TED – Ideas Worth Spreading
Professor David Laibson:
I’m going to tell you about some research that I did recently with a few collaborators – Keith Ericson, Sam McClure, George Loewenstein and Jonathan Cohen. We asked our subjects to not drink any fluids for three hours, and then when they came into the laboratory, we fed them potato chips. So they’re pretty thirsty at this point, and then we asked him a pretty devilish question: ‘Would you like one tiny sip of juice now, or two tiny sips of juice in five minutes?’ Think about that, try to put yourself in that position. What would you decide?
Our participants mostly want the earlier sip. They’re not willing to wait five minutes. We also asked them another kind of question in the same environment with the same level of thirst: How do you feel about the choice between one sip in twenty minutes, and two sips in 25 minutes? Same five minute delay, really the same question, but with this twenty minute add-on. Everything’s moved twenty minutes into the future. Now put yourself in that position. How do you feel in that case? Same preference for the earlier one, not willing to wait the extra five?
Very different this time. Now our participants say, “No, no, no – twenty versus twenty five minute wait, I am willing to wait the extra five minutes. I prefer the 25 minute double shot to the twenty minute single shot.
Now there’s lots and lots of ways in which the addition of a time horizon seems to fundamentally change the way we make decisions. Think about food, think about that chocolate soufflé – I feel the same way; I sit there at dessert and I say yeah, this chocolate soufflé is fantastic, I’m going to order it, but next week I’m definitely starting my diet. Or I think about entertainment, and I say yeah I’m exhausted, tonight I want to watch an action thriller sexy flick, I just can’t do anything serious, but next week I’m going to get that DVD off the shelf that I ordered from Netflix six months ago, you know the one about general relativity and string theory, I really want to watch that thing… next week.
So we’re full of these good intentions about what we’re going to do in the future, and they seem to fundamentally conflict with what we do right now. I want to talk about a way of explaining all that called present bias. Very simple mathematical theory of decision making in a time context. In this present bias world, we give full weight to rewards and costs that happen right now, and we give half weight to rewards and costs that happen in the future. For the future might even be a half hour away, certainly a day away, that’s the future. Full weight now, half weight on everything that’s not now.
Let’s play out present bias in the realm of exercise, a personal struggle. Imagine that exercise has an upfront cost of six, I hate exercising, I hate all that work, it’s effortful. But exercising is good for me, or at least so my spouse and physician tell me. There’s a delayed health benefit. Let’s say the benefit is eight, but it comes later in time, I’ll look better, I’ll feel better, over the next few years.
Well the world where I’m discounting everything that comes later by a half, I don’t want to exercise now, immediate cost of six, delayed benefit of eight, discount the benefit by a half, discounted total benefit times two. Because the eight gets discounted by a half. But how do you feel – how do I feel? Standing here today, planning to exercise tomorrow. There’s no upfront cost. Now the cost of exercise tomorrow is not in the present and the benefit in the even more delayed future is not in the present, so cost and benefits are on the same level playing field. So what I am going to do when I stand here today and I make my plan for tomorrow? Well there’s no upfront cost or benefit, that’s a zero, there are delayed costs of six, delayed benefit of eight, they both get discounted by a half – net benefit to me, plus one. Yes! I will exercise tomorrow. That’s my life.
About three years ago, my wife and I went to a health club and she talked us both into joining. We paid the health club one thousand dollars each for dual one-year memberships, two thousand dollars in total. And as I was signing the paperwork, giving them my credit card number, I believed I was going to exercise about twice a week for the next 52 weeks of membership. A few of you here know me, you can guess how often I exercised over the next 52 weeks. Dean, what do you think? [Laughing] Excellent, he guessed five, that’s exactly right. I went five times in 52 weeks, for a per-visit cost of 200 dollars per visit. [Laughing] My wife went a lot more often, maybe she went 12 times.
Let’s think about this example in the context of some real evidence. So Stefan and studied gyms in the Boston area. And they actually measured attendance, they had swipe card data, they compared that to the cost of being a member of these gyms, average membership cost 75 dollars per month. Average number of visits, four visits per month. Average cost per visit, 19 dollars per visit; kind of high, but not necessarily ridiculous. But the real interesting element of this study was the alternative to being a member. You could simply pay on a per-visit basis, and never pay a membership fee. So what was the cost of paying on a per-visit basis to these gyms? Just ten dollars a visit. So there seems to be a profound self-control problem when it comes to food, when it comes to exercise… what about savings, the topic of today’s event? Well here too, there seems to be something going on, it’s about the gap between our good intentions and our actual actions.
So here’s a study that I did with James Choi, and Brigitte Madrian and Andrew Metrick. We surveyed workers at a white collar firm and we asked them about their savings adequacy, are they saving the right amount, are they saving too much, are they saving too little? About two out of three workers say they’re saving too little, and about 24 out of 100 say that they’re going to raise their savings rate in the next two months. That’s their plan. But we actually had access to their administrative records. We knew exactly what they did. Now we gave them the benefit of the doubt; we didn’t check for four months. They plan was over a two month window, we gave them four months and we checked in. How many of these workers actually raised their savings rate over the next four months? Any guesses?
Three, on average out of a hundred. An awareness that there’s probably a savings problem, a lot of good intentions about changing the savings behavior, and terrible follow-through.
Now luckily we’ve got some solutions, we’ve got a lot of different strategies that nudge people to better outcomes. I want to talk about that now and set up a kind of a schematic. So let me begin by talking about an undesired behavior, like non-participation in a savings plan. And contrast that with a desired behavior, like being in a savings plan, joining your 401(k) or contributing to your IRA. There’s a barrier – all that effort involved in reading the prospectus, reading the paperwork, calling the human resources department, signing up, figuring it out, contemplating your future, thinking about it all. We create a lot of costs, I mean they’re not huge now, about two hours of work for a typical American household, but it is two hours of work. And in a world where people have present bias, where they want to delay work, they want to push stuff off to next week, you’ve got a cycle of procrastination. Why join today, I’ll join the savings plan next week. What’s the big deal, I’ll lose a week of savings. That’s nothing in a lifetime of work. But of course, every week it postpones again. So the obvious solution is to not set up the savings system like this; to switch the location of the good outcome, participation. So automatic enrollment or opt-out systems do exactly that. They put the good outcome to the left of the hurdle – now you’re automatically in the savings plan – and if you don’t like it, opt out. But now procrastination is working in your favor; it’s holding you inside the plan, you’ve got to overcome procrastination to get out.
Now that’s well and good if you’ve got an ounce of paternalism in your gut and you’re willing to kind of stack the deck in this way. What if you’re a libertarian and you don’t really want to create an unlevel playing field and you’ve got this unfortunate situation where we’ve got the bad outcome as the starting point and the good outcome as the thing that’s behind the barrier with a little bit of work. What could you do in this case? You’re a libertarian, you don’t want to advantage one or the other. Well, something natural really. Require that people make an active choice. Put both participation and non-participation on a level playing field. Move them both to the left of the hurdle. And now tell people, ‘You’ve got to choose’. Just like your health insurance, you’ve got to choose. You come into our firm, you’ve got to tell us whether you want or don’t want health insurance. A non-response is not an option. It’s too important a decision to let passivity create an outcome for you. To let procrastination dictate what happens. We’re going to put both options on the left of the hurdle, we’re going to require you to actively choose.
There’s yet a third way to approach this problem. We’ve got this barrier, two hours of paperwork and forms and prospectus, things that we might understand barely but that the typical American household has a hard time understanding. So what are you going to do? Well, something kind of obvious. Take two hours of work and turn it into a one minute check a box exercise. If they want to they can do all that work, or, here’s a postcard that has 27 words on it – check this box and we’ll enroll you with a five percent savings rate and we’ll put you in a diversified mutual fund. Drop it in the mail, it’s pre-posted. One minute of work. Eliminate the barrier that’s standing in the way of people not doing the right thing.
So how do all these systems play out in reality? We have to measure these things. So let’s take a look at the data. And here I’m going to aggregate across many, many different firms. At this point, hundreds of firms. When people are told, it’s up to you to opt in, you begin outside the plan, you’ve got to opt in, if you want to save, typically there’s 40 percent participation at the end of one year of work. When you make enrollment simple, quick enrollment; they’re still outside the plan but they can check a box to get in, now enrollment goes up to 50 percent. When you tell them passivity is not an option, you must choose, either choose to actively be in or actively be out, but you must make an active choice, you cannot do nothing, you must fill in the form. Tell us one way or the other what you want, now we get 70 percent enrollment. And finally, when you put people in the plan and tell them, if you don’t like it, opt out, now you get 90 percent enrollment.
In principle none of this should matter. But of course it makes a night and day difference in the real world. These tiny little barriers, these tiny little costs of where you start, what the paperwork is, makes all the difference in a decision that’s going to be one of the five most important decisions you’ll make in your life – how much to save for retirement.
Now I want to go back to the earlier example about exercise and talk about the tension between what we want to do and what we actually do. I want to think about this topic of commitment. How do we commit ourselves, are we willing to commit ourselves, do we recognize these self-control problems? I want to talk about a study that was actually motivated by some of the work that Dean Karlan did. This is related study that discusses the way to design commitment contracts that people want to use. In this study, jointly authored with John Beshears, James Choi, Brigitte Madrian and John Sakong, we gave our subjects a hundred dollars, and we asked them to divide that money across two accounts: a freedom account that works like any other account, perfectly liquid, and a goal account that says up to a particular goal date your money is basically illiquid; you can get it, maybe, but there’s an element of restrictiveness. And after the goal date, you can do whatever you want with it. Both accounts have the exact same interest rate. Well, from a classical perspective, why would you ever put your money in the more restrictive account? That makes no sense. Let’s see what our participants did in the study.
So first, we set up a goal account that had a 10% penalty for early withdrawal. That sounds familiar. What is that account? It’s the 401 (k), it’s the modern IRA, the way that 11 trillion dollars in the US economy is currently being saved. And in that situation 35% of the money in our study actually ended up in the goal account. Our participants were willing to tie their hands in that way and voluntarily accept a 10 percent penalty for early withdrawal. And here there’s no tax benefit, here there’s no max, so they’re just tying their hands gratuitously.
Now, in the next part of the study, a different set of subject pulled from the same pool, we increase the penalty. We made the penalty 20 percent for early withdrawal. What happened now? How would you respond? Higher penalty, less access to your money. Does that make you want to put your money into this account more, or less? Yeah, as an economist I kind of feel that it’s got to be less. But of course, the real world doesn’t exactly sync with economists. Now 43 % of the money goes into the goal account.
And then we did a third condition, where we had the goal account be completely illiquid. Absolutely no access to your money, you can plead, you can cry, we don’t care your house gets foreclosed, the money is locked up til the goal date. How much money do you want to put into the goal account? Again the economist says, well nothing really. Our subjects now put 56 percent of their money into the goal account. People seem to know that they have a bit of a self-control problem. People seem to want to tie their own hands sometimes.
Now this study doesn’t answer the really important policy question of what we should do with 11 trillion dollars of defined contribution money that’s now in accounts basically at 10 percent penalty for early withdrawal. But it does raise the question of whether that 10 percent penalty, which was probably set randomly in the 1970s when these accounts were created, by some staffer late at night who was thinking about these things being for high level executives, it suggest that we should revisit that calculation.
Now I don’t know whether a ten percent penalty is the right way to go in the IRA – DC world, or whether a “lockbox” IRA is the right way to go. But I do know that we should think about it a lot harder. And I also know that everyone’s different. To get me to exercise, I need a tennis partner waiting for me on the court, and I’ll show up. Other people will go just by being a member of the health club. Other people need a personal trainer. Everyone needs their own form of commitment to get them to exercise. And yet, while we admit that in the realm of exercise, and while we admit that we need to have 200 flavors of ice cream, and 50 kinds of coffee, we only have one flavor of commitment in the world of retirement savings. The single IRA, 401 (k), with a ten percent penalty for early withdrawal. We should think a lot harder about increasing the set of choices that enables us to tie our own hands in the ways that suit each of us, instead of putting everyone in the same, identical, homogeneous, plain vanilla product. What would the world look like if people had a set of choices that gave the flexibility that they probably want to tie their hands when it comes to saving in the way that works best for them.