September 23, 2008

In part 1, I suggested the future President of the United States promise to bring fairness to the people.

In all fairness, he might consider the health of financial markets a higher priority. I can hardly help him though with the current situation. Nobody can. He will not take office for several months while the best experts are hard put today to plan more than a few days in advance. Long term though, I feel as competent as all those professionals humbled by the current crisis. Besides the less one knows, the simpler the story one has to tell.

Money enables one to both store choice making capacity for the future and take bets on the future. John Kay credits Martin Wolf for speaking of utilities and casinos (*). How apt a metaphor if it did not imply one can and should segregate the two functions of money into two distinct industries. Alas, money aggregates usage indifferently. In fact storing choice is but a bet and, in the end, choice goes to all of us bettors according to our luck.

Worse, by nature, money makes the future a necessary factor to all financial transactions. When shared information gradually leads all present participants to share the same opinion, the future becomes the only real market counterpart to the present. With limited means but no say, the future is prey to bubbles, which develop as its limits seem far away and burst when they present themselves. Alas, all bubbles start as reasonable opinions.

Advanced mathematics guides many professional money managers today. Poorly mastered, these tools are as dangerous to markets as individual opinions turned into shared beliefs. By evolving a whole world of interdependent bets, financial risk modeling has created a complex environment whose very scope defies its being globally modeled. Limited in power, current modeling tools are also deficient in principle, "disregarding the possibility of sharp jumps and discontinuities" as prophetically argued Benoit Mandelbrot and Nassim Taleb (**).

Let then the future US President recognize he cannot compute the future, eliminate bubbles nor reliably identify bad from good bettors, except in hindsight. Modesty calls for simplicity. Give the proper independent federal institution the mission to specify two numbers. The first one, Rmax, represents the maximum amount a non federal entity can bet on its own. The second, Fcap, is the additional multiple this entity can bet, based on the cash reserves it deposits into the safekeeping of the Federal government. Assuming an entity deposits C, it can bet up to: Bets= Rmax +Fcap*C.

It is not for me, nor the future US President, to set and monitor numeric values for Rmax and Fcap. But he would find it simple to explain the new regulation. Fcap limits lending leverage and Rmax risk taking. In Martin Wolf's words, above a certain size, all casinos must increasingly behave like utilities. Such a rule has a fourfold advantage. It shuns measuring risk, the total value of each bet bluntly taken to be at risk. It frees innovation while putting it in its place, actors small enough to fail. It gives cash to the lender of last resort. It allows the federal monitor to broadcast its sentiment.

As Martin Wolf puts it (***), "attention must be paid to behaviour that may appear rational for each institution, but cannot be rational if all [...] are engaged in it at the same time". As a bubble inflates, the federal monitor would lower both Rmax and Fcap at the level at which the global need for credit balances the global need for prudence in lending. John Kay predicts "the industry will successfully resist". But why resist if highly leveraged bettors do not gain from economies of scale? If anything history shows that past a certain threshold, size turns leverage into a long term liability.

If this recalls the setting of the Federal funds rate and the reserve ratio by the Federal Reserve, it is so intended. But this rule would be the same for all economic entities, not just banks, and in principle cover individual bettors big enough to qualify. A threshold-breaking bettor would pay a hefty daily charge until it brought risk down to size, plus fines if the break occurred for increasing risks rather than having the threshold lowered too fast.

Such a simple system would be fair but fairness is not just about taking defensive measures. It should also be wielded to improve efficiencies. The current crisis may have been precipitated by the bursting of a real estate bubble. Its malignancy comes from the failure to date to find a fair market for those new fangled financial instruments. What is required is to create new, appropriate information-based value markets.

To trade risk instruments, one ought to account for all relevant factors, first among them how they were put together. So called risk securitization does spread risk but in a way conducive to general listeria, the poisoning of an entire supply chain by some toxic sources in the mix. The task at hand, to clean up the present and strengthen the future, is but a belated formalization of the role of third party recommendations.

Whether it concerns jobs or credit derivatives, a recommender can either be interested in the resulting transaction or not. The future US President would be wise to provide efficient mechanisms in both cases and insure the corresponding value markets reference all recommenders explicitly.

When an interested recommender is paid to issue insurance on a bad transaction, he should be listed as an abettor and subject to report as a bettor. More dangerous are simple intermediaries, like loan originators. Unless forced in an appropriate way to insure the bets they recommend for a living, their self-interest is to boost quantity at the expense of quality and promptly disappear with their fistful of fees.

Disinterested recommenders are as interesting. Local agents personally known to both parties and small enough to be threatened by competition will keep errors to a minimum lest they lose their reputation. But what about global recommenders who share a de facto monopoly, such as the top credit rating agencies, the top consumer credit report bureaus or the top search engines? Do they care?

I recommend they be subject to a yearly license exam based on a quantitative performance audit. I can hear the howls from major rating agencies. A bond notation is not an insurance against default, only an opinion on the risk of default. Take them at their word. Turn their twenty fancy codes into a probability of default within a year from 0% to 100% plus or minus 2.5% and verify their claim in hindsight. Their very size guarantees meaningful averages. If they underestimate default rates by more than 10%, cancel their license. They would suddenly care, would they not?

Our Information Age ties together markets, information and eprivacy. In this complex context, fair, decentralized, confidential mechanisms can provide the future US President with the right tools to protect and foster democratic prosperity.

Time is a great healer, John Gapper recalls (****). "Given time and analysis, [toxic securities] can be revalued and traded again". The future US President has little time left before the next mid-term elections but good analysts should be easy to find, given job losses on Wall Street. Why not take this opportunity to jump start a dense network of competing and competent recommenders, whose reputations would give risk a good name?

Give credit to Grant Wood's farmers. They would have caught the message. They bet the farm with their local banker based on mutual reputation.

Philippe Coueignoux

• (*) ......... Taxpayers will fund another run on the casino, by John Kay (Financial Times) - September 17, 2008
• (**) ...... A focus on the exceptions that prove the rule, by Benoit Mandelbrot and Nassim Taleb (Financial Times supplement) - March 24, 2006
• (***) .... The end of lightly regulated finance has come far closer, by Martin Wolf (Financial Times) - September 17, 2008
• (****) .. The King's men must put themselves together again, by John Gapper (Financial Times) - September 20, 2008