[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: A problem with anonymity


Tim May said in article <[email protected][]>:
> At 11:32 PM 8/31/95, Scott Brickner wrote:
> >he sells its assets to his own identity at a fraction of their worth,
> >and defaults on the liabilities.
>      ^^^^^^^^^^^^^^^^^^^^^^^^^^^
> ideally, one never "trusts" an agent with a transaction greater than the
> value of the reputation capital he will lose if he defaults.
> Whether cryptographic protocols (cf. the "encrypted open books" proposal by
> eric Hughes for one approach which may be useful) solve this problem is not
> known at this time. But the non-crypto world has of course not solved this
> problem, either.

I've often thought that in a system of digital pseudonyms, where no-one need
trade with a negative reputation (a reputation deficit?), something like
Akerlof's Market for Lemons will arise, and _all_ pseudonyms will be treated
as (reputationally) worthless.

[Akerlof, if I remember my economics right (and I am confident that I will
be corrected if I don't) analysed a market for used cars. There were two
types of cars: good ones, and lemons. A purchaser couldn't tell the difference
until she had bought the car. Since the expected value of a used car was less
than the value of a good car, purchasers wouldn't pay the good car price.
But that would mean owners of good cars wouldn't offer them for sale (in
this market). So the only cars for sale would be lemons :-)]

As Tim points out, this is a non-crypto problem as well, and devices such as
bonds or (which are game-theoretically similar) expensive advertising or
plush premises [if they spent an unrefundable $20million on the Rolling
Stones, they're not likely to throw it all away by ripping you off for
$100 ;-)] are used to convince potential customers of one's bona fides.

How these transfer to the world of cyber-finance, I'm not sure, but I suspect
it leaves a role for True Names in the management of credit risk: as escrow
agents, middlemen, clearing houses etc. [Although, having said that, if the
Akerlof analysis applies, you just *can't* grant (unsecured) credit to
pseudonyms - the percentage of defaulters will be 100...] But these Names are
True only in the sense that they are juridically persistent (that is, if
they transact today, they can be sued tomorrow), and need not be traceable
to any True People (Warm Bodies?) - anonymously held corporations, for

If you can't rely on the unsecured promise of a digital pseudonym, and you
can't accept reputation as 'security', how do you extend credit?


Version: 2.6.2