WEBVTT 00:00:00.000 --> 00:00:01.932 [SQUEAKING] 00:00:01.932 --> 00:00:03.864 [RUSTLING] 00:00:03.864 --> 00:00:07.728 [CLICKING] 00:00:10.630 --> 00:00:13.400 All right, I'm going to get started. 00:00:13.400 --> 00:00:18.820 This is lecture 22 on 14.13 on happiness and mental health. 00:00:18.820 --> 00:00:21.670 OK, so what are we going to talk about today? 00:00:21.670 --> 00:00:23.320 Happiness and subjective well-being 00:00:23.320 --> 00:00:24.982 and a little bit of mental health. 00:00:24.982 --> 00:00:27.190 So we're going to talk about rationality and revealed 00:00:27.190 --> 00:00:29.320 preferences and what can we learn from choices, 00:00:29.320 --> 00:00:32.020 and what does that tell us about happiness. 00:00:32.020 --> 00:00:35.123 We also talked about utility and how we can measure utility. 00:00:35.123 --> 00:00:36.915 In particular, how do we measure happiness, 00:00:36.915 --> 00:00:40.030 and is that a good or bad idea? 00:00:40.030 --> 00:00:41.770 And then at the very end, we'll have, 00:00:41.770 --> 00:00:44.690 hopefully, a llama or a goat visit, 00:00:44.690 --> 00:00:48.420 which will at least improve my happiness, hopefully yours too. 00:00:48.420 --> 00:00:52.060 I've had some trouble communicating by email 00:00:52.060 --> 00:00:57.400 with the llama, goat, and their keepers. 00:00:57.400 --> 00:00:59.320 So I'm not exactly sure when they're 00:00:59.320 --> 00:01:01.210 going to show up, so they're scheduled-- 00:01:01.210 --> 00:01:04.879 the problem is, so they had only different certain time slots. 00:01:04.879 --> 00:01:07.990 The only time slot that was available was at 2:30, 00:01:07.990 --> 00:01:12.940 so I asked them to show up earlier, 2:20 or so. 00:01:12.940 --> 00:01:16.000 I hope that's actually going to happen, 00:01:16.000 --> 00:01:18.220 but I haven't heard back, so I'm not exactly sure. 00:01:18.220 --> 00:01:19.720 So it might have to be at like 2:30. 00:01:19.720 --> 00:01:22.660 I hope you don't have to run. 00:01:22.660 --> 00:01:24.460 There was also some confusion about which 00:01:24.460 --> 00:01:26.740 Zoom room code the llama should go into, 00:01:26.740 --> 00:01:30.540 and hopefully, they find our room, 00:01:30.540 --> 00:01:32.430 even though they don't have and MIT ID. 00:01:32.430 --> 00:01:34.320 So anyway, that's that. 00:01:34.320 --> 00:01:36.510 And then on Monday, we're going to talk about policy 00:01:36.510 --> 00:01:39.540 with people agents, which is very much the idea that, 00:01:39.540 --> 00:01:42.360 well, now if people are potentially making mistakes 00:01:42.360 --> 00:01:46.350 or have certain biases and we might 00:01:46.350 --> 00:01:48.390 think that we know about it, then 00:01:48.390 --> 00:01:52.140 perhaps we can do policies to improve their behaviors. 00:01:52.140 --> 00:01:55.500 And in some cases, that's straightforward. 00:01:55.500 --> 00:01:59.460 If you think people have biased beliefs or just 00:01:59.460 --> 00:02:02.190 have wrong information based on which they make information, 00:02:02.190 --> 00:02:04.440 then you might say, well, it should have just improved 00:02:04.440 --> 00:02:07.260 their information, and then they make better choices. 00:02:07.260 --> 00:02:09.330 In other situations, it's much trickier, 00:02:09.330 --> 00:02:12.180 because A, people might have certain preferences that 00:02:12.180 --> 00:02:14.100 are not stable over time, and then 00:02:14.100 --> 00:02:18.090 it's very tricky to figure out what is the objective function 00:02:18.090 --> 00:02:20.710 that we should maximize. 00:02:20.710 --> 00:02:25.530 Second, it might be tricky because who knows what we 00:02:25.530 --> 00:02:27.420 know versus what people know? 00:02:27.420 --> 00:02:30.030 So who knows that the government or anybody else 00:02:30.030 --> 00:02:33.480 knows actually better what's good for other people? 00:02:33.480 --> 00:02:35.640 So in some sense, if you want to intervene, 00:02:35.640 --> 00:02:38.580 you better be reasonably confident that what you're 00:02:38.580 --> 00:02:41.610 doing is actually well founded and you 00:02:41.610 --> 00:02:43.230 know better what's good for people 00:02:43.230 --> 00:02:44.670 than they know about themselves. 00:02:44.670 --> 00:02:47.070 And that gets you in tricky situations, 00:02:47.070 --> 00:02:49.010 because potentially, if you get it wrong, 00:02:49.010 --> 00:02:50.970 or maybe if you just misunderstand 00:02:50.970 --> 00:02:54.630 people's preferences, or their choices, or behaviors, 00:02:54.630 --> 00:02:56.070 then you might intervene and make 00:02:56.070 --> 00:02:59.700 things actually [AUDIO OUT] because you just misunderstood 00:02:59.700 --> 00:03:01.670 [AUDIO OUT]. 00:03:01.670 --> 00:03:02.640 OK. 00:03:02.640 --> 00:03:08.460 So what is rationality in classical economics? 00:03:08.460 --> 00:03:10.740 How do we [AUDIO OUT]? 00:03:10.740 --> 00:03:14.550 So broadly, we think that beliefs, preferences, 00:03:14.550 --> 00:03:19.680 and actions are rational if they are mutually consistent. 00:03:19.680 --> 00:03:20.900 What do we mean by that? 00:03:20.900 --> 00:03:22.770 Well, we mean by that that if we see 00:03:22.770 --> 00:03:26.010 certain behavior or certain actions, 00:03:26.010 --> 00:03:29.290 we can measure people's beliefs in certain ways. 00:03:29.290 --> 00:03:31.440 And then we essentially can find preferences 00:03:31.440 --> 00:03:33.870 that could rationalize these actions. 00:03:33.870 --> 00:03:36.000 And so that's a very odd definition, 00:03:36.000 --> 00:03:39.330 and without that definition, we can essentially 00:03:39.330 --> 00:03:43.230 say it's possible to be a rational cocaine addict. 00:03:43.230 --> 00:03:45.540 It's possible to rationally commit suicide. 00:03:45.540 --> 00:03:48.150 It's possible to rationally marry somebody 00:03:48.150 --> 00:03:49.970 who you met six hours ago. 00:03:49.970 --> 00:03:52.950 And it's also possible to be rational violent offender. 00:03:52.950 --> 00:03:54.810 If there's preferences that people have, 00:03:54.810 --> 00:03:57.870 you could write down some preferences that rationalize 00:03:57.870 --> 00:03:59.190 all of those things. 00:03:59.190 --> 00:04:02.850 And so [AUDIO OUT] in mainstream economics very much 00:04:02.850 --> 00:04:05.080 a main assumption. 00:04:05.080 --> 00:04:07.530 So the researchers' jobs is to identify 00:04:07.530 --> 00:04:11.170 preferences that are consistent with observed human behavior. 00:04:11.170 --> 00:04:13.070 So when you see people [AUDIO OUT] way, 00:04:13.070 --> 00:04:14.910 so then the question is like, what 00:04:14.910 --> 00:04:19.200 are preferences that rationalize that behavior? 00:04:19.200 --> 00:04:20.839 Now what's the idea? 00:04:20.839 --> 00:04:22.390 We talked about this already before. 00:04:22.390 --> 00:04:24.750 So actors make choices, economists 00:04:24.750 --> 00:04:27.630 observe their choices, and then [AUDIO OUT] 00:04:27.630 --> 00:04:33.720 generate these choices if an actor were perfectly rational. 00:04:33.720 --> 00:04:37.080 [AUDIO OUT] tend to give these imputed preferences 00:04:37.080 --> 00:04:38.010 normative meaning. 00:04:38.010 --> 00:04:41.070 That's essentially, if I know your preferences, that 00:04:41.070 --> 00:04:42.375 [AUDIO OUT]. 00:04:42.375 --> 00:04:47.610 And that's what you should be doing, as in, if you prefer B, 00:04:47.610 --> 00:04:49.980 I should be helping you to do that. 00:04:49.980 --> 00:04:53.610 That would be a good thing for you. 00:04:53.610 --> 00:04:56.910 Now just to give you an example of where 00:04:56.910 --> 00:04:59.115 this could lead you potentially astray, 00:04:59.115 --> 00:05:02.940 or where you might argue, is that really rational 00:05:02.940 --> 00:05:06.900 versus not, for example, take Jack, who prefers 00:05:06.900 --> 00:05:09.000 taking cocaine over quitting. 00:05:09.000 --> 00:05:13.915 Jack has lots of speeches about wanting to quit, but in a way, 00:05:13.915 --> 00:05:15.040 he doesn't actually act on. 00:05:15.040 --> 00:05:18.150 That's just cheap talk. 00:05:18.150 --> 00:05:20.760 Jack might, in fact, be [AUDIO OUT] if he were clean. 00:05:20.760 --> 00:05:23.045 But getting clean is too costly, because there 00:05:23.045 --> 00:05:25.590 are withdrawal costs. 00:05:25.590 --> 00:05:30.060 Jack probably didn't [INAUDIBLE] [AUDIO OUT] tried cocaine, 00:05:30.060 --> 00:05:32.460 but this bad outcome was sufficiently unlikely that his 00:05:32.460 --> 00:05:34.830 early experiments with it made sense. 00:05:34.830 --> 00:05:37.560 So finally, cocaine should be legalized unless it 00:05:37.560 --> 00:05:39.120 generates externalities. 00:05:39.120 --> 00:05:42.210 That's essentially, people make choices. 00:05:42.210 --> 00:05:45.720 They make-- if Jack started taking it, 00:05:45.720 --> 00:05:49.605 it must have been that he preferred cocaine 00:05:49.605 --> 00:05:52.260 over not taking cocaine, even though there's 00:05:52.260 --> 00:05:54.180 a chance that [AUDIO OUT] get really addicted 00:05:54.180 --> 00:05:56.220 and really be miserable now. 00:05:56.220 --> 00:06:00.570 And probably even if he keeps taking cocaine and [AUDIO OUT] 00:06:00.570 --> 00:06:02.730 to quit-- well, if he doesn't do it, 00:06:02.730 --> 00:06:05.600 well, then it must be a rational choice for him 00:06:05.600 --> 00:06:08.430 to, not because it must be that the withdrawal 00:06:08.430 --> 00:06:10.110 costs are really high. 00:06:10.110 --> 00:06:12.690 Now we talked a little bit about this theory 00:06:12.690 --> 00:06:15.750 of rational addiction, much about Becker and Murphy. 00:06:15.750 --> 00:06:18.690 We talked about already that there's 00:06:18.690 --> 00:06:21.340 present bias or other factors involved, 00:06:21.340 --> 00:06:22.500 including biased beliefs. 00:06:22.500 --> 00:06:24.150 That might not be the case. 00:06:24.150 --> 00:06:27.450 [INAUDIBLE] understand that traditional economics, 00:06:27.450 --> 00:06:31.680 until very recently, would say that the rational thing to do 00:06:31.680 --> 00:06:33.810 with Jack here-- 00:06:33.810 --> 00:06:41.940 in fact, no good reason to make certain or any drugs illegal on 00:06:41.940 --> 00:06:42.780 externalities. 00:06:42.780 --> 00:06:45.285 So externalities would be like if Jack causes 00:06:45.285 --> 00:06:49.320 behavior that's bad for others. 00:06:49.320 --> 00:06:51.270 Then well, we shouldn't let him do 00:06:51.270 --> 00:06:54.590 that, because other people are harmed 00:06:54.590 --> 00:06:56.700 and you might not internalize [INAUDIBLE].. 00:06:56.700 --> 00:07:01.900 But unless that's the case, we shouldn't intervene in any way. 00:07:01.900 --> 00:07:06.910 OK, but so now, you might say, but what did we talk about 00:07:06.910 --> 00:07:09.110 in all [INAUDIBLE]? 00:07:09.110 --> 00:07:12.460 How is that consistent with what we talked about before? 00:07:12.460 --> 00:07:15.610 And the key of that question is the question of, 00:07:15.610 --> 00:07:18.830 do people act in their best interest? 00:07:18.830 --> 00:07:22.150 And so what we essentially assume often 00:07:22.150 --> 00:07:24.640 is that there is a rational relationship between people's 00:07:24.640 --> 00:07:28.540 choices, what they actually do, and the hedonic consequences 00:07:28.540 --> 00:07:30.760 of those choices, which are true well-being. 00:07:30.760 --> 00:07:32.260 And so what we assume is essentially 00:07:32.260 --> 00:07:35.500 that you make certain choices that maximize your well-being. 00:07:35.500 --> 00:07:38.110 Essentially what you learned in 14.01 or 14.03 00:07:38.110 --> 00:07:40.450 is essentially utility maximization, 00:07:40.450 --> 00:07:44.620 and that's at the heart of the rationality assumptions. 00:07:44.620 --> 00:07:46.073 So economists then tend to believe 00:07:46.073 --> 00:07:48.490 that most of the time, people act, at least approximately, 00:07:48.490 --> 00:07:50.650 in their best interest. 00:07:50.650 --> 00:07:53.110 Now as we talked about already, and this 00:07:53.110 --> 00:07:58.603 is our previous lecture about Nisbett and Wilson 00:07:58.603 --> 00:08:01.020 telling more than we can know, we should be actually quite 00:08:01.020 --> 00:08:02.700 skeptical about the fact that people actually 00:08:02.700 --> 00:08:04.290 know what they're doing and why they're doing it, 00:08:04.290 --> 00:08:07.200 because [INAUDIBLE],, as I showed you last time, in some sense, 00:08:07.200 --> 00:08:09.450 people often seem to have essentially no clue 00:08:09.450 --> 00:08:10.988 of why they do what they do. 00:08:10.988 --> 00:08:12.780 So it seems like a bit of an odd assumption 00:08:12.780 --> 00:08:16.890 to say, well, people are maximizing 00:08:16.890 --> 00:08:18.880 what's best for them. 00:08:18.880 --> 00:08:21.510 So now how can we check whether this assumption is appropriate? 00:08:21.510 --> 00:08:22.080 What would you do? 00:08:22.080 --> 00:08:23.640 So I'm telling you here some behaviors. 00:08:23.640 --> 00:08:25.260 How can you actually check whether this assumption 00:08:25.260 --> 00:08:26.177 is an appropriate one? 00:08:47.646 --> 00:08:48.660 Lauren. 00:08:48.660 --> 00:08:51.120 AUDIENCE: We could check their well-being 00:08:51.120 --> 00:08:54.300 after they make these decisions and if they're 00:08:54.300 --> 00:08:57.180 better or worse off. 00:08:57.180 --> 00:08:59.670 FRANK SCHILBACH: Yeah, and we have to somehow figure out 00:08:59.670 --> 00:09:03.750 some situations where we can track people's well-being. 00:09:03.750 --> 00:09:06.840 Ideally, we would have some sort of experiment. 00:09:06.840 --> 00:09:10.560 There's group A and B, and in group A [INAUDIBLE],, 00:09:10.560 --> 00:09:12.760 you let them choose, and in the other groups-- 00:09:12.760 --> 00:09:15.220 [INAUDIBLE] group C-- in some group, 00:09:15.220 --> 00:09:17.828 we essentially tell people, you should be doing A, 00:09:17.828 --> 00:09:19.620 doing one thing, and another group, we say, 00:09:19.620 --> 00:09:21.030 you should be doing another thing. 00:09:21.030 --> 00:09:22.470 And then there's another group where we 00:09:22.470 --> 00:09:23.762 have people essentially choose. 00:09:23.762 --> 00:09:24.845 And then you can look at-- 00:09:24.845 --> 00:09:26.040 track people's well-being. 00:09:26.040 --> 00:09:27.960 And then we could figure out, is the group 00:09:27.960 --> 00:09:29.760 that gets to choose actually doing better 00:09:29.760 --> 00:09:32.052 than maybe some groups where we actually just tell them 00:09:32.052 --> 00:09:33.390 what to do? 00:09:33.390 --> 00:09:39.450 And in some ways, or the work by Dan Ariely and others, 00:09:39.450 --> 00:09:41.460 if I can essentially manipulate you 00:09:41.460 --> 00:09:44.820 into making very drastically different choices by changing 00:09:44.820 --> 00:09:46.410 small things that really, arguably, 00:09:46.410 --> 00:09:48.270 should be irrelevant for your choice, 00:09:48.270 --> 00:09:51.000 like for example, the Social Security number-- 00:09:51.000 --> 00:09:52.800 for example, if I could manipulate 00:09:52.800 --> 00:09:55.750 you with anchoring you with your social security number 00:09:55.750 --> 00:09:58.920 and then say, would you like to purchase this item for $10, 00:09:58.920 --> 00:10:01.440 yes or no, and that affects your choice, 00:10:01.440 --> 00:10:04.180 well, then presumably, that's a clue in some ways 00:10:04.180 --> 00:10:06.780 that something is not going right. 00:10:06.780 --> 00:10:09.630 And so you are not necessarily optimizing, 00:10:09.630 --> 00:10:12.690 precisely because the Social Security number really 00:10:12.690 --> 00:10:15.042 should be entirely irrelevant, arguably. 00:10:15.042 --> 00:10:16.500 Some forms of what we talked about, 00:10:16.500 --> 00:10:19.140 like preference reversal, or some form of demand 00:10:19.140 --> 00:10:20.573 from commitment or the like. 00:10:20.573 --> 00:10:22.740 If that's the case, it seems to be that then there's 00:10:22.740 --> 00:10:25.530 self control problems involved or biased beliefs 00:10:25.530 --> 00:10:27.490 in various ways that we all talked about, 00:10:27.490 --> 00:10:29.550 which seems then sent to say, well, you 00:10:29.550 --> 00:10:30.900 keep saying you want one thing. 00:10:30.900 --> 00:10:34.610 You keep spending money on, say, healthy eating, but then you 00:10:34.610 --> 00:10:35.440 never-- 00:10:35.440 --> 00:10:37.410 the food all goes bad in your fridge, 00:10:37.410 --> 00:10:39.545 and you keep eating donuts. 00:10:39.545 --> 00:10:40.920 That seems to be something that's 00:10:40.920 --> 00:10:45.425 not quite-- don't seem to be maximizing your utility. 00:10:45.425 --> 00:10:46.800 But in general, it would be great 00:10:46.800 --> 00:10:48.383 if you could measure behavior and then 00:10:48.383 --> 00:10:50.640 the hedonic consequences of those behaviors, people's 00:10:50.640 --> 00:10:53.020 well-being, and if you then could say, 00:10:53.020 --> 00:10:56.850 well, option A or option B is better for people in general, 00:10:56.850 --> 00:10:59.190 or specifically when you ask them to choose, 00:10:59.190 --> 00:11:02.570 they can figure out what's best for them. 00:11:02.570 --> 00:11:05.420 Notice that you need to know the counterfactual here. 00:11:05.420 --> 00:11:08.420 You can't-- if I see you doing one thing and then track 00:11:08.420 --> 00:11:10.730 your well-being, it's not enough, 00:11:10.730 --> 00:11:12.723 because I don't know the counterfactual, 00:11:12.723 --> 00:11:14.890 whether you would be happy otherwise if you were not 00:11:14.890 --> 00:11:17.550 to choose that, OK. 00:11:17.550 --> 00:11:19.760 So great idea. 00:11:19.760 --> 00:11:21.330 So now, how do we actually do that? 00:11:21.330 --> 00:11:25.400 So it's useful to have some terminology here 00:11:25.400 --> 00:11:27.940 that, in particular, that Kahneman and others have 00:11:27.940 --> 00:11:28.630 introduced. 00:11:28.630 --> 00:11:33.290 And Kahneman, remember, this is from Tversky and Kahneman, 00:11:33.290 --> 00:11:35.450 the Nobel Prize winner, who worked 00:11:35.450 --> 00:11:39.080 a lot on loss aversion and prospect theory 00:11:39.080 --> 00:11:40.377 that we talked about. 00:11:40.377 --> 00:11:42.710 Then Kahneman also has quite a bit of work on well-being 00:11:42.710 --> 00:11:44.370 and happiness and so on. 00:11:44.370 --> 00:11:47.700 So and he introduces very useful terminology, 00:11:47.700 --> 00:11:51.440 which is-- they call decision utility and experience utility. 00:11:51.440 --> 00:11:53.330 So now what's decision utility? 00:11:53.330 --> 00:11:56.060 So economists tend to use the word utility or utility 00:11:56.060 --> 00:11:58.940 function to describe the preferences that rationalize 00:11:58.940 --> 00:12:00.620 observed choices. 00:12:00.620 --> 00:12:03.810 Essentially, if you took 14.01, 14.03, et cetera, 00:12:03.810 --> 00:12:05.720 people talk about the utility function. 00:12:05.720 --> 00:12:09.830 Essentially, it's the function that you write down that maps 00:12:09.830 --> 00:12:13.952 your choices into well-being. 00:12:13.952 --> 00:12:15.410 And the function that you maximize, 00:12:15.410 --> 00:12:19.040 or by maximizing it, essentially say, well, given this utility 00:12:19.040 --> 00:12:21.620 function, given your behavior, it 00:12:21.620 --> 00:12:24.620 must be this utility function or that utility function 00:12:24.620 --> 00:12:26.510 allows me to rationalize what you're doing. 00:12:26.510 --> 00:12:28.580 That actually allows me then to explain why 00:12:28.580 --> 00:12:30.800 you choose apples over bananas. 00:12:30.800 --> 00:12:34.040 And so Kahneman calls these revealed preferences 00:12:34.040 --> 00:12:35.720 the decision utility. 00:12:35.720 --> 00:12:38.810 That's essentially the preferences 00:12:38.810 --> 00:12:39.980 that rationalize decisions. 00:12:39.980 --> 00:12:43.940 When people make choices, that's the utility that they perceive 00:12:43.940 --> 00:12:48.260 or that they think they will perceive or receive 00:12:48.260 --> 00:12:50.150 as a consequence of their choices. 00:12:50.150 --> 00:12:55.920 That's essentially-- you might call that wanting or choosing. 00:12:55.920 --> 00:12:59.760 So quickly, for an addict, the decision utility 00:12:59.760 --> 00:13:03.720 of drug consumption exceeds the decision utility of quitting. 00:13:03.720 --> 00:13:08.670 So if that person makes the decision of taking drugs, 00:13:08.670 --> 00:13:11.970 it must be that their decision utility is higher. 00:13:11.970 --> 00:13:16.500 When making those choices, they prefer one option 00:13:16.500 --> 00:13:18.870 over the other, which, again, doesn't necessarily 00:13:18.870 --> 00:13:22.113 then mean that afterwards, the experience is actually better. 00:13:22.113 --> 00:13:24.030 But when making the choice of taking the drug, 00:13:24.030 --> 00:13:27.180 they think that's a better option for them. 00:13:27.180 --> 00:13:30.540 Now in contrast, Kahneman also talks 00:13:30.540 --> 00:13:34.920 about experience neutrality, which implies or talks 00:13:34.920 --> 00:13:37.830 about the hedonic consequences of choices. 00:13:37.830 --> 00:13:42.510 He calls these hedonic experiences experience utility. 00:13:42.510 --> 00:13:44.550 That's essentially the preferences 00:13:44.550 --> 00:13:46.620 that coincide with doing. 00:13:46.620 --> 00:13:49.950 So while you actually experience the whole consequences 00:13:49.950 --> 00:13:53.190 of your choice, what experience do you 00:13:53.190 --> 00:13:55.680 have, and what utility, or what's your well-being 00:13:55.680 --> 00:13:56.490 during that? 00:13:56.490 --> 00:13:58.380 That's actually going back a long time ago. 00:13:58.380 --> 00:14:02.280 If you look at some of the wards in economics, Bentham 00:14:02.280 --> 00:14:05.730 and others, they talk a lot about pleasure and pain, which 00:14:05.730 --> 00:14:07.230 is very much about the experience 00:14:07.230 --> 00:14:09.990 people have in their life, as opposed to utility, 00:14:09.990 --> 00:14:13.860 which is, again, the construct that you have at the time 00:14:13.860 --> 00:14:15.420 of [INAUDIBLE]. 00:14:15.420 --> 00:14:19.290 Now how do we measure these hedonic experiences 00:14:19.290 --> 00:14:20.723 of well-being? 00:14:20.723 --> 00:14:22.140 So there's two questions, I guess. 00:14:22.140 --> 00:14:23.710 One is, how can we measure this? 00:14:23.710 --> 00:14:26.610 If I wanted to know, what is your hedonic experience 00:14:26.610 --> 00:14:29.062 of this lecture right now, how would I do this? 00:14:29.062 --> 00:14:31.020 How do you enjoy, how much do you enjoy things? 00:14:31.020 --> 00:14:34.860 And then second, well, I can measure your experience 00:14:34.860 --> 00:14:37.340 right now, maybe this moment, maybe this hour, 00:14:37.340 --> 00:14:40.210 but somehow I need to aggregate those over time. 00:14:40.210 --> 00:14:43.847 I need to know, you took class A versus class B. 00:14:43.847 --> 00:14:45.930 Now I can measure you at different points of time, 00:14:45.930 --> 00:14:47.580 but somehow, then I need to aggregate 00:14:47.580 --> 00:14:49.020 this overall experience. 00:14:49.020 --> 00:14:51.640 And how people are doing that is [INAUDIBLE] some questions. 00:14:51.640 --> 00:14:53.140 Let's start with the first question, 00:14:53.140 --> 00:14:56.020 how do you measure people's well-being? 00:14:56.020 --> 00:14:58.540 If you wanted to learn about your classmates' 00:14:58.540 --> 00:15:01.630 experience in any type of situations, what would you do? 00:15:01.630 --> 00:15:03.610 What would you measure? 00:15:03.610 --> 00:15:06.520 One is you can ask them verbally, like, 00:15:06.520 --> 00:15:09.220 how did you like activity A, how do you like B, 00:15:09.220 --> 00:15:12.460 or feel [AUDIO OUT] now? 00:15:12.460 --> 00:15:15.050 So this is asking what people want. 00:15:15.050 --> 00:15:20.950 But also just look at their facial features or the like 00:15:20.950 --> 00:15:27.290 and try to see how much they're smiling or the like. 00:15:27.290 --> 00:15:30.280 Notice that that, you need to map into some happiness 00:15:30.280 --> 00:15:31.240 as well. 00:15:31.240 --> 00:15:35.140 In some sense, I could be smiling but be deeply unhappy, 00:15:35.140 --> 00:15:37.210 and then you would misinterpret that. 00:15:37.210 --> 00:15:41.170 So you'd have to have some function from facial feature 00:15:41.170 --> 00:15:44.020 to some underlying well-being that's somewhat 00:15:44.020 --> 00:15:45.370 grounded in some assumptions. 00:15:45.370 --> 00:15:47.787 It seems reasonable to think that when people are smiling, 00:15:47.787 --> 00:15:50.170 they're happy, and when they're crying, they're unhappy, 00:15:50.170 --> 00:15:54.520 but I'm just pointing out that that requires some underlying 00:15:54.520 --> 00:15:56.230 data to be able to map those things. 00:15:56.230 --> 00:16:00.700 One is the physiological measures such as heart rates, 00:16:00.700 --> 00:16:04.310 [INAUDIBLE], et cetera. 00:16:04.310 --> 00:16:07.780 And so there, again, I think commonsensical stuff, 00:16:07.780 --> 00:16:11.560 like if you're really afraid and your heart goes up, presumably 00:16:11.560 --> 00:16:12.980 that's not a good thing. 00:16:12.980 --> 00:16:15.938 But if you're really excited and your heart rate goes up, 00:16:15.938 --> 00:16:17.230 presumably that's a good thing. 00:16:17.230 --> 00:16:20.410 So one has to add some assumptions or some structure 00:16:20.410 --> 00:16:21.640 on these measures. 00:16:21.640 --> 00:16:25.690 In some sense, you're just measuring them by themselves. 00:16:25.690 --> 00:16:27.760 In general, it might be misleading, 00:16:27.760 --> 00:16:31.793 but with some structure, [INAUDIBLE] use that well. 00:16:31.793 --> 00:16:34.210 The other part is measuring-- there's lots of neuroscience 00:16:34.210 --> 00:16:35.950 and other work on this. 00:16:35.950 --> 00:16:38.500 Measuring people's brain and parts of the brain, 00:16:38.500 --> 00:16:42.710 and seeing which parts light up at which times 00:16:42.710 --> 00:16:43.820 could be another option. 00:16:43.820 --> 00:16:46.120 Of course, again there, you need to map that 00:16:46.120 --> 00:16:47.560 into something else. 00:16:47.560 --> 00:16:49.030 So you need to then be able to say, 00:16:49.030 --> 00:16:52.210 suppose I do something nice for you or [INAUDIBLE],, 00:16:52.210 --> 00:16:55.540 and I can then figure out which parts of the brain 00:16:55.540 --> 00:16:57.400 are affected. 00:16:57.400 --> 00:17:00.100 And then I can use that then to later, 00:17:00.100 --> 00:17:02.350 now if I wanted to figure out, suppose I could measure 00:17:02.350 --> 00:17:04.390 your brains during lecture. 00:17:04.390 --> 00:17:06.760 If I had figured out earlier which parts of the brain 00:17:06.760 --> 00:17:09.069 are associated with happy versus unhappy things, 00:17:09.069 --> 00:17:11.015 and now I could see what your brains are doing 00:17:11.015 --> 00:17:12.640 during lecture, I could then figure out 00:17:12.640 --> 00:17:14.589 whether you are happy or unhappy. 00:17:14.589 --> 00:17:19.720 But notice you need to have some mapping, because otherwise, 00:17:19.720 --> 00:17:23.170 it's hard to interpret what's going on. 00:17:23.170 --> 00:17:25.810 But of course, people have already worked a lot 00:17:25.810 --> 00:17:27.549 on that, that kind of mapping. 00:17:27.549 --> 00:17:30.160 And just to be clear, so the revealed 00:17:30.160 --> 00:17:34.384 preference measures are-- 00:17:34.384 --> 00:17:36.540 hold on, let me just share what I have here. 00:17:36.540 --> 00:17:38.040 But the revealed preference measures 00:17:38.040 --> 00:17:41.070 are useful for some things, but in a way, 00:17:41.070 --> 00:17:44.850 if you wanted to measure whether revealed preferences give you 00:17:44.850 --> 00:17:48.930 a good answer, then of course, the real preference measure 00:17:48.930 --> 00:17:49.800 is not helpful. 00:17:49.800 --> 00:17:53.820 Or put differently, you want to check precisely. 00:17:53.820 --> 00:17:58.020 Suppose if I ask you, do you want lecture A or lecture B, 00:17:58.020 --> 00:18:01.230 and then everybody says they want lecture A-- or whatever, 00:18:01.230 --> 00:18:05.340 good A or good B, and everybody says they want good A-- 00:18:05.340 --> 00:18:08.010 the econ assumption is, by revealed preference, 00:18:08.010 --> 00:18:11.160 it must be that A makes you happier than B. 00:18:11.160 --> 00:18:14.730 But if I wanted to the test, is that actually true, then I 00:18:14.730 --> 00:18:16.980 need to have your revealed preference, 00:18:16.980 --> 00:18:19.080 but then I also need to know your counterfactual 00:18:19.080 --> 00:18:19.663 in some sense. 00:18:19.663 --> 00:18:21.930 So I'd say give you good A, and I'd give you good B. 00:18:21.930 --> 00:18:24.090 How would that make you? 00:18:24.090 --> 00:18:25.590 So I need the measures that we have 00:18:25.590 --> 00:18:27.890 here in the slides, a version of that, 00:18:27.890 --> 00:18:30.600 to say, OK, I'm going to look at your facial expression 00:18:30.600 --> 00:18:32.490 and see how happy you look when I give you 00:18:32.490 --> 00:18:35.258 good A versus good B. Or I look at your brain 00:18:35.258 --> 00:18:37.050 and see which parts of it-- do the pleasure 00:18:37.050 --> 00:18:39.785 parts of your brain light up? 00:18:39.785 --> 00:18:41.160 And if that's the case, I'm going 00:18:41.160 --> 00:18:44.040 to conclude that you're happier with good A versus good B, 00:18:44.040 --> 00:18:46.470 or if that's more so the case for good A. 00:18:46.470 --> 00:18:48.570 And then I can look at your revealed preference. 00:18:48.570 --> 00:18:51.690 Is it really the case that you choose A over B, 00:18:51.690 --> 00:18:56.910 if the experienced utility says A makes you happier than B? 00:18:56.910 --> 00:18:59.790 Does that make sense? 00:18:59.790 --> 00:19:00.300 OK. 00:19:00.300 --> 00:19:02.750 So what are these techniques? 00:19:02.750 --> 00:19:05.760 [AUDIO OUT] I think we covered most of [AUDIO OUT].. 00:19:05.760 --> 00:19:10.660 There's the observer ratings, facial [AUDIO OUT],, 00:19:10.660 --> 00:19:13.570 reports of mood, pain, pleasure, happiness, and so on. 00:19:13.570 --> 00:19:15.420 This is what [AUDIO OUT]. 00:19:15.420 --> 00:19:19.510 There's autonomic measures, which 00:19:19.510 --> 00:19:22.600 is the respiratory, cardiovascular, your blood 00:19:22.600 --> 00:19:26.740 pressure, or your pulse, and so on, and so forth-- sweat. 00:19:26.740 --> 00:19:29.497 There's vocal measures as well, which is pitch, loudness, tone. 00:19:29.497 --> 00:19:31.330 And you can probably detect whether somebody 00:19:31.330 --> 00:19:32.710 is angry or excited and so on. 00:19:32.710 --> 00:19:34.990 And again, you can apply some mapping here. 00:19:34.990 --> 00:19:37.540 You can look at people's brains. 00:19:37.540 --> 00:19:40.900 There's also some responses to emotion sensitive tasks. 00:19:40.900 --> 00:19:44.950 So if I ask you, do you want to talk to a friend 00:19:44.950 --> 00:19:47.050 or do you want us to do something nice, 00:19:47.050 --> 00:19:49.560 and if the answer is, no, absolutely not, 00:19:49.560 --> 00:19:51.340 it could obviously be you're busy. 00:19:51.340 --> 00:19:53.680 But when people are in a bad mood, 00:19:53.680 --> 00:19:55.840 they tend to not want to do things that 00:19:55.840 --> 00:20:01.010 likely should make them happy. 00:20:01.010 --> 00:20:03.080 OK, so we can try to do that. 00:20:03.080 --> 00:20:07.250 Now why might decision utility and experienced utility differ? 00:20:07.250 --> 00:20:09.500 That's of course what we discussed 00:20:09.500 --> 00:20:13.040 at length in the class already. 00:20:13.040 --> 00:20:17.442 There's various different explanations. 00:20:17.442 --> 00:20:19.650 One is-- we're going to talk about this a little bit, 00:20:19.650 --> 00:20:22.530 which is inaccurate memories of past experiences. 00:20:22.530 --> 00:20:26.030 So you might just misremember your past experiences, 00:20:26.030 --> 00:20:28.280 and based on that, you might have 00:20:28.280 --> 00:20:32.900 poor forecasts of your future preferences or your utility. 00:20:32.900 --> 00:20:36.350 So this is essentially saying people's beliefs might 00:20:36.350 --> 00:20:38.450 be wrong in various ways, perhaps 00:20:38.450 --> 00:20:43.220 because people's memories are tricking them in some ways. 00:20:43.220 --> 00:20:44.930 People might also have wrong beliefs 00:20:44.930 --> 00:20:49.940 because they have some failures to anticipate adaptation, which 00:20:49.940 --> 00:20:51.573 is very much like projection bias 00:20:51.573 --> 00:20:53.240 that we talked about already previously. 00:20:53.240 --> 00:20:55.578 I'm going to talk about this a little bit more. 00:20:55.578 --> 00:20:57.620 There could be things like emotions, which again, 00:20:57.620 --> 00:21:01.830 is in some ways often like projection bias, which 00:21:01.830 --> 00:21:07.560 is about things like people might be really angry, 00:21:07.560 --> 00:21:10.400 or people might be hungry. 00:21:10.400 --> 00:21:14.235 [AUDIO OUT] essentially it's very hard for them to make 00:21:14.235 --> 00:21:18.470 certain choices, failing to anticipate their anger 00:21:18.470 --> 00:21:24.283 or their hunger or other impulses [INAUDIBLE] over time. 00:21:24.283 --> 00:21:25.950 There's plenty funding of other reasons, 00:21:25.950 --> 00:21:28.313 but we talked about this quite a lot. 00:21:28.313 --> 00:21:29.730 And so you could think about a lot 00:21:29.730 --> 00:21:31.320 of what we discussed in the course 00:21:31.320 --> 00:21:34.050 about disconnects between decision utility 00:21:34.050 --> 00:21:37.310 and experienced utility. 00:21:37.310 --> 00:21:42.730 Now I mentioned this already a little bit. 00:21:42.730 --> 00:21:46.150 The third part in people's utility 00:21:46.150 --> 00:21:49.180 is what Kahneman would call remembered utility, which 00:21:49.180 --> 00:21:51.160 essentially is to say, well, when you think 00:21:51.160 --> 00:21:54.730 about a past experience, you might not actually 00:21:54.730 --> 00:21:56.530 evaluate it correctly. 00:21:56.530 --> 00:21:57.760 In a sense, you might-- 00:21:57.760 --> 00:22:03.980 as I said, I might look at your experienced utility using 00:22:03.980 --> 00:22:05.900 these kinds of features, and then 00:22:05.900 --> 00:22:07.670 I might ask you afterwards, how did you 00:22:07.670 --> 00:22:09.140 like certain experiences? 00:22:09.140 --> 00:22:12.830 And then I can test, in a way, how are you aggregating, 00:22:12.830 --> 00:22:15.800 or how appropriately are you aggregating 00:22:15.800 --> 00:22:18.440 your experienced utility when I just ask you afterwards, 00:22:18.440 --> 00:22:21.830 how did you like certain experiences? 00:22:21.830 --> 00:22:25.460 And one quite interesting feature in people's remembered 00:22:25.460 --> 00:22:27.764 utility is what's called-- excuse me-- 00:22:27.764 --> 00:22:29.660 duration neglect. 00:22:29.660 --> 00:22:32.570 People tend to remember the quality, rather than 00:22:32.570 --> 00:22:34.760 the length of the experience. 00:22:34.760 --> 00:22:38.060 And it seems to be also what's called the peak-end rule, which 00:22:38.060 --> 00:22:42.560 essentially says that retrospective evaluations are 00:22:42.560 --> 00:22:44.360 very much predicted by two things, 00:22:44.360 --> 00:22:48.740 or an average of the peak affective response recorded 00:22:48.740 --> 00:22:53.090 during an episode and the end value recorded just 00:22:53.090 --> 00:22:55.020 before the termination of the episode. 00:22:55.020 --> 00:22:56.270 And how do I think about this? 00:22:56.270 --> 00:23:02.000 Well, mostly, I think this as being very salient. 00:23:02.000 --> 00:23:03.590 So when you think about and I ask you 00:23:03.590 --> 00:23:07.790 about a semester or a certain lecture or a certain experience 00:23:07.790 --> 00:23:10.010 that you had, maybe a TV show or the like, 00:23:10.010 --> 00:23:12.770 when you think about it and try to remember how much you 00:23:12.770 --> 00:23:16.850 enjoyed that experience, well, what's most salient, of course, 00:23:16.850 --> 00:23:18.380 is your latest experience. 00:23:18.380 --> 00:23:19.730 That's what comes to mind. 00:23:19.730 --> 00:23:22.700 The last thing you remember is the last thing you experienced. 00:23:22.700 --> 00:23:27.290 And if the last episode of a TV show was terrible, 00:23:27.290 --> 00:23:28.480 that's what comes to mind. 00:23:28.480 --> 00:23:30.188 And you might say, well, the whole series 00:23:30.188 --> 00:23:32.270 was really bad, even though maybe previously, 00:23:32.270 --> 00:23:33.690 it was quite good. 00:23:33.690 --> 00:23:35.330 And then the other thing, I guess, 00:23:35.330 --> 00:23:37.100 that seems to be really important 00:23:37.100 --> 00:23:40.235 is the peak, both positive or negative. 00:23:40.235 --> 00:23:42.360 So if something was really, really great or really, 00:23:42.360 --> 00:23:44.060 really bad, people really, really 00:23:44.060 --> 00:23:47.552 remember that, as opposed to stuff in between. 00:23:47.552 --> 00:23:49.010 And again, I think that's something 00:23:49.010 --> 00:23:51.740 that's just very salient in people's minds, 00:23:51.740 --> 00:23:56.450 and then in thinking about these experiences, that's 00:23:56.450 --> 00:23:57.560 what comes to mind. 00:23:57.560 --> 00:23:59.180 And in studies that-- 00:23:59.180 --> 00:24:00.560 so what are these studies doing? 00:24:00.560 --> 00:24:03.080 Well, there are these studies that tend to measure 00:24:03.080 --> 00:24:05.690 experienced utility in certain ways, 00:24:05.690 --> 00:24:11.570 either through just measuring people's-- 00:24:11.570 --> 00:24:14.300 tracking people's well-being over time 00:24:14.300 --> 00:24:18.102 and then trying to predict their remembered utility, how 00:24:18.102 --> 00:24:19.310 did you like that experience. 00:24:19.310 --> 00:24:20.420 That's one version. 00:24:20.420 --> 00:24:24.670 Another version is-- let me show you. 00:24:24.670 --> 00:24:25.990 Hang on one second-- 00:24:25.990 --> 00:24:30.010 is where people do certain tasks specifically, 00:24:30.010 --> 00:24:32.590 and they manipulate the experience deliberately. 00:24:32.590 --> 00:24:34.660 For example, there are these tasks 00:24:34.660 --> 00:24:36.940 which is called a cold pressor task, which 00:24:36.940 --> 00:24:39.760 is they ask people to put their hands into cold water 00:24:39.760 --> 00:24:43.450 and then vary the temperature, or they vary the duration. 00:24:43.450 --> 00:24:45.190 And then essentially, you can then 00:24:45.190 --> 00:24:49.360 try to predict people's remembered utility 00:24:49.360 --> 00:24:53.260 and then show things such as duration neglect or the like. 00:24:53.260 --> 00:24:55.870 And let me show you what that looks like. 00:24:55.870 --> 00:24:57.940 For example, this is a famous study 00:24:57.940 --> 00:25:00.690 by Schreiber and Kahneman. 00:25:00.690 --> 00:25:03.180 They did a short trial, which is you 00:25:03.180 --> 00:25:08.220 put your hand into 14-degree water 00:25:08.220 --> 00:25:10.380 for 60 seconds, which is very cold, 00:25:10.380 --> 00:25:14.940 or a long trial where you put it into water for 60 second, 00:25:14.940 --> 00:25:17.610 and then the temperature rises to 15 degrees. 00:25:17.610 --> 00:25:19.110 I think this-- 00:25:19.110 --> 00:25:20.970 I hope this is Celsius, not Fahrenheit. 00:25:20.970 --> 00:25:23.520 But anyway, it's cold. 00:25:23.520 --> 00:25:25.590 But the point here is that putting 00:25:25.590 --> 00:25:30.360 your hand for 60 seconds into 14-degree water, and then 00:25:30.360 --> 00:25:32.430 another 30 seconds into 15-degree water, 00:25:32.430 --> 00:25:34.230 these are both unpleasant experiences. 00:25:34.230 --> 00:25:37.020 So if you did each of these things in separation, 00:25:37.020 --> 00:25:43.240 people would say, [AUDIO OUT] just 60 seconds of cold water. 00:25:43.240 --> 00:25:47.860 I also don't like 30 seconds of 15-degree water. 00:25:47.860 --> 00:25:50.950 These are both unpleasant experiences. 00:25:50.950 --> 00:25:52.950 So neither of them are actually desirable. 00:25:52.950 --> 00:25:56.633 If you ask people, if you give people [AUDIO OUT] 00:25:56.633 --> 00:25:59.050 do you want to experience this again, people say, hell no. 00:25:59.050 --> 00:26:02.950 I really don't like this, for both of these, in separation. 00:26:02.950 --> 00:26:05.740 But once you put them together, the 30 seconds 00:26:05.740 --> 00:26:08.710 of 15-degree water is more pleasant 00:26:08.710 --> 00:26:10.630 than the 14-degree water. 00:26:10.630 --> 00:26:14.140 So when you do that, people say they prefer the long trial 00:26:14.140 --> 00:26:16.735 over the short trial. 00:26:16.735 --> 00:26:18.110 And the reason being is what they 00:26:18.110 --> 00:26:21.140 remember is the last experience, which 00:26:21.140 --> 00:26:24.340 was not that painful, when things got better over time. 00:26:24.340 --> 00:26:26.960 So people's remembered utility is 00:26:26.960 --> 00:26:31.550 violating any sort of rules of what you would usually 00:26:31.550 --> 00:26:34.460 think in terms of rationality and economics, 00:26:34.460 --> 00:26:38.120 which is precisely because they remember the end rather 00:26:38.120 --> 00:26:39.800 than the overall duration. 00:26:39.800 --> 00:26:41.330 So essentially, the utility does not 00:26:41.330 --> 00:26:44.120 aggregate in the way we would think they would, 00:26:44.120 --> 00:26:47.000 because the experience at each point in time [AUDIO OUT] 00:26:47.000 --> 00:26:49.970 negative. 00:26:49.970 --> 00:26:53.930 There's more unpleasant studies for colonoscopy. 00:26:53.930 --> 00:26:55.970 Don't ask me about how they got ethical approval 00:26:55.970 --> 00:26:57.000 for those things. 00:26:57.000 --> 00:27:01.850 But anyway, a study [INAUDIBLE] where 00:27:01.850 --> 00:27:07.820 getting an extra minute where nothing bad happens. 00:27:07.820 --> 00:27:09.560 And essentially, the treatment group 00:27:09.560 --> 00:27:13.790 has better experience and memories of those kinds of-- 00:27:13.790 --> 00:27:15.350 of the overall experience. 00:27:15.350 --> 00:27:17.810 Now that sounds kind of like torturing people, 00:27:17.810 --> 00:27:21.380 but in fact, if you think about it, well, if-- 00:27:21.380 --> 00:27:25.220 in particular in medical procedures that are unpleasant 00:27:25.220 --> 00:27:30.680 overall, if we can help people have better memories of those 00:27:30.680 --> 00:27:33.290 experiences, maybe they will be a lot likely 00:27:33.290 --> 00:27:36.080 to come back and get future treatments or the like. 00:27:36.080 --> 00:27:38.317 So in a way, you might say, what are these studies? 00:27:38.317 --> 00:27:39.650 Why are people torturing people? 00:27:39.650 --> 00:27:43.040 Why are researchers torturing people? 00:27:43.040 --> 00:27:45.390 I agree with that notion, but in fact, 00:27:45.390 --> 00:27:47.660 if you think about if you wanted to design, 00:27:47.660 --> 00:27:52.010 and if you think people are under-using certain medical 00:27:52.010 --> 00:27:54.680 procedures or certain medical tests that they should be 00:27:54.680 --> 00:27:55.520 doing-- 00:27:55.520 --> 00:27:58.610 think of dentists-- if there's some [AUDIO OUT] 00:27:58.610 --> 00:28:01.520 people's overall memory or memory 00:28:01.520 --> 00:28:04.040 of their overall experience better by having 00:28:04.040 --> 00:28:09.662 the end of that experience slightly better 00:28:09.662 --> 00:28:12.860 than the previous parts, that could [AUDIO OUT] 00:28:12.860 --> 00:28:17.510 helpful in generatings of behavior that's good for people 00:28:17.510 --> 00:28:21.060 in terms of their health, even though, of course, 00:28:21.060 --> 00:28:23.373 the last minute of the actual experience 00:28:23.373 --> 00:28:24.290 would not be pleasant. 00:28:24.290 --> 00:28:26.060 Nobody wants to say, would you rather 00:28:26.060 --> 00:28:28.220 have another minute of colonoscopy? 00:28:28.220 --> 00:28:29.420 Absolutely not. 00:28:29.420 --> 00:28:32.390 But maybe the overall experience that you remember 00:28:32.390 --> 00:28:35.390 is, in fact, better than if not. 00:28:35.390 --> 00:28:37.030 So there's other types of experience 00:28:37.030 --> 00:28:40.115 that shows this duration neglected 00:28:40.115 --> 00:28:41.680 and peak-end evaluations. 00:28:41.680 --> 00:28:48.520 People have shown pleasant and unpleasant movies or contents 00:28:48.520 --> 00:28:50.120 of movies. 00:28:50.120 --> 00:28:54.910 They have done these experiments with sounds or noise 00:28:54.910 --> 00:28:56.830 of various loudness and duration. 00:28:56.830 --> 00:29:02.110 There's also animal [INAUDIBLE],, where essentially animals are 00:29:02.110 --> 00:29:05.440 given unpleasant experiences, and then you look at what-- 00:29:05.440 --> 00:29:07.100 there's different types of experiences 00:29:07.100 --> 00:29:10.653 where the animal can go into one direction or the other, press 00:29:10.653 --> 00:29:12.070 one lever versus another, and then 00:29:12.070 --> 00:29:14.153 bad stuff happens in both of them, and one of them 00:29:14.153 --> 00:29:15.410 is worse than the other. 00:29:15.410 --> 00:29:16.993 And then you can look at, essentially, 00:29:16.993 --> 00:29:21.940 do animals have similar duration or other peak-end evaluations 00:29:21.940 --> 00:29:22.790 of experiences? 00:29:22.790 --> 00:29:27.610 And it seems to be very much a phenomenon that's 00:29:27.610 --> 00:29:30.160 also true for animals. 00:29:30.160 --> 00:29:31.630 Any questions on this? 00:29:45.546 --> 00:29:47.190 OK. 00:29:47.190 --> 00:29:51.250 So then, of course, if that's the case, 00:29:51.250 --> 00:29:56.920 if people remember things differently, 00:29:56.920 --> 00:30:00.880 that will lead to violations of the decision utility 00:30:00.880 --> 00:30:03.550 maximizing the experienced utility. 00:30:03.550 --> 00:30:05.710 If you think about the experienced utility 00:30:05.710 --> 00:30:11.590 or the overall experienced utility of a certain choice 00:30:11.590 --> 00:30:14.590 that you make, in this case, I guess, the cold pressor task, 00:30:14.590 --> 00:30:16.660 think about this like the integral over, 00:30:16.660 --> 00:30:19.580 say, in this case, I guess, 90 seconds. 00:30:19.580 --> 00:30:21.160 In the first case, it's 60 seconds 00:30:21.160 --> 00:30:24.410 plus 30 seconds, 60 seconds feeling bad, 00:30:24.410 --> 00:30:27.070 versus the 30 seconds of putting your hand out, 00:30:27.070 --> 00:30:30.020 which is somewhat of a neutral experience. 00:30:30.020 --> 00:30:33.530 If you're integrating that, presumably, the short trial 00:30:33.530 --> 00:30:36.080 would give you a higher valuation 00:30:36.080 --> 00:30:41.227 of higher, better, more positive experience overall. 00:30:41.227 --> 00:30:43.310 But then if you ask people, which one do you want, 00:30:43.310 --> 00:30:47.450 because people remember the long trial as being less bad, 00:30:47.450 --> 00:30:52.520 people might choose the longer trial over the short trial 00:30:52.520 --> 00:30:53.372 overall. 00:30:53.372 --> 00:30:54.830 So then that leads to, essentially, 00:30:54.830 --> 00:30:57.178 violations of the decision utility maximizing 00:30:57.178 --> 00:30:59.720 the experienced utility, because essentially, people remember 00:30:59.720 --> 00:31:04.090 things in a biased way. 00:31:04.090 --> 00:31:04.870 OK. 00:31:04.870 --> 00:31:08.260 So then we talked about this very briefly. 00:31:08.260 --> 00:31:11.620 So so far, we talked about different ways of measuring. 00:31:11.620 --> 00:31:14.050 One specific thing that people have done a lot 00:31:14.050 --> 00:31:18.130 is just asking people simply about their happiness and life 00:31:18.130 --> 00:31:19.940 satisfaction. 00:31:19.940 --> 00:31:21.430 So simply, you ask people. 00:31:21.430 --> 00:31:23.310 There's different ways. 00:31:23.310 --> 00:31:25.570 There are these types of ladder questions, 00:31:25.570 --> 00:31:29.260 how satisfied are you with your life as a whole these days? 00:31:29.260 --> 00:31:31.300 There's also affect questions about like, 00:31:31.300 --> 00:31:33.388 did you experience certain emotions yesterday? 00:31:33.388 --> 00:31:34.180 Did you feel angry? 00:31:34.180 --> 00:31:34.990 Did you feel happy? 00:31:34.990 --> 00:31:35.890 Did you feel sad? 00:31:35.890 --> 00:31:36.940 And so on. 00:31:36.940 --> 00:31:41.710 The ladder question is a mixture of-- some people think 00:31:41.710 --> 00:31:43.240 of this as life satisfaction. 00:31:43.240 --> 00:31:45.550 Some people interpret it as happiness. 00:31:45.550 --> 00:31:47.710 Others would argue that it's rather 00:31:47.710 --> 00:31:49.860 about social comparisons. 00:31:49.860 --> 00:31:51.652 So the ladder question has this notion of-- 00:31:51.652 --> 00:31:53.693 I'm going to show you the definition in a second, 00:31:53.693 --> 00:31:54.580 or in a bit. 00:31:54.580 --> 00:31:58.168 The ladder definition is very much just, 00:31:58.168 --> 00:32:02.252 on a ladder from 1 to 10, how great could your life be? 00:32:02.252 --> 00:32:03.460 Where are you in this ladder? 00:32:03.460 --> 00:32:07.340 So it has very much this flavor of some social comparison, 00:32:07.340 --> 00:32:11.320 which is not quite ideal, because it very well makes 00:32:11.320 --> 00:32:13.817 it like, well, on the ladder, how far and low are you 00:32:13.817 --> 00:32:14.400 at the ladder? 00:32:14.400 --> 00:32:16.275 It makes it very salient that some people are 00:32:16.275 --> 00:32:18.110 higher or lower than others. 00:32:18.110 --> 00:32:19.960 So some of the responses that you 00:32:19.960 --> 00:32:23.590 see in those kinds of ladder questions are about happiness, 00:32:23.590 --> 00:32:26.320 but they potentially are also about people comparing 00:32:26.320 --> 00:32:33.250 each other to others in society and so on, which might not 00:32:33.250 --> 00:32:36.070 be necessarily the same as, how happy are you right now 00:32:36.070 --> 00:32:37.670 [INAUDIBLE]. 00:32:37.670 --> 00:32:40.240 Now what's problematic about these happiness questions, 00:32:40.240 --> 00:32:45.680 or why are economists skeptical, or what 00:32:45.680 --> 00:32:47.010 instances do we [AUDIO OUT]? 00:32:59.560 --> 00:33:02.230 So just to be clear, so some people would argue, 00:33:02.230 --> 00:33:05.590 we should have a World Happiness Report and GDP. 00:33:05.590 --> 00:33:09.730 We should not maximize GDP or look at GDP growth 00:33:09.730 --> 00:33:12.070 as an overall performance of society. 00:33:12.070 --> 00:33:14.350 Instead, we should measure subjective well-being, 00:33:14.350 --> 00:33:16.750 and that's what we should essentially maximize. 00:33:16.750 --> 00:33:18.760 And policy should be-- the objective 00:33:18.760 --> 00:33:21.620 should be to maximize that. 00:33:21.620 --> 00:33:24.450 So what's the pushback on that? 00:33:24.450 --> 00:33:27.240 So part of that is people and stuff 00:33:27.240 --> 00:33:32.820 that we talked already about, about remembered utility, 00:33:32.820 --> 00:33:36.815 or people just might not know what's good for them. 00:33:36.815 --> 00:33:38.190 But what's wrong with just asking 00:33:38.190 --> 00:33:39.360 people, how happy are you? 00:33:39.360 --> 00:33:41.140 And are people just going to tell you, 00:33:41.140 --> 00:33:42.515 or what's happening with that? 00:33:42.515 --> 00:33:45.040 Yeah, so one problem, broadly speaking, 00:33:45.040 --> 00:33:49.817 is when you look at different countries, different societies, 00:33:49.817 --> 00:33:51.900 particularly, for example, if you look at Germans, 00:33:51.900 --> 00:33:55.638 Germans tend to be pretty grumpy and not super happy. 00:33:55.638 --> 00:33:57.180 So when you ask a German, how are you 00:33:57.180 --> 00:34:00.840 doing, the typical German would say, OK, by which they mean, 00:34:00.840 --> 00:34:02.010 actually pretty good. 00:34:02.010 --> 00:34:05.610 If you ask a typical American, how are you doing, 00:34:05.610 --> 00:34:06.570 they say, amazing. 00:34:06.570 --> 00:34:09.000 They probably mean, pretty good. 00:34:09.000 --> 00:34:12.389 So when people give you certain statements, 00:34:12.389 --> 00:34:14.800 there's lots of cultural and other aspect 00:34:14.800 --> 00:34:17.760 that make comparison, particularly across societies, 00:34:17.760 --> 00:34:19.080 kind of tricky. 00:34:19.080 --> 00:34:21.239 You might think people in Latin America tend 00:34:21.239 --> 00:34:23.270 to report higher levels of happiness 00:34:23.270 --> 00:34:25.770 because that's a cultural thing, which might be that they're 00:34:25.770 --> 00:34:27.812 actually happier, or it might be that that's just 00:34:27.812 --> 00:34:30.780 the thing that you're supposed to exhibit. 00:34:30.780 --> 00:34:35.110 While people in some other states, Scandinavia, Germany, 00:34:35.110 --> 00:34:37.150 et cetera, might be, in fact, quite happy, 00:34:37.150 --> 00:34:41.429 but what they report to you is just some moderate level 00:34:41.429 --> 00:34:43.080 of enthusiasm overall. 00:34:43.080 --> 00:34:48.927 So one is cultural comparison is quite hard 00:34:48.927 --> 00:34:50.219 using those types of questions. 00:34:50.219 --> 00:34:52.500 There's lots of Gallup surveys that people 00:34:52.500 --> 00:34:55.330 have done at large scales. 00:34:55.330 --> 00:34:58.470 It's the Gallup world happiness, or whatever, surveys. 00:34:58.470 --> 00:35:01.927 There's huge amounts of data that are being collected. 00:35:01.927 --> 00:35:03.760 They're a nationally representative surveys, 00:35:03.760 --> 00:35:04.780 and so on and so forth. 00:35:04.780 --> 00:35:06.270 So those data are available. 00:35:06.270 --> 00:35:10.475 Of course, there's questions of, in some places, 00:35:10.475 --> 00:35:11.850 it's much harder to collect these 00:35:11.850 --> 00:35:13.820 than in others, and so on and so forth. 00:35:13.820 --> 00:35:15.812 And some people don't reply to surveys, 00:35:15.812 --> 00:35:16.770 and so on and so forth. 00:35:16.770 --> 00:35:19.020 For example, the elderly might be much less 00:35:19.020 --> 00:35:22.260 likely to answer some surveys, or some people 00:35:22.260 --> 00:35:24.460 without phones or internet, and so on and so forth, 00:35:24.460 --> 00:35:25.530 you might not be able to reach them. 00:35:25.530 --> 00:35:27.697 But in principle, you can solve all of those issues. 00:35:27.697 --> 00:35:29.772 But I think that it's just cost of use. 00:35:29.772 --> 00:35:31.980 The reference dependence is a very interesting issue, 00:35:31.980 --> 00:35:34.680 because in some ways, you might say, well, now your income, 00:35:34.680 --> 00:35:36.600 but then your reference point goes off. 00:35:36.600 --> 00:35:39.360 Something happens to your reported happiness? 00:35:39.360 --> 00:35:42.220 But presumably you're happier than before? 00:35:42.220 --> 00:35:44.190 So that [AUDIO OUT]. 00:35:44.190 --> 00:35:47.915 On the other hand, utility might, in fact, 00:35:47.915 --> 00:35:49.290 be very much reference-dependent. 00:35:49.290 --> 00:35:53.450 For example, you might compare yourself to others. 00:35:53.450 --> 00:35:55.270 And then inequality, for example, 00:35:55.270 --> 00:35:57.960 might be extremely important for people's happiness. 00:35:57.960 --> 00:36:01.090 And one of the things that people argue is that, 00:36:01.090 --> 00:36:04.440 for example, [INAUDIBLE] the GDP, also for the US-- 00:36:04.440 --> 00:36:06.270 GDP in many places has gone up quite a bit, 00:36:06.270 --> 00:36:09.490 but happiness has not increased. 00:36:09.490 --> 00:36:12.270 And people argue that's because, in part, inequality also 00:36:12.270 --> 00:36:13.200 went up a lot. 00:36:13.200 --> 00:36:16.490 And if then inequality really makes people deeply unhappy, 00:36:16.490 --> 00:36:19.410 well, that is important, and we should 00:36:19.410 --> 00:36:22.830 incorporate that in some ways in our considerations overall. 00:36:22.830 --> 00:36:26.860 But you see, that causes tricky issues, because in one case, 00:36:26.860 --> 00:36:29.220 doubling everybody's income, and presumably, people 00:36:29.220 --> 00:36:32.730 should be happier, because their living standard goes up. 00:36:32.730 --> 00:36:36.227 So presumably, there, even if happiness doesn't go up, 00:36:36.227 --> 00:36:38.310 presumably, because living standards have gone up, 00:36:38.310 --> 00:36:40.560 and something is going wrong if I don't see increased 00:36:40.560 --> 00:36:42.600 happiness in terms of measurement-- 00:36:42.600 --> 00:36:46.170 on the other hand, there might be very much real issues about, 00:36:46.170 --> 00:36:49.830 inequality is quite important, and reference-dependent matters 00:36:49.830 --> 00:36:51.930 for people's happiness and well-being, 00:36:51.930 --> 00:36:54.480 in the sense that what really might matter a lot 00:36:54.480 --> 00:36:57.750 is relative income rather than absolute income, 00:36:57.750 --> 00:37:02.530 and we should capture that in some meaningful way. 00:37:02.530 --> 00:37:03.030 Oops. 00:37:03.030 --> 00:37:04.440 Sorry. 00:37:04.440 --> 00:37:06.150 In particular, there's lots of issues 00:37:06.150 --> 00:37:08.760 with, depending on what I focus you 00:37:08.760 --> 00:37:10.530 on when I ask certain questions, I 00:37:10.530 --> 00:37:12.150 can get very, very different answers. 00:37:12.150 --> 00:37:14.930 And one, for example, example of this 00:37:14.930 --> 00:37:16.800 is a paper by Strack et. al. that 00:37:16.800 --> 00:37:19.350 looked at the correlation between general happiness 00:37:19.350 --> 00:37:21.270 and happiness with dating. 00:37:21.270 --> 00:37:23.375 And so what they did is a very simple thing. 00:37:23.375 --> 00:37:24.750 They essentially just manipulated 00:37:24.750 --> 00:37:25.868 the order of questions. 00:37:25.868 --> 00:37:27.660 So first, they asked people about how happy 00:37:27.660 --> 00:37:29.028 are you in general. 00:37:29.028 --> 00:37:30.570 And then they asked people, how happy 00:37:30.570 --> 00:37:32.063 are you with your dating life? 00:37:32.063 --> 00:37:33.480 And in a different subsample, they 00:37:33.480 --> 00:37:35.590 asked people, how happy are you with the dating life, 00:37:35.590 --> 00:37:37.440 and then asked people about their general happiness. 00:37:37.440 --> 00:37:38.898 And what that does, essentially, it 00:37:38.898 --> 00:37:42.270 focuses people very much about dating as part of their life 00:37:42.270 --> 00:37:44.310 being important and for their happiness. 00:37:44.310 --> 00:37:47.448 And similarly, I think this is exactly what Maya 00:37:47.448 --> 00:37:48.240 was saying as well. 00:37:48.240 --> 00:37:50.650 We need to ask, as you know, about the ladder question, 00:37:50.650 --> 00:37:53.130 which is about other questions about, depending 00:37:53.130 --> 00:37:56.942 on how I phrase the question, I can get very different answers 00:37:56.942 --> 00:37:57.900 and what you're saying. 00:37:57.900 --> 00:37:59.695 So that makes it very tricky to figure out 00:37:59.695 --> 00:38:01.320 which questions should we use, how much 00:38:01.320 --> 00:38:02.820 weight should be put on it. 00:38:02.820 --> 00:38:07.500 And even things like the order of questions matter. 00:38:07.500 --> 00:38:09.845 Matters quite a bit, which makes economists very nervous 00:38:09.845 --> 00:38:11.970 and so on, and then that's why, in part, economists 00:38:11.970 --> 00:38:15.070 would argue that we really should have revealed preference 00:38:15.070 --> 00:38:16.210 and choice. 00:38:16.210 --> 00:38:20.470 Now on the other hand, I very much think the questions, 00:38:20.470 --> 00:38:23.240 when you just ask people and say, I'm deeply unhappy, 00:38:23.240 --> 00:38:26.420 that has some content, and there's information there. 00:38:26.420 --> 00:38:28.738 And so then the question is, well, 00:38:28.738 --> 00:38:30.280 how do we weigh these considerations? 00:38:30.280 --> 00:38:32.110 On the one hand, the survey instrument 00:38:32.110 --> 00:38:34.910 is really just very fragile, and we don't measure very much. 00:38:34.910 --> 00:38:38.110 On the other hand, there's useful information here 00:38:38.110 --> 00:38:40.510 that we should be incorporating somehow. 00:38:40.510 --> 00:38:43.440 So I just wanted to have you-- 00:38:43.440 --> 00:38:44.610 that caveat in mind. 00:38:44.610 --> 00:38:46.410 And so let's now look at some data 00:38:46.410 --> 00:38:47.850 that people have collected. 00:38:47.850 --> 00:38:50.640 So what's a very nice source overall 00:38:50.640 --> 00:38:54.750 is what's called Our World in Data, which 00:38:54.750 --> 00:38:56.865 has lots of interesting graphs and figures 00:38:56.865 --> 00:38:58.740 where you can look at happiness and all sorts 00:38:58.740 --> 00:39:01.300 of things data on the world. 00:39:01.300 --> 00:39:04.180 And when you look at life satisfaction around the globe, 00:39:04.180 --> 00:39:07.020 this is very much the ladder question here, which you see, 00:39:07.020 --> 00:39:09.210 exactly as I was saying. 00:39:09.210 --> 00:39:12.540 There's the ladder from zero to 10. 00:39:12.540 --> 00:39:15.180 And again, the ladder question is very much something 00:39:15.180 --> 00:39:18.570 hierarchical and makes inequality or relative 00:39:18.570 --> 00:39:20.100 comparisons quite salient. 00:39:20.100 --> 00:39:22.680 Even if it only asks you about your worst or your best 00:39:22.680 --> 00:39:26.370 possible life for yourself, it does have the feeling 00:39:26.370 --> 00:39:28.890 that really, it's about comparing yourself to others. 00:39:28.890 --> 00:39:31.707 But anyway, that's very much used in many cases. 00:39:31.707 --> 00:39:33.540 Often people will call it life satisfaction, 00:39:33.540 --> 00:39:36.170 but even here, they also call it happiness overall. 00:39:36.170 --> 00:39:38.520 This is from the World Happiness Report. 00:39:38.520 --> 00:39:41.130 And what you see is that, in general, 00:39:41.130 --> 00:39:42.930 rich countries tend to be-- 00:39:42.930 --> 00:39:45.120 report higher happiness, Scandinavian countries 00:39:45.120 --> 00:39:45.780 in particular. 00:39:45.780 --> 00:39:47.460 The US is reasonably happy. 00:39:47.460 --> 00:39:49.050 Europe is quite happy. 00:39:49.050 --> 00:39:50.940 And then if you go to poorer countries, 00:39:50.940 --> 00:39:53.730 in particular in Africa, they tend 00:39:53.730 --> 00:39:57.570 to report a lot lower happiness level. 00:39:57.570 --> 00:39:59.700 When you look at comparisons across countries, 00:39:59.700 --> 00:40:03.270 you see a very clear relationship. 00:40:03.270 --> 00:40:06.540 Rich countries tend to be a lot-- 00:40:06.540 --> 00:40:10.620 report much higher life satisfaction 00:40:10.620 --> 00:40:11.910 compared to poor countries. 00:40:11.910 --> 00:40:14.190 That's a very clear association. 00:40:14.190 --> 00:40:16.290 That's true, both across countries 00:40:16.290 --> 00:40:18.340 but also true within countries. 00:40:18.340 --> 00:40:19.740 So what this graph here shows you 00:40:19.740 --> 00:40:23.200 is the average income in a country. 00:40:23.200 --> 00:40:26.140 And then these arrows show you the gradient. 00:40:26.140 --> 00:40:27.820 Within the slope of the arrows is 00:40:27.820 --> 00:40:29.230 the gradient within countries. 00:40:29.230 --> 00:40:31.015 And what you see is essentially-- 00:40:31.015 --> 00:40:32.890 so this is comparing people within countries, 00:40:32.890 --> 00:40:35.560 rich versus poor countries. 00:40:35.560 --> 00:40:40.150 If you increase income within a country by $1,000 or something, 00:40:40.150 --> 00:40:44.180 how much does people's happiness go up? 00:40:44.180 --> 00:40:46.990 And what you see is all of those countries have-- 00:40:46.990 --> 00:40:49.600 almost all countries have increasing gradients, 00:40:49.600 --> 00:40:52.330 which tells you that, both within and across countries, 00:40:52.330 --> 00:40:56.410 richer people report higher life satisfactions 00:40:56.410 --> 00:40:58.790 compared to poorer people. 00:40:58.790 --> 00:41:01.340 Similarly, that's also true for mental health. 00:41:01.340 --> 00:41:03.910 So these are all questions about subjective well-being, 00:41:03.910 --> 00:41:07.240 how happy are you overall, asking the overall population. 00:41:07.240 --> 00:41:10.930 But if you try to look at mental health conditions 00:41:10.930 --> 00:41:16.330 in terms of particularly depression and anxiety, so 00:41:16.330 --> 00:41:22.840 using clear definitions and psychiatric disorders 00:41:22.840 --> 00:41:26.080 for depression and anxiety and also for other conditions, 00:41:26.080 --> 00:41:28.480 you find that the poor in any given location 00:41:28.480 --> 00:41:32.530 are more likely to suffer from depression and/or anxiety. 00:41:32.530 --> 00:41:36.620 Interestingly, the prevalence of depression, 00:41:36.620 --> 00:41:40.150 in fact, also and anxiety, is higher in rich countries. 00:41:40.150 --> 00:41:42.400 There's a lot of questions about measurement problems, 00:41:42.400 --> 00:41:44.470 like is it really measuring the same thing, 00:41:44.470 --> 00:41:45.843 and so on and so forth. 00:41:45.843 --> 00:41:48.010 But it could also be there's other factors involved, 00:41:48.010 --> 00:41:50.303 in particular, issues such as inequality. 00:41:50.303 --> 00:41:51.970 And it could just be that really, what's 00:41:51.970 --> 00:41:54.580 really important is relative income as opposed 00:41:54.580 --> 00:41:56.200 to absolute income. 00:41:56.200 --> 00:42:01.480 Now we know also that anti-poverty programs improve 00:42:01.480 --> 00:42:03.910 mental health in various ways, including 00:42:03.910 --> 00:42:11.170 things like cash transfers and other anti-poverty programs. 00:42:11.170 --> 00:42:13.270 So this is an overview of studies. 00:42:13.270 --> 00:42:16.030 This is essentially a recent meta-analysis 00:42:16.030 --> 00:42:19.480 that we did for some [INAUDIBLE] paper of mine. 00:42:19.480 --> 00:42:21.430 And what you see here is a number 00:42:21.430 --> 00:42:24.010 of different interventions, both cash transfers 00:42:24.010 --> 00:42:26.290 and other anti-poverty programs, which 00:42:26.290 --> 00:42:28.330 are more broader programs. 00:42:28.330 --> 00:42:31.258 And what you see here is the treatment effect estimated 00:42:31.258 --> 00:42:32.300 in the different studies. 00:42:32.300 --> 00:42:36.110 These are overall poverty anti-poverty programs. 00:42:36.110 --> 00:42:37.960 These are cash transfer programs. 00:42:37.960 --> 00:42:41.170 But overall, essentially, the overwhelming answer 00:42:41.170 --> 00:42:44.260 is that once you give people cash or reduce their poverty, 00:42:44.260 --> 00:42:46.750 that improves their mental health, as measured 00:42:46.750 --> 00:42:50.860 by different-- these are PWB, means psychological well-being 00:42:50.860 --> 00:42:51.730 indices. 00:42:51.730 --> 00:42:54.650 This is also depression screening scores and so on. 00:42:54.650 --> 00:42:57.340 So by various measures of people's mental health, 00:42:57.340 --> 00:43:00.460 giving people more money improves their mental health. 00:43:00.460 --> 00:43:04.900 So that's true at the extreme in terms of really mental illness, 00:43:04.900 --> 00:43:07.790 depressive and anxiety disorders. 00:43:07.790 --> 00:43:09.670 It's also true for other measures 00:43:09.670 --> 00:43:11.110 of psychological well-being. 00:43:11.110 --> 00:43:14.620 If you just give people cash, they report higher happiness, 00:43:14.620 --> 00:43:17.410 in addition to what I told you before, 00:43:17.410 --> 00:43:20.710 this being true in the cross section. 00:43:20.710 --> 00:43:23.520 Now interestingly, we also tend to think 00:43:23.520 --> 00:43:25.840 that others are less happy than they say they are. 00:43:25.840 --> 00:43:28.140 So if we just ask people, how happy 00:43:28.140 --> 00:43:29.970 are other people in your country, 00:43:29.970 --> 00:43:31.565 almost every country in the world, 00:43:31.565 --> 00:43:32.940 what you find is that people tend 00:43:32.940 --> 00:43:35.460 to say the fraction that people actually say that they're 00:43:35.460 --> 00:43:39.270 very happy or rather happy is lower 00:43:39.270 --> 00:43:46.680 than they actually say they are, which is an interesting fact. 00:43:46.680 --> 00:43:49.830 Now this-- as we talked about this already a little bit-- 00:43:49.830 --> 00:43:51.990 how is life satisfaction and happiness related 00:43:51.990 --> 00:43:53.293 to life events? 00:43:53.293 --> 00:43:55.460 In terms of when you look at different life events-- 00:43:55.460 --> 00:43:57.750 it's quite interesting that a lot of life events 00:43:57.750 --> 00:44:03.720 tend to affect people's reported well-being or life 00:44:03.720 --> 00:44:05.850 satisfaction a lot. 00:44:05.850 --> 00:44:09.780 For many events, including stuff like widowhood or divorce 00:44:09.780 --> 00:44:14.070 or marriage and other types of events, or winning the lottery, 00:44:14.070 --> 00:44:17.550 these effects are very much transitory, which is to say, 00:44:17.550 --> 00:44:19.470 when something really bad to somebody happens, 00:44:19.470 --> 00:44:21.490 their happiness goes down a lot. 00:44:21.490 --> 00:44:23.580 But it tends to recover quite a bit. 00:44:23.580 --> 00:44:25.330 This is what we talked about before. 00:44:25.330 --> 00:44:28.350 People are quite adaptive to many changes, 00:44:28.350 --> 00:44:29.980 and they adjust quite well. 00:44:29.980 --> 00:44:32.970 This is what people call the psychological immune system, 00:44:32.970 --> 00:44:35.310 except for a few things, including unemployment. 00:44:35.310 --> 00:44:38.080 And one potential reason is that, for example, unemployment 00:44:38.080 --> 00:44:40.055 might also cause things like depression. 00:44:40.055 --> 00:44:41.430 And once you're really depressed, 00:44:41.430 --> 00:44:44.190 that doesn't really go back over [INAUDIBLE] that easily. 00:44:44.190 --> 00:44:45.180 That's true for women. 00:44:45.180 --> 00:44:49.200 It's also true for men overall. 00:44:49.200 --> 00:44:50.700 There's also some other issues which 00:44:50.700 --> 00:44:52.825 you want to be careful of whether to interpret some 00:44:52.825 --> 00:44:57.450 of this data, which is there's called this evolution of latent 00:44:57.450 --> 00:45:02.040 situations, which is when you look at things like divorce, 00:45:02.040 --> 00:45:03.480 it tends to be that people-- 00:45:03.480 --> 00:45:05.303 when you look at people's divorce, 00:45:05.303 --> 00:45:06.720 you might say, well, divorce makes 00:45:06.720 --> 00:45:08.340 them happier in some ways. 00:45:08.340 --> 00:45:10.800 But really, what seems to be-- 00:45:10.800 --> 00:45:12.420 or you would look at this and say, 00:45:12.420 --> 00:45:13.878 marriage makes people less happy. 00:45:13.878 --> 00:45:15.420 But really, what seems to be the case 00:45:15.420 --> 00:45:18.150 is that people build up to a positive event, 00:45:18.150 --> 00:45:22.020 and then that peaks positive or negatively and then essentially 00:45:22.020 --> 00:45:23.490 goes back to what's before. 00:45:23.490 --> 00:45:26.370 For example, unemployment, things tend to go worse, 00:45:26.370 --> 00:45:32.070 even before people become unemployed, which seems to say, 00:45:32.070 --> 00:45:34.110 people tend to-- these underlying events, 00:45:34.110 --> 00:45:36.270 people tend to anticipate them already 00:45:36.270 --> 00:45:39.330 psychologically in some ways. 00:45:39.330 --> 00:45:41.490 Now, this is what I was saying before. 00:45:41.490 --> 00:45:44.930 When you look at comparisons over time, 00:45:44.930 --> 00:45:47.790 you get sometimes these situations where, particularly 00:45:47.790 --> 00:45:49.730 in places including in China, but also 00:45:49.730 --> 00:45:53.780 in the US, where real income has risen a lot in many places, 00:45:53.780 --> 00:45:56.240 but yet people's life satisfaction 00:45:56.240 --> 00:45:58.250 does not tend to increase. 00:45:58.250 --> 00:45:59.000 Why is that? 00:45:59.000 --> 00:46:02.170 Well, that's, in part, I think, because of inequality going up, 00:46:02.170 --> 00:46:03.770 but [INAUDIBLE] inequality really 00:46:03.770 --> 00:46:06.588 staying the same and relative income really being important. 00:46:06.588 --> 00:46:07.880 Could be other things going on. 00:46:07.880 --> 00:46:10.310 For example, pollution is really bad for people's health, 00:46:10.310 --> 00:46:12.830 mental health, but also their happiness and so on. 00:46:12.830 --> 00:46:15.315 It could also just be that people's reference points 00:46:15.315 --> 00:46:17.690 adjust, which is just, you make everybody twice as happy, 00:46:17.690 --> 00:46:19.850 and now everybody's-- the reference point goes up, 00:46:19.850 --> 00:46:22.788 even though their overall objective well-being standards, 00:46:22.788 --> 00:46:24.830 in terms of how much can they consume, et cetera, 00:46:24.830 --> 00:46:28.160 has gone up a lot. 00:46:28.160 --> 00:46:34.950 Now this is quite interesting data now on ceiling effects. 00:46:34.950 --> 00:46:37.860 So there's some controversy over this. 00:46:37.860 --> 00:46:39.620 Some people are arguing that when 00:46:39.620 --> 00:46:41.370 you give people more money, at some point, 00:46:41.370 --> 00:46:42.700 there's a ceiling effect. 00:46:42.700 --> 00:46:44.940 It doesn't really make them happier anymore. 00:46:44.940 --> 00:46:47.180 Now when you look at this graph here-- let 00:46:47.180 --> 00:46:48.180 me go back for a second. 00:46:48.180 --> 00:46:54.278 When you look at these graphs here, this-- 00:46:54.278 --> 00:46:56.420 whoops, sorry. 00:46:56.420 --> 00:46:56.990 [INAUDIBLE] 00:46:56.990 --> 00:46:59.795 One second. 00:46:59.795 --> 00:47:01.170 When you look at this graph here, 00:47:01.170 --> 00:47:03.170 it just doesn't look at all like ceiling effect. 00:47:03.170 --> 00:47:06.600 This looks like, essentially, self-reported life satisfaction 00:47:06.600 --> 00:47:09.930 as measured by these ladder question, 00:47:09.930 --> 00:47:14.610 essentially, is increasing, and there seems to be no ceiling, 00:47:14.610 --> 00:47:16.643 or it doesn't seem to be flattening in any way. 00:47:16.643 --> 00:47:18.060 When you look at this one as well, 00:47:18.060 --> 00:47:23.890 it seems to be if you increase income by $1,000, at any level, 00:47:23.890 --> 00:47:26.980 it seems to be that reported life satisfaction goes up. 00:47:26.980 --> 00:47:29.760 So again, that doesn't look like a ceiling effect. 00:47:29.760 --> 00:47:36.780 And that's true as well if you look at these questions here. 00:47:36.780 --> 00:47:38.918 Gallup had done these surveys in the US. 00:47:38.918 --> 00:47:40.710 When you look at the ladder question, which 00:47:40.710 --> 00:47:43.350 is this one here, when you look at annual income in the US, 00:47:43.350 --> 00:47:45.270 once you increase people's income, 00:47:45.270 --> 00:47:48.750 their mean ladder score, which is here 00:47:48.750 --> 00:47:51.660 on the axis on the right, tends to go up 00:47:51.660 --> 00:47:52.752 pretty much monotonically. 00:47:52.752 --> 00:47:54.210 Maybe it's a little bit flattening, 00:47:54.210 --> 00:47:55.377 but essentially, it goes up. 00:47:55.377 --> 00:47:58.020 So people report higher life satisfaction 00:47:58.020 --> 00:47:59.650 when they have higher income. 00:47:59.650 --> 00:48:02.460 But when you look at people's affect, both things 00:48:02.460 --> 00:48:04.980 like positive affect, which is like happiness, smiling, 00:48:04.980 --> 00:48:08.250 enjoyment, not being blue, being stress-free and so on, 00:48:08.250 --> 00:48:11.220 what you see is-- in particular, this is for the US. 00:48:11.220 --> 00:48:13.860 When you go from something like $10,000 to $20,000 00:48:13.860 --> 00:48:18.000 to $40,000 per year of income, there's 00:48:18.000 --> 00:48:19.180 a clear increase over time. 00:48:19.180 --> 00:48:25.290 So being poor is really-- or poor in the US, 00:48:25.290 --> 00:48:28.720 it is really, really bad for your affect 00:48:28.720 --> 00:48:33.330 and for your reported well-being overall. 00:48:33.330 --> 00:48:37.560 But once you reach something like $50,000, $60,000, $70,000 00:48:37.560 --> 00:48:40.260 in the US, it seems to me that there's 00:48:40.260 --> 00:48:43.410 no effect on positive affect, feeling blue, stress-free 00:48:43.410 --> 00:48:44.320 and so on. 00:48:44.320 --> 00:48:48.900 So one interpretation of that is that if we take these affect 00:48:48.900 --> 00:48:52.200 questions as more like a measure of experienced utility 00:48:52.200 --> 00:48:55.110 and happiness in any given life, essentially, 00:48:55.110 --> 00:48:59.250 having something like $70,000, $80,000 per year in the US-- 00:48:59.250 --> 00:49:01.470 granted, this is from 2009, so maybe this number 00:49:01.470 --> 00:49:03.330 has gone a little bit higher. 00:49:03.330 --> 00:49:06.600 But that number is sufficient to make people sufficiently happy, 00:49:06.600 --> 00:49:08.540 and beyond that, doubling of income 00:49:08.540 --> 00:49:10.830 will not do very much on how much you smile every day 00:49:10.830 --> 00:49:13.350 and how happy you are as a measure of those things. 00:49:13.350 --> 00:49:15.910 In contrast, if I ask you the ladder question, which again, 00:49:15.910 --> 00:49:18.960 is very much about comparison, about how you feel yourself 00:49:18.960 --> 00:49:22.260 compared to somebody else in society, 00:49:22.260 --> 00:49:27.940 going from $100,000 to $200,000 will make, potentially, still 00:49:27.940 --> 00:49:30.830 quite a big difference. 00:49:30.830 --> 00:49:32.740 Any questions on that? 00:49:36.508 --> 00:49:37.450 OK. 00:49:37.450 --> 00:49:39.100 So then let me-- 00:49:39.100 --> 00:49:43.960 this is among the most interesting parts of these 00:49:43.960 --> 00:49:46.730 kinds of studies, is the question about what, in fact-- 00:49:46.730 --> 00:49:47.707 so if not just income-- 00:49:47.707 --> 00:49:49.540 so if you know income essentially-- overall, 00:49:49.540 --> 00:49:51.370 if you're rich, you're going to be happier 00:49:51.370 --> 00:49:52.840 compared to if you're poor. 00:49:52.840 --> 00:49:56.170 But again, it seems to be really, 00:49:56.170 --> 00:49:59.270 if you want to not be unhappy, having 00:49:59.270 --> 00:50:02.050 income above $50,000 in the US is really important. 00:50:02.050 --> 00:50:05.780 Beyond that, it doesn't seem to do very much anymore. 00:50:05.780 --> 00:50:08.120 But now what other things matter quite a bit? 00:50:08.120 --> 00:50:09.680 And this is a very simple exercise. 00:50:09.680 --> 00:50:13.900 So what they did is, this is [INAUDIBLE].. 00:50:13.900 --> 00:50:15.970 They looked at the Gallup survey questions 00:50:15.970 --> 00:50:18.730 from 450,000 Americans. 00:50:18.730 --> 00:50:20.800 And what they did is, for each of these four 00:50:20.800 --> 00:50:23.740 measures of psychological well-being, 00:50:23.740 --> 00:50:26.710 which is these here, positive affect, not blue, stress-free, 00:50:26.710 --> 00:50:30.460 and the ladder questions, they were first 00:50:30.460 --> 00:50:33.520 just trying to predict, what is the regression 00:50:33.520 --> 00:50:34.787 coefficient of high income? 00:50:34.787 --> 00:50:36.370 So if you just split the sample in two 00:50:36.370 --> 00:50:39.130 and look at high income versus low income, 00:50:39.130 --> 00:50:41.560 what is the regression coefficients of that? 00:50:41.560 --> 00:50:44.450 They're trying to interpret that as causal effect, which, 00:50:44.450 --> 00:50:47.830 of course, is a little problematic. 00:50:47.830 --> 00:50:49.580 But you can think of this as the question, 00:50:49.580 --> 00:50:53.140 if I wanted to try to predict how happy you are, what do I 00:50:53.140 --> 00:50:54.697 need to know about you? 00:50:54.697 --> 00:50:57.280 What are the factors that I need to know about you if I wanted 00:50:57.280 --> 00:51:00.550 to predict how happy you are in your life on any point in time? 00:51:00.550 --> 00:51:03.010 If you wanted to know, 20 years from now, 00:51:03.010 --> 00:51:04.460 are you going to be happy or not, 00:51:04.460 --> 00:51:07.700 what are the things that you need to know about yourself? 00:51:07.700 --> 00:51:11.110 And so it turns out that high income is predictive. 00:51:11.110 --> 00:51:13.262 So there's some coefficient that says 00:51:13.262 --> 00:51:15.220 that if you have higher income, you're somewhat 00:51:15.220 --> 00:51:17.380 more likely to have positive affect, 00:51:17.380 --> 00:51:20.410 less likely to feel blue or less likely to be stressed, 00:51:20.410 --> 00:51:24.820 and a lot more likely to give a higher answer in the ladder 00:51:24.820 --> 00:51:25.370 questions. 00:51:25.370 --> 00:51:27.460 So in some sense, higher income is good. 00:51:27.460 --> 00:51:29.500 And then what the table does is it 00:51:29.500 --> 00:51:32.350 looks at-- it normalizes high income to 1 00:51:32.350 --> 00:51:34.970 and then shows you other characteristics and says, 00:51:34.970 --> 00:51:40.230 OK, if I knew another about you, are you religious versus not 00:51:40.230 --> 00:51:43.930 and so on, what is the relative coefficient of that compared 00:51:43.930 --> 00:51:45.580 to a high versus low income? 00:51:45.580 --> 00:51:46.210 OK? 00:51:46.210 --> 00:51:49.670 So for example, if you're old, you're-- 00:51:49.670 --> 00:51:52.150 so a positive coefficient means you're 00:51:52.150 --> 00:51:55.060 happier or more likely to have positive affect here, 00:51:55.060 --> 00:51:58.370 and a negative, the opposite, I guess, for positive affect 00:51:58.370 --> 00:51:59.120 here. 00:51:59.120 --> 00:52:01.040 So if you're old, for example, older people, 00:52:01.040 --> 00:52:06.553 at least in the US, tend to report higher positive affect. 00:52:06.553 --> 00:52:07.720 So that's good news for you. 00:52:07.720 --> 00:52:11.520 In 20, 30 years, you'll be likely to be happier. 00:52:11.520 --> 00:52:13.690 It turns out that religious people, for example, 00:52:13.690 --> 00:52:16.660 tend to be more-- 00:52:16.660 --> 00:52:18.310 report more positive affect compared 00:52:18.310 --> 00:52:20.300 to non-religious people. 00:52:20.300 --> 00:52:23.170 And the magnitude of these effects is pretty large. 00:52:23.170 --> 00:52:24.100 That is to say-- 00:52:24.100 --> 00:52:27.880 1.16 is to say, if you compare religious-- 00:52:27.880 --> 00:52:30.880 if I knew about you, are you religious versus not, 00:52:30.880 --> 00:52:33.280 that is as predictive as knowing about you, 00:52:33.280 --> 00:52:35.980 are you rich versus not, are you in the upper or lower half 00:52:35.980 --> 00:52:37.233 of this income distribution? 00:52:37.233 --> 00:52:38.650 So that's quite important in terms 00:52:38.650 --> 00:52:40.840 of trying to predict whether people are happy. 00:52:40.840 --> 00:52:45.010 Now what's particularly striking here when you look at this 00:52:45.010 --> 00:52:50.402 here is alone, smoker, and headache, and particularly 00:52:50.402 --> 00:52:52.030 pain. 00:52:52.030 --> 00:52:54.250 I want to emphasize the alone here. 00:52:54.250 --> 00:52:58.810 Social relationships or people being in relationships 00:52:58.810 --> 00:53:02.950 with friends or in family or with partners 00:53:02.950 --> 00:53:05.980 is extremely predictive of people's well-being. 00:53:05.980 --> 00:53:07.840 This is what I was talking about before when 00:53:07.840 --> 00:53:09.970 it came to social preferences. 00:53:09.970 --> 00:53:13.180 Investing in social relationships in one way is-- 00:53:13.180 --> 00:53:14.240 A, of course. 00:53:14.240 --> 00:53:16.780 Making other people happy and being nice is good. 00:53:16.780 --> 00:53:19.570 But in a way, you think about investing 00:53:19.570 --> 00:53:22.588 or being nice to others and investing 00:53:22.588 --> 00:53:23.630 in a social relationship. 00:53:23.630 --> 00:53:29.221 It's very much an investment in your future well-being. 00:53:29.221 --> 00:53:33.920 And the magnitudes here are enormous. 00:53:33.920 --> 00:53:37.580 Now another way you could do is-- 00:53:37.580 --> 00:53:40.620 so here, you could say, well, what 00:53:40.620 --> 00:53:43.200 should I do to make me happier in the future? 00:53:43.200 --> 00:53:45.790 So one thing you could do is just look at data. 00:53:45.790 --> 00:53:47.540 So one thing you could do is look at data. 00:53:47.540 --> 00:53:50.070 And well, what are these predictors [AUDIO OUT] 00:53:50.070 --> 00:53:50.810 people happy? 00:53:50.810 --> 00:53:52.290 These are all correlations. 00:53:52.290 --> 00:53:54.411 You want to be careful. 00:53:54.411 --> 00:53:56.730 [AUDIO OUT] tells you something about, if you 00:53:56.730 --> 00:54:00.030 want to be happy in 20 years, things you want to invest in, 00:54:00.030 --> 00:54:05.050 these numbers are probably a pretty good start. 00:54:05.050 --> 00:54:08.910 [AUDIO OUT] your health seems to be quite a good idea. 00:54:08.910 --> 00:54:12.780 Smoking, for example, not so much. 00:54:12.780 --> 00:54:16.960 Now another thing you could do is just ask people 00:54:16.960 --> 00:54:20.310 at the above and just [AUDIO OUT] 00:54:20.310 --> 00:54:23.920 about what they should have or could have done in their life. 00:54:23.920 --> 00:54:25.170 What do you think people said? 00:54:25.170 --> 00:54:26.460 So this is an Australian nurse who 00:54:26.460 --> 00:54:28.350 recorded the experience from palliative care 00:54:28.350 --> 00:54:32.130 as end of life experiences and asked people about-- 00:54:32.130 --> 00:54:34.440 or she did not simply ask, but she was just recording 00:54:34.440 --> 00:54:35.430 what people are saying. 00:54:35.430 --> 00:54:37.240 What are the top five regrets? 00:54:37.240 --> 00:54:38.490 What would you think they are? 00:54:43.327 --> 00:54:45.285 AUDIENCE: Not spending enough time with family? 00:54:58.390 --> 00:54:59.390 FRANK SCHILBACH: Mm-hmm. 00:54:59.390 --> 00:55:01.265 Wished I had worked less, somebody else says. 00:55:14.480 --> 00:55:15.950 Good enough. 00:55:15.950 --> 00:55:20.570 So essentially, what people tend to say is a combination of they 00:55:20.570 --> 00:55:22.950 wished they had been more social, 00:55:22.950 --> 00:55:26.030 they wished they had expressed their feelings more, 00:55:26.030 --> 00:55:30.230 they had had been more true to themselves in some ways and not 00:55:30.230 --> 00:55:31.910 done things that others wanted them 00:55:31.910 --> 00:55:37.020 to do but rather spend more time with their friends and family 00:55:37.020 --> 00:55:37.740 and so on. 00:55:37.740 --> 00:55:39.230 Now you want to be a little bit careful with this. 00:55:39.230 --> 00:55:41.147 Again, there's issues with remembered utility. 00:55:41.147 --> 00:55:45.870 It's also issues with if you are alone at the end of your life, 00:55:45.870 --> 00:55:48.090 you might be quite unhappy because of that. 00:55:48.090 --> 00:55:50.625 Be careful how to interpret that. 00:55:50.625 --> 00:55:53.000 But I think we can actually learn quite a bit about this. 00:55:53.000 --> 00:55:55.040 In some ways, if you think about certain choices 00:55:55.040 --> 00:55:58.700 you make in life, understanding some of these issues 00:55:58.700 --> 00:56:01.170 seem to be quite important overall. 00:56:01.170 --> 00:56:03.320 And when you think about what should you maximize, 00:56:03.320 --> 00:56:08.361 it seems to be that maximizing income, if that comes 00:56:08.361 --> 00:56:10.850 at the cost of not having friends, 00:56:10.850 --> 00:56:13.220 seems to be not a good idea if you want 00:56:13.220 --> 00:56:15.447 to be happy in the long run. 00:56:15.447 --> 00:56:17.030 Another thing that people-- by the way 00:56:17.030 --> 00:56:20.250 here, this is not in here-- tend to emphasize 00:56:20.250 --> 00:56:23.150 quite a lot is having a meaningful job 00:56:23.150 --> 00:56:25.700 or having work that they believe in in terms of something 00:56:25.700 --> 00:56:28.730 meaningful as opposed to maximizing income, 00:56:28.730 --> 00:56:33.130 which seems quite important. 00:56:33.130 --> 00:56:33.880 OK. 00:56:33.880 --> 00:56:36.550 So then what kinds of things then could 00:56:36.550 --> 00:56:38.350 you do to make yourself happier? 00:56:38.350 --> 00:56:41.538 And I have a few things written down here. 00:56:41.538 --> 00:56:43.330 So one of the thing things that I really is 00:56:43.330 --> 00:56:46.390 invest in and maintain social relationships. 00:56:46.390 --> 00:56:48.820 Small acts can make big differences, 00:56:48.820 --> 00:56:51.670 as we talked about, of letters of gratitude 00:56:51.670 --> 00:56:52.930 or random acts of kindness. 00:56:52.930 --> 00:56:55.300 These are very small things that you could do. 00:56:55.300 --> 00:56:58.552 Of course, you can do also really important things. 00:56:58.552 --> 00:57:00.760 It matters quite a lot how much you help your friends 00:57:00.760 --> 00:57:03.670 or how much you invest in others. 00:57:03.670 --> 00:57:06.130 Again, that's partially reflective of just being 00:57:06.130 --> 00:57:08.380 pro-social and being nice to others, 00:57:08.380 --> 00:57:10.990 but you can think of this very much as-- 00:57:10.990 --> 00:57:12.990 neurotically in some ways, you can think of this 00:57:12.990 --> 00:57:17.150 also as investing in your own happiness in the future. 00:57:17.150 --> 00:57:19.750 Choosing meaningful work over money seems really important. 00:57:22.525 --> 00:57:24.400 If you think about what to do with your life, 00:57:24.400 --> 00:57:27.190 there's so much talent at MIT, and I sometimes 00:57:27.190 --> 00:57:29.710 wish that talent would go more into more meaningful things, 00:57:29.710 --> 00:57:32.770 or because it's society or making society better 00:57:32.770 --> 00:57:36.670 overall, as opposed to some perhaps potentially 00:57:36.670 --> 00:57:39.970 [INAUDIBLE] that pay you a lot of money. 00:57:39.970 --> 00:57:42.820 That's, in some sense, potentially socially optimal 00:57:42.820 --> 00:57:44.470 in terms of just use MIT's talents 00:57:44.470 --> 00:57:47.320 or MIT students' talents for useful things in the world. 00:57:47.320 --> 00:57:49.780 But if you're just selfish and want 00:57:49.780 --> 00:57:52.360 to make yourself happy in 20 years from now, 00:57:52.360 --> 00:57:55.260 it seems to me that just maximizing income is just not 00:57:55.260 --> 00:57:56.510 what's going to get you there. 00:57:56.510 --> 00:57:59.080 So it seems to be like being rich is certainly 00:57:59.080 --> 00:58:01.190 glamorous and exciting, but in fact, 00:58:01.190 --> 00:58:05.320 if that comes at the cost of your health, [INAUDIBLE],, 00:58:05.320 --> 00:58:07.810 and your friendships, most likely, 00:58:07.810 --> 00:58:09.940 or at the cost of having a meaningful job 00:58:09.940 --> 00:58:12.370 that you actually believe in and think is useful, 00:58:12.370 --> 00:58:14.600 that's not going to make you happy in the long run. 00:58:14.600 --> 00:58:16.990 So for those of you thinking about what kind of work 00:58:16.990 --> 00:58:20.930 to pursue in the future, keep that in mind. 00:58:20.930 --> 00:58:22.445 And then I have two more things here 00:58:22.445 --> 00:58:23.945 in my list, which is seeking support 00:58:23.945 --> 00:58:27.340 to improve your mental health and reduce social media usage. 00:58:27.340 --> 00:58:28.490 So what do I mean by that? 00:58:28.490 --> 00:58:30.580 So psychotherapies have very much 00:58:30.580 --> 00:58:34.220 been shown to be effective in reducing depression, anxiety, 00:58:34.220 --> 00:58:34.900 and so on. 00:58:34.900 --> 00:58:37.380 Yet people tend to not make use of those services. 00:58:37.380 --> 00:58:38.830 Now why is that? 00:58:38.830 --> 00:58:41.490 There's obvious reasons, which is stigma, shame, and on. 00:58:41.490 --> 00:58:44.030 There's also potentially misperceptions, projection 00:58:44.030 --> 00:58:44.530 bias. 00:58:44.530 --> 00:58:46.910 Also depression itself, for example, 00:58:46.910 --> 00:58:49.878 could precisely generate these kinds of beliefs. 00:58:49.878 --> 00:58:51.670 There could also be other behavioral biases 00:58:51.670 --> 00:58:53.878 like other health conditions, where people just don't 00:58:53.878 --> 00:58:55.490 like to see doctors and so on. 00:58:55.490 --> 00:58:57.170 Now another way to view this is to say, 00:58:57.170 --> 00:58:59.372 well, you might think that psychotherapy 00:58:59.372 --> 00:59:01.330 is really important for depression and anxiety, 00:59:01.330 --> 00:59:04.300 but psychotherapy could also be viewed as a coach 00:59:04.300 --> 00:59:05.050 to make you happy. 00:59:05.050 --> 00:59:07.323 Just think about this as a happiness coach. 00:59:07.323 --> 00:59:09.490 It might help you figure out your objective function 00:59:09.490 --> 00:59:12.370 in life, what really makes you happy, and how to pursue that. 00:59:12.370 --> 00:59:14.380 And that's regardless of depression, anxiety, 00:59:14.380 --> 00:59:15.380 or any mental disorders. 00:59:15.380 --> 00:59:17.470 It's to say, think about sports. 00:59:17.470 --> 00:59:19.720 In sports, you get a coach to be a better tennis 00:59:19.720 --> 00:59:22.603 player or whatever, soccer or whatever sports you do. 00:59:22.603 --> 00:59:24.020 That's a very natural thing to do. 00:59:24.020 --> 00:59:25.420 People help you to do better. 00:59:25.420 --> 00:59:27.922 Now why not have a coach for many other things in life, 00:59:27.922 --> 00:59:29.380 including how to be happier and how 00:59:29.380 --> 00:59:34.117 to lead a life that makes you satisfied? 00:59:34.117 --> 00:59:35.575 In some sense, if you're maximizing 00:59:35.575 --> 00:59:37.617 your wrong objective function, which again, could 00:59:37.617 --> 00:59:40.780 be you just maximize money over other things, 00:59:40.780 --> 00:59:43.440 if a therapist helped you do that or helped you 00:59:43.440 --> 00:59:48.130 correct that, that could be extremely valuable and helpful. 00:59:48.130 --> 00:59:52.630 And again, that's in the absence of any serious mental illness. 00:59:52.630 --> 00:59:54.220 That's just for the average person. 00:59:54.220 --> 00:59:57.610 Having somebody to talk to to help you optimize your life 00:59:57.610 --> 01:00:00.640 and mental well-being seems to be extremely 01:00:00.640 --> 01:00:04.800 important and worth trying. 01:00:04.800 --> 01:00:06.500 Second, about social media. 01:00:06.500 --> 01:00:09.310 So there's a recent paper that's quite intriguing. 01:00:09.310 --> 01:00:10.810 Or there's actually a couple papers, 01:00:10.810 --> 01:00:12.270 but let me just mention one, which 01:00:12.270 --> 01:00:17.220 is by Alcott et al that randomizes paying students 01:00:17.220 --> 01:00:19.360 to stay off Facebook for a month. 01:00:19.360 --> 01:00:22.830 This was before the 2018 election, midterm election, 01:00:22.830 --> 01:00:27.120 which is, in some sense, perhaps relevant in some ways. 01:00:27.120 --> 01:00:31.260 The timing might matter for some of these results. 01:00:31.260 --> 01:00:33.617 So what they find is that getting people to pay-- you 01:00:33.617 --> 01:00:35.450 have to pay students quite a lot to do this. 01:00:35.450 --> 01:00:37.580 So students, you have to pay them like $50, 01:00:37.580 --> 01:00:39.800 and otherwise, they're not willing to do this. 01:00:39.800 --> 01:00:44.520 When you do that, people reduce their online activities. 01:00:44.520 --> 01:00:47.460 They also reduce their factual new knowledge, 01:00:47.460 --> 01:00:49.710 political polarization and so on, all kinds of things 01:00:49.710 --> 01:00:51.330 that you might expect. 01:00:51.330 --> 01:00:53.730 Then you see increased subjective well-being, 01:00:53.730 --> 01:00:56.220 both happiness but also reduced depression. 01:00:56.220 --> 01:01:00.300 So students are happier, quite a bit happier 01:01:00.300 --> 01:01:03.300 by doing that Moreover, there seem 01:01:03.300 --> 01:01:06.060 to be large, persistent reductions 01:01:06.060 --> 01:01:08.320 in post-experiment Facebook usage. 01:01:08.320 --> 01:01:10.740 That is to say, once the experiment is over, 01:01:10.740 --> 01:01:13.800 lots of students who have been paid to stay off Facebook 01:01:13.800 --> 01:01:16.080 tend to do that, or to continue doing 01:01:16.080 --> 01:01:19.403 that, which very much seems to say that it's a habit good. 01:01:19.403 --> 01:01:20.820 Perhaps there's also some learning 01:01:20.820 --> 01:01:26.940 involved as students learn that they're now happier overall. 01:01:26.940 --> 01:01:29.040 Now that always leads to the question, well, 01:01:29.040 --> 01:01:31.950 why are people then on Facebook if it doesn't make them happy? 01:01:31.950 --> 01:01:34.170 And does it have to do with peer pressure? 01:01:34.170 --> 01:01:36.120 Does it have to do with habit formation? 01:01:36.120 --> 01:01:39.360 Does it have to do with self-control problems? 01:01:39.360 --> 01:01:41.950 Does it have to do with biased beliefs and so on? 01:01:41.950 --> 01:01:44.110 But overall in some ways, I think the reason 01:01:44.110 --> 01:01:46.920 that I show this evidence is, when you think about your life 01:01:46.920 --> 01:01:48.570 or things that you do in your life, 01:01:48.570 --> 01:01:51.030 you want to think about what are the things that you 01:01:51.030 --> 01:01:51.730 do in your life? 01:01:51.730 --> 01:01:54.660 What are the things that make you happy or not? 01:01:54.660 --> 01:01:57.270 And perhaps experimenting with those kinds of things, 01:01:57.270 --> 01:01:59.520 or just looking at data from studies 01:01:59.520 --> 01:02:01.638 seems like a very reasonable thing to do. 01:02:01.638 --> 01:02:03.180 So it could well be that social media 01:02:03.180 --> 01:02:05.070 makes you really happy in connecting with your friends. 01:02:05.070 --> 01:02:07.680 In particular right now, you think that's really helpful, 01:02:07.680 --> 01:02:09.270 because you can stay in touch if you 01:02:09.270 --> 01:02:11.230 need to talk online and so on and so forth, 01:02:11.230 --> 01:02:13.230 and that's truly a thing that's worth doing. 01:02:13.230 --> 01:02:17.850 So that could really be very beneficial for your happiness 01:02:17.850 --> 01:02:20.970 overall, but it might also be that there's 01:02:20.970 --> 01:02:24.780 large costs of not, coming from perhaps social comparison 01:02:24.780 --> 01:02:27.060 or just seeing everybody happy on Facebook, 01:02:27.060 --> 01:02:28.680 or whatever people post selectively 01:02:28.680 --> 01:02:33.780 makes it look like people are way happier than they actually 01:02:33.780 --> 01:02:34.572 are. 01:02:34.572 --> 01:02:36.780 And then that makes people feel bad about themselves, 01:02:36.780 --> 01:02:39.150 seeing others post overall. 01:02:39.150 --> 01:02:40.770 But I think the reason why I show 01:02:40.770 --> 01:02:42.870 you this is I want you to think about or be more 01:02:42.870 --> 01:02:45.810 conscious about, what are the things that do make you, 01:02:45.810 --> 01:02:49.980 in fact, happier and try to invest in those that seem 01:02:49.980 --> 01:02:51.147 to make you happier overall. 01:02:51.147 --> 01:02:52.938 So some of these things you could obviously 01:02:52.938 --> 01:02:53.670 experiment with. 01:02:53.670 --> 01:02:55.295 For example, you can obviously turn off 01:02:55.295 --> 01:02:57.210 your Facebook or Instagram, whatever, 01:02:57.210 --> 01:02:58.932 very quickly and easily. 01:02:58.932 --> 01:03:00.390 Other things, of course-- the stuff 01:03:00.390 --> 01:03:04.560 that I showed you in the very long run, 01:03:04.560 --> 01:03:13.332 like the Kahneman here, these kinds of data, of course, 01:03:13.332 --> 01:03:15.040 it's much harder to experiment with that. 01:03:15.040 --> 01:03:18.280 It's harder to say, I'm going to be religious for a month, 01:03:18.280 --> 01:03:21.520 from now on, and then see, are you happier? 01:03:21.520 --> 01:03:23.230 But it is worth to think about, what 01:03:23.230 --> 01:03:25.022 are the things that you're really pursuing, 01:03:25.022 --> 01:03:27.250 and are these things going to be making you happy? 01:03:27.250 --> 01:03:29.028 Not just right now and, is this just 01:03:29.028 --> 01:03:30.820 doing some things because your friends want 01:03:30.820 --> 01:03:32.980 it or some other influence that you have, 01:03:32.980 --> 01:03:35.525 as opposed to what are the things that you really want, 01:03:35.525 --> 01:03:37.900 and what are the things that are going to make you happy, 01:03:37.900 --> 01:03:41.390 not just this month or next month, but five 10, 01:03:41.390 --> 01:03:42.555 20 years from now? 01:03:42.555 --> 01:03:43.930 Because a lot of choices that you 01:03:43.930 --> 01:03:48.940 make in the next couple of years will 01:03:48.940 --> 01:03:51.580 be very persistent in terms of depending on what jobs 01:03:51.580 --> 01:03:54.880 you choose, what kinds of friendships you have, 01:03:54.880 --> 01:03:56.740 what kind of partners you have, and so on, 01:03:56.740 --> 01:03:58.810 will be very much persistent for a long time 01:03:58.810 --> 01:04:02.520 in terms of determining your long run happiness. 01:04:02.520 --> 01:04:07.620 And so that's a very useful thing to think about. 01:04:07.620 --> 01:04:09.210 Let me see. 01:04:09.210 --> 01:04:12.690 So what do people think? 01:04:12.690 --> 01:04:14.480 I guess, at some point, hopefully-- 01:04:14.480 --> 01:04:16.710 so Cubek is monitoring the llama situation, I hope. 01:04:18.943 --> 01:04:20.610 We'll see about that, because hopefully, 01:04:20.610 --> 01:04:21.780 that makes people happier. 01:04:21.780 --> 01:04:23.515 But what do people think is-- 01:04:23.515 --> 01:04:25.890 why are people on Facebook if it doesn't make them happy? 01:04:25.890 --> 01:04:27.330 So it could be essentially issues 01:04:27.330 --> 01:04:32.340 with present bias or the like, that any point in time, 01:04:32.340 --> 01:04:35.010 there's some temptations of perhaps posting or the like, 01:04:35.010 --> 01:04:38.730 and then that makes you happy in the very short run. 01:04:38.730 --> 01:04:41.010 But in the long run overall, then that's 01:04:41.010 --> 01:04:44.250 not really helping you overall, and there's essentially 01:04:44.250 --> 01:04:47.865 some self-control problems in some ways. 01:04:47.865 --> 01:04:49.740 In particular right now-- and so to be clear, 01:04:49.740 --> 01:04:51.907 the study that was done here-- and I actually talked 01:04:51.907 --> 01:04:54.660 to [INAUDIBLE] some of the authors of the study would also 01:04:54.660 --> 01:04:57.940 tell me, if you did the study right now, 01:04:57.940 --> 01:05:00.760 we would find the opposite, because now, in fact, 01:05:00.760 --> 01:05:02.280 Facebook or any other social media 01:05:02.280 --> 01:05:05.363 helps people stay in touch, and there's some positive benefits. 01:05:05.363 --> 01:05:07.530 So what you're saying is, well, presumably, Facebook 01:05:07.530 --> 01:05:11.750 was developed early on. 01:05:11.750 --> 01:05:13.125 Or the reason why it's successful 01:05:13.125 --> 01:05:14.875 is because it really helped people connect 01:05:14.875 --> 01:05:16.230 to each other in some ways. 01:05:16.230 --> 01:05:18.480 And people thought there were some benefits from that. 01:05:18.480 --> 01:05:20.640 And that's particularly true if you're not 01:05:20.640 --> 01:05:22.980 in the same place physically. 01:05:22.980 --> 01:05:25.570 And particularly right now, then there 01:05:25.570 --> 01:05:28.948 are these benefits overall. 01:05:28.948 --> 01:05:30.490 And then there's a question of, well, 01:05:30.490 --> 01:05:32.430 what are the costs and benefits of that? 01:05:32.430 --> 01:05:40.590 And maybe people misperceive some of these or the like. 01:05:40.590 --> 01:05:42.450 What I'm trying to encourage you is to say, 01:05:42.450 --> 01:05:46.100 well, there's ways to learn about this. 01:05:46.100 --> 01:05:48.260 And Facebook is actually a very good example 01:05:48.260 --> 01:05:50.510 where you could learn very easily, where you just say, 01:05:50.510 --> 01:05:53.000 let's turn it off for a month and see what happens. 01:05:53.000 --> 01:05:54.950 You could also-- there's some other activities where you say, 01:05:54.950 --> 01:05:57.656 let's do more sports, or let's get up earlier in the morning 01:05:57.656 --> 01:05:58.880 or sleep more at night. 01:05:58.880 --> 01:06:02.420 There's some things that are quite easy to experiment with. 01:06:02.420 --> 01:06:04.850 Other stuff, as I said, becoming religious versus not, 01:06:04.850 --> 01:06:07.940 is much trickier to experiment with, or being healthy 01:06:07.940 --> 01:06:08.540 versus not. 01:06:08.540 --> 01:06:13.730 In some sense, it's just a very much long run effect. 01:06:13.730 --> 01:06:16.550 But exactly as you said, people have never actually experienced 01:06:16.550 --> 01:06:18.560 any of this, so how would you know 01:06:18.560 --> 01:06:20.610 how you're going to be happier? 01:06:20.610 --> 01:06:24.450 There's also some public goods issues and some issues 01:06:24.450 --> 01:06:26.640 of externalities [INAUDIBLE]. 01:06:26.640 --> 01:06:29.280 If all of [AUDIO OUT] are on social media and you are not, 01:06:29.280 --> 01:06:34.410 you're a bit of an outsider, and that will make you likely 01:06:34.410 --> 01:06:35.240 very-- 01:06:35.240 --> 01:06:38.160 or likely not increase your happiness. 01:06:38.160 --> 01:06:42.780 But it might well be that if everybody decides to go off 01:06:42.780 --> 01:06:45.307 social media, then in some sense, 01:06:45.307 --> 01:06:46.890 you don't have to rely on social media 01:06:46.890 --> 01:06:49.140 to go to parties or whatever and so on. 01:06:49.140 --> 01:06:51.217 And so there might be also some coordination 01:06:51.217 --> 01:06:52.050 issues, potentially. 01:06:52.050 --> 01:06:54.720 So conditional on everybody being on Facebook around you, 01:06:54.720 --> 01:06:57.780 you might want to be on Facebook or whatever other social media. 01:06:57.780 --> 01:07:00.520 But if that's not the case, then-- 01:07:00.520 --> 01:07:02.260 so it could be for every single person. 01:07:02.260 --> 01:07:05.310 Every single person could be happier if everybody else were 01:07:05.310 --> 01:07:07.890 not on social media, but conditional on that everybody 01:07:07.890 --> 01:07:11.100 is on social media, that's not the case. 01:07:11.100 --> 01:07:14.555 And what I'm saying here is, we tend to do the same things over 01:07:14.555 --> 01:07:15.180 and over again. 01:07:15.180 --> 01:07:17.820 As Maya was saying, is what if we just 01:07:17.820 --> 01:07:22.380 experiment more off of this and we did costs, perhaps long run, 01:07:22.380 --> 01:07:22.980 benefits. 01:07:22.980 --> 01:07:25.560 There's default effects, of course, and so on that 01:07:25.560 --> 01:07:27.000 keep us from experimenting. 01:07:27.000 --> 01:07:29.250 But if you think about it, if you could make yourself 01:07:29.250 --> 01:07:31.960 just a little bit happier every day for many, many years, 01:07:31.960 --> 01:07:34.350 it seems to be worth experimenting a lot. 01:07:34.350 --> 01:07:37.530 I would very much encourage you to go and try out 01:07:37.530 --> 01:07:41.400 new things in life that potentially make you happier. 01:07:41.400 --> 01:07:43.235 [INAUDIBLE] has as a nice-- 01:07:43.235 --> 01:07:44.860 I hope this is still working, the link, 01:07:44.860 --> 01:07:46.402 but it has a very nice New York Times 01:07:46.402 --> 01:07:50.880 article that argues that [INAUDIBLE] experiment more 01:07:50.880 --> 01:07:54.640 and try to figure out what makes us happier. 01:07:54.640 --> 01:07:56.965 Any other thoughts on social media or Facebook? 01:08:20.899 --> 01:08:21.399 OK. 01:08:21.399 --> 01:08:28.130 So that's all I have for now. 01:08:28.130 --> 01:08:30.020 Next time, we're going to talk about policy. 01:08:30.020 --> 01:08:31.510 That's the last lecture. 01:08:31.510 --> 01:08:35.394 Please do read Thaler and Sunstein. 01:08:35.394 --> 01:08:38.200 Peluk tells me the llama is not here yet, 01:08:38.200 --> 01:08:45.654 neither in my private room where the llama might be otherwise. 01:08:45.654 --> 01:08:46.779 I'll sit here just to wait. 01:08:46.779 --> 01:08:49.340 I think the announced time is 2:30, 01:08:49.340 --> 01:08:50.573 so that's in four minutes. 01:08:50.573 --> 01:08:52.490 So if you have to run, of course, that's fine. 01:08:52.490 --> 01:08:54.615 Otherwise, I'm just going to stick around and wait. 01:08:54.615 --> 01:08:56.229 If you want to talk about more, I'd 01:08:56.229 --> 01:08:59.920 love to know more about why people are experimenting more. 01:08:59.920 --> 01:09:03.753 I also would love to know more about how people think. 01:09:03.753 --> 01:09:05.170 So one thing I was mentioning here 01:09:05.170 --> 01:09:07.295 is that, while social media might be bad 01:09:07.295 --> 01:09:08.920 and there might be detrimental effects, 01:09:08.920 --> 01:09:12.350 in particular coming from social comparisons, which is to say, 01:09:12.350 --> 01:09:13.270 look. 01:09:13.270 --> 01:09:14.830 You see everybody on Instagram looks 01:09:14.830 --> 01:09:17.870 like they have a glamorous life and are happy all the time. 01:09:17.870 --> 01:09:20.960 You see that, and that might make you feel depressed. 01:09:20.960 --> 01:09:22.654 By the way, that might also lead you 01:09:22.654 --> 01:09:24.920 to [AUDIO OUT] glamorous pictures yourself. 01:09:24.920 --> 01:09:28.040 So you might reinforce the whole issue overall. 01:09:28.040 --> 01:09:30.710 I think there surely are some downs, actually. 01:09:30.710 --> 01:09:32.240 But there's also potential upsides 01:09:32.240 --> 01:09:35.490 of connecting people online, in particular 01:09:35.490 --> 01:09:37.180 that I'm trying to work on. 01:09:37.180 --> 01:09:41.920 [AUDIO OUT] if anybody has some small things [INAUDIBLE] could 01:09:41.920 --> 01:09:48.040 do right now online to help people be more connected, 01:09:48.040 --> 01:09:51.660 and less lonely, and less depressed.