Posts Tagged ‘taleb’

To what extent can we quantify human action?

Given rapidly expanding computational capacity and the proliferation of cheap sensors, there is a large, distributed trend towards human quantification. Wired Magazine confidently threw it on the cover of its July 2009 issue, celebrating self-tracking as popularized by the Nike + iPod system. It’s certainly a trend with a lot of momentum, and I imagine a lot of success will be had by people building businesses around it, but I’m increasingly worried about the way in which the concept is treated in gushing terms and without an understanding of its limitations.

To be clear, I think there is solid evidence that making explicit certain types of information can induce better behavior. One of the best examples is the feedback loop created by displaying energy use in real-time to homeowners, a practice that has been shown to reduce energy consumption. In fact, anyone who has driven a Prius can probably attest to their effort to keep the real-time MPG monitor in the higher numbers, an effort that changed my driving habit for the more efficient.

But I think too much exuberance for human quantification runs the risk of falling prey to a form of techno-utopianism that has already stricken many fields.

This post was catalyzed by two recent articles. The first, by copyright crusader turned political reformer, Larry Lessig, is an extended critique of the drive towards more transparency in government. He writes,

“We are not thinking critically enough about where and when transparency works, and where and when it may lead to confusion, or to worse. And I fear that the inevitable success of this movement–if pursued alone, without any sensitivity to the full complexity of the idea of perfect openness–will inspire not reform, but disgust.”

For Lessig, the work of groups like the Sunlight Foundation or MAPLight (“money and politics: illuminating the connection”) are too simplistic. The problems these groups rightfully seek to fix are too complex to be solved by transparency alone.

“This is the problem of attention-span. To understand something–an essay, an argument, a proof of innocence– requires a certain amount of attention. But on many issues, the average, or even rational, amount of attention given to understand many of these correlations, and their defamatory implications, is almost always less than the amount of time required. The result is a systemic misunderstanding–at least if the story is reported in a context, or in a manner, that does not neutralize such misunderstanding. The listing and correlating of data hardly qualifies as such a context. Understanding how and why some stories will be understood, or not understood, provides the key to grasping what is wrong with the tyranny of transparency.”

Now, perhaps the people who need to know the entire story – the trial judges, the political decision-makers – will take the time to look past a simplistic “money + politician = bribe” equation, but I think the worry is legitimate. The reason for the worry is the seductiveness of simplicity.

This is a point Paul Krugman makes strongly in his recent essay about failure of economists to predict or avoid the current recession. As he chronicles the shortcomings of modern economic thought, he writes,

“As I see it, the economics profession went astray because economists, as a group, mistook beauty, clad in impressive-looking mathematics, for truth.”

The “impressive-looking mathematics” are the economic models that academics have conceived and investors embraced. The short-comings of these have been noted time-and-again by critics like Nassim Nicholas Taleb or Richard Bookstaber (whose book I reviewed here), but these are minority voices. Bookstaber’s Congressional testimony is actually quoted by Krugman:

One thing that seems clear is that risk models that are designed to function in normal market conditions should not be relied upon to predict outcomes in times of crisis. On this account, VaR doesn’t kill banks; executives who don’t recognize the limits of VaR [the value-at-risk financial model] kill banks. As Bookstaber put it, “one has to look beyond VaR, to culprits such as sheer stupidity or collective management failure: The risk managers missed the growing inventory [of risky assets], or did not have the courage of their conviction to insist on its reduction, or the senior management was not willing to heed their demands. In other words, models succeed because they meet the needs of real human beings, and VaR was just what they needed during the boom.

This, to me, is the same point that Lessig was making – technologically-induced simplicity (in the form of “money + politician = bribe” or VaR) is seductive and likely to be misinterpreted to the detriment of society.

Certainly some people understand this. Carl Malamud, who has led an impressive effort to opening up government, responded to Lessig’s article as such,

“Lessig’s point is that transparency, naked and by itself, with no broader and deeper aims, will not automatically produce good results, and may indeed produce randomness in our government or far worse. Merely revealing data is not enough. One must work with it, work with policy, and monitor effects. Transparency without a long-term commitment to policy is transparency without context, transparency that is merely naked…”

The parallel for Krugman’s world is, very likely, the work of behavioral economists who are placing humanity’s knack for irrational activity within the framework of economic thought. However, in his vehement response to Krugman’s essay, U Chicago professor John Cochrane writes,

“The sad fact is that few in Washington pay the slightest attention to modern macroeconomic research…”

This is what Lessig calls “the problem of attention-span,” and even were Krugman and Cochrane to coalesce into a sophisticated macroeconomic theory that took into account the limits of human quantification, I fear the simple, erroneous models will win the day (again).

Update: Tim Wu responds to Lessig’s piece with an important reminder that civic virtue is the key ingredient, not technology.

[Image Credit 1 and 2]