Re: poly: Pondering Privacy

Hal had asked:
>Could you elaborate on this? What is the basis for the conclusion that
>most secrecy could be eliminated and leave both sides better off?

It's easier to explain this case by case in the context of each of the
situations where it seems to be true. And it would take us quite some
time to go through all the cases.

Here's an example: insurance. The possibility of adverse selection
happens because you might know things about your risk that insurance
companies don't. Adverse selection can result in all insurees buying
less insurance than they otherwise would, which hurts them all without
helping insurance companies.

If you insist on a general intuition, try this. When there are secrets,
the need for "incentive compatibility" limits the feasible institutions,
and hence the feasible outcomes. If you want outcomes to be correlated
with the secret info people possess (and you usually do), you have to
make it in people's interest to tell you what they know. Without secrets,
this constraint doesn't apply.

Another angle: A "better" situtation is one which everyone would prefer.
When there are secrets, there are many different versions of each person,
each version knowing a different secret value. With secrets, it is
harder to find a better situation because you have to please not just all
the people, but all the versions of all people.

Another angle: Consider a (perhaps hypothetical) point in time before
people got their secret info. If they can make deals at this point in time,
they shouldn't pick a certain deal if there is some other deal which makes
everyone better off. The remaining deals are called the "Pareto frontier".
If people can't or don't make deals then, but instead wait and make deals
after they get their secret info, however, things can only get worse.
In fact, all feasible later deals can be behind the original Pareto frontier.

Robin Hanson
hanson@econ.berkeley.edu http://hanson.berkeley.edu/
RWJF Health Policy Scholar, Sch. of Public Health 510-643-1884
140 Warren Hall, UC Berkeley, CA 94720-7360 FAX: 510-643-8614
Received on Wed May 20 22:40:46 1998