The beginning of April brings with it, among other things, a new batch of NBER papers, and here is one that is mildly amusing but probably not an April Fool's prank, although it is called "Tailspotting: How Disclosure, Stock Prices and Volatility Change When CEOs Fly to Their Vacation Homes," so, y'know, maybe. Anyway it's by Stern professor David Yermack and ... if it is fake, it's still probably correct:
Companies disclose favorable news just before CEOs leave for vacation and delay subsequent announcements until CEOs return, releasing news at an unusually high rate on the CEO’s first day back. When CEOs are away, companies announce less news than usual and stock prices exhibit sharply lower volatility. Volatility increases immediately when CEOs return to work. CEOs spend fewer days out of the office when their ownership is high and when the weather at their vacation homes is cold or rainy.
Except their ski chalets? I don't know. Also this sentence should win some sort of award: "However, a bivariate probit model presented below indicates that news disclosures appear to be linked to CEOs’ vacations even after using weather variables to control for endogeneity of the vacation schedule."*
The basic result is not that surprising; if I learned one thing from this paper it's less "CEO activities are an important part of the short-term news relating to a stock" and more "it is possible to use vacation home property records and FAA databases to find out when CEOs visit their vacation homes, and that that is what your Stern tuition is paying for, which, okay!" Still here are two passages to maybe think about:
Since the 1930s U.S. authorities have established detailed ground rules for the timing of company disclosures by enacting rules such as Regulation FD and the Sarbanes-Oxley Act (SOX); since it became effective in 2004, SOX has required companies to disclose a wide range of material events on Form 8-K within either two or four business days. Notwithstanding these regulations, my results strongly suggest that companies coordinate public news disclosures with the personal schedules of their CEOs. In particular, companies appear to empty their queues of news announcements just before CEOs leave for vacation, and then delay subsequent disclosures until CEOs are back in the office.
I'll just interject that - this doesn't necessarily mean they're doing Bad Illegal Things. Like, the importance of events is not always obvious at once, earnings are firmed up over the days after the quarter closes, merger negotiations get more and more serious, whatever, and you make a materiality judgment that is maybe just a little bit colored by the boss's tee time. There's no particular reason to be troubled by the possibility that companies adjust the timing of announcements to fit vacation schedules. But also:
Regardless of the direction of causation, the movement of a company’s aircraft to and from a CEO’s vacation residence provides a very visible signal of pending news announcements and silences. With a trivial amount of research and monitoring, investors could observe flights of corporate aircraft in real time between the headquarters airport and a CEO’s vacation locale, either by monitoring live FAA data on the Internet or stationing scouts for “tailspotting” of tail numbers of planes that land at leisure airports favored by CEOs such as Nantucket, MA, or Naples, FL. This information could support straightforward trading strategies, such as using derivatives to bet on declines in volatility when a CEO arrives at his vacation airport and increases in volatility when he departs.
April Fool's! Ha ha. No he really wrote that and it is I guess true: overall average volatility of the stocks in his sample is about 45% when the CEO is at work, but 38% when he or she is on a long (5+ day) vacation, so, hey, there's some chance that 7 vols is outside of the bid-ask spread and you can monetize that difference over five days and ... I guess? Also there's some predictive power for returns: there are 16-17bps of excess returns on average in the three days before and after a vacation.**
Anyway, so Rajat Gupta is maybe going to jail because he and/or someone else maybe tipped Raj Rajaratnam about Goldman and P&G earnings before they were publicly announced. Because those earnings numbers were - I guess everyone agrees - "material nonpublic information," meaning that (1) they weren't publicly disclosed and (2) they were likely to move the stock in some amorphous way. (The standard is "whether a reasonable investor would have viewed the undisclosed information as having significantly altered the total mix of information made available.")
A question you could ask is, what if Rajat had called Raj and said "hey, Lloyd is going on a long-overdue two-weeker starting next Monday" - something that is (apparently) statistically correlated with higher returns this week and lower volatility over the following two weeks? And suppose Raj had traded on that? Unfair advantage? Uneven playing field? Material nonpublic information?
Meh. But you might conclude that our disclosure and insider trading rules came from a simpler time - a time where long-term investors had a total mix of information that included basic operational and financial data and, like, mergers, but data mining for short-term performance signals with 16-17bps of alpha and then sending an observer to watch airplane tail numbers to capture that alpha was not a ... well, where none of those words had any meaning. (Thus the "mosaic theory," which makes no sense, but was arguably a thing at a time when 16bps of alpha was not "16bps of alpha" but "part of the mosaic.") So you have companies that - maybe! - move around earnings and other big-ticket announcements to accomodate CEO vacation schedules, while still probably complying with disclosure rules. But you also have proof - maybe! regressions, anyway! - that investors could generate "unfair" "edge" with nonpublic information that seems at first glance - and, probably, to a court - to be trivial.
* TBF this is a real thing. You could have two models of how CEOs operate: either they structure their lives around their companies and hop on the plane to Key Largo only when they've got a quiet moment at work, or they structure their companies around their lives and put off press releases / mergers / bankruptcy filings depending on the whims of the ballroom dancing competition calendar. And the paper is I guess weakly supportive of the second, rather less flattering view; my intuition would go the other way but take that bivariate probit model for what it's worth.
* Maybe because:
companies announce good news just before the CEO leaves for a long trip, then announce very little while he is gone, and finally announce more good news on his return. ... Bad news announcements do not seem to occur in proximity to the CEO’s longer vacations. Since many CEOs begin or end vacations in the first month of the year, these data may have a plausible connection to the well-known “January effect” of stocks performing unusually well in the first weeks of a new year.