Microsoft's Massively Misplaced Incentives

Behavioral economics research explains what was so bad about Microsoft's employee ranking system.

Microsoft's decline from technology superpower tooffice utility has many causes, but none has received more focus than the management technique known inside the company as "stack ranking."

Everyone on every team is divided into three groups. A few are top performers, and are eligible for promotion and larger bonuses. Most are average. The rest are denied bonuses and often asked to leave. The performance reviews that determine these rankings occur every six months.

As Kurt Eichenwald explained in an article in Vanity Fair magazine, the system encourages workers to undermine each other, to avoid taking risks and to spend more time sucking up to bosses than working. Even worse, it removes many of the incentives for collaboration.

Despite these downsides, ranking systems are surprisingly popular atbig companies. That's all the more remarkable considering that research shows that performance-based pay is only good for people who are doingroutine tasks. That's because these workers don't benefit much from cooperating. Other workers, however, perform better when they aren't judged against their peers.

Economists used to think it was obvious that people would work better if you give them greater financial incentives to succeed, although psychologists knew that choking under pressure was a real phenomenon. High stakes often produce nervousness and anxiety. Think of all the public-speaking advice you hear about imagining audience members in their underwear. In 2005, several economists combined this intuition with someclever experiments and found that "the highest levels of rewards produced lower performance."

Even more relevant for Microsoft is a recentpaper that studies how teams perform under three different compensation plans: one where there were no bonuses for individual effort, one where there were small bonuses and one where each team member was paid almost exclusively for performance. They found that "the presence of a high relative reward seems to crowd out voluntary cooperation (i.e., the willingness to exert effort above the individual rational level)," although this is somewhat offset by the fact that each team member works harder to grab the biggest possible share of the bonus pool.

The upshot for Microsoft and other companies is that they could benefit from a more cooperative compensation model. Microsoft's new management might be able to realize the company'spotential with the help of some behavioral economics.

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

    To contact the author on this story:
    Matthew C Klein

    Before it's here, it's on the Bloomberg Terminal.