Uri Gneezy is a Professor of Economics and Strategy and the Epstein/Atkinson Chair in Behavioral Economics at UC San Diego Rady School of Management. Gneezy received his Ph.D. from the Center for Economic Research in Tilburg.
Below, Uri shares five key insights from his new book, Mixed Signals: How Incentives Really Work. Listen to the audio version—read by Uri himself—in the Next Big Idea App.
1. Incentives send signals.
When my son Ron reached the age that he was able to effectively communicate with us, like other kids, he also started to experiment with lying. We told him he shouldn’t lie and that being honest is what separates the good guys from the bad guys, but this moral lesson soon got me into trouble.
On a trip to Disney World one nice day in July, while in line waiting to pay, I noticed the sign that said that kids under three years old entered for free, whereas three-year-olds cost $117. When it was our time to pay, the smiling cashier asked me how old Ron was, and I replied, “almost three.” Technically I wasn’t lying, he was almost three, but from the wrong side; his third birthday was a couple of months earlier. I paid for myself while the cashier kept smiling, and we went in to have fun. What happened half an hour later was an important lesson. Ron said, “Daddy, I’m confused. You told me only bad guys lie, but you just did.” I tried the “do as I say, not as I do” approach, but I don’t think it worked.
Ron received two contradicting signals from me: what I said versus what I did in the face of a $117 incentive. In a nutshell, you must always try to avoid mixed signals. Ensure that you don’t say one thing, but when the incentives present themselves, do something else. The key is to understand that incentives send signals. Too often there’s a conflict between what you say and what your incentives signal. You can tell everyone that you care about honesty, but talk is cheap. For the claim to be credible, you need to back it up by taking actions that are costly to you, such as paying the full entrance fee. If you align what you say with the incentives you offer, the signal will be credible and easy to understand.
When you understand signals, you learn how to make incentives more effective. For example, what could Disney World do to reduce the number of people lying about their age? Disney could ask that the child be present when you buy the ticket. This would force you to face a mixed-signal dilemma like I did: Am I going to lie in front of my kid or pay $117? Disney could even take it a step further and ask the child to tell their own age, for example, by answering the question, “Did you already have your third birthday?” You can clearly coach your child to lie to the cashier, but boy, that is one strong and costly signal that lying is acceptable.
2. “Pay to Quit” Strategy: Encourage your employees to quit.
My day job is teaching negotiation. I tell my students they should never lie, with one exception: when interviewing for a new job. In such a negotiation you should act excited about the prospective company even if you’re not because your value to the company is higher when you are excited to work for them.
“The goal of this strategy is to learn which of your employees are fully committed to their work and encourage the others to quit.”
How can the employer tell who’s faking it? You need to create situations in which, instead of asking people questions that will probably yield answers meant to please you, they’ll be incentivized to reveal their true preferences. One such way is using a “Pay to Quit” strategy—offer your employees a financial incentive to leave their job voluntarily. You can tell them something like, “We’d like you to stay with us, but if you want to leave, we’ll give you a $10,000 gift and depart as friends.” The goal of this strategy is to learn which of your employees are fully committed to their work and encourage the others to quit.
It sounds crazy, right? But it’s a clever way of putting employees’ excitement to the test by forcing them to put their money where their mouth is. “Pay to Quit” might be enough to push some unhappy employees out the door in a friendly way, such that both the company and the employee are ending up on good terms. This reduces hard feelings and the risk of sabotage. It also benefits the company in two other ways, one, because the employees who turned down the temptations credibly signal that they are committed to the company. Two, the psychology of the situation suggests that sunken costs can affect subsequent behavior. By giving up the bonus, the employees signal to themselves that they are serious and motivated. They will feel a need to justify to themselves that it was worth it to give up the bonus and stay by working harder and being committed to a long-term goal.
3. Why incentives for recycling could backfire.
There are two types of signaling: social signaling, which is how incentives affect what others think of us, and self-signaling, what we infer about ourselves from our behavior. These two signals interact in quite interesting and surprising ways. Consider the following scenario: On a freezing morning, you see your neighbor Sara carrying a large bag full of cans to the recycling center. You observe Sara’s behavior and create a narrative: “Wow, Sara is awesome! She cares about the environment and is willing to sacrifice time and expend effort to help protect it.” Sara socially signals to others that she is environmentally conscious. In all likelihood, her trip to the recycling center also serves as a positive self-signal. She could have placed the cans in the trash, but instead, took the time and effort to recycle them in the bitter cold. She’s probably quite happy with herself.
Now, consider the same scenario, only this time there’s an incentive program aimed at encouraging people to recycle soda cans: for each soda can Sara recycles, she will receive five cents. Does your original narrative that Sara is awesome still work? On day one of my economics training, I learned that more money is better than less, hence paying Sara should make her happy and incentivize her to recycle more. What could be wrong with that?
“Sara is doing the exact thing you previously admired her for, only now she receives a couple of dollars in return.”
Well, along with creating a financial gain (the economic effect), the incentive also changed the signals and the story. Sara is doing the exact thing you previously admired her for, only now she receives a couple of dollars in return. So, instead of seeing her as a devoted environmentalist, you now see her as your cheap neighbor. In other words, the incentive changed the social signal that recycling sends and the presence of incentives may also change how Sara feels about herself. Rather than feeling good every time she makes a trip to the recycling center, she may now wonder if it’s worth going through all the trouble for a couple of bucks: the incentive changed the self-signal that recycling sends.
4. How to use the insights from recycling incentives to incentivize blood donation.
Let’s look at Jane. Jane donates blood and it makes her feel good about herself. She also enjoys some social signaling when she mentions her donation while having dinner with friends. The economics of blood donation are very interesting, because billions of dollars change hands in this market every year, but the donors are not paid. The demand is greater than the supply so the economic solution in such a case is simple right? Pay donors, and the supply will increase. If so, then why isn’t Jane paid something every time she donates blood? If Jane was offered $50 each time she donated blood, Jane would be all the happier to do it, right?
However, monetary compensation may alter the signal sent by donating blood. Jane enjoys the positive signal in the absence of any financial incentives, but offering her $50 changes everything. Over dinner with friends, she can no longer mention her nice behavior without her friends’ judgmental assumptions, because they’d think she must be underpaid or, worse yet, cheap to do something like that for $50. As in the recycling example, the incentive might even make Jane doubt her own motives for donating: is she doing it because she’s nice, or because of the money she gets?
“The economics of blood donation are very interesting, because billions of dollars change hands in this market every year, but the donors are not paid.”
An interesting alternative incentive is giving Jane a coffee mug with the blood bank logo. Jane doesn’t need another coffee mug; however, with this one, she can casually bring it out in a meeting, and more easily signal to people that she’s a blood donor. The coffee mug also enhances her self-signal. After all, it is nice to be reminded every time she picks up her coffee mug just how much of a good person she is.
5. How “Pay what you want” sends a unique signal that can benefit everyone.
Ayelet Gneezy, Gerhard Riener, Leif Nelson, and I studied “Pay what you want” (PWYW) pricing using a field experiment in a Pakistani buffet restaurant. When customers are done eating, they choose how much to pay for the meal; they can even pay nothing at all. Before you begin to seriously question the business model of PWYW, let me assure you, this pricing model works—at least to some degree.
Our field experiment had two main variations. In the first variation, called “observable,” customers paid for their meal directly to the waiter. In the second variation, called “unobserved,” customers put their payment into a sealed envelope and dropped it into a box near the entrance so that no one knew how much they paid.
A simple explanation of why customers might pay in the observable case is that they don’t want to look cheap in front of the waiter. But if that’s the case, why would they pay in the unobserved case? Signaling to others clearly cannot be at play. As a result, one might expect that if the customers only care about social signaling, they will pay nothing at all. But, what if they also care about their self-image? Eating a nice lunch and leaving without paying might make them feel bad. Therefore, for self-signaling, they may still pay even though their payment goes unobserved.
In this experiment, customers in the unobserved variant paid more than those in the observed variant; customers paid more when no one saw how much they paid than when the waiter saw. This result demonstrates that in the unobserved case, the self-signal became stronger. Putting money in the envelope is done only because I’m a nice guy, not because I care about what people think about me. As such, the boost to my self-image from giving money, the anonymous case is much stronger. We know how incentives send signals, now the key to successful incentives is mastering these signals.
To listen to the audio version read by author Uri Gneezy, download the Next Big Idea App today: