Of course, there are other techniques of control, used by both organizations and individuals, which involve keeping track of expenditures. In organizations, these techniques are called accounting. Similarly, as we saw earlier, Humans use mental accounting, with the help of envelopes, mason jars, and retirement savings plans, to accomplish the same purpose. Notice that the failure to treat various pots of money as fungible, as Econs would do, is what makes such accounting strategies feasible.
I should stress that Shefrin and I did not think that there were actually two different people inside your head. Ours is an “as if” model that is meant to provide a useful way of thinking about self-control problems. We did include a footnote in our second paper noting that one could think of the planner as residing in the prefrontal cortex region of the brain, which is associated with conscious, rational thinking, whereas the doer can be associated with the limbic system. For those who are familiar with the two-system model such as the one Kahneman describes in Thinking, Fast and Slow, it is reasonable to think of the planner as the slow, reflective, contemplative System 2 while the doers are the fast, impulsive, intuitive System 1. Recent research in neuro-economics offers some support for this interpretation. But for practical purposes, it does not matter whether the model has a physiological basis. It is a metaphor that helps us think about how to incorporate self-control into economics.
I still find the planner–doer model the most useful way to think about self-control problems, but it has not proven to be the favorite formal model of the next generation of behavioral economists. David Laibson, a behavioral economist at Harvard, pioneered what has turned out to be the model of choice in his PhD dissertation, published in 1997. Two other behavioral economic theorists, Matthew Rabin and Ted O’Donoghue, elaborated on this approach, which most economists now just refer to using the two Greek letters that represent the important variables: beta (β) and delta (δ). The subtleties of the model are difficult to explain without going into some detail, but references to the key papers are provided in the endnotes. The crucial advantage that the beta–delta model has over the planner and the doer is mathematical simplicity. It is the smallest possible modification of Samuelson’s basic model that can capture the essential aspects of self-control.
Here is a simple way of thinking about how the beta–delta model works. Suppose that for any time period far enough away to be considered “later,” a person does not discount time at all, meaning that the discount rate is zero. But anything considered “now” is privileged and tempting, and anything considered “later” is worth only half as much. In the Wimbledon example discussed earlier, the first-round match that would be valued at 100 this year would only be worth 50 next year or any year thereafter. Such preferences are “present-biased” since they put so much weight on now versus later, and they lead to time-inconsistent choices.
Even in this highly simplified version of the model, it is possible to illustrate many interesting subtleties about intertemporal choice; these subtleties depend in part on whether people are aware of their self-control problems. When David Laibson wrote his first paper on this subject he assumed that agents were “sophisticated,” meaning that they knew they had this pattern of time preferences. As a graduate student trying to get a job with a paper on behavioral economic theory (a category that was then essentially unknown), it was clever of David to characterize the model this way. David’s agents were pure Econs except for one detail; they had problematic time preferences. When O’Donoghue and Rabin decided to join the party they considered a more radical approach, in which agents have present-biased preferences but are unaware of their affliction. Such agents are considered “naïve.”
Not surprisingly, neither of these simple formulations portrays a fully accurate description of behavior. I share a view held by all three authors that the “truth” is somewhere in between the two extremes: partial naiveté. Most of us realize that we have self-control problems, but we underestimate their severity. We are naïve about our level of sophistication. In particular, we suffer from what George Loewenstein has called “hot-cold empathy gaps.” When we are in a cool, reflective mood—say, contemplating what to eat at dinner on Wednesday after just having finished a satisfying brunch on Sunday—we think we will have no trouble sticking to our plan to eat healthy, low-calorie dinners during the week. But when Wednesday night comes along and friends suggest going out to a new pizza place featuring craft beers, we end up eating and drinking more than we would have predicted on Sunday, or even on Wednesday before arriving at the restaurant with its tempting aromas wafting from the wood-burning oven, not to mention an intriguing list of special brews to sample. For such cases we may need a planner to have established a rule—no midweek beer and pizza outings—and then to think of a way of enforcing that rule.
In the time since I first removed that bowl of cashews, behavioral scientists have learned a lot about self-control problems. This knowledge is proving important in dealing with many of society’s biggest problems, as we will see later on.
________________
* In truth, Odysseus was not clever enough to think up this plan himself. He got some good advice from Circe, a goddess who specialized in herbs and drugs. Go figure.
† Some researchers have tried a version of the marshmallow/Oreo experiment on animals. Most go for the immediate reward, but one particularly clever African gray parrot named Griffen was shown to display better self-control than most preschoolers (Zielinski, 2014).
‡ The two-system model articulated by Kahneman in Thinking, Fast and Slow was not the way he and Tversky originally thought about their research. One of Danny’s main reasons for writing the book was because he thought recasting their original work using the framework of a fast, automatic system and a slow, reflexive system offered an insightful new perspective on their earlier findings.
§ Tom Schelling started writing on this topic soon after I did. Our views were very much in the same spirit, but he was less convinced than I was that the longsighted set of preferences is more likely to be “right.” See, for example, Schelling (1984).
¶ Amos always referred to the planner as a female. I will do likewise in his honor. And since men are generally more like doers than women are, I will use the male pronoun for doers. Just call me a sexist.
# For the sake of simplicity, I will ignore the possibility that the one-energy-bar-per-day diet makes the doers increasingly hungry as time passes.
** Such technology does exist. The Kitchen Safe (kitchensafe.com) is a plastic container than the user can lock for any period of time. The manufacturers recommend it for anything tempting, from candy to smartphones to car keys. I received one from a thoughtful student. Naturally it was filled with cashews. In a world of Econs, there would be no demand for such a product.
INTERLUDE
13
Misbehaving in the Real World
If behavioral economics is supposed to offer a more realistic description of how people behave, then it should be helpful in practical settings. Although most of my time early in my career was devoted to academic research about mental accounting and self-control, I did have the occasional opportunity to venture out into the real world, or as close as one can get to that in Ithaca. I soon found that these ideas had practical business applications, especially relating to pricing. Here are two examples.
Greek Peak
At Cornell I got to know a student, David Cobb, who encouraged me to meet his brother, Michael. A local to the area and an avid skier, Michael was determined to make a career in the ski business and had landed a job as marketing director at Greek Peak, a family-owned operation near Ithaca. At the time, the resort was in serious financial difficulties. A few winters with less than the usual amount of snow and a tough economy had created a situation where the company had to borrow heavily to get through the off-season, and this was at a time when interest rates were high, even for good credit risks, which Greek Peak was not. The resort simply had to increase revenues a
nd decrease debt, or it would go bankrupt. Michael needed help and suggested a barter exchange. He would give me and my kids lift tickets and set the kids up with ski equipment. In return, I would try to help him get the business back in the black.
It quickly became apparent that Greek Peak would have to increase prices if they were going to turn a profit. But any increase large enough to generate a profit would put their ticket prices nearly on par with well-known ski resorts in Vermont or New Hampshire. Operational costs per skier were not much different than they were at those bigger resorts, but Greek Peak had only five chair lifts and less skiable terrain. How could we justify charging a similar price to the larger resorts, and do so without significantly reducing the number of skier visits? And how could we retain the price-sensitive local market, including students at Cornell and other nearby colleges?
In mental accounting terms, the lift ticket prices of the famous Vermont ski resorts would be a salient reference point for Greek Peak customers, and they would expect to pay significantly less since the product was distinctly inferior. What Greek Peak had going for it was proximity. It was the nicest place to ski in central New York, and getting to Vermont was a five-hour drive. Greek Peak was also the closest option for people living due south, including in Scranton, Philadelphia, and even Washington DC. Busloads of skiers would arrive from these cities every weekend.
I urged Michael to rethink Greek Peak’s revenue model, making use of principles from behavioral economics. The first problem to solve was how to raise the ticket price without losing too many customers. We adopted a plan of gradually raising the price over a period of years, thus avoiding a sudden jump that might create backlash. To partially justify the higher prices we tried to improve the skier experience, to make the purchase seem less of a rip-off.* I remember one early idea I had along these lines. There was a short racecourse on the side of one of the trails, where a skier could run through a series of slalom gates and receive an official time that was broadcast over loudspeakers. Younger skiers enjoyed the competitive aspect of this, and the gates were close enough together that speeds were safe. The price charged to use the racecourse was one dollar. A dollar was not a lot to pay, but the fee was a damn nuisance. Getting access to your money while on a ski hill is a pain. You have to take off your thick, clumsy gloves and dig down to whichever layer you are keeping your money in. Then, in this case, you had to feed a one-dollar bill into a vending machine–style slot. Given how well those machines work in the best of circumstances, you can imagine the failure rate once exposed to the elements.
I asked Michael and the owner, Al, how much money they were making from the racecourse. It was a small amount of money, perhaps a few thousand dollars a year. Why not make this free, I asked? We can improve the skier experience at a trivial cost. This was a no-brainer. And it got Michael and Al thinking about other things they could do to improve the quality and, importantly, the perceived value of their product.
Another example involved ski instructors. The instructors’ main business was teaching new skiers, especially groups of schoolkids—obviously an important way to grow the customer base. But instructors had a lot of downtime. Someone got the clever idea to set up a free ski clinic on the mountain. A skier would wait at a designated spot on the trail, and then ski though a few gates with the action captured on video. An instructor stationed at the bottom would show the skier a replay of the video and offer a few pointers. “Free lessons!”
Even if these enhancements were making higher lift ticket prices more palatable, we still had to worry about the price-sensitive local market. Here we had a nice existing model to work from. The resort offered university students a package of six weekday lift tickets at a heavily discounted price if purchased by October 15. These were popular and provided a good source of early revenue. I suspect the students also liked the fact that the deal was called a six-pack. Even subtle beer references appeal to the college crowd.
We wondered whether we could offer something like the six-pack to the local non-student market as well. The goal was to offer the locals a deal that would not be available to the out-of-town skiers who drove in once or twice a year. To these skiers, the price of the lift ticket was only a small portion of the trip’s expense, which included transportation, food, and lodging. A few dollars more for the lift ticket was unlikely to sway the decision about whether to make the trip, especially given the lack of nearby competition. We ended up with a solution called the ten-pack. It included five weekend tickets and five weekday tickets and was sold at 40% off the retail price when purchased by October 15.
Ten-packs turned out to be wildly popular with the locals. There are a few behavioral factors that explain their popularity. The first is obvious: 40% off sounds like a great deal. Lots of transaction utility. Second, the advance purchase decoupled the purchase decision from the decision to go skiing. As with wine mental accounting, the initial purchase could be viewed as an “investment” that saves money, making a spur-of-the-moment decision to go skiing on a sunny Friday after a recent snowfall costless to implement. That the customer may have gone out for a nice dinner the previous weekend would not put the recreation mental account in the red; the skiing was “free.” And from the resort’s point of view, it was better than free—it was a sunk cost.† As the season progressed, skiers would be eager to use some of their tickets to avoid wasting the money invested in the ten-pack, and they might bring along a friend who would pay full price. (The tickets were not transferrable.)
Ten-packs also were popular because skiing is one of those activities that people resolve to do more of next year. “Last year I only got out three times, which is ridiculous given that Greek Peak is so nearby. This year, I am going to take off a few days from work and go when it isn’t crowded.” As with paying for a gym membership to encourage more exercise, the skiers’ planners liked the idea of committing to ski more often this winter. Buying the ten-pack was a good way to do that and save money at the same time.
After a few years, six-packs, ten-packs, and season passes accounted for a substantial portion of the resort’s revenue, and this early money eliminated the need to borrow to stay afloat until the start of the season in December. Selling all these tickets in advance also hedged against a warm winter without much snow. While ski resorts can make snow, it has to be cold enough for the machines to work. Also—and this drives ski resort owners crazy—even if it has been cold, if there is no snow on the ground in town, people are less likely to think about going skiing, regardless of conditions at the resort.
After three years of selling the ten-packs, Michael did some analysis and called me with the results. Recall that ten-packs were sold at just 60% of the regular season retail price. “Guess what percentage of the tickets is being redeemed?” Michael asked. “Sixty percent!” The resort was selling the tickets at 60% of the retail price but only 60% of them were being redeemed. In essence they were selling the tickets at full price and getting the money several months earlier: a huge win.
This outcome did not seem to upset the clientele, most of whom repurchased ten-packs the following year. Even those who did not use many of their tickets would blame themselves, not the resort. Of course, there were customers that would end up with nearly all their tickets unused at the end of the season. Some would ask, hopefully, whether they could use the tickets the following season. They were politely told no, the tickets were explicitly sold as being good only this year. But Al designed a special offer for these customers. They were told that if they bought a ten-pack again this year, their unused tickets from the previous year would remain valid. Of course a customer who only went skiing two or three times last year is unlikely to go more than ten times this year, but the offer sounded good. Although I don’t think many people were foolish enough to buy another ten-pack simply for this reason, they did seem to appreciate that the resort was making an effort to be “fair,” something we will soon see can be important to keeping customers happy.
A final p
ricing challenge for Greek Peak was to figure out what to do early in the season, when, shortly after the first snowfall, the resort would open but often with only one lift running. Avid skiers who had been waiting since the previous March would show up for the first runs of a new season. What price should they be charged? Al’s policy had been to look out his window at the mountain and the weather, and then tell the ticket sellers the price, often half off the regular price. Of course, most of the skiers who arrived had no idea what the price would be; they only knew the retail price. Only true diehards might have been able to unravel Al’s pricing strategy for the early season. I call this a “secret sale.” A customer comes up to the cash register prepared to pay retail and the seller says, “Oh, that item is on sale for 50% off.” It might generate goodwill, but it is not a brilliant pricing strategy, because the customer was ready to pay the full price. Reducing the price only makes sense if it increases current sales or perhaps future sales by building customer loyalty.
Michael and I came up with a new strategy. Early in the season, or for that matter, any time only part of the mountain was open for skiing, pricing followed a set formula. Skiers would pay full price to ski that day, but would get a coupon good for up to 50% off their next visit, depending on how many chair lifts were operating. Since customers were expecting to pay full price, this offer seemed generous, and the coupon might induce them to come back, and perhaps buy lunch and a beer as well.
Michael once told me a story that captures how popular these coupons were. A guy shows up for his first ski outing of the year and has picked up a brand new ten-pack. He is standing in line to exchange one of these coupons for a lift ticket and overhears the ticket seller explain to the customer in front of him that she will get a 50%-off coupon that she can use toward her next purchase. This sounds so good to him that he puts the ten-pack back in his pocket and shells out for a full-priced ticket. I have always wanted to know whether he used that half-off coupon before he finished his ten-pack. We will never know.
Misbehaving: The Making of Behavioral Economics Page 13
