WEBVTT 00:00:00.000 --> 00:00:03.444 align:middle line:90% [AUDIO LOGO] 00:00:03.444 --> 00:00:18.722 align:middle line:90% 00:00:18.722 --> 00:00:20.680 align:middle line:84% FRANK SCHILBACH: All right, welcome, everybody. 00:00:20.680 --> 00:00:21.880 align:middle line:90% I'm Frank Schilbach. 00:00:21.880 --> 00:00:24.340 align:middle line:90% I'm a faculty member at MIT. 00:00:24.340 --> 00:00:28.390 align:middle line:84% And I teach behavioral and development economics. 00:00:28.390 --> 00:00:31.540 align:middle line:84% In particular, I'm interested in issues related to poverty, 00:00:31.540 --> 00:00:34.320 align:middle line:84% in particular, psychological issues, which I'll mostly 00:00:34.320 --> 00:00:37.270 align:middle line:84% tell you about next time on Wednesday 00:00:37.270 --> 00:00:39.627 align:middle line:84% when we talk about financial strains or people worrying 00:00:39.627 --> 00:00:41.710 align:middle line:84% about money, when we talk about sleep deprivation, 00:00:41.710 --> 00:00:45.628 align:middle line:84% and in particular, mental health. 00:00:45.628 --> 00:00:47.920 align:middle line:84% Today, we're going to talk about behavioral development 00:00:47.920 --> 00:00:50.740 align:middle line:90% economics as a field overall. 00:00:50.740 --> 00:00:54.573 align:middle line:84% And I'm trying to give you a broad overview. 00:00:54.573 --> 00:00:55.990 align:middle line:84% Most of what I'm going to tell you 00:00:55.990 --> 00:00:59.930 align:middle line:84% is based on a handbook chapter on behavioral development 00:00:59.930 --> 00:01:04.569 align:middle line:84% economics with Michael Kremer and Gautam Rao. 00:01:04.569 --> 00:01:07.240 align:middle line:90% So what's the plan? 00:01:07.240 --> 00:01:08.537 align:middle line:90% I'll give you a broad overview. 00:01:08.537 --> 00:01:10.120 align:middle line:84% And we'll kind of discuss a little bit 00:01:10.120 --> 00:01:13.000 align:middle line:84% what is behavioral development economics, a little bit 00:01:13.000 --> 00:01:15.910 align:middle line:84% of the history and how to think about this, and some caveats. 00:01:15.910 --> 00:01:20.020 align:middle line:84% Then we'll talk about one framework, one broad overview, 00:01:20.020 --> 00:01:22.990 align:middle line:84% and how to think about a lot of issues related 00:01:22.990 --> 00:01:25.930 align:middle line:84% to behavioral economics and development economics, which 00:01:25.930 --> 00:01:28.383 align:middle line:84% I'll call the Euler equation puzzle. 00:01:28.383 --> 00:01:29.800 align:middle line:84% And then we're going to apply this 00:01:29.800 --> 00:01:31.810 align:middle line:84% to one specific topic, which is health, 00:01:31.810 --> 00:01:33.850 align:middle line:84% which you will, later in the course, 00:01:33.850 --> 00:01:35.980 align:middle line:90% talk much more about anyway. 00:01:35.980 --> 00:01:37.960 align:middle line:84% And then, time permitting, I'll also 00:01:37.960 --> 00:01:40.480 align:middle line:84% tell you a little more about time preferences 00:01:40.480 --> 00:01:43.330 align:middle line:84% and demand for commitment in part based on some work 00:01:43.330 --> 00:01:44.890 align:middle line:90% that I've done on that topic. 00:01:44.890 --> 00:01:46.630 align:middle line:84% And then on Wednesday, we'll talk 00:01:46.630 --> 00:01:49.270 align:middle line:84% about the psychology of poverty, in particular, 00:01:49.270 --> 00:01:52.840 align:middle line:84% of the idea of the potential of psychological poverty traps. 00:01:52.840 --> 00:01:55.300 align:middle line:84% I know you have talked about poverty traps 00:01:55.300 --> 00:01:58.930 align:middle line:84% more broadly with Esther last week. 00:01:58.930 --> 00:02:02.860 align:middle line:84% And then we'll talk about three different topics 00:02:02.860 --> 00:02:03.850 align:middle line:90% that I have worked on. 00:02:03.850 --> 00:02:07.270 align:middle line:84% One is financial strain, which is the idea that people 00:02:07.270 --> 00:02:08.919 align:middle line:90% think a lot about money. 00:02:08.919 --> 00:02:11.530 align:middle line:84% Money is very much on people's minds 00:02:11.530 --> 00:02:14.140 align:middle line:84% and affects people's decisions, choices, 00:02:14.140 --> 00:02:15.760 align:middle line:90% and their productivity. 00:02:15.760 --> 00:02:17.703 align:middle line:84% We'll talk about sleep deprivation. 00:02:17.703 --> 00:02:19.120 align:middle line:84% And this is the paper that I think 00:02:19.120 --> 00:02:22.630 align:middle line:84% you're supposed to read until Wednesday, 00:02:22.630 --> 00:02:26.230 align:middle line:84% as well as mental health, which is what I'm currently 00:02:26.230 --> 00:02:27.760 align:middle line:90% focusing my research on. 00:02:27.760 --> 00:02:30.010 align:middle line:84% But there's lots of issues and open questions 00:02:30.010 --> 00:02:33.190 align:middle line:90% that one might want to tackle. 00:02:33.190 --> 00:02:36.580 align:middle line:84% Before I get started, any questions? 00:02:36.580 --> 00:02:38.280 align:middle line:90% OK, perfect. 00:02:38.280 --> 00:02:41.850 align:middle line:84% So when thinking about behavioral development 00:02:41.850 --> 00:02:44.610 align:middle line:84% economics, it's useful to take a somewhat broader 00:02:44.610 --> 00:02:48.030 align:middle line:84% historical perspective on development economics 00:02:48.030 --> 00:02:51.190 align:middle line:84% to try and understand where we came from. 00:02:51.190 --> 00:02:55.260 align:middle line:84% And so if you go back decades, when 00:02:55.260 --> 00:02:57.750 align:middle line:84% you think about historical views of development 00:02:57.750 --> 00:03:00.000 align:middle line:84% in different fields, in sociology, but also 00:03:00.000 --> 00:03:02.472 align:middle line:84% in economics, there was this view and development 00:03:02.472 --> 00:03:03.930 align:middle line:84% that people were thought to be very 00:03:03.930 --> 00:03:08.820 align:middle line:84% different before and after the event of modernity. 00:03:08.820 --> 00:03:11.160 align:middle line:84% The idea there's a pre-capitalist 00:03:11.160 --> 00:03:14.400 align:middle line:84% versus capitalist society or tradition versus rationalism 00:03:14.400 --> 00:03:17.430 align:middle line:84% or mechanical versus organic solidarity. 00:03:17.430 --> 00:03:20.220 align:middle line:84% And there was this idea, there was modernization theory 00:03:20.220 --> 00:03:23.220 align:middle line:84% that viewed modernization as a process 00:03:23.220 --> 00:03:25.740 align:middle line:84% of radical social change, but also 00:03:25.740 --> 00:03:28.060 align:middle line:84% change in ways of thinking and seeing the world. 00:03:28.060 --> 00:03:32.460 align:middle line:84% So the idea was we can look at poor countries 00:03:32.460 --> 00:03:34.380 align:middle line:90% in very different ways. 00:03:34.380 --> 00:03:36.990 align:middle line:84% And what we know about from economics in rich countries 00:03:36.990 --> 00:03:39.930 align:middle line:90% might not necessarily apply. 00:03:39.930 --> 00:03:41.610 align:middle line:84% So it's the idea that things are very 00:03:41.610 --> 00:03:43.770 align:middle line:84% different in poor countries among the poor. 00:03:43.770 --> 00:03:47.170 align:middle line:84% And so we first have to modernize the countries. 00:03:47.170 --> 00:03:49.440 align:middle line:84% And then afterwards, once things are modernized, 00:03:49.440 --> 00:03:52.440 align:middle line:84% we can apply traditional economics. 00:03:52.440 --> 00:03:55.800 align:middle line:84% And development economics, in part at least, 00:03:55.800 --> 00:03:58.500 align:middle line:84% emerged as a critical response to this view. 00:03:58.500 --> 00:04:02.940 align:middle line:84% It's saying like, no, poor people are people like you 00:04:02.940 --> 00:04:04.080 align:middle line:90% and I. They make choices. 00:04:04.080 --> 00:04:05.700 align:middle line:90% They make productive decisions. 00:04:05.700 --> 00:04:07.620 align:middle line:84% They're very much like rational actors, 00:04:07.620 --> 00:04:10.980 align:middle line:90% and economists would model them. 00:04:10.980 --> 00:04:13.320 align:middle line:84% And for example, farmers are viewed as, essentially, 00:04:13.320 --> 00:04:16.920 align:middle line:84% rational capitalists that may face some market failure 00:04:16.920 --> 00:04:18.959 align:middle line:90% or some other constraint. 00:04:18.959 --> 00:04:21.990 align:middle line:84% But people really are making choices like you and I, 00:04:21.990 --> 00:04:24.870 align:middle line:84% but they have certain different constraints. 00:04:24.870 --> 00:04:29.430 align:middle line:84% And so that view strongly rejects seemingly unfalsifiable 00:04:29.430 --> 00:04:32.340 align:middle line:84% cultural explanation, saying there's a Hindu rate of growth. 00:04:32.340 --> 00:04:36.030 align:middle line:84% And therefore, India is not growing as fast 00:04:36.030 --> 00:04:38.550 align:middle line:84% as perhaps some other countries in Asia. 00:04:38.550 --> 00:04:44.010 align:middle line:84% And it's very much modeling people's choices 00:04:44.010 --> 00:04:46.320 align:middle line:84% and decision-making as poor by deficient. 00:04:46.320 --> 00:04:50.400 align:middle line:84% And this is this famous expression from Shultz in 1964 00:04:50.400 --> 00:04:52.720 align:middle line:84% that's saying, is people are poor, 00:04:52.720 --> 00:04:54.870 align:middle line:90% but they're making choices. 00:04:54.870 --> 00:04:56.400 align:middle line:90% They're making rational choices. 00:04:56.400 --> 00:05:00.180 align:middle line:84% And we can apply the entire toolbox of economics, 00:05:00.180 --> 00:05:06.120 align:middle line:84% of macroeconomics, et cetera, to the applications 00:05:06.120 --> 00:05:08.400 align:middle line:90% of studying poverty. 00:05:08.400 --> 00:05:13.060 align:middle line:84% And so this view started to change in the past two decades. 00:05:13.060 --> 00:05:14.640 align:middle line:84% So previously, people were saying, 00:05:14.640 --> 00:05:16.320 align:middle line:84% look, people react to incentives. 00:05:16.320 --> 00:05:19.060 align:middle line:84% People react to price changes and so on. 00:05:19.060 --> 00:05:21.840 align:middle line:84% And we're essentially modeling people strongly as 00:05:21.840 --> 00:05:24.540 align:middle line:84% saying, like, no, people are rational actors that 00:05:24.540 --> 00:05:26.250 align:middle line:90% optimize fully. 00:05:26.250 --> 00:05:28.950 align:middle line:84% But with the rise of behavioral economics 00:05:28.950 --> 00:05:33.090 align:middle line:84% in other fields, other subfields of economics, 00:05:33.090 --> 00:05:36.480 align:middle line:84% in development economics, people started 00:05:36.480 --> 00:05:39.240 align:middle line:84% thinking about behavioral development economics 00:05:39.240 --> 00:05:44.640 align:middle line:84% and trying to model a more psychologically realistic view 00:05:44.640 --> 00:05:48.630 align:middle line:84% of human behavior in development economics as well. 00:05:48.630 --> 00:05:51.600 align:middle line:84% And that's to say, just like for the rich, 00:05:51.600 --> 00:05:54.360 align:middle line:84% there are systematic deviations from standard models 00:05:54.360 --> 00:05:56.820 align:middle line:84% in preferences, beliefs, and decision-making. 00:05:56.820 --> 00:05:58.920 align:middle line:84% For example, people might have social preferences, 00:05:58.920 --> 00:05:59.820 align:middle line:90% care about others. 00:05:59.820 --> 00:06:02.620 align:middle line:84% People might loss, might be loss averse 00:06:02.620 --> 00:06:06.750 align:middle line:84% and care more about losses than about gains. 00:06:06.750 --> 00:06:10.150 align:middle line:84% People might be present biased or present focused and so on. 00:06:10.150 --> 00:06:13.920 align:middle line:84% And this is based on universal insights 00:06:13.920 --> 00:06:15.960 align:middle line:84% from psychology about human behavior. 00:06:15.960 --> 00:06:19.380 align:middle line:84% That's to say anybody has potential behavioral biases. 00:06:19.380 --> 00:06:21.510 align:middle line:84% That's true for the rich and for the poor. 00:06:21.510 --> 00:06:24.480 align:middle line:84% And let's try to understand people's choices 00:06:24.480 --> 00:06:27.630 align:middle line:84% better by using these insights from psychology 00:06:27.630 --> 00:06:30.690 align:middle line:84% and other fields in understanding people's 00:06:30.690 --> 00:06:32.020 align:middle line:90% decision-making. 00:06:32.020 --> 00:06:36.210 align:middle line:84% And so most of the work that's in the handbook or people 00:06:36.210 --> 00:06:39.270 align:middle line:84% have studied so far is relying on universal insights 00:06:39.270 --> 00:06:40.920 align:middle line:84% from psychology about human behavior. 00:06:40.920 --> 00:06:43.380 align:middle line:84% That's to say, for example, social preferences, 00:06:43.380 --> 00:06:45.720 align:middle line:84% most people care about others in some ways. 00:06:45.720 --> 00:06:48.630 align:middle line:84% And that's not specific about the poor, about development 00:06:48.630 --> 00:06:50.370 align:middle line:90% economics. 00:06:50.370 --> 00:06:51.990 align:middle line:84% Recently, there has been somewhat more 00:06:51.990 --> 00:06:54.270 align:middle line:84% increasing attention to differences in psychology 00:06:54.270 --> 00:06:56.530 align:middle line:84% across cultures and across the rich and poor. 00:06:56.530 --> 00:06:58.530 align:middle line:84% And that's what I'm going to talk about with you 00:06:58.530 --> 00:07:03.870 align:middle line:84% on Wednesday, which is about poverty might affect people's 00:07:03.870 --> 00:07:07.500 align:middle line:84% decision-making in ways that are specific to poverty, that 00:07:07.500 --> 00:07:10.620 align:middle line:84% could be treatment effects of poverty specifically. 00:07:10.620 --> 00:07:13.080 align:middle line:84% And then more recently is also interactions 00:07:13.080 --> 00:07:16.590 align:middle line:84% between behavioral biases with institutions or market specific 00:07:16.590 --> 00:07:17.850 align:middle line:90% to developing economies. 00:07:17.850 --> 00:07:20.910 align:middle line:84% You might say, well, everybody has similar behavioral biases. 00:07:20.910 --> 00:07:22.560 align:middle line:84% But if people are credit constrained 00:07:22.560 --> 00:07:25.230 align:middle line:84% or if people have other constraints or less guidance 00:07:25.230 --> 00:07:27.450 align:middle line:84% from institutions in their decision-making, 00:07:27.450 --> 00:07:29.070 align:middle line:84% then some of these behavioral biases 00:07:29.070 --> 00:07:32.350 align:middle line:90% might be more pronounced. 00:07:32.350 --> 00:07:33.600 align:middle line:90% Let me stop here for a second. 00:07:33.600 --> 00:07:35.070 align:middle line:90% Are there any questions so far? 00:07:35.070 --> 00:07:38.240 align:middle line:90% 00:07:38.240 --> 00:07:38.780 align:middle line:90% OK. 00:07:38.780 --> 00:07:42.090 align:middle line:84% So then before I get into the specific material, 00:07:42.090 --> 00:07:46.430 align:middle line:84% let me give you a few caveats and common critiques 00:07:46.430 --> 00:07:50.390 align:middle line:84% of behavioral development economics upfront 00:07:50.390 --> 00:07:52.400 align:middle line:90% and some responses to that. 00:07:52.400 --> 00:07:55.430 align:middle line:84% So one broad feature of behavioral economics 00:07:55.430 --> 00:07:57.440 align:middle line:84% more broadly is we're not trying to say, 00:07:57.440 --> 00:08:01.700 align:middle line:84% economics is stupid and doesn't have any applicability 00:08:01.700 --> 00:08:03.440 align:middle line:90% or like the models are wrong. 00:08:03.440 --> 00:08:04.970 align:middle line:84% Instead, what behavioral economics 00:08:04.970 --> 00:08:07.137 align:middle line:84% and behavioral development economics is trying to do 00:08:07.137 --> 00:08:11.390 align:middle line:84% is trying to attempt to augment and improve rather than replace 00:08:11.390 --> 00:08:12.380 align:middle line:90% existing models. 00:08:12.380 --> 00:08:14.450 align:middle line:84% That's to say, let's take some of the assumptions 00:08:14.450 --> 00:08:18.410 align:middle line:84% that people make in neoclassical economic models 00:08:18.410 --> 00:08:20.060 align:middle line:84% and then try to say, well, can we 00:08:20.060 --> 00:08:24.470 align:middle line:84% improve the predictability of our models, 00:08:24.470 --> 00:08:27.860 align:middle line:84% and perhaps resulting policies, by improving some 00:08:27.860 --> 00:08:29.780 align:middle line:84% of the underlying assumptions and trying 00:08:29.780 --> 00:08:31.130 align:middle line:90% to make them more realistic. 00:08:31.130 --> 00:08:33.812 align:middle line:84% But that's to say, prices still matter. 00:08:33.812 --> 00:08:35.270 align:middle line:84% Incentives still matter, et cetera. 00:08:35.270 --> 00:08:36.860 align:middle line:84% Everything from economics applies. 00:08:36.860 --> 00:08:39.409 align:middle line:84% But perhaps utility function looks slightly different 00:08:39.409 --> 00:08:42.913 align:middle line:84% or people might have biased beliefs. 00:08:42.913 --> 00:08:45.080 align:middle line:84% Behavioral developmental economics is also very much 00:08:45.080 --> 00:08:47.600 align:middle line:84% not denying the importance of institutions for development. 00:08:47.600 --> 00:08:49.370 align:middle line:84% In fact, you might say institutions 00:08:49.370 --> 00:08:50.510 align:middle line:90% are even more important. 00:08:50.510 --> 00:08:54.652 align:middle line:84% If people make potentially biased choices in various ways, 00:08:54.652 --> 00:08:56.360 align:middle line:84% well, then there's perhaps more of a role 00:08:56.360 --> 00:08:59.720 align:middle line:84% for institutions to try and improve people's choices. 00:08:59.720 --> 00:09:04.400 align:middle line:84% As I said, it's not to say that real incentives and constraints 00:09:04.400 --> 00:09:05.120 align:middle line:90% don't matter. 00:09:05.120 --> 00:09:07.850 align:middle line:84% So it's not like anything goes and people are just irrational. 00:09:07.850 --> 00:09:09.350 align:middle line:84% There's a lot of structure and a lot 00:09:09.350 --> 00:09:14.000 align:middle line:84% of thought about carefully modeling people's behavior 00:09:14.000 --> 00:09:16.910 align:middle line:90% with a lot of discipline. 00:09:16.910 --> 00:09:19.340 align:middle line:84% And number 4-- and this is particularly important 00:09:19.340 --> 00:09:21.440 align:middle line:84% for the next lecture on Wednesday-- nobody 00:09:21.440 --> 00:09:24.110 align:middle line:84% is trying to blame the poor for their poverty, 00:09:24.110 --> 00:09:30.080 align:middle line:84% for perhaps making mistakes or other choices that 00:09:30.080 --> 00:09:32.960 align:middle line:84% might be suboptimal in some ways, in part 00:09:32.960 --> 00:09:35.900 align:middle line:84% because usually, as I said, we're 00:09:35.900 --> 00:09:40.400 align:middle line:84% typically studying universal psychological factors 00:09:40.400 --> 00:09:42.440 align:middle line:90% that apply to anybody. 00:09:42.440 --> 00:09:46.640 align:middle line:84% And second, behavioral biases, or in particular, differences 00:09:46.640 --> 00:09:49.473 align:middle line:84% in people's utility function are not blameworthy in any way. 00:09:49.473 --> 00:09:51.140 align:middle line:84% It's to say, people might have reference 00:09:51.140 --> 00:09:52.198 align:middle line:90% dependent preferences. 00:09:52.198 --> 00:09:53.990 align:middle line:84% And that's just what their utility function 00:09:53.990 --> 00:09:55.707 align:middle line:90% looks like in any way. 00:09:55.707 --> 00:09:57.290 align:middle line:84% And then, of course, there is the idea 00:09:57.290 --> 00:10:02.690 align:middle line:84% that poverty itself might cause some potentially 00:10:02.690 --> 00:10:06.260 align:middle line:84% suboptimal choices, again, which we'll talk about on Wednesday. 00:10:06.260 --> 00:10:10.250 align:middle line:84% Then you might say, well, isn't that a very paternalistic 00:10:10.250 --> 00:10:14.600 align:middle line:84% to be trying to tell people, and in particular the poor, what 00:10:14.600 --> 00:10:15.750 align:middle line:90% to do or not to do. 00:10:15.750 --> 00:10:18.380 align:middle line:84% And I think there is some truth to that critique in some ways. 00:10:18.380 --> 00:10:20.690 align:middle line:84% And they're saying, well, once you take a stance 00:10:20.690 --> 00:10:25.460 align:middle line:84% and say, well, people are making mistakes in their choices, 00:10:25.460 --> 00:10:29.660 align:middle line:84% people are often then engaging in potentially paternalistic 00:10:29.660 --> 00:10:30.620 align:middle line:90% policies. 00:10:30.620 --> 00:10:34.310 align:middle line:84% The best of behavioral economics tends 00:10:34.310 --> 00:10:36.350 align:middle line:84% to have what's called libertarian paternalism. 00:10:36.350 --> 00:10:38.360 align:middle line:84% It's trying to, essentially, design policies, 00:10:38.360 --> 00:10:41.480 align:middle line:84% for example, defaults or other types of choices where 00:10:41.480 --> 00:10:45.020 align:middle line:84% you help people who have behavioral biases while not 00:10:45.020 --> 00:10:48.960 align:middle line:84% interfering with other freedom of choice overall. 00:10:48.960 --> 00:10:51.170 align:middle line:84% So if I set a default in certain ways 00:10:51.170 --> 00:10:53.087 align:middle line:84% and say if you don't do anything, 00:10:53.087 --> 00:10:54.920 align:middle line:84% you get into the right insurance plan that's 00:10:54.920 --> 00:10:56.990 align:middle line:84% best for everybody on average, then 00:10:56.990 --> 00:10:59.732 align:middle line:84% if you still want to opt out of it, then you can do that. 00:10:59.732 --> 00:11:01.190 align:middle line:84% So that's kind of like the best of, 00:11:01.190 --> 00:11:05.330 align:middle line:84% perhaps, what's called libertarian paternalism. 00:11:05.330 --> 00:11:07.730 align:middle line:84% And then, in fact, it's not [INAUDIBLE] 00:11:07.730 --> 00:11:11.668 align:middle line:84% to say that everything that's found in lab experiments 00:11:11.668 --> 00:11:12.710 align:middle line:90% is going to be important. 00:11:12.710 --> 00:11:16.100 align:middle line:84% There are some examples, which are discussed in the handbook 00:11:16.100 --> 00:11:20.160 align:middle line:84% chapter where some lab experimental evidence, in fact, 00:11:20.160 --> 00:11:24.270 align:middle line:84% does not seem relevant in the real world. 00:11:24.270 --> 00:11:27.330 align:middle line:84% And so there are some examples of that as well. 00:11:27.330 --> 00:11:31.250 align:middle line:84% With that said, let me give you a brief overview before we then 00:11:31.250 --> 00:11:33.510 align:middle line:90% talk about the Euler puzzle. 00:11:33.510 --> 00:11:36.590 align:middle line:84% So what are the topics that behavioral economists 00:11:36.590 --> 00:11:38.930 align:middle line:84% or behavioral development economists have studied? 00:11:38.930 --> 00:11:43.400 align:middle line:84% You can think of this as three broad categories 00:11:43.400 --> 00:11:46.040 align:middle line:84% of different issues from behavioral economics 00:11:46.040 --> 00:11:47.510 align:middle line:90% that have been applied. 00:11:47.510 --> 00:11:49.220 align:middle line:90% One is nonstandard preferences. 00:11:49.220 --> 00:11:52.088 align:middle line:84% We're going to talk about this in a second, which is to say, 00:11:52.088 --> 00:11:53.630 align:middle line:84% people might have a different utility 00:11:53.630 --> 00:11:57.500 align:middle line:84% function from the typical neoclassical model. 00:11:57.500 --> 00:11:59.600 align:middle line:84% People might have different time preferences. 00:11:59.600 --> 00:12:01.430 align:middle line:84% People have different risk preferences 00:12:01.430 --> 00:12:03.260 align:middle line:90% or social preferences. 00:12:03.260 --> 00:12:05.300 align:middle line:84% Second, people's beliefs might be 00:12:05.300 --> 00:12:09.770 align:middle line:84% biased in certain ways that might be people have trouble 00:12:09.770 --> 00:12:11.090 align:middle line:90% engaging in Bayesian learning. 00:12:11.090 --> 00:12:13.340 align:middle line:84% Bayesian learning is really hard and figuring this out 00:12:13.340 --> 00:12:14.960 align:middle line:90% exactly is hard to do. 00:12:14.960 --> 00:12:17.690 align:middle line:84% People might be naive about their future behavior. 00:12:17.690 --> 00:12:20.360 align:middle line:84% They might be more optimistic about their own abilities. 00:12:20.360 --> 00:12:23.240 align:middle line:84% They might be more optimistic about overcoming present bias 00:12:23.240 --> 00:12:28.370 align:middle line:84% in the future or about how good they 00:12:28.370 --> 00:12:31.988 align:middle line:84% are, for example, at dealing with temptation in the future. 00:12:31.988 --> 00:12:34.280 align:middle line:84% They might also have what's called motivated reasoning. 00:12:34.280 --> 00:12:36.020 align:middle line:84% They might want to think that they're 00:12:36.020 --> 00:12:37.280 align:middle line:90% better than they actually are. 00:12:37.280 --> 00:12:39.488 align:middle line:84% They might want to think that the healthier than they 00:12:39.488 --> 00:12:43.560 align:middle line:84% actually are because they derive utility from that. 00:12:43.560 --> 00:12:47.250 align:middle line:84% And that leads to systematically overconfident beliefs. 00:12:47.250 --> 00:12:49.660 align:middle line:84% And finally, there's nonstandard decision-making, 00:12:49.660 --> 00:12:51.900 align:middle line:84% which is people might have limited attention 00:12:51.900 --> 00:12:54.575 align:middle line:84% or memory or mental accounting or default effects. 00:12:54.575 --> 00:12:55.950 align:middle line:84% You can think of this as anything 00:12:55.950 --> 00:12:57.480 align:middle line:84% that doesn't fit into one and two 00:12:57.480 --> 00:12:59.290 align:middle line:90% about preferences and beliefs. 00:12:59.290 --> 00:13:02.550 align:middle line:84% Even if you understand people's preferences and beliefs, 00:13:02.550 --> 00:13:05.880 align:middle line:84% people might still, for example, just not pay attention 00:13:05.880 --> 00:13:10.320 align:middle line:84% to certain options or they might react strongly 00:13:10.320 --> 00:13:13.140 align:middle line:84% to framing or default effects that can't really 00:13:13.140 --> 00:13:16.110 align:middle line:84% necessarily be explained by differences 00:13:16.110 --> 00:13:18.210 align:middle line:90% in preferences and beliefs. 00:13:18.210 --> 00:13:20.880 align:middle line:84% Now what makes behavioral development 00:13:20.880 --> 00:13:22.140 align:middle line:90% particularly exciting? 00:13:22.140 --> 00:13:24.750 align:middle line:84% Well, I think there are different features 00:13:24.750 --> 00:13:27.160 align:middle line:84% of different fields that come together. 00:13:27.160 --> 00:13:29.617 align:middle line:84% On the one hand, there's development economics, 00:13:29.617 --> 00:13:31.200 align:middle line:84% which of course, you all are studying, 00:13:31.200 --> 00:13:34.590 align:middle line:84% which is the experimental approach to tackling 00:13:34.590 --> 00:13:37.920 align:middle line:84% high stakes and policy relevant questions in field settings. 00:13:37.920 --> 00:13:40.290 align:middle line:90% So that's still maintained. 00:13:40.290 --> 00:13:43.080 align:middle line:84% But then there's influence from experimental economics, which 00:13:43.080 --> 00:13:46.860 align:middle line:84% are very careful procedures and methods 00:13:46.860 --> 00:13:51.090 align:middle line:84% and controlled lab settings to focus on mechanisms to measure 00:13:51.090 --> 00:13:53.760 align:middle line:84% beliefs, preferences, and so on very carefully 00:13:53.760 --> 00:13:57.930 align:middle line:84% to test specific theories about how people make choices. 00:13:57.930 --> 00:14:00.720 align:middle line:84% And usually that's in pretty somewhat contrived settings 00:14:00.720 --> 00:14:03.060 align:middle line:84% and experimental lab settings that you can't really 00:14:03.060 --> 00:14:05.640 align:middle line:84% necessarily replicate in the field. 00:14:05.640 --> 00:14:09.000 align:middle line:84% But behavioral development economics 00:14:09.000 --> 00:14:11.010 align:middle line:84% tries to use at least part of this 00:14:11.010 --> 00:14:13.393 align:middle line:84% in and transport it into field settings. 00:14:13.393 --> 00:14:15.810 align:middle line:84% I think there's always going to be some compromise because 00:14:15.810 --> 00:14:19.140 align:middle line:84% at the end of the day, if you do a field experiment somewhere 00:14:19.140 --> 00:14:22.500 align:middle line:84% in rural India, it's not going to be as cleanly controlled 00:14:22.500 --> 00:14:24.690 align:middle line:84% as if you do like a lab experiment 00:14:24.690 --> 00:14:29.010 align:middle line:90% here at MIT or any other place. 00:14:29.010 --> 00:14:31.500 align:middle line:84% But the idea of carefully controlling 00:14:31.500 --> 00:14:33.210 align:middle line:84% the environment and people's choices 00:14:33.210 --> 00:14:35.790 align:middle line:84% and measuring beliefs, preferences, et cetera, 00:14:35.790 --> 00:14:38.400 align:middle line:84% is a pretty strong idea in behavioral development 00:14:38.400 --> 00:14:39.510 align:middle line:90% economics. 00:14:39.510 --> 00:14:43.500 align:middle line:84% And then third, there are lots of ideas from psychology, 00:14:43.500 --> 00:14:47.640 align:middle line:84% other fields, for example, psychiatry, sleep medicine, 00:14:47.640 --> 00:14:50.100 align:middle line:84% et cetera that you can essentially 00:14:50.100 --> 00:14:51.960 align:middle line:90% use an arbitrage in some ways. 00:14:51.960 --> 00:14:54.210 align:middle line:84% That's to say, there's lots of interesting things 00:14:54.210 --> 00:14:57.210 align:middle line:84% that other fields have studied that you might want to bring 00:14:57.210 --> 00:14:58.590 align:middle line:90% into development economics. 00:14:58.590 --> 00:15:03.060 align:middle line:84% And usually, many of them are related to psychology. 00:15:03.060 --> 00:15:04.980 align:middle line:84% There's also, more recently, work 00:15:04.980 --> 00:15:07.450 align:middle line:84% in sociology or anthropology, et cetera. 00:15:07.450 --> 00:15:10.860 align:middle line:84% There's essentially lots of ideas or novel concepts 00:15:10.860 --> 00:15:12.360 align:middle line:84% where you can write models about it. 00:15:12.360 --> 00:15:15.790 align:middle line:84% You can test these models and so on. 00:15:15.790 --> 00:15:18.750 align:middle line:84% So there's lots of space for creativity and novel ideas 00:15:18.750 --> 00:15:19.650 align:middle line:90% or insights. 00:15:19.650 --> 00:15:21.750 align:middle line:84% And these insights then can be applied 00:15:21.750 --> 00:15:25.290 align:middle line:84% to lots of different topics in development economics. 00:15:25.290 --> 00:15:29.490 align:middle line:84% So here's a broad overview of this handbook chapter. 00:15:29.490 --> 00:15:33.305 align:middle line:84% And it's very long, so we didn't require you to read it. 00:15:33.305 --> 00:15:35.680 align:middle line:84% But if you're interested in any of the topics, hopefully, 00:15:35.680 --> 00:15:38.830 align:middle line:84% you can pick and choose what you're most interested in. 00:15:38.830 --> 00:15:41.910 align:middle line:84% And so it's essentially organized 00:15:41.910 --> 00:15:45.390 align:middle line:84% by canonical topics of development economics 00:15:45.390 --> 00:15:50.910 align:middle line:84% as you will go through mostly in your class with Esther and Ben. 00:15:50.910 --> 00:15:53.350 align:middle line:84% Of course, not everything, but most of it. 00:15:53.350 --> 00:15:55.920 align:middle line:84% And so you can think of this as like using insights 00:15:55.920 --> 00:15:57.990 align:middle line:84% from behavioral economics and applying them 00:15:57.990 --> 00:16:01.360 align:middle line:84% to each of the topics in development economics. 00:16:01.360 --> 00:16:04.440 align:middle line:84% So we're going to talk about the conceptual framework 00:16:04.440 --> 00:16:05.350 align:middle line:90% in a second. 00:16:05.350 --> 00:16:09.960 align:middle line:84% But then the topics to apply these insights to 00:16:09.960 --> 00:16:13.620 align:middle line:84% are health or health behavior, savings, risk insurance, 00:16:13.620 --> 00:16:16.500 align:middle line:84% technology adoption, and labor markets, 00:16:16.500 --> 00:16:18.060 align:middle line:90% as well as behavioral firms. 00:16:18.060 --> 00:16:20.430 align:middle line:84% And then there's like two broad other topics, which 00:16:20.430 --> 00:16:22.080 align:middle line:90% just don't fit quite into this. 00:16:22.080 --> 00:16:25.080 align:middle line:84% One is about social preferences, culture, and development. 00:16:25.080 --> 00:16:27.210 align:middle line:84% This is the type of work that Nathan [INAUDIBLE] 00:16:27.210 --> 00:16:31.170 align:middle line:84% and Ben [INAUDIBLE] and other people at Harvard 00:16:31.170 --> 00:16:32.550 align:middle line:90% are working on. 00:16:32.550 --> 00:16:34.680 align:middle line:84% And the psychology of poverty, which I'm going 00:16:34.680 --> 00:16:37.140 align:middle line:90% to tell you about on Wednesday. 00:16:37.140 --> 00:16:39.220 align:middle line:84% Let me stop for a second before I get started. 00:16:39.220 --> 00:16:40.290 align:middle line:90% Any questions? 00:16:40.290 --> 00:16:43.140 align:middle line:90% 00:16:43.140 --> 00:16:45.710 align:middle line:90% OK, right. 00:16:45.710 --> 00:16:48.260 align:middle line:84% So the starting point of all of this 00:16:48.260 --> 00:16:51.950 align:middle line:84% is that there are high returns to capital in many contacts. 00:16:51.950 --> 00:16:55.190 align:middle line:84% And so Esther and [INAUDIBLE] have a very nice handbook 00:16:55.190 --> 00:16:57.320 align:middle line:84% chapter in the Handbook of Economic Growth that 00:16:57.320 --> 00:17:00.560 align:middle line:84% talks about different pieces of evidence. 00:17:00.560 --> 00:17:02.360 align:middle line:84% There is borrowing at high rates. 00:17:02.360 --> 00:17:05.359 align:middle line:84% People have borrowed 70% to 100% annual rates 00:17:05.359 --> 00:17:06.960 align:middle line:90% and more in many settings. 00:17:06.960 --> 00:17:08.819 align:middle line:84% There's also some evidence, for example, 00:17:08.819 --> 00:17:13.849 align:middle line:84% from Chennai, where there are small scale fruit 00:17:13.849 --> 00:17:16.920 align:middle line:84% vendors in Chennai who borrow at daily rates of about 5%. 00:17:16.920 --> 00:17:21.500 align:middle line:84% So what they do is to have about 1,000 rupees of goods 00:17:21.500 --> 00:17:23.599 align:middle line:84% and fruits and vegetables that they're 00:17:23.599 --> 00:17:24.980 align:middle line:90% trying to sell on any day. 00:17:24.980 --> 00:17:28.369 align:middle line:84% They, in the morning, borrow these 1,000 rupees. 00:17:28.369 --> 00:17:30.740 align:middle line:84% And then in the evening, they pay back 50 rupees 00:17:30.740 --> 00:17:33.410 align:middle line:90% of that, which is 5% daily. 00:17:33.410 --> 00:17:34.980 align:middle line:90% And they do that every day. 00:17:34.980 --> 00:17:39.630 align:middle line:84% And so that's like an extremely high interest rate, 00:17:39.630 --> 00:17:41.700 align:middle line:84% which suggests that if you're doing that, 00:17:41.700 --> 00:17:45.950 align:middle line:84% you must have really high returns to capital. 00:17:45.950 --> 00:17:47.570 align:middle line:84% On the flip side, there's high returns 00:17:47.570 --> 00:17:48.740 align:middle line:90% to small business grants. 00:17:48.740 --> 00:17:50.365 align:middle line:84% So there's randomized controlled trials 00:17:50.365 --> 00:17:54.170 align:middle line:84% where people are randomly given some money versus not. 00:17:54.170 --> 00:17:56.070 align:middle line:84% There's high returns to inventory, 00:17:56.070 --> 00:17:58.730 align:middle line:84% so some work by Michael Kremer and co-authors on that. 00:17:58.730 --> 00:18:00.770 align:middle line:84% There's also a predictable large increase 00:18:00.770 --> 00:18:03.860 align:middle line:90% in prices between seasons. 00:18:03.860 --> 00:18:06.920 align:middle line:84% With these fluctuations of prices, 00:18:06.920 --> 00:18:11.302 align:middle line:84% if you just had a lot of money and perhaps storage capacities, 00:18:11.302 --> 00:18:12.885 align:middle line:84% then you could arbitrage, essentially, 00:18:12.885 --> 00:18:14.135 align:middle line:90% and make quite a bit of money. 00:18:14.135 --> 00:18:19.070 align:middle line:84% And Lauren [INAUDIBLE] was visiting us this semester-- 00:18:19.070 --> 00:18:21.500 align:middle line:84% and Marshall Burke and others have a very nice paper 00:18:21.500 --> 00:18:24.020 align:middle line:90% that discusses this issue. 00:18:24.020 --> 00:18:26.607 align:middle line:84% Now you might say, OK, that's great, 00:18:26.607 --> 00:18:28.940 align:middle line:84% but what does this have to do with behavioral economics? 00:18:28.940 --> 00:18:33.650 align:middle line:84% Well, thinking about the Euler equation 00:18:33.650 --> 00:18:35.700 align:middle line:84% is quite helpful in some sense to say, OK, 00:18:35.700 --> 00:18:38.690 align:middle line:84% what does the neoclassical model have to say about explaining 00:18:38.690 --> 00:18:40.470 align:middle line:90% high returns to capital? 00:18:40.470 --> 00:18:42.700 align:middle line:90% So what is the Euler equation? 00:18:42.700 --> 00:18:44.770 align:middle line:84% I've talked a lot, so I'd like to hear from you. 00:18:44.770 --> 00:18:54.740 align:middle line:90% 00:18:54.740 --> 00:18:58.160 align:middle line:84% Who's taking first year macro right now, or has taken? 00:18:58.160 --> 00:19:01.890 align:middle line:90% 00:19:01.890 --> 00:19:02.790 align:middle line:90% Yeah? 00:19:02.790 --> 00:19:04.380 align:middle line:84% AUDIENCE: It's like your consumption 00:19:04.380 --> 00:19:08.288 align:middle line:84% today versus tomorrow based on your production of your assets. 00:19:08.288 --> 00:19:09.330 align:middle line:90% FRANK SCHILBACH: Exactly. 00:19:09.330 --> 00:19:10.788 align:middle line:84% So the Euler equation, what it does 00:19:10.788 --> 00:19:13.770 align:middle line:84% it relates your consumption today to your consumption 00:19:13.770 --> 00:19:17.280 align:middle line:84% tomorrow, more precisely your marginal utility of consumption 00:19:17.280 --> 00:19:19.900 align:middle line:84% today to your marginal consumption tomorrow 00:19:19.900 --> 00:19:21.420 align:middle line:90% or in any future period. 00:19:21.420 --> 00:19:23.370 align:middle line:84% And as you say, it depends, essentially, 00:19:23.370 --> 00:19:26.160 align:middle line:84% on what are the factors that link these two. 00:19:26.160 --> 00:19:30.143 align:middle line:84% You said something about returns to capital. 00:19:30.143 --> 00:19:32.310 align:middle line:84% AUDIENCE: Yeah, like how productive your assets are. 00:19:32.310 --> 00:19:32.730 align:middle line:90% FRANK SCHILBACH: Exactly. 00:19:32.730 --> 00:19:34.730 align:middle line:84% So it's essentially about the return to capital, 00:19:34.730 --> 00:19:35.340 align:middle line:90% if you invest. 00:19:35.340 --> 00:19:38.220 align:middle line:84% If you don't consume today, you can save it in some ways 00:19:38.220 --> 00:19:40.480 align:middle line:90% and invest it in certain assets. 00:19:40.480 --> 00:19:42.120 align:middle line:90% So that's the return to capital. 00:19:42.120 --> 00:19:46.380 align:middle line:84% And then there is, as well, a delta. 00:19:46.380 --> 00:19:47.820 align:middle line:90% What is delta? 00:19:47.820 --> 00:19:48.960 align:middle line:90% Or what is this doing here? 00:19:48.960 --> 00:19:53.350 align:middle line:90% 00:19:53.350 --> 00:19:57.400 align:middle line:84% So a delta is often also used for depreciation-- not here. 00:19:57.400 --> 00:19:57.940 align:middle line:90% Yeah? 00:19:57.940 --> 00:19:58.590 align:middle line:90% AUDIENCE: Discounting. 00:19:58.590 --> 00:20:00.548 align:middle line:84% FRANK SCHILBACH: Yeah, so the discounting rate. 00:20:00.548 --> 00:20:03.010 align:middle line:84% If you think about, should I consume today versus tomorrow? 00:20:03.010 --> 00:20:04.660 align:middle line:84% If you think about any marginal units 00:20:04.660 --> 00:20:06.880 align:middle line:84% that you might want to consume versus save, 00:20:06.880 --> 00:20:10.210 align:middle line:84% you either get your marginal utility of consumption today 00:20:10.210 --> 00:20:11.380 align:middle line:90% or you save it. 00:20:11.380 --> 00:20:12.500 align:middle line:90% Now you can invest it. 00:20:12.500 --> 00:20:14.440 align:middle line:90% So we have, in some sense-- 00:20:14.440 --> 00:20:16.960 align:middle line:84% we wrote this more generally as like a production function 00:20:16.960 --> 00:20:20.850 align:middle line:84% you can invest it in certain ways into your capital stock. 00:20:20.850 --> 00:20:23.530 align:middle line:84% You might be buying machines or like invest it 00:20:23.530 --> 00:20:26.540 align:middle line:90% in your shop or the like. 00:20:26.540 --> 00:20:32.080 align:middle line:84% And that leads to the marginal product of capital tomorrow 00:20:32.080 --> 00:20:34.480 align:middle line:84% times the marginal utility of consumption tomorrow, 00:20:34.480 --> 00:20:36.760 align:middle line:84% so whatever you additionally produce tomorrow, 00:20:36.760 --> 00:20:38.105 align:middle line:90% you can consume now. 00:20:38.105 --> 00:20:40.480 align:middle line:84% And that needs to be discounted because people might have 00:20:40.480 --> 00:20:42.910 align:middle line:84% discounting or different discounting between today 00:20:42.910 --> 00:20:44.650 align:middle line:90% and tomorrow. 00:20:44.650 --> 00:20:47.250 align:middle line:90% Is this clear what's here? 00:20:47.250 --> 00:20:50.910 align:middle line:84% All right, so now I'm saying, it implies unrealistically high 00:20:50.910 --> 00:20:52.560 align:middle line:90% consumption growth rates. 00:20:52.560 --> 00:20:54.910 align:middle line:90% Why is that? 00:20:54.910 --> 00:20:59.020 align:middle line:84% So what did I tell you about F prime of K? 00:20:59.020 --> 00:21:02.580 align:middle line:90% 00:21:02.580 --> 00:21:03.443 align:middle line:90% Yeah? 00:21:03.443 --> 00:21:04.860 align:middle line:84% AUDIENCE: It's really high so that 00:21:04.860 --> 00:21:11.710 align:middle line:84% in order to make up for the fact that [INAUDIBLE] for really 00:21:11.710 --> 00:21:14.878 align:middle line:84% more [INAUDIBLE] consumption to compete with that. 00:21:14.878 --> 00:21:15.920 align:middle line:90% FRANK SCHILBACH: Exactly. 00:21:15.920 --> 00:21:17.610 align:middle line:90% So you can rewrite this. 00:21:17.610 --> 00:21:19.850 align:middle line:84% And this is a little bit fast here on the slide. 00:21:19.850 --> 00:21:23.210 align:middle line:84% But I think if you can write this to convince yourself-- 00:21:23.210 --> 00:21:28.730 align:middle line:84% for example, if you have log utility and F prime of K of 50% 00:21:28.730 --> 00:21:30.470 align:middle line:84% annually, which seems reasonable, 00:21:30.470 --> 00:21:34.100 align:middle line:84% and delta of 0.96, which is as you know people slightly 00:21:34.100 --> 00:21:37.910 align:middle line:84% discount the future about by 4% or 5%, 00:21:37.910 --> 00:21:39.680 align:middle line:84% then you get essentially a consumption, 00:21:39.680 --> 00:21:42.005 align:middle line:84% implied consumption growth of 44%. 00:21:42.005 --> 00:21:44.750 align:middle line:90% 00:21:44.750 --> 00:21:47.250 align:middle line:84% And so you might say, well, is log utility the right utility 00:21:47.250 --> 00:21:47.750 align:middle line:90% function? 00:21:47.750 --> 00:21:49.470 align:middle line:84% You can have different utility function. 00:21:49.470 --> 00:21:51.740 align:middle line:84% If you have a constant into temporal elasticity 00:21:51.740 --> 00:21:54.950 align:middle line:84% of substitution, utility where sigma equals 2, 00:21:54.950 --> 00:22:00.200 align:middle line:84% you get something like implied consumption growth of 20%. 00:22:00.200 --> 00:22:04.970 align:middle line:84% But essentially, it implies very unreasonable increases 00:22:04.970 --> 00:22:07.430 align:middle line:84% or consumption growth in the future. 00:22:07.430 --> 00:22:09.140 align:middle line:84% You might say, well, in one year, 00:22:09.140 --> 00:22:11.000 align:middle line:84% increasing your consumption by 20% 00:22:11.000 --> 00:22:13.430 align:middle line:84% or whatever, 40%, is not that large. 00:22:13.430 --> 00:22:16.250 align:middle line:84% But once you move that forward to 20 years, you get, 00:22:16.250 --> 00:22:20.370 align:middle line:84% essentially, people's consumption would explode. 00:22:20.370 --> 00:22:24.150 align:middle line:90% So something is not right here. 00:22:24.150 --> 00:22:27.200 align:middle line:84% So now what could we do to deal with this apparent puzzle? 00:22:27.200 --> 00:22:32.728 align:middle line:90% 00:22:32.728 --> 00:22:35.100 align:middle line:84% AUDIENCE: [INAUDIBLE] at a discount rate. 00:22:35.100 --> 00:22:36.600 align:middle line:84% FRANK SCHILBACH: Yeah, so I gave you 00:22:36.600 --> 00:22:38.370 align:middle line:90% like a discount rate of 0.96. 00:22:38.370 --> 00:22:40.170 align:middle line:84% But you might say, wow, why 0.96? 00:22:40.170 --> 00:22:42.628 align:middle line:90% That's not the right one. 00:22:42.628 --> 00:22:43.170 align:middle line:90% That's right. 00:22:43.170 --> 00:22:46.470 align:middle line:90% And that's one solution. 00:22:46.470 --> 00:22:48.000 align:middle line:90% And what else could be done? 00:22:48.000 --> 00:22:50.070 align:middle line:84% Or what else we try and fix or change? 00:22:50.070 --> 00:22:54.540 align:middle line:90% 00:22:54.540 --> 00:22:55.382 align:middle line:90% Yeah? 00:22:55.382 --> 00:22:56.738 align:middle line:90% AUDIENCE: Borrowing constraints? 00:22:56.738 --> 00:23:02.475 align:middle line:90% 00:23:02.475 --> 00:23:03.350 align:middle line:90% FRANK SCHILBACH: Yes. 00:23:03.350 --> 00:23:04.382 align:middle line:90% So it could be-- 00:23:04.382 --> 00:23:05.840 align:middle line:84% which is related to this, but yeah, 00:23:05.840 --> 00:23:11.180 align:middle line:84% people might just not have any money. 00:23:11.180 --> 00:23:13.040 align:middle line:84% But using sometimes the Euler equation still 00:23:13.040 --> 00:23:15.050 align:middle line:84% needs to hold at the end of the day. 00:23:15.050 --> 00:23:17.810 align:middle line:84% But I'll get back to borrowing constraints in a bit. 00:23:17.810 --> 00:23:19.140 align:middle line:90% 00:23:19.140 --> 00:23:20.630 align:middle line:90% What about the F prime of K? 00:23:20.630 --> 00:23:22.190 align:middle line:84% How can we get that down, perhaps? 00:23:22.190 --> 00:23:26.785 align:middle line:90% 00:23:26.785 --> 00:23:28.910 align:middle line:84% If I invest in my business and make a lot of money, 00:23:28.910 --> 00:23:31.640 align:middle line:90% what happens? 00:23:31.640 --> 00:23:32.600 align:middle line:90% STUDENT: [INAUDIBLE] 00:23:32.600 --> 00:23:35.690 align:middle line:84% FRANK SCHILBACH: Yes, I get taxed by either-- 00:23:35.690 --> 00:23:37.850 align:middle line:84% the government taxes might be really high 00:23:37.850 --> 00:23:40.070 align:middle line:84% or there might be some other informal taxes. 00:23:40.070 --> 00:23:42.560 align:middle line:84% Might be my neighbors, or my uncle, or my friends, 00:23:42.560 --> 00:23:44.840 align:middle line:84% or whatever, if I make a lot of money, perhaps 00:23:44.840 --> 00:23:46.170 align:middle line:90% they take it all away from me. 00:23:46.170 --> 00:23:49.110 align:middle line:84% And then I might not have that much left. 00:23:49.110 --> 00:23:51.440 align:middle line:84% So you might see that there's a high return to capital, 00:23:51.440 --> 00:23:53.420 align:middle line:84% but that return to capital does not go to me, 00:23:53.420 --> 00:24:00.200 align:middle line:84% so I don't have any necessary incentives to invest. 00:24:00.200 --> 00:24:03.560 align:middle line:84% Now, this is a little bit some technical points. 00:24:03.560 --> 00:24:05.630 align:middle line:84% And I don't want to spend too much time on it. 00:24:05.630 --> 00:24:07.505 align:middle line:84% And maybe I will have some time at some point 00:24:07.505 --> 00:24:09.560 align:middle line:84% to talk about this in recitation. 00:24:09.560 --> 00:24:12.050 align:middle line:84% But so the puzzle persists even when 00:24:12.050 --> 00:24:14.610 align:middle line:84% you have non-concave production functions. 00:24:14.610 --> 00:24:16.400 align:middle line:84% So one thing I told you to start with is, 00:24:16.400 --> 00:24:19.350 align:middle line:84% say, the production function is concave, which you know 00:24:19.350 --> 00:24:21.740 align:middle line:84% is a very reasonable assumption in many settings. 00:24:21.740 --> 00:24:24.170 align:middle line:84% But perhaps Esther was telling you last week, 00:24:24.170 --> 00:24:27.325 align:middle line:84% perhaps it's, in fact, non-concave. 00:24:27.325 --> 00:24:29.450 align:middle line:84% And then you might say, well, what's going on then? 00:24:29.450 --> 00:24:31.640 align:middle line:84% Well, then you'll have, essentially, 00:24:31.640 --> 00:24:34.730 align:middle line:84% multiple steady states and sustained poverty 00:24:34.730 --> 00:24:36.320 align:middle line:90% below a certain threshold. 00:24:36.320 --> 00:24:39.020 align:middle line:84% But importantly, the observed initial conditions, 00:24:39.020 --> 00:24:40.580 align:middle line:84% in particular the steady state, also 00:24:40.580 --> 00:24:43.710 align:middle line:84% needs to be consistent with the model. 00:24:43.710 --> 00:24:45.720 align:middle line:84% So the Euler equation needs to be satisfied, 00:24:45.720 --> 00:24:47.820 align:middle line:84% the first-order condition that I just showed you. 00:24:47.820 --> 00:24:52.730 align:middle line:84% And so it can't be that people are always in transition. 00:24:52.730 --> 00:24:55.032 align:middle line:84% It might be that some people are moving from one place 00:24:55.032 --> 00:24:56.990 align:middle line:84% to another, from one steady state to the other, 00:24:56.990 --> 00:24:59.060 align:middle line:84% but actually in steady state in the end, 00:24:59.060 --> 00:25:02.660 align:middle line:84% it must be that the Euler equation holds. 00:25:02.660 --> 00:25:04.790 align:middle line:84% There's also some other implications. 00:25:04.790 --> 00:25:08.210 align:middle line:84% For example, poverty trap models suggest transformative effects 00:25:08.210 --> 00:25:12.530 align:middle line:84% of credit, which didn't seem to be the case in many settings, 00:25:12.530 --> 00:25:16.820 align:middle line:84% at least not for a large fraction of people. 00:25:16.820 --> 00:25:19.310 align:middle line:84% Second, you might say, well, what about stochastic income 00:25:19.310 --> 00:25:20.310 align:middle line:90% and risk aversion? 00:25:20.310 --> 00:25:22.820 align:middle line:90% So it's not like-- 00:25:22.820 --> 00:25:24.620 align:middle line:84% what I had said, written down here, 00:25:24.620 --> 00:25:26.518 align:middle line:90% there's no risk in this model. 00:25:26.518 --> 00:25:28.560 align:middle line:84% And you might say, well, things are really risky. 00:25:28.560 --> 00:25:30.935 align:middle line:84% And if people think about investing in fertilizer, buying 00:25:30.935 --> 00:25:33.050 align:middle line:84% machines, and the like, maybe the rate of return 00:25:33.050 --> 00:25:35.190 align:middle line:90% is really volatile. 00:25:35.190 --> 00:25:39.330 align:middle line:84% And so then, it might be just not worth investing. 00:25:39.330 --> 00:25:43.410 align:middle line:84% And so it turns out that once you write this down-- 00:25:43.410 --> 00:25:47.600 align:middle line:84% you can write down a model with several assets 00:25:47.600 --> 00:25:49.820 align:middle line:84% and risk for each of these assets-- 00:25:49.820 --> 00:25:52.880 align:middle line:84% you've got a stochastic version of the Euler equation 00:25:52.880 --> 00:25:58.500 align:middle line:84% and some more nuanced predictions, where essentially, 00:25:58.500 --> 00:26:01.970 align:middle line:84% it depends on whether the assets positively or negatively covary 00:26:01.970 --> 00:26:04.590 align:middle line:90% with consumption. 00:26:04.590 --> 00:26:06.080 align:middle line:84% And so can think about this model. 00:26:06.080 --> 00:26:09.770 align:middle line:84% But I think one thing that then violates this model pretty 00:26:09.770 --> 00:26:14.030 align:middle line:84% starkly is the idea that if things were really risky, 00:26:14.030 --> 00:26:16.280 align:middle line:84% then people should anticipate that. 00:26:16.280 --> 00:26:19.310 align:middle line:84% People should have buffer stocks and should start saving, 00:26:19.310 --> 00:26:23.310 align:middle line:84% and make sure that they can deal with that type of risk. 00:26:23.310 --> 00:26:26.300 align:middle line:84% So you would essentially build up a large buffer stock 00:26:26.300 --> 00:26:31.530 align:middle line:84% to deal with these kinds of risks that might apply. 00:26:31.530 --> 00:26:35.407 align:middle line:84% And then if you have a large buffer stock that's safe, 00:26:35.407 --> 00:26:36.990 align:middle line:84% then your consumption should, in fact, 00:26:36.990 --> 00:26:40.100 align:middle line:84% not systematically vary with income or income volatility 00:26:40.100 --> 00:26:42.870 align:middle line:84% because you've essentially smoothed it out. 00:26:42.870 --> 00:26:44.840 align:middle line:84% But when you look at how much people 00:26:44.840 --> 00:26:46.550 align:middle line:84% are in fact saving in many settings, 00:26:46.550 --> 00:26:48.740 align:middle line:84% and Esther will talk with you specifically 00:26:48.740 --> 00:26:50.630 align:middle line:84% in a lecture about savings, you will 00:26:50.630 --> 00:26:53.840 align:middle line:84% see that not a lot of people have precautionary savings 00:26:53.840 --> 00:26:59.390 align:middle line:84% or purchase insurance or the like. 00:26:59.390 --> 00:27:06.410 align:middle line:84% And so there seem to be a number of incorrect predictions. 00:27:06.410 --> 00:27:10.990 align:middle line:84% And in part, [INAUDIBLE] buffer stocks are even often modest, 00:27:10.990 --> 00:27:13.120 align:middle line:90% so people are not saving a lot. 00:27:13.120 --> 00:27:15.920 align:middle line:84% As I said, consumption often varies with income, 00:27:15.920 --> 00:27:18.100 align:middle line:84% including with predictable income. 00:27:18.100 --> 00:27:20.987 align:middle line:84% And then rainfall insurance increases fertilizer use, 00:27:20.987 --> 00:27:23.320 align:middle line:84% which seems to say, well, if that's really a big effect, 00:27:23.320 --> 00:27:24.823 align:middle line:90% people should purchase that. 00:27:24.823 --> 00:27:26.740 align:middle line:84% It's not consistent with people not purchasing 00:27:26.740 --> 00:27:28.940 align:middle line:90% that in the first place. 00:27:28.940 --> 00:27:32.113 align:middle line:84% Now, if we make people, instead, impatient-- 00:27:32.113 --> 00:27:34.030 align:middle line:84% and I think that's what you suggested before-- 00:27:34.030 --> 00:27:38.780 align:middle line:84% then we can, perhaps, match these types of predictions. 00:27:38.780 --> 00:27:42.220 align:middle line:84% So what we need is something that creates 00:27:42.220 --> 00:27:43.990 align:middle line:90% a high effective discount rate. 00:27:43.990 --> 00:27:45.580 align:middle line:84% One brief thing-- people have also 00:27:45.580 --> 00:27:48.460 align:middle line:84% tried to look at kin taxes or other types of taxation. 00:27:48.460 --> 00:27:51.637 align:middle line:84% It seems like while you can explain some of that puzzle, 00:27:51.637 --> 00:27:52.970 align:middle line:90% you don't get actually that far. 00:27:52.970 --> 00:27:56.110 align:middle line:84% So quantitatively, these numbers are not large enough 00:27:56.110 --> 00:27:57.760 align:middle line:90% to explain everything. 00:27:57.760 --> 00:28:00.920 align:middle line:90% But it's at least helpful. 00:28:00.920 --> 00:28:03.850 align:middle line:84% So let's then say, let's change the time preferences. 00:28:03.850 --> 00:28:08.140 align:middle line:84% Let's say people's discount factor is actually much lower. 00:28:08.140 --> 00:28:12.260 align:middle line:84% I said 0.96 but it could be just 0.5. 00:28:12.260 --> 00:28:13.205 align:middle line:90% What about that? 00:28:13.205 --> 00:28:19.827 align:middle line:90% 00:28:19.827 --> 00:28:21.660 align:middle line:84% Now, if you go back to our earlier equation, 00:28:21.660 --> 00:28:24.720 align:middle line:90% we say put in delta equals 2.5. 00:28:24.720 --> 00:28:25.480 align:middle line:90% Problem solved. 00:28:25.480 --> 00:28:26.313 align:middle line:90% Thank you very much. 00:28:26.313 --> 00:28:34.990 align:middle line:90% 00:28:34.990 --> 00:28:35.710 align:middle line:90% Yes. 00:28:35.710 --> 00:28:38.440 align:middle line:90% STUDENT: [INAUDIBLE]. 00:28:38.440 --> 00:28:42.790 align:middle line:84% FRANK SCHILBACH: Yeah, so what you'll get, essentially, is-- 00:28:42.790 --> 00:28:46.170 align:middle line:90% 00:28:46.170 --> 00:28:49.620 align:middle line:90% so what that implies is-- 00:28:49.620 --> 00:28:52.500 align:middle line:84% you might say, today versus next year, maybe 50% 00:28:52.500 --> 00:28:54.210 align:middle line:90% is fine in some cases. 00:28:54.210 --> 00:28:56.740 align:middle line:84% That's warranted if you have a really high return. 00:28:56.740 --> 00:28:59.310 align:middle line:84% But if you take that model seriously, and say what about 00:28:59.310 --> 00:29:01.638 align:middle line:84% in 10 years from now, what about in 20 years from now? 00:29:01.638 --> 00:29:03.930 align:middle line:84% What you get is essentially if you iterate this forward 00:29:03.930 --> 00:29:06.900 align:middle line:84% is to say you would be willing to give up $1 today 00:29:06.900 --> 00:29:09.550 align:middle line:90% for $1 billion in 30 years. 00:29:09.550 --> 00:29:13.980 align:middle line:84% And that's just clearly implausible or counterfactual. 00:29:13.980 --> 00:29:16.147 align:middle line:84% So no one would own land, or get an education, 00:29:16.147 --> 00:29:17.730 align:middle line:84% or brush their teeth, for that matter. 00:29:17.730 --> 00:29:20.700 align:middle line:84% It's just nobody would invest in anything 00:29:20.700 --> 00:29:22.480 align:middle line:90% that's happening in the future. 00:29:22.480 --> 00:29:25.710 align:middle line:84% So essentially, what this tells us-- and this is, 00:29:25.710 --> 00:29:28.320 align:middle line:84% I think a key point about behavioral economics-- 00:29:28.320 --> 00:29:31.680 align:middle line:84% you can't just fix one parameter and say, look, here's 00:29:31.680 --> 00:29:33.810 align:middle line:84% some behavior that we can fix in isolation, 00:29:33.810 --> 00:29:35.580 align:middle line:90% and say, OK, now we're done. 00:29:35.580 --> 00:29:41.940 align:middle line:84% One has to have a systematic and structured or disciplined 00:29:41.940 --> 00:29:44.550 align:middle line:84% approach to say, can we actually match 00:29:44.550 --> 00:29:45.960 align:middle line:90% several facts and the data? 00:29:45.960 --> 00:29:50.440 align:middle line:84% Does this seem to be broadly the right model overall? 00:29:50.440 --> 00:29:52.770 align:middle line:84% And so now, an alternative hypothesis 00:29:52.770 --> 00:29:54.270 align:middle line:84% that people have studied quite a bit 00:29:54.270 --> 00:29:59.230 align:middle line:84% is present bias, or present focus as people have called it, 00:29:59.230 --> 00:30:02.610 align:middle line:84% which essentially is the idea that you can separate 00:30:02.610 --> 00:30:05.760 align:middle line:84% the discount rate between today and tomorrow, 00:30:05.760 --> 00:30:08.820 align:middle line:84% or today and next year, from the long-run discount 00:30:08.820 --> 00:30:10.500 align:middle line:90% rate in the future. 00:30:10.500 --> 00:30:14.190 align:middle line:84% That's to say I might be really patient between 10 years 00:30:14.190 --> 00:30:15.850 align:middle line:90% from now and 11 years from now. 00:30:15.850 --> 00:30:17.670 align:middle line:84% But I might be really impatient when 00:30:17.670 --> 00:30:19.150 align:middle line:84% it comes to today versus tomorrow 00:30:19.150 --> 00:30:21.640 align:middle line:90% or today versus next week. 00:30:21.640 --> 00:30:24.630 align:middle line:84% And this is essentially what people have called 00:30:24.630 --> 00:30:26.880 align:middle line:90% quasi-hyperbolic discounting. 00:30:26.880 --> 00:30:30.720 align:middle line:84% You can see here how in exponential discounting-- 00:30:30.720 --> 00:30:36.390 align:middle line:84% this is the pink, or whatever color that is, line here. 00:30:36.390 --> 00:30:40.950 align:middle line:84% You can see essentially that the exponential discounting 00:30:40.950 --> 00:30:45.330 align:middle line:84% function-- it's very hard to create impatience 00:30:45.330 --> 00:30:47.910 align:middle line:84% in the short run and patience in the long run 00:30:47.910 --> 00:30:52.530 align:middle line:84% because the function just has does not have enough curvature. 00:30:52.530 --> 00:30:56.040 align:middle line:84% Instead, what present bias or quasi-hyperbolic discounting 00:30:56.040 --> 00:30:56.550 align:middle line:90% does-- 00:30:56.550 --> 00:30:58.800 align:middle line:84% so instead of having essentially delta being 00:30:58.800 --> 00:31:02.070 align:middle line:84% the short-term discount factor and the long-term discount 00:31:02.070 --> 00:31:06.510 align:middle line:84% factor, what quasi-hyperbolic discounting does is to say, 00:31:06.510 --> 00:31:08.730 align:middle line:84% let's add another parameter, which is beta. 00:31:08.730 --> 00:31:10.620 align:middle line:84% And beta is the short-term discount factor 00:31:10.620 --> 00:31:13.240 align:middle line:84% and then delta is the long-term discount factor, 00:31:13.240 --> 00:31:14.760 align:middle line:84% so essentially trying to separate 00:31:14.760 --> 00:31:16.980 align:middle line:84% short-term and long-term discounting. 00:31:16.980 --> 00:31:23.200 align:middle line:84% And beta, then, is measuring people's present bias. 00:31:23.200 --> 00:31:24.490 align:middle line:90% Any questions on this? 00:31:24.490 --> 00:31:31.220 align:middle line:90% 00:31:31.220 --> 00:31:33.200 align:middle line:84% So what I'm essentially doing is I'm saying, 00:31:33.200 --> 00:31:39.280 align:middle line:84% well, the Euler equation is mostly about the short run. 00:31:39.280 --> 00:31:41.020 align:middle line:84% And so now what I said before is, 00:31:41.020 --> 00:31:43.140 align:middle line:84% I need a really low delta to match this. 00:31:43.140 --> 00:31:44.820 align:middle line:84% But if I get a really low delta, I 00:31:44.820 --> 00:31:48.010 align:middle line:84% can't really match any long-run implications. 00:31:48.010 --> 00:31:51.660 align:middle line:84% So now what the beta delta model or the hyperbolic discounting 00:31:51.660 --> 00:31:54.150 align:middle line:84% model does, it essentially says, let's just 00:31:54.150 --> 00:31:56.820 align:middle line:84% have two parameters that try to match the world better. 00:31:56.820 --> 00:31:58.380 align:middle line:84% One is about short-run discounting, 00:31:58.380 --> 00:32:00.600 align:middle line:84% so people are really impatient in the short run. 00:32:00.600 --> 00:32:02.370 align:middle line:84% And one is about long-run discounting. 00:32:02.370 --> 00:32:03.870 align:middle line:84% And so now, we are able to at least, 00:32:03.870 --> 00:32:05.578 align:middle line:84% perhaps, better measure why people do not 00:32:05.578 --> 00:32:07.230 align:middle line:90% invest in the short run today. 00:32:07.230 --> 00:32:10.020 align:middle line:84% But perhaps, then, we don't get these counterfactual 00:32:10.020 --> 00:32:13.360 align:middle line:90% implications for the future. 00:32:13.360 --> 00:32:14.280 align:middle line:90% Any questions? 00:32:14.280 --> 00:32:21.730 align:middle line:90% 00:32:21.730 --> 00:32:25.750 align:middle line:84% And that, in fact, helps us address 00:32:25.750 --> 00:32:28.330 align:middle line:90% quite a few of these puzzles. 00:32:28.330 --> 00:32:30.040 align:middle line:84% Present-focused agents will rapidly 00:32:30.040 --> 00:32:32.650 align:middle line:84% spend down liquid assets, becoming effective liquidity 00:32:32.650 --> 00:32:33.650 align:middle line:90% constraints. 00:32:33.650 --> 00:32:35.800 align:middle line:84% So that's to say liquidity constraints are 00:32:35.800 --> 00:32:39.260 align:middle line:84% in some ways like an equilibrium outcome, or are endogenous, 00:32:39.260 --> 00:32:39.760 align:middle line:90% in a sense. 00:32:39.760 --> 00:32:42.070 align:middle line:90% It's not they fall from the sky. 00:32:42.070 --> 00:32:43.810 align:middle line:84% Of course, sometimes there shocks. 00:32:43.810 --> 00:32:46.750 align:middle line:84% But people can, of course, affect 00:32:46.750 --> 00:32:48.160 align:middle line:90% how much liquidity they have. 00:32:48.160 --> 00:32:50.110 align:middle line:84% And in fact, liquidity constraints 00:32:50.110 --> 00:32:53.530 align:middle line:84% are very much a sign of present bias or of, essentially, 00:32:53.530 --> 00:32:55.690 align:middle line:90% present focus preferences. 00:32:55.690 --> 00:32:59.090 align:middle line:84% And so the result is people will be liquidity constrained. 00:32:59.090 --> 00:33:00.970 align:middle line:84% And then might not be able, precisely 00:33:00.970 --> 00:33:02.530 align:middle line:84% because of liquidity constraints, 00:33:02.530 --> 00:33:05.380 align:middle line:84% to make higher-return investments. 00:33:05.380 --> 00:33:08.230 align:middle line:84% Now, then, people-- at the same time, they 00:33:08.230 --> 00:33:10.450 align:middle line:84% might build up a stock of illiquid assets 00:33:10.450 --> 00:33:12.400 align:middle line:84% that's paying off in the distant future 00:33:12.400 --> 00:33:14.290 align:middle line:84% to essentially constrain themselves 00:33:14.290 --> 00:33:15.207 align:middle line:90% from spending it down. 00:33:15.207 --> 00:33:17.498 align:middle line:84% So there's some interesting work on, for example, cows. 00:33:17.498 --> 00:33:18.970 align:middle line:84% Why do people hold cows as opposed 00:33:18.970 --> 00:33:20.380 align:middle line:90% to other types of assets? 00:33:20.380 --> 00:33:23.710 align:middle line:84% Well, it's perhaps because their indivisible and illiquid, 00:33:23.710 --> 00:33:25.450 align:middle line:90% in some ways, at least. 00:33:25.450 --> 00:33:28.570 align:middle line:84% Then, they will leave high rates of return investments 00:33:28.570 --> 00:33:30.880 align:middle line:84% on the table if they're essentially effectively 00:33:30.880 --> 00:33:32.105 align:middle line:90% to the liquidity constraints. 00:33:32.105 --> 00:33:34.480 align:middle line:84% If you don't have any money and I offer you a high return 00:33:34.480 --> 00:33:37.000 align:middle line:84% investment that you might be able to engage in, 00:33:37.000 --> 00:33:39.002 align:middle line:84% like say fertilizer, but you can't do 00:33:39.002 --> 00:33:40.210 align:middle line:90% that if you don't have money. 00:33:40.210 --> 00:33:44.440 align:middle line:84% And then you might leave out high returns to capital. 00:33:44.440 --> 00:33:47.200 align:middle line:84% And then, you also might not be able to smooth consumption, 00:33:47.200 --> 00:33:50.110 align:middle line:84% as consumption will move with income shocks, 00:33:50.110 --> 00:33:52.930 align:middle line:84% even with predictable income variation. 00:33:52.930 --> 00:33:55.420 align:middle line:84% Now, one key parameter here is, then, 00:33:55.420 --> 00:33:57.568 align:middle line:84% the sophistication of present bias actors that 00:33:57.568 --> 00:33:59.860 align:middle line:84% will determine the degree of procrastination and demand 00:33:59.860 --> 00:34:02.600 align:middle line:84% for commitment devices, which I'll talk about in a bit. 00:34:02.600 --> 00:34:06.080 align:middle line:84% So the model-- it looks like we can explain a lot of stuff. 00:34:06.080 --> 00:34:09.370 align:middle line:84% We can, in some ways, help explain the Euler equation 00:34:09.370 --> 00:34:09.969 align:middle line:90% puzzle. 00:34:09.969 --> 00:34:11.739 align:middle line:84% It turns out that's actually somewhat more 00:34:11.739 --> 00:34:16.070 align:middle line:90% complicated in many settings. 00:34:16.070 --> 00:34:18.489 align:middle line:84% But is the intuition, clear how present bias 00:34:18.489 --> 00:34:24.820 align:middle line:84% might help with dealing with the overall Euler equation puzzle? 00:34:24.820 --> 00:34:25.780 align:middle line:90% Any questions? 00:34:25.780 --> 00:34:29.600 align:middle line:90% 00:34:29.600 --> 00:34:33.440 align:middle line:84% All right, so now, another approach-- and one thing 00:34:33.440 --> 00:34:35.010 align:middle line:84% I've done now so far is to say, well, 00:34:35.010 --> 00:34:37.670 align:middle line:84% what about discounting where discounting might be different? 00:34:37.670 --> 00:34:39.957 align:middle line:84% Another approach might be to say, 00:34:39.957 --> 00:34:42.290 align:middle line:84% there might be other things in people's utility function 00:34:42.290 --> 00:34:45.350 align:middle line:84% that's different, for example, loss aversion. 00:34:45.350 --> 00:34:47.600 align:middle line:84% Well, how might loss aversion help? 00:34:47.600 --> 00:34:51.170 align:middle line:84% Loss aversion is the idea that people care more about losses 00:34:51.170 --> 00:34:53.270 align:middle line:90% than they care about gains. 00:34:53.270 --> 00:34:55.580 align:middle line:84% And so if I think about investing in capital 00:34:55.580 --> 00:34:58.130 align:middle line:84% or investing in certain assets, what I need to do 00:34:58.130 --> 00:35:00.020 align:middle line:84% is I need to take some money today, 00:35:00.020 --> 00:35:04.130 align:middle line:84% and I get some gains potentially in the future. 00:35:04.130 --> 00:35:06.690 align:middle line:84% But I might also get some losses. 00:35:06.690 --> 00:35:08.270 align:middle line:84% And so if I'm loss averse, A, I don't 00:35:08.270 --> 00:35:09.920 align:middle line:84% want to invest today because I actually 00:35:09.920 --> 00:35:12.680 align:middle line:84% have to give up some money today and there's potentially 00:35:12.680 --> 00:35:16.190 align:middle line:84% some losses in income or consumption in the present. 00:35:16.190 --> 00:35:18.860 align:middle line:84% But also, in the future, if there's risk, 00:35:18.860 --> 00:35:21.080 align:middle line:84% that might amplify the effect that risk 00:35:21.080 --> 00:35:22.550 align:middle line:90% has on people's choices. 00:35:22.550 --> 00:35:24.797 align:middle line:84% But I might be really worried about a bad harvest. 00:35:24.797 --> 00:35:27.380 align:middle line:84% I might be really worried about other bad shocks in the future 00:35:27.380 --> 00:35:30.830 align:middle line:84% because that might imply not only is my general consumption 00:35:30.830 --> 00:35:33.500 align:middle line:84% going down, but perhaps the investment in fertilizer 00:35:33.500 --> 00:35:34.460 align:middle line:90% is also going down. 00:35:34.460 --> 00:35:36.590 align:middle line:84% And then, my consumption in the future 00:35:36.590 --> 00:35:38.990 align:middle line:84% is going down plus my consumption in the present. 00:35:38.990 --> 00:35:43.440 align:middle line:84% And if I'm loss averse, I might put a lot of weight on that. 00:35:43.440 --> 00:35:45.890 align:middle line:84% So there's some interesting work, 00:35:45.890 --> 00:35:48.800 align:middle line:84% partially on lab experiments and lab evidence. 00:35:48.800 --> 00:35:51.860 align:middle line:84% People have been trying to take some of that 00:35:51.860 --> 00:35:55.910 align:middle line:84% into the field, including Michael Kremer and coauthors, 00:35:55.910 --> 00:36:00.020 align:middle line:84% as well as Kevin Carney and others, trying to understand, 00:36:00.020 --> 00:36:03.320 align:middle line:84% is it the case that people are not investing in certain assets 00:36:03.320 --> 00:36:05.600 align:middle line:90% if people are loss averse? 00:36:05.600 --> 00:36:07.280 align:middle line:84% And then that has some implications. 00:36:07.280 --> 00:36:09.380 align:middle line:84% For example, perhaps you can provide 00:36:09.380 --> 00:36:11.540 align:middle line:84% insurance or other types of ways of helping 00:36:11.540 --> 00:36:16.590 align:middle line:90% people mitigate shocks. 00:36:16.590 --> 00:36:21.860 align:middle line:84% And that might lead to at least partially closing this gap. 00:36:21.860 --> 00:36:25.630 align:middle line:90% 00:36:25.630 --> 00:36:31.800 align:middle line:84% And so that's a broad overview of the topics 00:36:31.800 --> 00:36:35.110 align:middle line:90% of the broad puzzle as a whole. 00:36:35.110 --> 00:36:37.260 align:middle line:84% Now you can see how we can invest. 00:36:37.260 --> 00:36:40.140 align:middle line:84% We can think about all types of topics and development 00:36:40.140 --> 00:36:43.860 align:middle line:84% economics as an application of this Euler puzzle. 00:36:43.860 --> 00:36:47.220 align:middle line:84% You might say, health behavior or health investments 00:36:47.220 --> 00:36:51.155 align:middle line:84% could be really high investments in people's future. 00:36:51.155 --> 00:36:52.530 align:middle line:84% These are high-return investments 00:36:52.530 --> 00:36:54.300 align:middle line:84% because if you fall sick, you might 00:36:54.300 --> 00:36:56.760 align:middle line:84% have large income losses and the like, 00:36:56.760 --> 00:36:59.040 align:middle line:84% which you really would like to avoid. 00:36:59.040 --> 00:37:02.310 align:middle line:84% When you think about savings, people could save a lot, 00:37:02.310 --> 00:37:05.700 align:middle line:84% and often have high returns to savings in various ways, 00:37:05.700 --> 00:37:06.610 align:middle line:90% or perhaps not. 00:37:06.610 --> 00:37:07.950 align:middle line:90% And the question is, why not? 00:37:07.950 --> 00:37:10.950 align:middle line:84% People could invest in insurance. 00:37:10.950 --> 00:37:14.525 align:middle line:84% And insurance would really have high returns because often, 00:37:14.525 --> 00:37:15.900 align:middle line:84% when they're not able to mitigate 00:37:15.900 --> 00:37:18.060 align:middle line:90% shocks, that's really costly. 00:37:18.060 --> 00:37:20.520 align:middle line:84% In particular to technology adoption, which surely we'll 00:37:20.520 --> 00:37:23.790 align:middle line:84% talk about, why are people not investing in fertilizer? 00:37:23.790 --> 00:37:27.233 align:middle line:84% Why are people not investing in other types of technologies? 00:37:27.233 --> 00:37:29.400 align:middle line:84% Does this have to do with whether people are present 00:37:29.400 --> 00:37:29.900 align:middle line:90% biased? 00:37:29.900 --> 00:37:33.630 align:middle line:84% Does it have to do with people are loss averse or the like? 00:37:33.630 --> 00:37:35.712 align:middle line:84% Now, then, there's other types of explanations, 00:37:35.712 --> 00:37:37.170 align:middle line:84% which we'll talk about a little bit 00:37:37.170 --> 00:37:39.017 align:middle line:84% when it comes to health, which is beliefs. 00:37:39.017 --> 00:37:41.100 align:middle line:84% Maybe people don't know that the returns are high. 00:37:41.100 --> 00:37:43.680 align:middle line:84% Or people might have trouble learning from their neighbors 00:37:43.680 --> 00:37:44.340 align:middle line:90% or the like. 00:37:44.340 --> 00:37:46.710 align:middle line:84% But broadly, I think the idea that there's 00:37:46.710 --> 00:37:49.170 align:middle line:84% high-return investments that you could take advantage of, 00:37:49.170 --> 00:37:51.570 align:middle line:84% and people are not doing that as much as we perhaps 00:37:51.570 --> 00:37:53.880 align:middle line:84% would like them to or that would perhaps be good 00:37:53.880 --> 00:37:57.710 align:middle line:84% for them, trying to understand this apparent decision of here 00:37:57.710 --> 00:37:59.980 align:middle line:84% are high returns, should you invest or not? 00:37:59.980 --> 00:38:04.680 align:middle line:84% And if not, if it looks like people should be investing 00:38:04.680 --> 00:38:07.240 align:middle line:84% and people are not doing it, why is that? 00:38:07.240 --> 00:38:10.800 align:middle line:84% Or what can we do to help understand these issues? 00:38:10.800 --> 00:38:13.440 align:middle line:84% And that's not to say that people are making mistakes 00:38:13.440 --> 00:38:14.220 align:middle line:90% per se. 00:38:14.220 --> 00:38:16.447 align:middle line:84% It might well be that their preferences [INAUDIBLE] 00:38:16.447 --> 00:38:18.780 align:middle line:84% in certain ways, that it's very reasonable not to invest 00:38:18.780 --> 00:38:19.697 align:middle line:90% if you're loss averse. 00:38:19.697 --> 00:38:22.140 align:middle line:84% Because you just worry that if you incur losses, 00:38:22.140 --> 00:38:23.800 align:middle line:90% that's really costly for you. 00:38:23.800 --> 00:38:25.713 align:middle line:84% So we're trying to just better understand. 00:38:25.713 --> 00:38:27.630 align:middle line:84% And then based on that improved understanding, 00:38:27.630 --> 00:38:31.470 align:middle line:84% trying to see can we improve policies to try to understand 00:38:31.470 --> 00:38:35.032 align:middle line:90% people's decisions. 00:38:35.032 --> 00:38:36.115 align:middle line:90% Any questions or comments? 00:38:36.115 --> 00:38:39.820 align:middle line:90% 00:38:39.820 --> 00:38:40.320 align:middle line:90% Yeah. 00:38:40.320 --> 00:38:41.739 align:middle line:90% STUDENT: [INAUDIBLE]. 00:38:41.739 --> 00:38:45.050 align:middle line:90% 00:38:45.050 --> 00:38:49.080 align:middle line:84% So maybe it's not the [INAUDIBLE] preferences. 00:38:49.080 --> 00:38:51.690 align:middle line:84% Maybe it's the fact that they need to eat two times a day. 00:38:51.690 --> 00:38:55.910 align:middle line:90% So it's not how [INAUDIBLE]. 00:38:55.910 --> 00:38:58.900 align:middle line:84% So in the short run, the [INAUDIBLE] 00:38:58.900 --> 00:39:01.570 align:middle line:84% want money because they have less money 00:39:01.570 --> 00:39:03.070 align:middle line:90% and they have to eat. 00:39:03.070 --> 00:39:04.660 align:middle line:84% FRANK SCHILBACH: Yeah, so why is it 00:39:04.660 --> 00:39:06.160 align:middle line:84% that people are not holding assets? 00:39:06.160 --> 00:39:07.243 align:middle line:90% Why are people not saving? 00:39:07.243 --> 00:39:09.550 align:middle line:84% And Esther will talk about savings, 00:39:09.550 --> 00:39:11.410 align:middle line:90% I think, in a few lectures. 00:39:11.410 --> 00:39:17.140 align:middle line:84% And so one important issue is that even if you're very poor, 00:39:17.140 --> 00:39:20.620 align:middle line:84% the Euler equation should still hold. 00:39:20.620 --> 00:39:22.430 align:middle line:90% What does that say? 00:39:22.430 --> 00:39:25.870 align:middle line:84% Well, that's saying, even if you're very poor, 00:39:25.870 --> 00:39:28.630 align:middle line:84% you should think about should I eat today versus should 00:39:28.630 --> 00:39:30.880 align:middle line:90% I eat tomorrow. 00:39:30.880 --> 00:39:35.020 align:middle line:84% Now, you might say, people are having so little money 00:39:35.020 --> 00:39:40.150 align:middle line:84% that they're not able to even save anything. 00:39:40.150 --> 00:39:42.670 align:middle line:84% But in particular, actually, the paper 00:39:42.670 --> 00:39:48.250 align:middle line:84% by Colin, Mullainathan, and Roth is making these calculations 00:39:48.250 --> 00:39:50.830 align:middle line:84% for food vendors that say, if you 00:39:50.830 --> 00:39:54.490 align:middle line:84% are paying 5% of your inventory as a rate of-- 00:39:54.490 --> 00:40:00.670 align:middle line:84% if you're paying 50 rupees every day as interest, 00:40:00.670 --> 00:40:02.620 align:middle line:84% that implies, essentially, that your rate 00:40:02.620 --> 00:40:04.360 align:middle line:84% of return of any additional savings, 00:40:04.360 --> 00:40:07.763 align:middle line:84% if it's just 1 rupee every day, would be extremely high. 00:40:07.763 --> 00:40:09.430 align:middle line:84% So sometimes you still have to trade off 00:40:09.430 --> 00:40:10.840 align:middle line:90% today versus tomorrow. 00:40:10.840 --> 00:40:13.900 align:middle line:84% And in many situations, in particular when you talk about 00:40:13.900 --> 00:40:17.230 align:middle line:84% nutrition, it's not that people are-- in many places, 00:40:17.230 --> 00:40:18.280 align:middle line:90% for example in India-- 00:40:18.280 --> 00:40:20.710 align:middle line:84% it's not that people are starving per se, in the sense 00:40:20.710 --> 00:40:24.250 align:middle line:84% that they can't afford enough consumption. 00:40:24.250 --> 00:40:26.785 align:middle line:84% And they often, for example, drink a lot of tea. 00:40:26.785 --> 00:40:28.930 align:middle line:84% You might say, what if you had one 00:40:28.930 --> 00:40:33.490 align:middle line:84% tea lasts for the next 30 days, that costs five or 10 rupees? 00:40:33.490 --> 00:40:37.420 align:middle line:84% You would be able to save a lot going forward 00:40:37.420 --> 00:40:40.090 align:middle line:84% because you are essentially avoiding the high interest 00:40:40.090 --> 00:40:40.947 align:middle line:90% that you're paying. 00:40:40.947 --> 00:40:43.030 align:middle line:84% So essentially, even if you're at very low levels, 00:40:43.030 --> 00:40:44.920 align:middle line:84% you still trade off consumption today 00:40:44.920 --> 00:40:48.080 align:middle line:84% versus tomorrow, even very small amounts, 00:40:48.080 --> 00:40:49.660 align:middle line:84% especially for divisible investments, 00:40:49.660 --> 00:40:54.130 align:middle line:90% you should be able to make. 00:40:54.130 --> 00:40:56.440 align:middle line:84% And then if the rate of return is really high, 00:40:56.440 --> 00:41:02.770 align:middle line:84% you would be able to increase your consumption quickly 00:41:02.770 --> 00:41:03.493 align:middle line:90% over time. 00:41:03.493 --> 00:41:05.410 align:middle line:84% The same is true, for example, for fertilizer. 00:41:05.410 --> 00:41:08.050 align:middle line:84% Fertilizer is divisible, so fertilizer doesn't really 00:41:08.050 --> 00:41:10.160 align:middle line:90% fit the poverty trap model. 00:41:10.160 --> 00:41:12.400 align:middle line:84% So you should be able to invest very small amounts, 00:41:12.400 --> 00:41:16.630 align:middle line:84% and then increasingly increase your consumption and output 00:41:16.630 --> 00:41:19.290 align:middle line:90% over time. 00:41:19.290 --> 00:41:20.240 align:middle line:90% Any other questions? 00:41:20.240 --> 00:41:24.790 align:middle line:90% 00:41:24.790 --> 00:41:29.050 align:middle line:84% OK, so then let's talk about one such application. 00:41:29.050 --> 00:41:32.700 align:middle line:84% So one broad fact-- and again, very surely you'll talk about 00:41:32.700 --> 00:41:36.690 align:middle line:90% this later in the classes-- 00:41:36.690 --> 00:41:40.530 align:middle line:84% that people are under-investing in preventative health. 00:41:40.530 --> 00:41:44.340 align:middle line:84% And so you can think of this as a really high return investment 00:41:44.340 --> 00:41:48.120 align:middle line:84% opportunity because if your kids get sick, 00:41:48.120 --> 00:41:51.630 align:middle line:84% if you don't invest in, say, deworming, if you, 00:41:51.630 --> 00:41:53.760 align:middle line:84% yourself get sick, you forgo a lot of income. 00:41:53.760 --> 00:41:55.740 align:middle line:84% And that's a very high return investment. 00:41:55.740 --> 00:41:57.540 align:middle line:84% For example, Michael Kremer and co-authors 00:41:57.540 --> 00:42:00.960 align:middle line:84% have shown that the rate of return on deworming children 00:42:00.960 --> 00:42:04.110 align:middle line:90% is extremely high. 00:42:04.110 --> 00:42:07.080 align:middle line:84% But for many types of preventative investments-- 00:42:07.080 --> 00:42:09.690 align:middle line:84% vaccinations, deworming, bed nets, water treatments, 00:42:09.690 --> 00:42:14.700 align:middle line:84% hypertension, et cetera, we observe relatively low rates 00:42:14.700 --> 00:42:16.230 align:middle line:90% of investments. 00:42:16.230 --> 00:42:18.270 align:middle line:84% And there's different types of stylized facts 00:42:18.270 --> 00:42:19.920 align:middle line:90% that we have seen. 00:42:19.920 --> 00:42:23.140 align:middle line:84% They have low willingness to pay for preventative health. 00:42:23.140 --> 00:42:27.080 align:middle line:84% There's high expenditures for treatment of acute conditions. 00:42:27.080 --> 00:42:29.500 align:middle line:84% So it's not the case that people spend nothing on health 00:42:29.500 --> 00:42:31.090 align:middle line:90% or can't afford any of that. 00:42:31.090 --> 00:42:34.010 align:middle line:84% But as I say, people spend a lot on acute conditions, 00:42:34.010 --> 00:42:37.210 align:middle line:84% but much less on preventative prevention of those. 00:42:37.210 --> 00:42:40.120 align:middle line:84% And in part because people spend so much on acute conditions, 00:42:40.120 --> 00:42:42.760 align:middle line:84% there would be a really high return of prevention 00:42:42.760 --> 00:42:44.410 align:middle line:84% because you could, perhaps, avoid 00:42:44.410 --> 00:42:47.170 align:middle line:90% some of these acute conditions. 00:42:47.170 --> 00:42:49.000 align:middle line:84% And then third, there's high sensitivity 00:42:49.000 --> 00:42:51.890 align:middle line:84% of health investments to price and convenience. 00:42:51.890 --> 00:42:54.610 align:middle line:84% So if you just change the prices a little bit, 00:42:54.610 --> 00:42:58.310 align:middle line:84% people's investments seem to change quite a lot. 00:42:58.310 --> 00:43:00.910 align:middle line:84% And so the [INAUDIBLE] from a recent 00:43:00.910 --> 00:43:04.420 align:middle line:84% also handbook chapter by Pascaline Dupas and Ed Miguel, 00:43:04.420 --> 00:43:06.820 align:middle line:84% an overview where we see people's demand 00:43:06.820 --> 00:43:09.500 align:middle line:84% for different health investments at different prices. 00:43:09.500 --> 00:43:12.730 align:middle line:84% And what you see is, A, the fraction of people 00:43:12.730 --> 00:43:16.300 align:middle line:84% who demand these investments at positive prices 00:43:16.300 --> 00:43:20.446 align:middle line:84% tends to be often quite low compared to what you think, 00:43:20.446 --> 00:43:21.640 align:middle line:90% given the returns. 00:43:21.640 --> 00:43:25.030 align:middle line:84% And second, the price sensitivity 00:43:25.030 --> 00:43:29.000 align:middle line:84% seems to be high, as in the slopes are very steep. 00:43:29.000 --> 00:43:33.730 align:middle line:84% So once you go from, say $0.50 to $1 to $2, 00:43:33.730 --> 00:43:38.510 align:middle line:84% the investments tend to fall a lot. 00:43:38.510 --> 00:43:39.810 align:middle line:90% Yeah. 00:43:39.810 --> 00:43:43.680 align:middle line:84% STUDENT: Of those [INAUDIBLE] facts, how unique 00:43:43.680 --> 00:43:45.420 align:middle line:90% are the developing countries? 00:43:45.420 --> 00:43:49.060 align:middle line:84% Because all 3.6 [INAUDIBLE] also is appropriate for the United 00:43:49.060 --> 00:43:49.560 align:middle line:90% States. 00:43:49.560 --> 00:43:51.518 align:middle line:84% FRANK SCHILBACH: Yeah, that's a great question. 00:43:51.518 --> 00:43:54.660 align:middle line:84% So the question is, is this specific 00:43:54.660 --> 00:43:57.930 align:middle line:84% to developing countries or is it true 00:43:57.930 --> 00:43:59.260 align:middle line:90% anywhere else in the world? 00:43:59.260 --> 00:44:01.593 align:middle line:84% And I think that's an important general question about-- 00:44:01.593 --> 00:44:03.927 align:middle line:84% and maybe I should have started with that in some ways-- 00:44:03.927 --> 00:44:06.570 align:middle line:84% to say that's precisely what I was trying to say earlier about 00:44:06.570 --> 00:44:11.280 align:middle line:84% we're trying to study universal factors, 00:44:11.280 --> 00:44:14.700 align:middle line:84% the psychological factors that affect people's behavior. 00:44:14.700 --> 00:44:19.000 align:middle line:84% They might be more pronounced in certain ways in poor settings. 00:44:19.000 --> 00:44:21.570 align:middle line:84% For example, if you don't have, say, health insurance 00:44:21.570 --> 00:44:24.450 align:middle line:84% or other types of medical treatments 00:44:24.450 --> 00:44:27.270 align:middle line:84% that are free for you, then maybe your individual choices 00:44:27.270 --> 00:44:28.322 align:middle line:90% become more important. 00:44:28.322 --> 00:44:30.030 align:middle line:84% I think you're exactly right, that people 00:44:30.030 --> 00:44:32.250 align:middle line:84% have low willingness to pay for preventative health 00:44:32.250 --> 00:44:33.330 align:middle line:90% in many settings. 00:44:33.330 --> 00:44:36.540 align:middle line:84% People often spend a lot on acute conditions. 00:44:36.540 --> 00:44:38.190 align:middle line:84% The sensitivity of health investments 00:44:38.190 --> 00:44:40.260 align:middle line:84% of price and convenience might be more 00:44:40.260 --> 00:44:42.420 align:middle line:90% specific to poorer settings. 00:44:42.420 --> 00:44:43.880 align:middle line:90% But it's a very good point. 00:44:43.880 --> 00:44:45.630 align:middle line:84% It's by no means that we're trying to say 00:44:45.630 --> 00:44:48.840 align:middle line:90% that this is only the case-- 00:44:48.840 --> 00:44:50.640 align:middle line:84% this is unique to, say, poor settings. 00:44:50.640 --> 00:44:55.650 align:middle line:84% But rather, we're trying to say that this is an important issue 00:44:55.650 --> 00:44:56.582 align:middle line:90% that we observed. 00:44:56.582 --> 00:44:59.040 align:middle line:84% That might also be true-- and in this case, it's the case-- 00:44:59.040 --> 00:45:00.930 align:middle line:84% that might also be true in rich settings, 00:45:00.930 --> 00:45:04.980 align:middle line:84% but using, then, insights from behavioral economics 00:45:04.980 --> 00:45:07.750 align:middle line:90% might be able to help with that. 00:45:07.750 --> 00:45:09.500 align:middle line:90% Does that make sense? 00:45:09.500 --> 00:45:10.328 align:middle line:90% Yeah. 00:45:10.328 --> 00:45:12.370 align:middle line:84% STUDENT: This I think applies both here, and also 00:45:12.370 --> 00:45:16.330 align:middle line:84% to the general proposition of [INAUDIBLE].. 00:45:16.330 --> 00:45:19.570 align:middle line:84% How do you separate a present bias story 00:45:19.570 --> 00:45:25.500 align:middle line:84% for [INAUDIBLE] too much from a story 00:45:25.500 --> 00:45:28.670 align:middle line:84% where I don't have much information [INAUDIBLE] 00:45:28.670 --> 00:45:29.630 align:middle line:90% eating unhealthily. 00:45:29.630 --> 00:45:31.713 align:middle line:84% I just don't, because everyone around me is young. 00:45:31.713 --> 00:45:33.373 align:middle line:84% I don't know how bad diabetes might be. 00:45:33.373 --> 00:45:34.790 align:middle line:84% So I continue to eat like a beast. 00:45:34.790 --> 00:45:37.080 align:middle line:84% FRANK SCHILBACH: Yeah, that's a good question. 00:45:37.080 --> 00:45:40.250 align:middle line:84% So the question is, how do we separate preferences 00:45:40.250 --> 00:45:41.630 align:middle line:90% from information? 00:45:41.630 --> 00:45:46.050 align:middle line:84% It might well be that you observe these types of choices, 00:45:46.050 --> 00:45:48.380 align:middle line:84% it looks like people don't care about the future, 00:45:48.380 --> 00:45:50.300 align:middle line:84% but in reality, it might just be that people 00:45:50.300 --> 00:45:52.370 align:middle line:84% don't know that the returns are really high, 00:45:52.370 --> 00:45:54.470 align:middle line:84% and therefore, they don't invest. 00:45:54.470 --> 00:45:57.950 align:middle line:84% But if you only told them, they would invest. 00:45:57.950 --> 00:46:01.100 align:middle line:84% And I'll actually get back to that in a second. 00:46:01.100 --> 00:46:03.470 align:middle line:84% In some ways, I think that's partially, I think, 00:46:03.470 --> 00:46:05.300 align:middle line:84% where the literature in part is going. 00:46:05.300 --> 00:46:08.000 align:middle line:84% Many people have studied, to start with, 00:46:08.000 --> 00:46:12.680 align:middle line:84% preference-based explanations and studied last beliefs. 00:46:12.680 --> 00:46:15.200 align:middle line:84% But I think people more and more are increasingly 00:46:15.200 --> 00:46:16.880 align:middle line:90% interested in beliefs. 00:46:16.880 --> 00:46:20.060 align:middle line:84% Bear with me for a couple of slides and I'll get to this. 00:46:20.060 --> 00:46:25.010 align:middle line:84% So overall, here I'm saying [INAUDIBLE] the information 00:46:25.010 --> 00:46:27.050 align:middle line:90% part, many people do-- 00:46:27.050 --> 00:46:28.898 align:middle line:84% so one answer to your question is, 00:46:28.898 --> 00:46:30.440 align:middle line:84% many people may say that they believe 00:46:30.440 --> 00:46:32.270 align:middle line:90% in a biomedical explanation. 00:46:32.270 --> 00:46:34.610 align:middle line:84% So it's not clear that people necessarily 00:46:34.610 --> 00:46:38.390 align:middle line:84% think that, say, deworming or bed nets or the like 00:46:38.390 --> 00:46:39.050 align:middle line:90% don't work. 00:46:39.050 --> 00:46:42.510 align:middle line:84% But I'll get back to that in a second. 00:46:42.510 --> 00:46:45.980 align:middle line:84% Now, you might say, there might be some-- 00:46:45.980 --> 00:46:47.810 align:middle line:84% so there's one thing that I need to say, 00:46:47.810 --> 00:46:51.170 align:middle line:84% is to say that small and time-limited incentives 00:46:51.170 --> 00:46:54.470 align:middle line:84% for vaccination or other types of incentives 00:46:54.470 --> 00:46:57.110 align:middle line:84% for, say, collecting HIV tests can lead 00:46:57.110 --> 00:46:58.830 align:middle line:90% to large changes in behavior. 00:46:58.830 --> 00:47:01.530 align:middle line:84% So not only is it the case that people are price sensitive. 00:47:01.530 --> 00:47:04.580 align:middle line:84% If you vary prices at regular rates and give some discounts, 00:47:04.580 --> 00:47:07.670 align:middle line:84% people are price sensitive, but it's also 00:47:07.670 --> 00:47:09.570 align:middle line:84% once you provide incentives, when you say, 00:47:09.570 --> 00:47:12.440 align:middle line:84% OK, I'll give you a bag of lentils for vaccination, 00:47:12.440 --> 00:47:17.390 align:middle line:84% people react a lot to those types of incentives. 00:47:17.390 --> 00:47:22.320 align:middle line:84% Now, you might say, is there a knife-edge balance 00:47:22.320 --> 00:47:25.110 align:middle line:84% between costs and benefits of prevention? 00:47:25.110 --> 00:47:27.630 align:middle line:84% Maybe people are just all really indifferent. 00:47:27.630 --> 00:47:30.450 align:middle line:84% It could be lots of people are indifferent 00:47:30.450 --> 00:47:32.580 align:middle line:90% at this specific price. 00:47:32.580 --> 00:47:34.410 align:middle line:84% That seems unlikely because there's lots 00:47:34.410 --> 00:47:36.140 align:middle line:90% of variation across people. 00:47:36.140 --> 00:47:37.890 align:middle line:84% And that seems to be quite unlikely that's 00:47:37.890 --> 00:47:44.200 align:middle line:84% the case for everybody at the same time. 00:47:44.200 --> 00:47:46.920 align:middle line:84% So now, you might say, maybe present bias can help. 00:47:46.920 --> 00:47:49.920 align:middle line:84% And this is the traditional preference-based explanation. 00:47:49.920 --> 00:47:52.510 align:middle line:84% Maybe we can make some progress with that. 00:47:52.510 --> 00:47:54.870 align:middle line:84% And there's two ways in which present bias might 00:47:54.870 --> 00:47:56.010 align:middle line:90% generate under-investment. 00:47:56.010 --> 00:47:57.530 align:middle line:90% One might be procrastination. 00:47:57.530 --> 00:48:00.030 align:middle line:84% This is to say, oh, I'm going to go to the doctor next week. 00:48:00.030 --> 00:48:01.710 align:middle line:84% I'm going to go to the dentist next year, 00:48:01.710 --> 00:48:02.710 align:middle line:90% and so on, and so forth. 00:48:02.710 --> 00:48:05.375 align:middle line:84% And then next year comes, and you never actually do it. 00:48:05.375 --> 00:48:06.750 align:middle line:84% The second explanation might be-- 00:48:06.750 --> 00:48:08.140 align:middle line:84% and this is what we discussed before-- 00:48:08.140 --> 00:48:10.320 align:middle line:84% which is liquidity constraints, which is to say, 00:48:10.320 --> 00:48:11.862 align:middle line:84% I just don't have any money right now 00:48:11.862 --> 00:48:13.390 align:middle line:90% and I can't invest right now. 00:48:13.390 --> 00:48:16.230 align:middle line:84% And the reason for that is because I was present biased 00:48:16.230 --> 00:48:18.570 align:middle line:84% yesterday or two days ago, and I spent too much money 00:48:18.570 --> 00:48:19.690 align:middle line:90% on other things. 00:48:19.690 --> 00:48:22.500 align:middle line:90% And now, I can't invest. 00:48:22.500 --> 00:48:24.090 align:middle line:84% And that also explains, potentially, 00:48:24.090 --> 00:48:25.818 align:middle line:90% the high price sensitivity. 00:48:25.818 --> 00:48:28.360 align:middle line:84% If you give it to me for free, yes, sure, I'm going to do it. 00:48:28.360 --> 00:48:29.777 align:middle line:84% But once you charge me some money, 00:48:29.777 --> 00:48:32.700 align:middle line:84% I might not be able to do it because I just 00:48:32.700 --> 00:48:36.720 align:middle line:90% don't have any money available. 00:48:36.720 --> 00:48:39.330 align:middle line:90% So then, it turns out-- 00:48:39.330 --> 00:48:43.600 align:middle line:84% so on the one hand, we can explain quite a few things. 00:48:43.600 --> 00:48:45.330 align:middle line:84% So present bias and procrastination 00:48:45.330 --> 00:48:49.230 align:middle line:84% can explain essentially, there's immediate utility costs 00:48:49.230 --> 00:48:50.760 align:middle line:90% of many different investments. 00:48:50.760 --> 00:48:53.070 align:middle line:84% For example, hassle costs of going to the doctor, 00:48:53.070 --> 00:48:55.440 align:middle line:84% walking to a faraway water source, 00:48:55.440 --> 00:48:59.760 align:middle line:84% using somewhat inconvenient chlorine solution, 00:48:59.760 --> 00:49:02.670 align:middle line:84% changing your diet, learning painful news 00:49:02.670 --> 00:49:04.740 align:middle line:84% about your health status, and so on. 00:49:04.740 --> 00:49:08.160 align:middle line:84% All of these kinds of investment have some immediate cost. 00:49:08.160 --> 00:49:10.530 align:middle line:84% You have to invest in it in some ways 00:49:10.530 --> 00:49:12.990 align:middle line:84% and incur some costs in the short run that 00:49:12.990 --> 00:49:15.640 align:middle line:84% leads to potential benefits in the future. 00:49:15.640 --> 00:49:19.860 align:middle line:84% Now, you might say, then people might just procrastinate. 00:49:19.860 --> 00:49:22.560 align:middle line:84% Notice that that actually already requires 00:49:22.560 --> 00:49:25.770 align:middle line:84% some assumption about people's beliefs. 00:49:25.770 --> 00:49:28.210 align:middle line:84% That is to say, not only do I have to say next year, 00:49:28.210 --> 00:49:29.958 align:middle line:84% I'm going to go to the doctor, but I also 00:49:29.958 --> 00:49:31.500 align:middle line:84% need to think that I'm actually going 00:49:31.500 --> 00:49:34.080 align:middle line:84% to-- so if I am fully sophisticated, 00:49:34.080 --> 00:49:38.790 align:middle line:84% and I know that next year I'm also going to procrastinate, 00:49:38.790 --> 00:49:41.460 align:middle line:84% models have a hard time explaining 00:49:41.460 --> 00:49:43.170 align:middle line:90% sustained procrastination. 00:49:43.170 --> 00:49:47.370 align:middle line:84% That is to say, I cannot just fool myself every-- 00:49:47.370 --> 00:49:50.220 align:middle line:84% if I'm sophisticated, if I have rational expectations-- 00:49:50.220 --> 00:49:52.380 align:middle line:84% I cannot just fool myself over a period and say, 00:49:52.380 --> 00:49:53.920 align:middle line:90% next time I'm going to do it. 00:49:53.920 --> 00:49:56.935 align:middle line:84% So it requires already some naivete or some wrong beliefs, 00:49:56.935 --> 00:49:58.560 align:middle line:84% in some ways, about my future behavior. 00:49:58.560 --> 00:50:02.400 align:middle line:84% Otherwise, you just will not get sustained procrastination. 00:50:02.400 --> 00:50:08.070 align:middle line:84% Now, that can explain things like time-limited incentives 00:50:08.070 --> 00:50:10.380 align:middle line:84% or the effects of those, the effects of reducing 00:50:10.380 --> 00:50:14.950 align:middle line:84% hassle costs, for example, water dispensers, or the like. 00:50:14.950 --> 00:50:16.830 align:middle line:84% Notice that you would not procrastinate 00:50:16.830 --> 00:50:19.200 align:middle line:84% an acute condition, since the benefits 00:50:19.200 --> 00:50:20.160 align:middle line:90% are actually immediate. 00:50:20.160 --> 00:50:21.618 align:middle line:84% So that actually fits in some ways, 00:50:21.618 --> 00:50:24.360 align:middle line:84% as well, since people seem to actually make those. 00:50:24.360 --> 00:50:27.480 align:middle line:84% Now, then, as I said, present bias 00:50:27.480 --> 00:50:29.425 align:middle line:84% can lead to liquidity constraints. 00:50:29.425 --> 00:50:31.050 align:middle line:84% And that's actually an important point. 00:50:31.050 --> 00:50:33.690 align:middle line:84% It's not to say that liquidity constraints and present bias 00:50:33.690 --> 00:50:35.460 align:middle line:84% are different types of explanations. 00:50:35.460 --> 00:50:37.440 align:middle line:84% It's precisely people are liquidity constrained 00:50:37.440 --> 00:50:38.860 align:middle line:90% because they're present biased. 00:50:38.860 --> 00:50:40.277 align:middle line:84% And then the question is, perhaps, 00:50:40.277 --> 00:50:42.570 align:middle line:84% how can you help people save more differently, 00:50:42.570 --> 00:50:47.400 align:middle line:84% or invest in insurance, or the like. 00:50:47.400 --> 00:50:49.533 align:middle line:84% Now, one thing I want to point out 00:50:49.533 --> 00:50:51.450 align:middle line:84% is that these present-biased explanations also 00:50:51.450 --> 00:50:52.683 align:middle line:90% have lots of limits. 00:50:52.683 --> 00:50:54.850 align:middle line:84% So it's not like you can actually explain that much. 00:50:54.850 --> 00:50:58.740 align:middle line:84% And this is what the question that was just came up 00:50:58.740 --> 00:51:01.920 align:middle line:84% was about, as well, which is to say 00:51:01.920 --> 00:51:05.040 align:middle line:84% you might be able to explain certain behaviors where there 00:51:05.040 --> 00:51:06.090 align:middle line:90% is no deadline. 00:51:06.090 --> 00:51:08.742 align:middle line:84% If you ask me go to the dentist, and there's no deadline, 00:51:08.742 --> 00:51:10.950 align:middle line:84% I can go to the dentist next year, and the next year, 00:51:10.950 --> 00:51:13.080 align:middle line:84% and so on, I can procrastinate forever, 00:51:13.080 --> 00:51:14.820 align:middle line:90% if I'm sufficiently naive. 00:51:14.820 --> 00:51:16.528 align:middle line:84% That's to say, I'm really sure next year, 00:51:16.528 --> 00:51:17.820 align:middle line:90% I'm going to go to the dentist. 00:51:17.820 --> 00:51:19.920 align:middle line:84% And next year, I'm, like, actually, not this year. 00:51:19.920 --> 00:51:21.810 align:middle line:84% I'm going to procrastinate for another year, 00:51:21.810 --> 00:51:23.850 align:middle line:84% and for another year, and for another year. 00:51:23.850 --> 00:51:27.330 align:middle line:84% You can get really large welfare losses, potentially, 00:51:27.330 --> 00:51:29.957 align:middle line:84% if there's no deadline or no fixed point 00:51:29.957 --> 00:51:32.040 align:middle line:84% where I say, OK Frank, today you go to the dentist 00:51:32.040 --> 00:51:33.687 align:middle line:90% or never again in your life. 00:51:33.687 --> 00:51:35.520 align:middle line:84% Even if I'm present biased, I would probably 00:51:35.520 --> 00:51:39.870 align:middle line:84% go to the dentist because I know that the costs of not doing it 00:51:39.870 --> 00:51:40.830 align:middle line:90% are very large. 00:51:40.830 --> 00:51:43.420 align:middle line:84% But if I can fool myself every year to say, 00:51:43.420 --> 00:51:45.300 align:middle line:84% I'm going to do it in the future, that might 00:51:45.300 --> 00:51:48.360 align:middle line:90% lead to large welfare losses. 00:51:48.360 --> 00:51:51.000 align:middle line:84% On the other hand, when you see one-shot investments 00:51:51.000 --> 00:51:53.130 align:middle line:84% with deadlines, if you say today, 00:51:53.130 --> 00:51:54.660 align:middle line:84% you're going to vaccinate your kids 00:51:54.660 --> 00:51:58.590 align:middle line:84% or never again, present bias does actually not 00:51:58.590 --> 00:51:59.670 align:middle line:90% explain very much. 00:51:59.670 --> 00:52:02.160 align:middle line:90% It doesn't have a lot of bite. 00:52:02.160 --> 00:52:02.832 align:middle line:90% Why is this? 00:52:02.832 --> 00:52:04.290 align:middle line:84% Well, present bias-- usually people 00:52:04.290 --> 00:52:06.600 align:middle line:84% think that people's present bias is about 0.7, 00:52:06.600 --> 00:52:08.370 align:middle line:84% or 0.6, or the like, which is to say 00:52:08.370 --> 00:52:12.930 align:middle line:84% I discount everything that's in the future by about 30%, 40%. 00:52:12.930 --> 00:52:15.840 align:middle line:84% But if I have a rate of return of, say, 500% or something 00:52:15.840 --> 00:52:19.260 align:middle line:84% on investing in my children, then present bias 00:52:19.260 --> 00:52:20.350 align:middle line:90% doesn't get you very far. 00:52:20.350 --> 00:52:24.100 align:middle line:84% It's just not large enough to create large welfare losses. 00:52:24.100 --> 00:52:27.270 align:middle line:84% So there would need to be something that's essentially-- 00:52:27.270 --> 00:52:28.382 align:middle line:90% that can't be deadlines. 00:52:28.382 --> 00:52:30.090 align:middle line:84% And it needs to be something where people 00:52:30.090 --> 00:52:32.160 align:middle line:84% are naive in some ways, people have wrong beliefs 00:52:32.160 --> 00:52:34.510 align:middle line:90% about their future behavior. 00:52:34.510 --> 00:52:36.510 align:middle line:84% So I think that's quite important to understand, 00:52:36.510 --> 00:52:39.630 align:middle line:84% that while it's tempting to say, present bias helps 00:52:39.630 --> 00:52:42.390 align:middle line:84% us solve the Euler equation puzzle, in some ways you 00:52:42.390 --> 00:52:44.185 align:middle line:84% actually have to be quite sophisticated. 00:52:44.185 --> 00:52:45.310 align:middle line:90% The models are complicated. 00:52:45.310 --> 00:52:48.210 align:middle line:84% And you have to be very specific to be able to match 00:52:48.210 --> 00:52:49.140 align:middle line:90% the data that we see. 00:52:49.140 --> 00:52:50.640 align:middle line:84% And it actually matters quite a bit 00:52:50.640 --> 00:52:53.850 align:middle line:84% what some of these small details are. 00:52:53.850 --> 00:52:56.250 align:middle line:84% Now, next, then, is biased beliefs. 00:52:56.250 --> 00:52:58.500 align:middle line:84% And I think this is in part where some of the frontier 00:52:58.500 --> 00:53:01.230 align:middle line:84% is in behavioral development economics, 00:53:01.230 --> 00:53:02.730 align:middle line:84% not only when it comes to health, 00:53:02.730 --> 00:53:06.240 align:middle line:84% but also, say, technology adoption or other types 00:53:06.240 --> 00:53:11.380 align:middle line:84% of issues, which is to say learning is really hard. 00:53:11.380 --> 00:53:13.900 align:middle line:84% Learning about health, in particular, is really hard. 00:53:13.900 --> 00:53:15.550 align:middle line:84% So trying to understand, for example, 00:53:15.550 --> 00:53:18.330 align:middle line:84% whether a certain medication works, 00:53:18.330 --> 00:53:20.550 align:middle line:84% whether you got better if you got therapy 00:53:20.550 --> 00:53:22.350 align:middle line:90% or other types of drugs-- 00:53:22.350 --> 00:53:25.770 align:middle line:84% the reason why that's hard is because lots of health 00:53:25.770 --> 00:53:27.420 align:middle line:90% conditions are self-limiting. 00:53:27.420 --> 00:53:31.020 align:middle line:84% For example, I have some work on psychotherapy. 00:53:31.020 --> 00:53:34.830 align:middle line:84% And it turns out that, in fact, depression, 00:53:34.830 --> 00:53:37.200 align:middle line:84% even if not treated, often remits. 00:53:37.200 --> 00:53:39.460 align:middle line:84% Remission from depression is actually quite high. 00:53:39.460 --> 00:53:43.230 align:middle line:84% So if you look at people who are depressed last year, 00:53:43.230 --> 00:53:45.150 align:middle line:84% a large fraction in many settings-- 00:53:45.150 --> 00:53:48.090 align:middle line:84% 50% or more-- are actually not depressed anymore, regardless 00:53:48.090 --> 00:53:49.510 align:middle line:90% of whether they saw treatment. 00:53:49.510 --> 00:53:51.820 align:middle line:84% They might get depressed again in the future. 00:53:51.820 --> 00:53:54.150 align:middle line:84% But in fact, remission from these types of conditions 00:53:54.150 --> 00:53:55.932 align:middle line:90% is, in fact, quite high. 00:53:55.932 --> 00:53:57.390 align:middle line:84% Now, if you have gotten treatment-- 00:53:57.390 --> 00:54:00.390 align:middle line:84% suppose I went to a therapist and got treatment-- 00:54:00.390 --> 00:54:03.300 align:middle line:84% and I thought the therapist was great and I got better, 00:54:03.300 --> 00:54:06.930 align:middle line:84% how do I know whether that was because of the therapist? 00:54:06.930 --> 00:54:09.030 align:middle line:84% [INAUDIBLE] got better anyway, and it's just 00:54:09.030 --> 00:54:10.950 align:middle line:90% nice to chat with the therapist. 00:54:10.950 --> 00:54:13.910 align:middle line:84% It's extremely hard to actually learn this on their own. 00:54:13.910 --> 00:54:15.660 align:middle line:84% And that's true for many other conditions. 00:54:15.660 --> 00:54:18.285 align:middle line:84% If you think about other health conditions that you might have, 00:54:18.285 --> 00:54:21.210 align:middle line:84% you kind of know there were RCTs or trials that 00:54:21.210 --> 00:54:23.820 align:middle line:84% showed that whatever antibiotics, et cetera, 00:54:23.820 --> 00:54:25.290 align:middle line:90% are effective. 00:54:25.290 --> 00:54:27.720 align:middle line:84% But can you actually tell yourself? 00:54:27.720 --> 00:54:30.120 align:middle line:84% Even things like the COVID vaccine, 00:54:30.120 --> 00:54:34.290 align:middle line:84% you might say, on the individual level, 00:54:34.290 --> 00:54:37.170 align:middle line:84% it's actually very hard to learn this on your own. 00:54:37.170 --> 00:54:39.660 align:middle line:84% Of course, you can see larger, broader data 00:54:39.660 --> 00:54:42.930 align:middle line:84% where you see people who got vaccinated versus not got sick 00:54:42.930 --> 00:54:43.560 align:middle line:90% versus not. 00:54:43.560 --> 00:54:46.120 align:middle line:84% And that helps you learn whether it's, in fact, effective. 00:54:46.120 --> 00:54:48.620 align:middle line:84% But if you think about just from your individual level, what 00:54:48.620 --> 00:54:51.330 align:middle line:84% you observe around you without necessarily having access 00:54:51.330 --> 00:54:54.570 align:middle line:84% to broader data, learning about health conditions, 00:54:54.570 --> 00:54:57.910 align:middle line:84% about the effects of health treatments, is very hard. 00:54:57.910 --> 00:55:01.980 align:middle line:84% So people might have essentially inaccurate beliefs 00:55:01.980 --> 00:55:04.720 align:middle line:90% about health investments. 00:55:04.720 --> 00:55:07.570 align:middle line:84% And there's actually, surprisingly, 00:55:07.570 --> 00:55:10.840 align:middle line:84% little good evidence about that, about what exactly 00:55:10.840 --> 00:55:12.010 align:middle line:90% do people believe. 00:55:12.010 --> 00:55:14.408 align:middle line:84% It's kind of hard to measure in some ways. 00:55:14.408 --> 00:55:16.450 align:middle line:84% And that's partially what the reason for that is. 00:55:16.450 --> 00:55:19.030 align:middle line:84% But partially, I think people have just not studied it 00:55:19.030 --> 00:55:20.380 align:middle line:90% as much as perhaps they should. 00:55:20.380 --> 00:55:22.750 align:middle line:84% So trying to measure, understand the role of beliefs, 00:55:22.750 --> 00:55:26.170 align:middle line:84% and measuring the returns to health investments, 00:55:26.170 --> 00:55:29.830 align:middle line:84% and trying to understand why they believe what they believe 00:55:29.830 --> 00:55:30.980 align:middle line:90% is quite important. 00:55:30.980 --> 00:55:33.430 align:middle line:84% Why are they not learning from others or other data that 00:55:33.430 --> 00:55:35.770 align:middle line:90% might be available? 00:55:35.770 --> 00:55:39.450 align:middle line:84% And then, there's another factor that's really difficult 00:55:39.450 --> 00:55:42.750 align:middle line:84% or interferes potentially with people's correct beliefs 00:55:42.750 --> 00:55:45.532 align:middle line:84% or people's learning, which is motivated beliefs. 00:55:45.532 --> 00:55:47.490 align:middle line:84% This is what I mentioned earlier already, which 00:55:47.490 --> 00:55:49.920 align:middle line:84% is the idea that people derive utility 00:55:49.920 --> 00:55:51.783 align:middle line:84% from thinking that they're healthy, 00:55:51.783 --> 00:55:53.700 align:middle line:84% that they're good looking, that they're smart, 00:55:53.700 --> 00:55:55.710 align:middle line:90% that they're funny, and so on. 00:55:55.710 --> 00:55:59.100 align:middle line:84% Because this is essentially their self image. 00:55:59.100 --> 00:56:01.470 align:middle line:84% They like to think they derive utility 00:56:01.470 --> 00:56:03.480 align:middle line:84% from a good self image of themselves 00:56:03.480 --> 00:56:05.640 align:middle line:84% or thinking that their future is bright. 00:56:05.640 --> 00:56:07.630 align:middle line:84% And a lot of health information, of course, 00:56:07.630 --> 00:56:08.642 align:middle line:90% is potentially negative. 00:56:08.642 --> 00:56:11.100 align:middle line:84% So if I don't do anything, I might say, I'm really healthy. 00:56:11.100 --> 00:56:12.310 align:middle line:90% Why should I see the doctor? 00:56:12.310 --> 00:56:13.510 align:middle line:90% I'm doing just fine. 00:56:13.510 --> 00:56:18.420 align:middle line:84% So people try to fool themselves into being healthy, 00:56:18.420 --> 00:56:19.590 align:middle line:90% potentially. 00:56:19.590 --> 00:56:24.030 align:middle line:84% This is not specific at all to poor settings, again. 00:56:24.030 --> 00:56:25.650 align:middle line:84% But that might be also interfering 00:56:25.650 --> 00:56:27.390 align:middle line:90% with people's learning. 00:56:27.390 --> 00:56:30.180 align:middle line:84% People might be systematically overoptimistic about the health 00:56:30.180 --> 00:56:32.160 align:middle line:90% status. 00:56:32.160 --> 00:56:34.620 align:middle line:84% Or they might want to believe my children are going 00:56:34.620 --> 00:56:36.750 align:middle line:84% to be happy, and healthy, and so on anyway 00:56:36.750 --> 00:56:38.370 align:middle line:84% because they want to believe that. 00:56:38.370 --> 00:56:42.210 align:middle line:84% And that then lowers the return to potential investments 00:56:42.210 --> 00:56:44.160 align:middle line:84% because you're going to do fine anyway. 00:56:44.160 --> 00:56:46.530 align:middle line:84% And why worry about bad things in the future 00:56:46.530 --> 00:56:48.940 align:middle line:84% if that makes me unhappy and worried, 00:56:48.940 --> 00:56:50.420 align:middle line:90% which I wouldn't like to do. 00:56:50.420 --> 00:56:52.965 align:middle line:84% And so there's a literature on that that's relatively small, 00:56:52.965 --> 00:56:54.340 align:middle line:84% but I think it's quite important. 00:56:54.340 --> 00:56:56.048 align:middle line:84% And it might actually explain quite a bit 00:56:56.048 --> 00:56:59.220 align:middle line:84% of interfering with people's learning about health 00:56:59.220 --> 00:57:01.210 align:middle line:90% and health conditions. 00:57:01.210 --> 00:57:03.600 align:middle line:84% And then, maybe more radically in some way-- 00:57:03.600 --> 00:57:05.190 align:middle line:84% here you might say, people might just 00:57:05.190 --> 00:57:07.232 align:middle line:84% underestimate the rate of return in certain ways. 00:57:07.232 --> 00:57:09.787 align:middle line:84% But you can have a neoclassical framework of here's 00:57:09.787 --> 00:57:12.120 align:middle line:84% the rate of return, that's the perceived rate of return, 00:57:12.120 --> 00:57:13.530 align:middle line:90% and that's a little bit lower. 00:57:13.530 --> 00:57:16.710 align:middle line:84% But people might just also have very different mental models 00:57:16.710 --> 00:57:18.420 align:middle line:90% of the world. 00:57:18.420 --> 00:57:20.712 align:middle line:84% People might just have the wrong theories of the world. 00:57:20.712 --> 00:57:23.045 align:middle line:84% They might not even think of certain types of investment 00:57:23.045 --> 00:57:23.970 align:middle line:90% as health investments. 00:57:23.970 --> 00:57:28.290 align:middle line:84% They might not even think that certain types of factors 00:57:28.290 --> 00:57:29.640 align:middle line:90% affect their choices. 00:57:29.640 --> 00:57:31.740 align:middle line:84% And there's some really interesting work 00:57:31.740 --> 00:57:34.890 align:middle line:84% by Josh Schwartzstein and coauthors 00:57:34.890 --> 00:57:38.908 align:middle line:84% on saying people might just have very different models 00:57:38.908 --> 00:57:39.450 align:middle line:90% of the world. 00:57:39.450 --> 00:57:43.650 align:middle line:84% It might be that certain factors might not even 00:57:43.650 --> 00:57:46.320 align:middle line:84% enter their decision making in certain ways. 00:57:46.320 --> 00:57:49.180 align:middle line:84% That is to say, I might not even think-- so for example, 00:57:49.180 --> 00:57:51.750 align:middle line:84% when people are allergic against certain types of food, 00:57:51.750 --> 00:57:54.405 align:middle line:84% you might get headaches or other types of-- 00:57:54.405 --> 00:57:57.547 align:middle line:84% you might be really tired on some days. 00:57:57.547 --> 00:57:59.880 align:middle line:84% But if you don't pay attention to the food that you eat, 00:57:59.880 --> 00:58:01.880 align:middle line:84% if you don't make that mental connection between 00:58:01.880 --> 00:58:03.383 align:middle line:84% your allergies-- or, say, gluten, 00:58:03.383 --> 00:58:05.550 align:middle line:84% or whatever types of allergies that you might have-- 00:58:05.550 --> 00:58:07.350 align:middle line:84% and your decision making, you're never 00:58:07.350 --> 00:58:09.810 align:middle line:84% even going to pay attention to those types of factors. 00:58:09.810 --> 00:58:12.930 align:middle line:84% And if you never pay attention to those types of factors, 00:58:12.930 --> 00:58:15.510 align:middle line:84% how are you going to ever going to learn that certain food is 00:58:15.510 --> 00:58:16.680 align:middle line:90% bad for you or good for you? 00:58:16.680 --> 00:58:18.240 align:middle line:84% Somebody might point it out for you. 00:58:18.240 --> 00:58:20.657 align:middle line:84% Somebody might say, actually, you could have this allergy. 00:58:20.657 --> 00:58:22.320 align:middle line:84% And then you might want to experiment. 00:58:22.320 --> 00:58:24.760 align:middle line:84% But as long as nobody pointed out to you 00:58:24.760 --> 00:58:29.610 align:middle line:84% or explains to you what factors might affect your utility 00:58:29.610 --> 00:58:33.030 align:middle line:84% or your future health and so on, then there's 00:58:33.030 --> 00:58:35.740 align:middle line:84% no chance to actually learn on your own. 00:58:35.740 --> 00:58:37.812 align:middle line:84% And so they might be broadly incorrect-- 00:58:37.812 --> 00:58:39.270 align:middle line:84% so there might be just-- say people 00:58:39.270 --> 00:58:40.675 align:middle line:90% might miss certain factors. 00:58:40.675 --> 00:58:43.050 align:middle line:84% People also might have just very different mental models. 00:58:43.050 --> 00:58:45.587 align:middle line:84% People might be very, very superstitious. 00:58:45.587 --> 00:58:47.670 align:middle line:84% There might be religious or other types of beliefs 00:58:47.670 --> 00:58:49.320 align:middle line:90% and other traditions. 00:58:49.320 --> 00:58:53.400 align:middle line:84% Where if you think about it, if you grow up or live 00:58:53.400 --> 00:58:56.620 align:middle line:84% in a village somewhere in rural areas, 00:58:56.620 --> 00:58:59.788 align:middle line:84% how do you know that modern medicine is better 00:58:59.788 --> 00:59:01.080 align:middle line:90% than the healer in the village? 00:59:01.080 --> 00:59:04.500 align:middle line:84% Or how do you know that I should not believe in witchcraft? 00:59:04.500 --> 00:59:05.670 align:middle line:90% That's a serious question. 00:59:05.670 --> 00:59:07.800 align:middle line:84% It's very hard to actually-- somebody 00:59:07.800 --> 00:59:11.070 align:middle line:84% needs to teach you that or you need to learn it in some ways. 00:59:11.070 --> 00:59:14.490 align:middle line:84% And you need to actually believe that. 00:59:14.490 --> 00:59:16.920 align:middle line:84% And maybe-- of course, for you guys, this is clear. 00:59:16.920 --> 00:59:20.010 align:middle line:84% You have taken a biology or learned about medicine, 00:59:20.010 --> 00:59:21.340 align:middle line:90% and so on, and so forth. 00:59:21.340 --> 00:59:23.880 align:middle line:84% But without necessarily being exposed 00:59:23.880 --> 00:59:26.880 align:middle line:84% to all of those kinds of scientific studies, 00:59:26.880 --> 00:59:29.190 align:middle line:84% it's not obvious what the right model of the world 00:59:29.190 --> 00:59:32.520 align:middle line:84% is from a person's perspective, in 00:59:32.520 --> 00:59:34.770 align:middle line:84% particular in rural areas with less access 00:59:34.770 --> 00:59:38.110 align:middle line:84% to, say, good medical or other types of information. 00:59:38.110 --> 00:59:40.680 align:middle line:84% And so there's some really interesting work by Nava Ashraf 00:59:40.680 --> 00:59:43.920 align:middle line:84% and others on, for example, maternal risk in Zambia 00:59:43.920 --> 00:59:46.920 align:middle line:84% and widespread belief about marital-- 00:59:46.920 --> 00:59:49.230 align:middle line:84% not martial, marital-- infidelity, 00:59:49.230 --> 00:59:51.720 align:middle line:84% and complications during childbirth. 00:59:51.720 --> 00:59:55.920 align:middle line:84% And so there, if you think that this has to do with infidelity, 00:59:55.920 --> 00:59:57.810 align:middle line:84% you can see then how people are very 00:59:57.810 --> 01:00:00.390 align:middle line:84% hesitant to learn about other things of what's going on 01:00:00.390 --> 01:00:03.120 align:middle line:84% because essentially, there's widespread beliefs 01:00:03.120 --> 01:00:04.560 align:middle line:90% about other factors. 01:00:04.560 --> 01:00:07.720 align:middle line:84% And those are actually very hard to refute in some ways. 01:00:07.720 --> 01:00:12.450 align:middle line:84% So I think in some ways, you might have biased beliefs 01:00:12.450 --> 01:00:14.940 align:middle line:84% in a structured way and a neoclassical way, 01:00:14.940 --> 01:00:16.320 align:middle line:84% where we say people make choices. 01:00:16.320 --> 01:00:17.820 align:middle line:84% There's an Euler equation and people 01:00:17.820 --> 01:00:21.810 align:middle line:90% make choices like anybody else. 01:00:21.810 --> 01:00:23.970 align:middle line:84% But their beliefs might be biased in certain ways. 01:00:23.970 --> 01:00:25.800 align:middle line:84% And we can put this into our Euler equation 01:00:25.800 --> 01:00:28.860 align:middle line:84% and change it or then match it in some ways, 01:00:28.860 --> 01:00:32.138 align:middle line:84% by having essentially lower perceived rates of returns. 01:00:32.138 --> 01:00:33.930 align:middle line:84% But if people have incorrect mental models, 01:00:33.930 --> 01:00:35.970 align:middle line:84% they might make choices just very differently. 01:00:35.970 --> 01:00:37.512 align:middle line:84% And the Euler equation, in some ways, 01:00:37.512 --> 01:00:40.018 align:middle line:84% doesn't even apply in some ways, and people make 01:00:40.018 --> 01:00:41.310 align:middle line:90% choices in very different ways. 01:00:41.310 --> 01:00:43.170 align:middle line:84% That's a more radical deviation, if you 01:00:43.170 --> 01:00:46.200 align:middle line:84% want, from neoclassical economics, 01:00:46.200 --> 01:00:49.560 align:middle line:84% but I think extremely interesting to study. 01:00:49.560 --> 01:00:52.370 align:middle line:90% 01:00:52.370 --> 01:00:53.420 align:middle line:90% Questions on this? 01:00:53.420 --> 01:01:01.882 align:middle line:90% 01:01:01.882 --> 01:01:03.340 align:middle line:84% All right, no question either means 01:01:03.340 --> 01:01:05.548 align:middle line:84% everybody is really confused or everybody understood. 01:01:05.548 --> 01:01:08.630 align:middle line:84% So hard to tell, especially with masks. 01:01:08.630 --> 01:01:10.357 align:middle line:90% Yes. 01:01:10.357 --> 01:01:12.940 align:middle line:84% STUDENT: I was wondering if you see any heterogeneity in terms 01:01:12.940 --> 01:01:15.070 align:middle line:84% of people's willingness to pay and how much they're 01:01:15.070 --> 01:01:18.250 align:middle line:84% willing to invest, or if people that are educated in biology, 01:01:18.250 --> 01:01:20.020 align:middle line:84% or something like this might be-- 01:01:20.020 --> 01:01:21.850 align:middle line:84% FRANK SCHILBACH: So can you repeat 01:01:21.850 --> 01:01:23.275 align:middle line:90% the first part of the question? 01:01:23.275 --> 01:01:25.150 align:middle line:84% STUDENT: If there's any heterogeneity that we 01:01:25.150 --> 01:01:27.580 align:middle line:84% see in terms of willingness to pay 01:01:27.580 --> 01:01:32.490 align:middle line:84% or people's willingness to invest in their health. 01:01:32.490 --> 01:01:35.930 align:middle line:84% FRANK SCHILBACH: In terms of how much they know about biology 01:01:35.930 --> 01:01:37.100 align:middle line:90% or other types of medicine? 01:01:37.100 --> 01:01:39.960 align:middle line:84% STUDENT: Yeah, or I guess even more generally in the data, 01:01:39.960 --> 01:01:41.045 align:middle line:90% if we see that most-- 01:01:41.045 --> 01:01:42.920 align:middle line:84% everyone is like this or maybe there's some-- 01:01:42.920 --> 01:01:44.250 align:middle line:84% FRANK SCHILBACH: Yeah, I think that's a good question. 01:01:44.250 --> 01:01:46.370 align:middle line:84% So I think in general, education and health 01:01:46.370 --> 01:01:50.510 align:middle line:84% are very much correlated, like in the US and any other place. 01:01:50.510 --> 01:01:54.980 align:middle line:84% Rich people and educated people are a lot healthier. 01:01:54.980 --> 01:01:58.670 align:middle line:84% And there's lots of work in health economics on the health 01:01:58.670 --> 01:02:01.460 align:middle line:84% education or health income gradient. 01:02:01.460 --> 01:02:03.410 align:middle line:84% But of course, that comes along with-- 01:02:03.410 --> 01:02:06.230 align:middle line:84% being rich or educated comes along with lots of things. 01:02:06.230 --> 01:02:08.060 align:middle line:84% Comes along with health insurance. 01:02:08.060 --> 01:02:11.720 align:middle line:84% It comes along with access to doctors. 01:02:11.720 --> 01:02:14.450 align:middle line:84% It comes along with peer effects of other people 01:02:14.450 --> 01:02:18.500 align:middle line:84% who have access to doctors and talk about health, and so on. 01:02:18.500 --> 01:02:22.190 align:middle line:84% I don't know good studies, but I might be the wrong person 01:02:22.190 --> 01:02:25.820 align:middle line:84% to ask, about specifically isolating the causal effect. 01:02:25.820 --> 01:02:28.160 align:middle line:84% What if you taught a class of, say, 01:02:28.160 --> 01:02:31.850 align:middle line:90% biology or medicine, et cetera? 01:02:31.850 --> 01:02:34.160 align:middle line:84% And what are the returns to doing that? 01:02:34.160 --> 01:02:36.672 align:middle line:84% There's lots of information interventions 01:02:36.672 --> 01:02:37.880 align:middle line:90% that people have tried to do. 01:02:37.880 --> 01:02:39.690 align:middle line:84% I talk about this somewhere in the slides. 01:02:39.690 --> 01:02:41.010 align:middle line:90% I think I may have skipped it. 01:02:41.010 --> 01:02:41.510 align:middle line:90% Here. 01:02:41.510 --> 01:02:46.610 align:middle line:84% Information interventions can have large impacts 01:02:46.610 --> 01:02:50.490 align:middle line:84% on health outcomes in some contexts, but not in others. 01:02:50.490 --> 01:02:52.430 align:middle line:84% So if you look at this handbook chapter 01:02:52.430 --> 01:02:56.220 align:middle line:84% by Pascaline Dupas and Miguel, that talks about that. 01:02:56.220 --> 01:02:59.040 align:middle line:84% And it's somewhat puzzling overall why that is. 01:02:59.040 --> 01:03:00.770 align:middle line:84% And it might have to do precisely 01:03:00.770 --> 01:03:03.140 align:middle line:84% with differences in backgrounds or differences 01:03:03.140 --> 01:03:04.580 align:middle line:90% in type of information. 01:03:04.580 --> 01:03:06.620 align:middle line:84% It might also have to do with motivated beliefs. 01:03:06.620 --> 01:03:08.210 align:middle line:84% In some cases, people don't even want 01:03:08.210 --> 01:03:10.880 align:middle line:84% to learn because they want to think they're healthy. 01:03:10.880 --> 01:03:14.900 align:middle line:84% But I would start here, with that handbook 01:03:14.900 --> 01:03:16.580 align:middle line:90% and trying to understand that. 01:03:16.580 --> 01:03:18.380 align:middle line:84% But that's the right thinking, saying, 01:03:18.380 --> 01:03:20.450 align:middle line:90% can we teach people overall? 01:03:20.450 --> 01:03:23.900 align:middle line:84% Or can people learn about different mental models 01:03:23.900 --> 01:03:25.880 align:middle line:90% of, say, medicine? 01:03:25.880 --> 01:03:28.940 align:middle line:84% And you have to think, then, a lot about trust and about ways 01:03:28.940 --> 01:03:29.780 align:middle line:90% of being compelling. 01:03:29.780 --> 01:03:31.550 align:middle line:84% Because again, if there's a healer 01:03:31.550 --> 01:03:35.210 align:middle line:84% in the village, and a traditional doctor, 01:03:35.210 --> 01:03:37.820 align:middle line:84% it's actually very hard to figure out who's competent. 01:03:37.820 --> 01:03:41.090 align:middle line:84% And why should you believe the traditional doctor versus not? 01:03:41.090 --> 01:03:44.570 align:middle line:90% 01:03:44.570 --> 01:03:46.550 align:middle line:84% So let me then briefly talk a little bit 01:03:46.550 --> 01:03:49.350 align:middle line:84% about time preferences and specific measurements. 01:03:49.350 --> 01:03:52.640 align:middle line:84% So far, we talked about essentially the Euler equation 01:03:52.640 --> 01:03:55.070 align:middle line:84% puzzle and some applications to health. 01:03:55.070 --> 01:03:57.200 align:middle line:84% Now, we talked a lot about time preferences 01:03:57.200 --> 01:04:00.410 align:middle line:84% and how time preferences might be able to explain 01:04:00.410 --> 01:04:01.598 align:middle line:90% certain phenomena. 01:04:01.598 --> 01:04:03.890 align:middle line:84% And now I want to talk a little bit specifically about, 01:04:03.890 --> 01:04:04.730 align:middle line:90% how do we actually measure this? 01:04:04.730 --> 01:04:06.480 align:middle line:84% And how we actually make progress on this? 01:04:06.480 --> 01:04:08.930 align:middle line:84% And what are some of the limits of the literature overall? 01:04:08.930 --> 01:04:11.510 align:middle line:84% As I said, Esther will talk later about savings, 01:04:11.510 --> 01:04:13.950 align:middle line:90% as well, in a later lecture. 01:04:13.950 --> 01:04:16.700 align:middle line:84% And she will actually talk about quasi-hyperbolic discounting, 01:04:16.700 --> 01:04:19.215 align:middle line:84% but mostly focused on savings behavior, 01:04:19.215 --> 01:04:21.840 align:middle line:84% and particularly, why do people not save as much as they could? 01:04:21.840 --> 01:04:24.620 align:middle line:84% And what does, for example, providing savings accounts 01:04:24.620 --> 01:04:26.538 align:middle line:90% to people do, et cetera? 01:04:26.538 --> 01:04:28.580 align:middle line:84% I'm going to talk about something specific, which 01:04:28.580 --> 01:04:30.680 align:middle line:84% is about the measurement of time preferences 01:04:30.680 --> 01:04:32.270 align:middle line:84% and then demand for commitment, which 01:04:32.270 --> 01:04:34.020 align:middle line:90% I'll talk about in a second. 01:04:34.020 --> 01:04:37.460 align:middle line:84% So as it turns out, there's no commonly accepted and easily 01:04:37.460 --> 01:04:41.340 align:middle line:84% implementable approach to measuring time preferences, 01:04:41.340 --> 01:04:43.670 align:middle line:90% in particular in field settings. 01:04:43.670 --> 01:04:46.522 align:middle line:84% So you might say, there's different approaches 01:04:46.522 --> 01:04:47.480 align:middle line:90% that people have taken. 01:04:47.480 --> 01:04:49.320 align:middle line:84% There are broadly three approaches. 01:04:49.320 --> 01:04:51.230 align:middle line:84% One is-- and there's a long literature 01:04:51.230 --> 01:04:53.150 align:middle line:84% on this-- about offering people a choice 01:04:53.150 --> 01:04:56.150 align:middle line:84% between monetary payments today or sometimes 01:04:56.150 --> 01:04:57.680 align:middle line:84% in the present versus the future, 01:04:57.680 --> 01:04:59.085 align:middle line:90% or at different points in time. 01:04:59.085 --> 01:05:00.710 align:middle line:84% That's to say I'm asking you, would you 01:05:00.710 --> 01:05:03.950 align:middle line:84% like to have $100 now versus $200 in the future, 01:05:03.950 --> 01:05:08.930 align:middle line:84% or tomorrow, in 5 days, in 10 days, or next year, and so on. 01:05:08.930 --> 01:05:12.170 align:middle line:84% And using those choices I can then calibrate or elicit 01:05:12.170 --> 01:05:13.370 align:middle line:90% what your preferences are. 01:05:13.370 --> 01:05:16.190 align:middle line:84% Because then I essentially say how much more do you prefer 01:05:16.190 --> 01:05:17.930 align:middle line:90% money now versus in the future? 01:05:17.930 --> 01:05:21.260 align:middle line:84% And does your impatience vary across time horizon? 01:05:21.260 --> 01:05:25.430 align:middle line:84% That is to say, if you're impatient from today 01:05:25.430 --> 01:05:28.430 align:middle line:84% versus tomorrow, are you also impatient from a year 01:05:28.430 --> 01:05:30.560 align:middle line:84% from now versus a year from now plus one day? 01:05:30.560 --> 01:05:35.810 align:middle line:84% And then I can try to understand what does your discounting 01:05:35.810 --> 01:05:37.460 align:middle line:90% function look like. 01:05:37.460 --> 01:05:40.250 align:middle line:84% Now, the issue with that is that choices over money 01:05:40.250 --> 01:05:43.250 align:middle line:84% may not reveal time preferences because the marginal propensity 01:05:43.250 --> 01:05:44.930 align:middle line:90% to consume does not equal 1. 01:05:44.930 --> 01:05:46.700 align:middle line:84% That is to say, if I give you money, 01:05:46.700 --> 01:05:49.113 align:middle line:84% you might not use that money to actually consume. 01:05:49.113 --> 01:05:50.780 align:middle line:84% You might just save it or you might just 01:05:50.780 --> 01:05:52.500 align:middle line:90% borrow differently, and so on. 01:05:52.500 --> 01:05:54.230 align:middle line:84% So what I'm estimating is actually 01:05:54.230 --> 01:05:56.330 align:middle line:84% not people's time preferences, but rather 01:05:56.330 --> 01:05:59.090 align:middle line:84% something that has to do with the interest rate or people's 01:05:59.090 --> 01:06:01.045 align:middle line:90% liquidity constraints. 01:06:01.045 --> 01:06:02.420 align:middle line:84% There's a long literature on that 01:06:02.420 --> 01:06:03.780 align:middle line:84% and people are trying to address that. 01:06:03.780 --> 01:06:05.750 align:middle line:84% So there's a very nice review by Cohen et al. 01:06:05.750 --> 01:06:08.420 align:middle line:90% that discusses that issue. 01:06:08.420 --> 01:06:11.600 align:middle line:84% Second, you might give people choices between consumption, 01:06:11.600 --> 01:06:13.430 align:middle line:84% for example, between food that you 01:06:13.430 --> 01:06:16.790 align:middle line:84% eat at certain points in time, or between drinks or whatever, 01:06:16.790 --> 01:06:19.490 align:middle line:84% or between effort at different points in time, where I say, 01:06:19.490 --> 01:06:22.172 align:middle line:84% would you rather do some work right now versus next week? 01:06:22.172 --> 01:06:23.630 align:middle line:84% And the next week, I ask you again, 01:06:23.630 --> 01:06:25.463 align:middle line:84% would you like to do work next week 01:06:25.463 --> 01:06:26.630 align:middle line:90% and then two weeks from now? 01:06:26.630 --> 01:06:28.970 align:middle line:84% And I can also ask you today about work next week 01:06:28.970 --> 01:06:30.053 align:middle line:90% and in two weeks from now. 01:06:30.053 --> 01:06:33.020 align:middle line:84% And by looking at your choices over time and how they vary, 01:06:33.020 --> 01:06:36.330 align:middle line:84% I can then elicit people's time preferences. 01:06:36.330 --> 01:06:39.270 align:middle line:84% Now, that's conceptually the right object to elicit. 01:06:39.270 --> 01:06:42.270 align:middle line:84% That's what you want to measure because that's essentially 01:06:42.270 --> 01:06:45.240 align:middle line:84% affecting people's utility right away. 01:06:45.240 --> 01:06:46.780 align:middle line:90% There's two issues with that. 01:06:46.780 --> 01:06:48.540 align:middle line:84% And I've spent a lot of time trying 01:06:48.540 --> 01:06:51.990 align:middle line:84% to do this in various setting, including actually in the sleep 01:06:51.990 --> 01:06:55.390 align:middle line:84% paper that I think some of you might have read already. 01:06:55.390 --> 01:07:00.150 align:middle line:84% And so one issue is that people might change their behavior 01:07:00.150 --> 01:07:01.650 align:middle line:90% outside of the experiment. 01:07:01.650 --> 01:07:03.150 align:middle line:84% For example, if I ask you, would you 01:07:03.150 --> 01:07:05.220 align:middle line:84% like to have food today versus next week, 01:07:05.220 --> 01:07:08.268 align:middle line:84% or next week versus two weeks from now, once you show up 01:07:08.268 --> 01:07:09.810 align:middle line:84% next week in my experiment, you might 01:07:09.810 --> 01:07:11.220 align:middle line:90% have eaten already versus not. 01:07:11.220 --> 01:07:16.620 align:middle line:84% And that confounds, then, what I'm trying to estimate. 01:07:16.620 --> 01:07:19.350 align:middle line:84% Or if I ask you, would you like to exert effort today 01:07:19.350 --> 01:07:21.660 align:middle line:84% versus next week or in two weeks from now, 01:07:21.660 --> 01:07:24.412 align:middle line:84% you might also change whatever effort or types 01:07:24.412 --> 01:07:26.370 align:middle line:84% of activities you have in the rest of your life 01:07:26.370 --> 01:07:28.560 align:middle line:90% as a response to my elicitation. 01:07:28.560 --> 01:07:30.460 align:middle line:84% So I have to get you in a setting 01:07:30.460 --> 01:07:32.460 align:middle line:84% where I'm closely controlling what you're doing, 01:07:32.460 --> 01:07:36.685 align:middle line:84% ideally all day, and then try to measure your choices 01:07:36.685 --> 01:07:38.310 align:middle line:84% and then measure also what you're doing 01:07:38.310 --> 01:07:39.943 align:middle line:90% outside of the experiment. 01:07:39.943 --> 01:07:41.610 align:middle line:84% And you can see how that's tricky to do. 01:07:41.610 --> 01:07:44.430 align:middle line:84% If you have people in the lab for two weeks entirely, 01:07:44.430 --> 01:07:46.180 align:middle line:90% you could do that pretty easily. 01:07:46.180 --> 01:07:49.230 align:middle line:84% But if people go about their lives in field experiments-- 01:07:49.230 --> 01:07:52.170 align:middle line:84% you cannot spend seven hours eliciting time preferences-- 01:07:52.170 --> 01:07:54.000 align:middle line:90% then it gets very hard to do. 01:07:54.000 --> 01:07:58.320 align:middle line:84% This tends to be logistically quite challenging, 01:07:58.320 --> 01:08:00.420 align:middle line:84% and in many cases, relatively hard 01:08:00.420 --> 01:08:02.590 align:middle line:84% to do, in part because there's also 01:08:02.590 --> 01:08:03.840 align:middle line:90% a lot of comprehension issues. 01:08:03.840 --> 01:08:06.270 align:middle line:90% It's hard to do specifically. 01:08:06.270 --> 01:08:09.460 align:middle line:84% Now in addition, people have done self-assessed survey 01:08:09.460 --> 01:08:09.960 align:middle line:90% answers. 01:08:09.960 --> 01:08:15.630 align:middle line:84% They essentially just ask you, how 01:08:15.630 --> 01:08:18.300 align:middle line:84% likely are you to delay gratification 01:08:18.300 --> 01:08:19.950 align:middle line:90% on a scale from 1 to 10? 01:08:19.950 --> 01:08:22.830 align:middle line:84% Or how willing are you to wait for good things that 01:08:22.830 --> 01:08:25.120 align:middle line:90% happen to you in the future? 01:08:25.120 --> 01:08:27.255 align:middle line:84% Now, that's a self-assessed answer. 01:08:27.255 --> 01:08:29.130 align:middle line:84% That might actually have a lot of information 01:08:29.130 --> 01:08:32.309 align:middle line:84% because people know actually pretty well what 01:08:32.309 --> 01:08:33.434 align:middle line:90% kinds of choices they make. 01:08:33.434 --> 01:08:39.990 align:middle line:90% 01:08:39.990 --> 01:08:41.450 align:middle line:90% We're back. 01:08:41.450 --> 01:08:43.790 align:middle line:84% However, of course, these are like self reports. 01:08:43.790 --> 01:08:46.130 align:middle line:84% And economists are very notoriously skeptical 01:08:46.130 --> 01:08:49.609 align:middle line:84% of self reports, and sometimes for good reason. 01:08:49.609 --> 01:08:51.710 align:middle line:84% Because there might be demand effects. 01:08:51.710 --> 01:08:54.170 align:middle line:84% You might want to tell me that you're really patient 01:08:54.170 --> 01:08:57.172 align:middle line:84% and look good in front of the experimenter. 01:08:57.172 --> 01:08:59.630 align:middle line:84% But they might have also other issues, such as social image 01:08:59.630 --> 01:09:00.689 align:middle line:90% concerns or the like. 01:09:00.689 --> 01:09:02.330 align:middle line:84% You might want to think of yourself 01:09:02.330 --> 01:09:05.640 align:middle line:84% as a patient person, which in reality might not be true. 01:09:05.640 --> 01:09:08.569 align:middle line:84% And so then, what you're telling me might not necessarily 01:09:08.569 --> 01:09:10.729 align:middle line:84% reflect how patient you actually are, 01:09:10.729 --> 01:09:17.100 align:middle line:84% but what you'd like yourself to be. 01:09:17.100 --> 01:09:20.060 align:middle line:84% And so there's no easy answer here what to do. 01:09:20.060 --> 01:09:22.979 align:middle line:84% I think in some ways, if you asked me five years ago, 01:09:22.979 --> 01:09:24.979 align:middle line:84% I would have said this is the right thing to do. 01:09:24.979 --> 01:09:27.439 align:middle line:84% This is conceptually the right object to elicit. 01:09:27.439 --> 01:09:29.632 align:middle line:84% Right now, I'm much less convinced of that 01:09:29.632 --> 01:09:32.090 align:middle line:84% partially because people have trouble logistically actually 01:09:32.090 --> 01:09:34.800 align:middle line:84% implementing it with high comprehension rates. 01:09:34.800 --> 01:09:36.500 align:middle line:84% So I think by now, I would say, maybe 01:09:36.500 --> 01:09:38.149 align:middle line:84% having a combination of those measures 01:09:38.149 --> 01:09:39.830 align:middle line:84% actually measures things pretty well. 01:09:39.830 --> 01:09:43.370 align:middle line:84% You can try to do convex time budgets in some ways that 01:09:43.370 --> 01:09:46.015 align:middle line:84% elicit these choices, but you want 01:09:46.015 --> 01:09:48.140 align:middle line:84% to be careful with making sure that people actually 01:09:48.140 --> 01:09:51.490 align:middle line:90% understand what they're doing. 01:09:51.490 --> 01:09:52.560 align:middle line:90% Any questions on this? 01:09:52.560 --> 01:09:56.980 align:middle line:90% 01:09:56.980 --> 01:09:59.320 align:middle line:84% So now, this is about time preference, 01:09:59.320 --> 01:10:02.283 align:middle line:84% if I wanted to estimate specific time preference parameters. 01:10:02.283 --> 01:10:04.450 align:middle line:84% If I wanted to know what is your beta and your delta 01:10:04.450 --> 01:10:07.600 align:middle line:84% and the utility function, that's what I want to do in some way, 01:10:07.600 --> 01:10:09.340 align:middle line:90% at least approach 1 and 2. 01:10:09.340 --> 01:10:11.830 align:middle line:84% Now, another way to do this is to say, 01:10:11.830 --> 01:10:14.060 align:middle line:84% how do we measure whether people are present biased? 01:10:14.060 --> 01:10:16.840 align:middle line:84% We can offer people commitment devices. 01:10:16.840 --> 01:10:18.170 align:middle line:90% What's a commitment device? 01:10:18.170 --> 01:10:20.890 align:middle line:84% Here's a definition-- it's an arrangement entered 01:10:20.890 --> 01:10:24.100 align:middle line:84% into by an agent who restricts his or her future choice set 01:10:24.100 --> 01:10:27.310 align:middle line:84% by making certain choices more expensive, perhaps infinitely 01:10:27.310 --> 01:10:29.020 align:middle line:90% more expensive. 01:10:29.020 --> 01:10:30.770 align:middle line:90% What does that mean? 01:10:30.770 --> 01:10:33.250 align:middle line:84% That's a relatively convoluted definition. 01:10:33.250 --> 01:10:35.950 align:middle line:84% Essentially, it's to say present bias preferences are time 01:10:35.950 --> 01:10:36.670 align:middle line:90% inconsistent. 01:10:36.670 --> 01:10:41.320 align:middle line:84% That is to say, if I am present biased, I want a lot of utility 01:10:41.320 --> 01:10:44.230 align:middle line:84% right now or I want to eat donuts or not 01:10:44.230 --> 01:10:46.040 align:middle line:90% exercise so on right now. 01:10:46.040 --> 01:10:47.900 align:middle line:84% But in the future, I would like to do it. 01:10:47.900 --> 01:10:50.067 align:middle line:84% I would say, next week, I'm going to exercise a lot. 01:10:50.067 --> 01:10:52.750 align:middle line:84% I'm going to get up at 6:00 in the morning, write a poem, 01:10:52.750 --> 01:10:54.760 align:middle line:90% and go running. 01:10:54.760 --> 01:10:57.430 align:middle line:84% Now, if next week comes at 6:00 AM, 01:10:57.430 --> 01:11:00.610 align:middle line:84% I might just want to sleep and have no interest in doing so. 01:11:00.610 --> 01:11:02.320 align:middle line:84% So my preferences are inconsistent. 01:11:02.320 --> 01:11:05.770 align:middle line:84% That's to say my plans change even if nothing else changes, 01:11:05.770 --> 01:11:07.790 align:middle line:84% there's no information, or the like. 01:11:07.790 --> 01:11:12.310 align:middle line:84% And so different selves want different things for themselves 01:11:12.310 --> 01:11:13.810 align:middle line:90% in the future. 01:11:13.810 --> 01:11:16.990 align:middle line:84% And in particular, people, present biased individuals, 01:11:16.990 --> 01:11:19.360 align:middle line:84% want themselves to be more patient in the future. 01:11:19.360 --> 01:11:21.043 align:middle line:84% They realize that in the future-- 01:11:21.043 --> 01:11:22.960 align:middle line:84% sophisticated individuals realize that they're 01:11:22.960 --> 01:11:24.190 align:middle line:90% impatient in the future. 01:11:24.190 --> 01:11:26.560 align:middle line:84% And they want to then change their prices 01:11:26.560 --> 01:11:31.150 align:middle line:84% or restrict their options in the future in an effort to change 01:11:31.150 --> 01:11:32.590 align:middle line:90% their future behavior. 01:11:32.590 --> 01:11:36.490 align:middle line:84% For example, maybe I want to set myself incentives to engage 01:11:36.490 --> 01:11:38.290 align:middle line:84% in certain behaviors because I know 01:11:38.290 --> 01:11:40.720 align:middle line:84% I want to set myself incentives to exercise. 01:11:40.720 --> 01:11:44.680 align:middle line:84% So I tell somebody else, every time I exercise, pay me $10 01:11:44.680 --> 01:11:49.270 align:middle line:84% or I can give that person money, with the hope 01:11:49.270 --> 01:11:51.940 align:middle line:84% that in the future I'll change my behavior. 01:11:51.940 --> 01:11:55.000 align:middle line:84% Notice that that requires sophistication. 01:11:55.000 --> 01:11:59.810 align:middle line:84% You need to know that you're present biased in the future. 01:11:59.810 --> 01:12:02.110 align:middle line:84% So you need to understand your behavior, as well. 01:12:02.110 --> 01:12:03.340 align:middle line:90% Let me tell you-- yeah. 01:12:03.340 --> 01:12:05.215 align:middle line:84% STUDENT: Wouldn't that also require some sort 01:12:05.215 --> 01:12:06.057 align:middle line:90% of [INAUDIBLE]? 01:12:06.057 --> 01:12:07.390 align:middle line:90% FRANK SCHILBACH: Say that again. 01:12:07.390 --> 01:12:08.848 align:middle line:84% STUDENT: Wouldn't that also require 01:12:08.848 --> 01:12:11.903 align:middle line:90% some effort in the present? 01:12:11.903 --> 01:12:13.320 align:middle line:84% FRANK SCHILBACH: Well, it depends. 01:12:13.320 --> 01:12:14.650 align:middle line:90% You could just say-- 01:12:14.650 --> 01:12:19.020 align:middle line:84% for example, I could tell Ed, next week I'm 01:12:19.020 --> 01:12:20.610 align:middle line:84% going to have a draft of this paper. 01:12:20.610 --> 01:12:22.800 align:middle line:84% And if not, I'm going to pay $100. 01:12:22.800 --> 01:12:25.560 align:middle line:84% Here's my credit card and you can make some expense 01:12:25.560 --> 01:12:26.940 align:middle line:90% for something that I don't want. 01:12:26.940 --> 01:12:28.523 align:middle line:84% I don't need to do anything right now. 01:12:28.523 --> 01:12:30.810 align:middle line:84% I'm just saying next week, I'm going 01:12:30.810 --> 01:12:32.530 align:middle line:90% to work really hard on my paper. 01:12:32.530 --> 01:12:37.140 align:middle line:84% And if I don't do it, I'm going to lose $100 to the Pro Smoking 01:12:37.140 --> 01:12:38.710 align:middle line:90% Association or something. 01:12:38.710 --> 01:12:41.190 align:middle line:84% And so now, there's no effort right now. 01:12:41.190 --> 01:12:43.830 align:middle line:84% It's just to say next week if I don't do it, 01:12:43.830 --> 01:12:44.880 align:middle line:90% bad things will happen. 01:12:44.880 --> 01:12:47.100 align:middle line:84% If Ed is ruthless and implements, 01:12:47.100 --> 01:12:49.308 align:middle line:84% then I might change my future behavior 01:12:49.308 --> 01:12:51.600 align:middle line:84% because I know next week, I'm going to slack off or not 01:12:51.600 --> 01:12:53.070 align:middle line:90% work on my draft. 01:12:53.070 --> 01:12:53.820 align:middle line:90% But not right now. 01:12:53.820 --> 01:12:55.060 align:middle line:90% I don't do anything. 01:12:55.060 --> 01:12:57.833 align:middle line:84% So some commitment devices are-- there different types 01:12:57.833 --> 01:12:58.750 align:middle line:90% of commitment devices. 01:12:58.750 --> 01:13:00.750 align:middle line:84% But some are essentially to say purely, 01:13:00.750 --> 01:13:03.390 align:middle line:84% you're going to just try to change future prices to affect 01:13:03.390 --> 01:13:05.770 align:middle line:90% your behavior. 01:13:05.770 --> 01:13:07.900 align:middle line:84% So let me briefly tell you a little bit about this. 01:13:07.900 --> 01:13:10.840 align:middle line:84% I know I've five minutes left, so let me briefly 01:13:10.840 --> 01:13:12.020 align:middle line:90% tell you about this. 01:13:12.020 --> 01:13:14.980 align:middle line:84% So one example about commitment contracts from a paper of mine 01:13:14.980 --> 01:13:17.860 align:middle line:90% is about drinking in Chennai. 01:13:17.860 --> 01:13:20.380 align:middle line:84% These are rickshaw pullers in Chennai, 01:13:20.380 --> 01:13:23.230 align:middle line:84% where a large fraction of low-income men drink daily. 01:13:23.230 --> 01:13:26.500 align:middle line:84% They spend high fractions of their income on alcohol. 01:13:26.500 --> 01:13:29.080 align:middle line:84% There seem to be severe consequences, at least reported 01:13:29.080 --> 01:13:31.265 align:middle line:84% by people, for these men and their families 01:13:31.265 --> 01:13:33.890 align:middle line:84% in terms of health, in terms of money that's available to them, 01:13:33.890 --> 01:13:34.450 align:middle line:90% and so on. 01:13:34.450 --> 01:13:36.550 align:middle line:84% And lots of people say they want to drink less. 01:13:36.550 --> 01:13:39.460 align:middle line:84% So they say they want to change their future behavior. 01:13:39.460 --> 01:13:43.510 align:middle line:84% But everybody drinks a lot, or many people drink a lot, 01:13:43.510 --> 01:13:44.290 align:middle line:90% in the present. 01:13:44.290 --> 01:13:46.150 align:middle line:84% Here is a fraction of people who are 01:13:46.150 --> 01:13:48.040 align:middle line:84% inebriated or drunk during regular work 01:13:48.040 --> 01:13:49.150 align:middle line:90% hours during the day. 01:13:49.150 --> 01:13:51.940 align:middle line:84% About half of them seem to be inebriated or drunk 01:13:51.940 --> 01:13:54.350 align:middle line:90% during regular work hours. 01:13:54.350 --> 01:13:57.010 align:middle line:84% And so what I then did is I did a field experiment 01:13:57.010 --> 01:14:00.592 align:middle line:84% for three weeks with these drivers 01:14:00.592 --> 01:14:02.800 align:middle line:84% where everybody would come to the office between 6:00 01:14:02.800 --> 01:14:03.820 align:middle line:90% and 10:00 PM. 01:14:03.820 --> 01:14:05.920 align:middle line:84% I did a short survey under breathalyzer tests, 01:14:05.920 --> 01:14:08.500 align:middle line:84% and people were given financial incentives for sobriety 01:14:08.500 --> 01:14:11.260 align:middle line:84% for some random subset of individuals. 01:14:11.260 --> 01:14:14.020 align:middle line:84% And then I wanted to measure people's demand for commitment. 01:14:14.020 --> 01:14:15.447 align:middle line:84% As I said, if you just ask people, 01:14:15.447 --> 01:14:17.780 align:middle line:84% most people say, yeah, yeah, I would like to drink less. 01:14:17.780 --> 01:14:21.010 align:middle line:84% But of course, that's the right answer, in a sense, 01:14:21.010 --> 01:14:23.360 align:middle line:84% in terms of the socially desired answer. 01:14:23.360 --> 01:14:27.410 align:middle line:84% So I wanted to have hard evidence on that. 01:14:27.410 --> 01:14:31.030 align:middle line:84% So what I then did is offer people dominated contracts 01:14:31.030 --> 01:14:32.560 align:middle line:90% of the following sort-- 01:14:32.560 --> 01:14:37.210 align:middle line:84% so people were given the choices between either incentives 01:14:37.210 --> 01:14:39.800 align:middle line:84% for sobriety-- that's to say, every time you show up, 01:14:39.800 --> 01:14:42.340 align:middle line:84% you get $1 regardless of how drunk you are. 01:14:42.340 --> 01:14:44.860 align:middle line:84% Then there's a breathalyzer test, 01:14:44.860 --> 01:14:48.280 align:middle line:84% and anybody who is sober with a blood alcohol content of 0 01:14:48.280 --> 01:14:49.420 align:middle line:90% gets another dollar. 01:14:49.420 --> 01:14:51.790 align:middle line:90% So 60 rupees is about $1. 01:14:51.790 --> 01:14:53.350 align:middle line:84% So that's just straight up incentive. 01:14:53.350 --> 01:14:56.090 align:middle line:84% You get $1 for any time in addition, 01:14:56.090 --> 01:14:59.320 align:middle line:84% if you are not inebriated, if your blood alcohol 01:14:59.320 --> 01:15:00.640 align:middle line:90% content is 0. 01:15:00.640 --> 01:15:03.550 align:middle line:84% And option B-- this is what you see on the right side-- 01:15:03.550 --> 01:15:05.110 align:middle line:84% is people get payments regardless 01:15:05.110 --> 01:15:06.940 align:middle line:90% of their blood alcohol content. 01:15:06.940 --> 01:15:08.350 align:middle line:84% Now, they're given three choices. 01:15:08.350 --> 01:15:11.470 align:middle line:84% These are different rows, row 1, 2, and 3. 01:15:11.470 --> 01:15:14.920 align:middle line:84% So the first choice where people choose between incentives 01:15:14.920 --> 01:15:16.990 align:middle line:84% and the outside option of 90 rupees, 01:15:16.990 --> 01:15:19.240 align:middle line:84% notice that is not demand for commitment. 01:15:19.240 --> 01:15:22.600 align:middle line:84% Because people are-- you might choose that option 01:15:22.600 --> 01:15:26.147 align:middle line:84% even if you just don't want to change your drinking. 01:15:26.147 --> 01:15:27.730 align:middle line:84% You might just think it's a good deal. 01:15:27.730 --> 01:15:29.260 align:middle line:84% Next week, I'm going to drink less. 01:15:29.260 --> 01:15:31.177 align:middle line:84% It's not because I want to change my drinking, 01:15:31.177 --> 01:15:33.010 align:middle line:84% I just have plans to not drink very much. 01:15:33.010 --> 01:15:35.440 align:middle line:84% So now, choosing option A is actually a good deal 01:15:35.440 --> 01:15:38.170 align:middle line:84% because most of the time, I'm going to show up sober anyway. 01:15:38.170 --> 01:15:40.030 align:middle line:90% So I'm going to choose option A. 01:15:40.030 --> 01:15:43.450 align:middle line:84% But then in particular, option three that people were offered 01:15:43.450 --> 01:15:47.200 align:middle line:84% is to say, you get either these incentives of 60 rupees 01:15:47.200 --> 01:15:51.070 align:middle line:84% if you show up and are drunk, 120 rupees if you are sober 01:15:51.070 --> 01:15:54.740 align:middle line:84% and show up, or the other option is any time you show up, 01:15:54.740 --> 01:15:57.100 align:middle line:90% you get 150 rupees anyway. 01:15:57.100 --> 01:15:59.410 align:middle line:84% Now notice, if you choose now option A, 01:15:59.410 --> 01:16:01.100 align:middle line:90% that's a dominated contract. 01:16:01.100 --> 01:16:03.550 align:middle line:84% That's a contract that's worse in terms of study payments. 01:16:03.550 --> 01:16:05.825 align:middle line:84% You're always going to get less money regardless 01:16:05.825 --> 01:16:06.700 align:middle line:90% of how drunk you are. 01:16:06.700 --> 01:16:08.440 align:middle line:84% Even if you show up sober all the time, 01:16:08.440 --> 01:16:10.240 align:middle line:90% you're going to get less money. 01:16:10.240 --> 01:16:11.650 align:middle line:90% So now the only reason-- 01:16:11.650 --> 01:16:13.480 align:middle line:84% that's essentially a commitment contract. 01:16:13.480 --> 01:16:17.230 align:middle line:84% The only reason why we would choose this option, option A, 01:16:17.230 --> 01:16:19.870 align:middle line:84% is because you want to set yourself incentives to drink 01:16:19.870 --> 01:16:22.040 align:middle line:90% less in the future. 01:16:22.040 --> 01:16:25.960 align:middle line:84% And so that's essentially the idea of this experiment, 01:16:25.960 --> 01:16:28.120 align:middle line:84% or of eliciting commitment contract, 01:16:28.120 --> 01:16:30.430 align:middle line:84% is to say you reveal people's preferences 01:16:30.430 --> 01:16:31.810 align:middle line:90% to change their future prices. 01:16:31.810 --> 01:16:34.210 align:middle line:84% Are people willing to give up money in some ways 01:16:34.210 --> 01:16:35.860 align:middle line:84% to change their prices and affect 01:16:35.860 --> 01:16:38.060 align:middle line:90% their behavior in the future? 01:16:38.060 --> 01:16:40.550 align:middle line:90% So let me skip some of this. 01:16:40.550 --> 01:16:43.840 align:middle line:84% People have very high demand for these incentives. 01:16:43.840 --> 01:16:46.570 align:middle line:84% We look at over there, about a third to half of people 01:16:46.570 --> 01:16:49.240 align:middle line:84% are willing to give up a substantial amount of money 01:16:49.240 --> 01:16:53.570 align:middle line:84% to receive incentives for sobriety. 01:16:53.570 --> 01:16:56.180 align:middle line:90% Now, let me skip this. 01:16:56.180 --> 01:16:59.680 align:middle line:84% So that's the case, so now, we have 01:16:59.680 --> 01:17:03.040 align:middle line:84% shown that people have demand for commitment, 01:17:03.040 --> 01:17:04.540 align:middle line:84% where this literature, in some ways, 01:17:04.540 --> 01:17:07.090 align:middle line:84% is that people often then engage in commitment devices. 01:17:07.090 --> 01:17:09.250 align:middle line:84% So people demand commitment devices, 01:17:09.250 --> 01:17:10.690 align:middle line:90% but then they actually fail. 01:17:10.690 --> 01:17:12.640 align:middle line:84% So in my case, in my paper, what happened 01:17:12.640 --> 01:17:15.710 align:middle line:84% is that people, in fact, were showing up sober at 6:00 01:17:15.710 --> 01:17:17.920 align:middle line:84% PM or 7:00 PM at the study office. 01:17:17.920 --> 01:17:23.180 align:middle line:84% They demanded these commitment contracts. 01:17:23.180 --> 01:17:25.180 align:middle line:84% They were actually drinking less during the day, 01:17:25.180 --> 01:17:27.055 align:middle line:84% but then they were substituting into drinking 01:17:27.055 --> 01:17:28.510 align:middle line:90% more later at night. 01:17:28.510 --> 01:17:30.020 align:middle line:90% So that didn't help, either. 01:17:30.020 --> 01:17:31.600 align:middle line:84% So as a policy, that's not working. 01:17:31.600 --> 01:17:33.460 align:middle line:84% As a proof of concept that self-control 01:17:33.460 --> 01:17:36.010 align:middle line:84% seems really important for people's behavior, that's 01:17:36.010 --> 01:17:38.740 align:middle line:84% really helpful, but it's not really a useful policy 01:17:38.740 --> 01:17:42.460 align:middle line:84% that actually helps people reduce their drinking overall. 01:17:42.460 --> 01:17:46.090 align:middle line:84% And similarly, I've written down some open research questions, 01:17:46.090 --> 01:17:50.170 align:middle line:84% but broadly, there's a notion that many people often 01:17:50.170 --> 01:17:52.510 align:middle line:84% demand commitment but then don't follow through, 01:17:52.510 --> 01:17:55.570 align:middle line:84% the reason being that they essentially are partially 01:17:55.570 --> 01:17:56.560 align:middle line:90% naive in some ways. 01:17:56.560 --> 01:17:59.950 align:middle line:84% They think that great, I have some self-control problems. 01:17:59.950 --> 01:18:02.800 align:middle line:84% Let me purchase this commitment device. 01:18:02.800 --> 01:18:05.410 align:middle line:84% But then they overestimate how useful that commitment device 01:18:05.410 --> 01:18:07.347 align:middle line:84% is in the future, and then fail, and actually 01:18:07.347 --> 01:18:09.430 align:middle line:84% are potentially worse off than if they hadn't even 01:18:09.430 --> 01:18:12.230 align:middle line:84% purchased a commitment device in the first place. 01:18:12.230 --> 01:18:14.920 align:middle line:84% So there's a very nice paper by Liang Bai and co-authors 01:18:14.920 --> 01:18:18.670 align:middle line:84% about low demand for commitment devices for health visits 01:18:18.670 --> 01:18:21.010 align:middle line:84% in India, where demand for commitment 01:18:21.010 --> 01:18:22.390 align:middle line:90% is relatively low overall. 01:18:22.390 --> 01:18:24.460 align:middle line:84% But even among people who demand commitment, 01:18:24.460 --> 01:18:26.323 align:middle line:90% few people follow through. 01:18:26.323 --> 01:18:27.490 align:middle line:90% So where is this literature? 01:18:27.490 --> 01:18:30.970 align:middle line:84% The literature essentially is somewhat inconclusive 01:18:30.970 --> 01:18:32.380 align:middle line:84% in a sense, to say there's proofs 01:18:32.380 --> 01:18:35.960 align:middle line:84% of concept of high demand for commitment in some settings. 01:18:35.960 --> 01:18:38.140 align:middle line:84% But in other settings demand for commitment 01:18:38.140 --> 01:18:39.437 align:middle line:90% is, in fact, quite low. 01:18:39.437 --> 01:18:41.020 align:middle line:84% So one broad question is, why is that? 01:18:41.020 --> 01:18:43.000 align:middle line:90% Why is there so much variation? 01:18:43.000 --> 01:18:46.270 align:middle line:84% And more broadly, how can we design commitment contracts, 01:18:46.270 --> 01:18:50.350 align:middle line:84% in particular in development policy contexts, that 01:18:50.350 --> 01:18:54.130 align:middle line:84% are actually useful and help people improve outcomes? 01:18:54.130 --> 01:18:56.230 align:middle line:84% We have shown, as a proof of concept, 01:18:56.230 --> 01:18:57.975 align:middle line:84% that people have self-control problems. 01:18:57.975 --> 01:19:00.100 align:middle line:84% But we have shown much less that these are actually 01:19:00.100 --> 01:19:03.230 align:middle line:84% useful policies that we can do to improve outcomes. 01:19:03.230 --> 01:19:06.770 align:middle line:84% And there are some questions-- commitment savings accounts-- 01:19:06.770 --> 01:19:08.270 align:middle line:84% could those be helpful and the like? 01:19:08.270 --> 01:19:12.580 align:middle line:84% But how can we modify these types of contracts 01:19:12.580 --> 01:19:14.990 align:middle line:90% to help people improve outcomes? 01:19:14.990 --> 01:19:16.930 align:middle line:84% And so that's where this literature is. 01:19:16.930 --> 01:19:18.970 align:middle line:84% I think it's a really interesting, exciting 01:19:18.970 --> 01:19:21.370 align:middle line:84% literature, but it's a little bit stuck at this point 01:19:21.370 --> 01:19:23.080 align:middle line:90% in trying to understand. 01:19:23.080 --> 01:19:26.170 align:middle line:84% So beyond showing that people have self-control problems, 01:19:26.170 --> 01:19:29.080 align:middle line:84% how do we actually help people do better? 01:19:29.080 --> 01:19:31.468 align:middle line:84% And there could be interesting things to do. 01:19:31.468 --> 01:19:33.760 align:middle line:84% There's some work on, for example, incentives to sleep, 01:19:33.760 --> 01:19:36.760 align:middle line:84% incentives to do therapy or at least try therapy, 01:19:36.760 --> 01:19:38.740 align:middle line:84% incentives for sobriety in other settings 01:19:38.740 --> 01:19:41.123 align:middle line:84% that have more sophisticated measurement devices, 01:19:41.123 --> 01:19:41.665 align:middle line:90% and the like. 01:19:41.665 --> 01:19:44.270 align:middle line:90% 01:19:44.270 --> 01:19:45.910 align:middle line:84% And then showing that this is actually 01:19:45.910 --> 01:19:49.270 align:middle line:84% a policy, in some ways, that can help people do better 01:19:49.270 --> 01:19:52.060 align:middle line:90% would be the next thing to do. 01:19:52.060 --> 01:19:54.250 align:middle line:84% So then, as I said, the handbook chapter 01:19:54.250 --> 01:19:56.950 align:middle line:84% has lots of other exciting topics. 01:19:56.950 --> 01:19:59.230 align:middle line:84% My favorite are 6 and 8, perhaps. 01:19:59.230 --> 01:20:00.730 align:middle line:84% One is about technology adoption, 01:20:00.730 --> 01:20:02.680 align:middle line:84% which is about behavioral social learning. 01:20:02.680 --> 01:20:03.580 align:middle line:90% How do people learn? 01:20:03.580 --> 01:20:07.450 align:middle line:84% And what things interfere with people's learning decisions? 01:20:07.450 --> 01:20:10.420 align:middle line:84% As well as A, behavioral firms that say, 01:20:10.420 --> 01:20:14.620 align:middle line:84% it's not just people or consumers are potentially 01:20:14.620 --> 01:20:17.200 align:middle line:84% affected by behavioral biases, but also firms, 01:20:17.200 --> 01:20:18.730 align:middle line:84% in particular shopkeepers-- so types 01:20:18.730 --> 01:20:22.878 align:middle line:84% of businesses that might make potentially wrong decisions. 01:20:22.878 --> 01:20:24.420 align:middle line:84% As I said, in the next lecture, we're 01:20:24.420 --> 01:20:26.337 align:middle line:84% going to talk about the psychology of poverty, 01:20:26.337 --> 01:20:29.350 align:middle line:84% in particular about financial strain, sleep 01:20:29.350 --> 01:20:32.015 align:middle line:90% deprivation, and mental health. 01:20:32.015 --> 01:20:33.140 align:middle line:90% I'm a little bit over time. 01:20:33.140 --> 01:20:34.630 align:middle line:90% So I'm going to stop here. 01:20:34.630 --> 01:20:38.440 align:middle line:84% Happy to chat more after lecture, or of course, 01:20:38.440 --> 01:20:39.190 align:middle line:90% next time. 01:20:39.190 --> 01:20:41.190 align:middle line:90% Thank you so much. 01:20:41.190 --> 01:20:54.000 align:middle line:90%