Sunday, May 1, 2011

Killing people and regulating OTC forex swaps

France Inter had a two-part podcast on the death of the Pope John-Paul I, who was a pope for a total of 30 days. One of the commentators was comparing his death to the death of Pius XI, because both of them died the day before historical speeches: John Paul I was going to announce some reforms to alter the potential connections between the Vatican and the mafia while Pius XI was going to attack fascism and racism. This was one of the argument for conspiracy theories: those two deaths have been officially considered to be natural deaths, but a lot of conspiracy theories are floating around.

Likewise, in their editorial on the Obama's administration lobbying to exclude foreign currency swaps for over-the-counter regulation in the Dodd-Frank bill, the NYT made it clear that the administration is deliberately trying to do it under the counter:
In an announcement on Friday afternoon — the time slot favored by officials eager to avoid scrutiny(...)
Now, here is the link between the two stories. If I want to kill a guy because he's going to make a speech that I don't like, killing him on the day before is suspicious. So I should probably do it earlier, and the coincidence in the popes' story is meaningless.  However, if the police is seeing it this way, I actually have an incentive to kill the guy the day before the speech, since the police would say "meh. If it had been intentional, the guy would never have killed him/her the day before". Then again, the police could take this into account, etc... This is where a model would be useful: what day should I kill the pope(and how that day is informative for the police).

Same thing for the administration's announcement. If everybody knows that fancy stuff is going to happen on Friday afternoons, how comes it is still used? Journalists should be ready on Friday afternoons, so that it is not optimal for the administration to make announcement at this time...

On both counts, the administration's and the killer's best strategy seems to be complete randomization so as not to convey any information. 

No comments:

Post a Comment