economic Shocks

I am going to revisit territory that we already discussed at this year’s Austin Symposium: America’s incredible geographic disparities. This figure is from a paper I wrote with Larry Summers and Ben Austin. I believe that America's largest unsolved social problem is the rise of prime age joblessness, where the census defines prime age as between twenty-five and fifty-four. When I was born in 1967, roughly one in twenty prime age males were jobless. For most of the past fifteen years, more than fifteen percent of prime age males have been jobless. This is a big problem because every available piece of evidence shows that for men in particular, joblessness is vastly more catastrophic than being a low wage worker.

A job is not just about earnings: it is about purpose and social connection. The evidence shows that there are dangers associated with joblessness, including misery, suicide, and opioid abuse. Compared to being a low wage worker, joblessness appears to be catastrophic.

I differentiate between men and women, because women not formally in the labor force are still working by providing family care and incredible social value. I know you have seen articles on The New York Times website that praise the wonderful house husband, who is doing great work caretaking for his kids and parents. But the American Time Use Survey shows that jobless men have not increased their time in caring for others to any meaningful degree. They do, however, watch much more television. The New York Times has found an ideal, but it does not describe the norm for unemployed American men.

And male joblessness is not spatially neutral. There is a joblessness arc that courses through America's Eastern heartland. It begins down in Louisiana and Mississippi and runs up through Appalachia. It ends in the American Rust Belt, where often more than one in four prime age men is unemployed. This catastrophic situation reflects a failure to respond to different two shocks. First, these areas did not respond to the deindustrialization shock. (In some places, this is also the China shock.) The second shock is generous social benefits, including state level social benefits. Sometimes these men’s girlfriends will bail them out or their parents will let them sleep on their couches. More than a third of prime age men are living in their parents’ homes. There is an informal social safety net that accompanies the federal social welfare assistance and helps make this possible.

These two shocks have led to joblessness. Two entrepreneurial failures have meant that our country has not been flexible enough in response to those shocks. One failure is that entrepreneurs have failed to create jobs for less skilled men who are not otherwise working in their home regions. The second failure is that builders have been unable to make enough space for them in the country’s most productive areas. There are jobs in San Francisco, Los Angeles, and Atlanta, but these men are not moving to these productive places partially because the cost of living is too high.

These are pictures of often sympathetic-seeming groups that figured out how to organize to block building in America. On the left are the pioneering leaders of the Save the Bay Foundation in greater San Francisco. One of these women is the wife of Berkeley’s former Chancellor Clark Kerr. On the right are activists hoping to save Penn Station.

In America’s lower density areas, the language of NIMBY-ism—not in my backyard-ism—is environmental. This NIMBY-ism continues even when not building is worse for the environment than building. My work with Matthew Kahn finds that carbon emissions are vastly lower in coastal California because of its innately moderate temperature. If we really wanted to lower America’s carbon footprint, we would build energetically by the San Francisco Bay. But too many environmental activists have decided just to say no to building.

By contrast, NIMBY-ism in the urban core focuses on historic preservation. Regardless of how architecturally undistinguished a building is, someone will argue that it needs to be preserved, and the neighborhood needs to be protected from development. This provides the ideological cover for saying no to change. It is not obvious how much of this antipathy towards change is rational. Some people do not want their home prices to decrease, and others want to avoid the inconvenience of new people moving into the neighborhood. Some of it may be blind fear of the unknown.

A crucial element is that highly educated people are particularly good at preventing building. Consequently, the barriers to growth are most severe in America’s most educated areas. Those educated regions, however, would be the best places to build if you wanted to promote upward mobility. Along this figure’s horizontal axis, I have ranked areas based on Raj Chetty’s measure of how well poor children are rising into middle income adulthood. Along the vertical axis, I show how restrictive the building code is according to the Wharton Urban Land Use Survey. The areas that would offer the most upward mobility are also those that say no to building most often.

I now want to highlight how land use restrictions restrict American dynamism. I took this graph from a Journal of Economic Perspectives article that I wrote together with Joseph Gyourko. Along the horizontal axis is the amount of new construction permitted in the metropolitan area between 2000 and 2013 relative stock of housing in 2000. Along the vertical axis is the gap between how much it costs to buy a house and how much it costs to build a house. The number three, for example, shows that in Honolulu, buying a house is three times more expensive than building a house.

This graph shows that there is no way to repeal the laws of supply and demand. If demand for an area is robust, then this will either lead to more building, higher prices, or both. Making building very easy, like Austin has traditionally done, leads to wildly increasing population and moderate prices. Blue state Massachusetts may claim to care about affordable housing, but red state Texas actually provides vast more successfully affordable housing, just by unleashing private developers. There is plenty of demand to live in San Francisco and Los Angeles, but because it is so difficult to build there, that demand shows up in higher prices, not more homes.

The Slowdown in Construction Productivity

(From joint work with Leo D’Amico, Joe Gyourko, Bill Kerr, and Giacomo Ponzetto)

In my past work on land use controls with Wharton’s Joe Gyourko, we refer to the gap between building costs and purchasing costs as the “regulation tax.” In our work, we found that about one-half of a New York City apartment’s cost is due to regulation. That is the gap between how much it costs to add one more story to a building and how much that extra space is worth in the market. But what if regulations also shape the physical building costs? What if, over the long haul, we have created a building industry that is far less dynamic and productive than it could be because we have created these rules make it difficult to efficiently supply new housing?

This figure, which is from a working paper I jointly authored with Leo D’Amico, Joe Gyourko, Bill Kerr, and Giacomo Ponzetto, shows the ratio between housing units built and construction sector employees over the long haul. The graph shows that there is a period from 1940 where homes per worker remained relatively flat. Between the late 1940s and 1970, homes per worker soared. But since 1970, homes per worker has fallen, which echoes Austan Goolsbee and Chad Syverson’s important work on productivity in the construction sector.

What happened? In the immediate post-war period, mass production made the construction sector more productive. The Levitts and other builders were copying Henry Ford and figuring out how to reap scale economies in housing production. They did not build the home in factories, but rather on site. Still, these builders figured out how to create something like an assembly line, where construction workers went up and down the street doing their bit for each house and then moving on. This enabled builders to supply the housing that returning American soldiers were eager to buy, especially with the help of thirty-year Veterans’ Administration mortgages.  

But starting in 1970, the projects start getting smaller. Communities get much better at saying no, and the available land parcels tended to shrink a bit. As the projects get smaller, the companies get smaller, and productivity decreased. Smaller companies invested less in R&D, and the whole industry stagnated. In our paper, we highlight the difference between regulation of entry, which means that you have a few dominant firms such as Detroits’s big three in 1950, versus regulation of process, which means that instead of having big firms and big projects, you get small firms and small projects.

The next set of figures make the point that America’s builders are tiny. The first figure shows the size of establishments that make single-family home construction and the size distribution of establishments in other core economic sectors. Most industries have most of their employment in establishments with more than 500 employees, but most workers in single-family residential construction are in establishments with fewer than ten workers. These are tiny firms. It is unreasonable to expect mom-and-pop operations to effectively build nearly all our single-family homes.

The next figure is a map showing the size distribution of residential, single-family builders across American states. The dark blue states average more than 4.56 workers per establishment, which is the national average. In Texas, there are more than 4.5 workers per establishment, but all the lighter places have fewer than 4.5 workers on average in the median establishment. Those are tiny firms.

The next table shows the correlation over space between one measure of regulation, the Wharton Residential Land Use Regulation Index (WRLURI). The top panel shows correlations with the actual WRLURI measure. The bottom panel shows correlations with a projected regulation index, which is formed by using demographics to predict the index where we have the index and then forming predicted values for a much larger range of metropolitan areas. All four columns tell a similar story: the most regulated areas have the least employment in large construction firms.  

Higher housing regulation levels are associated with smaller firms.  As a direct measure of innovation in construction, we look at patenting activity in the next figure.

The red line shows patenting in construction. The green line shows patenting in manufacturing and the gray line shows other industries. In all cases, we have indexed patenting levels so that all three series equal one in 1930. The three series track each other almost perfectly through the late 1940s. Between 1950 and 1970, a slight window opens, in which construction begins to lag. Then the series completely diverged after 1970, which is compatible with the view that construction became far less innovative after that time period.

As a final look at U.S. construction productivity, this table uses data from the company Turner and Townsend and compares the costs of building in different U.S. cities and a few comparison places. This data is meant to reflect the physical cost of building a twenty-story office building in different regions of the world. There is a big gap between less regulated Houston, where pure physical construction costs are cheaper, and more regulated San Francisco or New York. Our costs are also higher than our global comparison cities. One reason why America is less dynamic, is that we have gotten much worse at building structures and infrastructures. Regulatory sludge bears some blame for this, but our desire to ensure no one is ever inconvenienced is also a partial cause. We blanket the builders and procurement agents with rules, which makes it difficult to be nimble and innovative.

Responding to Disease-Related Shocks

Let us turn to our ability to respond to disease-related shocks, like the COVID-19 pandemic that hit the world in 2020. Cities have been enduring pandemics for thousands of years. Our first well documented urban pandemic is the Plague of Athens which began in 430 B.C. Fifth-century Athens was doing everything you could imagine wanting a city to do. A genius stood on every street corner. Euripides, Aeschylus, and Sophocles were writing dramatic plays that still resonate with us today. Phidias was revolutionizing sculpture. Herodotus and Thucydides, the two fathers of history, lived and worked there. Democracy itself was being born and on top of that, Athens was an economic and military powerhouse.

All this success excited the envy of Athens’ land based, non-urban, highly militarily successful rival: Sparta. Sparta demanded that Athens stand down from its leadership of the Delian League. In his famous Funeral Oration, a speech that would have made Sam Houston himself proud, Pericles tells the Spartans that he would guff from no man, and that kickstarted the Peloponnesian War.

Pericles’ plan was to summon the Athenians and their Attic allies behind Athens’ walls, trusting those walls to keep out the superior Spartan army. Then he would send the vastly superior Athenian fleet to harass the Peloponnese coastline. As military strategy, it worked perfectly well: the walls held against the hoplites. But walls that can keep out enemy warriors may not manage to keep out a virus or bacteria.

Cities are vulnerable to pandemics for two reasons. First, they are the nodes on our global travel and trade lattice; they are the entry ports for goods, people, ideas, and viruses. It was so in New York City in 2020 and its was so with the Athenian port of Piraeus in 430 B.C. Something entered the city, probably from Asia, and started laying waste.

Cities’ second vulnerability is the proximity of its inhabitants. People living close together can often be an asset, but it also enables easy travel for bacteria and viruses. Thucydides described a city that had run amok, in which people live only for the day, because they expect to die tomorrow. The plague caused the deaths of around twenty-five percent of Athens’ population, a death rate one hundred times that of COVID-19. Athens’ greatness allowed it to keep fighting the Peloponnesian War for another twenty-three years before losing, but in some sense the plague dimmed its luster forever. When Athens emerged from the war, it slipped from being the New York City of the eastern Mediterranean to the Philadelphia, and then the New Haven.

Over much of the past 650 years, our cities have been relatively resilient to pandemic. This figure shows the path of New York City’s death rates over the last two hundred years. The early 19th century was a period of proto-globalization and proto-globalization meant proto-plague. The early decades of the 19th century experienced yellow fever, a mosquito borne illness that emerged out of Africa in the late 17th century. Sailing ships carried it along to the Caribbean and from there it moved north to East Coast cities. Yellow fever’s death rate was two and a half percent, making it ten times more fatal than COVID-19 and one-tenth as fatal as the Plague of Athens.

Another plague was a particularly virulent strand of Cholera, which emerged in the Ganges Delta in 1817. The British Army carried it over the sea and from there it made its way to North American ports. Cholera’s death rates reached almost five percent.

But despite these death levels, cities continued to grow. They were just that productive. If your alternative to living in a city was dying of starvation in Ireland, facing the plague risk in New York might not be too bad. Cities also continued to grow because their governments started investing in infrastructure that mattered. In Survival of the City, David Cutler and I highlight this as a hinge of history. This is the point at which governments started doing something other than killing people. Eighteenth century and earlier Western governments were basically in the death business.

As much as we may celebrate Frederick the Great and Voltaire’s eloquent correspondence or his patronage of the mathematician Lagrange, we must remember that Frederick the Great's day job was waging war and stealing Silesia from the Empress Maria Theresa. But starting in the 19th century, city governments suddenly started investing in clean water, sewers, and other infrastructure improvements that save lives. As a Chicago Ph.D., I am predisposed to the view that there are many things that governments should not do, like overregulate building or entrepreneurship, but there are certainly things that governments should do. Reducing the risk that your citizens will die of Cholera seems like a reasonable task for governments to undertake.

The health-related investments of the nineteenth century were functional. The were the product of a bottom-up movement, in which leading citizens built coalitions to support infrastructure improvements. Bond markets made it possible for cities to borrow enough to make these investments. Another theme of this talk is that we want our governments to make our roads smoother, to effectively defend America, and to improve our education systems, although I am certainly open to vouchers and charter schools. As a result of these 19th century investments in clean water and sewers, cities’ health improved. Indeed, our cities became so much healthier between 1919 and 2019 that we almost forgot that cities could be killing fields. Then we suffered another plague.

What is going to protect us against 21st century pandemic? Amazingly, education seems to be as protective against illness as it was earlier against the deindustrialization shocks of the 1960s and 1970s. This figure shows the relationship between total COVID death rates from 2020 to 2023 and the share of the adult population with a college degree. The amazing thing is not that there are fewer deaths in more educated metropolitan areas, but that the slope is so extreme: there is a four-fold difference among cities like Las Vegas, San Antonio and Oklahoma, where roughly four-tenths of one percent of the population died, to cities like San Jose and San Francisco where one-tenth of one percent of the population died. More people died in Seattle of fentanyl overdoses during the COVID pandemic than died of COVID.

This negative relationship also reflects the fact that more educated cities had the most extreme work from home practices and policies, which may have saved some lives, but also created an economic hangover. The education gap in working from home is enormous. In May 2020, during the pandemic, 68.9 percent of Americans with advanced degrees were working remotely while only five percent of high school dropouts were working remotely. What downtowns are still the most underused today? The hyper-educated metropolises, like San Francisco and Seattle.

This figure shows data from Kastle Technologies from February, 2024. This measure of office vacancy is based on key swipes for people entering large office buildings. These fancy downtown offices are still at about fifty percent occupancy relative to pre-pandemic. Nationally, work from home is a much smaller phenomenon. About ten percent of workers nationwide are permanently work from home and around twelve percent are partially work from home.

Cities will manage to respond to the office vacancy shock only if they are flexible enough to change.  This is not the first time America’s cities have been under stress. In the 1970s, all of America's older, colder cities were heading to history’s trash heap because of massive changes in transportation technology. Transportation costs shaped the destinies of all America’s older cities. Each of the twenty largest U.S. cities in 1900 was on a major waterway. The oldest—New York and Boston—were built where the river meets the sea. Minneapolis, the newest, was built at the northernmost navigable point on the Mississippi River. These cities grew up as nodes on this great transportation network, but over the course of the twentieth century, transportation costs plummeted.

The inflation-adjusted cost of moving a ton per mile by rail fell by ninety percent between 1890 and 2000. Container ships, the highway system, and air freight were causing even more amazing things to happen. Detroit's location with great waterways, railroads and proximity to America’s agricultural heartland was a huge bonus in 1910 when Henry Ford started modern manufacturing. By 1960, however, Detroit was just a really cold place. Americans relocated.

January temperature is the best predictor of twentieth century metropolitan area growth for several reasons. Warmer places are more likely to be right-to-work states and they attracted industry after World War II. Tom Holmes’s work found counties on the right-to-work sides of state borders experienced large increases in industrial employment after 1947 relative to neighboring counties in pro-union states. Warmer places have also made it much easier to build. But let's face it: Americans don't like driving on the ice for three months out of the year. As a New England parent who raised three kids amidst all that ice, that it shows a lack of character on America’s part not to like those icy roads.

These are iconic images from my youth. This is President Ford in New York during the 1970s fiscal crisis. It reflects that in addition to deindustrialization, crime had gone crazy. Fiscal responsibility was completely lacking. President Ford quite understandably refused to bail the city out and it really looked as if it was heading for catastrophe. The other photo is Jimmy Carter wandering through the wasteland South Bronx had become. The city looked doomed.

In 1971, two jokers put up a billboard on the highway leaving Seattle, asking the last person to leave to please turn off the lights. Just as no one could imagine Detroit with a smaller General Motors, no one could imagine Seattle with a smaller Boeing, and Boeing had been cutting back on jobs. But Seattle recovered, and urban density came back.

In 1980, Alvin Toffler, the hip futurist of his age, wrote a book predicting that all this fancy information technology—fax machines, personal computers—would make urban offices just as obsolete as containerships had made New York’s garment district decades earlier. He looked forward to a world in which we would have empty skyscrapers that we would have to occupy as warehouses. His arguments were similar to what we have heard in today’s urban doom loop literature.

That did not happen, and instead, some American regions experienced an urban renaissance. The global economy’s tidal forces came to favor face-to-face contact. There has been a vast rise in the return to skills and innovation. As a social species, we get smarter when we live and work around other smart people. If you do not know that face-to-face contact is incredibly valuable for learning, then you are not taking advantage of UT Austin. Our universities are designed to be machines for collaborative learning and innovation. Globalization adds to this by radically increasing the returns to innovation, partially because any great idea can be sold or built anywhere on the planet.

The left image is the wall-less office at Mike Bloomberg’s City Hall, which is based on the wall-less office at Bloomberg LLP., which is based on the Salomon Brothers trading floor. Jamie Dimon was eager to bring trading floors back during the pandemic, and trading floors provide a great example of why face-to-face contact is so valuable: a little bit of extra information can make you a fortune in minutes on a trading floor. Finance is all about knowing what is going to happen. Traders learn a great deal amid the action. Neither COVID nor Zoom changed that fact. If information technology were making face-to-face contact obsolete, then why would Google buy the Googleplex? Why would they buy millions of square feet in downtown Manhattan if they did not think it was important to have young workers working together and learning from each other?

These benefits of face-to-face interaction did not enable Detroit, Cleveland, or many of the older, colder cities to recover from deindustrialization. They never reinvented themselves in the 1970s. If a city does not have Texas’s pro-business environment and sunshine, it at least needs to have skills. This figure shows the relationship between share of the population with a B.A. and earnings. As the work of Jim Rauch and Enrico Moretti shows, the place-level connection between skills and economic success does not just reflect the fact that your skills make you more productive, it also reflects the fact that your neighbors’ skills make you more productive. Economists call this phenomena “human capital externalities.”  

Typical estimates are that as the share of adults in your metropolitan area with a college degree goes up by ten percentage points, your earnings also go up by ten percentage points, holding years of schooling constant. If you live in highly skilled environments, then chances are that your neighbors have taught you something smart, or helped you find a job, or bought something from you. Having skilled people around you is valuable.

This shows that population growth has concentrated in highly skilled areas. The shift towards skilled areas has been particularly important in the colder northern parts of the country. The skills that matter are not necessarily the skills that schools teach; we tend to focus on those skills because they are measurable. In some sense, the most important skills for urban regenerations are those talents that relate to running a business, producing new innovations, and being functional in the world. Our colleges do not necessarily teach those.

Sixty years ago, economist Benjamin Chinitz compared New York and Pittsburgh and noted that New York appeared to be more resilient than Pittsburgh even then. This was because, he argued, New York had an entrepreneurship culture that had been inculcated in the city’s massive garment industry. The garment sector was the U.S.’s largest industrial cluster in 1950s, significantly larger than Detroit’s automobile industry. Moreover, garment production was unlike automobile production because there were no barriers to entry and very weak scale economies. Consequently, anyone with an innovative idea and a couple of sewing machines could start a business, and the garment industry became a school for entrepreneurs.

These entrepreneurs were flexible; they would start with garments and then start a movie studio, or build skyscrapers, or they would open a bank, because that's how entrepreneurial human capital works. It created a place that was flexible when shocked, which is of course this talk’s fundamental theme. Even though New York City had a government that was dumping regulatory sludge on everyone, it had human capital that was maximally adaptive, which is what enabled it to change. Today’s question is whether New York still retains that spirit sixty years later.

It is remarkable, given how mediocre our measures of entrepreneurial human capital are, that they predict urban growth so well. Economists use two measures of entrepreneurial human capital: average establishment size in an initial year and the share of employment in new establishments. Both are enormously predictive of which cities managed to have employment growth over the next thirty-three years. There is a fourfold difference between the places with the biggest establishments and the places with the smallest establishments. In past papers, I have documented that this is not just a result of industry choice or regional choice, but it reflects a real tendency of places with lots of nimble firms to add employment more quickly.

My prognosis is that Zoom is not going to mean an end to face-to-face contact, but it certainly does mean a more competitive talent environment. Imagine a hypothetical San Jose-based fifteen-person startup. I do not think a single employee would say “since we can Zoom, we should never bother to meet in-person again. We’ll go to our parents’ houses and save on housing costs.” What seems more likely to me is that some employees would decide they want to live in Austin to avoid paying state taxes and still be around skilled people. Some would move to Vail because they like skiing or some to Honolulu because they like surfing. The market for talent has never been hotter. This is what our local governments face: an environment where they need to figure out how to get physical or regulatory systems that were designed for a different world to adapt to this new, highly competitive setting.

So which cities managed to do this well? We start the second edition of Survival of the City by measuring which cities did well and which did poorly. We chose four measures, primarily because of their relatively high frequency: changes in earnings and employment, which is coming from the quarterly census of employment wages; repeat home sales prices, from the FHFA repeat sales index; and permit data, meaning how much new construction is happening, including quantity of permits, prices, and quantity of employment wages. Then we divided these factors by the standard deviation of those factors across areas and created a z-score. And then we just took a simple arithmetic average of those four series.

What are the patterns? January temperature is predictive of positive outcomes, like employment change. We expected this because people want to be in safe, warm areas where they can spend time outside. Places with higher January temperatures also had fewer lockdown rules and enough people probably valued that too.

But these arguments cannot explain the positive relationship between January temperature and earnings growth the next figure shows.

If people just want to live in warm areas, then the number of people moving to an area should be going up, but wages should be going down. A labor supply shift caused by attractive amenities is supposed to cause populations to rise and earnings to fall. But the fact that wages are going up is telling you this is not just about wanting to live in warmer places: firms also want to produce in warmer places. Warmer parts of America seem to have proven to be more dynamic and more adaptable in both the labor demand side and the supply side.

For most of the past forty years, skills have strongly positive predicted housing price growth. Over the COVID period, however, the opposite is true, perhaps precisely because of the link between skills and adaptability. Skilled workers have used Zoom to move away from expensive places like San Francisco, San Jose, Boston, and Washington D.C. Skilled workers have also experienced the largest shifts to work from home, which may make them less willing to pay for real estate that is close to their jobs.

Austin is the cover image on the hardcover edition of our book. Our editor did not choose it because he knew the numbers, but because he just knew that special things were happening here. Austin is so remarkable because it combines tremendous human capital with Texas’s relatively pro-business environment. Those two factors make it a Sun Belt superstar.

When we put together our entire index, Austin ranked first by a considerable margin. One reason that Austin is so high on this chart because of incredible housing price growth. This will probably mean-revert at some point in time because it's relatively easy to build here. But changing housing permits way up relative to robust employment growth shows that most of the top twenty-five cities are in the Sun Belt.

The Sun Belt superstars completely dominate the top twenty-five. There are a lot of Rust Belt cities on the bottom twenty-five, but the real bottom are places that you wouldn't normally think of as being catastrophes: New York, for example, is down there. For most of the past forty years, New York has been soaring over most of the Rust Belt because its vast office market was an asset, rather than a curse. Now that vast, almost segregated office market has become a curse, in part because it takes people so long to get to work.

Washington D.C. is another example. Government is a stable employer, but for the last three years, it has been catastrophic for the city. I suspect the Federal government has fairly generous work from home policies. For many government agencies there's little competitive pressure to get people to show up. Of course, San Francisco is on the absolute bottom.

How can local governments respond to this dynamic and become flexible? In a world where talent is mobile, you need to provide an environment where people want to live. This is the idea of the consumer city: you want to attract people to vibrant downtowns. Government’s job is not to micromanage industrial growth; it is to attract and train smart people and then get out of their way. That is not laissez-faire, because attracting and training smart people requires public safety, the management of city streets, and other core public services.

One of the ways that governments can make cities more attractive is to stop over-regulating urban entrepreneurs. An appalling thing about America is that we regulate the entrepreneurship of the poor so much more tightly than we regulate the entrepreneurship of the rich. If you want to start your internet phenomenon in your Harvard College dorm, you can have a billion users before any regulator knows you exist. Starting a small convenience store that sells milk products requires dozens of permits. People without fancy computer science degrees tend to be entrepreneurs in the real world, not cyberspace, and this leads to a very different level of regulatory oversight.

Governments must provide safety. In the 1960s and 1970s, New York decided they wanted to run a local welfare system. New York’s leaders wanted their cops to stop being mean to people, which is entirely understandable. But there is danger in a leftward local shift at a moment when exit—to suburbia and the sunbelt—had become easier. Companies do not like paying high local taxes. Upper middle-class urbanites do not like crime. By the thousands they voted with their feet and left cities.

I fear a repeat of this phenomenon today. There is an understandable desire to right historic wrongs, but remote work and the ability to relocate makes cities especially vulnerable. Cities must prioritize public safety and quality of life.

In the aftermath of the terrible murder of George Floyd, a bevy of unwise public policies came to the fore. The defund the police movement made little sense then or now. The idea that it was acceptable to have autonomous zones where the police did not interfere is similarly bizarre.

The right lesson to take from George Floyd’s tragic murder is that we must expect our police to fulfill two mandates. The first mandate is that police must ensure our safety when we walk the streets at night. Typically, the rich are at less risk from crime, and so failing to stop crime particularly harms the most vulnerable urbanites.

The second mandate is that police must treat all our citizens with respect. We are all worthy of dignity. We want to make sure that our police are respecting people. This dual mandate for police—safety and dignity—is doable. I am proud of the way the Boston Police Department has evolved over the decades to manage to deliver both things. They were very different in the 1980s. But this evolution occurred not by spending less, but by spending more.

There is no free lunch in government service. If you want a dual mandate, you are going to need to spend more rather than less. You are also going to need to manage the force better and measure not just crime levels, but how the police treat citizens. Cities should also consider engaging in audit studies in which you check how the police are treating people. Body cameras are a great tool for monitoring officer behavior.

Managing our Infrastructure

(Joint with Lindsey Currier and Gabriel Kreindler)

Road quality and infrastructure are significant governmental tasks. Infrastructure may be less exciting than other topics we have discussed, but our infrastructure’s quality matters. For the last forty years, the American Society of Civil Engineers has told us that our highways had been getting worse. In the 1980s, they gave us a B-. Now they have downgraded our score to a D. The Department of Transportation, however, has been sending out trucks at night with accelerometers to measure road quality. They have shown that our highways have gotten a lot smoother. Over the same period that the ASCE has been arguing that we need to spend trillions and trillions of dollars more on infrastructure, our highways have gotten smoother.

In my daily commute, my highway drives are smooth, but when I get off the highway, particularly after months of winter, the roads are rough. Yet we have not had widespread measures of bumpiness until now.

We have gotten access to the universe of Uber drivers’ accelerometers’ data for August 2021. Our cell phones have accelerometers that re-right your cell phone when you start tilting it in different directions. All the Uber drivers beam that information back to Uber headquarters where their engineers have figured out how to filter from the data a measure of much the phone is bumping up or down.

From the data we can get a usable measure. This is a time series of a single drive from downtown Chicago to O'Hare: bumpy at the beginning, smooth on the highway, bumpy at the end. We can then go from there to measures of local road roughness across America. From this, we can tell you which metropolitan areas are rougher.

The coasts are generally rougher than the interior. Houston and New Orleans are terrible, possibly because of the clay in the soil, which absorbs water, expands, and ruins the road. We can then estimate how much American drivers dislike bumpiness. We all can smooth our drives by slowing down, but that is costly because we waste time. The extent to which drivers slow down in response to salient changes in roughness generates a measure of the willingness to pay to avoid roughness. We looked at town borders to calculate the relationship between speeds and roughness.

This figure represents the border between Chicago and its wealthy neighbors like Evanston. The richer suburbs’ roads are much smoother than Chicago’s. This should be clear from the shift in “predicted roughness” on the left. On the right, we show the shift in speeds when the roads get smoother. From that shift, we can estimate a willingness-to-pay-for-smooth-roads. We are not measuring  a total social willingness to pay, but the drivers’ dislike of bumpiness. There may be positive or negative externalities to others from driving more quickly.

These are our estimates. The baseline cost of driving is about seventy cents per mile, which represents the drivers’ time plus the gas cost. Bumpiness typically adds about fifty percent to the cost. We can then correlate the bumpiness costs with neighborhood income and demographics. We estimate that if a household drives about three thousand miles per year on local roads, then moving from a one hundred percent white neighborhood to a one hundred percent African-American neighborhood is associated with drivers experiencing $318 worth of extra pain in road roughness. These differences are almost all across town rather than within town. Poorer people live in towns and towns tend to have rougher roads.

In big cities, typically rich people have roads that are just as rough as poor people. This figure shows New York City. The lighter areas are really bumpy. In Tribeca, the area close to Wall Street, there are very rough roads, and they also have really rich people. Often rich people live near and in the urban core to save commute times, but those are also the areas with the highest number of trucks and the bumpiest roads. In Chicago, you have a slightly different pattern, because there is a truck-intensive industrial area in the west that is also really bumpy. There is more of a gradient with poverty there.

It is possible to fit this into a cost benefit analysis and figure out how much we should be repaving, but it turns out there is a much simpler way to improve America’s roads.

This figure shows the probability of being repaved over the eight months after we measured roughness as a function of roughness at a pre-period. In three of the four, there is virtually no correlation between probability of being repaved and the initial roughness. We seem to be doing nothing to target rougher roads in these cities. The simple formula for making America's local roads smoother is to be smarter about where you repave. Even in New York, where there is a slight correlation, it is weak.

We surveyed 120 cities and towns. Typically, towns decide when to repave their roads by having some measure for documenting roughness, which may be a visual measure that is only weakly correlated with speeds. Then they have a mandate which says all the roads that fail on this test are supposed to be repaved. Now what percent of roads are repaved in these cities? Sixty-two percent of our U.S. sample of cities say that less than thirty percent of the roads that need repaving are being repaved. Our discussions with public works groups suggests that there is little targeting of that thirty percent. The mandate is unrealistic. Our big less for improving American infrastructure is just to repave the bumpy roads first.

Conclusion

Dynamism in America requires two changes. First, governments need to decrease unwise regulation, including regulation of large or small businesses and building. Second, our governments need to do their work more effectively.

For these changes to occur, our citizens must ask their governments to do more. A simple example of government underperformance is poor targeting of road repaving. But we also need to figure out why our roads are so expensive. We care a lot about eliminating corruption in road contracts, but we care not at all about reducing prices. We end up adding a lot of sludge that makes it impossible to get more competitive bidding and firms’ easier entry into this process.

Policing also needs to improve. We need to make sure that our streets are safe, and that simultaneously, our police treat everyone with respect. We should incarcerate fewer people for minor crimes.

Public schooling is also a major problem. We need to ensure America continues to have the level of opportunity that we want it to have. Education has been our traditional engine for that. I am open to other engines, but skills are the bedrock on which individual town, city, or national success rests. We are not going to be a strong nation unless our skill levels lead the world.

But here in Texas it is impossible not to feel confident and optimistic. Despite the challenges we face, we should not count this country out yet. Tremendous dynamism is possible. But we need to take on the hard work of tearing down the regulations that make us weak and improving the quality of public services that will support our natural strengths.

Readers are asked to make allowances for what was originally an oral presentation delivered on May 8, 2024 at The Austin Symposium.

Ed Glaser is the Fred and Eleanor Glimp Professor of Economics at Harvard University.