August 7, 2012

"[Retail investors] don't do enough research and tend to sell on emotion rather than logic". With this powerful and alas accurate indictement, Steven M. Davidoff sends a shot across the bow of decentralization (*). If the latter were such an unmitigated good, wouldn't a crowd of small time investors be smarter at pricing the market than a paucity of savvy investors?

I claim no expertise in finance. Besides, four years into its repressed depression, the current sorry state of the economy is proof enough that, past a certain point, problems defy easy solutions. While the patient is still alive, the endless parade of experts called to its sick bed reminds one of Molière's physicians, whose ineptitude was in proportion to their use of fake latin. Et tu quoque Libor!

Yet why not take stock of what markets can teach us? It's a cheap way to simulate the future of our Information Age in which data is like money.

The irony of it all is that, as long as one does not look for a cure, the diagnostic is pretty simple. On the one hand the larger the crowd, the more dangerous its herd behavior. On the other hand the minute amateur investors delegate their tasks to expert professionals, the latter are bound to serve themselves first, the so-called principal-agent problem (1).

Classical bubbles are good examples of herd behavior. If the price of a house appears certain to increase over time, everybody wants one and prices do go up until cheap credit runs out. As for illustrating the principal-client problem, nothing beat the London 2012 Olympic Games.

Sam Borden stresses athletes "trained to win a medal" (**). Last week some badminton players found "a better chance at winning the tournament [in losing] the final group game". So they threw the game, only to be thrown out themselves to placate the crowd. Belatedly the officials realized they had misaligned their incentives with the demands of the public. As principal, it paid to be entertained, not fooled by professional athletes, the agents.

Take away the jargon, bankers behave like badminton players. They go for the gold and, if in the process, they leave society penniless, they ask "where exactly did [they] veer off into corruption? They did not organize the tournament. They did not arrange the draw." The naive public "should also not be so foolish". Inside their own opaque bubble, agents laugh at the principal's expense.

There is of course a small difference. To sustain an economy, one must be able to store present value for future use and allocate current resources to the most valuable projects. One can stop bad badminton at little cost to the public at large, not so with the flow of money across space and time.

The same imperative applies to information. When people worried about lack of eprivacy stop using personal data on the Internet, it is as if they converted all their assets into gold to hid in their mattresses. Possible perhaps but doubtfully desirable on a large scale.

Assuming data banking companies are not yet as well entrenched as their financial peers as to defy reasonable remedies, what should be done then?

The idea that people are too stupid to be responsible for themselves is alluring. It is also dystopian to the extreme. It leads to a world run by professional players and their wonder machines as typified by the like of Lehman Brothers, MF Global and Knight Capital Group. The latter is reeling after having "revealed a $440m pre-tax loss from erroneous trading positions triggered by a software glitch", Arash Massoudi reports (***).

In hindsight, decentralization remains a good thing. But we must not forget to be on the look out for herd behavior and check it before it is too late.

Gillian Tett asks "whether th[e] 2.6tn money market fund sector is vulnerable to "runs"" (****). To avoid deadly stampedes, create safety exits and limit malicious speech. Do First Amendment rights allow one to falsely shout "fire" in a crowded, darkened movie theater? When it comes to information, rumors mirror panics, runs become swarms. China bashing is not the best way to discuss how to spot and manage virtual mobs.

Still Steven M. Davidoff's warning carries over to the world of data. Despite their shortcomings, agents are there to stay. Can they be tamed? When dealing with animals, classify. Until all the facts have been revealed, it is not so easy to separate the goats from the sheep and bears from bulls. But surely we should see to sort tigers from horses before picking a mount. Will the segmentations proposed by John Kay and Gillian Tett be of help?

John Kay distinguishes between "transactions and trading over trust relationships" (2). Popularizing anthropologist Michael Thompson's views, Gillian Tett suggests to sort the environment according to whether "anybody is in charge" and whether "the power structure [is] benign" (*****).

One should not rely on trust in a jungle. Where "the laws govern the poor, and the rich govern the laws" according to Dr Primrose's apt definition of pronaocracy (3), expect mostly tigers out there. As they love to lie in wait in impenetrable legislative thickets, threaten them with transparency. Start with defoliation, follow with an open road network. Simplify the laws, clarify the rules, mandate interoperability, you will create a more benign milieu.

In such a milieu, real trust can exist based on mutual dependence. The latter precludes the existence of monopolies, cozy oligopolies, any organization too big to be tamed or whose services are too addictive to be switched off. To flourish, trust relationships thus need strong anti-trust enforcement. Lest the latter descend into industry warfare, competition by means of the government, the people, not the crowd, must be in charge.

Trust however is slow. Like any good shock absorber, it dissipates energy. When agents have to be given a fuller rein, we must leave the realm of trust for the world of transactions. Such a world needs an even stronger government to make sure horses do not slip their harness altogether. Roles must be severely limited, lest conflicts of interest arise. Speed must be taxed, lest systems become runaway. Successful solutions must become open lest they achieve this feat, so contrary to both biology and our interests, the metamorphosis of horses into tigers, markets into rackets.

So what should we do? The best source of trust springs from local recommenders. It should be encouraged. Second vertical integration and diversification are bound to create monsters. All large technology companies today fall in this case. They should be given the AT&T treatment (4).

To work, decentralization needs a strong democracy. Too bad citizens "don't do enough research and tend to [vote] on emotion rather than logic".

Philippe Coueignoux

• (*) ......... In Picking Facebook Shares, Repeating the Past's Mistakes, by Steven M. Davidoff (New York Times) - August 1, 2012
• (**) ....... The Goal Is Winning Gold, Not Winning Every Match, by Sam Borden (New York Times) - August 3, 2012
• (***) ..... Software glitch leaves brokerage Knight nursing loss of $440m, by Arash Massoudi (Financial Times) - August 3, 2012
• (****) ... The Achilles heel of America's financial system, by Gillian Tett (Financial Times) - July 31, 2012
• (*****) . Anthropologists join actuaries to teach us all about risk, by Gillian Tett (Financial Times) - August 3, 2012
• (1) for more details, see the Principal-agent problem in the wikipedia
• (2) Finance needs trusted stewards, not toll collectors, by John Kay (Financial Times) - July 23, 2012
• (3) for more details, see the The Vicar of Wakefield in the wikipedia
• (4) for more details, see the break up of the old AT&T in the wikipedia