Hacker News Reader

Comment by yummyfajitas | Hacker News Reader

I love the classical "he said, she said".

The fact is, Piketty's numbers appear to be incorrect. It is no longer remotely close to clear that wealth inequality is going up, particularly at the top, at least from what we see in Piketty's book. The FT corrections show that inequality has not increased in the UK, and different weighting schemes (e.g. population weighting across countries) can completely make his effects vanish.

Krugman showed that income inequality, particularly below the 99'th percentile, is going up. Income (including capital income) != wealth.

Now maybe other sources prove him correct. If so, this response is far from convincing - why not just cite those other sources? And if that's the case, why did anyone care about his book to begin with?

In any case, Piketty is to be commended for releasing his data/methodology. Just think - if he didn't do this, it would have taken years for his errors to be discovered.

View on HN

Piketty's numbers appear incorrect

Er...some of his numbers appear incorrect. From this you incorrectly assume that this invalidates the conclusion. It may invalidate the conclusion if the errors are significant. However, what little I've seen so far does not seem to be all that significant.

From the fine article:

'Scott Winship, an economist at the right-leaning Manhattan Institute for Policy Research, and notable critic of Piketty's analysis, said the FT's allegations weren't "significant for the fundamental question of whether Piketty's thesis is right or not".'

So there seem to be highly qualified sources that say that the differences are not significant. That's not the same as "he said/she said".

Do you have any evidence that the errors are significant?

You also blatantly mischaracterize Krugman's point: he showed that inequality in capital income is going up, not (just) general income. Capital income has this property that it usually derives from capital (as the name says). So unless there is evidence that the top 1% or so systematically get significantly better returns on the same capital, an increase in capital income let's us include that there is an increase in base capital from which to derive that income.

On the other side of the coin, income leads to wealth, and unless you have evidence that high-income people save less of their income than lower-income people, then wealth will diverge. All the evidence I have seen so far (and common sense) say that it is the other way around: consumption saturates, so people with higher incomes actually save more and spend less, at least in relative terms.

So both sides of this equation quite trivially show the link between income inequality (especially for, but not limited to capital income) that you deny.

So both the FT criticism (= nit-picking in non-substantive details) and your criticism are unfounded, as far as we can tell.

Again, if you have some actual substantive evidence that the errors are significant, I'd love to see it.

On Scott Winship, see one of my top level comments (I made two top level comments as I only started investigating this after I posted once). Whatever was printed in the article, he has said on Twitter that he does doubt Piketty's conclusions, and also that the FT allegations are significant.

"...and also that the FT allegations are significant" (my emphasis).

Original Tweet:

"3/But UK--&, I now believe, US--methods need to be explained by #Piketty, else looks pretty bad. Just not CONSEQUENTIAL 4 overall thesis" (again, my emphasis)

https://twitter.com/swinshi/status/470880036233699328

"are significant" vs "not CONSEQUENTIAL" seem like exact opposites to me.

EDIT: missing parenthesis

you are misreading the context.

According to Winship, Piketty's results are wrong/irrelevant for other, independent reasons. But not everyone shares this view, so Winship is clarifying that, for people who were initially convinced by Piketty's analysis, the FT criticism is important. But it is not important for Winship because even without the FT criticism, he has independent reasons to discounting Piketty's analysis.

What the FT showed is that "It is no longer remotely close to clear that wealth inequality is going up...". This is because when they corrected Piketty's errors, and adopted different (but similarly plausible) weighting schemes, Piketty's major claim of inequality being "high, then low, then high" vanished in Europe.

If by "significant" you mean something else, please say so.

So unless there is evidence that the top 1% or so systematically get significantly better returns on the same capital, an increase in capital income let's us include that there is an increase in base capital from which to derive that income.

You are simply wrong on this: https://news.ycombinator.com/item?id=7795331

I'm not sure why you describe me as "blatantly mischaracterizeing" Krugman since I explicitly noted it was capital income.

...unless you have evidence that high-income people save less of their income than lower-income people, then wealth will diverge.

Again, this is incorrect. They might achieve a lower rate of return (including being more vulnerable to large shocks).

In order for income inequality to not lead to wealth inequality, we need to believe that poor people save more of their income than wealthy people - which seems rather difficult.

I think it is clear from other sources that wealth inequality is going up, and the FT is pushing its findings too far.

>In order for income inequality to not lead to wealth inequality, we need to believe that poor people save more of their income than wealthy people - which seems rather difficult.

Yes, but that's actually what the FT refutes (center of 3rd § of the article):

> The newspaper concluded there was little evidence in Piketty's original sources to bear out his theory that the richest were accumulating more wealth

In order for income inequality to not lead to wealth inequality, we need to believe that poor people save more of their income than wealthy people - which seems rather difficult.

Or that wealthy people achieve diminishing returns on investment as their wealth grows. I.e., my startup might grow 10-fold, but Facebook is unlikely to.

Even if you assume Piketty's "r > g" macro-level hypothesis, there are a lot of microeconomic theories completely compatible with it. Not all of them result in increasing inequality.

http://www.chrisstucchio.com/blog/2014/piketty_and_inequalit...

(But now we are moving into the topic of Piketty's theoretical weaknesses, which are supposedly irrelevant because his data is so great.)

I think it is clear from other sources...

[citation needed]

And if other sources prove this point, why was Piketty's book such a sensation?

>I.e., my startup might grow 10-fold, but Facebook is unlikely to.

Yes, startups, an investment famous for attracting only small scale investors of small means…

What is Piketty's theoretical weakness?

See this article, especially toward the end on his understanding of capital: https://mises.org/daily/6736/Thomas-Piketty-on-Inequality-an...

The main criticism there is that the concept of return on capital is meaningless. I think that is fringe view and I don't find it compelling.

>I think that is fringe view and I don't find it compelling.

That's a vacuous non-argument and appeal to popular opinion. Not compelling as well.

Many complex systems, hard to describe on a micro level, have straightforward macro statistics. I think return on capital can be described but such statistics even if individual investments are highly variable. Capturing the return is achieved by having a highly diversified portfolio.

His lack of a clearly stated theory. Ask yourself this question: in a few equations, what is his fundamental claim? What specific data (irrespective of technical details about collecting it) would prove or disprove his theory?

The closest he comes to having a clearly stated theory is that "r > g => growth of inequality". That specific claim is completely wrong.

http://www.chrisstucchio.com/blog/2014/piketty_and_inequalit...

He has a clearly stated theory which can be judged independently of his empirical measurements and is not on its face "completely wrong". I can't understand how you have spent so much time virulently criticizing Piketty and so little time doing even the most basic research of his work, going so far as to ignore material that nullifies your (imagined) criticism and which has been explicitly pointed out to you.

Had you simply spent five minutes reading documents that were helpfully found for you then you would know exactly what is his model.

Didn't the fact that you thought you could refute an recognized expert in the field's work in a few minutes with the most basic criticism give you pause?

I read his book. He talks about Balzac and Jane Austen a lot and devotes a small amount of time handwaving about the micro-scale of inequality. No equations (beyond a few accounting identities), representative agents, or anything else you'd expect to see.

Of course, if you want to include a link to someplace where it was "explicitly pointed out"...I'm not holding my breath.

And since you seem to disagree with the claims I made in my blog post, what claim do you disagree with? What's the specific error I made beyond doubting the Great Piketty?

Here is where (a PhD economist) "explicitly pointed out" references to you after you derisively made the exact same demand you do now. https://news.ycombinator.com/item?id=7658808#up_7661247 Section 5 of HID especially.

As that economist said models and equations are not included in the book presumably because it would make it inaccessible to many people. On the other hand I don't know how someone so interested in those aspects not only spent essentially no effort in finding them but ignored references they themselves requested and were provided with.

As for the blog post what I disagree with is that it represents Piketty's claim, whether that claim is true or not. It doesn't.

Your tirade against Piketty is completely unfounded, bizarrely rabid and seemingly prejudiced.

Those papers are more of the same - no theory is clearly stated, just a few accounting identities that describe wealth/income ratios.

As for the blog post, I could be wrong. Does Piketty not claim that r > g is the cause of inequality? If so, why does Piketty have a subsection called "The Fundamental Force for Divergence: r > g" in his book?

In any case, confusion like this is why it's so important to clearly write down your main theory at the very beginning. If Piketty did this, you could simply point at a set of equations and show how my counterexamples don't disprove them.

I don't know why you are getting hostile or accusing me of being "rabid". Kind of goes against the spirit of careful intellectual discussion. Makes me think you have an emotional attachment to Piketty being right.

> I don't know why you are getting hostile or accusing me of being "rabid". Kind of goes against the spirit of careful intellectual discussion.

Initially, I felt similarly about the general 'tone' of your statements. Then I browsed through your comment history. Here are the most 'rabid' statements I could find (emphasis mine):

> What's the specific error I made beyond doubting the Great Piketty?

> I imagine Piketty would consider that a problem if he knew India existed.

followed, to be fair, by pointing out at least twice:

> It's an attempt at poetic exaggeration

To my surprise, you've been acting more in 'the spirit of careful intellectual discussion' than I initially thought, and I suspect it was the surrounding comments that agreed with yours that 'polluted' my view (as well as a slight personal bias that I have).

Perhaps it's good to completely avoid poetic exaggeration, or to maybe consider the context of you comments and adjust to make your comments a bit more palatable to the many who strongly disagree.

Or perhaps you're doing fine as it is, and your 'karma' attests to that.

Either way, I figured I'd post my observations for those who might similarly, and I believe erroneously, feel that you are being 'rabid'.

I'd also like to thank you for your informative comments. I'm following all discussions on Pikkety with great interest, and I'm trying to stay aware of my bias of wanting his results to be correct.

(and my apologies for not contributing something more substantial at this point in time)

It is pretty brazen to say no theory exists when you were directed not only to the paper but the section where that theory is described.

Since you can't be bothered to read it, section 5 of Piketty's HID paper describes his model of long-run wealth distribution, assuming a Cobb-Douglas utility function and random multiplicative shocks to an individual's savings rate, as a Pareto distribution where the tail concentration is a function of r-g.

I am spoon feeding you at this point and I won't do it anymore.

Actually you are right - I missed it the first time. My mistake. Based on a quick skim, it looks like Piketty is assuming far more than r > g (which is of course necessary), but there does appear to be some sort of distributional theory.

Based on a first skim (I'll take a more detailed look later), inequality appears to be a function of (r - g). I.e., r > g does not imply increasing inequality, the magnitude of the inequality merely determines the level of inequality. Seems to disagree a bit with the newspaper summary of Piketty, but I'll have to read more carefully to be sure.

Reply on news.ycombinator.com