When questioned, the vast majority of U.S. workers say that the number one factor determining their pay is their individual performance. I know, because in a recent survey I asked over 1,000 of them. Among a list of factors including seniority, experience, education and company profits, nothing outranks individual performance in workers’ accounts of why they get paid what they do: Fully two-thirds believe it is a very important determinant of the number on their paycheck.
Despite this widely held belief, most of us aren’t paid based on our individual performance, and the fraction that is has actually declined over time. Research finds that after increasing during the 1980s, jobs with performance-related pay structures peaked in the early 2000s and fell in the following two decades. In a 2009 review of a variety of incentive pay practices, a team of researchers found that “relatively few workers have pay that varies in a direct formulaic way with their productivity, and that the share of such workers is probably declining.” Even among workers whose pay is partly tied to measures of individual productivity, that share is usually a small portion of total compensation.
What accounts for the disconnect between what workers believe influences their pay and what actually does? Three factors stand out: fairness concerns, the difficulty in measuring individual performance in many jobs, and disagreements over what performance actually means.
In 2008, the Sacramento Bee created a website publicizing the pay of state employees. A team of researchers took advantage of the sudden disclosure to track how workers reacted. Those who used the website and discovered they were underpaid compared with similarly situated peers expressed anger and engaged in more job searching—predictable enough. Workers who discovered they were overpaid, however, experienced no symmetrical emotional response. They weren’t suddenly more loyal to their bosses or more satisfied in their workplaces. Their behavior didn’t change at all.
Why? A longstanding finding in social psychology and economics is that most of us think we’re pretty good at what we do—indeed, better than most of our colleagues. It’s a real-world version of the fantasy town of Lake Wobegon in Prairie Home Companion, a place “where all the children are above average,” only here we’re talking about adults.
Our overestimation of ourselves goes way back. One survey from the 1960s asked professionals and managers how their performance measured up against their peers. The mean self-rating was the 77th percentile, meaning that, on average, respondents rated their performance as better than three-quarters of their co-workers. Only two of the 100 or so respondents said that their performance was below average.
That’s a mathematical impossibility, and it poses a conundrum for employers who would like to differentiate pay based on a measure of productivity. Such a move would upset many workers who incorrectly believe that they are more productive than their peers. As a result, most employers don’t try it, thus avoiding the fairness concerns that arise when workers think that they deserve more than their colleagues.
Paying for individual performance also assumes that reliable measures of performance exist. In some fields they do: Trucking, for instance, involves moving as many goods as possible from point A to point B, quickly and safely. But in most other jobs, there are no comparable measures.
Take your own job: Is there a widely agreed-upon metric that captures your productivity? The difficulty in measuring one’s individual contribution is especially acute in the collaborative knowledge-worker jobs that have expanded over the past decades, such as management consultants, market researchers and midlevel managers of all kinds. There are millions of these jobs today, and in each of them distilling individual performance into one quantifiable metric exceeds our capabilities—not because we haven’t discovered the right measurement but because no such measurement exists.
Even in occupations where a general standard of performance is widely agreed upon, measurement can be difficult. For physicians, the goal is to improve patients’ health. But does that mean doctors should be paid on the basis of their patients’ overall health, rather than for performing a specialized task correctly, such as a hip replacement? What if there were other, less invasive options available, but the physician only received payment for major surgery? Is that how we want to structure the incentives for our healthcare system?
In the U.K. under Tony Blair’s government, the National Health Service increased compensation to hospitals that reduced their wait times. It’s true that nobody likes sitting around the doctor’s office waiting to be seen, so wait time, while not a direct measure of patient health, does capture an important component of patient well-being. But an issue quickly arose: In an effort to hit waiting time targets, hospitals held patients in ambulances before transferring them to waiting rooms. Why? Time spent in an ambulance didn’t count as waiting, according to this particular performance measure. As a result of these complications, performance measures for doctors differ markedly both across and within national healthcare systems.
The definition of productivity in many jobs involves choices and trade-offs, and there is no such thing as a single, objective measure awaiting discovery. In 2014, Time Inc., which then owned over 90 publications, began to evaluate its journalists’ performance at producing content that was “beneficial to advertiser relationship.” The reaction, not only among Time Inc. employees but across the industry, was dismay, with journalists upset that their craft could be reduced to a measure so crass. But how should one measure journalistic productivity? Stories filed per week? Words per hour? What about time spent cultivating sources for deep reporting?
Definitional disputes roil occupation after occupation. Some journalists, for instance, would say that the true purpose of journalism is to produce award-winning investigations, while others believe that generating revenue is the core mission. In my own field, higher education, some professors prioritize teaching while others believe a university rises or falls on its research. In other words, definitional disputes often stem from disagreements about what an organization’s “product” should be. If we disagree about our organization’s product, we’ll disagree about each worker’s individual contribution to it, so compensating workers based on one productivity measure will upset others who are wed to an alternate vision.
Fairness concerns, imperfect measures and definitional disputes: As a result of these factors, most of us aren’t actually paid for our individual performance, despite what we tend to think.
