This article appears in the Spring 2018 issue of The American Prospect magazine. Subscribe here.
When the Democratic “blue wall” stretching from Wisconsin through Michigan and Ohio to Pennsylvania fell on November 8, 2016, its breach reflected a growing socioeconomic gulf between the prosperous coastal states and depressed non-metro America. The vast majority of economic growth since 2008 has flowed to the coasts, while the Midwest and rural America have seen spikes in deaths of despair, divorce, an opiate crisis, and a moribund economy.
It wasn’t always like this. Once upon a time, from roughly 1880 to 1980—an era Trump’s white supporters may have in mind when they demand America be Made Great Again—incomes of different regions more nearly converged. Much of U.S. manufacturing and its supply chain was based in the Midwest. Thanks to unionization, a good deal of basic industry paid decent wages. Meanwhile, other industries such as textiles and apparel migrated from New England to the Southeast, raising earnings there. Regional development policies, everything from the Tennessee Valley Authority to the geographic dispersion of war production plants, tended to spread the wealth around.
Main Street, Ottumwa, Iowa
But since 1980, that converging trend leveled off and then began to reverse. The Economic Innovation Group reported that sparsely populated counties with fewer than 100,000 residents were home to 32 percent of new businesses founded in the wake of the 1991 recession, and 15 percent following the 2001 recession—but zero percent of all new businesses created in the recovery from the 2008 recession. The more rural states tend to be in the Midwest; the more urban states, with the stronger economies, tend to be on the coasts.
There is more to the story of Democratic regional decline, of course. It includes the gradual Republican takeover of the white South, reversal of voting rights, and the “sagebrush rebellion” in once strongly Democratic states of the mountain West. But the economic wipeout of the once-Democratic industrial heartland is a crucial and relatively recent element.
On its face, the story of declining middle America and the prospering coastal hubs is simple: As manufacturing jobs left the United States, economic growth was increasingly driven by a handful of sectors—notably tech and finance—that have agglomerated in coastal metropolises and that look to global communities of interest and supply chains rather than domestic ones. Just Washington, D.C., New York City, San Francisco, San Jose, Boston, and Los Angeles account for nearly 70 percent of all IT-industry venture capital, for instance. AnnaLee Saxenian, a professor in the Department of City and Regional Planning at the University of California, Berkeley, told me that these cities have “accumulated a skill base with a depth unparalleled anywhere in the world.”
There, the economy is booming. Highly skilled workers spend locally the money that streams into their business from across the globe, with multiplier effects that benefit the entire region. In his book The New Geography of Jobs, Enrico Moretti estimates that while a traditional manufacturing job creates 1.6 service-sector jobs in the community around it, what he calls an “innovation job” creates about 5 service jobs in the community around it, and those tend to be higher-paying than the average service-sector job to boot!
But why this dense spatial concentration of advanced economic sectors? And can we do anything to spread the tech wealth around, or to promote more balanced regional development generally?
IN THE 1990S, THERE WAS some optimism that the internet and other digital technologies would lead to decentralization. Code can be written and delivered just as well from cheap rural Iowa as it can be from a Stanford University computer lab located in one of the world’s most expensive ZIP codes. A computer scientist living in astronomically expensive Palo Alto might telecommute from Santa Rosa in Sonoma County, where a comparable house costs perhaps one-fourth of its Palo Alto equivalent. But most don’t. In Shelburne Falls, Massachusetts, a charming and hip country town two hours away from downtown Boston, a beautiful four-bedroom home can be had for about $400,000, maybe one-fifth the cost of a comparable house in Newton. But for the most part, techies evidently like to live near other techies. And some of this reflects not just tastes but the economic benefit of a local, spatial network. The tech industry seems to depend on the regional concentration of highly specialized experts for its success.
“Tech,” Saxenian says, “is hard to define.” It refers to products and processes closest to the cutting edge of innovation, the ones that take the highest levels of education to design and the most precise expertise to produce. Amazon, Facebook, Microsoft, Apple, and Alphabet (Google’s parent) are part media company, part software innovator, part platform; Amazon is part delivery service and supermarket. All five have large and active merger and acquisition divisions and diversify their portfolios as quickly as new opportunities emerge. What they have in common with computer-chip manufacturer Qualcomm, the online messaging platform Skype, or a startup that intends to connect your at-home juice-maker to the internet isn’t immediately obvious. But all these companies spend a large portion of their budget on R&D, and they mostly outsource manufacturing. As technology advances, they all employ a hyper-educated, hyper-specialized class of workers. What these companies have in common is dependence on a skill base that evolves very quickly.
Tech employees include computer scientists in Silicon Valley, but also biomedical scientists in Boston or physicists employed by Raytheon in Washington, D.C. The broader ecology includes design and marketing professionals, legions of financiers, merger-and-acquisition lawyers, and patent attorneys. All of these jobs that orbit around the tech industry require a high level of education and specialization. And they tend to be located in a handful of metros on the coasts of the United States.
TECH’S ASCENDANCE HAS occurred against the backdrop of manufacturing’s steady decline. After holding relatively stable for a number of decades, the number of Americans employed in manufacturing declined precipitously after the 2001 recession and the entry of China into the World Trade Organization. By 2010, a sector that in 1980 employed one in every five Americans employed one in every ten, and one in three American manufacturing jobs disappeared between 1997 and 2010. The decline was harshest in the Midwest and South.
The decline of manufacturing, its causes and its consequences, has been expounded upon at length in the pages of this magazine and elsewhere. Some of its causes were inevitable and outside the power of the U.S. government to alter, like the productivity gains driven by automation. Others were the consequence of U.S. policy choices. Free-trade deals incentivized large corporations with activist shareholders to outsource their supply chains and subcontract American jobs overseas. Except for key defense industries, the United States disdained industrial policies as a matter of ideology.
The tech company is an avatar for post-manufacturing America. Tech companies tend to be more horizontally organized, in contrast with the vertically integrated firms of yesteryear. With a few exceptions, such as medical devices and aircraft, they increasingly subcontract their manufacturing overseas. Foxconn is the best known of these subcontractors: Headquartered in Taipei, Foxconn workers assemble 40 percent of all consumer electronic goods, including the iPhone, iPad, Kindle, PlayStation, Xbox, and Wii. Today, Apple employs around a hundred thousand people in the United States; in 2012, the last year the data was made publicly available, they had seven times that number of subcontractors working for them, mostly in Southeast Asia.
The tech jobs that do stay in America tend to be very good jobs. The problem: They are concentrated in a handful of metros, and don’t seem likely to spread across the country anytime soon. Those that do spread—like Amazon warehouse jobs, or legal and financial backroom work—tend to be the least-desirable jobs, the ones that carry the smallest multipliers with them (if any at all).
Scholars have disagreed over why tech has refused to disperse, although they are unanimous in their assessment of its concentration. Richard Florida famously argued that regional concentration is substantially due to the cultural appeal of cosmopolitan, tolerant, and hip cities. In his influential and best-selling 2002 book The Rise of the Creative Class, Florida defines the creative class as spanning from tech professionals to artists to writers, who tend to cluster around cities like San Francisco, Seattle, Boston, and Minneapolis whose economic strength depends on their presence. Of the top 20 cities with the largest creative class (Florida calls them “creative centers”), 13 are also among the top 20 cities with the largest high-tech industry. Technicians may be geeks in some respects, but evidently they like to live in the hippest metropolises.
Similar arguments have been made by other scholars observing the importance of the “marriage market” for young people. Tech workers tend to be young because of the need for up-to-date education to succeed in the industry, and so it follows that tech firms would locate in the densely populated areas that young people prefer.
But economic factors may well be more important in the clustering of tech, as even Florida has lately acknowledged. One of the most persuasive of these factors is the nature of what economists like Moretti call “thick labor markets” and their appeal for tech. As the speed of technological advance has accelerated, so has the specialization of the experts in the field. In the hunt for the self-driving car, for example, there are only an estimated 10,000 artificial intelligence and computer scientists in the world who have the requisite technical know-how to be of help to Google, Apple, and Uber as they frantically race to be the first to reach what is sure to be a multibillion-dollar market.
In a sector that requires a labor market to be both specialized and dense, it benefits employer and employee alike for the market to be geographically clustered. With a thick labor pool, the employee can shop the handful of employers against one another, and the employer can do likewise with prospective employees. The marketing, legal, and financial expertise associated with tech specializes, too. Biotech patent law requires a hyper-specialized skill set distinct from the law associated with Uber and Waymo’s recent intellectual property dispute. All of this encourages the formation of spatially concentrated thick pools of specialized labor. Even before its selection of a second headquarters, Amazon in late February added 2,000 jobs to its workforce in Boston, a center of specialized tech workers. Them that has, gets.
American tech firms increasingly subcontract their manufacturing overseas, to places like China's Foxconn Technology Group, the world's largest electronics contractor.
By contrast, the manufacturing economy relied mainly on semi-skilled assembly workers, who could be trained in a matter of weeks, some of them right off the farm. No specialized labor pool was needed. This meant factories could be built where it was cheap to do business, organically driving a convergence of wealth between the urban, expensive-to-do-business-in major metro areas and the rural, cheaper-to-do-business-in countryside. The more highly skilled workers, such as tool and die makers or engineers, went where the work was. But with tech, the pool needs to cluster, to keep replicating itself, and the virtuous cycle of regional dispersion has disappeared from the landscape of American life.
Rates of technological change are accelerating, notably in the consumer market: The telephone existed for 35 years before it was in a quarter of American households. The television, 26 years. The personal computer, 16. The iPhone? Under ten. The ability of a firm to anticipate and adapt to technological change is essential to a firm’s success in the 21st century, and this is nowhere more true than in Silicon Valley, where the ability to conform to a trend and beat competitors to market is life or death for startups. A thick labor market, full of specialized tech workers and the lawyers or financiers who facilitate their work, all living within miles of each other, is necessary for flexible firms to retool at the drop of a hat.
ANOTHER NON-CULTURAL FACTOR that may be contributing to the regional concentration of tech industries is more pernicious: the use of monopoly power to crush rivals and potential ones. The “big five” tech companies are infamous for buying out or driving out competitors. Some of those potential challengers would be located outside the top tech hubs.
In her book Profit Cycles, Oligopoly, and Regional Development, Ann Markusen describes how industrial cycles interact with geographic dispersion. After a first moment of innovation and high profits where industry leaders dominate the competition, an industry tends to be regionally concentrated. It’s in the second phase of the profit cycle, where normal competition sets in, that the industry disperses geographically. Significantly, if the first moment doesn’t transition into normal competition where new firms enter the market and drive profits down, and instead a few firms oligopolize the industry, there is less of an incentive for the industry to disperse geographically. This may be occurring in tech.
The big five have come under antitrust scrutiny over the past year. Google controls more than 60 percent of desktop online searches and 94 percent of mobile online searches. Half of all online commerce happens through Amazon. Facebook and Google split online advertising; Apple and Google split the smartphone market. When a new company with promise emerges in tech, it’s either swallowed or smothered. Snapchat is an instructive example: After Facebook offered to acquire the company for several billion dollars and was rebuffed, the half-trillion-dollar super-company with two billion monthly active users cloned Snapchat’s features and depressed their market share. Venture capitalist Tomasz Tunguz explained to The Guardian that the big five were “financing the next generation [of] research at a scale that no one else can afford.” Remember those artificial intelligence computer scientists working on the self-driving car? Their average salary at Google is $345,000. Good luck holding onto specialists with expertise in a market the big five is interested in controlling.
Today, the number of new tech businesses is at a 30-year low. Oligopolization has set in, and it’s fair to assume that some of those new businesses that the big five’s anti-competitive behavior is keeping locked out of the market would have been born in metros less expensive than San Francisco, Boston, or Seattle. Instead, as there are fewer and fewer tech IPOs and more and more acquisitions by the big five, it seems as though many startups angle to be acquired by those super-rich companies located in those super-rich metros.
Another oft-cited factor is the role of a handful of universities. Elite research universities like MIT, Stanford, Carnegie Mellon, or Caltech are important both in concentrating those specialized labor pools and in replenishing them with federal research and development dollars that in turn help incubate advanced tech companies. A great many startups are spinoffs from these and a few other universities. Silicon Valley is where it is today in large part because William Shockley chose to open his silicon semiconductor firm near Stanford.
But initial locations also reflect chance. Enrico Moretti estimates that Microsoft alumni have created some 4,000 businesses in the greater Seattle area. But why did Bill Gates and Paul Allen move their young company from cheap New Mexico to Seattle? In part because that’s where they were born. And once a city has gotten that first foothold in the tech world, agglomeration is soon to follow. Seattle’s population has boomed over the last few decades and its per capita GDP has soared—even as inequality and housing prices have grown, too.
Today, cities from Memphis to New York to a small community in Georgia that offered to rename itself after Amazon are offering billions in subsidies and whole departments of city employees dedicated just to servicing Amazon’s needs in the hope of becoming home to Amazon’s “HQ2.” Larry Hogan, the Republican governor of Maryland, offered Amazon $5 billion in incentives and called HQ2 “the single greatest economic development opportunity in a generation.” This is what a winner-take-all economic geography looks like: When the most important and valuable industry in the country agglomerates in just a handful of metros, cities are more than willing to race to the bottom in the hope of getting to join the chosen few.
LIKE TECH, FINANCE HAS ALSO tended to cluster. As the banking industry has become more concentrated, more and more of its wealth has been located in a few ZIP codes. Manhattan, despite its astronomical rents, is still home to most investment banking. Hedge fund moguls flock together in places like Greenwich, Connecticut. America, with its long-standing hostility to financial concentration, once prohibited interstate banking and even discouraged branch banking. In 1966, the United States was home to almost 14,000 banks, more than 10,000 of which had only a single branch. By 2014, there were fewer than 6,000 banks and fewer than 1,000 of them were single-branch banks. The five largest institutions have nearly 50 percent of all banking assets. And while there are financial services operations scattered throughout the country, they often take the form of badly paid back-office operations.
When South Dakota agreed to end its usury laws in the 1970s, the big New York banks moved their back-office operations there, keeping the executive jobs in New York, close to other bankers and an ecology of specialized law and accounting firms. As with tech, financial market concentration has led to geographic concentration—making the world of 21st-century finance Thomas Jefferson’s nightmare: As bankers cluster around the Big Apple, Midwestern and rural America has fewer small, regional banks geographically proximate and culturally sympathetic to the small businesses that need investment and lines of credit.
OUR RELATIVELY EQUITABLE regional past wasn’t the result of an economy behaving according to its own whims. In preceding centuries of American life, the country embarked on huge, centrally planned, regional development projects sponsored by state and national government. In the early 19th century, the Hamilton-Clay “American System” combined a high tariff on foreign imports with investments on infrastructure projects like roads and canals to improve the American economy. In 1862, the Homestead Act incentivized thousands of American families to move west. This was also the era of government-subsidized land-grant colleges under the 1862 Morrill Act, plus transcontinental rail lines underwritten by cheap public land, all of which served to spread economic activity.
At the turn of the 20th century, the Populist Party represented a new regional consciousness in rural America. Framing their arguments in the language of anti-monopoly, they argued that wealthy city dwellers controlled and drained the wealth from rural America. They won victories that saw Congress regulate the railroad, an industry that played an important role in defining the economic geography of the country. Government efforts to spread economic activity expanded in the New Deal. The Tennessee Valley Authority was an explicit regional development program to bring electrification and economic opportunity to a neglected rural region of the country. Likewise the public dams of the Northwest. Eisenhower’s Federal Highway Act, along with being the largest public works program in American history, made it cheaper and easier to do business in parts of the country that were formerly geographically isolated. And airline regulation assured that isolated regions and small cities would have affordable air transport.
Since then, America hasn’t lacked regional development aid. Instead, we’ve had a patchwork of poorly coordinated measures. Trade Adjustment Assistance, intermittently available to workers whose jobs are lost to trade, offers skills retraining to move workers into more-competitive sectors. But the program is flawed: Not only is it underfunded, but as layoffs happen in waves, retrained workers flood the employment market and many are still left unemployed even after the retraining. Reacting to trade deals like this, instead of proactively investing in regional development, leaves many workers and regions behind.
One of the largest disguised regional investment programs is federal Defense Department spending, which funnels hundreds of billions of dollars into procurement, manufacturing, research, and development projects across the country. In the 1940s and 1950s, this money was dispersed in a pattern that followed the spread of manufacturing across the United States: The biggest recipients were in a belt stretching from Pennsylvania to Michigan. But from the 1950s to the 1980s, American defense spending shifted focus from tanks and shipbuilding to aeronautics and high-tech research, and defense spending began to redistribute in a pattern eerily similar to contemporary distributions of high-tech metropoles. By the 1990s, states in New England received a disproportionate share of defense contracts, with Massachusetts receiving twice the national average per capita, and Connecticut receiving triple. Reagan’s military buildup deliberately favored his (mostly non-union) political base in the Sun Belt. Because it’s not thought of as an explicit regional development program, defense spending has often mirrored trends in the distributions of high-tech research, development, and production.
As the competition for Amazon's HQ2 demonstrates, cities are more than willing to race to the bottom in the hope of getting to join the chosen few. Here, construction cranes tower over the Amazon Campus Grand Plaza in Seattle.
Bringing growth back to rural America will not be easy. But many Midwestern American cities have competitive advantages that smart regional development policies could build on. Toledo was once the world leader in the production of glass, and was on the verge of becoming a global leader in solar panel manufacturing too—until China entered the market, producing subsidized solar panels much cheaper on a per watt basis than American brands, eventually forcing the closure of a Toledo-area solar panel factory. Rochester, New York, is home to Xerox—which has been acquired by Fujifilm Holdings, a photography and multimedia company headquartered in Tokyo. Rochester, like Toledo, feels like a missed opportunity. Both were R&D leaders in distinct fields, and both had manufacturing plants in the United States.
Many close observers of industrial development have argued that R&D benefits from being close to manufacturing—managers learn how to efficiently design a product from watching it be made regularly. Given Apple’s success in outsourcing to East Asia, other students of design and supply chains contend otherwise. Yet there are missed opportunities here. While Toledo and Rochester will never be able to compete with San Francisco and Boston in the most prosperous areas of the tech economy, they could leverage their particular competitive advantages in their specific fields into vibrant local economies, with spillover to the surrounding rural areas, given more supportive trade and regional development policies. It’s not enough to say, as our president has, that rural Americans “can leave” and move somewhere else. Investment and industrial policy could save the regions of the country those Americans grew up in from economic decline.
OTHER COUNTRIES WITH ROBUST regional development strategies offer useful examples. Germany’s regional inequality, despite the challenge of reunification with the much poorer East since the fall of the Berlin Wall, is much lower than that of the United States. A number of factors contribute to this. Offshoring and subcontracting have been resisted, both by corporate leaders and by national policy. Germany, despite high labor costs, has the world’s largest trade surplus.
Germany also has smart regional development policies that earmark funds for economically underdeveloped regions, helping those regions invest in infrastructure and leverage local competitive advantages into vibrant economies, spreading economic growth across the country. This commitment to regional equality is enshrined in the German constitution itself, which grants the federal government the right to provide the states funding to “equalize differing economic capacities.”
This constitutional commitment has been put into practice by a variety of policy tools. At the federal level, Germany after unification instituted what’s called in German the “Länderfinanzausgleich” or a nationwide pool of tax money that poorer states can draw on for the purposes of regional development. Since 1989, Germany has spent more than 1.3 trillion euros on the development and integration of the former DDR. The sum of money available is enormous—about 5 percent of all German taxes—and is earmarked for use only by poor states to invest locally to make their regions more competitive. These funds flow into both infrastructure and innovation, spreading the wealth. This tax system is joined with the Joint Federal Government/Länder Scheme for the Improvement of Regional Economic Structures, a pool of funding instruments put at the disposal of weaker regions that present concrete plans for economic development.
Another tool Germany uses to foster regionally equitable growth is its development bank, the “Kreditanstalt für Wiederaufbau,” or Reconstruction Credit Institute. Formed with funding from the United States as part of the Marshall Plan in the wake of World War II, the bank is publicly owned and controls assets valued at around $500 billion. The bank makes loans with an eye toward regional development and equity. It has played an essential role in Germany’s energy transition from coal to cleaner energies, providing funding for energy-efficient technologies and housing. Similar banks exist in China, South Korea, and Japan. Ironically, each of these was set up initially by the United States, with an eye toward the threat that regional inequality could pose to postwar political stability, in an era when planning was not a dirty word. We have been more willing to create regionally equitable development strategies for countries around the world than we are here at home.
Germany is also home to a complex network of publicly owned banks with public service mandates. “Landesbanken” are regional state-run banks that boast annual investment budgets in the hundreds of millions. There are eight Landesbanken operating at the federal state level. On an even more local level are the “Sparkassen,” 386 publicly owned banks that together control more than 1.1 trillion euros in assets, and that loan only to local businesses and customers. This network of banks has kept finance decentralized and diffused across Germany. In contrast to the United States, where finance is increasingly concentrated in New York, banks in Germany are geographically proximate to the small businesses they service.
Germany’s economy, in short, is regionally equitable by design, not chance. In addition to all of the above tools, Germany also benefits from the European Union’s regional policy, which aims to avoid regional disparities within the EU, and an active and strong labor movement (that recently won the right to a shorter work week). Dr. Philipp Steinberg, the director general for economic policy in the German Federal Ministry for Economic Affairs and Energy, says the whole cluster of different tools Germany uses to stave off regional inequality is essential to “keeping everyone on board. As we adjust to globalization, regional economic policies give everyone in Germany the feeling that the state really cares for them.”
If America, instead of sending jobs overseas, took inspiration from other countries’ programs to invest in regional competitive advantages, it could embark on another grand regional development program for the 21st century. The U.S. government could use any number of those tools—a development bank, a network of small publicly owned banks, a tax set-aside explicitly for regional development—to direct a flow of investment toward regions with potential competitive advantages that are now being smothered in the crib. Not only would this broaden America’s economic prosperity, it would go a long way toward re-knitting America’s social fabric, which was torn long before the election of Donald Trump, and further torn afterward.
Rural America, especially in the Midwest, is in a crisis driven in part by the agglomeration of the most prosperous sectors of the American economy. That crisis won’t solve itself: The speed at which tech develops demands thick labor pools concentrated around a handful of cities. And it doesn’t look likely that tech’s prominence will fade in the coming years. But that doesn’t mean the economy is condemned to soaring rates of regional inequality. Smart policy, using tools both from America’s own experience and from models in other countries, can help alleviate the pain and spread some of that prosperity—but only if voters demand that elected leaders choose to do so.