October 21, 2008

In Heisenberg's universe, objects do not, at the same time, have a precise position and momentum (1). However frustrating this model may be, it has freed physicists to focus their energy on more useful subjects. Reading Jennifer Hughes and Nikki Tait's dispatches from the front in the "fair value accounting" war (*) (**), one wonders if times are not ripe for quantum accounting.

Assets are to this new theory (2) what particles are to physics. Following Heisenberg, we declare one cannot observe both the ownership and the monetary value of an asset beyond some fundamental uncertainty level.

For instance take a so-called toxic collateralized debt obligation (CDO). If you observe it is legally in the hands of bank B, you should not hope to assign it too precise a value. Since there is no working price-based market on which to trade it, it is unfair to force on it a pseudo market price and call it a fair price at that. Neither is price to maturity a legitimate measure unless you assume the holder need not report its balance sheet until then.

The opposite tack may prove even more persuasive. At bottom a CDO is a complex instrument binding a number of actors to a set of real physical assets, whose total value could be measured with great precision at any time if people cared to. It is however impossible to allocate this value among the several participating actors with any degree of precision as their claims are future, not current, i.e. the asset location cannot be ascertained exactly.

If I may mix my metaphores, financial risk tends to turn asset price into a relative notion. In the value market approach I advocate, price is but one among the many factors considered by the two parties negotiating the trade of a complex security. The transaction occurs not because the two parties agree on how to price the security but because, at that price, the buyer assigns a higher value to the security than the seller.

In physics, the Heisenberg uncertainty principle is no reason for measurements not to account for all known relevant factors. Similarly, while the transaction price observed on a value market is not absolutely accurate, it reflects better than any other pricing mechanism the corresponding asset value. Instead of wasting precious resources on some "illusory value accounting" to end what I propose to dub the "war of the tranches", accounting bodies and commissions ought to encourage the design of such value markets.

In keeping with the primary purpose of Philippe's fillips, one will appreciate that price accuracy on a value market rises when participants are limited by design to consider the objective factors of a deal and no other. This calls for a concept which is the raison d'être of ePrio: total confidentiality.

If participants on a value market are not kept at arm's length, prices can easily be manipulated. If third parties, first and foremost market operators and regulators, can observe price discovery, nothing prevents one such third party to take advantage of this information. Whether directly or through confederates, whether sanctioned by the hierarchy or perpetrated by a rogue employee, such abuses would undermine market trust. On a value market, negotiators resort to game theory and the game is either confidential or a confidence game.

How much prompting does Warren Buffett need to poke fun at illusory precision? Accountants should not fear to lose face if they bow to reality. Besides there is enough work to do to capture business reality as it is.

Take Ben Worthen's account of the growing spread of the new data privacy laws which mandate businesses to encrypt "personally-identifiable customer data" (***). Set aside the fact that all customer data is personally identifiable, when in sufficient quantity. The point here is that these laws are shown to put a financial burden on business. That's how regulation undeservedly gets a bad reputation.

Imagine for a minute that accounting bodies recognize the fact that holding customer data creates a liability, just like holding customer money. Then I would be surprised if encrypting such data would not cost less than the decrease in the corresponding liability, a financial benefit for business. Good accounting rules can go far in revealing true value.

Will privacy rights ever receive their due? Realistically this outcome is highly uncertain. It would require such a quantum leap in accounting.

Philippe Coueignoux

• (*) ..... Debate to ease fair value accounting gathers pace, by Jennifer Hughes (Financial Times) - October 16, 2008
• (**) ... Accounting rule changes could dent confidence, say analysts, by Jennifer Hughes and Nikki Tait (Financial Times) - October 21, 2008
• (***) . New Data Privacy Laws Set For Firms, by Ben Worthen (The Wall Street Journal) - October 16, 2008
• (1) see the uncertainty principle in the wikipedia
• (2) and liabilities are to assets what antiparticles are to particles.
• (3) for more information, see the documentation for our lecture on Handling of Medical Records