From: Robert Hettinga <rah@shipwright.com>
Date: Tue, 6 May 1997 08:34:36 +0800
To: cypherpunks@toad.com
Subject: Re: Bypassing the Digicash Patents
In-Reply-To: <v02140b01af925860a90b@[10.0.2.15]>
Message-ID: <v0302096faf941f7ddba2@[139.167.130.246]>
MIME-Version: 1.0
Content-Type: text/plain


At 4:23 pm -0400 on 5/4/97, Steve Schear wrote:


> Although I stongly share your overall beliefs in the superiority of digital
> bearer certificates (DBC), I am much less certain about the intrinsic
> financial cost differentials.

Only time, and actual data, will tell, as I've said elsewhere.

> Let's first consider credit cards (CCs). Consumers using CCs, if they
> choose their bank/agent carefully and settle their monthly statements in
> full and promptly, pay nothing (directly) for the priviledge of using the
> cards.  One can argue that merchants build the cost of their CC transaction
> clearance costs into their price structure and therefore every patron,
> whether using cash or CC, pay for the costs of CC usage (few merchants
> offer cash discounts).  Most walk-in retail merchants, in good standing,
> pay about 2% in discount (fees).  On-line and mail order merchants pay
> about 3% as do those with a higher rate of chargebacks and/or in grey
> businesses (e.g., on-line gaming and adult entertainment).  These fees
> represent the costs of operating the CC clearing system (MC, VISA, AMEX,
> etc.) the bank's overhead and profit.

So, you've refuted that one for me. :-).

> Now let's look at Automatic Clearing House (ACH).

Funny you should mention that. It's one of my ideas for an e$Lab company.

I claim that ACH transactions can be driven to 2.5 cents with something
like the FSTC electronic check project when it comes on line. And, that
once you have enough bank servers on the net operating as "deposit windows
on the information superhighway", Metcalfe's law kicks in, and those
servers can clear against each other directly on the net, bypassing the ACH
system altogether. Probably bringing the price of clearing an internet
check to .0025 cents, which would leave the ACH record format as the Roman
wagon-wheel rut (which begat the standard european rail guage) of the
information age, and giving us the three or four orders of magnitude I was
talking about.

> It is not clear that the back office costs of running a mint are any
> cheaper than a book-entry system (e.g., control and auditing requirements
> to meet Generally Accepted Accounting Principles and establish/maintain
> trust)

Funny you should mention that one too. :-). The mint itself, being run by
an underwriter, probably doesn't need anything else in the way of record
keeping but a list of expired certificates, certificate issues, and keys.
The only place GAAP would really apply would be the trustee, who needs GAAP
anyway as a requirement of the book-entry world it lives in. And,
fortunately, the system already accounts for things like ATM transactions
anyway, which is how I see digital cash collateral getting on and off the
net for the most part. No net accounting load at all, to my mind, except,
of course, from the long-phase, but temporary, increase caused over time as
the net becomes a better place to do business than meatspace. Again, at
some point, when on-net trustees can keep assets in digital bearer
certificates, GAAP may hold for some things, but probably not too much for
the collateral pool of digital bearer assets, because those can be checked
by some kind of ZKP, anyway, right? There's also Eric Hughes' encrypted
open books idea, which would probably follow GAAP, granted. Anyway, maybe
the data format for ATM systems would become another Roman axlewidth to
measure commerce by.


> and you still need the book-entry system to occassionally transfer
> value; to the extent that digital bearer certificates become accepted as
> legal tender costs should decline, but the regulatory issues increase.

It's hard for me to see the regulatory load increase in a transaction
clearing system collateralized entirely by digital bearer certificates.
They can be anonymously held, remember? :-).

> However, the ACH and CC system's prices reflect the substantial Fed and CC
> backbone cost and profit to run their private networks, and it is here that
> costs for on-line digital bearer certificate systems will have and edge for
> some time.

Exactly. See my points about wagonruts, above. :-).


> When one combines the transactional economic advantages of DBC: no customer
> accounts, statement generation, little or no need for dispute resolution
> and the economies of using the Net for settlement, with the still
> considerable marketing costs, the likely differential between DBC and
> CC/ACH is closer to one order of magnitude at best.

Again, I still hold out for three or four.

> This is still
> considerable, however, and if properly branded and marketed could
> significantly displace current competition, for lower value transactions.
> The overhead of CC discounts is keenly watched by merchants.  If a trusted
> DBC issuer/agent offered 1.0% fees (especially to on-line merchants) it
> would get noticed quickly.

Agreed. However, when it happens, it'll probably be more significant than
that. I think removing 3 or 4 extra zeroes from the status quo is probably
what people should try to do when developing this stuff. I think that the
technology offers us at least that much slack, and probably more.

Remember, it *is* supposed to change the world. :-).

Cheers,
Bob Hettinga


-----------------
Robert Hettinga (rah@shipwright.com), Philodox
e$, 44 Farquhar Street, Boston, MA 02131 USA
Lesley Stahl: "You mean *anyone* can set up a web site and compete
               with the New York Times?"
Andrew Kantor: "Yes."  Stahl:  "Isn't that dangerous?"
The e$ Home Page: http://www.shipwright.com/