02 January 2007

Gladwell on Enron


Malcolm Gladwell has an article in the latest New Yorker on the Enron incident, in which he draws attention to the differences between incidents that unfold like a mystery and incidents that unfold like a puzzle. His argument is that the unfolding of the Enron case did not proceed like a puzzle, in which investigators realized that shareholders were given incomplete information on the shady side-deals of Enron executives and proceeded to find the missing information they needed to explain Enron's sky-high stock price. Rather, the Enron case unfolded like a mystery, in which all of the information shareholders needed in order to see how Enron was cooking the books was fully disclosed from the beginning, but obscured in the white noise of complicated accounting and financial practices that firms engage in today.

Solving the Enron case meant discerning the evidence that mattered from the vast white noise of data. As Gladwell puts it, solving Enron was like diagnosing prostate cancer in an externally healthy patient or predicting the fall of the Berlin Wall from intelligence collection. All of the information you need is right there, but since there are multiple inferences you can draw from surface-level symptoms, you can't linearly deduce the origins of the problem from the ends. You solve the mystery by making predictions about the probable outcomes of critical junctures where causes meet their contingent effects, not by finding some missing piece of information, some missing link.

Gladwell's argument is really about the complexification of systems in modern society, and the inundation of puzzle pieces we have at our fingertips in today's Information Age. Intelligence, financial reports, results from medical tests--the data we need to answer our questions is nearly always out there, lost in a tangled mass of less useful data. The intelligence community is more concerned with sifting useful data from the noise of useless data. Doctors are concerned not with developing more tests for as yet undetected measures, but with developing more sophisticated explanations for the relationships between test results and diagnostic causes.

It's a great and useful distinction, this difference between problems that are puzzles and problems that are mysteries. And the argument seems to be that whereas in the past, big media cases unfolded like puzzles, today, in our complex markets, politics and societies, big scandals unfold like mysteries. The Watergate scandal and Deep Throat's provision of the key missing link to the case are a thing of the past; today, we have complex mysteries of multiple possible causes for each observed outcome. Problems are harder to solve today than they were yesterday.

But were there ever that many problems that need to be solved like a puzzle rather than a mystery? Markets, politics and societies are undeniably more complex systems now than before, but haven't problems always had multiple contingencies and multiple causes? Haven't problems always been mysteries, regardless of how we treat them?

Why did democracy work in revolutionary American but collapse in revolutionary France? Why do more immigrants become citizens in Canada than in the U.S.? Why do labor unions form in newly globalizing regions?* It's not that we need missing information or missing links to answer these questions. It's that we need to reorganize the data out there to highlight causal explanations and built theories to construct a foundation for these explanations. We might collect more missing data in order to show that hey, the problem isn't that labor unions form in globalizing regions, it's that labor unions form in democratic regions, which are more likely to be globalizing as well. Finding the confounding variable in the problem is a missing link approach to a puzzle. But you're still stuck with the question of why unions form in democratic countries (not necessarily a true statement). And this question is a mystery, not a puzzle. I think the fact is, most problems have always been mysteries, not puzzles.

Maybe what has changed is not the nature of today's problems, but the treatment of these problems in circles like journalism, medicine, and intelligence analysis. Radiologists forty years ago were exuberant about the CAT scan, which allowed them to collect missing information like changes in cranial pressures and visual evidence of tumors, but this missing information didn't provide explanations for the origins and causes of these problems. They still had to fill in the gaps between evidence and outcome with narrative explanations of what actually happened.

Maybe the difference between a puzzle and a mystery is really a question of what kind of evidence you find. Do you see external symptoms of the underlying problem (puzzle), or do you see a theory of the developments leading to the outcome (mystery)? Maybe we need to see every problem as a puzzle until we have sufficient evidence to isolate the actual mechanism occurring (it's democracy, not globalization, that causes a proliferation in labor union formation); and only then can we build an explanatory narrative theorizing the nuts and bolts of the mechanism. Either way, the question of whether Gladwell's right that puzzles become mysteries with time and modernization is itself a puzzle. Or, if you'd like, a mystery. I'm not ready to attribute the move from puzzle-like approaches to solving problems to mystery-like approaches to solving problems to time and modernization. I think it's a question of kinds of evidence and methodology.

*All these mysteries are lifted from academic sociology books: Tocqueville's Democracy in America, Bloemraad's Becoming a Citizen, Silver's Forces of Labor.


Post a Comment

<< Home