RE: [ubl-dev] Modeling multiple possible invoice totals based on payment means

["Andrew Chilcott" <andrew.chilcott@stpsolutions.com>]

Invoiceless trading does work exceptionally well, but in very restricted
circumstances. The best example that I have is the UK securities industry.
In 1979, when the London Stock Exchange introduced its Talisman system,
invoices (contact notes) between stockbrokers and jobbers(principal dealers)
were abolished. The basic principle being that each party to a transaction
inputs their version of the deal that they have agreed to a central
settlement counterparty and, as long as the primary (usually about 6) fields
match, the settlement counterparty/bank is authorised to move funds between
the parties on the agreed settlement date.

In the early 1990's the LSE tried to update this system with a replacement
called Taurus. This was an industry-wide project that ended up being pulled
one week before its launch date. The ramifications of this were immense. The
project had been too complex, trying to build in industry bad habits as
'market norms'. The Bank of England came riding to the rescue and in a bold
industry meeting in 1993, IT directors and Chief Executives were told that
the B of E would introduce its CREST settlement system in 1995, and that
they better all accept this or go out of business. CREST was to be a very
basic settlement engine exchanging money for the delivery of stock. The
system had few bells and whistles when it started and, miraculously, was
brought in on time and on budget.

Even with the complexities of delivering title to securities, CREST settles
in excess of 300,000 trades every single day - some £30bn in value each day!
These trades typically settle in full 3 days after trade date and in some
circumstances on the day following trade date. It is hugely efficient and
one of the great payment infrastructure successes of the last decade.
Compare this with the 30+ days that industrial/commercial companies take to
process invoices. Why was it such a success? In my opinion, this was because
there were a finite number of participants in the industry (less than
 200 initial CREST members). They were all trading an external, industry
defined product that was built into all of their back office systems.
Legislation was passed to enable legal title in the securities to pass at
exactly the moment of settlement. All participants were financial
institutions and had solid balance sheets. All participants were told from
the outset that they had no option other than to join and that if they
didn't they would be out of business the day after CREST was introduced.
Everybody joined, everybody invested in systems and as a result a whole
industry benefitted.

In essence, there was no disconnect between the payment system of one
participant and the receivables system of another as outlined in Fulton's
response as being the major problem in invoice settlement processing.

I am aware that some major financial institutions are advocating the
creation of settlement hubs for commercial/industrial companies based on
similar lines to CREST. This will be difficult unless all participants in
the supply chain join, unless there is a commonly defined product to trade,
unless the delivery of the product and therefore the legal liability to pay
can be evidenced within the settlement system and unless all of the back-end
systems have automated connectivity to the settlement system.

Andrew