From: tcmay@netcom.com (Timothy C. May)
Date: Wed, 21 Sep 94 16:23:55 PDT
To: cypherpunks@toad.com
Subject: Re: THOSE DARNED HIT MEN
Message-ID: <aaa672850202100411bd@DialupEudora>
MIME-Version: 1.0
Content-Type: text/plain


[I'm back to getting mail late, sometimes by days, as Netcom's overloaded
servers bounce incoming mail, which means it gets resent some time later.
For example, as I write this, I've seen Sandy's "HIT MEN" response to David
Merriman, but not Merriman's post. This straggly situation puts me at a
disadvantage, but I shall strive to overcome it.]

Sandy Sandfort writes:

>I don't buy it.  Anonymous digital assassins, murder escrows and
>all that work fine in an artificial, abstract, game-theory world.
>I doubt we'll see much--if any--of it in the Real World.  Why?
>Because the market is too richly textured not to come up with its
>own cultural, ethical and digital fixes.

Of course the real world will evolve complex, richly-textured constructs.
We all know that. I never presented a claim that the exact, and simple,
behavior could be predicted. Markets just don't work that way.

>Would you do business with a escrow that was the bag man for
>contract murders?  I wouldn't, you wouldn't and the vast majority
>of people in the world wouldn't.  That wouldn't stop some escrows
>from performing that service, but it would run up the cost.  When
>the costs rise high enough, profitable opportunities are created
>for false escrows to enter the business.  That's just one problem
>that can interfere with such an odious endeavor; there are more.

I can't follow Sandy's logic here. That I wouldn't use such services, that
Sandy wouldn't, etc., is hardly persuasive. Contract killings happen today,
after all. Sandy says this "would run up the cost." But from what basis?
I've made no predictions about the costs, either with or without the
participation in such markets by Sandy or me!

What the costs will be is unknown to me, and I don't plan to try to
forecast the costs. All I claim is that anonymous escrow services "solve"
the specific problem raised earlier about one or more of the parties
welshing on the contract. A kind of 'clearing' mechanism.

In any case, there are in fact "escrow agents" today for contract murders.
Mob families act in this way, putting the "full faith and credit" of their
organizations behind such hits. (I'm of course not saying that welshing
never occurs, that snags never develop, etc. Like any market, imperfections
exist.)

The mob families are not cryptographically pseudonymous, naturally, but to
the extent the code of omerta applies, the internal transactions and
discussions are cut off from outside observation.

Reputations matter. If it becomes known that Frankie the Lip took money and
didn't make the hit he contracted to do, he'll not get many more jobs (and
his Don may send him to sleep with the fishes, for undermining the market
value of his own rep). And so on.

I won't belabor the point about how organized crime works, except to say
that contracts are routinely enforced by a mixture of things, with
reputation an important constituent. Yes. the threat of ultimate violence
is paramound, and this is of course lacking in the crypto case.

So we have to examine areas where only "reputation" matters. I've done this
in earlier posts on this very topic.

>I think the best way to illustrate that the anonymous murder
>business is nothing more than a bugaboo, is to set the best minds
>on the planet--Cypherpunks--to work on the problem.  Let's all
>put our thinking caps on, and come up with answers to the follow
>hypothetical situation:
>
>Let us assume a world with totally anonymous communications and
>payments (strong crypto, remailers, digital cash, etc.).
>
>1)  How would YOU scam money from the system without actually
>    knocking anyone off?  Or in the alternative,
>
>2)  How would YOU use technology to address the problem from
>    police/private investigator perspective?
>
>I'm betting that with no more than a few moments of thought,
>Cypherpunks will come up with a ton of hacks.  To get the ball
>rolling, here are two from me:
>
>SCENARIO ONE
>
>I set up a meat-and-potatoes escrow business.  I keep my nose
>clean.  I honor my obligations.  I build up a good reputation.
>At some point, I'll be approached by a murderer and the person
>who is hiring him or her.  I'll accept the payment.  When the
>murder is committed, I won't pay off.  The murderer will (a) sue
>me (I don't think so), (b) damage my reputation (I'll leave this
>one as an exercise for the student), or (c) murder *me* (ah, but
>first he has to find me; in the world we posit, that won't be

(a) The party to the escrow transaction posts a transcript of the
communications from the escrow agent, including his digitally signed
statements, and produces proof that he upheld his end of the bargain.

He then says: "Al's Anonymous Escrow" announced they were holding money for
this job, as you can see. I did the job, as you can see from the digitized
images I took at the scene, and now Al won't pay up. I call him a liar. I
plan to move my business to "Murder, Incorporated," which seems to have a
much better attitude."

Al cannot deny that the escrow arrangement was made, due to the digital
signatures (all handled via anonymous pools or similarly untraceable means,
it should be emphasized, despite the obviousness). Al can of course claim
that the hit was not made, that the presenter of the evidence was not the
actual hitter, etc. (if the party to the signed transaction is also the
presenter of the digitized image of the murder scene, for example, that
would be mighty compelling evidence that the party was in fact centrally
involved).

(b) damage to reputation. Sandy leaves it as "exercise for the student,"
but I think the point I just made shows that fully pseudonymous agents can
still present evidence to the court of public opinion and have their
reputations influenced positively or negatively. Reputations will still
matter.

(Again, nothing in my arguments presumes to speak to what the market costs
will be, how long it will take reputations to evolve, etc. I have some
ideas, but won't make them right now.)

(c) retaliating physically against the escrow agent. This is straw man, as
we all know.

>very easy).  If this scenario happens very often, it'll take all
>the profit out of the murder business.

If an escrow agent does this very often, he'll lose all his business.
Opportunity for an agent who takes his reputation more seriously to then
gain market share.

Just as with Swiss banks who can claim an account was closed by the
customer. Since signatures are so easy to forge, relatively speaking, this
ought to happen a lot, right? Of course, it happens almost not at all (so
far as I've ever heard), because of the points about reputations, future
business, etc. Lots of points here, and I'm not planning to get into a
massive discussion of why and how illegal gambling (bookies, for example)
works this way.

(I'll just make the aside that Sandy's arguments apply to bookies the same
way: bookmaking can't thrive, because some or most bookies will cheat their
customers and their customers can't sue them, can't affect their
reputation, and can't physically attack them. The key is that cheated
customers can and will "spread the word." This applies, with some minor
(but interesting) wrinkles, to crypto-mediated bookies. They're all closely
related issues...)


>SCENARIO TWO
>
>I set up a phoney murder-for-hire business.  Someone contracts
>with me to bump-off their rich uncle.  The client deposits my
>payment with a reputable escrow company, "Murder Escrows R Us." I
>go to the uncle and tell him the whole deal.  Using digital
>technology, bribed coroners, etc., we fake his death.  When the
>news hits the Net, the escrow pays me off.  The uncle comes back
>to life, disinherits whomever he suspects wanted him dead.  And I
>laugh all the way to the digital bank.  I create a new pseudonym,
>place another murder-for-hire ad, and do it all again.  Given our
>Brave New World, nobody can touch me.

This just says that standards of proof will be a factor, naturally, and
that markets will take these into account. The phoney murder for hire
business, call it "Sandy's Salvage Company," will start of with a very low
reputation, as with any new outfit with little track record. The standards
for proof, the fees paid, etc., will be proportionately affected.

However, "Tim's Tribunal," which has had a 5-year record of "really and
truly" offing dozens, and which has not been "caught" as being in any of
the scams Sandy described, will demand and get a proportionately higher
fee, and will face fewer delays in being paid.

(Similar arguments apply to any of the parties, which is why I've followed
Sandy's lead in switching the focus from how to handle cheating escrow
agents to cheating contract killers.)


I've written enough. I'm not persuaded by Sandy's arguments that the threat
of cheaters is sufficient to derail these markets. It hasn't derailed them
in the real world. It won't in the less traceable but even more
reputation-critical crypto world.

Again, I've made no claims to how ubiquitous such markets will be, or what
the market dynamics will be. Only that strong crypto makes possible certain
types of markets which are now very illiquid.

The issue of "untraceable cheaters" comes up in many more areas than just
contract killings: information markets, pseudonymous consulting, etc.

This is why reputation, so important in the physical/legal world, is also
so important in the crypto world.


--Tim May


..........................................................................
Timothy C. May         | Crypto Anarchy: encryption, digital money,
tcmay@netcom.com       | anonymous networks, digital pseudonyms, zero
408-688-5409           | knowledge, reputations, information markets,
W.A.S.T.E.: Aptos, CA  | black markets, collapse of governments.
Higher Power: 2^859433 | Public Key: PGP and MailSafe available.
"National borders are just speed bumps on the information superhighway."