Re: e$: Cypherpunks Sell Concepts

At  5:41 PM 8/6/94 -0700, Eric Hughes wrote:

>The obstacles are certainly not for electronic money, which the Fed's
>been using for some time now, but rather for electronic cash, which
>includes anonymity.

We've chased each other around a tree like this one before... Let's see
what the differences are this time. I've been doing some thinking about
this...

Anonymity can come out of retail settlement of e$, if the transactions
aren't tracked.  We've talked here before about how you think that the
tracking of those transactions at the retail level is pretty trivial, so
the cost to the user of traceable e$ may be meaningless.  I'm not so sure
that that's the case, and I think (I hope!) I remember Perry agreeing with
me on that point. But if we fiat the argument just to see where it takes
us, we come to the sheer volume of transaction records themselves.

Is it possible to accurately estimate the cash transaction load of an
economy?  I bet that if we could, you'd see that the data from each
transaction would cause the problem news servers have by several orders of
magnitude.  The information would get dumped pretty frequently.  This is
probably the same problem the NSA has now picking out signals to listen in
on, but running down an audit trail is different, it's a historical
process.  Since you don't know whose transactions you need, you need to
keep them all. True, this doesn't keep TLAs from trying trying to drink
from a firehose, or more to the point, to free-dive to the bottom of the
Marianas Trench (if they could keep all of the data), or high-dive into a
wading pool (if they couldn't). Hmmm...

>The USA provides a fair amount of financial
>privacy to everyone but the government, particularly law enforcement.
>So the _business_ case for privacy is largely felt to be already
>satisfied by the regulators.

When *every* business transaction can be scrutinized (as much as physically
possible, per above) at any time, for any reason the government deems
necessary, it makes a sizable business case *for* traceable electronic
cash.  This is probably the place to put the lever on the business
community.

>The Treasury department, among others, really _doesn't_ want
>non-recorded transactions.  Unless the banking community as a united
>front _does_, I don't think it will happen domestically (USA) before
>other deployments.  If there's not a united front, it'll be divide and
>conquer.

Non-recorded transactions exist already. It's keeping them from
dissapearing that we're really talking about here.

It's quite possible to get banks to present a united front.  They have one
of the largest lobbies in Washington.  They have fought reporting
requirements tooth and nail with some considerable success, but every time
they get greedy (S&Ls) the noose tightens.  It might be the threat of
international deployment and regulatory arbitrage which brings them around,
and fires up the lobbying apparatus on our side of the issue.  It has
worked before (gold, et.al.). On the other hand if those reporting
requirements are frictionless, they don't *need* to fight it, do they...
It's time to leave the ring.  Somebody tag me.  My brain hurts...


Now to plug the topic of the thread a bit, how receptive would people in
the crypto community be to participating in an annual dog&pony/schmoozefest
for the suits?  Who should chair the morning "primer" session?  *E-mail* me
with your ideas, everyone.


Thanks,
Robert Hettinga


-----------------
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

Robert Hettinga says:
> Is it possible to accurately estimate the cash transaction load of an
> economy?  I bet that if we could, you'd see that the data from each
> transaction would cause the problem news servers have by several orders of
> magnitude.  The information would get dumped pretty frequently.  This is
> probably the same problem the NSA has now picking out signals to listen in
> on, but running down an audit trail is different, it's a historical
> process.  Since you don't know whose transactions you need, you need to
> keep them all. True, this doesn't keep TLAs from trying trying to drink
> from a firehose, or more to the point, to free-dive to the bottom of the
> Marianas Trench (if they could keep all of the data), or high-dive into a
> wading pool (if they couldn't). Hmmm...

It is perfectly feasable to track all financial transactions in the
U.S., down to the "quarter for a phone call" level, without
eliminating all capacity to use the data or placing more than, say,
another several percent burden on the cost of all transactions. I know
how to architect such a system, and I'm sure that I'm not the only
one. It would be a big job, but not an impossible one, especially not
with modern computer systems. A several percent burden on the economy
would be devistating, but from the point of view of the bureaucrats it
probably isn't such a bad thing. I feel that it is inevitable that the
folks in Washington will eventually come to the conclusion that such
systems are needed -- the boys at FINCEN will start bawling for them,
and the drug warriors will want them, and the rest of us are all just
a bunch of folks who are upset that we couldn't go to woodstock
because we had to do our trig homework...

Perry

> It is perfectly feasable to track all financial transactions in the
> U.S., down to the "quarter for a phone call" level, without
> eliminating all capacity to use the data or placing more than, say,
> another several percent burden on the cost of all transactions.
>
> Perry
> 

	Already, at least here in the northeast, virtually all credit
card transactions are on-line verified - it would take relatively little
additional effort to capture additional transaction details including ID
from our spiffy new national ID card and a more specific description of
what was bought.  And many supermarkets around here now do a substantial
part of their business via debit or credit cards and checks - the added
burden of converting everything over to watchable on-line electronic
transactions is probably not measured in percent per transaction but in
fractions of a percent.  The major investment in on line retail
infrastructure has already been made in most cases, what needs to be added
is just some additional software and a more legally binding ID card.

	One suspects that the cost of physically handling cash,
providing security for it and so forth is actually quite comparable to
costs of such a cashless electronic regime.  Outlawing cash is indeed
(unfortunately) quite practical. 

	If I had to guess as to what *the major* domestic target of wideband
electronic surveillance and monitoring by the TLAs is licit or illict, I
would name the credit card authorization data streams.  Probably that
and interbank wire and check clearing transfers consitute much the
largest cross section of data being watched regularly.   And I am unclear
as to whether such surveillance, with the tacit consent of the banks and
credit card companies of course, is obviously and specifically illegal. 

						Dave Emery

Is it possible to accurately estimate the cash transaction load of an
   economy?

I have some 1992 USA figures on this.  The number of checks was 58
billion (58 * 10^9).  The number of card transactions was 12 billion.
There were about 2 billion other electronic transfers.  72 billion
total.  Cashless transactions are about a tenth (roughly, this is from
memory) of the total.

So as a first cut, assume about one trillion (10^12) transactions to
be tracked per year.  Assume 1/8 Kbyte per transaction (that's a lot).
If you stored transactions on 8 Gbyte tapes, that's 2^40 xact * 2^7
bytes/xact * 2^-33 tapes/byte = 2^14 tapes, or about 16 thousand.  A
robotic retreival device for 16 thousand tapes is certainly feasible;
I've seen a similar system for about 2 thousand 9-track tapes -- it
was feeding a Cray 2 at Livermore in their fusion center.

Now that's just storage, not the whole system.  But it's apparent from
these estimates that a real system is certainly affordable, and,
possibly, relatively inexpensive as such totalitarian devices go.

Remember, "suspects" (10^-3 of the population) can be filtered out
before hitting tape and stored on about 128 Gbytes of hard disk, for
very fast retreival and realtime analysis.

   When *every* business transaction can be scrutinized (as much as physically
   possible, per above) at any time, for any reason the government deems
   necessary, it makes a sizable business case *for* traceable electronic
   cash.  This is probably the place to put the lever on the business
   community.

It might be, but remember that in making the case to business, the
financial privacy, such as it exists today, is _not_ "at any time, for
any reason".  It might be in the future, but then you're making a
perceived-weaker argument.

   Non-recorded transactions exist already. It's keeping them from
   dissapearing that we're really talking about here.

The number of non-recorded transactions, however, is dropping.  The
largest class, cash, got some reporting requirements clamped on it
recently.  We are talking about both ensuring that the current
non-recorded transactions stay that way and allowing for non-recorded
electronic transactions in the future.

   It might be the threat of
   international deployment and regulatory arbitrage which brings them around,
   and fires up the lobbying apparatus on our side of the issue.  

With that in mind, shouldn't you have your first conference in London,
invite a bunch of US bankers, and raise the issue explicitly?  As soon
as you can get different countries competing for revenue, you're more
than halfway home.

   On the other hand if those reporting
   requirements are frictionless, they don't *need* to fight it, do they...

Nope.  And remember, the divide-and-conquer is likely already
starting.  The first bank to provide FINCEN with a live transaction
feed will likely see some regulatory hurdles fall, no?

Eric