Chris Stucchio recommended Matt Rognlie's criticisms of Piketty (paper, summary, Voxsplainer).

Rognlie starts by saying that Piketty didn't correctly account for capital depreciation (ie capital losing value over time) in his calculations. This surprises me, because Piketty says he does in his book (p. 55) but apparently there are technical details I don't understand. When you do that, the share of capital decreases, and it becomes clear that 100% of recent capital-share growth comes from one source: housing.

I can't find anyone arguing that Rognlie is wrong. I do see many people arguing about the implications, all the way from "this disproves Piketty" to "this is just saying the same thing Piketty was".

I think it's saying the same thing Piketty was in that housing is a real thing, and if there's inequality in housing, then that's real inequality. And landlords are a classic example of the rentiers Piketty is warning against.

But it's saying a different thing in that most homeowners use their homes by living in them, not by renting them out. That means they're not part of Piketty's rentier class, and so using the amount of capital to represent the power of rentiers is misleading. Rentiers are not clearly increasing and there is no clear upward trend in rentier-vs-laborer inequality. I think this does disprove Piketty's most shocking thesis.

Rognlie also makes an argument for why increasing the amount of capital will decrease the returns on capital, leading to stable or decreasing income from capital. Piketty argues against this on page 277 of his book, but re-reading it Piketty's argument now looks kind of weak, especially with the evidence from housing affecting some of his key points.

Grendel Khan highlights the role of housing with an interesting metaphor:

Did someone say housing?

As an illustration, the median homeowner in about half of the largest metros made more off the appreciation of their home than a full-time minimum-wage job. It's worst in California, of course; in San Jose, the median homeowner made just shy of $100 per working hour.

See also Richard Florida's commentary. See also everything about how the housing crisis plays out in micro; it is precisely rentier capitalism.

In the original post, I questioned Piketty's claim that rich people and very-well-endowed colleges got higher rates of return on their investment than ordinary people or less-well-endowed colleges. After all, why can't poorer people pool their money together, mutual-fund-style, to become an effective rich person who can get higher rate of return? Many people tried to answer this, not always successfully.

brberg points out that Bill Gates – one example of a rich person who's gotten 10%+ returns per year – has a very specific advantage:

Not sure about Harvard's endowment, but it's worth noting that the reason Gates, Bezos, Zuckerberg, and other self-made billionaires have seen their fortunes grow so quickly is that each of them has the vast majority of their wealth invested in a single high-growth company.

This is an extremely high-risk investment strategy that has the potential to pay off fantastically well in a tiny percentage of cases, but it's not really dependent on the size of the starting stake. Anyone who invested in Microsoft's IPO would have seen the same rate of return as Gates.

This is a good point, but most of Piketty's data focuses on college endowments. How do they do it?

Briefling writes:

I'm not sure you can take the wealth management thing at face value. The stock market since 1980 has 10% annualized returns. Instead of trying to replicate whatever Harvard and Yale are doing, why don't you just put your money in the stock market?

Also a good point, but colleges seem to do this with less volatility than the stock market, which still requires some explanation.