A Techtonic Shift In Economic Geography

Remote work weakened Silicon Valley's grip on the US economy. In early 2022, the Federal Reserve Board finished the job:

Because the Fed [had] distorted asset prices by making cash free, tech companies did not focus on cash earnings. Now we’re in an environment when cash earnings and dividends matter. Tech was in an environment that incentivised growth. Now all that has changed.

Big Tech's growth paradigm concentrated venture capital near Stanford University. This 20-minute rule informed what economist Enrico Moretti titled, The New Geography of Jobs:

If you look at the economic map of America today, you do not see just one country. You see three increasingly different countries. On one hand there are cities like Seattle, San Francisco, Raleigh-Durham or Austin, with a strong innovation-based economy and workers who are among the most creative and best paid on the planet. At the other extreme are former manufacturing centers like Detroit, Flint or Cleveland, where jobs and salaries are plummeting. In the middle, there is the rest of America, apparently undecided on which direction to take.

Regions are either cool city tech hubs or Rust Belt. Everywhere else was in the midst of economic restructuring. Moretti didn't offer much hope for the places still relatively dependent on manufacturing employment. But "undecided" America could aspire to attract young, college-educated adults. The best someone in Cleveland could do was move away.

Remote work threw the old geography of jobs a bone. VC firm Andreessen Horowitz moved to the cloud (but not out of Silicon Valley). Not only could someone living in Cleveland work in Silicon Valley, a startup could stay in Cleveland and thumb its nose at the 20-minute rule. In this sense, the forces of economic convergence were gathering.

However, Silicon Valley's Detroit moment was not yet at hand. During the pandemic, Big Tech got drunk on talent. Telecommuting cuts both ways. Apple or Google could offer software engineers in the Rust Belt San Francisco wages. Local companies could not compete in that labor market. For a spell, the growth of Facebook seemed limitless.

When the Fed started jacking up interest rates to tame inflation, the music stopped playing for startups first. Instead of rapid growth, venture capital went searching for positive cash flows. Then investors turned their attention to public companies and, eventually, Big Tech. Perhaps Amazon could produce more value with less workers. If you read Moretti's book or understand what happened to US manufacturing, that should sound familiar. The new geography of jobs is starting to resemble the old geography of jobs.

Annotating an Interview With Pittsburgh Mayor Bill Peduto About Uber

I'm commenting on this interview with Pittsburgh Mayor Bill Peduto in the Washington Post, "Inside Pittsburgh’s driver’s-seat move on Uber’s autonomous-car testing." I'll post the question asked and then annotate Peduto's answer using the lens I helped to develop in "From Metal to Minds: Economic Restructuring in the Rust Belt."

Q: Pittsburgh will be the first U.S. city to have commuters riding in self-driving cars. How did it happen?

Burgh Diaspora Comment: Why did Uber think of Pittsburgh as the site of a contemporary Manhattan Project? What's the attraction? Pittsburgh is one of "five significant cities in the world" with an artificial intelligence knowledge cluster. Uber already has a presence in one of them, the Bay Area. The other three are Boston, Zurich, and Beijing. Recently, professor Ruslan Salakhutdinov left the University of Toronto for Carnegie Mellon University (located in the City of Pittsburgh) because the machine learning program at CMU, as he put it, is "a huge powerhouse." If Uber is a moth, then CMU knowledge production is the flame.

Q: Then what happened?

Comment: The relationship between Uber and Mayor Peduto evolved very quickly, over the course of a year. Both the City and the State of Pennsylvania have been accommodating, forming what the Washington Post characterized as, "Pittsburgh might be the exact environment that innovators love to leap into — a legal void that can be defined by technologists, not bureaucrats."

This attitude of deregulation is a product of the political familiarity with university research at both the city and state level. Before serving as Pittsburgh's mayor, Tom Murphy helped to create the Ben Franklin Technology Partnership as a state legislator. This experience would serve him well while guiding Pittsburgh through its early stages of economic restructuring in the wake of the steel industry implosion in the 1980s. Murphy said the following at a conference I attended last April:

The first person I have ever heard to mention the idea of the university, as the steel mills are collapsing, that universities could be the economic driver of Pittsburgh was Doctor Cyert, who was the president of CMU.

As I have learned, Richard Cyert is the person most responsible for Pittsburgh's economic transformation that Mayor Peduto enjoys today. As Stanford birthed the powerhouse of Silicon Valley, so has CMU lured the likes of Google, Disney, and Uber. I used to think the talent produced in Pittsburgh was the main attraction. Almost exactly a year ago, the New York Times Magazine set me straight:

Carnegie Mellon’s experience is a familiar one in the world of high-tech research. As a field matures, universities can wake up one day to find money flooding the premises; suddenly they’re in a talent war with deep-pocketed firms from Silicon Valley. The impacts are also intellectual. When researchers leave for industry, their expertise winks off the map; they usually can’t publish what they discover — or even talk about it over drinks with former colleagues. In the long run, raids can generate symbiotic relationships; researchers who return to academia years later bring their real-world experience into the classroom and can draw on their network of wealthy industry contacts to fund university research. But as Carnegie Mellon’s roboticists are finding, reaching that end point can make for a bumpy ride.

The robotics knowledge produced at CMU matured to the point where it would interest Uber. The group of applied scientists could converse with basic researchers and a market-facing tech company such as Uber. The National Robotics Engineering Center, founded in 1995, would serve as the translator between "early-stage inquiry" and the making of "commercial products". Uber poached the workers the Engineering Center, not the scholars at the Robotics Institute (created in 1979).

Basic research isn't of much use to anyone without these translators. Without them, the Industrial Revolution wouldn't have spread from Britain to France. Without them, Mayor Peduto is managing decline instead of serving as a model for Rust Belt revitalization.

Q: What was the agreement? Do they assume all the risk if there’s an accident?

Comment: Peduto states, "There is no formal agreement." Many researchers simply moved from CMU to Uber, a matter of blocks not time zones. Uber's CEO is on the Mayor's turf, not the other way around. When he was a member of the City Council, Peduto represented the East End where CMU is located. He doesn't have to trust Travis Kalanick. He can visit with a robot scientist in his own backyard.

Q: Why is Pittsburgh a good place for this experiment?

Comment: The short answer to this question, even in Peduto's words, is Carnegie Mellon University. He also references another knowledge cluster in addition to AI/robotics, cybersecurity. Somehow overlooked, CMU's Software Engineering Institute (a legacy of the Cyert era) might receive over $1.7 billion from the Department of Defense during the next 10 years for cybersecurity R&D. After World War II, economic development concentrated in places with federal labs and/or top-tier research universities. Pittsburgh is blessed with both. Uber in Pittsburgh is no accident, some 60 years in the making. Peduto doesn't forget Pittsburgh's former industrial prowess, uttering "advanced manufacturing", which is a buzz term used by margin-challenged producers to keep the subsidies coming. The region continues to deleverage from manufacturing employment, much for the better.

Q: Uber has been a job creator in Pittsburgh by providing work for drivers. Now, the city is poised to become a leader in building a technology that could make those jobs obsolete. Does that concern you?

Comment: In some ways, Mayor Peduto throws advance manufacturing under the bus in this answer. The future economy concerns autonomous vehicles, not drivers providing a service. Even the cars as a thing have ceased to be a commodity of value. The car isn't an item for purchase. It's a service and a data collector. I'd reckon that even the service takes a backseat to the data produced.

Key quote, "Pittsburgh is one of the only cities that has traffic signals that can learn." The reason for that, once again, is CMU. The new jobs, the ones we can't imagine today, will show up in Pittsburgh because of Carnegie Mellon, not because Rust Belters make things.

Q: But people are upset. Kalanick has said he hopes to replace all of Uber’s human drivers with technology.

Comment: Peduto quips, "We would hope to replace drivers lost with advanced manufacturing jobs." Pittsburgh won't be a hub of making autonomous vehicles. I would bet that ends up in Mexico. Pittsburgh will make the tech that allows autonomous vehicles to operate everywhere. Human drivers will have to take advantage of new industries that stem from the presence of autonomous vehicles. Cybersecurity looks good.

Q: Still, there could be blowback.

Comment: I love Peduto's response. He recounts his journey from Rust Belt shame to Rust Belt chic. The key to revitalizing urban manufacturing centers is to transition from the psychogeography of managing decline to managing growth. From "The Robots That Saved Pittsburgh":

“The big challenge now is figuring out a way to disseminate the city’s new prosperity to all of its residents, not just the high-tech guys,” says Peduto, 49, who just ran and won the mayor’s job on a promise to create “the first progressive administration for a Rust Belt city in America.”

Paradoxically, all the urban poverty and blight is still there. Pittsburgh still has major Rust Belt problems. Yet the main issue is inequality and displacement. Peduto will be judged on how well he manages the renaissance.

Q: Is there an application for self-driving cars in public transit?

Comment: Peduto describes Rust Belt demographics in one sentence, "Think about someone who is elderly who lives a mile away from the bus stop." Among shrinking cities struggling with an aging population, Pittsburgh is exceptional. The exodus from the region in the 1980s blew a giant generational hole in the birth rates that still hobbles the economy (and municipal finances) in 2016. This unique experience allowed Peduto to imagine the benefits of autonomous vehicles for his constituents, more so than anywhere else.

Q: This is coming quickly. Like all of a sudden — boom, driverless cars! How can that be?

Comment: Peduto doesn't answer the question. Earlier this month, the head of CMU's National Robotics Engineering Center did:

Robotics is exploding, and the adoption curve, the tech disruption curve—these are getting steeper and steeper. I’ve been doing robotics for two decades, and if you look at the progress made between 1995 to 2005, in the past couple years we’ve seen more progress than in that whole decade.

Not only is the place right, so is the time. Pittsburgh as "Roboburgh" is old hype without much substance. Many regional watchers are still waiting for the supposed startup boom. Attracting part of Uber or Google doesn't count as a success story. Pittsburgh doesn't have Microsoft or Facebook (both companies were founded somewhere other than where they are headquartered now) to call its own. I don't know if that will change. It could change along with the acceleration in robotics. The knowledge cluster is two years old, not thirty seven years old.

Q: Have you taken a ride in a self-driving car?

Comment: "The day they announced, [Kalanick] picked me up from work and took me home," said Mayor Peduto. Again, I would stress Peduto's familiarity with CMU as critical to this relationship. The trust was there before Uber existed.

Q: A lot of people are scared of self-driving cars. Have you talked to Kalanick about that?

Comment: Peduto name drops, "I had dinner with him one night. One person at the table was the guy who created Google Maps, and the other person is the reason you don’t see the fail whale on Twitter anymore." 20 years ago, such a story wouldn't have been odd in Pittsburgh. This is old hat because CMU links the city to Silicon Valley.

Q: It sounds like they were sidestepping the issue.

Comment: Great line from Peduto, "we didn’t make steel, we made the middle class." From the cradle of robber barons, income equality was born. Pittsburgh doesn't make robots. It makes the upper middle class.

The New Geopolitical Significance of Pittsburgh, Pennsylvania

I'm celebrating an anniversary. Thursday, June 1st, 2006, I wrote my first blog post. What you are reading now is my 3000th pithy effort. But enough about me and onto my muse, Pittsburgh. In 2006, the city and the region were a Rust Belt basket case. In 2016, Pittsburgh is the blueprint for Rust Belt urban revitalization. What happened?

Poetically (not ironically), one of Pittsburgh's failures is its greatest success. In the mid-1980s, the Software Engineering Institute (SEI) was supposed to put Pittsburgh on the new economy map:

"It`s Silicon Valley for chips, it`s Boston for electronics and it`s going to be Pittsburgh for software", predicted Dr. Angel Jordan, provost of Carnegie-Mellon, the intellectual pivot of this turnaround.

That's a great article written by one of the greatest journalists ever to cover economic development and globalization, Richard C. Longworth. He captures Pittsburgh's Andrew Carnegie-like ambitions at a nadir of the steel collapse. Pittsburgh was in a tailspin that continues to haunt the region's demography to this day (read Null Space by Chris Briem, who had his own decade anniversary a few weeks ago). Fair to say, the SEI didn't pan out for Pittsburgh. Or, did it?

I've talked to a few people who know better than I do that SEI didn't save, as the fish the did, Pittsburgh. In the late 1990s, the policy powers that be accepted that fate. I accept that fate then and now. Software landed in Seattle.

The SEI couldn't fly. This turkey would prove to be a feast. What the SEI looked like last summer:

Carnegie Mellon University's Software Engineering Institute will continue research and development work for the Department of Defense thanks to the extension of a contract that could reach $1.73 billion.

The contract ensures that SEI, the only federally funded research and development center dedicated to cybersecurity and software engineering, will continue work under the Office of the Under Secretary of Defense for Acquisition, Technology and Logistics for at least the next five years, with an option to extend an additional five years.

Burgh Diaspora drops the mic.

1960, Peak Cohabitation for Young Adults in the United States

If the 1950s were peak city (in terms of population) in the United States, the decade also represented peak cohabitation for young adults (18-34 year-olds). "For First Time in Modern Era, Living With Parents Edges Out Other Living Arrangements for 18- to 34-Year-Olds":

Dating back to 1880, the most common living arrangement among young adults has been living with a romantic partner, whether a spouse or a significant other. This type of arrangement peaked around 1960, when 62% of the nation’s 18- to 34-year-olds were living with a spouse or partner in their own household, and only one-in-five were living with their parents.

Back in the 1880s, an agricultural economy tipped industrial. In the 1950s, a blue collar economy tipped white collar. In the 2010s, a white collar economy is tipping white coat. These economic shifts are reflected in the household composition changes for young adults.

For women, the trend since 1960 is one of increasing labor market share. For men, it is a steady fall from a zenith of economic power. One important result is a dearth of two young adults sharing a one bedroom apartment in a city.

Those couples sharing a bed are likely college-educated and well remunerated, thus punishing young adults requiring solitude for slumber in the urban rental market. Bottom line, the same number of people in 1960 would require more city rooms in 2014. The sun will shine on two couples, all four working, sharing a two bedroom flat. They will have the pick of the real estate litter.

The social arrangements during peak city (1950s) define today's housing cost challenges. Why pack into tenements when inner ring suburbs drown in affordable single family housing? Immigrants have figured this out. The native born would rather pay the cool city price or move back in with mom and dad.

The Middle Class in the United States Is Dying

A successful real estate developer demanded we get to the bottom of a Rust Belt paradox. Why is there no land for housing when the population is so much smaller than it was during the 1950s? As decades go, the 1950s remain a zenith for many US cities. That trend provides a clue for those who see the world through demographic goggles. A typical household might have six residents with five dependents (four children and a stay-at-home mom). In 2010, that same typical household will have one resident, who is a dependent. Welcome to grandma's house. She isn't leaving her urban digs until the paramedics take her out on a gurney. For the time being, her home isn't vacant. But the population under the roof is 5/6 smaller than it was in 1950.

In the hands of a dilettante, albeit an expert real estate developer, population change obfuscates more than it illuminates. If he has seen one population boom, he hasn't seen them all. Rural-to-urban migration brings with it a birth rate dividend. Those who would have been field hands crowd tenement buildings. Whereas the boom of suburban brats in urban chic may never yield children within city limits. Without immigration, the 1950s won't happen anywhere in the United States. The birth rates are too low.

When in doubt, dig into the data. Interrogating the struggling middle class in the Federal Reserve Bank's Ninth District:

Here’s a simplistic example to illustrate the effect of household composition changes. Suppose there are 10 people (five men and five women) each making $34,000 per year, and together they make up six households: four married couples with household incomes of $68,000, one male- and one female-headed household, both with income of $34,000. The middle income observation for these six households is $68,000, the income of a typical married couple.

The next year, one of the married couples is divorced. That leaves three married households and four single-headed households—two headed by men and two headed by women. Now let’s say every person gets a 10 percent pay raise. Is everyone better off? You certainly could argue that. But because of the compositional shift, the median for these seven households plummets to $37,400 because the middle-ranked observation (with the raise) is now a single-headed household.

This example exaggerates the actual decline in married couples, but it demonstrates why the overall household median can be misleading. In the district, every state saw a decline in married households roughly in line with the national trend.

The "compositional shift" is just one example of how the median income metric (like population change) can distort our understanding of how the middle class is doing. Everyone gets a raise. The median income goes down. I see this mistake all the time in arguments about how to address affordable housing. Experts and dilettantes alike abuse the data to score ideological points. Misleading traps abound.

Remember rule 1: When in doubt, dig into the data. Rule number 2 concerns disaggregation. The Fed's Ninth District shows how:



Simply put, the Fed controls for the household compositional shift problem. What is good for median income is good for population, particularly in Rust Belt cities. Over long time periods, a household is anything but a constant. Don't be fooled.

Flat World of Finance

Whether global (more recently) or national, the location of certain professional services have defined the urban hierarchy. At the top are cities with F.I.RE. (finance, insurance, and real estate) industries. Finance in particular has come to define the latest era of globalization, with a few large cities dominating the sector. No more, FIRE is dying:

Forget the bright lights and fast pace of living in two of the world’s greatest metropolises, city living for a new generation of financial workers is now more Jacksonville in Florida and Warsaw in Poland than New York and London.

Instead of serving to attract top talent, New York and London price out financial firms in search of fatter bottom lines. High-end banking is no longer tethered to the strictures of the urban hierarchy. The world is, effectively, flat. Convergence defines the economic geography.

If FIRE is dying, then what kind of proximity logic is taking its place? What kind of industries are economically divergent? ICE:

New York is reinventing itself once again, drawing on a different source of advantage, the unique assets that, regardless of the FIRE sector's future, will sustain it as a world city. Our intellectual, cultural and educational (ICE) strengths—already among the world's greatest—are becoming the essence of New York's global identity. The creative economy of New York is once again a gateway economy, as ICE assets originate here to influence and shape our world.

The hub of ICE ("intellectual, cultural, and educational") is the urban research university. White coats replace white collars. Private industry will flock to be near knowledge production, like a moth to a flame.

New York City does not dominate this world as Boston does. Cambridge's Kendall Square makes Manhattan look bush league by comparison:

Speaking in one of his shiny new labs on the banks of Manhattan’s East River, André Choulika, chief executive of Cellectis, says he is a “big Boston fan”, but that the cut-throat competition for talent there would make it tough for a small company like his.

“In New York, you don’t have your employees hired when they go to buy their sandwich from the food truck, like in Boston,” he says.

For biotech, New York is to Boston as Jacksonville is to New York for finance. The city has enough assets to be a player. But it isn't the top dog in this ICE industry. It's the Small Apple.

Is the World Still Spiky?

Richard Florida insists the world is spiky. Geography still matters. Blah, blah, blah, blah. Touting that geography still matters is a straw man. Geography matters whether or not the world is spiky. Spiky is one pattern (economic divergence). The flat world pattern (economic convergence) is another. Urbanists fawn over Richard Florida's normative spiky geography. Denser cities are the best way to organize space. Sprawl (flat world) is bad. If the world is more spiky (spikier?), urbanism is winning. Suburbanism is losing. Urbanism is dying:

It’s been dubbed the e-change movement, where people move from the city to the country, using the internet to make a living. Garlick is not alone in making the move. The recent Super Connected Lifestyle Locations (SCLL) report identified more than 600 “lifestyle towns”, from Kiama in NSW to Ballarat in Victoria. With the rise of silicon suburbs in regional areas outside our capital cities, demographer Bernard Salt says, “We are witnessing a quiet lifestyle revolution in suburban Australia.”

Since I own the nauseating X place is dying meme, I get to explain. There is no zero-sum game between urban and rural. There is no zero-sum game between urban and suburban. Spiky and sprawl are not mutually exclusive geographic patterns. Sometimes, the world is flatter because it is so spiky. London is flatter because it is so spiky:

I’d expect future London émigrés to consider the European sunbelt, which stretches from Lisbon to the delightful Albanian coast. The Mediterranean is the perfect place to live as long as you aren’t part of the local economy. London is the opposite: the ideal is to be part of the economy without living there. The logical thing to do (presuming Brexit doesn’t happen) is to arbitrage the two: live on the Med and work for a London company. This will be an obvious decision for the next generation of knowledge workers, for whom the concept of a daily commute to an office will probably seem baffling.

For the next generation of knowledge workers, it will be dubbed the e-change movement. Work in London, live in Lisbon. Technology will make today's daily commute seem baffling. The world is flat for workers who have a remuneration connection to spiky London.

One can't arbitrage London unless one lives and toils there. The world is spiky. To live and toil in London is to desire escape. The world is flat.

The Rust Belt is spiky. The Sun Belt is flat. The spiky Rust Belt made the flat Sun Belt possible. For African-Americans, Atlanta is Chicago sprawl.