SAN FRANCISCO – There is an apartment for rent in a renovated former warehouse here, across the street from the Caltrain commuter rail station. It has two bedrooms and two bathrooms and nearly 700 square feet of space. It comes furnished. It is, the online advertisement proclaims, a "perfect place" for someone, like so many young startup employees today, who works in Silicon Valley but lives in the city. It could be yours for $8,525 a month.

Rents get more bearable in farther flung parts of the city, and in bedroom communities down the Peninsula, but only slightly. A deluge of cash has soaked the engineers and executives of the Bay Area, pushing up home prices and leaving San Francisco with levels of income inequality typically seen in developing nations.

Paul Graham, a venture capitalist and one of the founders of the startup incubator Y Combinator, would have you believe this rising inequality is a good thing. Or, at very worst, the inevitable consequence of a good thing. "You can't prevent great variations in wealth without preventing people from getting rich," he wrote in an essay that went viral online last week, "and you can't do that without preventing them from starting startups."

Graham's piece happened to light up the Internet just as thousands of economists descended on San Francisco for the annual conference of the American Economic Association. The drizzly three days of the gathering featured what appears to have been the largest focus on inequality in the organization's history. The 70 inequality-themed papers presented here wove a nest of new research demonstrating how and why inequality has increased, and what side effects appear to have accompanied it.

Taken together, they make the case that Graham -- and others who wave off inequality as inconsequential -- has misread what's happened in the American economy. They suggest that everyone should worry about the drivers and consequences of inequality -- even venture capitalists.


Since the middle of the 1980s, a gap has been widening between what the best-paid Americans earn and what everyone else in the country earns. There has also been a gap between how much money those two groups spend, but for a long time that gap didn't grow as much, because "everyone else" borrowed heavily to try to keep up with the folks at the top.

The Great Recession cut off that borrowing. It also reduced real incomes for the bottom 90 percent of American income-earners. Smaller paychecks and less access to credit left those Americans with only one choice: spend less than they had been. The top earners didn't increase their own spending by nearly enough to make up the difference. That is a big reason the U.S. economy has recovered so sluggishly from the recession, the economists Barry Z. Cynamon and Steven M. Fazzari declared in a new paper presented at the conference.

"Rising income inequality," the wrote in their research paper, "is now a significant barrier to economic growth and full employment."

There was already some irony in that paper, and so many others on inequality, landing in the northern outpost of Silicon Valley. Graham's essay cranked that irony to foghorn decibels.

Inequality, Graham argues in his piece, is not bad "per se." It can stem from rich people exploiting tax loopholes or poor people being locked in prison at high rates, which are both bad things. Or, it can result from more tech entrepreneurs getting rich, which is a good thing, because those entrepreneurs are making valuable things for the economy.

The argument takes several turns, but it boils down to two points. First, whether inequality comes from good or bad sources, it does not, by itself, hurt anyone; just because the rich get richer does not mean the poor and middle class can't get richer, too. He concedes that some rich people got that way by taking money from the poor, but not most of them.

That's his second point: We shouldn't try to reduce inequality, because doing so would necessarily mean killing off the innovators and entrepreneurs who get rich for socially good reasons. He writes:

"Eliminating great variations in wealth would mean eliminating startups. Are you sure, hunters, that you want to shoot this particular animal? It would only mean you eliminated startups in your own country. Ambitious people already move halfway around the world to further their careers, and startups can operate from anywhere nowadays. So if you made it impossible to get rich by creating wealth in your country, the ambitious people in your country would just leave and do it somewhere else."

Research suggests Graham is both overestimating the importance of startups to inequality and underestimating the damage high inequality can inflict.

There is evidence to support his contention that returns to entrepreneurship are making inequality worse. The problem for his argument is that the effect doesn't appear to be large, comparatively.

At the conference, a group of five economists, including Philippe Aghion of College de France and David Hemous of the University of Zurich, unveiled a study that speaks directly to the Silicon Valley effect on inequality. They looked at levels of innovation (as measured by a particular kind of patent production) across individual states over time. They found a significant relationship between increased innovation in a state and the increased income share of the top 1 percent of earners in the state.

When the math cleared from their analyses, the economists estimated that 14 percent of the increase in the share of income going to the top 1 percent of Americans between 1975 and 2012 "may be explained by an increase in innovation." The economists say those increases are temporary, they generate economic growth and they're associated with stronger upward mobility. Those are all good things. Point, Graham.

If you take that as a proxy for the Silicon Valley effect, though, you're left with a problem: 86 percent of the recent inequality increase can't be explained by innovation. You're also stuck with the fact that startup formation for tech companies has been falling for more than a decade even as inequality has been widening, and not rising, as Graham implies. Total venture capital funding remains well below late-1990s levels, even before you adjust for inflation, according to data from the National Venture Capital Association.

In light of all that, it's difficult to conclude that startups are mostly driving the income gap.

Economists have suggested all sorts of other things that might be contributing to the gap. There's the "superstar effect," which says the opening of global markets is delivering big payouts to the very top players in a variety of industries now rolling in cash from around the globe. There's a broader theory that advancing technology is delivering higher payouts to the most skilled workers, and leaving lower-skilled workers to compete with robots and less-expensive foreign labor, pushing down their wages.

There is also mounting evidence that some, or even a lot, of inequality has nothing to do with superstars or skills -- at least not in the way that we tend to think about them. The culprit in this theory is what economists call rent-seeking. Really simply put, it's people making extra money not because they deliver a good or a service that the market values, but because they've bent the rules of some system to shuttle more compensation their way.

Examples of this would include chief executives who pad their profits by lobbying for government favors, and reap big bonuses as a result. Or financial pros who score huge fees helping clients game the tax code to pay less to the IRS. There are all sorts of other examples, but a basic rule is, if you're earning more money without doing anything to add additional value to the economy, you're rent-seeking.

Several papers presented in San Francisco hinted at, or directly identified, rent-seeking as a prime cause of growing inequality. Most forceful was the liberal economist Dean Baker, from the Center for Economic and Policy Research, who argued in a paper that rent-seeking in copyright protections and the financial sector, along with excess pay for CEOs and other highly educated professionals, explains "the bulk" of the growth in top incomes.

If his argument is correct, Baker writes, "it means that there is nothing intrinsic to capitalism that led to this rapid rise in inequality." It would mean you couldn't reduce inequality simply by raises taxes on the rich -- you'd need policies to straighten kinks in the market, by shrinking the financial sector or reducing copyright windfalls. It would also mean that reducing inequality shouldn't dampen entrepreneurship.

Other recent studies also suggest high levels of rent-seeking are driving inequality. Brian Bell and John Van Reenan, a pair of economists in Britain (which resembles the United States in many ways when it comes to inequality), reported in 2014 that increased bonuses for bankers accounted for two-thirds of the growth of top 1 percent incomes in Britain after 1999. There are all sorts of reasons to believe that premium financial sector pay is almost entirely rent-seeking; the British paper would suggest that at least two-thirds of inequality could be linked back to "bad" sources, in other words.

Oliver Denk, an economist for the Organization for Economic Co-operation and Development, expanded on that work in two papers for the conference. He dove into statistics for every country in Europe and pieced together a composite of the men -- they're almost all men -- who make up the top 1 percent of earners on the continent. Again, his findings give us a window on America, where inequality is even more pronounced than in Europe.

"Workers in the top 1 percent," Denk found, "tend to be 40 to 60 years old, be men, have tertiary education, work in finance or manufacturing and be senior managers." At worst, that's a group ripe for rent-seeking. At best, it doesn't sound much like the innovators Graham is so worried about discouraging. The liberal Economic Policy Institute has done similar work on America's top 1 percent, and found that at least a quarter of its increased income share since 1979 can be attributed to finance.

Taken together, that evidence suggests Graham may be overly worried about inequality "hunters" discouraging today's crop of startup founders; they just aren't that prevalent in the ranks of the super rich who are pulling away from everyone else.

Other evidence suggests he's wrong on his second point, too: Inequality appears to have consequences for the economy, no matter what's driving it.

At one point in his essay, Graham suggests replacing concern over inequality with concern over poverty and social mobility. He suggests those are distinct from the growth of incomes for the very rich. Americans, he writes, fall prey to "the pie fallacy: that the rich get rich by taking money from the poor." He says it's possible, but not always true, that the pie can grow such that everyone gets more, even if the rich get a lot more.

It is indeed possible for that to happen. It just doesn't appear that's what's been happening in America in recent years. Instead, it appears the very rich have been enriching themselves further at the expense of workers. That's what researchers from the International Monetary Fund found in a paper presented at the conference, which argues that the rise in top-level earnings is correlated with a decline in union membership; when workers get less, essentially, their bosses and shareholders get more.

A similar argument lies at the heart of the paper by Cynamon and Fazzarri, who are economists at the St. Louis Fed and Washington University in St. Louis. Their contention is that the lack of pie growth, if you will, for most Americans created a shortage of consumer spending that has kept the recovery from recession at historically weak levels. The stagnation of most incomes below the very rich "has opened a large gap in demand generation that has not been filled," they write.

Perhaps most resonant to Graham is the idea that inequality might be discouraging smart workers from working at all -- in startups or elsewhere. A paper from economists at the New York Fed and the University of Southern California finds a link between growing inequality and declining labor-force participation among highly educated women, whose exit from the workforce hurts the economy. The economists find that as some highly educated men's wages rose quickly, their highly educated wives earned and worked less.

If there is middle ground between Graham and this body of research, it's the idea that policymakers shouldn't go after inequality with blunt instruments, like big tax hikes just for the sake of soaking the rich. Perhaps, instead, they should target rent-seeking, which economists agree is bad for everyone who isn't a rent seeker.

That could mean taxing financial transactions to discourage excessive Wall Street trading, or taxing income from investments at the same rate as income earned on the job, in order to discourage expensive financial engineering. It could mean simplifying tax and regulatory codes to reduce companies' gains from lobbying -- which, by the way, big tech firms are increasingly chasing via ramped-up presences in Washington.

If you curbed rent-seeking -- or what conservatives like to call cronyism, in a solely government-focused form -- you'd almost certainly improve growth and help more people share in its returns. Graham's worry is that you wouldn't actually reduce inequality in that case. All the former rent-seekers, he writes, would flow into entrepreneurship: "If the only way left to get rich is to start startups, they'll start startups."

In economic terms, that would be a great problem to have.