From: rah@shipwright.com (Robert Hettinga)
Date: Wed, 22 Jun 94 08:46:57 PDT
To: perry@imsi.com
Subject: Re: e$: Geodesic Securities Markets
Message-ID: <199406221546.LAA27793@zork.tiac.net>
MIME-Version: 1.0
Content-Type: text/plain


Perry Metzger says,

>Its fairly easy to envision a system that directly
>matches orders for shares in IBM. Trying to match up buyers and
>sellers of swaps might not be that easy.

Agreed. I do want to talk my buddy about derivatives, however, because he
seems to think that there are ways that e$ protocols could help. Of course,
he didn't tell me in any way I could understand...  I think it had to do
with the authentication of e$ instruments, proving you are who you say you
are and you have what you say you have.  A WAG: Maybe zero-knowlege stuff
has something to do with this?

However, like you say it's easy to see how IBM could issue its own shares
directly into a geodesic market; maybe even a company could do it's own
IPO, with the right rules...

When you think about it that way, an investment banker could "sign" an
issue (for a fee...), to vouch for it. After all, using that kind of
"protocol" on large well established companies is what kept Morgan partners
in wood walls and leather wing-back chairs from the time of Glass-Stegall
all the way up until the early 70's. In the IPO markets, that's what
Hambrecht & Quist does (did? ;-)). The technology changes, but the practice
would be the same in a geodesic market. I mean, there's the risk of
underwriting the security, but that's pretty much nonexistant these days
anyway, and with some modification of the shelf registration idea, a
company could "underwrite" it's own offering. Of course when you're
underwriting yourself, the term "underwriting" ceases to have meaning,
really.

>In the foreign exchange market, for instance, most trading
>is done on blocks of millions to hundreds of millions of dollars worth
>of currency. In the current scheme of things people will only deal
>with entities that they know because fails are devistating. It is
>possible for third parties to guarantee credit to open up markets, but
>they will expect to be paid for this. You can't get rid of the banks
>-- someone has to guarantee that you have the money on hand.

It seems to me that the issue of capital concentration could be taken care
of by the mutual fund analog I used before, the portfolio manager in
Marblehead with an e-cash / fund "certificate" exchanger.  The restriction
of transaction size in the markets is still there, though it might be that
the economies of scale would diminish a bit, reducing that required
transaction size...

The whole idea of avoiding trusted entities is what the DigiCash algorithm
is about, right? e$ itself in proper form is proof it is what it says it
is. Just like it's possible to spot counterfeit money. It's a pointer to
some other stuff, but it can be moved around much better than the stuff
it's denominated in, or the market wouldn't create that particular form of
e$.

>...no matter what you hold, be it dollars,
>shares of IBM, or futures contracts for dried silkworm cocoons (a
>perfectly real commodity, by the way) you need a bank to hold the
>account and guarantee the existance of the thing being held, be it a
>figment of the computer's memory or a thing backed by a bar of gold.
>The banks will expect to be paid for this service. Try imagining a
>digital cash algorithm that DOESN'T involve a bank, and you will
>swiftly see that there is a small problem involved...

I'm not entirely sure we disagree, Perry. e$ protocols are just pointers
(with teeth) to the actual stuff being traded. e$ (cash, mutual fund
shares, cocoon futures) has to be issued by somebody and it has to be
convertable into something else to be of any use at all.

To take one business model that we've pretty much killed, it's easy to see
how someone could plug the ATM network into one end of a DigiCash server,
and pay out DigiCash to customers on the internet, using encrypted card
swipes and PIN numbers for authorization.  Banks are necessary in that
scenario, because the money the digicash was issued for has to be held in a
bank by the issuer, whoever that may be ;-)... Or maybe the issuer buys
t-bills (safely) and keeps the interest. Maybe the issuer charges a nominal
transaction fee for the conversion from one side of the gateway to another.
The money doesn't just disappear or get created.  Only governments (or
lending institutions) get to do that (both ways, in the case of Uncle Sam).

>
>This is not to say that transaction costs can't be radically reduced,
>and the role of intermediation in fully fungible goods reduced.
>However, transaction costs will not go to zero, and banks will not
>disappear.

Yes.  There will still be a need for *some* demand deposits. There will be
a need for *some* brokerage accounts. The timescale is certainly nebulous
here. But I think that like most successful new technology a superset of
the old capital market structure gets created. We still walk, but we fly
too.

Cheers,
Bob

-----------------
Robert Hettinga  (rah@shipwright.com) "There is no difference between someone
Shipwright Development Corporation     who eats too little and sees Heaven and
44 Farquhar Street                       someone who drinks too much and sees
Boston, MA 02331 USA                       snakes." -- Bertrand Russell
(617) 323-7923