I tend to share Rob's sense that the point about the relationship between r and g lacks the kind of formal proof that one is used to in economic presentations. But on the other hand, so what? P offers a plausible mechanism, so the onus is on people who don't like it to demonstrate that it can't work. Which, as far as I know, they haven't yet done.
I think that his future extrapolation is weaker than his reasoning about the past. Because the rapid growth in the second half of the c.20 isn't explained (much) as other than contingent, so any future projection is likely to be contingent as well. He places a lot of weight on low population growth as a factor in causing low economic growth in the c.21, but he hasn't, for example (unless I missed it) emphasised the baby boom as a driver of rapid growth after 1950.
A Marxist would argue that capitalism at the beginning of the c.20 faced a crisis of overproduction (mismatch between purchasing power outside the elite and the need for the elite to increase productivity to fuel growth), so, absent revolution, this can be resolved by massive destruction of capital (which happened) leaving room for reinvestment. But he doesn't go there either. It leaves his mechanisms looking a bit thin to be the basis of confident prediction, to me at least.
It's also a fact that the birthrate (in Britain anyway, I don't know about anywhere else) was extremely low among the elite in the c.18*. Was this a factor in holding down economic growth before the c.19, and if so, how important was it? Agricultural productivity, and to a lesser extent industrial, did grow markedly before the "take off" date, but was it constrained by the size of the market?
*Lawrence Stone did the grunt work on this IIRC. Elite families dying out left, right and centre.
Elite birthrates have been extraordinarily low in many societies. There is a huge section of Fisher's big book devoted to the impending catastrophe facing civilisation because the upper classes don't think they can afford to breed.
I don't think birthrate among the elite has all that much to do with economic growth, it's birthrate among the population generally. More labor available, more mouths to feed, and so on.
Overall, I agree with Rob and Chris -- I don't see the argument that r will once again be greater than g, beyond the fact that it has been greater for most of history and the last century has been an exception. (Which, to be clear, makes it seem perfectly plausible that things might go back to 'normal', but I've missed the basis for any certainty that they will.)
so the onus is on people who don't like it to demonstrate that it can't work. Which, as far as I know, they haven't yet done.
Summers takes a crack at it here:
Economists universally believe in the law of diminishing returns. As capital accumulates, the incremental return on an additional unit of capital declines. The crucial question goes to what is technically referred to as the elasticity of substitution. With 1 percent more capital and the same amount of everything else, does the return to a unit of capital relative to a unit of labor decline by more or less than 1 percent? If, as Piketty assumes, it declines by less than 1 percent, the share of income going to capital rises. If, on the other hand, it declines by more than 1 percent, the share of capital falls.
Economists have tried forever to estimate elasticities of substitution with many types of data, but there are many statistical problems. Piketty argues that the economic literature supports his assumption that returns diminish slowly (in technical parlance, that the elasticity of substitution is greater than 1), and so capital's share rises with capital accumulation. But I think he misreads the literature by conflating gross and net returns to capital. It is plausible that as the capital stock grows, the increment of output produced declines slowly, but there can be no question that depreciation increases proportionally. And it is the return net of depreciation that is relevant for capital accumulation. I know of no study suggesting that measuring output in net terms, the elasticity of substitution is greater than 1, and I know of quite a few suggesting the contrary.
My Summers heuristic, which has served me well, is don't trust the son of a bitch. But this argument resides right at the outer limit of my comprehension, and if it's wrong, I don't understand why (though I've got a few guesses).
The crucial argument is that charts like the one we see for r and g can explain the charts we see for wealth inequality. The primary argument here is simply that there is a plausible mechanism. As he says repeatedly, if the rate of growth is 1%, and the return on capital is 5%, capital owners only need to save a little more than a fifth of their income from capital in order to earn money faster than people who work for a living.
This may be a stupid question but why is it assumed this is a one-way relationship? Inequality can decrease growth through reducing demand. This is one counterargument to the Summers position, incidentally.
4: if anything, elite families dying out all the time might decrease inequality and thus increase economic growth...
Hmm. I have a policy idea now that I think bob could get behind.
Bob is busy wondering if he should register at LGM just to throw Koizumi Junichiro (Uchiyama Yu book) at their pathetic excuses for Obama. Don't think so. Don't care.
Meanwhile Dan Kervick wrote his ass off in May-June trying to get mainstream economists to not cheat with Piketty. At many other sites too. The June 18 post with its embedded links may be on today's topic.
But what happened to estates when family died out? If escheat still followed the feudal model at that time, from what I can tell it would have gone to the next aristocrat up the ladder. Or under a more modern system, I imagine the nonfinancial assets would have been auctioned off. I don't see inequality decreasing majorly as a result of either.
I think I can still do chapter 11. It may be a summary that does not refer to other chapters in great detail, though.
Sounds good. If you need to flake, just email me and I'll cover.
10. At what date and in what country?
1 et seq: I think Piketty acknowledges pretty straightforwardly that his tools are blunt instruments. His assumption of a 3-5% future return on capital seems to be based on the fact that absent upheaval, that's what it always seems to be. (This is part of the way I'd answer my own question in 5).
5: Summers is basically saying r will rapidly decrease as wealth accumulates, such that labour becomes more relatively valuable. He has some data to back him up, a number of papers but not big data sets as I understand it.
Piketty doesn't assume labour-capital elasticity >1, he gets that estimator based on his historical dataset. That is unusual, but he also has a larger dataset than others seem to (relying on others summaries of the state of the research). Basically the rate of return seems to go down as wealth accumulates, but not that fast.
Overall, I agree with Rob and Chris -- I don't see the argument that r will once again be greater than g, beyond the fact that it has been greater for most of history and the last century has been an exception
My complaint about the lack of detail in the argument for the connection between r-g and inequality was only for the historical portion of his thesis.
He's not firmly committed to his predictions at all. In fact, he's hoping people will act in ways that make his predictions false. This part of what he is doing is politics, not science. He's not saying "this number will be that, and if not, I'm wrong." The prediction is there to motivate action.
.18 Britain had a fully capitalist economy and legal system. You could leave your stuff to anybody you wanted. If you died intestate there were complex rules about dividing your estate among your relatives. If you had no offspring, then it was divided among your collateral relatives, going back to relatives of the person who first brought the property came into your family. If your family was so completely extinct that the lawyers couldn't find any relatives at all, however remote, I'd guess it was tied up in probate until its value had all been spent.
The arguments on the elasticity of capital and labor are on pages 220-222. What he says is actually that the capital labor elasticity was less than 1 in the pre-industrial era, and greater than 1 in the industrial era. As conflated says, he makes this claim based on his dataset, specifically the dynamics of α and β. Also the differences in the elasticity before and after industrialism makes sense, because in the industrial era there are many more uses for capital, while in the pre-industrial era, capital was just land.
I'm not sure any of this ultimately matters, though, because we see r-g driving inequality both prior to 1700 and in the period 1700-1914. If the capital-labor. Increase capital may have driven down the return on capital before 1700, but growth was so low r-g could still drive inequality. So really it looks like r-g could drive inequality no matter what the elasticity of labor and capital is.
18: Because of my Peter Wimsey reading, I know quite a bit more than I should about British probate laws in the early 20th century.
20. My Peter Wimsey reading confused me about them. There's a very late (and substandard) short story called "Talboys" set in 1942, in which it's mentioned that Peter's fortune in entailed. This is puzzling i. because of the late date and ii. because Lord Peter is a younger son, who wouldn't normally have inherited any entailed property- that would normally have gone to his older brother Gerald, Duke of Denver. I think Sayers had skimped on her homework there.
So most likely when a family dies out, its fortune gets distributed among relatives so distant they're not considered part of the same family, and maybe in part to lawyers. I still don't see how that decreases inequality appreciably (to 7) - maybe it disperses the fortune across 2-3 times as many people, but they're likely to be rich too.
22: At some point, his brother died and his nephew was killed during the Battle of Britain. Maybe it was after that?
22: Couldn't Peter's father have left him money/property out of the portion of his own estate that wasn't entailed in favor of Peter's older brother, but giving Peter only a life interest and entailing the property in favor of Peter's children? It would be a new entail -- Peter would be the first owner subject to it -- but I don't think there's anything implausible about that having happened.
24: Did Gerald die? I didn't think Peter was the Duke in anything I've read.
You're right. Wikipedia says he didn't die until 1951.
There's a very late (and substandard) short story called "Talboys" set in 1942, in which it's mentioned that Peter's fortune in entailed. This is puzzling i. because of the late date and ii. because Lord Peter is a younger son, who wouldn't normally have inherited any entailed property- that would normally have gone to his older brother Gerald, Duke of Denver. I think Sayers had skimped on her homework there.
No, it's mentioned in "Busman's Honeymoon" - Harriet is asking him, post-engagement but pre-marriage, about children and whether he wants any, and he at first expresses reluctance but then admits that under the terms of his own will he has left his property entailed. But that isn't property that he inherited through entailment - that's non-entailed property (mostly in London) which he inherited under his father's will and has built up. The entailed property is Duke's Denver, which Gerald got.
Slightly pwned by LB (there's a surprise).
25. He could, but he'd have to have done it not earlier than 1900, which would have been a bit eccentric. Entails weren't fashionable in the last century. Inheritance taxes were making rich bastards choose other instruments which were more likely to keep things intact.
So even if the 15th Duke had created a new entail for Lord Peter's descendants it's not likely that by 1942 he'd still expect to be bound by it. He'd break the entail as soon as his son was of age and put it in a Jersey trust or something. It was just a bit jarring to see Mary Sue Harriet Vane expecting that their older son would still be bound to inherit everything at so late a date.
[A lot of "old money" did find its way to the Treasury, as Piketty implies, because not every rich bastard had the sense to get decent legal advice.]
It was just a bit jarring to see Harriet Vane expecting that their older son would still be bound to inherit everything at so late a date.
Especially since, as I read it, this was Peter's decision, not his father's - look up Busman's Honeymoon on Amazon and search inside for "entailed".
Other Wimsey inheritance weirdness: Lady Mary also had a large personal fortune, but "to avoid the awkwardness which emerges when a poor man marries a rich woman", she put it all into a trust for her kids when she married the poor but virtuous Inspector Parker.
32. Well then he was more of a jerk than I realised. Are you still allowed to make a strict settlement? If not, when was it banned?
32: Looks like. On advice from his lawyer, who apparently wasn't up to date on the tax implications.
Murbles isn't your man for an up to date view of inheritance law. See "Unnatural Death".
35. On second thoughts it might work, provided he is the initial beneficiary of his own trust, which I don't know if it's allowed, and if the trust is set up to devolve on the hypothetical child of a hypothetical child, which I don't know if that's allowed either*. Still a jerk move though, unless he had reason to suppose he would only have one kid.
*Indefinite trusts entailing property forever were ruled out of order at a fairly early date.
unless he had reason to suppose he would only have one kid.
And they end up having two, at least.
I think 14 and 16 are correct.
To 17, 19, and the OP summary, r>g leads to more inequality, as a theory, isn't really an empirical or historical argument -- it's necessary conclusion (because, so long as r outpaces g, the increase in the savings rate will naturally lead to more capital accumulation, and we know that the distribution of capital is massively more inegalitarian than the distribution of income). I don't think anyone is seriously questioning the issue of whether it is true that if r>g we will have more inequality. The issue is whether r (as properly defined, net or not of depreciation) in fact will or is likely to outpace g into the future, or whether the accumulation of capital will sufficiently depress r so that it doesn't exceed g and the problem Piketty identifies doesn't happen.
On Piketty's side, he's got the argument that there were very long periods of history -- in fact, it seems, most of the industrial era -- in which r in this relevant sense did exceed g without declining sufficiently due to excess accumulation of capital. At the very least, I think, he's shown that there is no automatic, built-in mechanism in modern capitalism that prevents this problem from arising.
I wonder if a website could be established that tracks r-g annually.
To 17, 19, and the OP summary, r>g leads to more inequality, as a theory, isn't really an empirical or historical argument -- it's necessary conclusion
But we are talking about an historical hypothesis backed up by a causal mechanism. I really don't want to say that it can be proven a priori. That just seems like the worst of freshwater economics. You're saying that history must fit a certain pattern established by an idealized story the economist has made up.
I found it a bit funny that I was reading a deeply unserious book (The River of No Return by Bee Ridgway)* that explained entail in more detail than either Austen or Piketty.
*It's 19th-century month in my book club.
You're saying that history must fit a certain pattern established by an idealized story the economist has made up.
No, I'm not -- I'm just saying that if you use a proper definition of "r", and you make the further assumption that capital is distributed in a much more inegalitarian manner than income (which is an empirical fact, but seems to be an absolutely universal one), then if r>g you will automatically get more inequality of wealth. That's definitional. If a higher return consistently accumulates to capital than income over time, and capital is distributed in a more inegalitarian manner than income, and you don't have some other countervailing force affecting the distribution of wealth, then you will have increasing inequality based on the accumulation of capital by capitalists.
Whether or not real history actually fits this pattern, or those conditions currently obtain (or will obtain, or given the nature of markets likely could obtain), or whether there are other things that Piketty hasn't identified that makes this mechanism unimportant, are absolutely empirical questions -- though Piketty has a very plausible and important empirical and historical argument that (for whatever reason) they did obtain for most of the history of the industrialized west.
But what you're not going to demonstrate empirically is that you have a consistent pattern of r>g on his definitions, and a much more inegalitarian distribution of capital vs income, and no other countervailing measures without increasing inequality in the (pretax) distribution of wealth -- the result is definitional. What Summers and other critics are saying isn't that Piketty is wrong granting his assumptions as true; what they are saying is that there are other features of capitalism that make it likely that r (as properly understood, and including depreciation) will decline such that it does not in fact exceed g for these purposes in the future and thus that there are self-correcting mechanisms that will solve the problem Piketty identifies, or that there are other mechanisms that make the thing Piketty identifies not important or not likely to be important. Piketty's arguments against these criticisms is largely empirical.
40. I'd be astonished if Piketty hadn't had a grad student build one already, but it might not go public just yet.
I wasn't really trying to capture Summer's criticism so much as trying to press one of my own, which may be based on holding Piketty to a higher standard than is reasonable.
My question was whether he had really shown that r>g was the cause of greater inequality in the period of time he was looking at. You can grant that in general x is a cause of y, without saying that it was the cause of a particular y.
In fact, it is possible to say that both x and y were present at a given time, and that x's generally cause y's, but in this case x was not the cause of y. For instance, smoking in bed causes fires. It may be the case that Bob was smoking in bed the night his house burned down, and yet the fire was the work of an arsonist, and not caused by Bob dozing off with a lit cigarette.
I am entirely open to the idea that I am just holding Piketty up to philosophical standards involving crazy counterexamples that don't really belong in economic history.
Piketty's arguments against these criticisms is largely empirical.
Well, Piketty is using historical reasoning as well as standard economic reasoning. If you challenged him on it, he'd say "Yes. And your point is?" I think he either said or strongly implied in the introduction that he had come to the conclusion that mainstream economic reasoning was an inadequate tool for understanding macro behaviour, and he anted to go beyond it. And he has.
My question was whether he had really shown that r>g was the cause of greater inequality in the period of time he was looking at.
I don't know that he's "proven" this, but he certainly has shown that it is very plausible, at least historically. Most importantly, what he does seem to have shown is that r did not in fact decline over time, and that the capital/output ratio has substantially varied over time.
37 &c: Turn a glass over for Sarah Caudwell.
This seems to have ground to a premature halt, so, LB, how do I get my thing on chapter 12 to you? What format? What email?
LizardBreath at unfogged and so on works, as does elizardb at hotmail (although I'm sort of phasing the latter out, so use the former preferentially). Formatwise, if you do any html coding you want to show up (links and such) then I don't have to, but plain text is fine. Emails from some people show up with line breaks that are a pain to remove -- if you know how to make that not happen, that'd be nice. If you don't, I've had to remove them for three out of four posts anyway, so don't worry about it.
It's ok - I prepared an emergency Piketty limerick
A French economist named Piketty
Contemplated inequality
"It's apparent, you see
That r's greater than g
So the FT can suck on my comprehensive two century multi-country dataset". Itty.
Well, Piketty is using historical reasoning as well as standard economic reasoning. If you challenged him on it, he'd say "Yes. And your point is?"
I don't think the our esteemed metal singer and IP lawyer meant this as a criticism. Everyone finds Piketty's empiricism a breath of fresh air.
Hiketty-Piketty
Modern economist:
"Balzac and Austen show
r exceeds g;
"Stand by for trouble if
Maldistribution
Returns to the level in
op. cit. (Dickens, C.)"
From floundering haplessly if half-arsedly through some of the linked papers in the technical appendix, like this one
http://piketty.pse.ens.fr/files/PikettyZucman2014HID.pdf
... It seems that Piketty does show the model works in theory, as well as holding it up next to the historical data and saying, pretty close eh? He also takes initial condition data, runs it through simulations, compares the output to history, and says pretty close eh?
The Kunkel LRB article linked a few times in comments points out that 70 years of r
56 ... r < g is actually quite a lot of industrial capitalism, for a historical argument.
Actually, that Dan Kervick site that bob mentioned earlier notes that the problem isn't automatic when r > g; it's when r is sufficiently larger than g that the rentiers' savings from r, after taking out consumption and taxes, is still greater than g. That's what leads to increasing inequality. So r a little bit bigger than g might be ok, but r significantly bigger wouldn't be.
I thought r was supposed to be real returns after taxes.
A couple of delayed comments from chapter 9, in the section on "Wage Scales and the Minimum Wage." He notes that "The United States used the minimum wage to increase lower-end wages in the 1950s and 1960s but abandoned this tool in the 1970s." That didn't fit with my recollection of the 70s. According to this chart, the US minimum wage went from $1.60/hr to $2.90/hr over the course of the 70s. That may not have fully kept up with the inflation that was also a hallmark of the 70s, but it doesn't look like a country abandoning the idea to me.
Also, a bit later on, he notes that "The payment of a monthly rather than a daily wage was a revolutionary innovation that gradually took hold in all the developed countries during the twentieth century." I wondered if the shift from hourly to salaried work in some industries is also part of a shift to a more predictable paycheck. That is less of a global trend, though, and is offset by part-time hourly employees who have seen less predictable hours.
59: I believe r is supposed to be real returns, net of depreciation. It is calculated from the total amounts paid in royalties, dividends, etc. Computing it net of taxes would be harder, because that would depend on the tax situations of the recipients (although you could make some assumptions).
So r a little bit bigger than g might be ok, but r significantly bigger wouldn't be.
In the technical appendix for chapter 10, he argues that wealth inequality is extremely sensitive to differences in r-g. Small changes in r-g can explain a lot of the variation in a coefficient "b," which appears in Pareto equations and can be used to measure inequality. These Pareto coefficients, furthermore, measure inequality in Pikitty's preferred way, because they talk about the share of wealth that a given percentile has in a given larger group.
60: See this inflation-adjusted chart for a clearer picture - it doesn't contradict what you say, but note that the highest the real minimum wage ever was was 1968. I assume he means "increase" in the sense of raise from previous levels, not simply maintain the effect.
59, 61: On page 25 he says "r stands for the average annual rate of return on capital, including profits, dividends, interests, rents and other income from capital, expressed as a percentage of its total value."
As I understanding it, depreciation doesn't apply to most of these forms of returns. It would apply to the value of companies, because it applies to the value of the machinery, etc. they own. But that would just be included in the calculation of their profit. It isn't a later subtraction from the gross returns of a nation.
As for taxes, the relevant tax is capital gains, which doesn't exist in all the countries he is considering. As I recall correctly, he runs the number in the US two times, both with and without capital gains.
OK, now I remember, there's a chart comparing r-g with and without taxes, and by that metric g did exceed r but only in the Trente.