That distorted graph is pretty explainable by the fact that he has only 6 data-points, and that by default Excel's graphs just treated those dates as labels (instead of a time series).
And yet the chart is pretty clearly presented as a time-series. I think "well, he just used Excel defaults..." is a better argument for the prosecution than the defense.
Great project. I'm only about 15% into the book according to my Kindle, but it's interesting enough that I'm definitely going to check out the status of this project when I finish it. (could be a couple of months at the rate I'm going now)
The validity of the numbers depends on more than just the maths though, do you guys plan on going through the sources as well?
I'm also only about a quarter through, but so far it's interesting for non-economists. Some tolerance of technical explanations and simple formulas (i.e. not panicking at greek symbols) is required, but nothing unusual for the HN crowd.
This book has generated a lot of discussion, and so I went into it with my controversy-goggles strapped on. This is absolutely not required, as the book (based on what the author says he will cover and has covered so far, anyways) is more of a history and description of capital than political narrative or policy-making.
I agree that the tone of the first chapters is rather neutral and objective or at least not as controversial as expected. I think the author's recommendations which by definition are subjective are to be found in the last chapter which I have not read yet.
The website I have linked to was actually referenced in the book (if you had just only taken the time to read it ...). SO I thought, it was actually something the author of the post should consider when working on the dataset from the book.
Still I find it weird to be downvoted for a suggestion. :/
I think the book and the data is very interesting. But, I'm curious - why the effort? Isn't his book, data, and reputation as an economist / historian vetted by other experts in his field? Certainly they'd be the best people to say whether his methodology and conclusions are legit or not.
That is, do we really care about what the FT says? To me it's similar to the Washington Times making a statement about a book on evolution.
While not everything can be reduced to incontrovertible math-like proofs, I think we -- data scientists, statisticians, software engineers -- could build a world with a lot fewer legitimate arguments.
Imagine a Wikipedia of science, logic, and reasoning. Instead of summarizing consensus, you instead build a vast database of primary facts, data, and then on top of it formalize your assumptions and arguments so that it is plain to see how conclusions follow -- or don't -- from the bare facts and reasonable assumptions. You could build almost math-like proofs, with all arguments and counterarguments categorized and catalogued.
You could build a verifiable consensus. Given two different schools of thought, for instance, two different schools of economics, you could reduce each to the set of assumptions or faulty logic that distinguishes them. Some economists believe one thing to be true, some another, and they argue with one another and some people are convinced and some people aren't. Where does the difference arise? Is there a faulty chain of reasoning underlying one (or both!) of them, or is there a difference in the assumptions each group makes? If so, is one of these assumptions unfounded, or can it be tested and verified so that the assumptions that go into each school can be reconciled along with the conclusions that follow from them?
I think we have an opportunity, not to be smarter or more informed than the experts in their different fields, but to have each argument once, online, crowd-sourced, rather than having the same arguments crop up again and again and again between individual experts arguing with each other. We can bring organization and permanence.
I think you are in a blind alley. Yes, there is a singular epistemological error underlying every flavor of mathematical, scientistic economics. There is no actual "equilibrium point" where some say and there's no "coefficient of human volition" that we can calculate decisions like we could calculate the path of an atom.
After about 100 years of trying economics this way, there is still nothing close to a model that has reliable predictive power as we would expect with other natural sciences using the hypothetico-deductive method. I advocate instead the axiomatic-deductive method of the Austrian School. If you like, here is more on why what you want to do doesn't work from von Mises: http://mises.org/daily/3540
If your fine with groupthink, personality driven research and science as a priesthood, then sure, why question. If you believe Piketty has written a largely rhetorical position paper, then sure, why question. However if Piketty is endeavoring to to do evidence-based formal social science questioning is not only appropriate, it should be welcomed.
Is it really that cute? For example, I get the WSJ delivered daily because I value it's top-notch reporting and journalism. But, I usually skip its editorial pages: their editorials are written by people with strong ideologies and lots of rhetoric - things not very interesting to me. Similarly, FT is probably a top-notch news paper, but should I consider their editorial staff's opinions any more than the WSJ's editorial writers'?
I guess I have a bias against editorialists and a bias toward scientists, researchers, historians.
Furthermore, I am more surprised that more editorialists haven't slammed the book. I mean, given the book's subject and conclusions, I would have expected an avalanche of criticism, by now.
