Disclaimer: This is my attempt to read The Logic of Life: The Rational Economics of an Irrational world, I am but someone who tries to synthesize understanding from ideas presented in this book.
There are many problems in the office that make it a living hell: back-stabbing colleagues, bullying bosses and all sorts of seemingly irrational behaviors.
All of the problems stem from the same root. In a company, you’d need information about the employees: Who are talented, and who are hard-working, who are honest; in order to pay them accordingly. But these virtues are very hard to measure, which is why the absurdities often result from plans that attempt to sideway this problem.
Author of the freakonomic book
To see that this underlies the bad behaviors observed in the workplace, we can look at the workplace that lack such undesirable behaviors. The author of the book freakonomics – which caused a global phenomenon, were paid by royalties. The publisher could have convinced him that the book wouldn’t have done well and offered him a big upfront – yet negligible to the actual revenue of that book. This would cut the cost of operation – per se, for the publishers, but it wouldn’t be rational.
This is because no one wants to hire a writer who is not interested in the success of the book. If an advance was offered, then hard-working and creativity doesn’t serve as a mean to maximize economic well-being. As a result, most likely the author would produce an average book. This is not desirable for the publishers, because parts of the book’s revenue become theirs.
The book freakonomics went on to become a global phenomenon and sold over 2,000,000 copies. The author took home a lot of moneys, because he was paid based on the royalties. Maybe his ideas presented in the book are controversial, but his paycheck is not. There is no controversy or need for explanations about why he’s getting all this money.
Windshield installation/replacement company
The workers were paid an hourly wage, regardless of their performance. New managers came in and decided that the performance was slow. They wanted to speed things up. They also decided that the workers are rational, and would install more windshields if they get paid more to do so, and if they are made to work to redeem their mistakes, they would avoid making mistakes.
This scheme increased the performance of the workers by 50%. The first reason is that workers work harder. The second reason is that the most competent workers tend to stay with the company, while incompetent workers tend to drift away, which improves company’s performance overall. The quality increased and the number of reported mistakes fell.
Performance pay works if performance is straight-forward to measure
This is because in this case, performance is very straight-forward to measure, therefore a performance pay works. It gets more tricky in certain office jobs. For example, to measure the performance of an audit, we need to have another audit to take a look, but who’s gonna measure the one who measures ?
Poor Measurements can be easily manipulated
Even when attempts are made, manipulation tailored for these measurements is often easy. If you r job is to respond to customer’s complaints, and if your KPI is that no customers should wait longer than 10 days. You would ignore those who haven’t reached the 7, 8th day because it wouldn’t mean much to your KPI, and only prioritize ones that are near the deadline. As a result, the response time slows down on average.
So your managers propose a second KPI which is the response time must not be lower than some particular number. So this time you prioritize customers with easy-to-resolve complaints, and those with serious issues which would take time to investigate and respond would never get resolved. The average response time goes up though.
So a third KPI is introduced which is to both improve response time and resolving most critical issues. You of course can achieve this by working overtime. But working overtime means that the company must pay you more, so a fourth KPI which restricts the overtime comes into play.
As a result, you basically won’t do anything, and be a grunt. Therefore, sometimes it’s just hard to know who’s performing well on the job.
But more often, we know who’s performing well on the job, but are unable to reward them. A certain supermarket was a subject of an economic experiment. The economists measured the performance of the cashiers through the number of ‘beep’ per second.
They wanted to know if workers are more productive when surrounded by productive colleagues. The answer is that, we do tend to be more productive when surrounded by productive colleagues only when the productive ones observe us. They note that performance went up when cashiers are being watched by more productive cashiers – by means of arranging the checkout aisles, and did not change when cashiers watch more productive cashiers.
This information could have been used to reward the more productive cashiers if the supermarket has switched to a piece rates model, where cashiers get rewarded by the number of pieces they check out. However, this plan was never put into motion, because the labor union resisted the piece rates, and because workers valued fixed hours more than flexible hours but higher salary. Another important reason is that measuring performance by the number of ‘beep’ indeed would have incentivized better performance, but at the cost of minimizing away variables like correctness of the scan, or the attitudes toward customers.
Objective performance measures are risky
Usually, managers know when people are lazy or productive, but such knowledge will not be put into an agreement or any documentation, except to be used in annual review, promotion and let-gos. As the saying goes: ‘You get what you measure’, measuring certain aspects of performance would encourage de-optimizing other factors. Some jobs are easy to see if they are well-done or not, but in other jobs not so straight-forward. For these jobs, the rational manager would look for a more ‘holistic’ approach.
Rationally, this means managers would base the decisions about what is good work like porn: They can’t define it, but they know it when they see it. This is quiet a sensible line of reasoning which would align well with what people see as well, just that it can’t be objectively explained.
However, this has a problem – which is that employees don’t trust managers who claim that they know what good work looks like but cannot explain why.
Solution – turn workplace into a tournament
The idea is to promise a million dollars reward to the most-outstanding employees. This builds trust with the employees because the program of rewarding the most-outstanding ones is official, and there is little room to escape the promise. By design, the managers don’t have to specify what ‘most-outstanding’ means. It’s just an idea floating around, it is what it is.
It is hard to measure objectively, but it’s relatively easier to compare one employee to another employee in terms of perceived performance. Perceptions are easy to be simplified, and unified among the employees regarding who’s better. The tournament is structured so that the managers can incentivized employees by promises of rewards, while keeping the options open.
However, it also explains the seemingly irrational behaviors
From an employee’s perspective, in order to be the outstanding one – there are two ways: To do a really good job, or to make other employees look bad in comparison. One study showed that more backstabbing behaviors were observed when the big raises are promise to the best workers. Backstabbing behaviors in this context are perfectly rational.
Some would say that tournaments fail to motivate workers. This is false, tournaments do encourage workers to be more productive, however it also encourages workers to make other workers look bad. Therefore, to realize a tournament, a manager would (only) need to calculate (informally) wether efforts of employees in improving his/her own performances would be more or less than efforts to denigrate other. It is a fairly realistic scheme.