A couple of weeks ago, I read Enrico Moretti’s The New Geography of Jobs. I give it a semi-hearty recommendation. Moretti is an accomplished and serious economist; what you read in this book has the considerable advantage of being true. And he also writes surprisingly stylish prose. It provides a useful introduction to how to think about economic geography. But it won’t necessarily blow your mind.
So here’s what your correspondent has to say about this book. First, I’ve gotten really interested in economic geography in recent months for a very simple reason: for somebody who (like me) takes economic theory seriously and looks at the world through that lens a lot, it is surprising how much geography matters. Because according to the most very basic economic model of how the world works, the world should, over time, grow to look more and more the same, economically (or, in other words, geography should cease to matter, economically). The reason for this is that, in very broad generalization, when two regions are economically different, then various economic actors theoretically have an incentive to make them more alike, along various dimensions. Think of it this way: If region A and region B are identical in all ways except housing is much more expensive in region A, then more businesses and residents should want to move to region B to get cheaper rent. This increased demand for region B’s housing should, in turn, push housing costs up until region B’s housing costs equal those of region A, which will make the two regions economically the same. Or if regions C and D are identical in every way except doctors’ wages are higher in region C, then more doctors should move from region D to C, making doctors dearer in D, and cheaper in C.
Sure, some of our jobs depend on the natural features of the earth — mining jobs can only exist where there are minerals to mine. But even here, a lot of the economic features of mining vs. non-mining towns should still look the same (i.e., wages should still be equal, because if a mining boon raised wages in the mining town, more people would move there and compete for those jobs). And, then again, other natural features affect our economic decisions — people pay a premium in higher housing costs to live in California because the weather is so marvelous. But still, the theory suggests there should be a huge and powerful economic pressures making the the different parts of the world look more the same economically. And, intuitively, this should be even more so the case as we, increasingly, live in a technologically-advanced and symbolic services-based economy in which more and more work can be done from a laptop anywhere in the world. I, for example, have never held a job that did not consist solely of manipulating symbols (both words and data) in computers and offered the opportunity for remote work — so why did my employers give me relatively expensive offices in Manhattan and Boston?
So we know, empirically, that the world is not flat and all the same. This, even though the first theory we worked with suggested that it should be. So what’s our new theory? What theory explains why geography still matters? The full explanation to this question can, more or less, be captured in two concepts: (1) clusters and (2) frictions. By clusters we mean how, for example, all the tech stuff is done in Silicon Valley because, well, all the tech stuff is done in Silicon Valley (and also how all the hipsters are in Brooklyn because all the hipsters are in Brooklyn). By frictions we mean the things that make it the case that labor and capital are not perfectly mobile, and so don’t just go where they get the highest returns, but often stay put where they are for a while. For example, human beings are creatures who remain (depending on how you like to put it) attached to, or stuck in, their hometowns, even when their hometown economies are declining and there are better economic opportunities elsewhere.
Now, clusters are, almost by definition, features of a local economy that reinforce the idiosyncracies of that economy, and hence prevent the world from becoming all the same. The basic idea is that, in a lot of industries, any given company can gain big advantages from locating near its competitors. If you’re a tech startup, a big part of your success is going to be your Brilliant New Idea. But the rest of the equation involves venture capitalists and business consultants and employees who are accustomed to working with tech firms. And those people are largely in Silicon Valley because, well, that’s where all their business is. The same basic logic explains why the big auto companies located together in Detroit — each one had an incentive to go to and stay in Detroit because that’s where skilled auto-workers were, and the skilled auto-workers were there because that’s where good prospective employers were. There are also a lot of more subtle, social advantages to clustering — if you’re a consultant who lives in a city with lots of other smart consultants, you’ll absorb a lot of valuable knowledge even in night-club banter. And there’s also, obviously, a sort of ‘social clustering’ going on, too, though economists talk less about that. For example, young college-educated people largely want to live in cities with lots of other young, college-educated people whom they can prospectively befriend and/or have sex with. The more college-educated young people there are in a city, the more options you have, the greater your chances of finding an ideal person to befriend and/or have sex with — i.e., the ‘market’ for friends and partners is ‘thicker.’ (Why is it considered unprofessional and not part of Serious Economics to mention this incredibly obvious, basic fact about human decision-making?) This is all pretty obvious once you start thinking about it.
So the economics of clusters suggests that there are economic pressures which make distinct economies more like themselves instead of making the world flatter. The tech firms cluster together; the young and educated do; the hipsters do — because each entity in each group gets advantages from locating near the others. So these are the basic concepts we need to talk about why economic geography is still a thing.
Now, why does this matter? Well, clusters have implications for such fraught topics as inequality and social mobility in the U.S. Let’s use Silicon Valley as our example of the quintessential modern high-tech cluster. Silicon Valley is doing well. People there are high-paid — higher-paid than the people in Flint, Michigan. This raises a number of questions: Why is this? Wouldn’t it be awesome if all of the country were like Silicon Valley? How can we manage that? You might think, “People in Silicon Valley are so well-paid because high technology is a growing industry making huge margins and lots of profits!” Actually, not quite. The simple fact that Silicon Valley’s high technology industry is a high-margin, growing industry does not by itself explain why people in Silicon Valley make a lot of money. Look at it this way: If everyone in the world suddenly developed a rapacious desire for a particular mineral found only in Pennsylvania, then mining in Pennsylvania would suddenly be a huge growth industry with high margins and lots of profits. But wages in Pennsylvania wouldn’t actually be that high. Why not? Because the very success of the mines would attract a lot of workers, who would compete for jobs, driving down wages. Because mining isn’t a skilled job, and lots of people have the capability to do it, the basics of supply and demand mean that the wages it provides will never be much higher than those of other unskilled jobs.
Rather, the reason wages in Silicon Valley are higher than in the rest of the country is that Silicon Valley workers have talents and skills that are both sought after and rare — globally. Simply put, they can do stuff, like write program, read others’ programs, and do all the other things that depend on those abilities, that the rest of us can’t. The supply of such people is relatively low, meaning that the price of their labor must be very high. In other words, high-tech workers in Silicon Valley are paid well because they are highly-skilled and highly-educated, hence rare.
Now, there’s another pretty interesting and important thing we need to consider here as well: Not only are super-high skilled programmers in Silicon Valley getting paid well, but the lawyers in Silicon Valley get paid better than lawyers elsewhere, and even low-skilled barbers in Silicon Valley are getting paid better than barbers elsewhere. Why is this? Theoretically, this shouldn’t be the case, since the wage differential should prompt more barbers to move there. The explanation, as I understand it, is that since the main “tradable” industry (high tech) is doing so well, its employees generate a lot of extra demand for services from the “non-tradable” local services sector. A lot of rich people floating around equals a lot of people willing to pay some extra cash for even a marginal improvement in the quality of their haircut. But this only answers half the puzzle — this could explain how a wage differential could develop, but not how it is sustained. Why aren’t more barbers, then, moving to Silicon Valley? Part of the explanation, of course, has to do with housing costs, but even adjusting for those differentials, barbers in Silicon Valley are taking home more cash. So what is the darn explanation? Put simply: frictions. Americans simply are not moving around the country, toward economic opportunities, as much as economists expect them, too. Economists like Moretti think this is a bad thing.
So let’s put what we’ve learned so far together to answer the question every mayor in America is asking herself: What does a city need in order to be exceptionally prosperous? It needs highly skilled, highly educated workers in profitable, high-tech industries in which America enjoys a comparative advantage (Moretti calls these industries the “innovation sector”). Detroit still has a cluster of auto firms — but since auto jobs are low-skilled, and Chinese firms can make cheap cars, too, this cluster will never again bring high-paying jobs to its locality. You need a cluster that provides jobs that are in demand, but relatively rare, that can’t be outsourced to, or competed down by, China, and you need the workers to fill them — and this will drag your whole local economy up.
And the theory we’ve produced here is borne out by the data. Moretti has put out chapters and chapters of data showing that the major correlates of a city’s average income are (1) the share of its workers with college and advanced degrees and (2) the proportion of the population employed in what he calls the “innovation sector” (i.e., high tech, information technology, life sciences, clean tech, nanotechnology, robotics, high finance, etc. He writes, “What they all have in common is that they make intensive use of human capital and human ingenuity.”).
But we’ve sort of answered this question descriptively rather than causally. I.e., we’ve articulated the features that will define a successful city in the 21st century, but haven’t answered how a city can become this thing. And it’s a lot harder than it seems. There are chicken-and-egg problems. Which do you get first? The cluster, or the highly-skilled workers? A naive mid-range city mayor might, upon observing point (1) above, issue huge amounts of debt to pay for thousands of his citizens to get masters in computer science, only to see them move off to Silicon Valley once they’ve gotten their degrees. And clusters usually start very serendipitously, rather than through any substantial planning. Moretti documents how the city of Seattle has turned around since the 1970s, thanks largely to its own high-tech cluster. But this cluster developed simply because Bill Gates was from Seattle originally, and wanted to move back there . Since then, former Microsoft employees have left Microsoft and founded their own firms in their new home — Seattle. And so on.
And that seems to be one of the main takeaways of this book. We have a lot of good economic reasoning to explain why clusters matter, why they persist, and how they benefit their local populations and countries as a whole. But the major explanation for why most clusters come into being in the first place is pure serendipity. Nothing that a mayor or governor can really control. What’s more, it’s not even clear that this is a bad thing. The important thing is that clusters happen somewhere, and that we all have the freedom and ability to move to the opportunities they provide.
For an aside, I want to put on my ethicist had again. The truth is, I find a lot of the regional economics literature kind of morally ridiculous. And that’s because a lot of it is about offering advice to mayors on how to help their city succeed. But this is stupid. What matters is not that particular regions look nice, but that individual people have lots of economic opportunities. What an ethically serious person should be focusing on, then, is the prosperity of the world as a whole. The thing we should be doing right now is not trying to bring some new tech cluster to Buffalo, NY. That’s never going to happen. And even if it did, it wouldn’t be a good thing for the world as a whole. The tech cluster in Silicon Valley is really awesome and productive. It’s best for America and the world as a whole that new tech firms locate there, where they will learn a lot more, grow more quickly, and be more likely to do something really innovative and mind-blowing, than in Buffalo, where they will stagnate, and their young employees will have miserable social lives. Of course, the trouble is you don’t win mayoral elections on the campaign slogan “Well-managed decline.”
And one thing I like about Moretti is that he seems to get this. He suggests that one major response to growing regional inequality in the U.S. should not be to subsidize declining rust belt cities, but, rather, to offer “mobility vouchers” for people to move from areas with high unemployment to areas with low unemployment. That strikes me as the smartest, most important policy takeaway from the book.
Here are a couple of other things, in no particular order, that stuck out to me:
1. Toward the start of the book, Moretti explains the standard consensus among serious economists of what is a fraught issue in our political and cultural discourse: The decline of stable, middle-class manufacturing jobs. As Moretti explains, economists believe that it wasn’t rapacious financiers or private equity that was responsible for their decline. It was globalization and technological progress. Now that poorer Chinese people can do some of those jobs, and others can be fully automated, there’s no economic or ethical reason to expect that American workers should keep doing them.
2. We read often that real incomes have completely stagnated for the bottom half of America over the past 30 years. But this may not be true. In fact, this may be an artefact of the way we measure inflation. The truth is that different tiers of American society have been experiencing different levels of inflation in the prices of the goods they consume. Thanks to superstores like Wal Mart, and the fact that low-income people shop at these superstores more frequently than the rest of us, lower-income Americans have seen their cost of living rise more slowly than the general population. So dividing their real incomes by standard CPI measures can be deceptive — they have made real gains that are occluded by the fact that it is largely higher-income Americans who have seen their cost of living rise.
3. The internet is, somewhat obviously, both creating and destroying a lot of jobs. Amazon is putting a lot of local bookstores out of business. At the same time, it’s employing a lot of programmers, web designers, Kindle-single writers, etc. On the whole, Moretti claims that the internet is creating more jobs than it’s destroying. But, significantly, the jobs it is creating are geographically concentrated, while the jobs it is destroying are geographically dispersed. I.e., retailers are getting put out of business everywhere. But a hugely disproportionate share of the people gaining new employment opportunities from the internet are in coastal metropolitan areas — particularly New York, San Fran, and Seattle. So the lesson is: Don’t fear the internet; do provide people more opportunities for mobility.
4. One thing which is frustrating about a lot of the regional and urban economics literature, to which I think Moretti falls prey, is the treatment of all people of similar education levels as interchangeable. For example, a lot of regional and urban economists will point out that people with a high-school education earn more money in Boston than those in rural western Massachusetts. They suggest that this shows that density makes us more productive. I don’t doubt that that is partly the case. But it also seems pretty likely that, seeing Boston’s higher cost of living, it is only the high-school graduate who is more confident in her abilities who is likely to move to Boston in the first place. So simply comparing wages of high-school graduates in urban versus rural areas may not actually reveal the effect of density on productivity — it may be confusing cause and effect.
5. Moretti writes wonderfully empathetically about the effects of growing regional economic inequality. he notes, for example, that the life-expectancy in Baltimore is below that of Paraguay and Iran. Divorce rates in declining rust-belt towns are extraordinarily high, while they are getting lower among the affluent. Our increasing economic geographic segregation goes hand in hand with increasing political balkanization which makes it more difficult to achieve consensus. America is increasingly coming apart, not just economically, but socially and culturally.
6. Clusters are really hard to move. They’re hugely inert. One way to look at this is that it’s a bad thing: it means that middle America will continue a relative stagnation while the wealthiest highly educated metropolitan areas will continue to get wealthier. But another way to look at this is that the fact that America has been lucky enough to have gotten a critical mass of high-tech industry clusters means that it will, despite the blunders it will certainly continue to make, still win the 21st century.
7. Speaking of mobility: Another big divide in the U.S. right now is that college graduates are more likely to uproot themselves and move to opportunities than non-college graduates. This is a big friction that is decreasing less-educated workers’ ability to find work that suits them. It also means that states and localities hoping to improve their economic fortunes through investing more in higher education are partly kidding themselves: many or most of the graduates will move away. (But it still might be a good thing for the world as a whole, of course.)
8. The rent is too damn high in high-productivity urban areas, and this is a seriously bad thing for the country as a whole. It should be taken more seriously. And the bottom line is that restricting development = higher housing costs.
9. Moretti does a good job taking on Richard Florida of The Creative Class fame. He points to a number of cities — New Orleans, Santa Fe, and Miami — that offer a bohemian cultural lifestyle, but that have not attracted good jobs. Moretti also argues that Berlin — which has been touted as a success, since it is a fun place to be right now — is in fact an economic failure, with high unemployment and low wages. If you’re an American on vacation it’s great; if you’re a German trying to feed a family, it’s not so much. Moretti suggests that Berlin is unlikely to turn into the next Silicon Valley, and its supply of creatives will continue to outpace the demand for their labor there. And if the ‘be sexy and creative and bohemian and the jobs will come’ strategy is not working Berlin, Moretti notes with some derision, it is “hard to see how it could work for Flint” (or Buffalo for that matter).
10. Political incentives have a lot of deleterious effects on regional economics. Mayors and governors often try to attract big companies to them with carve-outs — special deductions, special goodies, etc. But often, the cost to the local government of these goodies is greater than the benefits that company will bring. But a mayor will face more outrage for failing to act, than acting at too much expense. So local governments too often overbid for company headquarters, etc. More broadly, “which mayor will bend over backward the most” is obviously not a rational criterion by which to find out where economic capital should be distributed, from the perspective of the country as a whole.
But my main takeaway is something I often get from books on the dismal science: There’s very little we can do. Our economic future is largely going to be determined by overwhelming economic forces over which we as a whole — and especially local governments — have very little control. We should be doing more to help people access opportunities, wherever those opportunities are, and do less to resist the economic forces that have redistributed economic opportunities out of Buffalo and into San Francisco.