Thanks for the summary.
I didn't understand from Piketty how s actually gets calculated from empirical data. Then again, I did not read all the footnotes from this chapter, and I have not read either wikipedia or a textbook chapter to learn more.
For the PhD != tractors case above, is there an assumption about growth or resource availability necessary for this condition to hold?
I liked P's comments about Marx in this chapter.
Embarrassingly, I haven't read this week's chapter yet. I may start commenting tomorrow.
Thanks for the summary, Conflated.
I read it, but I got bored and have very little in the way of substance to offer.
Good job, conflated, you killed the reading group.
I just want to say that these summaries are the perfect amount of vicarious Piketty for me. Kudos to all of you with more Pikettendurance than me.
Wait, we are supposed to read the footnotes too?
I've been keeping up, but this chapter and the last one have been pretty tough slogs, and I'm not sure that I am retaining much. I definitely couldn't handle more than one of these a week.
I guess the one thing I remember was returns on capital declining from the 4-5% range to the 3-4% range as a result of (I guess?) declining marginal productivity on capital investment.
People selling you financial services will tell you the stock market returns 8% on average over time. I'm assuming that's likely to decline as well. I imagine that number includes the return on capital, plus a bit of a risk premium (probably offset by bonds in the other direction). Also, not sure if that number is real or nominal, which seems kind-of important.
I've been consistently a chapter and a half behind, but I like the spoilers.
9. Nominal, and I think that's based on SP500 since some date after WWI. An overestimate, since no period of capital loss due to war or catastrophic change of government is included. I'm looking forward to anything he'll have to say about Germany 1922-1925.
Well, if its nominal, the real number is probably not far off from Piketty's range. Especially considering that basing it on the S&P 500 is bullshit, given that companies come and go from that index.
2:There's not a lot if detail on the calculation of elasticity in the book notes themselves, in fact I have a feeling he doesn't even mention it by variable name. I haven't tackled the technical appendix yet. He mentions elasticity is between 1.3 and 1.6 though so they do explicitly calculate it.
I have just realised I really should not have used s for elasticity when it is already used for the saving rate, though. Damn, that could be very confusing. I looked for a conventional variable name and stuffed up. Um, I can send a new version using the variable E? The whole last set of cases should be E > 0, etc.
12: Is that a meaningful distinction in practice, though? I thought indices were supposed to track overall stock performance, not be "picks". Checking, sure enough, the 5-year performances of Dow Jones, S&P 500, NYSE composite, and NASDAQ composite are all within 0.1% of each other (specifically, they all round to -0.1%).
There is a bit less maths in the chapter than in my summary, and I reorganised the elasticity sections in ascending order of elasticity. This is less narrative than the chapter itself. Piketty explains it clearly enough, but he sort of introduces the Cobb-Douglas case and then rounds back on the others as kind of rebuttals. I found it hard to summarise that without reordering, but may have sucked extra moisture out that way.
Oh, Piketty uses sigma for elasticity on p38 of the technical appendix. He works back from the capital share to get the calculated elasticity using a constant elasticity of substitution (CES). Then he says to remember that CES like Cobb-Douglas is rubbish and go read his technical book, "Capital is Back", p34-37.
14: Yes, in practice the distinction between S&P 500 and all publicly traded companies is totally unimportant. The value of the largest companies is most of the value of all publicly-traded companies. Even the Dow is an okay proxy, even though it has a weighting scheme designed by morons.
9: No, it's a real rate of return, as cited in my comment here in TFA. 8.7% real rate of return on US common stocks, 1926-77.
Now one difference between that computation and Piketty is that the former counts unrealized capital gains by marking the value of the (hypothetical) investment to market at the end of each period, while I believe that Piketty's r only looks at actual cash flows. There are arguments for each approach - the former is arguably more relevant for someone trying to decide between different investments, while Piketty's approach is more stable over time, smoothing out the bubble/crash mania of the market.
That difference may explain much of the difference in the measured rates of return. E.g., Piketty's approach would say that Larry Page (co-founder of Google) would have a much lower rate of return on his capital so far, because GOOG has never paid a cash dividend, and Page still owns most of his original stock. (Though it's still pretty spectacular, given that he was presumably starting from a pretty small base). The two calculations might converge over a sufficiently long period of time, but the required time may be measured in centuries instead of decades.
I found the part of the chapter about production functions kind of difficult to penetrate because it was missing equations. The definition of the Cobb-Douglas production function as a power law in the footnote wasn't very enlightening since it wasn't clearly explained what the function means.
I believe that Piketty's r only looks at actual cash flows
This would explain the difference between average stock returns calculated in Finance textbooks and what P writes. Would an r calculated this way show a loss in case of catastrophe or business collapse though? Such a loss seems like the opposite of an unrealized capital gain.
17: Yeah, if you go from the S&P 500 to the Wilshire 5000, you find that the top 500 companies constitute somewhere between 75-80% of the value of the top 5000. The guy I linked to argues that you should be investing in the broader-based index, but I take that as an argument that it doesn't make that much difference, unless the smaller companies are spectacularly better than the larger ones.
And now I'm starting to read Wikipedia entries on economics when I haven't finished writing my talk for tomorrow yet.
I thought indices were supposed to track overall stock performance, not be "picks".
My concern is not the contents of the index, rather that indexes tend to overlook the losses inherent to companies that fall out of the index. Sure, if your index consistently replaces the worst performers with strivers, its going to look good. But if, say, Enron is in you index, and the value ends up being 0, the index does't account for that loss. Rather, it replaces Enron with Google, and calls it a win.
If the S&P 500 knows to replace companies like Enron before they loses their value, they are better at stock picking than anyone else alive today.
There are reports that Costco is now selling Piketty's book. Maybe there's an oversupply.
23: I'm not sure what you think the issue is. When tracking index/fund performance, if the index drops company A and replaces it with company B, you treat it as a sale of company A on the last day it is part of the index, and a purchase of the appropriate number of shares of company B on the next day. If there's a difference in the price of the two blocks of shares, that becomes a gain or loss to the return of the index/fund over the period in question. Things get a little trickier if company A is eliminated because it is merging with company C that is also part of the index, but the basic idea is the same - you make sure you wind up with the right number of shares of B and C at the end, and treat any differences in the costs involved as a gain or loss to the fund.
26: It's reasonable to think there might be an issue; it just turns out that it's not that big of an issue in practice.
I guess I can buy that its not a huge issue in practice. But its just never smelled right to me. I'd be interested to see some data on what the actual effect is.
I have been inspired by Piketty to read this book:
http://www.amazon.com/Economic-Growth-Edition-David-Weil/dp/0321416627
Used copies are pretty cheap. It is basically anecdotes about different countries combined with algebra.
People like Cobb-Douglas production function because it is relatively simple. they like fixed shares of capitol and labor because that that stuff is in exponents. Even more humorously, they like to use the income share of capitol (alpha) as 1/3, rather than .35 or something, since an important equation derived from the Cobb-Douglas production function has an exponent of the form (alpha/1-alpha) and isn't it nice (especially before computers) when that is 1/2.
That's a lot of money for a building.
So why does Piketty say there's no self-correcting mechanism keeping the capital-income ratio steady? There isn't in practice now, agreed, but I thought it was said earlier that there's probably a point where there's so much capital it has a hard time accumulating as a percentage of national income; we're just nowhere near that point, which might be at a ridiculous ratio.
Yes this is about right ... he basically accepts there must be a theoretical maxima due to the oversupply of capital, but it's far beyond anything we've experienced historically. There's basically no sign of it.
I think this is the Larry Summers criticism of Piketty though.
I think that's right about Summers.
What would an oversupply of capital look like? Wouldn't it just result in a bidding up of the prices of every asset with a positive expected yield, making it so that only those with an ability to absorb large losses on their way to eking out a small average profit (that is, the very wealthy) could afford to invest?
Meaning eventually only the wealthy with their large risk-bearing capacity could profit from investing.
I guess 35 addresses the distribution of income from capital, not capital's share of the total income. Regarding that, I don't see why that would be limited in my scenario. The return to capital would drop but not it's total share.
34. I think there's more than one configuration. Sustained extremely low interest rates coupled with high unemployment looks like too much capital.
I'm still puzzled about 20-- if Piketty is using net profits (or maybe only dividends? ), for r, then I don't understand how the accounting works out when equity gets destroyed. There are businesses that generate zero or near zero profit-- are they invisible for P?
Low interest rates, sure, but why does high unemployment reflect too much capital? Or do you mean high unemployment in the presence of low interest rates? In which case, same question. In any case, if we currently have a glut of capital (and I agree that we do), do you see something to suggest that it is being prevented from increasing its share of income?
Thanks for this summary. Wish I could have participated more in this discussion when it was more active. This chapter really seems to get at the nub of the book and maybe potential problems with it. I think he agrees that the return to capital will go down as the capital to income ratio goes up -- the only question is whether it drops sufficiently to automatically result in a relatively constant capital vs. labor share of income, or whether there is something about capital such that increased capital will over time increase the total income from capital/income from labor ratio. Piketty says "yes there is something special about capital"; Larry Summers says "no there isn't and so we shouldn't expect an increasing capital/output ratio to produce an increase in capital's share of total income. On Summer's view, more capital just means that capital gets paid less and labor gets paid more, so the share of total income should remain about the same. I believe people who agree with Summers emphasize rising labor incomes in the late 19th century.
I don't think Piketty does a very good job of explaining why Summers is wrong. With that said, and this isn't based on anything other than intuition and I could be totally wrong, it seems to me pretty likely that the Summers theory won't hold in the future absent massive political and social change in rich countries/the United States, or rather that it won't hold except in a very narrow sense. First, the growth of labor-replacing technology seems likely to accelerate; in a world of robots, it may be that capital is just able to increase its share permanently. Second, whatever increased gains go to "labor" will in fact be manipulated by the very wealthy such that apparent gains to labor are things like the salaries paid to hedge fund managers and will effectively serve as a compensatory form of increased income in order to maintain class position of the very wealthy even in a world of declining rates of return to capital.
37: Wouldn't losses just be shown as negative cash flows?
38: Yes I thought unemployment would show an oversupply of labour. Low interest rates might just mean low-ish return on capital and low growth. Even on low returns to capital, remember for Piketty high postwar capital is an anomaly returning to lower values later.
39: Yes, the third way neoliberal critique (Summers) seems to focus on oversupply of capital.
The left critique seems to focus on the Cambridge Capital Controversy, ie marginal pricing vs the labour theories of value. Eg Jacobin and the LRB (this week) taking this tack.
I haven't gone through the critiques in detail yet, trying to focus on the book first. They don't at first pass seem terribly empirical. They are mostly attacks from theory ... Summers says supply and demand still applies to capital. Others say marginal pricing doesn't apply well over the long term.
Though it's a crucial research question, I think it also misses the point. Piketty doesn't need to tell you why it's happening for the lack of a constant capital income ratio to be a crucial, policy-impacting result. He just needs to show that it is, and with trucks full of data, he has.
I'm a bit confused how the capital supply critique can stand in the absence of this data. Unless they are pointing at the 19th century and saying "you get there eventually, just wait a decade, it's just around the corner".
The marginal pricing critique is saying Piketty's definitions are wrong I think, but it still seems a bit weird in the absence of data.
I'm reminded of the hoary old joke about French engineers looking at an American car and saying "I know it works in practice, but we want something that works in theory!" Except the with economists and role reversal and such.
So Summers was implicitly saying "supply and demand apply to capital too, therefore we are at an equilibrium point like in every graph any economist draws"?
FINAL EXCITING HOLD LIST UPDATE
The library now has 19 copies and I should be able to pick one up tomorrow. If there are still unclaimed chapters and if the reading group will still be going in a few weeks, I can claim a chapter. The nature of the hold list is such that I won't be able to make a renewal, so I either read it in the next three weeks or try to read the academic library ebook I have access to. Or, if I really want the book, I'll just buy it. Anyway, committing to write up notes will make it more likely that I read it.
2,43: The Summers review is not a bad read and it actually fits a missing puzzle piece
http://www.democracyjournal.org/33/the-inequality-puzzle.php?page=all
It is a fundamental disagreement, though he also suggests Piketty should get a Nobel, so it's not exactly pistols at dawn.
Summers says elasticity: 0 1. (This is the PhD != tractors case in my summary. So what does it look like? Today, if you are Larry Summers. It is not hard to think of parts of the economy where this is true. Like it doesn't matter how many pens you give a mathematician beyond perhaps 3: it doesn't make them discover theorems any faster. I find it hard to get a grip on what this would look like in the large though, which I feel better about now, as apparently I have company.
Larry says it looks like E>1 in Piketty's data because of globalisation and technological change, ie alpha has changed over the last few decades. This seems to be more his professional opining.
Secondly, he claims savings as a proportion should go down as wealth goes up. This sounds strange it me. If you have less urgent spending as a proportion of income I would think you would out more in the bank. Certainly going from poor to middle class, but I don't see why it wouldn't go all the way up until you hit wastrels. Even then what is the wastrel production rate vs merely gormless dynasts that maintain wealth of their parents? Clearly a topic for future research.
Summers calls for neo-Keynesian infra spending at the end.
Summers says elasticity: 0 < E < 1.
44: It will be ongoing, and we're only claimed up through eight. Pick a chapter at will from nine on forward.
HTML supremacists. Free less than! No more ampersand!
Picked up the library copy just a short while ago. It's got, like, a lot of words.
Even within the neoclassical model, I think one might expect capital's income share to be fairly invariant with respect to the supply. Summers' position requires the supply of capital for an investment to increase with a decrease in the risk, so a scaling up of the supply curve shifts the market clearing price in the direction of less income for capital. Presumably, the supply of capital increases with safety because there are multiple owners of capital with varying degrees of risk tolerance. But if you are risk averse you can always pool your resources with others. You probably have to, because otherwise you will always be outbid by those more able to bear risk. The result is that the only entities offering investment capital will be relatively homogenous in their risk tolerance, and so the supply curve will look like a step function, and the market clearing price won't shift with a scaling up of supply.
51 is probably gibberish, but at least it's gibberish buried in a dead thread.
On the subway to work, I sat next to a ridiculously appealing young man reading Capital. Big muscular kid with a deep tan and a face like an Irish boxer in a 30's movie. And he stood up to give his seat twice, to two successive old ladies. (The first one got off after a couple of stops, he sat back down, and then got up for another one.) I was considering telling him I was in love, but I figured that would be creepy.
(The woman sitting cattycorner from us was reading The Shock Doctrine. All in all, score one for the A train this morning.)
Sign me up for Chapter 11. In the Piketty book too.
The funny thing about Piketty is that he has a lot more faith in returns on invested capital than any professional investor I've ever met. It's actually very interesting about his book. This is exactly what you'd expect form a French socialist economist. He assumes it's really easy to put money in the market for 40 years or 80 years or 100 years and have it compound at these amazing rates. He never explains how that's supposed to happen.
Every investment manager I know is sweating the opposite problem, which is: what do I do? Where do I get the growth? I can't get into the public market, so I have to go into the private market. The problem in the private market is there isn't much growth. Maybe a dozen hedge funds. After that they're not that good. The returns degrade down to S&P 500 levels.
I don't get this criticism. The idea that something relatively understandable and predictable in the aggregate is actually quite complex at the level of the individual is supposed to be a rebuttal?
51: There is a kind of step function effect for people with small amounts of capital. Like if you have $1000, different investments are available to you vs having $100k or $10m.
However, Piketty points out wealth is pretty concentrated, so the effect doesn't make much of a difference to the overall average return on capital, s.
Either way, I think Summers and Piketty are both talking about the total amount of wealth washing around the economy, not a particular individual? And they both agree return decreases with increased supply, just disagree about how fast.
Or are you saying there is some economy-wide effect when a country gets rich enough?
56: It's even dumber than that; he's starts from the complaint that it's hard to get above average returns. It's a complete non sequitur.
57: I'm trying to rationalize my intuition that there isn't a strong limit to capital's share of income based on supply and demand. For there to be, for Summers to be right, the return to capital has to decease faster than one over the supply. That is, proportionally diminishing returns are equivalent to a fixed share of income, so at some point they have to decrease super-proportionally.
58: goddammit why did I go look up that interview? Andreessen is such a nitwit.
Summers needs there to be a continual supply of new investments with positive but decreasing yield made available as the supply of capital increases, so that the market clearing return drops and capital's surplus is decreased.
I think I'm getting even less coherent. I need to stop writing about economics when I'm really tiered but can't fall asleep. Besides, I'm interrupting tonight's teo and Smearcase show.
It's okay. No one watches the show anyway.
My intuition says that there isn't a supply of positive yield investments (or at least not detectable ones going untapped because of some shortage of capital), and that to a good approximation an increase in the supply of capital will instead just bid up prices.
Mark Andreesen is so full of shit in that interview that I was daydreaming of how I would write the definitive refutation. I didn't even get to the part about Piketty before I couldn't stomach it anymore.
42.last: Reagan made that exact same joke about economists when he was President.
The problem in the private market is there isn't much growth. Maybe a dozen hedge funds. After that they're not that good. The returns degrade down to S&P 500 levels.
Haha, he doesn't know the difference between alpha and absolute growth. What a clown.
Also from Andreessen:
" If the gains from the market exceed the gains from economic growth, the delta is productivity growth. That's the math."
What is this I don't even.
56,67: He's obviously wrong, but I think this does capture the social dynamic of reducing returns r as the capital supply increases. You have wealthy getting richer but working ever harder to chase the diminishing returns.
This feeds into a political dynamic too, where one group of people say inequality is increasing, the rich get richer, and people with even moderate wealth feel they have to work ever harder to keep getting any return on capital, so how could they be fat cats?
Ugh I've lost the habit of using letters as variables, I keep mixing them up.
I had no idea the teo and Smearcase show was tv rather than radio. My mind is blown. I'm still a little too sick and feverish and pitiful to keep up with Piketty, alas.
He seems to be describing a dynamic where the wealthy have access to higher r than the rest of us losers, who are stuck at S&P levels.
I had no idea the teo and Smearcase show was tv rather than radio.
Well, the visuals are just the text of our comments on the screen, so it's not really that exciting.
72: Yes. He's a venture capitalist though - his job pretty much is to chase higher returns through more risk. If he was talking about his job rather than the whole economy, that would be realistic.
18, 20: Going back to Chapter 5, it looks like Piketty does include corporate retained earnings in private capital, and the returns on retained earnings would be included in r. From page 157 of the Nook edition: "It is true that stock prices tend to rise more quickly than consumption prices over the long run, but the reason for this is essentially that retained earnings allow firms to increase their size and capital (so that we are looking at a volume effect rather than a price effect). If retained earnings are included in private savings, however, this price effect largely disappears... Under these conditions, it is better to treat retained earnings as savings realized on behalf of the firm's owners and therefore as a component of private saving."
So it looks like even though he's not treating Larry Page's unrealized capital gains as income, he is treating the retained earnings of Google to date as income and return on capital, hence part of r. This mitigates the effect. The finance model of stock valuation I was taught included a component that reflects the book value (including retained earnings) of the firm, and a component that reflects discounted expected future earnings. So in ignoring unrealized capital gains, P is still counting the former but not the latter.