WEBVTT 00:00:00.000 --> 00:00:02.410 [SQUEAKING] 00:00:02.410 --> 00:00:04.338 [RUSTLING] 00:00:04.338 --> 00:00:05.784 [CLICKING] 00:00:10.604 --> 00:00:12.100 FRANK SCHILBACH: Welcome, everyone. 00:00:12.100 --> 00:00:15.120 This is 14.13 Psychology and Economics, also known 00:00:15.120 --> 00:00:16.980 as Behavioral Economics. 00:00:16.980 --> 00:00:18.570 My name is Frank Schilbach. 00:00:18.570 --> 00:00:21.750 I'm a faculty member in the Economics Department, 00:00:21.750 --> 00:00:24.450 teaching and doing research in Behavioral Economics 00:00:24.450 --> 00:00:28.440 and Psychology and Economics and Development Economics. 00:00:28.440 --> 00:00:32.189 There's a syllabi over here in case anybody needs one. 00:00:32.189 --> 00:00:35.140 So let me sort of start with introducing ourselves, 00:00:35.140 --> 00:00:36.180 including myself. 00:00:36.180 --> 00:00:38.070 As I said, I'm Frank Schilbach. 00:00:38.070 --> 00:00:42.480 I'm an Economics PhD from Harvard recently. 00:00:42.480 --> 00:00:45.840 I'm from Germany in case-- as you may have noticed. 00:00:45.840 --> 00:00:48.115 I do research at the intersection of development 00:00:48.115 --> 00:00:48.990 behavioral economics. 00:00:48.990 --> 00:00:51.720 In particular, I'm interested in integrating 00:00:51.720 --> 00:00:54.570 psychological issues and helping us understand 00:00:54.570 --> 00:00:57.450 better the lives of the poor. 00:00:57.450 --> 00:01:00.270 And so I study all sorts of issues related to poverty, 00:01:00.270 --> 00:01:03.450 how poverty itself affects people's behavior, 00:01:03.450 --> 00:01:05.230 and how conditions of poverty, or things 00:01:05.230 --> 00:01:07.230 that are associated with poverty might feed back 00:01:07.230 --> 00:01:11.340 into people's decision-making and their productivity or labor 00:01:11.340 --> 00:01:12.420 market behaviors. 00:01:12.420 --> 00:01:15.660 And then perhaps, sort of lead to the persistence of poverty 00:01:15.660 --> 00:01:17.410 through those kinds of effects. 00:01:17.410 --> 00:01:19.740 So I have some work on financial constraints, 00:01:19.740 --> 00:01:22.920 how financial constraints affect people's behavior, in terms 00:01:22.920 --> 00:01:24.708 of thinking about money itself. 00:01:24.708 --> 00:01:26.250 I have some work on sleep deprivation 00:01:26.250 --> 00:01:27.450 among the urban poor. 00:01:27.450 --> 00:01:29.280 I'm thinking about pain and substance abuse 00:01:29.280 --> 00:01:32.100 and how that might affect people's choices. 00:01:32.100 --> 00:01:35.555 Most recently, I'm interested in things 00:01:35.555 --> 00:01:37.680 related to mental health, in particular, depression 00:01:37.680 --> 00:01:40.320 and loneliness, how those might affect people's well-being, 00:01:40.320 --> 00:01:44.130 and then their behaviors. 00:01:44.130 --> 00:01:48.752 I have office hours where you can sign up on my websites. 00:01:48.752 --> 00:01:50.460 Currently, they're on Tuesday afternoons, 00:01:50.460 --> 00:01:52.560 but if you cannot find a time set work for you, 00:01:52.560 --> 00:01:55.110 please email me and we'll find some time that works out. 00:01:55.110 --> 00:01:56.700 My assistant is Krista Moody in case 00:01:56.700 --> 00:01:58.530 you want to sort of get a hold of me 00:01:58.530 --> 00:02:00.660 and I'm not available for some reason, 00:02:00.660 --> 00:02:03.010 you can reach out to her. 00:02:03.010 --> 00:02:05.500 So now, let me give you sort of the overview of the class. 00:02:05.500 --> 00:02:07.982 I'm going to start sort of with four things. 00:02:07.982 --> 00:02:09.690 One, first, I'm going to sort of tell you 00:02:09.690 --> 00:02:11.770 what is this class, what is behavioral economics, 00:02:11.770 --> 00:02:12.990 or what is psychology economics? 00:02:12.990 --> 00:02:13.890 What do we mean by that? 00:02:13.890 --> 00:02:15.848 Is it sort of just putting two fields together, 00:02:15.848 --> 00:02:17.430 or is this something more specific, 00:02:17.430 --> 00:02:18.955 which I'll argue it is. 00:02:18.955 --> 00:02:20.580 I'm going to give you some example then 00:02:20.580 --> 00:02:24.450 about how you might use behavioral economics, 00:02:24.450 --> 00:02:27.150 and a specific example and a policy that 00:02:27.150 --> 00:02:28.260 might affect all of you. 00:02:28.260 --> 00:02:30.510 And in fact, it will because we have a specific laptop 00:02:30.510 --> 00:02:32.008 policy in this class. 00:02:32.008 --> 00:02:34.300 And I'll tell you a little bit how to think about this. 00:02:34.300 --> 00:02:35.967 There will also be a problem set that'll 00:02:35.967 --> 00:02:37.920 help you think about this a bit more. 00:02:37.920 --> 00:02:40.930 Then we talk about somewhat boring, of course, logistics. 00:02:40.930 --> 00:02:43.320 And at the end, we do a questionnaire, a quiz, 00:02:43.320 --> 00:02:45.157 for everybody, in part, because we 00:02:45.157 --> 00:02:46.740 want to know some information for you. 00:02:46.740 --> 00:02:48.115 In part, we want to sort of learn 00:02:48.115 --> 00:02:51.450 a little bit about some decisions that you might make. 00:02:51.450 --> 00:02:52.200 OK. 00:02:52.200 --> 00:02:53.690 Great. 00:02:53.690 --> 00:02:56.100 Sorry, this is-- 00:02:56.100 --> 00:02:56.600 OK. 00:02:56.600 --> 00:02:58.460 So what is psychology economics? 00:02:58.460 --> 00:02:59.870 So you may have heard of this. 00:02:59.870 --> 00:03:02.540 The class used to be called economics and psychology, 00:03:02.540 --> 00:03:04.340 or behavioral economics. 00:03:04.340 --> 00:03:07.500 For our purposes, that's kind of like all the same. 00:03:07.500 --> 00:03:10.970 Some economists sort of argue that all economics 00:03:10.970 --> 00:03:11.970 is about behavior. 00:03:11.970 --> 00:03:15.877 So like behavioral economics is sort of a bit of a weird word. 00:03:15.877 --> 00:03:17.960 So that's why partially I'm using sort of the word 00:03:17.960 --> 00:03:19.340 psychology and economics. 00:03:19.340 --> 00:03:21.830 In some sense, it's sort of broader than that. 00:03:21.830 --> 00:03:25.280 One definition is, "it's a field of academic research 00:03:25.280 --> 00:03:26.930 that studies the joint influences 00:03:26.930 --> 00:03:30.320 of psychological and economic factors on behaviors." 00:03:30.320 --> 00:03:31.890 You could be sort of broader and say 00:03:31.890 --> 00:03:33.890 we're trying to integrate insights from not just 00:03:33.890 --> 00:03:36.320 psychology, but also anthropology, sociology, 00:03:36.320 --> 00:03:39.950 medicine, psychiatry, et cetera, and so on, into economics. 00:03:39.950 --> 00:03:43.370 And trying with an attempt to make economic models more 00:03:43.370 --> 00:03:46.490 realistic, and therefore, more predictive and help us 00:03:46.490 --> 00:03:48.230 understand people's behavior better. 00:03:48.230 --> 00:03:51.170 And then help us make better policies, perhaps, 00:03:51.170 --> 00:03:54.740 in trying to influence people's behavior. 00:03:54.740 --> 00:03:58.100 Now, as I said, that's medicine, sociology, et cetera. 00:03:58.100 --> 00:04:00.350 Broadly, we're trying to use insight from those fields 00:04:00.350 --> 00:04:02.267 and try and understand how we missed something 00:04:02.267 --> 00:04:04.470 by making the fairly stark assumptions 00:04:04.470 --> 00:04:07.610 that economic models usually make. 00:04:07.610 --> 00:04:10.505 That leads me to what are then the standard economics models? 00:04:10.505 --> 00:04:13.130 In some sense, what we're going to try and study to some degree 00:04:13.130 --> 00:04:15.260 is deviations from the classical, 00:04:15.260 --> 00:04:17.362 or sort of the standard economic model. 00:04:17.362 --> 00:04:19.279 So then, of course, we need to understand what 00:04:19.279 --> 00:04:20.571 is the standard economic model? 00:04:20.571 --> 00:04:23.877 To start with, which leads me to impart some prerequisites 00:04:23.877 --> 00:04:25.460 of the class, or you should have taken 00:04:25.460 --> 00:04:27.050 at least some economics or-- 00:04:27.050 --> 00:04:29.150 some economics to start with because I'm 00:04:29.150 --> 00:04:32.030 going to talk a lot about sort of deviations from those models 00:04:32.030 --> 00:04:32.660 and sort of-- 00:04:32.660 --> 00:04:35.660 if you haven't taken any class before, 00:04:35.660 --> 00:04:37.760 don't really know that very well, 00:04:37.760 --> 00:04:39.350 then understanding the deviations 00:04:39.350 --> 00:04:41.840 will be a little bit tricky. 00:04:41.840 --> 00:04:44.730 So for those of you who have taken economic classes, 00:04:44.730 --> 00:04:46.860 what do you usually assume about people's behavior? 00:04:46.860 --> 00:04:48.568 What are some of the assumptions that you 00:04:48.568 --> 00:04:50.540 make about economic behavior? 00:04:52.940 --> 00:04:53.440 Yes. 00:04:53.440 --> 00:04:55.755 AUDIENCE: Stable preferences. 00:04:55.755 --> 00:04:58.230 FRANK SCHILBACH: Stable preferences, yes. 00:04:58.230 --> 00:05:00.270 So first, well-defined preferences 00:05:00.270 --> 00:05:03.090 in the sense of like people know what their preferences are. 00:05:03.090 --> 00:05:04.350 They can state them. 00:05:04.350 --> 00:05:06.840 And then they're also stable in the sense of when 00:05:06.840 --> 00:05:11.190 I ask you today would you rather have apples or bananas, 00:05:11.190 --> 00:05:13.200 that will not change unless there's 00:05:13.200 --> 00:05:16.060 sort of new information or your circumstances change. 00:05:16.060 --> 00:05:18.060 Of course, if I ask you tomorrow and you already 00:05:18.060 --> 00:05:21.210 have a bunch of apples, and then you'd say you want bananas, 00:05:21.210 --> 00:05:23.355 that doesn't mean your preferences of changed. 00:05:23.355 --> 00:05:25.980 But if I ask you about tomorrow, what would you like for lunch, 00:05:25.980 --> 00:05:28.620 you tell me a choice, you want apples rather than bananas. 00:05:28.620 --> 00:05:31.080 And then tomorrow, you show up and nothing else has changed 00:05:31.080 --> 00:05:32.460 and you say, now, I want bananas. 00:05:32.460 --> 00:05:35.280 Well then, your preferences are not stable. 00:05:35.280 --> 00:05:36.433 That's one assumption, yes. 00:05:36.433 --> 00:05:37.350 What else do you have? 00:05:37.350 --> 00:05:38.010 Yeah. 00:05:38.010 --> 00:05:40.315 AUDIENCE: More broadly, they're rational. 00:05:40.315 --> 00:05:41.190 FRANK SCHILBACH: Yes. 00:05:41.190 --> 00:05:43.581 And what does that mean? 00:05:43.581 --> 00:05:45.850 AUDIENCE: They behaved deterministically 00:05:45.850 --> 00:05:47.397 to optimize some utility function. 00:05:47.397 --> 00:05:48.730 FRANK SCHILBACH: Right, exactly. 00:05:48.730 --> 00:05:51.750 So people essentially optimize some utility function. 00:05:51.750 --> 00:05:53.830 So we say here, people have the utility function 00:05:53.830 --> 00:05:55.080 and know what that is. 00:05:55.080 --> 00:05:56.830 And you're sort of saying they maximize it 00:05:56.830 --> 00:05:58.220 to optimize in a certain way. 00:05:58.220 --> 00:06:01.910 They don't make mistakes in that maximization process. 00:06:01.910 --> 00:06:06.880 So that's to say if you tell me you like apples over bananas, 00:06:06.880 --> 00:06:09.172 and then you choose bananas, well then, something 00:06:09.172 --> 00:06:10.630 is sort of going wrong in some ways 00:06:10.630 --> 00:06:11.890 that we haven't fully understood. 00:06:11.890 --> 00:06:13.473 It might be that you sort of like just 00:06:13.473 --> 00:06:16.660 making mistakes that could be construed as irrational. 00:06:16.660 --> 00:06:19.160 It's hard to rationalize with a model. 00:06:19.160 --> 00:06:20.490 What else? 00:06:20.490 --> 00:06:21.780 AUDIENCE: Self-interest. 00:06:21.780 --> 00:06:23.330 FRANK SCHILBACH: Self-interest, yes. 00:06:23.330 --> 00:06:26.860 So like usually in a lot of models all like this sort of-- 00:06:26.860 --> 00:06:29.980 the easiest model is a very narrowly defined self-interest 00:06:29.980 --> 00:06:31.900 is what people care about themselves. 00:06:31.900 --> 00:06:34.030 They care about what they consume 00:06:34.030 --> 00:06:35.590 and not about what others consume, 00:06:35.590 --> 00:06:38.210 or what others think of them. 00:06:38.210 --> 00:06:42.250 So sort of narrowly defined self-interest. 00:06:42.250 --> 00:06:42.750 Yes? 00:06:42.750 --> 00:06:44.333 AUDIENCE: People have the self-control 00:06:44.333 --> 00:06:46.638 to [? smooth their ?] consumption over time. 00:06:46.638 --> 00:06:47.640 FRANK SCHILBACH: Yes. 00:06:47.640 --> 00:06:49.440 So essentially, perfect self-control 00:06:49.440 --> 00:06:52.003 usually is sort of one version of putting that. 00:06:52.003 --> 00:06:54.420 One version of putting that is in some sense sort of going 00:06:54.420 --> 00:06:56.170 back to what you saying earlier is sort of 00:06:56.170 --> 00:06:57.480 to say stable preferences. 00:06:57.480 --> 00:07:00.090 If I'm telling you I'd like to exercise tomorrow, 00:07:00.090 --> 00:07:02.670 and that's my preference, tomorrow I'm 00:07:02.670 --> 00:07:05.790 not going to be like, oh, yeah, actually, I changed my mind 00:07:05.790 --> 00:07:08.190 and now I'm just watching movies, or the like, right? 00:07:08.190 --> 00:07:10.770 So that's the version of sort of preferences being stable. 00:07:10.770 --> 00:07:13.020 But the inherent underlying issue there 00:07:13.020 --> 00:07:15.450 is like self-control problems in the sense of like, 00:07:15.450 --> 00:07:17.100 if I like certain things, or make 00:07:17.100 --> 00:07:20.190 certain plans for the future, I have the self-control 00:07:20.190 --> 00:07:21.682 to follow through on those plans. 00:07:21.682 --> 00:07:22.890 That's usually an assumption. 00:07:22.890 --> 00:07:27.470 And that then shows up as preferences being stable. 00:07:27.470 --> 00:07:27.970 What else? 00:07:27.970 --> 00:07:28.470 Yeah. 00:07:28.470 --> 00:07:30.251 AUDIENCE: They prefer consumption today 00:07:30.251 --> 00:07:31.334 than consumption tomorrow. 00:07:31.334 --> 00:07:33.143 [INAUDIBLE] 00:07:33.143 --> 00:07:34.060 FRANK SCHILBACH: Yeah. 00:07:34.060 --> 00:07:36.880 So usually, there is like a discount factor 00:07:36.880 --> 00:07:39.490 in terms of how much you like consumption today 00:07:39.490 --> 00:07:41.890 versus consumption tomorrow. 00:07:41.890 --> 00:07:43.900 That is usually in any economic model. 00:07:43.900 --> 00:07:47.350 Usually, we think of like the discount factor being constant, 00:07:47.350 --> 00:07:49.510 as like how you think about today versus tomorrow 00:07:49.510 --> 00:07:52.060 is the same as you think about today versus in two 00:07:52.060 --> 00:07:54.920 days from now, or like a year from now and a year and a day 00:07:54.920 --> 00:07:55.420 from now. 00:07:55.420 --> 00:07:58.030 So usually, there's a constant discount factor for the future. 00:07:58.030 --> 00:08:02.260 Usually, we sort of employ economic-- sorry-- 00:08:02.260 --> 00:08:06.280 constant discounting or exponential discounting. 00:08:06.280 --> 00:08:07.960 What else do we have? 00:08:07.960 --> 00:08:08.460 Yeah. 00:08:08.460 --> 00:08:10.260 AUDIENCE: Risk aversion. 00:08:10.260 --> 00:08:11.950 FRANK SCHILBACH: Yes, risk aversion 00:08:11.950 --> 00:08:16.260 in the sense of, I guess, people are risk averse, 00:08:16.260 --> 00:08:19.747 and people often sort of define their preferences over outcome. 00:08:19.747 --> 00:08:22.080 So they have a utility function that we tend to concave, 00:08:22.080 --> 00:08:23.880 and the utility function is-- and we'll 00:08:23.880 --> 00:08:26.200 get to this in a few lectures. 00:08:26.200 --> 00:08:28.290 But the utility function is concave 00:08:28.290 --> 00:08:31.800 and it's defined over outcome as opposed to over changes 00:08:31.800 --> 00:08:32.490 of outcomes. 00:08:32.490 --> 00:08:34.710 And that shows up usually as sort of like essentially 00:08:34.710 --> 00:08:38.460 like a strictly concave utility function, 00:08:38.460 --> 00:08:42.330 and that's risk aversion. 00:08:42.330 --> 00:08:45.543 What about information, how do people think about information? 00:08:45.543 --> 00:08:46.710 How do they use information? 00:08:46.710 --> 00:08:47.827 AUDIENCE: Bayesian. 00:08:47.827 --> 00:08:48.910 FRANK SCHILBACH: Bayesian. 00:08:48.910 --> 00:08:49.950 And what does that mean? 00:08:49.950 --> 00:08:51.420 AUDIENCE: They update [? their prior. ?] 00:08:51.420 --> 00:08:52.462 FRANK SCHILBACH: Exactly. 00:08:52.462 --> 00:08:56.293 So people essentially are perfect information processors. 00:08:56.293 --> 00:08:57.460 You can call them Bayesians. 00:08:57.460 --> 00:08:59.204 But essentially, it's sort of-- for those of you 00:08:59.204 --> 00:09:01.287 know that-- it's sort of did they use Bayes' rule, 00:09:01.287 --> 00:09:03.280 essentially, which is to say if you gave 00:09:03.280 --> 00:09:06.940 a statistician a problem how to update information, 00:09:06.940 --> 00:09:09.220 people are able to do that in their heads. 00:09:09.220 --> 00:09:10.832 And that's a pretty stark assumption. 00:09:10.832 --> 00:09:12.790 But essentially, it's like when you people have 00:09:12.790 --> 00:09:16.660 certain information, standard economics does not necessarily 00:09:16.660 --> 00:09:19.050 assume that everybody has full information in a sense 00:09:19.050 --> 00:09:20.740 that everybody knows everything. 00:09:20.740 --> 00:09:23.030 But if you give them some information, 00:09:23.030 --> 00:09:26.165 then they update their beliefs accordingly, right? 00:09:26.165 --> 00:09:28.540 So they-- essentially, they optimally use new information 00:09:28.540 --> 00:09:30.957 and then form their beliefs, their posterior beliefs based 00:09:30.957 --> 00:09:34.780 on having new information and what they believe previously, 00:09:34.780 --> 00:09:35.950 their priors. 00:09:35.950 --> 00:09:37.450 There's another assumption, usually, 00:09:37.450 --> 00:09:38.867 and I sort of can actually sort of 00:09:38.867 --> 00:09:42.880 put these out for you, which is, there's another assumption. 00:09:42.880 --> 00:09:46.360 That is, people have no taste for beliefs or information. 00:09:46.360 --> 00:09:47.360 What do we mean by that? 00:09:47.360 --> 00:09:49.030 We mean by that is, essentially, people 00:09:49.030 --> 00:09:51.800 use information only to make decisions. 00:09:51.800 --> 00:09:53.198 So if you tell me something about 00:09:53.198 --> 00:09:54.740 like what's going to happen tomorrow, 00:09:54.740 --> 00:09:57.310 or if you tell me something about my health status, 00:09:57.310 --> 00:10:00.320 or the like, I use that to make better decisions. 00:10:00.320 --> 00:10:03.820 So if you tell me I'm sick, I'm going to use that information 00:10:03.820 --> 00:10:05.730 to go to the doctor. 00:10:05.730 --> 00:10:07.910 But people usually don't assume that people 00:10:07.910 --> 00:10:09.660 have certain beliefs about whether they're 00:10:09.660 --> 00:10:11.940 sick or healthy, or whether they're smart or not, 00:10:11.940 --> 00:10:13.500 or they are good looking or not. 00:10:13.500 --> 00:10:16.230 Usually, the assumption that beliefs are only 00:10:16.230 --> 00:10:19.050 used-- information is only used to make better decision. 00:10:19.050 --> 00:10:21.795 Not that people get utility from beliefs. 00:10:21.795 --> 00:10:23.170 But instead, you might say, well, 00:10:23.170 --> 00:10:25.770 I'm really good looking and smart, 00:10:25.770 --> 00:10:28.047 and maybe smarter than you actually are. 00:10:28.047 --> 00:10:29.880 And the reason why people might believe that 00:10:29.880 --> 00:10:32.640 is because it makes them feel good about themselves. 00:10:32.640 --> 00:10:38.370 That's usually an assumption that standard economic models 00:10:38.370 --> 00:10:39.005 make. 00:10:39.005 --> 00:10:40.630 I think we mentioned almost everything. 00:10:40.630 --> 00:10:42.450 We sort of said Bayesian information processor-- 00:10:42.450 --> 00:10:44.010 essentially processing information 00:10:44.010 --> 00:10:47.130 optimally-- well-defined and stable preferences, 00:10:47.130 --> 00:10:50.250 maximize expected utility, which is sort of like, in some sense, 00:10:50.250 --> 00:10:54.090 rationality if you want, apply exponential discount weighting 00:10:54.090 --> 00:10:56.190 current and future well-being. 00:10:56.190 --> 00:11:00.790 People are narrowly defined in terms of their self-interest. 00:11:00.790 --> 00:11:04.270 And people have preferences for final outcomes, not changes. 00:11:04.270 --> 00:11:06.900 So what you care about is how the weather is today, not 00:11:06.900 --> 00:11:10.710 kind of how the weather changed between yesterday and today 00:11:10.710 --> 00:11:12.003 or tomorrow. 00:11:12.003 --> 00:11:13.420 And we'll talk about all of these. 00:11:13.420 --> 00:11:15.690 By the way, these are all sort of some terms that 00:11:15.690 --> 00:11:17.837 seem maybe unfamiliar to you. 00:11:17.837 --> 00:11:19.920 We'll talk about all of these kinds of assumptions 00:11:19.920 --> 00:11:23.350 and perhaps how to deviate from them specifically. 00:11:23.350 --> 00:11:23.850 OK. 00:11:23.850 --> 00:11:26.970 So now, that we have those assumptions, now 00:11:26.970 --> 00:11:29.820 what kinds of deviations when you look into the world-- 00:11:29.820 --> 00:11:31.350 and sort of one part how you think 00:11:31.350 --> 00:11:33.568 about behavioral economics is, in some sense, 00:11:33.568 --> 00:11:35.610 think about the world, and try to sort of observe 00:11:35.610 --> 00:11:36.900 what's going on in the world. 00:11:36.900 --> 00:11:38.850 And try to see which assumptions might be 00:11:38.850 --> 00:11:41.670 violated in an important way. 00:11:41.670 --> 00:11:43.685 And try to sort of improve our models. 00:11:43.685 --> 00:11:45.060 So when you think about the world 00:11:45.060 --> 00:11:46.643 and look at these assumptions, can you 00:11:46.643 --> 00:11:50.760 come up with real world examples that might sort of violate 00:11:50.760 --> 00:11:54.310 these assumptions? 00:11:54.310 --> 00:11:56.200 Yes. 00:11:56.200 --> 00:11:58.310 AUDIENCE: The last assumption that you 00:11:58.310 --> 00:12:03.100 listed, the one of taste and beliefs or information 00:12:03.100 --> 00:12:05.950 is violated on a constant basis, people 00:12:05.950 --> 00:12:08.870 tend to believe what they want to believe. 00:12:08.870 --> 00:12:12.080 FRANK SCHILBACH: And do you have an example, a specific? 00:12:12.080 --> 00:12:15.924 AUDIENCE: This is very common as it regards political matters. 00:12:15.924 --> 00:12:17.591 I'm not going to name a particular topic 00:12:17.591 --> 00:12:20.350 because obvious reasons. 00:12:20.350 --> 00:12:27.310 But people often choose to discount information that 00:12:27.310 --> 00:12:28.920 goes against their beliefs. 00:12:28.920 --> 00:12:31.200 FRANK SCHILBACH: Right. 00:12:31.200 --> 00:12:33.580 So one reason could be-- and there's quite a bit of work 00:12:33.580 --> 00:12:34.080 recently. 00:12:34.080 --> 00:12:36.330 We'll talk a little bit about in this class about sort 00:12:36.330 --> 00:12:38.910 of how political beliefs or other reasons 00:12:38.910 --> 00:12:41.910 why people might be motivated to believe certain things. 00:12:41.910 --> 00:12:45.292 One clear example would be climate change, not just 00:12:45.292 --> 00:12:46.500 because of political reasons. 00:12:46.500 --> 00:12:48.210 There's some interesting work about, 00:12:48.210 --> 00:12:50.263 when people live in flood areas, for example, 00:12:50.263 --> 00:12:51.930 what do they think about climate change? 00:12:51.930 --> 00:12:54.450 And you might sort of say, well, if you really 00:12:54.450 --> 00:12:57.150 live in an area that's potentially affected by floods, 00:12:57.150 --> 00:12:58.650 you might really want to know about climate change 00:12:58.650 --> 00:13:01.108 and know what's going on and really try to inform yourself. 00:13:01.108 --> 00:13:03.690 But what people tend to do is sort of try to ignore the issue 00:13:03.690 --> 00:13:07.590 and try to sort of be happy as long as sort of nothing 00:13:07.590 --> 00:13:08.590 happens. 00:13:08.590 --> 00:13:11.980 Now, that's one example, yes. 00:13:11.980 --> 00:13:13.510 Yes. 00:13:13.510 --> 00:13:15.850 AUDIENCE: Some purposes might not be well-defined. 00:13:15.850 --> 00:13:19.180 So say, I have a decision on what I want to eat. 00:13:19.180 --> 00:13:22.390 I want to eat steak over chicken if I'm 00:13:22.390 --> 00:13:24.550 presented with the opportunity. 00:13:24.550 --> 00:13:28.900 But I prefer chicken over pork, but I might else 00:13:28.900 --> 00:13:30.142 prefer pork over steak. 00:13:30.142 --> 00:13:31.100 FRANK SCHILBACH: Right. 00:13:31.100 --> 00:13:33.120 AUDIENCE: So it's not always well-defined. 00:13:33.120 --> 00:13:33.420 FRANK SCHILBACH: Right. 00:13:33.420 --> 00:13:35.420 So one part that you're telling is essentially-- 00:13:35.420 --> 00:13:38.340 so one part might be it might not just be well-defined. 00:13:38.340 --> 00:13:40.210 There's other issues, like, for example, 00:13:40.210 --> 00:13:42.990 that irrelevant other-- there's an assumption economics. 00:13:42.990 --> 00:13:45.480 Usually, as they say, if you choose between A and B, 00:13:45.480 --> 00:13:47.880 the availability of C should not matter 00:13:47.880 --> 00:13:51.370 for your choice between A and B. But often, that's not the case. 00:13:51.370 --> 00:13:52.890 So if I offer you apples and bananas 00:13:52.890 --> 00:13:54.570 and you also can get cherries, now 00:13:54.570 --> 00:13:56.580 suddenly your choice between apples and bananas 00:13:56.580 --> 00:13:59.510 might change even if you can't get the cherries-- 00:13:59.510 --> 00:14:03.060 sorry, even if you don't choose the cherries eventually. 00:14:03.060 --> 00:14:03.690 Exactly. 00:14:03.690 --> 00:14:05.482 There might be sort of peoples' preferences 00:14:05.482 --> 00:14:07.510 might not be well-defined. 00:14:07.510 --> 00:14:09.930 They might also not be stable. 00:14:09.930 --> 00:14:10.790 Any other example? 00:14:10.790 --> 00:14:11.730 Yes. 00:14:11.730 --> 00:14:14.160 AUDIENCE: Also, for the well-defined preferences, 00:14:14.160 --> 00:14:16.290 we have information processing constraints. 00:14:16.290 --> 00:14:19.830 If I'm given a menu of 100 choices, 00:14:19.830 --> 00:14:23.947 it's going to be difficult for me to know which one. 00:14:23.947 --> 00:14:25.280 FRANK SCHILBACH: Right, exactly. 00:14:25.280 --> 00:14:29.800 So people have lots of information around themselves. 00:14:29.800 --> 00:14:33.420 And if you think about to process that information, 00:14:33.420 --> 00:14:36.570 if you go to a supermarket, it's impossible to know 00:14:36.570 --> 00:14:39.160 all the prices, all the goods, and make all these choices. 00:14:39.160 --> 00:14:41.400 So in some sense, there's an abundance of information 00:14:41.400 --> 00:14:42.180 everywhere. 00:14:42.180 --> 00:14:47.140 And we have to sort of figure out how to deal with that. 00:14:47.140 --> 00:14:47.980 Any other example? 00:14:47.980 --> 00:14:49.450 Yes. 00:14:49.450 --> 00:14:55.010 AUDIENCE: Some people who might have preferences [INAUDIBLE] 00:14:55.010 --> 00:14:57.986 equal an outcome of getting $1,500 00:14:57.986 --> 00:15:00.130 would be different if you were initially 00:15:00.130 --> 00:15:03.637 promised $1,000 versus if you [INAUDIBLE].. 00:15:03.637 --> 00:15:04.970 FRANK SCHILBACH: Right, exactly. 00:15:04.970 --> 00:15:09.370 So one part is to say people have preferences 00:15:09.370 --> 00:15:11.170 of a final outcome, not changes, which 00:15:11.170 --> 00:15:16.180 is to say you might feel about the weather in certain ways, 00:15:16.180 --> 00:15:19.300 you know, depending on how the weather was yesterday. 00:15:19.300 --> 00:15:23.140 Another version of that would be you evaluate the outcomes 00:15:23.140 --> 00:15:24.968 that you get based on your expectations. 00:15:24.968 --> 00:15:26.510 And sort of if you thought, you know, 00:15:26.510 --> 00:15:30.285 today is going to be really nice or your day is going to be 00:15:30.285 --> 00:15:32.410 really good, but then it happened to be not as good 00:15:32.410 --> 00:15:35.350 for whatever reason-- the weather is bad or something bad 00:15:35.350 --> 00:15:37.300 happen to you today-- 00:15:37.300 --> 00:15:39.948 the typical assumption would be just say, well, 00:15:39.948 --> 00:15:41.740 you should just evaluate the final outcome. 00:15:41.740 --> 00:15:44.157 It shouldn't matter what you thought about previously what 00:15:44.157 --> 00:15:45.130 might happen. 00:15:45.130 --> 00:15:47.200 You're going to look at what happens at the end. 00:15:47.200 --> 00:15:52.160 And that's how you evaluate your well-being. 00:15:52.160 --> 00:15:53.816 Yes. 00:15:53.816 --> 00:15:56.940 AUDIENCE: You might need a commitment device, 00:15:56.940 --> 00:15:59.360 so that you-- 00:15:59.360 --> 00:16:01.644 you might avoid buying potato chips 00:16:01.644 --> 00:16:03.075 because you think you'll eat them. 00:16:03.075 --> 00:16:04.506 That didn't really make sense. 00:16:04.506 --> 00:16:05.102 [INAUDIBLE] 00:16:05.102 --> 00:16:06.060 FRANK SCHILBACH: Right. 00:16:06.060 --> 00:16:08.580 So people might on purpose restrict their choices 00:16:08.580 --> 00:16:11.160 in certain ways, right? 00:16:11.160 --> 00:16:12.660 And this is what gets me eventually 00:16:12.660 --> 00:16:14.770 to the laptop policy. 00:16:14.770 --> 00:16:16.337 But people might sort of essentially 00:16:16.337 --> 00:16:18.420 know that in the future they make certain choices. 00:16:21.580 --> 00:16:23.940 For example, in the laptop case, you might say, 00:16:23.940 --> 00:16:26.100 I'm going to use a laptop in class 00:16:26.100 --> 00:16:29.160 with all the best intentions to taking notes and paying 00:16:29.160 --> 00:16:29.850 attention. 00:16:29.850 --> 00:16:31.350 But you know, at the end of the day, 00:16:31.350 --> 00:16:33.600 you know, you're going to watch the football review 00:16:33.600 --> 00:16:34.800 from yesterday or whatever. 00:16:34.800 --> 00:16:36.030 Other stuff comes up. 00:16:36.030 --> 00:16:38.620 You're going to chat with your friends and so on and so forth. 00:16:38.620 --> 00:16:40.703 So now, in some sense, if you're as sophisticated, 00:16:40.703 --> 00:16:43.920 as in if you know what you're going to do in case 00:16:43.920 --> 00:16:45.930 you sort of make certain choices, you might say, 00:16:45.930 --> 00:16:51.000 I choose not to even allow myself to use a laptop at all. 00:16:51.000 --> 00:16:54.270 The reason being because I know that essentially I'm going 00:16:54.270 --> 00:16:55.930 to misbehave in the future. 00:16:55.930 --> 00:16:57.750 And that's we'll talk about this soon, 00:16:57.750 --> 00:17:00.000 I think in lecture three or four, which is essentially 00:17:00.000 --> 00:17:01.500 people's demand for commitment, as we 00:17:01.500 --> 00:17:03.292 call it, which is to say people have demand 00:17:03.292 --> 00:17:05.462 for restricting their choice. 00:17:05.462 --> 00:17:07.920 And in any neoclassical model, this doesn't make any sense. 00:17:07.920 --> 00:17:10.121 Because you're going to choose optimally anyway. 00:17:10.121 --> 00:17:12.329 And you're going to make the best choice for yourself 00:17:12.329 --> 00:17:12.829 in any case. 00:17:12.829 --> 00:17:13.920 That's an assumption. 00:17:13.920 --> 00:17:16.770 So why would you ever sort of shut down certain choices? 00:17:16.770 --> 00:17:18.420 Because, for example, in an emergency, 00:17:18.420 --> 00:17:19.619 or there might be something really important 00:17:19.619 --> 00:17:21.150 coming up in class where you might 00:17:21.150 --> 00:17:23.849 want to use your laptop regardless of what's happening. 00:17:23.849 --> 00:17:25.720 Why would you shut that down? 00:17:25.720 --> 00:17:26.220 Exactly. 00:17:26.220 --> 00:17:30.010 So demand for commitment is another example. 00:17:30.010 --> 00:17:31.490 Yes. 00:17:31.490 --> 00:17:35.720 AUDIENCE: People are also not anywhere near perfect Bayesian 00:17:35.720 --> 00:17:39.440 information processors either. 00:17:39.440 --> 00:17:43.040 Or as a general rule, people are pretty 00:17:43.040 --> 00:17:50.360 bad at updating their beliefs based on new information. 00:17:50.360 --> 00:17:53.750 And we frequently traffic in 100% and 0% 00:17:53.750 --> 00:17:59.930 certainties very often, which far more often than what 00:17:59.930 --> 00:18:03.542 you would expect from a Bayesian information processor. 00:18:03.542 --> 00:18:04.500 FRANK SCHILBACH: Right. 00:18:04.500 --> 00:18:05.670 AUDIENCE: Practically not [INAUDIBLE].. 00:18:05.670 --> 00:18:06.712 FRANK SCHILBACH: Exactly. 00:18:06.712 --> 00:18:09.180 So one summary of that is to say Bayes' 00:18:09.180 --> 00:18:11.460 rule, a lot of this sort of updating behavior, 00:18:11.460 --> 00:18:12.610 is really hard. 00:18:12.610 --> 00:18:13.837 It's actually tough to do. 00:18:13.837 --> 00:18:16.170 And for those of you who do statistics, math, et cetera, 00:18:16.170 --> 00:18:18.007 you might be able to do that very well. 00:18:18.007 --> 00:18:19.590 But in a way you're a very small share 00:18:19.590 --> 00:18:23.700 of the population who actually does that well. 00:18:23.700 --> 00:18:27.000 But even among those, there's a bunch of stuff 00:18:27.000 --> 00:18:28.793 that's really hard to compute in your head. 00:18:28.793 --> 00:18:30.960 You might be able to write it down and figure it out 00:18:30.960 --> 00:18:32.820 eventually, but usually the assumption 00:18:32.820 --> 00:18:35.112 is that people just do all of this stuff in their heads 00:18:35.112 --> 00:18:36.460 and really complicated problems. 00:18:36.460 --> 00:18:40.350 And that's sort of clearly not the case in many situations. 00:18:40.350 --> 00:18:44.340 And then, of course, that's sort of one part of that sort 00:18:44.340 --> 00:18:46.200 of among MIT students. 00:18:46.200 --> 00:18:48.750 But then, you know, there's other populations 00:18:48.750 --> 00:18:50.400 that are less educated. 00:18:50.400 --> 00:18:52.890 They might not even know what Bayes' rule is. 00:18:52.890 --> 00:18:54.890 They might be illiterate and so on and so forth. 00:18:54.890 --> 00:18:57.900 So the assumption of people perfectly processing 00:18:57.900 --> 00:19:02.680 information is even perhaps more tenuous. 00:19:02.680 --> 00:19:03.630 Yeah. 00:19:03.630 --> 00:19:05.547 AUDIENCE: A lot of models are usually hedonic, 00:19:05.547 --> 00:19:07.031 but people usually care about what 00:19:07.031 --> 00:19:08.755 other people are doing as well. 00:19:08.755 --> 00:19:12.240 So for example, I may not be willing to take a class. 00:19:12.240 --> 00:19:14.670 But if my friend takes a class or gets into the class, 00:19:14.670 --> 00:19:16.210 then my utility would go up. 00:19:16.210 --> 00:19:17.190 FRANK SCHILBACH: Right. 00:19:17.190 --> 00:19:20.073 So there's two parts to that. 00:19:20.073 --> 00:19:21.490 This is essentially the assumption 00:19:21.490 --> 00:19:23.323 of narrowly defined self-interest and people 00:19:23.323 --> 00:19:25.872 caring only about themselves or their own consumption. 00:19:25.872 --> 00:19:27.080 So there's two parts of that. 00:19:27.080 --> 00:19:28.480 One is you might care about your friends. 00:19:28.480 --> 00:19:29.860 You might give money to your friends. 00:19:29.860 --> 00:19:30.910 You might sort of help them out. 00:19:30.910 --> 00:19:32.327 You might help them with homework. 00:19:32.327 --> 00:19:35.590 You might help them in other situations. 00:19:35.590 --> 00:19:38.020 You might send money to charity and so on and so forth. 00:19:38.020 --> 00:19:40.187 So this is kind of like you care about other people. 00:19:40.187 --> 00:19:42.130 And the well-being of other people 00:19:42.130 --> 00:19:45.010 is in your utility function one way or the other. 00:19:45.010 --> 00:19:48.010 The second part is what you already 00:19:48.010 --> 00:19:50.410 discussed as well, which is kind of social influences, 00:19:50.410 --> 00:19:52.930 call it, which is essentially peer effects. 00:19:52.930 --> 00:19:55.630 You care a lot about what other people think about you. 00:19:55.630 --> 00:19:58.630 You care a lot about what other people do. 00:19:58.630 --> 00:19:59.650 People get jealous. 00:19:59.650 --> 00:20:00.670 People get angry. 00:20:00.670 --> 00:20:01.810 People get envious. 00:20:01.810 --> 00:20:03.490 And sort of their behavior about what 00:20:03.490 --> 00:20:07.810 they want at the end of the day is very volatile and malleable, 00:20:07.810 --> 00:20:10.760 essentially, based on influences from others, right? 00:20:10.760 --> 00:20:14.030 And that might sort of show up in all sorts of ways, 00:20:14.030 --> 00:20:16.347 including in terms of peer effects. 00:20:16.347 --> 00:20:18.430 People do all sorts of things because other people 00:20:18.430 --> 00:20:20.440 think that's cool. 00:20:20.440 --> 00:20:23.150 But it's not necessarily what's best for them 00:20:23.150 --> 00:20:25.250 if they had to choose on their own. 00:20:25.250 --> 00:20:25.750 OK, great. 00:20:25.750 --> 00:20:26.630 So let me stop here. 00:20:26.630 --> 00:20:29.060 There's a lot of other examples that we can come up with. 00:20:29.060 --> 00:20:29.810 I have a few here. 00:20:29.810 --> 00:20:31.880 So one is sort of limited self-control. 00:20:31.880 --> 00:20:34.480 That's one of my favorite pictures about gym attendance 00:20:34.480 --> 00:20:36.850 where you say, well, they seem to be 00:20:36.850 --> 00:20:39.340 people who have preferences for going to the gym. 00:20:39.340 --> 00:20:42.340 But then there's an escalator, and people 00:20:42.340 --> 00:20:43.450 not working out so much. 00:20:43.450 --> 00:20:46.895 So in some sense, something is perhaps amiss here. 00:20:46.895 --> 00:20:49.270 When you look at sort of-- and this is one of my favorite 00:20:49.270 --> 00:20:50.490 things to do is-- 00:20:50.490 --> 00:20:53.213 or not quite as-- it's a fun thing to do, but maybe not one 00:20:53.213 --> 00:20:54.130 of my favorite things. 00:20:56.830 --> 00:20:59.320 When you sort of Google terms, this 00:20:59.320 --> 00:21:02.770 is calories or Weight Watchers or the like, 00:21:02.770 --> 00:21:05.350 where you say, well, what is people's interest in googling 00:21:05.350 --> 00:21:07.690 those kinds of terms over the course of a year? 00:21:07.690 --> 00:21:09.190 And as you can imagine, what you see 00:21:09.190 --> 00:21:12.850 are these spikes that are essentially January 1st. 00:21:12.850 --> 00:21:15.250 People start sort of googling calories. 00:21:15.250 --> 00:21:17.110 They start googling Weight Watchers 00:21:17.110 --> 00:21:18.588 and so on and so forth. 00:21:18.588 --> 00:21:21.130 And now, of course, you can sort think about some models that 00:21:21.130 --> 00:21:23.530 sort of can rationalize this in some sense in saying, 00:21:23.530 --> 00:21:27.380 well, January is just a good way to start a diet and so on. 00:21:27.380 --> 00:21:29.770 But really the right thing to do is sort of a people, 00:21:29.770 --> 00:21:32.560 on January 1st, have all the best of intentions to lose 00:21:32.560 --> 00:21:37.070 weight, eat healthier, exercise more, and so on and so forth. 00:21:37.070 --> 00:21:40.240 And then the year starts, the semester 00:21:40.240 --> 00:21:41.900 happens and so on and so forth. 00:21:41.900 --> 00:21:45.040 And the good intentions sort of fade away, 00:21:45.040 --> 00:21:49.150 which is essentially limited self-control. 00:21:49.150 --> 00:21:51.640 There's some interesting things about demand 00:21:51.640 --> 00:21:55.510 for information, especially when it comes to health. 00:21:55.510 --> 00:21:57.100 You know, every time I ask this-- 00:21:57.100 --> 00:22:00.430 and I've taught this class a few years so far. 00:22:00.430 --> 00:22:02.680 Every time I ask this, I sort of realize how old I am. 00:22:02.680 --> 00:22:04.570 Who knows Dr. House? 00:22:04.570 --> 00:22:05.530 Oh, not bad. 00:22:05.530 --> 00:22:08.618 So can anybody explain to me Thirteen and Dr. House 00:22:08.618 --> 00:22:09.910 and what is sort of their deal? 00:22:09.910 --> 00:22:12.593 And what does this have to do with demand for information? 00:22:18.020 --> 00:22:18.830 So who's Thirteen? 00:22:18.830 --> 00:22:22.012 Or what is her health issues? 00:22:22.012 --> 00:22:23.970 AUDIENCE: She's got a genetic disease, I think. 00:22:23.970 --> 00:22:24.960 FRANK SCHILBACH: Yes, she has-- 00:22:24.960 --> 00:22:26.202 AUDIENCE: She has, like-- 00:22:26.202 --> 00:22:27.030 I'm not-- 00:22:27.030 --> 00:22:28.020 FRANK SCHILBACH: Huntington's, yes. 00:22:28.020 --> 00:22:29.070 AUDIENCE: Yeah, Huntington's. 00:22:29.070 --> 00:22:30.028 FRANK SCHILBACH: Right. 00:22:30.028 --> 00:22:31.410 AUDIENCE: He finds out later on. 00:22:31.410 --> 00:22:31.690 FRANK SCHILBACH: Right. 00:22:31.690 --> 00:22:34.660 So she has a disease, which is called Huntington's disease, 00:22:34.660 --> 00:22:38.220 which we'll also talk about more in the class, which essentially 00:22:38.220 --> 00:22:41.250 is a brain disease where over time 00:22:41.250 --> 00:22:43.080 your brain more or less degenerates. 00:22:43.080 --> 00:22:49.050 And it's a really bad condition where at age 40, 50 00:22:49.050 --> 00:22:49.950 that manifests. 00:22:49.950 --> 00:22:54.210 And it's a very serious condition. 00:22:54.210 --> 00:22:56.400 It's a genetic disease in the sense 00:22:56.400 --> 00:22:58.410 of, if your parents have that disease, 00:22:58.410 --> 00:22:59.950 the chance of you having it is way, 00:22:59.950 --> 00:23:03.960 way higher than in the general population. 00:23:03.960 --> 00:23:06.550 There's a test for it, but there's no cure. 00:23:06.550 --> 00:23:09.990 So now, the dilemma and the show then 00:23:09.990 --> 00:23:13.800 comes where she essentially is thinking about getting tested. 00:23:13.800 --> 00:23:16.230 And Dr. House, who in some ways very rational 00:23:16.230 --> 00:23:18.330 and others not so much, essentially 00:23:18.330 --> 00:23:21.540 encourages her to get that test. 00:23:21.540 --> 00:23:24.285 And any rational model would say, well, 00:23:24.285 --> 00:23:25.410 you should really get that. 00:23:25.410 --> 00:23:28.080 Any sort of neoclassical model would say, 00:23:28.080 --> 00:23:29.550 you should get that test. 00:23:29.550 --> 00:23:30.970 Why should you get the test? 00:23:30.970 --> 00:23:32.182 Why is that helpful? 00:23:32.182 --> 00:23:33.600 AUDIENCE: More information. 00:23:33.600 --> 00:23:35.933 FRANK SCHILBACH: And what's the information helpful for? 00:23:35.933 --> 00:23:38.015 AUDIENCE: [? It ?] [? would update ?] [? price ?] 00:23:38.015 --> 00:23:38.980 [? of action. ?] 00:23:38.980 --> 00:23:39.610 FRANK SCHILBACH: Right, exactly. 00:23:39.610 --> 00:23:40.960 You would say information is good. 00:23:40.960 --> 00:23:41.890 Why is information good? 00:23:41.890 --> 00:23:43.600 Because there's lots of important choices 00:23:43.600 --> 00:23:46.330 that you might make that depend on that information. 00:23:46.330 --> 00:23:49.120 You might think about how much money you want to save. 00:23:49.120 --> 00:23:52.790 You think about your choices about taking 00:23:52.790 --> 00:23:57.430 a vacation, about career choices, about partner choices, 00:23:57.430 --> 00:24:00.612 about having children, all sorts of issues 00:24:00.612 --> 00:24:01.570 about health behaviors. 00:24:01.570 --> 00:24:03.528 You might sort of take better care of yourself. 00:24:03.528 --> 00:24:05.800 There's all sorts of really important decisions 00:24:05.800 --> 00:24:08.860 that might hinge on the fact your life will be dramatically 00:24:08.860 --> 00:24:12.190 different whether you have Huntington's disease or not. 00:24:12.190 --> 00:24:14.062 Now, Thirteen does not want to-- 00:24:14.062 --> 00:24:15.520 I think she actually does the test, 00:24:15.520 --> 00:24:20.120 and then doesn't want to see the result. And why is that? 00:24:20.120 --> 00:24:20.620 Yes. 00:24:20.620 --> 00:24:22.328 AUDIENCE: Even if you have the condition, 00:24:22.328 --> 00:24:23.882 there's no cure for the disease. 00:24:23.882 --> 00:24:28.090 So it might help your quality of life if you [INAUDIBLE].. 00:24:28.090 --> 00:24:29.790 FRANK SCHILBACH: So that's one view. 00:24:29.790 --> 00:24:31.330 Just to view that, I think, in some sense, as I said, 00:24:31.330 --> 00:24:33.490 you could make other decisions, like economic choices. 00:24:33.490 --> 00:24:34.300 You could save more. 00:24:34.300 --> 00:24:35.550 You could see the doctor more. 00:24:35.550 --> 00:24:37.450 You could go more on vacations or whatever, 00:24:37.450 --> 00:24:38.830 do fun things in life. 00:24:38.830 --> 00:24:42.650 But you could say, well, maybe that's not that important. 00:24:42.650 --> 00:24:46.150 But the neoclassical or the classical economics model 00:24:46.150 --> 00:24:48.455 would still argue that you should get that information. 00:24:48.455 --> 00:24:49.330 It wouldn't hurt you. 00:24:49.330 --> 00:24:51.550 You wouldn't want to refuse it. 00:24:51.550 --> 00:24:52.720 But why is she refusing it? 00:24:52.720 --> 00:24:53.140 Yes. 00:24:53.140 --> 00:24:54.130 AUDIENCE: I think maybe she doesn't 00:24:54.130 --> 00:24:55.340 want to live the rest of her life 00:24:55.340 --> 00:24:56.560 knowing she has the disease. 00:24:56.560 --> 00:24:57.602 FRANK SCHILBACH: Exactly. 00:24:57.602 --> 00:25:00.820 So she essentially derives utility from information, 00:25:00.820 --> 00:25:02.590 if you want, from her beliefs. 00:25:02.590 --> 00:25:05.350 She likes to sort of pretend to herself that she's healthy 00:25:05.350 --> 00:25:08.480 or the chance of her being healthy is very high. 00:25:08.480 --> 00:25:11.350 Now, under some assumptions, you might 00:25:11.350 --> 00:25:14.710 want to not get tested and sort of see the result of that test. 00:25:14.710 --> 00:25:17.780 Because once you have the test, if it's positive, 00:25:17.780 --> 00:25:20.572 it's very hard to pretend to yourself 00:25:20.572 --> 00:25:22.030 that you're healthy because there's 00:25:22.030 --> 00:25:24.310 a hard test saying that. 00:25:24.310 --> 00:25:26.290 Before the test, she can still sort of think, 00:25:26.290 --> 00:25:28.248 try to forget about it, and try to sort of live 00:25:28.248 --> 00:25:31.930 a happy life sort of until the symptoms set in. 00:25:31.930 --> 00:25:35.350 But that's a clearer violation of the classical model 00:25:35.350 --> 00:25:37.660 of the assumptions that I was showing you above. 00:25:40.880 --> 00:25:43.700 The next part is default effects. 00:25:43.700 --> 00:25:45.770 So what I'm showing over here is the fraction 00:25:45.770 --> 00:25:49.310 of organ donors by country and type of default. 00:25:49.310 --> 00:25:54.020 There's two types of defaults, often, for organ donations. 00:25:54.020 --> 00:25:57.590 This is from a few years ago, but I think, overall, things 00:25:57.590 --> 00:25:59.630 have not changed that much. 00:25:59.630 --> 00:26:01.740 There's opt-in versus opt-out. 00:26:01.740 --> 00:26:04.235 So in the countries on the left, essentially 00:26:04.235 --> 00:26:05.840 you have an opt-in policy. 00:26:05.840 --> 00:26:07.460 In a sense, if you don't do anything, 00:26:07.460 --> 00:26:08.840 you're not an organ donor. 00:26:08.840 --> 00:26:11.730 So you have to actively say in some form 00:26:11.730 --> 00:26:15.840 or some form of declaration that you want to be an organ donor. 00:26:15.840 --> 00:26:18.347 On the right, there's opt-out policies, 00:26:18.347 --> 00:26:20.430 which means essentially, if you don't do anything, 00:26:20.430 --> 00:26:23.700 you'll be automatically registered or viewed 00:26:23.700 --> 00:26:24.870 as an organ donor. 00:26:24.870 --> 00:26:29.092 You can opt out, but only if you do so, 00:26:29.092 --> 00:26:30.300 you'll actually be opted out. 00:26:30.300 --> 00:26:32.217 Otherwise, if there's an accident or the like, 00:26:32.217 --> 00:26:34.710 you'll be viewed as an organ donor. 00:26:34.710 --> 00:26:36.310 Now, what's weird about this graph? 00:26:36.310 --> 00:26:39.330 Or why is this sort of potentially a violation 00:26:39.330 --> 00:26:42.735 of the neoclassical or the typical economics assumptions? 00:26:45.710 --> 00:26:46.500 Yes. 00:26:46.500 --> 00:26:49.850 AUDIENCE: Because assuming that the people have 00:26:49.850 --> 00:26:52.850 the same preference of whether they want to opt-in or opt-out, 00:26:52.850 --> 00:26:56.000 then there should be the same proportion across. 00:26:56.000 --> 00:26:59.030 But the difference is that the effort required 00:26:59.030 --> 00:27:02.023 to make that change or the lack of indifference 00:27:02.023 --> 00:27:04.190 causes there to be the imbalance between the opt-out 00:27:04.190 --> 00:27:04.817 and the opt-in. 00:27:04.817 --> 00:27:06.150 FRANK SCHILBACH: Right, exactly. 00:27:06.150 --> 00:27:09.632 So some people in-- 00:27:09.632 --> 00:27:11.590 I don't want to get in trouble, but some people 00:27:11.590 --> 00:27:14.240 would argue Germany and Austria is actually not that different. 00:27:14.240 --> 00:27:14.823 So I'm German. 00:27:14.823 --> 00:27:16.562 So some Austrians might disagree. 00:27:16.562 --> 00:27:19.020 But you would say, you know, overall, Germans and Austrians 00:27:19.020 --> 00:27:20.100 are not that different overall. 00:27:20.100 --> 00:27:21.960 And you think, you know, their preferences in particular, 00:27:21.960 --> 00:27:22.960 they might be different. 00:27:22.960 --> 00:27:24.768 But their preferences for organ donations 00:27:24.768 --> 00:27:27.060 are probably pretty similar across those two countries, 00:27:27.060 --> 00:27:28.890 which is a reasonable assumption. 00:27:28.890 --> 00:27:30.690 Second, lots of people care a lot 00:27:30.690 --> 00:27:33.080 about what happens to them after they die, right? 00:27:33.080 --> 00:27:35.448 So they, for whatever reasons, religious or other, 00:27:35.448 --> 00:27:36.990 they might actually care a lot about, 00:27:36.990 --> 00:27:40.060 at the end of the day, what happens to their body 00:27:40.060 --> 00:27:40.560 and so on. 00:27:40.560 --> 00:27:43.470 And that's obviously their choice to make. 00:27:43.470 --> 00:27:45.550 Now, if you assume those two things, 00:27:45.550 --> 00:27:48.450 then it's very hard to reconcile this picture. 00:27:48.450 --> 00:27:51.120 Because, essentially, the opt-in and opt-out 00:27:51.120 --> 00:27:52.830 is a very small change to make. 00:27:52.830 --> 00:27:55.150 It's very simple and easy to do. 00:27:55.150 --> 00:27:56.730 You just have to fill out some form. 00:27:56.730 --> 00:27:58.380 Maybe it takes you an hour, maybe two. 00:27:58.380 --> 00:28:00.480 You have to maybe go to some office or the like. 00:28:00.480 --> 00:28:03.570 But if you really care, you could easily do it. 00:28:03.570 --> 00:28:05.910 Yet you see huge differences in outcome 00:28:05.910 --> 00:28:09.810 based on this very simple small change, which 00:28:09.810 --> 00:28:12.030 is very hard to model or write down a model that sort 00:28:12.030 --> 00:28:13.930 of rationalizes this behavior. 00:28:13.930 --> 00:28:17.160 So somehow the default, sort of the way in which the decision 00:28:17.160 --> 00:28:20.220 is presented, seems to really affect people's behavior 00:28:20.220 --> 00:28:22.170 in a fundamental way. 00:28:24.950 --> 00:28:26.930 OK, next one is GlowCaps. 00:28:26.930 --> 00:28:30.680 This is help for people to take their medication, in particular 00:28:30.680 --> 00:28:32.450 for the elderly, which essentially 00:28:32.450 --> 00:28:35.200 is a reminder for people. 00:28:35.200 --> 00:28:37.700 You can sort of program this in certain ways, like every day 00:28:37.700 --> 00:28:40.220 or every few hours, GlowCaps. 00:28:40.220 --> 00:28:41.930 So it's either sort of making sounds 00:28:41.930 --> 00:28:44.210 or such blinking and the like and gets 00:28:44.210 --> 00:28:46.040 you to take your medication. 00:28:46.040 --> 00:28:49.070 And it's a very sort of simple proof of concept 00:28:49.070 --> 00:28:51.800 that reminders are important and can save people's lives. 00:28:51.800 --> 00:28:54.000 And people have done studies on that and so on. 00:28:54.000 --> 00:28:56.180 But essentially, it's a very simple proof of concept 00:28:56.180 --> 00:28:57.838 that memory is limited, right? 00:28:57.838 --> 00:29:00.380 If your memory was perfect, if you could remember everything, 00:29:00.380 --> 00:29:03.668 if you were a perfect information processor, 00:29:03.668 --> 00:29:05.460 you would not forget about your medication, 00:29:05.460 --> 00:29:07.127 especially if it's important, especially 00:29:07.127 --> 00:29:10.610 if it's medication that potentially saves your life. 00:29:10.610 --> 00:29:17.390 Yet we see that people's medication usage is strongly 00:29:17.390 --> 00:29:20.330 affected by sort of types of policies or products 00:29:20.330 --> 00:29:22.400 that can help people remember. 00:29:22.400 --> 00:29:26.810 So that's essentially a rejection of perfect memory. 00:29:26.810 --> 00:29:28.130 Next one is charity. 00:29:28.130 --> 00:29:32.090 So if you think about people in all sorts of ways 00:29:32.090 --> 00:29:36.830 seem to care about others in the sense of just giving money, 00:29:36.830 --> 00:29:39.050 here's one charity that's called GiveDirectly, which 00:29:39.050 --> 00:29:41.720 is a very nice charity in a sense that what it does, 00:29:41.720 --> 00:29:46.010 essentially, is a very simple way of helping the poor by just 00:29:46.010 --> 00:29:48.767 directly transferring money to poor countries. 00:29:48.767 --> 00:29:50.600 So on the right, you see an actual recipient 00:29:50.600 --> 00:29:52.460 from GiveDirectly. 00:29:52.460 --> 00:29:53.930 That person has a phone. 00:29:53.930 --> 00:29:56.367 They're sort of a mobile phone based transfers 00:29:56.367 --> 00:29:57.950 where, essentially, you can send money 00:29:57.950 --> 00:30:00.090 to people with mobile phones. 00:30:00.090 --> 00:30:04.100 So if you decided now to give $100 to that person-- 00:30:04.100 --> 00:30:06.570 it could be that person or a similar person-- 00:30:06.570 --> 00:30:09.200 95 of those or 90 of those dollars would then, in fact, 00:30:09.200 --> 00:30:12.140 arrive at the cell phone of that person. 00:30:12.140 --> 00:30:14.480 Now, why are people doing this? 00:30:14.480 --> 00:30:16.700 Presumably, because they care about others 00:30:16.700 --> 00:30:18.290 one way or the other. 00:30:18.290 --> 00:30:20.232 It could be either that, in some sense, that's 00:30:20.232 --> 00:30:21.440 part of the utility function. 00:30:21.440 --> 00:30:23.608 Just donating money to others makes you happier. 00:30:23.608 --> 00:30:26.150 And that's essentially a way in which you improve your being. 00:30:26.150 --> 00:30:29.675 You just have others well-being in your utility function. 00:30:29.675 --> 00:30:31.800 Or it could be things like what we said previously. 00:30:31.800 --> 00:30:34.160 It could you sort of think it's appropriate thing to do. 00:30:34.160 --> 00:30:35.120 Maybe other people do. 00:30:35.120 --> 00:30:36.110 It's an important thing to do. 00:30:36.110 --> 00:30:37.777 Maybe you can tell your friends about it 00:30:37.777 --> 00:30:39.030 and so on and so forth. 00:30:39.030 --> 00:30:40.760 But one way or the other, people give lots of money 00:30:40.760 --> 00:30:43.250 to charity, which in some sense says, one way or the other, 00:30:43.250 --> 00:30:45.920 you must care about others overall. 00:30:45.920 --> 00:30:50.070 So others are in your utility function. 00:30:50.070 --> 00:30:53.480 So now, attention, if you have seen this before, 00:30:53.480 --> 00:30:55.950 don't tell your friends about it. 00:30:55.950 --> 00:31:00.340 Here's a simple attention test that I want you to do. 00:31:00.340 --> 00:31:03.507 What's interesting about this is, in some ways, 00:31:03.507 --> 00:31:04.840 you might have seen this before. 00:31:04.840 --> 00:31:07.540 Some of you may have seen the gorilla before. 00:31:07.540 --> 00:31:10.030 Who didn't see the gorilla? 00:31:10.030 --> 00:31:10.732 That's OK. 00:31:10.732 --> 00:31:11.690 I didn't see it either. 00:31:11.690 --> 00:31:13.350 But some people have seen it before. 00:31:13.350 --> 00:31:15.100 There's a bunch of experiments that people 00:31:15.100 --> 00:31:18.760 have done that essentially show that, often, a large fraction 00:31:18.760 --> 00:31:20.783 of people don't see this gorilla. 00:31:20.783 --> 00:31:22.450 And there's various ways in which people 00:31:22.450 --> 00:31:24.100 do other types of experiments. 00:31:24.100 --> 00:31:27.310 There's types of experiments where essentially people, 00:31:27.310 --> 00:31:28.495 they go to some bank teller. 00:31:28.495 --> 00:31:31.120 The bank teller says, I'm going to go away and then comes back. 00:31:31.120 --> 00:31:33.245 And then there's a different person who comes back, 00:31:33.245 --> 00:31:34.750 and people don't notice. 00:31:34.750 --> 00:31:36.680 There's various versions of that. 00:31:36.680 --> 00:31:39.430 And so what's interesting about this in some sense is-- 00:31:39.430 --> 00:31:42.700 so A, it proves, in some ways, that attention is limited. 00:31:42.700 --> 00:31:46.760 You just focus your attention on some things and not others. 00:31:46.760 --> 00:31:48.790 And you might miss important things. 00:31:48.790 --> 00:31:50.230 Now, that's OK in some ways. 00:31:50.230 --> 00:31:52.330 There's sort of a version of that that says, well, 00:31:52.330 --> 00:31:54.790 it might be sort of rational inattention. 00:31:54.790 --> 00:31:57.703 In some sense, you focus on the important things in life. 00:31:57.703 --> 00:31:59.620 And you might miss some things, and that's OK. 00:31:59.620 --> 00:32:01.840 In some sense, I asked you, or the video asked you, 00:32:01.840 --> 00:32:03.900 to count the number of passes. 00:32:03.900 --> 00:32:05.990 So in some sense, the gorilla is irrelevant. 00:32:05.990 --> 00:32:07.568 And that's true in this case. 00:32:07.568 --> 00:32:09.610 But in many other cases, people might sort of not 00:32:09.610 --> 00:32:10.690 rationally pay attention. 00:32:10.690 --> 00:32:12.857 They might sort of miss some really important things 00:32:12.857 --> 00:32:15.580 in their lives in part because they're 00:32:15.580 --> 00:32:17.500 distracted, in part because they don't want 00:32:17.500 --> 00:32:19.930 to pay attention and so on. 00:32:19.930 --> 00:32:22.150 And some strand of behavioral economics 00:32:22.150 --> 00:32:25.930 which we'll talk about is sort of when 00:32:25.930 --> 00:32:28.060 people do not pay attention and then make 00:32:28.060 --> 00:32:30.100 mistakes because of that. 00:32:32.920 --> 00:32:33.460 OK. 00:32:33.460 --> 00:32:36.590 So I could go on and on with lots of different examples. 00:32:36.590 --> 00:32:39.680 But the bottom line here is that most researchers in psychology 00:32:39.680 --> 00:32:43.300 economics believe that the classical model of behavior, 00:32:43.300 --> 00:32:47.470 sort of the homo economicus, is too extreme in various ways. 00:32:47.470 --> 00:32:51.310 That person is too selfish, too rational, and too willful. 00:32:51.310 --> 00:32:54.490 And in some ways, we want to kind of understand 00:32:54.490 --> 00:32:56.830 how relaxing some of these assumptions 00:32:56.830 --> 00:32:59.260 might make economic models more realistic. 00:32:59.260 --> 00:33:02.078 Now, in some sense, no economist would, in fact, 00:33:02.078 --> 00:33:04.120 argue that the assumptions of the standard models 00:33:04.120 --> 00:33:05.680 are exactly correct. 00:33:05.680 --> 00:33:08.188 The questions are, are these deviations important? 00:33:08.188 --> 00:33:09.730 Do they actually matter for something 00:33:09.730 --> 00:33:11.740 important in explaining people's behavior? 00:33:11.740 --> 00:33:15.340 And which of those assumptions or deviations actually matter? 00:33:15.340 --> 00:33:18.550 And that's kind of the name of the game here. 00:33:18.550 --> 00:33:22.990 In fact, when you talk to cognitive scientist, 00:33:22.990 --> 00:33:26.140 psychologist, et cetera, they will tell you the world 00:33:26.140 --> 00:33:28.180 is full of cognitive biases. 00:33:28.180 --> 00:33:30.702 And you might not be able to read this. 00:33:30.702 --> 00:33:31.910 And that's kind of the point. 00:33:31.910 --> 00:33:34.120 There's so many different things in which we are biased. 00:33:34.120 --> 00:33:35.500 In some sense, every choice that we make 00:33:35.500 --> 00:33:36.875 or any things that we do, there's 00:33:36.875 --> 00:33:39.157 lots of biases that interfere with people's choices. 00:33:39.157 --> 00:33:40.740 And almost any choice you can think of 00:33:40.740 --> 00:33:42.640 or any behavior they can think about, 00:33:42.640 --> 00:33:44.590 there will be psychology or other experiments 00:33:44.590 --> 00:33:48.740 showing that people do not behave perfectly. 00:33:48.740 --> 00:33:54.520 Now, the key question then is, which of those assumptions 00:33:54.520 --> 00:33:55.750 are important? 00:33:55.750 --> 00:33:59.413 And which of those violations of assumptions should we focus on? 00:33:59.413 --> 00:34:01.330 And for that, I want to step back a little bit 00:34:01.330 --> 00:34:02.913 and say, OK, what is actually a model? 00:34:02.913 --> 00:34:04.992 What are economic models trying to do? 00:34:04.992 --> 00:34:05.950 And so what is a model? 00:34:05.950 --> 00:34:09.610 A model is a simplified representation of the world. 00:34:09.610 --> 00:34:12.610 And we know, in some sense, that the assumptions of the models 00:34:12.610 --> 00:34:13.179 are not true. 00:34:13.179 --> 00:34:16.330 They're sort of supposed to be approximately true and exactly 00:34:16.330 --> 00:34:18.020 false. 00:34:18.020 --> 00:34:20.139 So when you think about used models of the Earth, 00:34:20.139 --> 00:34:22.630 in some sense, they're flat, spherical, ellipsoid, 00:34:22.630 --> 00:34:24.940 and so on and so forth, models. 00:34:24.940 --> 00:34:28.000 Now, good models do not account for bumps and grooves 00:34:28.000 --> 00:34:28.870 and so on. 00:34:28.870 --> 00:34:32.260 A perfect replica of the Earth is not a useful model to use. 00:34:32.260 --> 00:34:34.120 You kind of want to simplify and capture 00:34:34.120 --> 00:34:37.100 the essence of what's important. 00:34:37.100 --> 00:34:38.590 So now, then what's a good model? 00:34:38.590 --> 00:34:41.257 Well, a good model-- and you can read a bit about this in Gabaix 00:34:41.257 --> 00:34:42.280 and Laibson's paper. 00:34:42.280 --> 00:34:44.360 A good model is supposed to be simple. 00:34:44.360 --> 00:34:46.570 It's supposed to be easy to work with and tractable. 00:34:46.570 --> 00:34:48.670 So in some sense, you have only a few variables, 00:34:48.670 --> 00:34:51.580 a few things that matter. 00:34:51.580 --> 00:34:54.050 It's conceptually insightful in the sense 00:34:54.050 --> 00:34:56.130 that it focuses on important things. 00:34:56.130 --> 00:34:57.880 It tells you about behavior that we really 00:34:57.880 --> 00:34:59.830 care about and important ideas. 00:34:59.830 --> 00:35:02.440 It's generalizable in a sense that, ideally, we 00:35:02.440 --> 00:35:06.040 are trying to look for some behaviors that are generally 00:35:06.040 --> 00:35:06.800 in some sense. 00:35:06.800 --> 00:35:10.420 So if I can just explain you one simple choice in one domain 00:35:10.420 --> 00:35:12.310 and write a model about it, but that model 00:35:12.310 --> 00:35:14.320 doesn't apply for anything else, that's 00:35:14.320 --> 00:35:16.900 not a good model to work with. 00:35:16.900 --> 00:35:19.058 The model is supposed to be falsifiable in a sense 00:35:19.058 --> 00:35:20.350 that we can actually test them. 00:35:20.350 --> 00:35:22.480 And that's kind of what we do in experimental and behavioral 00:35:22.480 --> 00:35:23.080 economics. 00:35:23.080 --> 00:35:26.860 We try to test people's theories and then empirical work 00:35:26.860 --> 00:35:30.580 and then sort of falsify or reject models that are wrong 00:35:30.580 --> 00:35:33.010 and sort of accept or do not reject 00:35:33.010 --> 00:35:36.522 models we think are better. 00:35:36.522 --> 00:35:38.980 There's supposed to be empirical consistency in a sense of, 00:35:38.980 --> 00:35:40.840 if I explain one behavior, I should also 00:35:40.840 --> 00:35:42.913 explain this over time or in different domains. 00:35:42.913 --> 00:35:44.830 And we should be able to make good predictions 00:35:44.830 --> 00:35:47.590 for people's behavior. 00:35:47.590 --> 00:35:50.050 Now, are the assumptions of the standard models 00:35:50.050 --> 00:35:51.680 true for most people? 00:35:51.680 --> 00:35:53.350 The answer is no. 00:35:53.350 --> 00:35:57.490 But the key insight here, the key properties of good models, 00:35:57.490 --> 00:35:58.840 is simplicity. 00:35:58.840 --> 00:36:01.863 So in some sense, assuming perfect rationality, 00:36:01.863 --> 00:36:04.030 selfishness and willpower is actually a simple thing 00:36:04.030 --> 00:36:05.020 to do in some sense. 00:36:05.020 --> 00:36:08.320 The reason why economists have made those assumptions 00:36:08.320 --> 00:36:10.870 to start with is not necessarily because they thought 00:36:10.870 --> 00:36:13.960 people are perfectly rational or that there 00:36:13.960 --> 00:36:16.773 are not these psychological issues going on. 00:36:16.773 --> 00:36:18.190 The reason is actually simplicity. 00:36:18.190 --> 00:36:19.370 It's an easy thing to do. 00:36:19.370 --> 00:36:21.640 You model how you think people should behave 00:36:21.640 --> 00:36:24.370 and how they behave perfectly. 00:36:24.370 --> 00:36:27.250 And making some of these models sort of richer 00:36:27.250 --> 00:36:29.680 in their psychology is actually complicated 00:36:29.680 --> 00:36:32.860 and makes the models more complex and harder to analyze. 00:36:32.860 --> 00:36:35.200 So that is just to say, well, then the question is, 00:36:35.200 --> 00:36:37.240 can we find some assumptions of economics? 00:36:37.240 --> 00:36:39.670 Can we make them more realistic in a tractable way? 00:36:39.670 --> 00:36:41.230 In a sense, can we find key things 00:36:41.230 --> 00:36:45.100 that we change that keep the models tractable 00:36:45.100 --> 00:36:47.380 while then also explaining important things better 00:36:47.380 --> 00:36:48.665 than we can before? 00:36:48.665 --> 00:36:50.290 So it's not about just taking all sorts 00:36:50.290 --> 00:36:52.270 of psychological issues that might be going on 00:36:52.270 --> 00:36:54.850 and say, well, economic assumptions are wrong 00:36:54.850 --> 00:36:58.630 or models of economics are wrong or the assumptions of those 00:36:58.630 --> 00:36:59.270 are wrong. 00:36:59.270 --> 00:37:01.190 We know that these assumptions are wrong. 00:37:01.190 --> 00:37:04.150 The question is, can we make somewhat simple assumptions 00:37:04.150 --> 00:37:06.040 or improvements of those assumptions 00:37:06.040 --> 00:37:07.810 based on insights from psychology 00:37:07.810 --> 00:37:10.600 and other fields that help us improve those models 00:37:10.600 --> 00:37:13.240 and then make better predictions and all 00:37:13.240 --> 00:37:15.650 of that in a tractable way? 00:37:15.650 --> 00:37:17.200 So now, this is very important here. 00:37:17.200 --> 00:37:20.710 A good behavior economist or a good student in this class 00:37:20.710 --> 00:37:22.660 is also a good economist. 00:37:22.660 --> 00:37:24.490 Behavioral economics is not sort of trying 00:37:24.490 --> 00:37:26.460 to replace standard economics. 00:37:26.460 --> 00:37:28.210 So I don't want you to go to my colleagues 00:37:28.210 --> 00:37:31.150 and say, you know, 1401, 1402, and so on, this is all garbage. 00:37:31.150 --> 00:37:34.155 Frank is telling you we do things differently. 00:37:34.155 --> 00:37:36.530 In fact, you very much need to know those kinds of things 00:37:36.530 --> 00:37:38.930 from standard economics to understand 00:37:38.930 --> 00:37:42.500 what the assumptions are and how to best to deviate from that. 00:37:42.500 --> 00:37:47.630 And so such then, key principles of mainstream economics 00:37:47.630 --> 00:37:49.310 continue to apply. 00:37:49.310 --> 00:37:51.470 Decision makers are still highly sophisticated. 00:37:51.470 --> 00:37:53.580 Markets and incentives matter. 00:37:53.580 --> 00:37:57.230 They, in fact, play a key role in shaping behavior. 00:37:57.230 --> 00:38:00.868 And markets allocate resources well most of the time. 00:38:00.868 --> 00:38:03.410 The question is, can we sort of focus on important deviations 00:38:03.410 --> 00:38:06.120 and try to fix those? 00:38:06.120 --> 00:38:08.210 And then, again, sort of methodological principles 00:38:08.210 --> 00:38:09.320 still apply. 00:38:09.320 --> 00:38:11.660 Use observational experimental data. 00:38:11.660 --> 00:38:14.785 Mathematical models are good and so on and so forth. 00:38:14.785 --> 00:38:16.160 And ideally, models would sort of 00:38:16.160 --> 00:38:18.320 finesse the special case of perfect 00:38:18.320 --> 00:38:21.755 rationality or perfectly sort of standard economic models. 00:38:21.755 --> 00:38:23.630 And they try to build in some parameters that 00:38:23.630 --> 00:38:25.255 deviate from that and try to understand 00:38:25.255 --> 00:38:27.740 can we make better predictions. 00:38:27.740 --> 00:38:30.050 And then so, often, prices are actually 00:38:30.050 --> 00:38:31.940 the most important aspects of choice. 00:38:31.940 --> 00:38:33.620 Here's one experiment by Ito et al. 00:38:33.620 --> 00:38:36.570 that try to essentially sort of change people's energy usage 00:38:36.570 --> 00:38:37.310 and so on. 00:38:37.310 --> 00:38:38.630 And they have two treatments. 00:38:38.630 --> 00:38:40.760 One is called moral suasion, which 00:38:40.760 --> 00:38:42.500 essentially is telling people-- 00:38:42.500 --> 00:38:44.673 sort of appealing to their morals in some ways. 00:38:44.673 --> 00:38:47.090 And the other one is essentially just financial incentive, 00:38:47.090 --> 00:38:49.932 straight up changing the price of people's choices. 00:38:49.932 --> 00:38:51.890 And what you see in this experiment essentially 00:38:51.890 --> 00:38:55.130 is there's, in red, you have the impact. 00:38:55.130 --> 00:38:57.800 These are all treatment effects over time comparing a treatment 00:38:57.800 --> 00:38:59.210 group and a control group. 00:38:59.210 --> 00:39:01.370 And you see, in red, the treatment effects 00:39:01.370 --> 00:39:02.300 of the incentives. 00:39:02.300 --> 00:39:05.000 And you see these are relatively large and persist over time. 00:39:05.000 --> 00:39:07.250 And you see the treatment effects of the moral suasion 00:39:07.250 --> 00:39:11.300 treatment, which happens to be reasonably large to start with, 00:39:11.300 --> 00:39:13.293 but essentially just goes away. 00:39:13.293 --> 00:39:14.460 What did we learn from that? 00:39:14.460 --> 00:39:16.400 We've learned from that prices matter. 00:39:16.400 --> 00:39:18.080 And maybe in this case, moral suasion 00:39:18.080 --> 00:39:19.400 is just not that important. 00:39:19.400 --> 00:39:20.580 And that's perfectly fine. 00:39:20.580 --> 00:39:23.450 We're trying to identify situations or cases 00:39:23.450 --> 00:39:26.810 where some of the underlying psychological issues 00:39:26.810 --> 00:39:27.590 are important. 00:39:27.590 --> 00:39:30.360 And we've learned from an experiment that, in some cases, 00:39:30.360 --> 00:39:31.310 it doesn't apply. 00:39:31.310 --> 00:39:32.060 Well, so be it. 00:39:32.060 --> 00:39:33.770 Then we should focus on other cases 00:39:33.770 --> 00:39:38.330 where the psychological issues are more important. 00:39:38.330 --> 00:39:38.840 OK. 00:39:38.840 --> 00:39:44.380 So then what's sort of our broad approach to each topic? 00:39:44.380 --> 00:39:46.900 We're trying to sort of start with an intuitive, empirical, 00:39:46.900 --> 00:39:49.060 and experimental examples how people 00:39:49.060 --> 00:39:51.293 behave in some situations. 00:39:51.293 --> 00:39:52.960 We try to think about their motivations, 00:39:52.960 --> 00:39:55.127 try to see about how they make choices and behaviors 00:39:55.127 --> 00:39:57.117 and how they behave in certain situations. 00:39:57.117 --> 00:39:59.200 And then we're going to try and sort of model this 00:39:59.200 --> 00:40:01.510 in a more precise way and try to consider 00:40:01.510 --> 00:40:03.700 how the modeling of that perhaps deviates 00:40:03.700 --> 00:40:07.678 from the neoclassical or the classical model of economics. 00:40:07.678 --> 00:40:09.220 Some other times, you also just start 00:40:09.220 --> 00:40:10.870 from the classical model of economics, 00:40:10.870 --> 00:40:13.425 look at the predictions, and say, well, can we reject that? 00:40:13.425 --> 00:40:14.800 Or do we make certain predictions 00:40:14.800 --> 00:40:15.640 that are just not true? 00:40:15.640 --> 00:40:17.140 And then say, can we sort of improve 00:40:17.140 --> 00:40:18.645 those types of assumptions? 00:40:18.645 --> 00:40:20.770 And then, overall, we're trying to think about then 00:40:20.770 --> 00:40:23.110 how these hypotheses or deviations might 00:40:23.110 --> 00:40:25.210 be able to explain how people behave in markets 00:40:25.210 --> 00:40:27.748 and what choices they make and how, perhaps, we 00:40:27.748 --> 00:40:30.040 can think about policies that might affect the people's 00:40:30.040 --> 00:40:35.310 well-being, their welfare, or any other consequences. 00:40:35.310 --> 00:40:36.820 Any questions so far? 00:40:42.310 --> 00:40:43.300 OK. 00:40:43.300 --> 00:40:47.400 So now, I want to give you one very simple, in some sense, 00:40:47.400 --> 00:40:50.740 stylized example which sort of demonstrates a little bit what 00:40:50.740 --> 00:40:54.070 sort of the neoclassical economic standard 00:40:54.070 --> 00:40:57.070 economic assumptions are and how sort of enriching 00:40:57.070 --> 00:41:00.250 a model with psychological considerations 00:41:00.250 --> 00:41:02.600 might be more powerful. 00:41:02.600 --> 00:41:05.800 So when you think about laptops in class, 00:41:05.800 --> 00:41:09.070 what are sort of standard economic considerations? 00:41:09.070 --> 00:41:10.120 Should you allow laptops? 00:41:10.120 --> 00:41:11.260 Should there be laptops in class? 00:41:11.260 --> 00:41:12.460 Is that a good thing or a bad thing? 00:41:12.460 --> 00:41:13.793 Or what do you think about that? 00:41:19.490 --> 00:41:22.280 And now, I'm asking for standard, non-psychological 00:41:22.280 --> 00:41:25.280 issues, just saying, if you went to a classical economist 00:41:25.280 --> 00:41:27.510 and say, should we allow laptops in class, 00:41:27.510 --> 00:41:30.680 what would be considerations for that? 00:41:30.680 --> 00:41:31.490 Yeah. 00:41:31.490 --> 00:41:33.448 AUDIENCE: Externalities-- if you're distracting 00:41:33.448 --> 00:41:34.575 people who are around you? 00:41:34.575 --> 00:41:35.450 FRANK SCHILBACH: Yes. 00:41:35.450 --> 00:41:40.070 So you could call that psychological or not. 00:41:40.070 --> 00:41:42.610 But essentially, it could distract people essentially. 00:41:42.610 --> 00:41:46.130 And externalities tend to be sort of, very 00:41:46.130 --> 00:41:49.100 much in classical economics-- essentially, if you smoke, 00:41:49.100 --> 00:41:49.850 it affects others. 00:41:49.850 --> 00:41:52.010 Similarly, if you had a laptop, it might affect others. 00:41:52.010 --> 00:41:53.635 It's a little bit sort of in the middle 00:41:53.635 --> 00:41:56.120 between psychological and economic considerations. 00:41:56.120 --> 00:41:59.840 Because in some sense, if somebody next to you 00:41:59.840 --> 00:42:01.890 would use a laptop, then I would just say, well, 00:42:01.890 --> 00:42:03.598 why don't you just focus on class anyway? 00:42:03.598 --> 00:42:05.478 You should be able to just ignore that. 00:42:05.478 --> 00:42:07.520 But setting that aside that exactly externalities 00:42:07.520 --> 00:42:10.150 are surely one considerations that we have. 00:42:10.150 --> 00:42:11.320 What's the pro-side? 00:42:11.320 --> 00:42:13.090 Or why would we allow laptops? 00:42:13.090 --> 00:42:15.740 What's good about laptops in class? 00:42:15.740 --> 00:42:16.240 Yeah. 00:42:16.240 --> 00:42:18.790 AUDIENCE: It could just be efficient for taking notes. 00:42:18.790 --> 00:42:20.380 Some people might need to use them. 00:42:20.380 --> 00:42:24.320 Some people may are dyslexic, stuff like that. 00:42:24.320 --> 00:42:25.240 So it's just good. 00:42:25.240 --> 00:42:26.282 FRANK SCHILBACH: Exactly. 00:42:26.282 --> 00:42:28.642 It's a useful technology to take notes. 00:42:28.642 --> 00:42:30.142 And you might just be better at that 00:42:30.142 --> 00:42:31.190 or might be more comfortable. 00:42:31.190 --> 00:42:32.320 It's easier for you to do. 00:42:32.320 --> 00:42:35.380 Maybe it also saves some paper or whatever. 00:42:35.380 --> 00:42:37.330 What else are laptops good for? 00:42:37.330 --> 00:42:38.087 Yeah. 00:42:38.087 --> 00:42:39.920 AUDIENCE: More choice is just always better. 00:42:39.920 --> 00:42:41.350 So if each person has their laptop, 00:42:41.350 --> 00:42:43.030 they can choose what's better for them to pay attention-- 00:42:43.030 --> 00:42:44.280 FRANK SCHILBACH: Yes, exactly. 00:42:44.280 --> 00:42:46.390 If you are so inclined and want to watch football 00:42:46.390 --> 00:42:49.700 from yesterday, the replay and so on, if you prefer that 00:42:49.700 --> 00:42:51.310 and that's good for you, you know, 00:42:51.310 --> 00:42:52.900 who should be stopping you, right? 00:42:52.900 --> 00:42:59.110 And so in some sense, laptops are useful for note taking. 00:42:59.110 --> 00:43:02.370 They're also useful for non-class activities. 00:43:02.370 --> 00:43:03.940 And as you say, each student should 00:43:03.940 --> 00:43:06.700 be able to choose for themselves what's good. 00:43:06.700 --> 00:43:12.797 And that may involve paying attention or not. 00:43:12.797 --> 00:43:15.130 Now, what are sort of some psychological considerations? 00:43:15.130 --> 00:43:16.672 We had one of them, which essentially 00:43:16.672 --> 00:43:19.640 distracting others, which essentially is the externality. 00:43:19.640 --> 00:43:20.870 What else? 00:43:20.870 --> 00:43:21.370 Yeah. 00:43:21.370 --> 00:43:24.987 AUDIENCE: Temptation and not valuing your future learning? 00:43:24.987 --> 00:43:26.320 FRANK SCHILBACH: Right, exactly. 00:43:26.320 --> 00:43:28.450 You might sort of have all the best 00:43:28.450 --> 00:43:31.092 intentions of taking notes, what I said previously. 00:43:31.092 --> 00:43:32.800 But then, you know, it gets a little dull 00:43:32.800 --> 00:43:34.922 and boring at minute 40. 00:43:34.922 --> 00:43:36.880 And you might sort of be inclined to sort think 00:43:36.880 --> 00:43:38.860 about other stuff and sort of start 00:43:38.860 --> 00:43:42.010 surfing an internet or the like or start chatting or whatever. 00:43:42.010 --> 00:43:44.170 That's sort of essentially limited self-control one 00:43:44.170 --> 00:43:45.380 way or the other. 00:43:45.380 --> 00:43:45.880 Yeah. 00:43:45.880 --> 00:43:47.755 AUDIENCE: Yeah. 00:43:47.755 --> 00:43:50.005 So I think even if you don't fall into the temptation, 00:43:50.005 --> 00:43:52.780 you kind of have to spend energy resisting the temptation 00:43:52.780 --> 00:43:54.280 and that might kind of distract you. 00:43:54.280 --> 00:43:55.238 FRANK SCHILBACH: Right. 00:43:55.238 --> 00:43:56.710 So there's some cognitive resources 00:43:56.710 --> 00:43:59.258 potentially from that. 00:43:59.258 --> 00:44:00.550 That's a very nice observation. 00:44:00.550 --> 00:44:01.120 Exactly. 00:44:01.120 --> 00:44:02.752 That also might be true. 00:44:02.752 --> 00:44:04.210 There's another part to that, which 00:44:04.210 --> 00:44:07.810 is people tend to overestimate how much they can multitask. 00:44:07.810 --> 00:44:10.060 And there's a large number of experiments 00:44:10.060 --> 00:44:12.220 that say, well, I'm actually paying attention. 00:44:12.220 --> 00:44:14.020 I'm just sort of reading some other stuff. 00:44:14.020 --> 00:44:16.300 I'm chatting with my friend. 00:44:16.300 --> 00:44:19.540 And people actually think they pay attention, and they learn. 00:44:19.540 --> 00:44:20.780 Trust me, they do not. 00:44:20.780 --> 00:44:24.100 So essentially, there's a large, large literature on people 00:44:24.100 --> 00:44:26.710 thinking that they can multitask when, in fact, they cannot. 00:44:26.710 --> 00:44:28.330 And this is a very human thing to do. 00:44:28.330 --> 00:44:31.420 You might think you can study for an exam by watching TV. 00:44:31.420 --> 00:44:34.610 Chances are you're not studying very well. 00:44:34.610 --> 00:44:39.730 So that's essentially some form of overconfidence. 00:44:39.730 --> 00:44:40.390 Right. 00:44:40.390 --> 00:44:41.890 And then there's another part, which 00:44:41.890 --> 00:44:45.590 is a sort of somewhat different psychological consideration, 00:44:45.590 --> 00:44:48.963 which is people tend to not like hard paternalism, as in, 00:44:48.963 --> 00:44:49.630 if you sort of-- 00:44:49.630 --> 00:44:52.360 we talk about this at the very end when we talk about policy. 00:44:52.360 --> 00:44:54.280 People don't like certain hard rules 00:44:54.280 --> 00:44:55.960 about you're allowed to do this, or you 00:44:55.960 --> 00:44:57.160 don't allow to do x or y. 00:44:57.160 --> 00:44:58.870 You have to come to class or whatever. 00:44:58.870 --> 00:45:00.632 People tend to not like that. 00:45:00.632 --> 00:45:02.340 That's more another sort of consideration 00:45:02.340 --> 00:45:07.110 if you think about what is the right policy to do. 00:45:07.110 --> 00:45:09.210 Now, what policy solutions could we do? 00:45:09.210 --> 00:45:10.678 What laptop policies have you seen? 00:45:10.678 --> 00:45:11.970 Or how do you think about them? 00:45:17.500 --> 00:45:19.790 Maybe somebody else? 00:45:19.790 --> 00:45:20.320 Yes. 00:45:20.320 --> 00:45:22.820 AUDIENCE: I haven't seen it, but let the people with laptops 00:45:22.820 --> 00:45:23.745 sit in the back. 00:45:23.745 --> 00:45:24.620 FRANK SCHILBACH: Yes. 00:45:24.620 --> 00:45:26.530 You could do that. 00:45:26.530 --> 00:45:27.970 Yeah. 00:45:27.970 --> 00:45:30.760 And sort of that's kind of minimizing the externality 00:45:30.760 --> 00:45:32.860 potentially. 00:45:32.860 --> 00:45:35.770 It's not really helping with self-control, I guess, right? 00:45:35.770 --> 00:45:37.840 Because particularly, if you sit in the back, 00:45:37.840 --> 00:45:40.000 nobody sees what you're doing. 00:45:40.000 --> 00:45:43.480 And you might be so inclined to do all sorts of things. 00:45:43.480 --> 00:45:44.420 Yes. 00:45:44.420 --> 00:45:47.070 AUDIENCE: I've had professors say a hard no on laptops, 00:45:47.070 --> 00:45:49.390 but if you have special needs or if you 00:45:49.390 --> 00:45:50.890 think you're a special case, you can 00:45:50.890 --> 00:45:52.240 tell the professor [INAUDIBLE]. 00:45:52.240 --> 00:45:53.198 FRANK SCHILBACH: Right. 00:45:53.198 --> 00:45:55.210 So that's sort of like hard paternalism 00:45:55.210 --> 00:45:59.360 with some exceptions potentially. 00:45:59.360 --> 00:46:00.800 Yeah. 00:46:00.800 --> 00:46:03.320 AUDIENCE: I've seen only devices that 00:46:03.320 --> 00:46:06.680 are flat on the table allowed for [INAUDIBLE].. 00:46:06.680 --> 00:46:09.730 FRANK SCHILBACH: For note-taking, yeah. 00:46:09.730 --> 00:46:12.460 AUDIENCE: I've seen TAs sitting in the back 00:46:12.460 --> 00:46:15.320 and keeping track of class participation. 00:46:15.320 --> 00:46:18.490 So if they see you doing something not academic, 00:46:18.490 --> 00:46:20.573 they could technically ding you for it. 00:46:20.573 --> 00:46:21.490 FRANK SCHILBACH: Yeah. 00:46:21.490 --> 00:46:22.490 That's what they're for. 00:46:22.490 --> 00:46:23.890 No. 00:46:23.890 --> 00:46:26.620 No. 00:46:26.620 --> 00:46:28.161 Anything else? 00:46:28.161 --> 00:46:30.036 AUDIENCE: I've seen classes where you're just 00:46:30.036 --> 00:46:31.869 allowed to use your laptop because, I guess, 00:46:31.869 --> 00:46:33.994 you're an adult. And you're responsible for letting 00:46:33.994 --> 00:46:34.510 you learn. 00:46:34.510 --> 00:46:35.200 FRANK SCHILBACH: Right, exactly. 00:46:35.200 --> 00:46:36.340 That's sort of laissez-faire and saying 00:46:36.340 --> 00:46:37.725 you know what's best for you. 00:46:37.725 --> 00:46:39.100 In some sense, since this is sort 00:46:39.100 --> 00:46:42.855 of a psychology and behavioral class, 00:46:42.855 --> 00:46:47.090 I tend to disagree with some of those assumptions. 00:46:47.090 --> 00:46:47.840 But, yes, exactly. 00:46:47.840 --> 00:46:51.330 That's laissez-faire. 00:46:51.330 --> 00:46:52.324 Any other thoughts? 00:46:57.070 --> 00:46:57.610 OK. 00:46:57.610 --> 00:46:58.960 So then here they are. 00:46:58.960 --> 00:47:00.295 So there's laissez-faire. 00:47:00.295 --> 00:47:02.920 There's educational intervention which essentially is providing 00:47:02.920 --> 00:47:06.580 people information about laptops or what 00:47:06.580 --> 00:47:08.160 laptops do for learning. 00:47:08.160 --> 00:47:10.390 There's some experiments I'll show you in a second. 00:47:10.390 --> 00:47:12.140 Educational interventions tend to actually 00:47:12.140 --> 00:47:13.340 not work particularly well. 00:47:13.340 --> 00:47:15.257 So essentially, just giving people information 00:47:15.257 --> 00:47:17.200 tends to not change behavior often 00:47:17.200 --> 00:47:19.030 in the way we'd like to do that. 00:47:19.030 --> 00:47:20.540 You could tax laptop use. 00:47:20.540 --> 00:47:22.060 You say, it's costly to, but that 00:47:22.060 --> 00:47:24.310 would be the typical sort of public economic sort 00:47:24.310 --> 00:47:24.933 of solution. 00:47:24.933 --> 00:47:27.100 I think I'm not allowed to take money from you guys. 00:47:27.100 --> 00:47:29.830 So I might not do that. 00:47:29.830 --> 00:47:32.075 You could ban laptops. 00:47:32.075 --> 00:47:34.450 That was a previous suggestion to say except for students 00:47:34.450 --> 00:47:36.250 with medical need. 00:47:36.250 --> 00:47:38.530 You could make a non-laptop section the default 00:47:38.530 --> 00:47:42.490 and let students opt out or opt in depending on that. 00:47:42.490 --> 00:47:44.890 It's essentially saying, the default choice 00:47:44.890 --> 00:47:46.810 is one, essentially no laptops. 00:47:46.810 --> 00:47:50.110 But you can opt into that by emailing to a TA 00:47:50.110 --> 00:47:51.790 and then use it in a certain section. 00:47:51.790 --> 00:47:54.040 We'll talk about, in the problem set, a little bit why 00:47:54.040 --> 00:47:55.805 that might be a good idea. 00:47:55.805 --> 00:47:58.180 You could also set up an active choice between the laptop 00:47:58.180 --> 00:48:00.910 and the no laptop sections. 00:48:00.910 --> 00:48:03.020 So here's the educational interventions. 00:48:03.020 --> 00:48:05.260 The evidence-- and there's very clear evidence 00:48:05.260 --> 00:48:07.510 from various settings that shows that laptops in class 00:48:07.510 --> 00:48:10.570 are not good for learning on average. 00:48:10.570 --> 00:48:13.250 There's a very nice article by Susan Dynarski in The New York 00:48:13.250 --> 00:48:13.750 Times. 00:48:13.750 --> 00:48:15.898 We'll put this online as well that you could read. 00:48:15.898 --> 00:48:17.440 But essentially, one of these studies 00:48:17.440 --> 00:48:19.180 is a randomized controlled trial that, 00:48:19.180 --> 00:48:21.970 actually, an MIT grad student, a former MIT grad student had 00:48:21.970 --> 00:48:24.190 done, in an intro econ class. 00:48:24.190 --> 00:48:25.485 This is Carter et al. 00:48:25.485 --> 00:48:26.860 And essentially what they find is 00:48:26.860 --> 00:48:29.470 that allowing computers in class reduced test scores 00:48:29.470 --> 00:48:31.360 by 0.18 standard deviations. 00:48:31.360 --> 00:48:33.430 That's quite a large number. 00:48:33.430 --> 00:48:36.610 That's essentially saying they did randomize across classes, 00:48:36.610 --> 00:48:39.580 like at the classroom level, whether they allowed laptops 00:48:39.580 --> 00:48:40.870 in class or not. 00:48:40.870 --> 00:48:42.850 Interestingly, they found negative effects 00:48:42.850 --> 00:48:46.120 both of unconstrained laptops, like on any laptop use, 00:48:46.120 --> 00:48:48.620 but also the flat tablet solution. 00:48:48.620 --> 00:48:51.070 So even the flat tablet solution was worse 00:48:51.070 --> 00:48:54.252 than sort of the note-taking with pen and paper. 00:48:54.252 --> 00:48:55.960 You know, and there's a bunch of evidence 00:48:55.960 --> 00:48:57.290 that sort of shows that. 00:48:57.290 --> 00:48:58.707 So in this class, what we're going 00:48:58.707 --> 00:49:01.120 to have a sort of a version of an opt-in policy, which is 00:49:01.120 --> 00:49:03.040 there's going to be a laptop section starting 00:49:03.040 --> 00:49:04.780 next class, which is going to be in the front on one 00:49:04.780 --> 00:49:05.860 side of the class. 00:49:05.860 --> 00:49:09.430 The reason being we're trying to sort of minimize externalities. 00:49:09.430 --> 00:49:11.620 It's in the front, not in the back. 00:49:11.620 --> 00:49:14.380 The reason being that, if you're in the very back, 00:49:14.380 --> 00:49:16.000 then there's no supervision. 00:49:16.000 --> 00:49:19.660 I'll, I guess, ask some TAs to sit sort of in the back. 00:49:19.660 --> 00:49:22.150 If there's too much other activities going on, 00:49:22.150 --> 00:49:24.907 maybe there won't be a laptop section anymore. 00:49:24.907 --> 00:49:26.740 Anyway, there will be sort of considerations 00:49:26.740 --> 00:49:28.085 about that in the problem set. 00:49:28.085 --> 00:49:29.710 But the point of all of that is to say, 00:49:29.710 --> 00:49:33.100 if you just had the standard economic considerations, 00:49:33.100 --> 00:49:35.710 you would make certain policies on very sort 00:49:35.710 --> 00:49:38.380 of simple considerations by missing important facts. 00:49:38.380 --> 00:49:40.845 And that might be sort of the psychological externality. 00:49:40.845 --> 00:49:44.050 It might be the self-control problems, the overestimation 00:49:44.050 --> 00:49:46.910 of people's ability to multitask and so on and so forth. 00:49:46.910 --> 00:49:48.460 And all of those things might sort of 00:49:48.460 --> 00:49:50.590 interfere with people's optimal choices. 00:49:50.590 --> 00:49:54.040 And that makes certain more paternalistic policies 00:49:54.040 --> 00:49:56.710 potentially more promising. 00:49:56.710 --> 00:49:58.990 But again, in the problem set, you'll think about some 00:49:58.990 --> 00:50:01.390 of these solutions. 00:50:01.390 --> 00:50:02.820 Any questions on the laptops? 00:50:02.820 --> 00:50:03.680 Yes. 00:50:03.680 --> 00:50:05.805 AUDIENCE: I might be thinking of a different study, 00:50:05.805 --> 00:50:08.080 but if this is the same one, I got 00:50:08.080 --> 00:50:11.110 that, in their other classes, they previously were not 00:50:11.110 --> 00:50:12.920 allowed to use laptops. 00:50:12.920 --> 00:50:17.230 So it might be that they had not optimized themselves 00:50:17.230 --> 00:50:20.180 for laptop usage during class. 00:50:20.180 --> 00:50:21.950 FRANK SCHILBACH: That's interesting. 00:50:21.950 --> 00:50:24.890 So I think there are several studies that show that. 00:50:24.890 --> 00:50:27.330 I don't know the details on that. 00:50:27.330 --> 00:50:31.070 But if you can send that to me, I'm happy to look at that 00:50:31.070 --> 00:50:32.100 and reconsider. 00:50:32.100 --> 00:50:34.790 I think my view overall is that sort of the existing evidence 00:50:34.790 --> 00:50:37.070 essentially shows laptops are bad. 00:50:37.070 --> 00:50:40.893 Since I'm interested in your learning, 00:50:40.893 --> 00:50:41.810 that's what I go with. 00:50:41.810 --> 00:50:44.930 But I'm happy to be educated. 00:50:44.930 --> 00:50:48.255 I'll try and put the slides online always the night 00:50:48.255 --> 00:50:50.630 before in case you want to print them out or look at them 00:50:50.630 --> 00:50:53.330 or whatever during class. 00:50:53.330 --> 00:50:56.150 You're welcome to do that. 00:50:56.150 --> 00:50:56.690 OK. 00:50:56.690 --> 00:50:58.550 So then let me tell you very briefly 00:50:58.550 --> 00:51:01.640 about different topics of the class 00:51:01.640 --> 00:51:03.750 and what we're going to cover. 00:51:03.750 --> 00:51:05.840 So first, we're going to have an introduction 00:51:05.840 --> 00:51:07.850 and overview of today's introduction lecture. 00:51:07.850 --> 00:51:10.080 The overview will be on Wednesday, 00:51:10.080 --> 00:51:12.883 which will providing you an overview of what 00:51:12.883 --> 00:51:15.050 the different topics are and how to think about them 00:51:15.050 --> 00:51:17.000 and what evidence do we actually have. 00:51:17.000 --> 00:51:18.600 So I was kind of a little bit handwavy 00:51:18.600 --> 00:51:21.980 in a sense of showing you lots of flashy pictures of people 00:51:21.980 --> 00:51:22.850 misbehaving. 00:51:22.850 --> 00:51:25.010 But in fact, I'm going to show you some more 00:51:25.010 --> 00:51:27.720 sort of rigorous evidence that we have that, in fact, 00:51:27.720 --> 00:51:30.920 some of the assumptions of the classical model 00:51:30.920 --> 00:51:33.750 are violated and sort of how we think about them 00:51:33.750 --> 00:51:35.420 and how we might sort incorporate them 00:51:35.420 --> 00:51:36.350 into economics. 00:51:36.350 --> 00:51:38.838 So that's going to be an overview on Wednesday. 00:51:38.838 --> 00:51:40.880 Then we're going to talk a lot about preferences. 00:51:40.880 --> 00:51:44.750 You can think about essentially, when people make choices, 00:51:44.750 --> 00:51:46.040 there's a utility function. 00:51:46.040 --> 00:51:49.070 And sort of one set of behavioral economics issues 00:51:49.070 --> 00:51:51.820 are changes to the utility function. 00:51:51.820 --> 00:51:55.020 This might be time preferences and people's self-control, 00:51:55.020 --> 00:51:56.510 which we already discussed. 00:51:56.510 --> 00:51:58.970 It might be risk preferences, how people think about risk, 00:51:58.970 --> 00:52:00.560 how they think about gains and losses 00:52:00.560 --> 00:52:03.260 and how their preferences are reference-dependent. 00:52:03.260 --> 00:52:07.130 This is what a student was saying earlier about how 00:52:07.130 --> 00:52:09.590 you evaluate something might depend 00:52:09.590 --> 00:52:12.260 a lot on your expectation of that outcome as opposed 00:52:12.260 --> 00:52:15.230 to just the outcome itself. 00:52:15.230 --> 00:52:17.820 Then we're going to talk quite a bit about social preferences. 00:52:17.820 --> 00:52:21.560 This is sort of like how much we weigh 00:52:21.560 --> 00:52:25.400 others' utility and our utility function 00:52:25.400 --> 00:52:28.040 or their consumption and our utility function. 00:52:28.040 --> 00:52:30.020 We'll do some experiments in class 00:52:30.020 --> 00:52:32.757 on that, which should be a lot of fun, 00:52:32.757 --> 00:52:34.340 and then discuss various issues about, 00:52:34.340 --> 00:52:35.972 A, how much people care about others 00:52:35.972 --> 00:52:37.430 and how they're affected by others' 00:52:37.430 --> 00:52:39.680 behavior and social influences. 00:52:39.680 --> 00:52:42.320 So that's kind of the first half of the class broadly 00:52:42.320 --> 00:52:45.710 about changes in people's preferences. 00:52:45.710 --> 00:52:48.410 Then we're going to talk more about beliefs broadly 00:52:48.410 --> 00:52:53.530 speaking and sort of how people view information in the world. 00:52:53.530 --> 00:52:55.160 We talk a little about emotions, what's 00:52:55.160 --> 00:52:57.020 called projection and attribution bias. 00:52:57.020 --> 00:52:59.220 These are issues about, for example, 00:52:59.220 --> 00:53:00.950 when people are hungry or tired, they 00:53:00.950 --> 00:53:03.780 might make quite different choices. 00:53:03.780 --> 00:53:06.960 And so that might affect their preferences. 00:53:06.960 --> 00:53:09.540 But it might also affect their beliefs in some sense. 00:53:09.540 --> 00:53:10.890 If you think about how you're going to behave, 00:53:10.890 --> 00:53:12.973 if you're hungry right now, it's very hard for you 00:53:12.973 --> 00:53:15.620 to believe or think about that you might not 00:53:15.620 --> 00:53:17.000 be hungry in the future. 00:53:17.000 --> 00:53:19.010 This is sort of the classical example is 00:53:19.010 --> 00:53:21.402 shopping on an empty stomach. 00:53:21.402 --> 00:53:23.360 There's lots of different applications of that. 00:53:23.360 --> 00:53:25.730 There's people buy convertibles when it's sunny 00:53:25.730 --> 00:53:29.080 and then return them when it rains. 00:53:29.080 --> 00:53:30.830 But there's also much more serious issues, 00:53:30.830 --> 00:53:32.330 in particular things like depression. 00:53:32.330 --> 00:53:33.955 For example, when people are depressed, 00:53:33.955 --> 00:53:36.170 it's very hard for them to think about how 00:53:36.170 --> 00:53:38.420 it might feel when they're not depressed anymore 00:53:38.420 --> 00:53:39.050 in the future. 00:53:39.050 --> 00:53:41.300 So we'll talk about sort of projection and attribution 00:53:41.300 --> 00:53:42.758 bias, which is kind of these biases 00:53:42.758 --> 00:53:45.483 and how to think about people states of the world. 00:53:45.483 --> 00:53:47.150 How about limited attention, that people 00:53:47.150 --> 00:53:49.040 sort of don't pay attention to certain things 00:53:49.040 --> 00:53:51.332 in the world and sort of how that might affect people's 00:53:51.332 --> 00:53:53.630 behavior, how we might sort of exploit that potentially 00:53:53.630 --> 00:53:55.490 if we're taxing them or how we might sort of direct 00:53:55.490 --> 00:53:56.865 their attention to certain things 00:53:56.865 --> 00:53:58.880 and might improve their behavior? 00:53:58.880 --> 00:54:01.280 Similarly, then we talk about beliefs and learning. 00:54:01.280 --> 00:54:03.170 This is kind of like what information 00:54:03.170 --> 00:54:04.655 do they have available-- 00:54:04.655 --> 00:54:08.120 A, how they update their beliefs when they get information. 00:54:08.120 --> 00:54:10.400 And are they able to process information well? 00:54:10.400 --> 00:54:12.440 B, is there demand for information? 00:54:12.440 --> 00:54:14.387 Do people get utility from beliefs? 00:54:14.387 --> 00:54:16.220 This is what I was talking about like health 00:54:16.220 --> 00:54:18.650 behaviors or the health information where people might 00:54:18.650 --> 00:54:20.180 have motivated beliefs in the sense 00:54:20.180 --> 00:54:24.890 that they like to believe certain things when, 00:54:24.890 --> 00:54:26.360 in fact, they're not true in part 00:54:26.360 --> 00:54:28.490 because it makes them happy or in part because they 00:54:28.490 --> 00:54:32.292 want to be right about something or their party or whatever. 00:54:32.292 --> 00:54:34.250 We'll talk a little about augmental accounting, 00:54:34.250 --> 00:54:37.893 which is people tend to sort of narrowly bracket their choices. 00:54:37.893 --> 00:54:40.310 They might sort of have certain accounts in their behavior 00:54:40.310 --> 00:54:43.190 and decide sort of separately as opposed to aggregating 00:54:43.190 --> 00:54:45.260 their behavior as a whole. 00:54:45.260 --> 00:54:47.690 And that's sort of an issue that's 00:54:47.690 --> 00:54:50.510 less researched, but quite interesting overall. 00:54:50.510 --> 00:54:52.220 Then we're going to move towards sort 00:54:52.220 --> 00:54:54.230 of more radical deviations, if you want, 00:54:54.230 --> 00:54:57.440 from the standard model, which is things malleability 00:54:57.440 --> 00:54:58.997 and accessibility of preferences, 00:54:58.997 --> 00:55:01.580 which is to say people might not actually know what they want. 00:55:01.580 --> 00:55:03.955 They might not even understand what their preferences are 00:55:03.955 --> 00:55:06.080 or their preferences are easily manipulable. 00:55:06.080 --> 00:55:10.610 I can sort make you choose A or B or A versus B or B versus A, 00:55:10.610 --> 00:55:13.340 depending on what kind of situation I put you in. 00:55:13.340 --> 00:55:17.150 And you might not even notice that I'm doing that. 00:55:17.150 --> 00:55:18.785 And that makes things a lot trickier. 00:55:18.785 --> 00:55:20.160 Because then, in some sense, it's 00:55:20.160 --> 00:55:22.000 much harder to sort of say should 00:55:22.000 --> 00:55:24.000 the government or any sort of other policy maker 00:55:24.000 --> 00:55:25.980 choose A or B if we don't even know 00:55:25.980 --> 00:55:28.397 what people's preferences are, when people don't even know 00:55:28.397 --> 00:55:29.730 what their own preferences are? 00:55:29.730 --> 00:55:31.890 We're going to talk about happiness and mental health. 00:55:31.890 --> 00:55:33.900 Just kind of broadly speaking, what makes people happy? 00:55:33.900 --> 00:55:35.900 And can we think about that, in particular, sort 00:55:35.900 --> 00:55:37.698 of financial choices and others? 00:55:37.698 --> 00:55:39.240 I'll tell you a little bit about some 00:55:39.240 --> 00:55:40.698 of the work on mental health that I 00:55:40.698 --> 00:55:43.920 have been doing, thinking about kind of how mental health might 00:55:43.920 --> 00:55:46.355 affect economic behaviors and choices 00:55:46.355 --> 00:55:48.480 and, in part, how people's demand for mental health 00:55:48.480 --> 00:55:52.900 interventions might be shaped by influences of others. 00:55:52.900 --> 00:55:56.190 We're going to talk about gender and racial discrimination, 00:55:56.190 --> 00:55:59.220 which sometimes there's sort of classical and neoclassical 00:55:59.220 --> 00:56:00.535 models of discrimination. 00:56:00.535 --> 00:56:02.160 But we, in particular, think about sort 00:56:02.160 --> 00:56:04.590 of issues of discrimination that are sort of not 00:56:04.590 --> 00:56:08.760 rational in the sense of think about unfair, in some sense, 00:56:08.760 --> 00:56:10.105 discrimination. 00:56:10.105 --> 00:56:11.730 Finally, we're going to talk about sort 00:56:11.730 --> 00:56:14.130 of policy and paternalism. 00:56:14.130 --> 00:56:16.170 One broad issue is about-- and this goes back 00:56:16.170 --> 00:56:19.740 to malleability and inaccessibility of preferences, 00:56:19.740 --> 00:56:22.680 which is essentially ways in which choices 00:56:22.680 --> 00:56:26.040 can be affected through frames, defaults, and nudges. 00:56:26.040 --> 00:56:28.920 And sometimes I say, I can manipulate your choice 00:56:28.920 --> 00:56:30.630 architecture in certain ways that 00:56:30.630 --> 00:56:32.570 make you choose certain things. 00:56:32.570 --> 00:56:35.070 You know, I might sort of send you letters to do your taxes. 00:56:35.070 --> 00:56:39.000 I can sort of frame or set certain defaults 00:56:39.000 --> 00:56:43.690 for organ donations or for savings choices and the like. 00:56:43.690 --> 00:56:47.175 And that might have profound effects on people's behavior. 00:56:47.175 --> 00:56:49.050 Once we go through that in a sense of showing 00:56:49.050 --> 00:56:52.110 you some evidence that it's possible to do that, we're 00:56:52.110 --> 00:56:54.450 going to talk a little bit about policy and paternalism 00:56:54.450 --> 00:56:57.810 in a sense of say, OK, now, if I know I can change 00:56:57.810 --> 00:57:00.540 your behavior in certain ways, what kinds of policies 00:57:00.540 --> 00:57:01.200 should we do? 00:57:01.200 --> 00:57:02.670 And should we, in fact, do that? 00:57:02.670 --> 00:57:04.770 Or should we tax people? 00:57:04.770 --> 00:57:07.350 Should we set certain frames and defaults? 00:57:07.350 --> 00:57:09.360 And you know, are we potentially making 00:57:09.360 --> 00:57:11.730 some people worse off while making some people better 00:57:11.730 --> 00:57:14.400 off by doing these kinds of policies? 00:57:14.400 --> 00:57:17.715 There's often also some ethical issues associated with that. 00:57:17.715 --> 00:57:19.840 And then, finally, I'm going to talk about poverty. 00:57:19.840 --> 00:57:23.010 This is sort of mostly the research that I do 00:57:23.010 --> 00:57:25.560 and thinking about poverty issues 00:57:25.560 --> 00:57:27.810 through the lens of psychology. 00:57:27.810 --> 00:57:30.697 Again, I'll tell you a little bit about the work I do myself. 00:57:30.697 --> 00:57:32.280 Partially, we get to talk a little bit 00:57:32.280 --> 00:57:35.280 about so how financial constraints or just thinking 00:57:35.280 --> 00:57:38.460 about money affects people's behavior and then, second, how 00:57:38.460 --> 00:57:40.530 others' issues related to poverty 00:57:40.530 --> 00:57:45.990 might shape people's choices, decision making, 00:57:45.990 --> 00:57:48.835 and their labor market outcomes, earnings, and so on 00:57:48.835 --> 00:57:50.460 and where there's potentially something 00:57:50.460 --> 00:57:53.430 that you might want to call a psychological or behavioral 00:57:53.430 --> 00:57:55.160 poverty trap. 00:57:55.160 --> 00:57:57.110 Any questions on these topics? 00:58:03.350 --> 00:58:03.850 OK. 00:58:03.850 --> 00:58:06.430 So then readings for next time, this is Wednesday. 00:58:06.430 --> 00:58:10.360 Please read-- there's a paper on the course website by Matthew 00:58:10.360 --> 00:58:14.560 Rabin, who is one of sort of the all-stars 00:58:14.560 --> 00:58:16.360 in behavioral economics. 00:58:16.360 --> 00:58:19.047 20 years ago, he wrote a very useful perspective 00:58:19.047 --> 00:58:21.130 on psychology and economics that sort of discusses 00:58:21.130 --> 00:58:23.740 a lot of the issues of behavioral economics, 00:58:23.740 --> 00:58:25.840 how to think about that. 00:58:25.840 --> 00:58:28.000 Again, read sections one and two. 00:58:28.000 --> 00:58:29.800 We're not going to test you specifically 00:58:29.800 --> 00:58:31.870 on very specific things, but you should sort of 00:58:31.870 --> 00:58:35.350 be able to sort of remember, at least roughly, what's 00:58:35.350 --> 00:58:37.170 in that paper.