tag:blogger.com,1999:blog-2761684730989137546.post2710123636520013330..comments2024-03-29T02:19:19.866-04:00Comments on Mike Norman Economics: Lars P Syll — Bob Rowthorn questions two of Piketty’s central assumptionsmike normanhttp://www.blogger.com/profile/03296006882513340747noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-2761684730989137546.post-20726985348524437772014-06-26T21:22:01.931-04:002014-06-26T21:22:01.931-04:00Thus, Piketty’s argument rests on two crucial assu...<i>Thus, Piketty’s argument rests on two crucial assumptions: β = K/Y and σ > 1.</i><br /><br />Piketty doesn't "assume" that β is equal to K/Y. That's just the definition of β. He also doesn't "assume" σ > 1. He attempts to measure the historical values of σ by combining the growth model framework with historical data about capital shares and capital-to-income ratios.<br /><br />Chapter 6 is the only chapter that contains any use of the CES growth model. Piketty, as I read him, relies on that model solely for "defensive purposes." His opponents have used that model in conjunction with empirically unfounded assumptions to argue that the rate of return on capital falls commensurately with rises in the capital-to-income ratio so as to keep the capital share constant. Piketty chief argument against this position is empirical: that's just not what the historical data shows. For anyone committed to the CES framework, the historical data yields the conclusion that elasiticities were slighly greater than 1 in the 19th century. But if one thinks all that elasticity stuff is BS, one is still left with the sheer historical data which seems enough to refute the balanced growth path conjecture. Piketty himself seems to be diffident. He is willing to work within the CES framework to interpret history in a way that accepts his opponents basic framework while showing why they are wrong. But in the online appeandix and lecture notes he says that the whole idea that one can measure the various ways in which capital can or cannot be substituted for labor with a single parameter σ is greatly oversimplified.Anonymousnoreply@blogger.com