Market rationality isn’t. The fact that if you made an index fund of only female-lead companies it would beat the pants out of the market has been been known for a really long time and still hasn’t been arbitraged away. 30 years or so of traders just leaving money on the sidewalk because of testosterone poisoning.
I don’t believe that, and you lay out exactly why one should not believe this claim for an instant: you seriously think that in the the $2.4 trillion+ hedge fund industry—stuffed full of the smartest hungriest slimiest most ambitious money-hungry people, men and women who would sell their own grandfather if that would provide collateral for a juicy short, who would encourage their employees to break the law and throw them to the wolves if they get caught, who are worse friends than sharks because at least sharks’ bellies can get full—that this entire industry would uniformly pass up almost doubling their return through a dead-simple legal strategy which would be discovered by their machine-learning algorithms even if they were blind to it—out of sexism? (How many Wall Street traders even know the gender of the CEOs whose associated hieroglyphics flash across their screens?) I have to say, you seem to have a much higher opinion of the moral principles of Wall Street than I do.
Having established that you are making an extraordinary claim which requires extraordinary evidence, let’s take a look at your evidence.
A link to a piece whose opening centerpiece is link to an informal analysis (‘Source: interactive data’) in Fortune magazine in July 2014, which mention that there are now 27 female CEOs in the Fortune 1000 and that ‘during their tenure’ they had returns of 103% vs 70%. Problems with your claim I can spot just from the Fortune writeup (although calling an infographic a writeup is a bit generous):
no indicating of volatility in total returns (how big is 30% anyway?) or other forms of risk-adjustment
small sample (27?)
no backtesting or cross-validation or out-of-sample tests
clear researcher degrees of freedom: why CEOs, why not Chairwomens or CEO+Chairwoman or either CEO or Chairwoman, or all varieties (one might expect an interaction effect when both positions are female or occupied by the same person)? Why the Fortune 1000, exactly? (The Fortune 500 is more prestigious, and if they wanted a larger sample, Factset should’ve been able to provide them a much bigger set of corporations to look at.)
unspecified data and time periods
other datamining: the stats they present look kind of random
nothing about issues of survivorship: companies move in and out of the Fortune 1000 regularly, and other work establishes women are more likely to become CEO after companies experience turmoil—so if the failed firms are excluded from the analysis, the selected firms will tend to experience high returns (as by definition they pass the crisis which depressed share prices) and also tend to be disproportionately female CEO-chaired
Academic papers regularly try to find and show violations of EMH, but the more careful a paper is, the smaller the violations become, so they typically find only small ones and are often still false positives due to any one of the reasons I give above and there are far more ways to go wrong than that. A full-blown paper which takes countermeasures against all the problems I mention may have begun to earn some reasonable probability of being correct. It’s a hard topic with many traps for the unwary, and some listoids or graphicles isn’t going to cut the mustard. One can safely predict that any research showing excess returns to female CEOs will either turn in meaninglessly small effects which could be due to minor methodological issues or the effect will quickly shrink to zero when tested out of sample and especially after the paper is published… (I particularly like the bogus results caused by the database company providing the data retroactively editing the database to remove low-performers. Which is relevant here, now that I think about it.)
Well, consider where the industrial revolution took of, and what was special about that time and place. It wasnt coal or literacy of technical expertise. China had that in abundance for thousands of years. It was all that and high wages
And why could English companies pay so much to workers? Because of high productivity. Maybe you should go reread Clark’s papers. Not that one can attribute the IR to simply ‘high wages’, which is a consequence, not a cause...
I don’t believe that, and you lay out exactly why one should not believe this claim for an instant: you seriously think that in the the $2.4 trillion+ hedge fund industry—stuffed full of the smartest hungriest slimiest most ambitious money-hungry people, men and women who would sell their own grandfather if that would provide collateral for a juicy short, who would encourage their employees to break the law and throw them to the wolves if they get caught, who are worse friends than sharks because at least sharks’ bellies can get full—that this entire industry would uniformly pass up almost doubling their return through a dead-simple legal strategy which would be discovered by their machine-learning algorithms even if they were blind to it—out of sexism? (How many Wall Street traders even know the gender of the CEOs whose associated hieroglyphics flash across their screens?) I have to say, you seem to have a much higher opinion of the moral principles of Wall Street than I do.
Having established that you are making an extraordinary claim which requires extraordinary evidence, let’s take a look at your evidence.
A link to a piece whose opening centerpiece is link to an informal analysis (‘Source: interactive data’) in Fortune magazine in July 2014, which mention that there are now 27 female CEOs in the Fortune 1000 and that ‘during their tenure’ they had returns of 103% vs 70%. Problems with your claim I can spot just from the Fortune writeup (although calling an infographic a writeup is a bit generous):
no indicating of volatility in total returns (how big is 30% anyway?) or other forms of risk-adjustment
small sample (27?)
no backtesting or cross-validation or out-of-sample tests
clear researcher degrees of freedom: why CEOs, why not Chairwomens or CEO+Chairwoman or either CEO or Chairwoman, or all varieties (one might expect an interaction effect when both positions are female or occupied by the same person)? Why the Fortune 1000, exactly? (The Fortune 500 is more prestigious, and if they wanted a larger sample, Factset should’ve been able to provide them a much bigger set of corporations to look at.)
unspecified data and time periods
other datamining: the stats they present look kind of random
nothing about issues of survivorship: companies move in and out of the Fortune 1000 regularly, and other work establishes women are more likely to become CEO after companies experience turmoil—so if the failed firms are excluded from the analysis, the selected firms will tend to experience high returns (as by definition they pass the crisis which depressed share prices) and also tend to be disproportionately female CEO-chaired
Academic papers regularly try to find and show violations of EMH, but the more careful a paper is, the smaller the violations become, so they typically find only small ones and are often still false positives due to any one of the reasons I give above and there are far more ways to go wrong than that. A full-blown paper which takes countermeasures against all the problems I mention may have begun to earn some reasonable probability of being correct. It’s a hard topic with many traps for the unwary, and some listoids or graphicles isn’t going to cut the mustard. One can safely predict that any research showing excess returns to female CEOs will either turn in meaninglessly small effects which could be due to minor methodological issues or the effect will quickly shrink to zero when tested out of sample and especially after the paper is published… (I particularly like the bogus results caused by the database company providing the data retroactively editing the database to remove low-performers. Which is relevant here, now that I think about it.)
And why could English companies pay so much to workers? Because of high productivity. Maybe you should go reread Clark’s papers. Not that one can attribute the IR to simply ‘high wages’, which is a consequence, not a cause...