Matt Rognlie is a fun guy and a very smart guy. An economics doctoral student at Massachusetts Institute of Technology, Rognlie had a blog, but ditched it in favor of writing long comments and e-mails to bloggers (including yours truly). As proof that blogs can be an important part of the research process, Rognlie turned a blog comment into the most biting critique of Thomas Piketty yet produced.

Piketty, remember, is the French economist who became a celebrity for his work on inequality. In his blockbuster book, "Capital in the Twenty-First Century," he asserts that the rate at which capital owners' wealth increases is almost always larger than the rate of economic growth. If that's true, it means that the rich get richer, and wealth inequality never stops increasing.

But young Rognlie has three observations that cast doubt on Piketty's big thesis.

The first is that Piketty doesn't take depreciation into account. As capitalists accumulate more and more machines, buildings and other hard assets they have to pay more and more to maintain that physical capital. Trucks need new tires. Offices need renovation. What Rognlie notices is that this upkeep cost has been increasing over time. Nowadays, more than in the past capital goods are often in the form of computers, software and other high-tech products that go obsolete very quickly. That means that capitalists have to spend more money replacing these things. A lot of what looks like more money going into owners' pockets is really just an increased cost of doing business.

Rognlie isn't the first to make this point — it has been made by James Hamilton of the University of California, San Diego, and by Benjamin Bridgman of the Bureau of Economic Analysis.

But Rognlie adds two other important points. His second point is that much of the income that went to capital owners in the last six decades has been from capital gains — from stock prices going up, rather than from an increase in book value (the total net value of the assets owned by companies). When you look at book value, the increase was much more modest. It might be that what looks like the start of a permanent explosion in the wealth of shareholders is really just the end of the "equity premium" that has fascinated financial economists for decades.

But Rognlie's third point is perhaps the most interesting. Economists combine a lot of different things into "capital," such as machines, buildings and land. Rognlie points out that almost all of the increase in the value of capital over Piketty's timeline comes from land, instead of from other forms of capital. In other words, it's landlords, not corporate overlords, who are sucking up the wealth in the economy. It's a dramatic, startling insight that was somehow overlooked before Rognlie came along.

This is a very different story from the one we usually think of. Didn't we relegate all-powerful landlords to the dustbin of history when we got rid of feudalism? Haven't productive corporations replaced rent-collecting landlords as the wealthy class in advanced societies?

Maybe not. Urban economists believe that as density increases, productivity increases. This is what is known as an "agglomeration economy." But as it becomes more valuable for people to work and live near each other, the value of central locations — of land — goes up. Landlords, who are producing no more than they used to, but who were sitting on advantageous locations, reap huge benefits.

In general, this will mean cities tend to be too small — the incentive to cluster together for higher productivity is choked off by the high price of land. The drain of urban income to landlords will tend to increase as the economy grows and the productivity advantage of cities increases. To see this in action today, just look at San Francisco, where the soaring price of land, and the accompanying surge in rent, has absorbed much of the wealth created by the new tech boom. Of course, this has been heavily exacerbated by the city's refusal to allow more housing construction, which may be a reflection of the political power of landlords.

Rognlie's results, and the theory of agglomeration economies, suggest that to fight wealth inequality, what we really need to do isn't to redistribute income from corporations, but to redistribute income from land. How do we do that? Well, allowing more development in urban areas is a good start. But the real weapon here is the Henry George tax, or land value tax (LVT). This is like a property tax, but it taxes only the value of land, not the value of the structures built on the land. It encourages efficient use of land, while providing the most efficient method of wealth redistribution. Milton Friedman called it the "least bad tax."

Rognlie shows that if we really want to counter the inequality that Piketty warns about, we will probably need tools like the LVT. Most importantly, our cities will need to find ways to counter the political power of landlords. The most important redistribution policy may now be urban land-use policy.

Noah Smith is an assistant professor of finance at Stony Brook University and a freelance writer for finance and business publications.