1997-11-05 - Re: Protocols for Insurance to Maintain Privacy

Header Data

From: “John Kelsey” <kelsey@plnet.net>
To: “cypherpunks” <cypherpunks@Algebra.COM>
Message Hash: 4212eb1274b633515b5ad2953313be6e959c5f51b7bc2c8f412cd893d6c4d00f
Message ID: <199711052249.QAA06301@email.plnet.net>
Reply To: N/A
UTC Datetime: 1997-11-05 22:56:59 UTC
Raw Date: Thu, 6 Nov 1997 06:56:59 +0800

Raw message

From: "John Kelsey" <kelsey@plnet.net>
Date: Thu, 6 Nov 1997 06:56:59 +0800
To: "cypherpunks" <cypherpunks@Algebra.COM>
Subject: Re: Protocols for Insurance to Maintain Privacy
Message-ID: <199711052249.QAA06301@email.plnet.net>
MIME-Version: 1.0
Content-Type: text/plain



-----BEGIN PGP SIGNED MESSAGE-----

[ To: cypherpunks ## Date: 11/03/97 ##
  Subject: Re: Protocols for Insurance to Maintain Privacy ]

>Date: Mon, 3 Nov 1997 10:26:03 -0700
>To: "John Kelsey" <kelsey@plnet.net>,
>    "cypherpunks" <cypherpunks@Algebra.COM>
>From: Tim May <tcmay@got.net>
>Subject: Re: Protocols for Insurance to Maintain Privacy

>Such insurance is now common. A boat owner doesn't buy
>insurance for iceberg collisions if he is never in arctic
>waters, a small plane pilot doesn't buy cargo insurance if
>he doesn't ferry cargo, and so on.

Yes, I understand this.  Nobody with a grain of sense
insures himself for things he knows will never happen.

>>the point of having insurance, which is to protect yourself
>>from low probability high cost things happening.  That is,

>I have a different view of what insurance is than John does.

>What insurance is, and how it is priced, is too long a topic
>to get into here. Suffice it to say that the insurance
>company makes its profits by charging more for coverage than
>it pays out. And the customer, of course, tends to lose the
>differential.

An insurance broker and a bookie are in the same business.
A bookie takes a bet from me, at 1000 to 1 odds, for some
unlikely to occur event--which he thinks has a probability
of less than 1/1000.  An insurance broker does the same
thing.  The only important difference is my purpose in
placing the bet.  When I take out medical insurance, I am
placing a bet that I will get sick enough to spend more than
my deductable on doctors.  The insurance broker estimates
(based on whatever information is available) the probability
that this will happen, and makes me a bet at odds that are
in his favor, if his estimate is true.  If I get sick, I get
some money (or my doctors do).  If I stay well, I lose the
bet.  This is just paying someone to take some risk off my
hands.

>Each side tries to get as much information as possible. If
>Joe Client knows he never pilots a cargo plane, he doesn't
>opt for cargo insurance. If Joe Client knows he never
>engages in unprotected sex with diseasy prostitutes, etc.,
>he skips HIV insurance. The fact that some "low probability
>events," like meteor strikes, are uncovered is part of the
>price of keeping Joe's premiums tolerable.

>>before I've taken the test for genetic disease X, my best
>>estimate of the probability that I will test positive is
>>very low.  Once I have taken it, I know the result.  If I
>>sign up for a-la-carte insurance for this disease, the
>>insurance company effectively knows I must have tested
>>positive for a predisposition to it, and so either won't
>>give me insurance, or will give me insurance only at an
>>extremely high rate (corresponding to a 1/10 chance of
>>getting the disease, rather than a 1/1,000,000 chance).

>This is the idea. It causes those with the predilections to
>the disease to pay the high coverage costs.

Right.  I don't think I expressed my point too clearly,
because you seem to have responded to something different
than what I was saying.  Suppose there is some genetic
disease that kills its victims on their 31st birthday,
unless they get a $1,000,000 treatment first.  Before I have
taken the test for this disease, I have to accept a certain
risk--if I find out I have the disease, I have to raise a
million dollars in the next few months.  After I have taken
the test and gotten back the results, there's no more risk
involved (assuming the test is perfect)--I either have the
disease or I don't.  After the test, insurance isn't useful
(unless I defraud my insurer).  Before the test, though,
insurance might be useful--I could essentially place a bet
with someone that I had the disease--I pay $1, and get a
million dollars back if my test comes back positive--just
enough to pay for my treatment.

>The alternative is not pretty: banning private testing
>(how?) and forcing insurance companies to cover all
>applicants for all conditions at a fixed rate.

I know.  Let me make it clear that I am not at all
interested in banning private testing, coercing insurance
companies or anyone else into agreements they don't want to
make, etc.  I am saying it would be nice if I could buy
insurance against the results of the tests before I took
them.  The problem is, I can't see a really workable way to
do this, because there's no way to keep people from taking
the test beforehand.

>--Tim May

   --John Kelsey, Counterpane Systems, kelsey@counterpane.com
 PGP 2.6 fingerprint = 4FE2 F421 100F BB0A 03D1 FE06 A435 7E36

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBNGDdikHx57Ag8goBAQE6/AP/Yub9LA0b1rvivDeiHW9tHtV7b/n4jmUA
aP50hPFTCn7EEMsoKdC2pUv0eq5F+s8AmKyaE+SmxTgBbalrvm+T+lPib4bbEXjH
iI9nHbnfzTSxiVbkVHp0p0/0urIgokTcrFBWbfIEnmOnGvI7OHI+KcPjqr3JBQeo
A53+dL0NBRQ=
=kx2R
-----END PGP SIGNATURE-----

   --John Kelsey, Counterpane Systems, kelsey@counterpane.com
 PGP 2.6 fingerprint = 4FE2 F421 100F BB0A 03D1 FE06 A435 7E36






Thread