From: hughes@ah.com (Eric Hughes)
Date: Fri, 19 Aug 94 17:08:09 PDT
To: cypherpunks@toad.com
Subject: In Search of Genuine DigiCash
In-Reply-To: <9408192037.AA10062@bilbo.suite.com>
Message-ID: <9408192343.AA13855@ah.com>
MIME-Version: 1.0
Content-Type: text/plain


   Can a case be made that anonymous digicash is less risky
   (to a bank) than NON-anonymous digicash? 

In certain circumstances, it might be.  Where a bank is at risk of
violence when it does not reveal transaction information, not
possessing such information poses less risk.  On the other hand, in
the USA a bank is at risk of violence when it does not possess
transaction information.

   Would a Chaum-style anonymous digital cash service be more profitable to a  
   bank than a NON-anonymous digital cash service?

Maybe.  It depends on what the demand curve for transaction services
of various kinds looks like and what the relative demand for privacy
is.

If there were already a fully identified digital money system, then
creating an anonymous digital cash system would grab you most of the
market which was willing to pay a premium for privacy services.  That,
by the way, is not everybody.  There will be at least a local maximum
at some large premium, simply because certain benefits of bank secrecy
are so large.

On the other hand, there is likely also a local maximum where the
premium is fairly small.  In this case you get not only all the people
above, but a large percentage of the people who are willing to pay
just a little more for privacy.

As to where these local maxima actually are, and which yields the
larger profits, I have no idea.

   Are the costs involved in offering and supporting anonymous digital cash  
   more, or less, than the costs associated with NON-anonymous digital cash?

The costs associated with anonymous digital cash may well be less that
for identified digital money systems.  There are additional services
being sold in most identified systems, including statements of
transaction logs, reversibility of transactions, delay in settlement,
and availability of logs to government.  This last service is sold to
the government with each transaction, a hidden fourth party which
taxes the bank with the requirement to offer this service, in order to
permit the bank to operate.  These additional service take resources
to operate.  Reversibility, I suspect, is the most expensive to
operate, since it's all human labor that can't be easily handled by
computer.

Digital cash, on the other hand, needs a redeemed note database, but
this is one of its only unique costs.  Since settlement is immediate,
reversibility is not an issue, and neither is any delay in settlement.
There are far fewer long term records to keep.

It is likely that digital cash is more efficient economically, since
it unbundles a bunch of previously linked services and allows them to
be purchased separately by those who actually need them.

Eric