TOC The Great Data Deception Your Turn

December 1, 2009

Call me old-fashioned, but I like clarity and logic.

Interviewed by Emma Jacobs, Biz Stone, co-founder of Twitter, fails the test (*). In his own words, Twitter is based on "giving away as much data as possible and being open" but lives on the principle "both parties must get something out of a business relationship or it won't last". So, finding no appreciable revenues, logic concludes Twitter won't last unless the ideal of "giving away data" and "being open" is a front for future broken promises or shady deals. Clarity must then assume Twitter's shareholders are content with participating in a vast Ponzi scheme to defraud latter investors or an elaborate confidence game to ensnare users. Were Biz Stone to cling to both logic and clarity, he would be self-indicting himself, wouldn't he?

Meanwhile drowning in doubtful debts, Dubai World tries to stay afloat as it sinks its neighbors' reputation. In her dispatch (**), Roula Khalaf can only clarify a present lack of clarity. "Bankers [...] are furious, feeling they were led on the wrong path". I feel their pain. A year ago they found out the cost of lending to the poor. Now it seems one cannot even lend to the rich. Are banking business models becoming another affront to logic?

Science of course brings clarity and logic to the domains it surveys and economics has long longed to belong. This aspiration is not without recalling Turkey's application to EU membership. Success requires some clever way to transcend an oxymoron, such as putting Asia Minor in Europe or, as Clive Cookson, Gillain Tett and Chris Cook's titled their analysis, developing "Organic mechanics" (***). Piling "biology envy" on top of "physics envy" may not be an appealing formula but borrowing ideas from both sciences is still legitimate as long as it is done eyes wide-opened.

In this perspective the authors quote George Soros in highlighting one crucial difference. Due to "reflexivity", "markets change in response to participants", i.e. consciousness trumps physics. I have myself pointed out that, to the extent evolution applies to markets, the appropriate theory is to be found in Lamark rather than Darwin. Unfortunately these two difficulties combine in a new caveat, characteristic of organic mechanics. Far from being linear, the march of scientific discovery involves many trials and errors. With reflexivity, such trials are turned into measures to better the current reality and with evolution, these measures, erroneous or not, irreversibly turned into a new, less tractable reality.

It would be an error to think this inexorable complexification of economic issues to be some new, personal insight. Neither new, nor personal, it was rather stated as early as 1972 by an expert in dynamic systems and control theory, Michael Athans, today MIT Professor, Emeritus (1).

As a result, whatever scientific truth there is in economics is in danger of decreasing with time. Example provided by the authors, the CDO squared (2) is said on average to depend on 93,750,000 mortgages. Science would provide investors access to the data on each mortgage. Instead bankers have raised complexity to the point one is compelled to resort to statistical modeling, replacing scientific truth with truth based on authority.

There is nothing wrong with truth by authority except that it perforce relies on our personal beliefs. These are shaken when favored recommenders are found to be greedy and ignorant, like credit rating agencies. Beliefs indeed are based on trust and trust betrayed turns to bust. If not, would Gillian Tett warns us of "the logical extreme", that "the debt fundamentals in Dubai are not necessarily so different from those in [...] the US" (****)? Commenting on Toyota's "accelerator pedal" recall on four million cars, Jonathan Soble, Louise Lucas and Paul Betts likewise make clear "the Japanese group is terrified of [seeing its] reputation for quality shredded in the US" (*****). Like cast iron, brands are strong and brittled both.

History quickly dispells the hope of finding refuge in the past. O tempora! O mores! The late Roman Republic was not exactly a model, when even the best orator could not prevail against raw personal power politics. Rather I seek three lessons for the present. Avoid throwing data away. If statistical evidence is unavoidable, remember it is based as much on reputation as on data. In nurturing reputation, promote clarity and logic.

A courageous reader who reads that far may be tempted to yawn. If this filip is pushing clarity and logic, will the next one be about motherhood and apple pie? How long will I abuse his patience with platitudes?

Let me be clear then. Whenever you rely on data, expect data to be corrupt. Who indeed would be naive enough to believe data gathered from Internet without further verification? Who would be naive enough to think the higher the stakes, the lower the potential for deception? Not when Rob Hughes and Eric Pfanner report how the police in Europe has "arrested the leaders of a far-reaching gambling cartel [accused of fixing] 200 soccer games in nine countries" (******). Not when one remember the best coaches may be revealed as ethically challenged as the worst players.

Not when Andrew C. Revkin details why pro climate change scientists no longer use tree rings "to track temperature fluctuations" (*******). "Tree rings and thermometers show a consistent rise in temperature until 1960, when some tree rings, for unknown reasons, no longer show that rise, while the thermometers continue to do so until the present". Let us be logical. When using statistical evidence, one cannot decide to drop data once it contradicts one's preferred conclusions without losing one's reputation. To reclassify tree rings from raw data to unreliable measurements is a trick.

The above judgement is of course strictly limited to the scientists who have been personally exposed by the security breach at the University of East Anglia. But climate change has become a brand and as such can be damaged by a reluctance to recall unreliable supporters. Clarity carries a cost.

Long term, the economic, political, human stakes behind predicted rises in world temperatures cannot be higher. They make the rise of healthcare costs no less real. While everybody concerned agree the latter cannot be contained without gathering more data, the danger would be to assume such data will neatly settle all doctrinal disputes and deliver decision makers from all doubts.

Unless extraordinary measures are taken in healthcare, the more data, the more it will be doctored, the more and the more heated the controversies. Prepare for the great data deception.

Philippe Coueignoux

  • (*) ............. 'It's like riding a rollercoaster', by Emma Jacobs (Financial Times) - November 27, 2009
  • (**) ........... Emirate will pay dearly for a long time to come, by Roula Khalaf (Financial Times) - November 27, 2009
  • (***) ......... Organic mechanics, by Clive Cookson, Gillain Tett and Chris Cook (Financial Times) - November 27, 2009
  • (****) ....... Greece and Dubai show the system is still unstable, by Gillain Tett (Financial Times) - November 27, 2009
  • (*****) ..... Toyota acts quickly to keep its grip on the PR wheel, by Jonathan Soble, Louise Lucas and Paul Betts (Financial Times) - November 27, 2009
  • (******) ... Prosecutors Hit Ring in Europe Fixing Soccer, by Rob Hughes and Eric Pfanner (New York Times) - November 20, 2009
  • (*******) . Hacked Climate-Change Data Stirs a Tempest for Scientists, by Andrew C. Revkin (New York Times) - November 21, 2009
  • (1) for further information, see Michael Athans' page at MIT
  • (2) see the types of CDO's in the wikipedia
December 2009
Copyright © 2009 ePrio Inc. All rights reserved.