Perverse Incentives

By , May 7, 2010

“Perverse incentives” is a phrase I seem to use quite often, lately. We create (or accept) the perverse incentives that create the perverse outcomes we detest.

Wikipedia defines the term as:

A perverse incentive is an incentive  that has an unintended and undesirable effect, that is against the interest of the incentive makers. Perverse incentives by definition produce negative unintended consequences.

Most recently, I’ve used the term to describe the bizarre incentives created under the “No Child Left Behind” law (and the current “Race to the Top” program).  The problem here is that the intent of the law is to improve public education, but because the law focuses on a very narrow range of “easily-testable” subjects, it actually degrades public education.  Since the law “grades” schools and teachers based solely on standardized math and reading test scores, it creates a strong incentive to focus on the “tested aspects” of those two subjects, to the exclusion of all other curriculum.  The result is an increase in standardized test scores, but declines in performance under all other measures of educational quality or outcomes.

In the context of the “mortgage meltdown,” the employees of banks and mortgage brokerage firms were evaluated exclusively on the number or dollar value of mortgages issued, not on quality or any reasonable estimate of the financial results in “the long term” (at one time, we’d think that “the long term” meant the lifetime of the mortgages, but within the financial industry the term was reinterpreted to mean a period of ten years, then five years, then three years, and finally “a few quarters”).   Reckless agents who “played the game” (by issuing “no-doc” or “liar loans,” or even by fraudulently altering mortgage-application documents) earned huge fees (which they were never asked to repay, even after the federal government bailed out the industry). Many of their more ethical and honest colleagues earned less — or were fired for their “poor performance” and their unwillingness to join the “team effort.”

Of course, I’ve also encountered many perverse incentives in the “internet industry,” dating back to the reckless days when companies that had never earned any revenue (much less any profit) were rewarded with initial public offerings (IPOs) in which the company was was valued at billions of dollars.  Since investors were looking only for “the next new thing,” without considering any aspects of the company’s financial prospects, executives at internet firms abandoned “profit” or “revenue” as meaningful benchmarks, and instead looked toward other measures, such as “traffic” or “adviews,” sometimes assigning absurd values (for example, using valuations of $3 to $10 per “visitor” on a site whose traffic generated no income). Of course, if “visitors” were assigned a value of $10 each by Wall Street, then it seemed logical to spend millions of dollars to attract new visitors, even if the “new visitors” acquired for $5 each were very different from the earlier visitors to the site.  (In the end, many firms found that they couldn’t generate revenue to justify a value of even one cent per visitor.)

Historically, advertising agencies charged fees based on a percentage of the dollar amount spent on the advertising which the agencies bought.  If the agency could persuade the client to spend more on advertising, the agency’s fees increased, even if the advertising didn’t increase the client’s sales or profits.  Even after technology existed to monitor and evaluate the performance of internet advertising, many advertising agencies have continued to command fees based primarily on “amount spent,” providing no incentive to improve the effectiveness of advertising — nor even to seek out efficient ways to monitor performance.

I always resist “perverse incentives.” In the late 90’s, I refused to work for clients who succumbed to “perverse incentives,” and in the past decade, I’ve refused to work for clients who won’t set up systems to evaluate the performance of their advertising.  When I worked as a high-school teacher (briefly), I sought to teach what students needed to learn, without focusing exclusively on the subjects they’d be “tested on,” and I feel some regret about the classroom time that I spent preparing my sophomore students for the “High School Exit Exam.”  And I now realize that my unwillingness to accept “perverse incentives” was a big factor in my career decisions as an attorney.

Evaluating “performance” is usually a very complex and difficult task.  Too often, we seek shortcuts by focusing on a subset of the work being evaluated, and based on a subset of outcomes that are “easy to measure,” even if we all agree that the result doesn’t fairly consider all of the aspects of the work (nor even the aspects of the work we consider most important).

We create (or accept) the perverse incentives that create the perverse outcomes we detest.

perverse incentive is an incentive that has an unintended and undesirable effect, that is against the interest of the incentive makers. Perverse incentives by definition produce negative unintended consequences.

Leave a Reply

*

OfficeFolders theme by Themocracy