WEBVTT 00:00:00.000 --> 00:00:03.888 align:middle line:90% [SQUEAKING][RUSTLING][CLICKING] 00:00:03.888 --> 00:00:11.147 align:middle line:90% 00:00:11.147 --> 00:00:11.980 align:middle line:90% ROBERT TOWNSEND: OK. 00:00:11.980 --> 00:00:19.000 align:middle line:84% So let me update where we are in terms of the calendar. 00:00:19.000 --> 00:00:22.360 align:middle line:84% I guess a notable thing is that we 00:00:22.360 --> 00:00:27.130 align:middle line:84% have reached lecture 10 today, Thursday, in terms 00:00:27.130 --> 00:00:29.590 align:middle line:90% of the reading list. 00:00:29.590 --> 00:00:32.334 align:middle line:84% We are-- having done all the risk-sharing. 00:00:32.334 --> 00:00:33.880 align:middle line:90% Yeah, we're at lecture 10. 00:00:33.880 --> 00:00:37.420 align:middle line:84% So there's a couple of things that are starred here, 00:00:37.420 --> 00:00:40.910 align:middle line:84% and they obviously both come up in the lecture today. 00:00:40.910 --> 00:00:43.210 align:middle line:84% One is the early version of this paper 00:00:43.210 --> 00:00:47.020 align:middle line:84% with Alvarez and Pawasutipaisit sit 00:00:47.020 --> 00:00:49.570 align:middle line:90% cash, which has a star on it. 00:00:49.570 --> 00:00:52.900 align:middle line:84% It's a little bit rough because it's kind of, even then, 00:00:52.900 --> 00:00:54.340 align:middle line:90% in the process of revision. 00:00:54.340 --> 00:00:57.850 align:middle line:84% And I have added new material in the slides that haven't even 00:00:57.850 --> 00:00:59.980 align:middle line:84% made their way into the working paper yet. 00:00:59.980 --> 00:01:02.620 align:middle line:84% And then there's Samphantharak and Townsend, 00:01:02.620 --> 00:01:04.930 align:middle line:84% which is describing the tools we're 00:01:04.930 --> 00:01:06.790 align:middle line:90% going to be using a lot today-- 00:01:06.790 --> 00:01:11.740 align:middle line:84% how to organize data, and think about flows and stocks in terms 00:01:11.740 --> 00:01:15.170 align:middle line:90% of financial accounts and so on. 00:01:15.170 --> 00:01:17.890 align:middle line:84% So, again, the two starred readings. 00:01:17.890 --> 00:01:20.950 align:middle line:84% Other things you might enjoy, time permitting, 00:01:20.950 --> 00:01:23.530 align:middle line:84% there's something on Kenya, M-Pesa. 00:01:23.530 --> 00:01:25.810 align:middle line:84% And I have one slide describing-- actually 00:01:25.810 --> 00:01:28.360 align:middle line:84% two slides describing that today; 00:01:28.360 --> 00:01:30.850 align:middle line:90% something on lifecycle planning. 00:01:30.850 --> 00:01:33.920 align:middle line:84% But none of those things have stars on them. 00:01:33.920 --> 00:01:34.420 align:middle line:90% OK. 00:01:34.420 --> 00:01:36.590 align:middle line:90% So that's the reading list. 00:01:36.590 --> 00:01:40.360 align:middle line:84% The last thing, turning to the study guide or the lecture 00:01:40.360 --> 00:01:44.590 align:middle line:84% from last time, which was "Risk and Return 00:01:44.590 --> 00:01:45.970 align:middle line:90% in Village Thailand." 00:01:45.970 --> 00:01:51.310 align:middle line:84% So explain intuitively in words why multiple technologies can 00:01:51.310 --> 00:01:54.190 align:middle line:84% be chosen to be operated simultaneously, 00:01:54.190 --> 00:01:55.990 align:middle line:84% and why the rate of return on some 00:01:55.990 --> 00:01:58.330 align:middle line:84% might be higher than the rate of return 00:01:58.330 --> 00:02:01.120 align:middle line:84% for the others that are also operating. 00:02:01.120 --> 00:02:05.380 align:middle line:84% And for that matter, why some projects are not chosen at all. 00:02:05.380 --> 00:02:08.410 align:middle line:84% If you can, try to use the language of risk premium, 00:02:08.410 --> 00:02:10.720 align:middle line:90% but don't use it as jargon. 00:02:10.720 --> 00:02:13.540 align:middle line:84% Try to imagine you're telling someone in your household 00:02:13.540 --> 00:02:17.840 align:middle line:84% or a friend kind of what we're studying in class. 00:02:17.840 --> 00:02:19.990 align:middle line:90% So I've been calling on people. 00:02:19.990 --> 00:02:22.150 align:middle line:84% Let me just ask for volunteers today. 00:02:22.150 --> 00:02:26.370 align:middle line:90% 00:02:26.370 --> 00:02:28.890 align:middle line:90% Of course, I can call on people. 00:02:28.890 --> 00:02:31.360 align:middle line:90% Any volunteers? 00:02:31.360 --> 00:02:34.350 align:middle line:84% Carrie, you want to take a crack at this? 00:02:34.350 --> 00:02:37.260 align:middle line:84% AUDIENCE: Is it because some technologies 00:02:37.260 --> 00:02:41.280 align:middle line:84% have a lower rate of return because they're riskier? 00:02:41.280 --> 00:02:44.550 align:middle line:84% And the probability of a high rate of return 00:02:44.550 --> 00:02:45.690 align:middle line:90% is lower for them? 00:02:45.690 --> 00:02:48.997 align:middle line:90% 00:02:48.997 --> 00:02:50.205 align:middle line:90% ROBERT TOWNSEND: That's true. 00:02:50.205 --> 00:02:53.320 align:middle line:90% 00:02:53.320 --> 00:02:56.230 align:middle line:84% At first blush, the rate of return 00:02:56.230 --> 00:02:59.950 align:middle line:84% is simply all the possible returns multiplied 00:02:59.950 --> 00:03:02.780 align:middle line:90% by the respective probability. 00:03:02.780 --> 00:03:05.560 align:middle line:84% So that's kind of inherent in the technology itself. 00:03:05.560 --> 00:03:09.060 align:middle line:90% 00:03:09.060 --> 00:03:11.960 align:middle line:84% But if that were the only thing going on, then why would 00:03:11.960 --> 00:03:15.542 align:middle line:84% they not choose the one with the highest expected return? 00:03:15.542 --> 00:03:17.000 align:middle line:84% AUDIENCE: Can I take a crack at it? 00:03:17.000 --> 00:03:18.440 align:middle line:90% ROBERT TOWNSEND: Yep. 00:03:18.440 --> 00:03:20.570 align:middle line:84% AUDIENCE: Well, I think they've got-- yeah, 00:03:20.570 --> 00:03:22.250 align:middle line:84% I'm a bit behind in the readings. 00:03:22.250 --> 00:03:24.770 align:middle line:84% But what it looks like here is that you 00:03:24.770 --> 00:03:27.350 align:middle line:84% have some technologies that have a higher expected rate 00:03:27.350 --> 00:03:29.870 align:middle line:84% of return but more risk, and some technologies that 00:03:29.870 --> 00:03:32.780 align:middle line:84% have the low rate of return and lower risk. 00:03:32.780 --> 00:03:36.660 align:middle line:84% And basically-- and some of them-- 00:03:36.660 --> 00:03:38.360 align:middle line:90% so some of them are-- 00:03:38.360 --> 00:03:40.860 align:middle line:84% some people-- it's like some households might be more like-- 00:03:40.860 --> 00:03:43.470 align:middle line:84% it's like, you can have like-- and if these aren't perfectly 00:03:43.470 --> 00:03:46.950 align:middle line:84% correlated, then you might be able to get a better-- 00:03:46.950 --> 00:03:50.430 align:middle line:84% sort of a less risky rate for the same-- less 00:03:50.430 --> 00:03:53.730 align:middle line:84% risk for the same rate of return by diversification. 00:03:53.730 --> 00:03:58.020 align:middle line:84% And the reason that you would have lower-- 00:03:58.020 --> 00:04:02.010 align:middle line:84% you would have-- you would have lower rate of return 00:04:02.010 --> 00:04:03.420 align:middle line:90% technologies in the same-- 00:04:03.420 --> 00:04:06.060 align:middle line:84% used at the same time as higher rate of return technologies, 00:04:06.060 --> 00:04:10.610 align:middle line:84% as the lower rate of return points, have a-- 00:04:10.610 --> 00:04:13.518 align:middle line:84% the lower rate of return points have a lower risk, 00:04:13.518 --> 00:04:15.060 align:middle line:84% and the higher rate of return points, 00:04:15.060 --> 00:04:17.070 align:middle line:84% you have to basically pay a risk-- 00:04:17.070 --> 00:04:18.649 align:middle line:84% there's risk premium there, which 00:04:18.649 --> 00:04:21.567 align:middle line:84% is why it's less desirable or something like that. 00:04:21.567 --> 00:04:22.500 align:middle line:90% ROBERT TOWNSEND: Yep. 00:04:22.500 --> 00:04:23.000 align:middle line:90% OK. 00:04:23.000 --> 00:04:23.792 align:middle line:90% That's pretty good. 00:04:23.792 --> 00:04:24.410 align:middle line:90% Thank you. 00:04:24.410 --> 00:04:27.970 align:middle line:90% 00:04:27.970 --> 00:04:28.870 align:middle line:90% Yeah. 00:04:28.870 --> 00:04:30.400 align:middle line:84% What you're describing is correct. 00:04:30.400 --> 00:04:33.730 align:middle line:84% There's like a mean variability trade-off 00:04:33.730 --> 00:04:38.260 align:middle line:84% here, a downward-sloping frontier, 00:04:38.260 --> 00:04:40.210 align:middle line:84% if you want to think about it that way. 00:04:40.210 --> 00:04:42.430 align:middle line:84% As you move across different projects 00:04:42.430 --> 00:04:45.580 align:middle line:84% they may have higher and higher means, 00:04:45.580 --> 00:04:49.130 align:middle line:84% but they come with higher and higher variance. 00:04:49.130 --> 00:04:50.305 align:middle line:90% So there's a trade-off. 00:04:50.305 --> 00:04:53.330 align:middle line:90% 00:04:53.330 --> 00:04:55.140 align:middle line:84% And since households are risk-averse, 00:04:55.140 --> 00:04:57.900 align:middle line:90% they don't like the variance. 00:04:57.900 --> 00:04:59.225 align:middle line:90% So they pay attention to that. 00:04:59.225 --> 00:05:02.550 align:middle line:90% 00:05:02.550 --> 00:05:04.340 align:middle line:90% And you said this also. 00:05:04.340 --> 00:05:08.090 align:middle line:84% You could have two technologies with the same mean return-- 00:05:08.090 --> 00:05:10.200 align:middle line:84% one with a higher variance than the other one, 00:05:10.200 --> 00:05:13.400 align:middle line:84% and they would never do the higher variability one 00:05:13.400 --> 00:05:16.270 align:middle line:84% because then there's no trade-off. 00:05:16.270 --> 00:05:18.580 align:middle line:84% Since they get the same mean but higher risk, 00:05:18.580 --> 00:05:21.010 align:middle line:90% they wouldn't choose that. 00:05:21.010 --> 00:05:23.920 align:middle line:84% There is a diversification aspect 00:05:23.920 --> 00:05:28.480 align:middle line:84% that we featured quite a bit in terms of land plots 00:05:28.480 --> 00:05:31.300 align:middle line:84% and how they want to diversify, or in village India 00:05:31.300 --> 00:05:34.870 align:middle line:84% over crops and income sources and so on. 00:05:34.870 --> 00:05:38.410 align:middle line:84% And that's kind of captured here in a new way, 00:05:38.410 --> 00:05:40.120 align:middle line:90% but it is in here. 00:05:40.120 --> 00:05:47.200 align:middle line:84% Namely, a guess of what technologies 00:05:47.200 --> 00:05:52.540 align:middle line:84% to operate overall, you can compute the variability. 00:05:52.540 --> 00:05:55.150 align:middle line:84% And then you look on the margin about whether to add 00:05:55.150 --> 00:05:58.390 align:middle line:90% or subtract one more. 00:05:58.390 --> 00:06:03.250 align:middle line:84% And you look at its variability relative to this village 00:06:03.250 --> 00:06:05.530 align:middle line:90% average. 00:06:05.530 --> 00:06:11.730 align:middle line:84% So if it's going to add to diversification, 00:06:11.730 --> 00:06:16.050 align:middle line:84% it's going to have a return stream which is low when-- 00:06:16.050 --> 00:06:19.570 align:middle line:84% well, say high when the rest of the village is low. 00:06:19.570 --> 00:06:21.690 align:middle line:84% So that's kind of the way the diversification 00:06:21.690 --> 00:06:28.370 align:middle line:84% aspect is coming out in this mean variance framework. 00:06:28.370 --> 00:06:29.040 align:middle line:90% But great. 00:06:29.040 --> 00:06:30.620 align:middle line:90% That's very helpful. 00:06:30.620 --> 00:06:31.120 align:middle line:90% All right. 00:06:31.120 --> 00:06:34.570 align:middle line:90% So here's another one. 00:06:34.570 --> 00:06:36.790 align:middle line:84% Explain how it can be that while the magnitude 00:06:36.790 --> 00:06:40.300 align:middle line:84% of idiosyncratic risk is higher than the magnitude 00:06:40.300 --> 00:06:45.760 align:middle line:84% of aggregate risk, nevertheless the idiosyncratic risk premium 00:06:45.760 --> 00:06:50.220 align:middle line:84% is smaller than the aggregate risk premium. 00:06:50.220 --> 00:06:53.630 align:middle line:90% So, again, I'll take volunteers. 00:06:53.630 --> 00:06:55.700 align:middle line:90% AUDIENCE: Yeah. 00:06:55.700 --> 00:07:00.440 align:middle line:84% So is it-- so I think it's because when 00:07:00.440 --> 00:07:02.330 align:middle line:84% it comes with idiosyncratic risk, 00:07:02.330 --> 00:07:07.260 align:middle line:84% it's possible for the village to distribute the returns 00:07:07.260 --> 00:07:08.180 align:middle line:90% amongst each other. 00:07:08.180 --> 00:07:09.950 align:middle line:90% It's like ex-post. 00:07:09.950 --> 00:07:12.740 align:middle line:84% Whereas, for aggregate risk, it's like 00:07:12.740 --> 00:07:14.240 align:middle line:84% is risk affecting the whole village. 00:07:14.240 --> 00:07:17.780 align:middle line:84% And so there's not-- and so redistribution can't really 00:07:17.780 --> 00:07:21.230 align:middle line:84% benefit anyone when it comes to things that 00:07:21.230 --> 00:07:23.230 align:middle line:90% affect the village as a whole. 00:07:23.230 --> 00:07:24.130 align:middle line:90% ROBERT TOWNSEND: Yep. 00:07:24.130 --> 00:07:25.620 align:middle line:90% Perfect. 00:07:25.620 --> 00:07:28.470 align:middle line:84% Yeah, the last part, which I didn't read out, 00:07:28.470 --> 00:07:31.030 align:middle line:84% explain in words what gifts and loans have to do with it, 00:07:31.030 --> 00:07:32.820 align:middle line:90% but you kind of did already. 00:07:32.820 --> 00:07:37.890 align:middle line:84% That's the way they pool away most, if not all, 00:07:37.890 --> 00:07:39.810 align:middle line:90% of the idiosyncratic risk. 00:07:39.810 --> 00:07:43.170 align:middle line:90% 00:07:43.170 --> 00:07:44.850 align:middle line:84% Whereas, in the aggregate, it's common. 00:07:44.850 --> 00:07:45.810 align:middle line:90% It's hitting everyone. 00:07:45.810 --> 00:07:49.270 align:middle line:84% So essentially, you can't pool it away. 00:07:49.270 --> 00:07:51.900 align:middle line:84% And this next part is related to that. 00:07:51.900 --> 00:07:55.230 align:middle line:84% Is there full insurance in these Thai village economies or not? 00:07:55.230 --> 00:07:57.900 align:middle line:90% And to give it away a bit, not. 00:07:57.900 --> 00:08:00.720 align:middle line:84% And, in particular, what is it in the data 00:08:00.720 --> 00:08:04.440 align:middle line:84% that is inconsistent with full risk-sharing? 00:08:04.440 --> 00:08:07.570 align:middle line:90% Volunteers? 00:08:07.570 --> 00:08:09.880 align:middle line:84% AUDIENCE: We find that in the data 00:08:09.880 --> 00:08:11.830 align:middle line:90% household consumption is still-- 00:08:11.830 --> 00:08:18.130 align:middle line:84% is positively correlated with the idiosyncratic shocks. 00:08:18.130 --> 00:08:22.180 align:middle line:84% So that shows that it's not full risk-sharing. 00:08:22.180 --> 00:08:23.680 align:middle line:84% But there's still some risk-sharing 00:08:23.680 --> 00:08:26.170 align:middle line:84% because the coefficient is really, really small 00:08:26.170 --> 00:08:28.302 align:middle line:90% compared to other coefficients. 00:08:28.302 --> 00:08:29.260 align:middle line:90% ROBERT TOWNSEND: Right. 00:08:29.260 --> 00:08:33.159 align:middle line:84% Compared to the coefficients on gifts and lending, 00:08:33.159 --> 00:08:36.740 align:middle line:84% which absorbs the bulk of the risk. 00:08:36.740 --> 00:08:38.020 align:middle line:90% So that's a perfect answer. 00:08:38.020 --> 00:08:40.000 align:middle line:90% I am so happy with this class. 00:08:40.000 --> 00:08:42.669 align:middle line:84% I enjoy teaching you tremendously. 00:08:42.669 --> 00:08:46.060 align:middle line:90% It's very gratifying. 00:08:46.060 --> 00:08:46.560 align:middle line:90% OK. 00:08:46.560 --> 00:08:47.100 align:middle line:90% Thank you. 00:08:47.100 --> 00:08:50.250 align:middle line:84% There's one more-- see if we can get it in. 00:08:50.250 --> 00:08:54.570 align:middle line:84% Describe one example from Thai villages of propagation 00:08:54.570 --> 00:08:57.090 align:middle line:84% through networks that would be inconsistent 00:08:57.090 --> 00:08:59.880 align:middle line:84% with full risk-sharing and investment efficiency. 00:08:59.880 --> 00:09:03.080 align:middle line:90% 00:09:03.080 --> 00:09:05.710 align:middle line:90% And I'm abbreviating. 00:09:05.710 --> 00:09:08.950 align:middle line:84% So this happened at the end of the lecture, just as a hint. 00:09:08.950 --> 00:09:12.607 align:middle line:84% Can someone described what we did at the end of the lecture? 00:09:12.607 --> 00:09:14.940 align:middle line:84% Well, it's kind of given away here in this bullet point, 00:09:14.940 --> 00:09:16.890 align:middle line:90% but I'll read it out anyway. 00:09:16.890 --> 00:09:19.590 align:middle line:84% How can an individual specific illness shock 00:09:19.590 --> 00:09:23.760 align:middle line:84% turn into an aggregate village-level downturn? 00:09:23.760 --> 00:09:24.620 align:middle line:90% Volunteers? 00:09:24.620 --> 00:09:31.830 align:middle line:90% 00:09:31.830 --> 00:09:34.810 align:middle line:84% AUDIENCE: I think at the end of the class, 00:09:34.810 --> 00:09:37.407 align:middle line:84% we were talking about how some of the networks-- 00:09:37.407 --> 00:09:39.990 align:middle line:84% how different kinds of networks are different from each other. 00:09:39.990 --> 00:09:42.420 align:middle line:84% I'm not sure which specific networks were being compared, 00:09:42.420 --> 00:09:45.387 align:middle line:90% but yeah. 00:09:45.387 --> 00:09:46.595 align:middle line:90% ROBERT TOWNSEND: That's true. 00:09:46.595 --> 00:09:49.290 align:middle line:90% 00:09:49.290 --> 00:09:50.870 align:middle line:84% So what happens, do you recall, when 00:09:50.870 --> 00:09:53.150 align:middle line:84% a household gets hit with an illness shock that 00:09:53.150 --> 00:09:54.860 align:middle line:90% increases its expenditure? 00:09:54.860 --> 00:09:58.660 align:middle line:90% 00:09:58.660 --> 00:10:01.070 align:middle line:90% AUDIENCE: Right. 00:10:01.070 --> 00:10:03.180 align:middle line:84% ROBERT TOWNSEND: Did nothing happen? 00:10:03.180 --> 00:10:04.680 align:middle line:90% Something? 00:10:04.680 --> 00:10:08.890 align:middle line:84% You remember, or anyone else remember? 00:10:08.890 --> 00:10:12.850 align:middle line:84% AUDIENCE: I remember that the households who 00:10:12.850 --> 00:10:17.950 align:middle line:84% are directly connected to the household who got illness 00:10:17.950 --> 00:10:22.720 align:middle line:84% shocks would [INAUDIBLE] down their working opportunity 00:10:22.720 --> 00:10:26.860 align:middle line:84% much more than households who are unconnected 00:10:26.860 --> 00:10:30.190 align:middle line:90% to those illness households. 00:10:30.190 --> 00:10:37.240 align:middle line:84% So through the network, households who are connected, 00:10:37.240 --> 00:10:43.360 align:middle line:84% who are related to each other much more maybe shock-- 00:10:43.360 --> 00:10:46.100 align:middle line:84% there was an actual shock between them. 00:10:46.100 --> 00:10:51.040 align:middle line:84% So sudden illness shocks to these households 00:10:51.040 --> 00:10:55.350 align:middle line:84% may transfer to the aggregate village 00:10:55.350 --> 00:10:58.270 align:middle line:90% through the [INAUDIBLE]. 00:10:58.270 --> 00:10:59.690 align:middle line:90% ROBERT TOWNSEND: OK. 00:10:59.690 --> 00:11:01.100 align:middle line:90% That's a pretty good. 00:11:01.100 --> 00:11:03.830 align:middle line:84% And it's a bit tricky, this question. 00:11:03.830 --> 00:11:07.820 align:middle line:84% Because there are different networks in the village. 00:11:07.820 --> 00:11:10.230 align:middle line:84% There's kind of the risk-sharing insurance network. 00:11:10.230 --> 00:11:11.610 align:middle line:90% There's the labor network. 00:11:11.610 --> 00:11:14.960 align:middle line:84% There's the supply chain production network. 00:11:14.960 --> 00:11:18.500 align:middle line:90% And they don't coincide. 00:11:18.500 --> 00:11:21.770 align:middle line:84% So if you got hit with an illness shock, 00:11:21.770 --> 00:11:25.730 align:middle line:84% but nevertheless, you're in this risk-sharing insurance network, 00:11:25.730 --> 00:11:28.790 align:middle line:84% then part of the-- or almost all of the shock 00:11:28.790 --> 00:11:32.780 align:middle line:84% gets covered, in come the gifts, just like we were describing 00:11:32.780 --> 00:11:36.220 align:middle line:90% with the production shock. 00:11:36.220 --> 00:11:40.990 align:middle line:84% However, when that illness shock is not covered by gifts, 00:11:40.990 --> 00:11:44.080 align:middle line:84% then the household has to respond on its own in some way. 00:11:44.080 --> 00:11:49.490 align:middle line:84% And it does this by cutting the number of laborers 00:11:49.490 --> 00:11:53.480 align:middle line:84% that it hires, and by shrinking-- 00:11:53.480 --> 00:11:57.720 align:middle line:84% saving costs by shrinking its production operations. 00:11:57.720 --> 00:12:01.250 align:middle line:84% And that is what spills over to other households 00:12:01.250 --> 00:12:04.120 align:middle line:90% along the production network. 00:12:04.120 --> 00:12:06.960 align:middle line:84% So networks could be good or bad. 00:12:06.960 --> 00:12:09.900 align:middle line:84% Networks are good for sharing risk, 00:12:09.900 --> 00:12:16.340 align:middle line:84% but they also can be bad because they propagate contagion. 00:12:16.340 --> 00:12:18.840 align:middle line:90% So that was-- 00:12:18.840 --> 00:12:24.260 align:middle line:84% AUDIENCE: So if there was a full risk-sharing, then the-- 00:12:24.260 --> 00:12:31.330 align:middle line:84% then there were no effect on the employment for the household 00:12:31.330 --> 00:12:33.670 align:middle line:90% who did not get any shock? 00:12:33.670 --> 00:12:36.100 align:middle line:90% Network-- I mean, network-- 00:12:36.100 --> 00:12:37.570 align:middle line:90% that was a factor. 00:12:37.570 --> 00:12:39.700 align:middle line:84% Those houses who did not get any shock 00:12:39.700 --> 00:12:45.640 align:middle line:84% because they are more related to the households with illness. 00:12:45.640 --> 00:12:49.180 align:middle line:84% But if there was full risk-sharing, 00:12:49.180 --> 00:12:52.480 align:middle line:84% those households will not have any effect? 00:12:52.480 --> 00:12:54.880 align:middle line:90% ROBERT TOWNSEND: That's right. 00:12:54.880 --> 00:12:59.710 align:middle line:84% So the emphasis here is on a very specific household 00:12:59.710 --> 00:13:03.450 align:middle line:90% idiosyncratic shock. 00:13:03.450 --> 00:13:07.440 align:middle line:84% And this-- I drew an analogy with COVID-19, 00:13:07.440 --> 00:13:08.460 align:middle line:90% but it's a bit unfair. 00:13:08.460 --> 00:13:12.090 align:middle line:84% Because COVID, as we know, has a contagion effect 00:13:12.090 --> 00:13:14.980 align:middle line:90% literally in the health sense. 00:13:14.980 --> 00:13:18.170 align:middle line:84% And if one person gets it, people around that person 00:13:18.170 --> 00:13:18.670 align:middle line:90% get it. 00:13:18.670 --> 00:13:24.930 align:middle line:84% So here, though, it's a different kind of illness. 00:13:24.930 --> 00:13:27.750 align:middle line:84% But it-- and it could be just an elderly person 00:13:27.750 --> 00:13:28.650 align:middle line:90% in the household. 00:13:28.650 --> 00:13:31.290 align:middle line:84% But still, there are big expenses-- 00:13:31.290 --> 00:13:33.160 align:middle line:90% double the normal expenses. 00:13:33.160 --> 00:13:35.910 align:middle line:84% So they don't have the budget to manage, 00:13:35.910 --> 00:13:39.150 align:middle line:84% and they have to start either getting a gift from someone 00:13:39.150 --> 00:13:41.160 align:middle line:90% or cutting. 00:13:41.160 --> 00:13:43.290 align:middle line:84% So when they start cutting expenses, 00:13:43.290 --> 00:13:46.950 align:middle line:84% that's when it spills over to other people that they hire 00:13:46.950 --> 00:13:47.700 align:middle line:90% or people. 00:13:47.700 --> 00:13:49.800 align:middle line:84% So this is very reminiscent, as I 00:13:49.800 --> 00:13:53.520 align:middle line:84% said in class, of what's going on in the supply chain 00:13:53.520 --> 00:13:58.590 align:middle line:84% networks in the US, for example, and many other countries. 00:13:58.590 --> 00:14:02.275 align:middle line:84% Although, that's more of a different kind of health shock. 00:14:02.275 --> 00:14:05.560 align:middle line:90% 00:14:05.560 --> 00:14:06.060 align:middle line:90% OK. 00:14:06.060 --> 00:14:11.170 align:middle line:84% Any questions about-- any other questions about the lecture? 00:14:11.170 --> 00:14:11.670 align:middle line:90% OK. 00:14:11.670 --> 00:14:15.600 align:middle line:84% So then we come to the lecture for today-- 00:14:15.600 --> 00:14:17.520 align:middle line:90% ledgers and management. 00:14:17.520 --> 00:14:20.460 align:middle line:84% In fact, there's a lot in the lecture today. 00:14:20.460 --> 00:14:22.020 align:middle line:84% I would say I'm going to give you 00:14:22.020 --> 00:14:25.990 align:middle line:84% a tool to think about, namely, financial accounts-- 00:14:25.990 --> 00:14:29.730 align:middle line:84% income and balance sheet accounts, in particular. 00:14:29.730 --> 00:14:33.360 align:middle line:84% And then we'll go to the Thai monthly data 00:14:33.360 --> 00:14:37.380 align:middle line:84% that we were just talking about, and use the tool 00:14:37.380 --> 00:14:40.950 align:middle line:84% to think about planning-- about wealth planning. 00:14:40.950 --> 00:14:44.670 align:middle line:84% Then we'll take a step back and add another account. 00:14:44.670 --> 00:14:48.690 align:middle line:84% I'll motivate that account by thinking about Bitcoin, 00:14:48.690 --> 00:14:54.330 align:middle line:84% as I promised in lecture one, and link the ledgers of Bitcoin 00:14:54.330 --> 00:14:56.370 align:middle line:90% to the statement of cash flow. 00:14:56.370 --> 00:14:59.100 align:middle line:84% And hence, back to the Thai villages. 00:14:59.100 --> 00:15:01.830 align:middle line:84% In particular, I'll talk about distributed ledgers, which 00:15:01.830 --> 00:15:04.470 align:middle line:90% is a big part of Bitcoin-- 00:15:04.470 --> 00:15:07.440 align:middle line:84% Bitcoin, blockchain, and distributed ledgers. 00:15:07.440 --> 00:15:10.140 align:middle line:84% And then having introduced that concept, 00:15:10.140 --> 00:15:14.250 align:middle line:84% I'll focus on the use almost exclusively 00:15:14.250 --> 00:15:18.030 align:middle line:84% of paper currency in Thailand, the use 00:15:18.030 --> 00:15:21.900 align:middle line:84% of almost exclusively electronic transfers in Sweden, 00:15:21.900 --> 00:15:24.850 align:middle line:84% and something in between in Kenya, 00:15:24.850 --> 00:15:28.290 align:middle line:84% which has two different kinds of monies, essentially-- 00:15:28.290 --> 00:15:30.870 align:middle line:84% privately issued e-money from a cell phone 00:15:30.870 --> 00:15:35.130 align:middle line:84% provider, and a public Fiat money, Kenyan shillings. 00:15:35.130 --> 00:15:38.670 align:middle line:84% And then I'll end-- and so these three things for Thailand, 00:15:38.670 --> 00:15:40.770 align:middle line:84% Sweden, and Kenya are about cash-- 00:15:40.770 --> 00:15:43.410 align:middle line:84% various concepts of cash and cash management. 00:15:43.410 --> 00:15:46.410 align:middle line:84% And then we'll come back to the ledgers 00:15:46.410 --> 00:15:49.710 align:middle line:84% again and talk about trusted third parties, some problems 00:15:49.710 --> 00:15:51.750 align:middle line:84% and management that may or may not 00:15:51.750 --> 00:15:53.880 align:middle line:84% show up at the individual level, but definitely 00:15:53.880 --> 00:15:56.730 align:middle line:84% show up at the wholesale level at the level 00:15:56.730 --> 00:15:58.650 align:middle line:90% of financial infrastructure. 00:15:58.650 --> 00:15:59.670 align:middle line:90% OK. 00:15:59.670 --> 00:16:04.170 align:middle line:90% So here's a statement of income. 00:16:04.170 --> 00:16:07.110 align:middle line:84% And forgive me if this is a review for some of you. 00:16:07.110 --> 00:16:10.570 align:middle line:84% These things aren't hard, but they're a bit tedious. 00:16:10.570 --> 00:16:14.250 align:middle line:84% So we talk about sources of income and uses of income. 00:16:14.250 --> 00:16:17.130 align:middle line:84% And the sources of income, as in the Thai data, 00:16:17.130 --> 00:16:20.430 align:middle line:84% come from cultivation, livestock, fish and shrimp. 00:16:20.430 --> 00:16:24.150 align:middle line:84% This is a wage-earning household primarily, 00:16:24.150 --> 00:16:25.440 align:middle line:90% a little bit of livestock. 00:16:25.440 --> 00:16:28.410 align:middle line:84% So the line entry here is almost entirely 00:16:28.410 --> 00:16:30.900 align:middle line:90% from wage, labor, income. 00:16:30.900 --> 00:16:32.040 align:middle line:90% Those are the sources. 00:16:32.040 --> 00:16:34.740 align:middle line:84% And then the uses, so to speak, or you could 00:16:34.740 --> 00:16:36.750 align:middle line:90% say revenues versus costs-- 00:16:36.750 --> 00:16:42.060 align:middle line:84% you have expenses from each of the activities-- livestock. 00:16:42.060 --> 00:16:44.490 align:middle line:84% If they're in debt, they have interest expenses. 00:16:44.490 --> 00:16:47.580 align:middle line:84% There's that depreciation that we talked about last time. 00:16:47.580 --> 00:16:50.140 align:middle line:84% And then this is charges against revenue. 00:16:50.140 --> 00:16:52.980 align:middle line:84% So in addition to having expenses for production, 00:16:52.980 --> 00:16:54.390 align:middle line:90% they might pay taxes-- 00:16:54.390 --> 00:16:55.260 align:middle line:90% these guys don't. 00:16:55.260 --> 00:17:00.150 align:middle line:84% But they can either consume the rest or save. 00:17:00.150 --> 00:17:01.830 align:middle line:84% Advantage of the financial accounts 00:17:01.830 --> 00:17:03.310 align:middle line:90% is that everything balances. 00:17:03.310 --> 00:17:05.099 align:middle line:84% So in this case, revenue is totally 00:17:05.099 --> 00:17:07.290 align:middle line:84% balanced by charges against revenue. 00:17:07.290 --> 00:17:09.810 align:middle line:84% And if you're-- as we did-- taking the Thai data 00:17:09.810 --> 00:17:12.540 align:middle line:84% and thinking of all the line items then creating 00:17:12.540 --> 00:17:15.030 align:middle line:84% the accounts, and if these left- and right-hand sides 00:17:15.030 --> 00:17:17.069 align:middle line:84% don't balance, we know we something wrong. 00:17:17.069 --> 00:17:20.130 align:middle line:84% And we have to go back and iterate and figure out-- 00:17:20.130 --> 00:17:23.339 align:middle line:84% excluded a variable, or put it in the inappropriate spot. 00:17:23.339 --> 00:17:26.130 align:middle line:84% So it's a great discipline device. 00:17:26.130 --> 00:17:30.660 align:middle line:84% The balance sheet is stocks of stuff at a point in time, 00:17:30.660 --> 00:17:34.620 align:middle line:90% namely, liabilities and assets. 00:17:34.620 --> 00:17:36.480 align:middle line:84% So the obvious liability would be 00:17:36.480 --> 00:17:39.600 align:middle line:90% debt, which this household has. 00:17:39.600 --> 00:17:41.460 align:middle line:84% This is really somebody's numbers. 00:17:41.460 --> 00:17:44.520 align:middle line:84% This is a real person, I want to remind you. 00:17:44.520 --> 00:17:46.540 align:middle line:90% More on that momentarily. 00:17:46.540 --> 00:17:49.440 align:middle line:84% And the debt's about 37,000 baht. 00:17:49.440 --> 00:17:52.770 align:middle line:84% And the assets that the household has 00:17:52.770 --> 00:17:54.915 align:middle line:90% could be real or financial. 00:17:54.915 --> 00:17:59.910 align:middle line:84% The financial assets are a big chunk of it in paper currency 00:17:59.910 --> 00:18:01.860 align:middle line:90% and some deposits at the bank. 00:18:01.860 --> 00:18:06.550 align:middle line:84% And real assets, in this case, household fixed assets. 00:18:06.550 --> 00:18:08.640 align:middle line:84% So we went through some examples last time 00:18:08.640 --> 00:18:11.190 align:middle line:84% when we talked about rates of return, because we 00:18:11.190 --> 00:18:14.100 align:middle line:84% had in the numerator profits, which 00:18:14.100 --> 00:18:17.790 align:middle line:84% is on the previous account revenue minus expenses. 00:18:17.790 --> 00:18:19.950 align:middle line:90% And here, profits over capital. 00:18:19.950 --> 00:18:25.590 align:middle line:84% And here we have the real and also financial capital. 00:18:25.590 --> 00:18:28.680 align:middle line:84% Now, again, these things balance. 00:18:28.680 --> 00:18:32.400 align:middle line:84% But here, there's a big item that I deliberately 00:18:32.400 --> 00:18:35.520 align:middle line:84% didn't mention at first to get your attention, which 00:18:35.520 --> 00:18:38.940 align:middle line:84% is if you take assets minus liabilities, 00:18:38.940 --> 00:18:42.270 align:middle line:90% hopefully net worth is positive. 00:18:42.270 --> 00:18:44.580 align:middle line:84% If liabilities were greater than assets, 00:18:44.580 --> 00:18:48.010 align:middle line:84% this guy's bankrupt, at least in the accounting sense. 00:18:48.010 --> 00:18:50.100 align:middle line:84% So for almost all these households, 00:18:50.100 --> 00:18:52.290 align:middle line:84% assets are greater than liabilities. 00:18:52.290 --> 00:18:56.040 align:middle line:84% And the difference is something we call equity or net worth. 00:18:56.040 --> 00:18:58.770 align:middle line:84% And that enters oddly on the right-hand side 00:18:58.770 --> 00:19:01.890 align:middle line:84% with liabilities, precisely so that the two sides 00:19:01.890 --> 00:19:04.950 align:middle line:84% of the financial account, or the balance sheet account, 00:19:04.950 --> 00:19:06.870 align:middle line:90% are equal to each other. 00:19:06.870 --> 00:19:07.770 align:middle line:90% OK. 00:19:07.770 --> 00:19:11.690 align:middle line:84% So now with those concepts, the statement and balance sheet, 00:19:11.690 --> 00:19:13.890 align:middle line:90% would take it to an application. 00:19:13.890 --> 00:19:16.250 align:middle line:84% So this is a wealth planner, which 00:19:16.250 --> 00:19:21.080 align:middle line:84% we implemented to that household and all of the other ones. 00:19:21.080 --> 00:19:24.710 align:middle line:84% What is the household we imagine to be doing? 00:19:24.710 --> 00:19:28.160 align:middle line:84% Namely, maximize discounted expected utility. 00:19:28.160 --> 00:19:30.590 align:middle line:84% So you've kind of seen versions of this before. 00:19:30.590 --> 00:19:34.250 align:middle line:84% You saw it first with respect to storage of seed 00:19:34.250 --> 00:19:37.620 align:middle line:84% and so on in the medieval village economy. 00:19:37.620 --> 00:19:40.610 align:middle line:84% We've got a finite horizon here that happens, 00:19:40.610 --> 00:19:42.660 align:middle line:90% but more on that momentarily. 00:19:42.660 --> 00:19:44.720 align:middle line:90% So it's discounted utility. 00:19:44.720 --> 00:19:48.260 align:middle line:84% This is constant relative risk aversion, with a risk averse 00:19:48.260 --> 00:19:50.450 align:middle line:90% coefficient of kappa. 00:19:50.450 --> 00:19:52.580 align:middle line:90% Delta is the discount rate. 00:19:52.580 --> 00:19:56.060 align:middle line:84% And the end of life, which is T, they 00:19:56.060 --> 00:19:59.250 align:middle line:84% may bequest assets to their kids. 00:19:59.250 --> 00:20:01.010 align:middle line:84% It looks like a similar function, 00:20:01.010 --> 00:20:02.600 align:middle line:84% but there's an extra term here, which 00:20:02.600 --> 00:20:05.060 align:middle line:90% is beta, almost for bequest. 00:20:05.060 --> 00:20:09.080 align:middle line:84% And again, the notation changes from one lecture to the next. 00:20:09.080 --> 00:20:11.060 align:middle line:84% But anyway, so that's another parameter. 00:20:11.060 --> 00:20:12.920 align:middle line:90% We have delta, kappa, and beta. 00:20:12.920 --> 00:20:15.420 align:middle line:84% We're going to have to figure out what they are because they 00:20:15.420 --> 00:20:16.940 align:middle line:90% will impact the solution. 00:20:16.940 --> 00:20:19.160 align:middle line:84% Subject to the big constraint-- now, 00:20:19.160 --> 00:20:22.520 align:middle line:84% again, you've seen a version of this in the medieval village 00:20:22.520 --> 00:20:26.240 align:middle line:84% economy because we had resources available today, which 00:20:26.240 --> 00:20:30.500 align:middle line:84% they could eat or store in two different ways, which 00:20:30.500 --> 00:20:33.470 align:middle line:84% would determine resources available next period. 00:20:33.470 --> 00:20:36.170 align:middle line:84% Here, they have wealth today, and they 00:20:36.170 --> 00:20:38.750 align:middle line:84% have resources available today, which 00:20:38.750 --> 00:20:42.230 align:middle line:84% is the beginning of period stocks, their wealth, 00:20:42.230 --> 00:20:43.400 align:middle line:90% plus income-- 00:20:43.400 --> 00:20:45.120 align:middle line:90% more on that momentarily. 00:20:45.120 --> 00:20:47.250 align:middle line:84% And they can either eat it-- and the difference 00:20:47.250 --> 00:20:49.940 align:middle line:84% between what they have and what they eat, they invest. 00:20:49.940 --> 00:20:54.080 align:middle line:84% So some fraction, Z, goes into risky assets, 00:20:54.080 --> 00:20:58.310 align:middle line:84% and the residual 1 minus Z fraction goes into safe assets. 00:20:58.310 --> 00:21:03.380 align:middle line:84% So this is a lifecycle portfolio allocation problem. 00:21:03.380 --> 00:21:05.840 align:middle line:84% A little bit more detail about these numbers-- 00:21:05.840 --> 00:21:09.290 align:middle line:84% so the income term is the second and third term. 00:21:09.290 --> 00:21:13.460 align:middle line:84% If gamma is equal to 1, then the mu, which is the mean income, 00:21:13.460 --> 00:21:14.460 align:middle line:90% cancels out. 00:21:14.460 --> 00:21:17.180 align:middle line:84% And this household is entirely on its own 00:21:17.180 --> 00:21:20.330 align:middle line:84% with income fluctuations around the mean. 00:21:20.330 --> 00:21:24.270 align:middle line:84% When gamma is equal to 0, this whole term is gone, 00:21:24.270 --> 00:21:26.670 align:middle line:84% and the household has its mean income. 00:21:26.670 --> 00:21:31.370 align:middle line:84% So gamma here, between 0 and 1, is capturing the risk-sharing 00:21:31.370 --> 00:21:34.340 align:middle line:84% that we know is prevalent in the Thai villages. 00:21:34.340 --> 00:21:36.740 align:middle line:84% So that's the way all that risk-sharing stuff entered 00:21:36.740 --> 00:21:38.180 align:middle line:90% into the individual program. 00:21:38.180 --> 00:21:41.420 align:middle line:84% It looks like self insurance, but it isn't. 00:21:41.420 --> 00:21:45.800 align:middle line:84% Because the income stream can be attenuated due to the gifts. 00:21:45.800 --> 00:21:48.050 align:middle line:84% But it can also vary over households. 00:21:48.050 --> 00:21:50.360 align:middle line:84% We have to elicit from the households 00:21:50.360 --> 00:21:52.220 align:middle line:90% how big or small gamma is. 00:21:52.220 --> 00:21:55.130 align:middle line:90% This is a premium if the-- 00:21:55.130 --> 00:21:59.150 align:middle line:84% sorry-- an indemnity if the household head dies, 00:21:59.150 --> 00:22:02.480 align:middle line:84% all these village villagers get together and contribute 00:22:02.480 --> 00:22:05.780 align:middle line:84% something to help the bereaved household with expenses 00:22:05.780 --> 00:22:07.730 align:middle line:90% and so on. c is consumption. 00:22:07.730 --> 00:22:09.990 align:middle line:90% Big C is committed consumption. 00:22:09.990 --> 00:22:12.680 align:middle line:84% So these are things the household tells us they're 00:22:12.680 --> 00:22:14.250 align:middle line:90% going to do no matter what. 00:22:14.250 --> 00:22:18.260 align:middle line:84% And so we enter as a particular expense. 00:22:18.260 --> 00:22:18.800 align:middle line:90% All right. 00:22:18.800 --> 00:22:22.640 align:middle line:84% So here is a bit of what we show the household. 00:22:22.640 --> 00:22:24.110 align:middle line:90% We explain to them-- 00:22:24.110 --> 00:22:26.120 align:middle line:90% we go and interview them. 00:22:26.120 --> 00:22:29.240 align:middle line:84% And that we have been getting data from them all along, 00:22:29.240 --> 00:22:30.830 align:middle line:90% but this is extra. 00:22:30.830 --> 00:22:33.740 align:middle line:84% So we tell them we want to know how much they're 00:22:33.740 --> 00:22:36.530 align:middle line:84% planning to give to their kids, if anything. 00:22:36.530 --> 00:22:39.350 align:middle line:84% Try to elicit their current income profile 00:22:39.350 --> 00:22:41.420 align:middle line:90% and future movements and income. 00:22:41.420 --> 00:22:45.775 align:middle line:84% Get their goals, and get their current wealth. 00:22:45.775 --> 00:22:47.150 align:middle line:84% And then we're going to tell them 00:22:47.150 --> 00:22:50.180 align:middle line:84% how much they should be saving, and how to save it 00:22:50.180 --> 00:22:52.430 align:middle line:90% in risky versus safe things. 00:22:52.430 --> 00:22:55.788 align:middle line:84% And particular, whether they might want to be borrowing. 00:22:55.788 --> 00:22:56.330 align:middle line:90% I don't know. 00:22:56.330 --> 00:22:58.880 align:middle line:84% Some people don't like these little cartoon characters. 00:22:58.880 --> 00:23:02.180 align:middle line:84% It's an attempt to communicate clearly. 00:23:02.180 --> 00:23:04.100 align:middle line:84% So we don't say, tell us everything. 00:23:04.100 --> 00:23:07.040 align:middle line:84% We say, let us demonstrate to you how this works. 00:23:07.040 --> 00:23:10.820 align:middle line:84% We'll show you the results of another villager. 00:23:10.820 --> 00:23:13.250 align:middle line:84% We asked that household whether or not 00:23:13.250 --> 00:23:15.080 align:middle line:84% they want to leave money to their kids. 00:23:15.080 --> 00:23:17.540 align:middle line:90% And this household said, no. 00:23:17.540 --> 00:23:19.850 align:middle line:84% My kids are going to be much richer than I am. 00:23:19.850 --> 00:23:21.650 align:middle line:84% They're going to work in Bangkok. 00:23:21.650 --> 00:23:25.100 align:middle line:84% I don't think I need to leave anything to my children 00:23:25.100 --> 00:23:26.760 align:middle line:90% after my death. 00:23:26.760 --> 00:23:29.780 align:middle line:84% So this household said zero baht. 00:23:29.780 --> 00:23:31.340 align:middle line:84% I made the mistake once of having 00:23:31.340 --> 00:23:34.100 align:middle line:84% this interview with the mother and the daughter 00:23:34.100 --> 00:23:36.770 align:middle line:84% present in the room at the same time. 00:23:36.770 --> 00:23:39.072 align:middle line:90% Obviously, that was awkward. 00:23:39.072 --> 00:23:41.390 align:middle line:90% [LAUGHS] OK. 00:23:41.390 --> 00:23:45.500 align:middle line:84% Collect inputs-- so we need to calibrate this income process. 00:23:45.500 --> 00:23:48.973 align:middle line:90% And we give, for example, Chen. 00:23:48.973 --> 00:23:50.390 align:middle line:84% Now you're going to see some Thai. 00:23:50.390 --> 00:23:50.930 align:middle line:90% Sorry. 00:23:50.930 --> 00:23:56.600 align:middle line:84% Anyway, wage earners earn a regular income of 10,000 baht. 00:23:56.600 --> 00:24:01.190 align:middle line:84% Expects to work to age 65, and then income 00:24:01.190 --> 00:24:04.520 align:middle line:84% drops to something very small, which is a pension. 00:24:04.520 --> 00:24:05.930 align:middle line:90% So that's this picture. 00:24:05.930 --> 00:24:08.390 align:middle line:84% 10,000 baht, age 65, the household's 00:24:08.390 --> 00:24:10.380 align:middle line:90% currently young, actually. 00:24:10.380 --> 00:24:11.160 align:middle line:90% Drops to zero. 00:24:11.160 --> 00:24:13.740 align:middle line:84% Or this guy-- different example-- 00:24:13.740 --> 00:24:15.960 align:middle line:84% has a little shop as a business-- 00:24:15.960 --> 00:24:18.780 align:middle line:90% very volatile income. 00:24:18.780 --> 00:24:22.080 align:middle line:84% Expects to keep its business running until age 60-- 00:24:22.080 --> 00:24:24.720 align:middle line:90% currently age 25. 00:24:24.720 --> 00:24:28.680 align:middle line:84% And then receive some small income, again, from the kids. 00:24:28.680 --> 00:24:29.250 align:middle line:90% Oh, yeah. 00:24:29.250 --> 00:24:31.930 align:middle line:84% Kids give money to the parents, you guys. 00:24:31.930 --> 00:24:33.330 align:middle line:90% You hear that? 00:24:33.330 --> 00:24:37.110 align:middle line:90% [LAUGHS] 00:24:37.110 --> 00:24:38.600 align:middle line:90% OK. 00:24:38.600 --> 00:24:41.770 align:middle line:84% Depends on where you're from whether you can relate to this. 00:24:41.770 --> 00:24:46.480 align:middle line:84% Another example-- this household has regular increasing income 00:24:46.480 --> 00:24:48.050 align:middle line:90% for 30 years. 00:24:48.050 --> 00:24:50.140 align:middle line:84% But, again, income starts to drop. 00:24:50.140 --> 00:24:53.050 align:middle line:84% However, they're going to go into a second occupation, 00:24:53.050 --> 00:24:54.760 align:middle line:90% and income only slowly. 00:24:54.760 --> 00:24:58.330 align:middle line:84% So I stay-- we used to stay in the village with the headman. 00:24:58.330 --> 00:25:01.090 align:middle line:84% But eventually, Thailand has motels. 00:25:01.090 --> 00:25:05.260 align:middle line:84% And you'll often see an elderly couple running the motel. 00:25:05.260 --> 00:25:07.900 align:middle line:84% That's now their new business, relative to what 00:25:07.900 --> 00:25:09.880 align:middle line:90% they used to do, et cetera. 00:25:09.880 --> 00:25:12.350 align:middle line:84% So these are just three examples. 00:25:12.350 --> 00:25:14.730 align:middle line:90% So what about goals? 00:25:14.730 --> 00:25:17.890 align:middle line:84% So do you expect to send your kids to college? 00:25:17.890 --> 00:25:19.870 align:middle line:90% Is that a planned expense? 00:25:19.870 --> 00:25:24.400 align:middle line:84% Or in the future open a business or buy a new TV? 00:25:24.400 --> 00:25:27.040 align:middle line:84% But what we want are expenses they're 00:25:27.040 --> 00:25:28.400 align:middle line:90% going to be no matter what. 00:25:28.400 --> 00:25:31.420 align:middle line:90% So yes or no, depending. 00:25:31.420 --> 00:25:33.940 align:middle line:84% In this particular example household, 00:25:33.940 --> 00:25:35.830 align:middle line:90% we give them car purchase. 00:25:35.830 --> 00:25:37.000 align:middle line:90% Was that obligatory? 00:25:37.000 --> 00:25:37.600 align:middle line:90% No. 00:25:37.600 --> 00:25:39.040 align:middle line:90% Send the kids to college? 00:25:39.040 --> 00:25:40.090 align:middle line:90% Yes. 00:25:40.090 --> 00:25:41.350 align:middle line:90% How many years away? 00:25:41.350 --> 00:25:43.720 align:middle line:90% 15 years-- kids are young. 00:25:43.720 --> 00:25:47.560 align:middle line:84% And it's going to cost $10,000 for three years-- 00:25:47.560 --> 00:25:49.300 align:middle line:90% 30,000 baht. 00:25:49.300 --> 00:25:52.990 align:middle line:84% That's about-- it's a lot less than US expenses. 00:25:52.990 --> 00:25:55.180 align:middle line:90% It's not an MIT tuition. 00:25:55.180 --> 00:25:55.780 align:middle line:90% All right. 00:25:55.780 --> 00:25:59.020 align:middle line:90% So now we have wealth, so w. 00:25:59.020 --> 00:26:00.010 align:middle line:90% This is tricky. 00:26:00.010 --> 00:26:03.850 align:middle line:84% Money available to save or spend or invest. 00:26:03.850 --> 00:26:06.910 align:middle line:84% So if you have a car that's worth 10,000 baht. 00:26:06.910 --> 00:26:10.130 align:middle line:84% You think-- it's used, but you could sell it easily. 00:26:10.130 --> 00:26:13.060 align:middle line:84% Then we would count that as a household asset or business 00:26:13.060 --> 00:26:14.710 align:middle line:90% asset-- could be sold. 00:26:14.710 --> 00:26:18.010 align:middle line:84% If they own land worth 100,000 baht, 00:26:18.010 --> 00:26:20.920 align:middle line:84% they may never plan to sell it, and may not even 00:26:20.920 --> 00:26:22.480 align:middle line:90% be productive land. 00:26:22.480 --> 00:26:25.570 align:middle line:84% So we don't count it unless they said it was. 00:26:25.570 --> 00:26:30.410 align:middle line:84% If it's productive, we want to avoid double counting. 00:26:30.410 --> 00:26:31.660 align:middle line:90% So here's the thing. 00:26:31.660 --> 00:26:33.220 align:middle line:84% You would count the asset, or you 00:26:33.220 --> 00:26:34.930 align:middle line:90% would count the return stream. 00:26:34.930 --> 00:26:36.775 align:middle line:90% You cannot count both. 00:26:36.775 --> 00:26:42.610 align:middle line:84% The value of land should be the discounted capitalized value 00:26:42.610 --> 00:26:45.100 align:middle line:90% of net profits from farming. 00:26:45.100 --> 00:26:47.380 align:middle line:84% That's how much the land is worth. 00:26:47.380 --> 00:26:49.960 align:middle line:84% You don't want to count both the land as an asset, 00:26:49.960 --> 00:26:53.170 align:middle line:84% and also ask the households about the income stream. 00:26:53.170 --> 00:26:55.600 align:middle line:84% So generally, we don't count assets 00:26:55.600 --> 00:26:59.050 align:middle line:84% that we are using in calculating the income stream unless they 00:26:59.050 --> 00:27:02.200 align:middle line:84% say, I'm going to stop farming at a certain date. 00:27:02.200 --> 00:27:04.180 align:middle line:84% And how much would the land be worth then? 00:27:04.180 --> 00:27:07.780 align:middle line:84% If we count income up to that point, then we count the land. 00:27:07.780 --> 00:27:09.910 align:middle line:90% So let's run the planner. 00:27:09.910 --> 00:27:11.320 align:middle line:90% I'll give you examples. 00:27:11.320 --> 00:27:14.260 align:middle line:84% We created an Excel spreadsheet with the answers 00:27:14.260 --> 00:27:16.360 align:middle line:84% to all the household's questions, 00:27:16.360 --> 00:27:20.060 align:middle line:84% and then enter it into the computer and run it. 00:27:20.060 --> 00:27:21.420 align:middle line:90% I'm not showing you all of them. 00:27:21.420 --> 00:27:23.840 align:middle line:84% The second household-- we told this household 00:27:23.840 --> 00:27:27.360 align:middle line:84% to cut its expenses on consumption quite a bit. 00:27:27.360 --> 00:27:30.290 align:middle line:84% They weren't saving enough, and to put almost all of it 00:27:30.290 --> 00:27:32.090 align:middle line:90% in safe assets. 00:27:32.090 --> 00:27:35.370 align:middle line:84% More interesting is the life cycle picture. 00:27:35.370 --> 00:27:37.670 align:middle line:84% So here's what the planner recommends 00:27:37.670 --> 00:27:39.980 align:middle line:90% that the household should do. 00:27:39.980 --> 00:27:43.100 align:middle line:90% Namely, let's start down here. 00:27:43.100 --> 00:27:45.050 align:middle line:90% We've got income in red. 00:27:45.050 --> 00:27:48.590 align:middle line:84% That's the elicited profile over time 00:27:48.590 --> 00:27:50.390 align:middle line:90% that I just told you about. 00:27:50.390 --> 00:27:54.350 align:middle line:84% The consumption is endogenous variable, obviously. 00:27:54.350 --> 00:27:55.490 align:middle line:90% It's in black. 00:27:55.490 --> 00:27:58.550 align:middle line:84% So here, we told the household to cut its consumption 00:27:58.550 --> 00:28:02.450 align:middle line:84% expenditure a lot and much lower than its current income 00:28:02.450 --> 00:28:04.770 align:middle line:90% in order to build up the assets. 00:28:04.770 --> 00:28:06.890 align:middle line:84% So here's the total level of assets. 00:28:06.890 --> 00:28:11.570 align:middle line:84% And then when income drops to little or nothing, 00:28:11.570 --> 00:28:13.410 align:middle line:90% they eat off the assets. 00:28:13.410 --> 00:28:15.800 align:middle line:84% So this is just saving for retirement. 00:28:15.800 --> 00:28:19.100 align:middle line:84% Most of it's in safe assets, not so much in risky, 00:28:19.100 --> 00:28:20.660 align:middle line:90% and there's no borrowing. 00:28:20.660 --> 00:28:24.710 align:middle line:84% Another household basically had committed consumption 00:28:24.710 --> 00:28:26.930 align:middle line:84% that is going to get it in big trouble, 00:28:26.930 --> 00:28:30.770 align:middle line:84% even though they said they were going to do it no matter what. 00:28:30.770 --> 00:28:33.410 align:middle line:84% When we ran the planner at the time 00:28:33.410 --> 00:28:35.300 align:middle line:84% that they have this high expense-- 00:28:35.300 --> 00:28:37.880 align:middle line:84% I think it was for starting a business-- they really didn't 00:28:37.880 --> 00:28:39.650 align:middle line:90% have the resources to cover it. 00:28:39.650 --> 00:28:44.270 align:middle line:84% Now, they try to increase their savings 00:28:44.270 --> 00:28:47.930 align:middle line:90% with income over consumption. 00:28:47.930 --> 00:28:54.270 align:middle line:84% But in this median simulation scenario, it's just not enough. 00:28:54.270 --> 00:28:57.590 align:middle line:84% So when that big expense comes, on average, they 00:28:57.590 --> 00:29:00.320 align:middle line:90% use up all of their assets. 00:29:00.320 --> 00:29:02.910 align:middle line:84% Consumption isn't dropping to zero. 00:29:02.910 --> 00:29:05.070 align:middle line:84% You remember, it's constant relative risk aversion. 00:29:05.070 --> 00:29:08.240 align:middle line:84% So marginal utility of consumption at zero 00:29:08.240 --> 00:29:09.140 align:middle line:90% is infinite. 00:29:09.140 --> 00:29:12.050 align:middle line:84% Although, you can't see this, but we gave them 00:29:12.050 --> 00:29:14.000 align:middle line:90% some subsistence bundle. 00:29:14.000 --> 00:29:15.650 align:middle line:84% So they don't really go bankrupt. 00:29:15.650 --> 00:29:16.670 align:middle line:90% They survive. 00:29:16.670 --> 00:29:19.520 align:middle line:84% Well, they don't die when they go bankrupt. 00:29:19.520 --> 00:29:24.080 align:middle line:84% They continue to survive because of that subsistence bundle. 00:29:24.080 --> 00:29:25.850 align:middle line:90% What else is going on here? 00:29:25.850 --> 00:29:27.710 align:middle line:90% They kind of know in advance-- 00:29:27.710 --> 00:29:31.550 align:middle line:84% the planner knows-- state by state, year by year 00:29:31.550 --> 00:29:33.780 align:middle line:90% that they're in trouble. 00:29:33.780 --> 00:29:36.320 align:middle line:84% So actually, they start borrowing. 00:29:36.320 --> 00:29:39.200 align:middle line:84% Of course, given our subsistence consumption rule, 00:29:39.200 --> 00:29:42.050 align:middle line:84% if the median scenario occurs, they're 00:29:42.050 --> 00:29:43.550 align:middle line:90% not going to pay back the loan. 00:29:43.550 --> 00:29:46.190 align:middle line:84% They get to default on the loan, and they know that. 00:29:46.190 --> 00:29:48.410 align:middle line:84% Call this gambling for redemption. 00:29:48.410 --> 00:29:51.350 align:middle line:84% They're starting to do risky things too, 00:29:51.350 --> 00:29:54.860 align:middle line:84% and borrowing, leveraging up, with the idea 00:29:54.860 --> 00:29:57.860 align:middle line:84% that there's a good chance they're not going to make it. 00:29:57.860 --> 00:30:00.680 align:middle line:84% Here, the planner is not the bank. 00:30:00.680 --> 00:30:04.010 align:middle line:84% Whether the bank would make this loan is another issue. 00:30:04.010 --> 00:30:06.320 align:middle line:84% Anyway, they survive this episode 00:30:06.320 --> 00:30:10.080 align:middle line:84% and continue to live another day with their fluctuating income. 00:30:10.080 --> 00:30:12.350 align:middle line:84% So the rest of this looks like the other diagram 00:30:12.350 --> 00:30:13.670 align:middle line:90% to some extent. 00:30:13.670 --> 00:30:14.180 align:middle line:90% OK. 00:30:14.180 --> 00:30:16.640 align:middle line:84% So we show the households these pictures, 00:30:16.640 --> 00:30:19.790 align:middle line:84% and try to explain to them what's going on. 00:30:19.790 --> 00:30:22.230 align:middle line:84% And there are all kinds of households in the data. 00:30:22.230 --> 00:30:25.040 align:middle line:84% And some of them, in fact, are very conservative, 00:30:25.040 --> 00:30:27.710 align:middle line:84% and they have fluctuations in income and consumption. 00:30:27.710 --> 00:30:31.640 align:middle line:84% Consumption is never more than the lowest possible income 00:30:31.640 --> 00:30:35.000 align:middle line:84% draw, and mostly, a lot lower than their income draw. 00:30:35.000 --> 00:30:38.450 align:middle line:84% So they're saving like crazy, and their net worth 00:30:38.450 --> 00:30:40.910 align:middle line:90% is growing at 20% per year. 00:30:40.910 --> 00:30:42.350 align:middle line:90% Sounds good to me, by the way. 00:30:42.350 --> 00:30:44.840 align:middle line:84% But they're probably over-saving. 00:30:44.840 --> 00:30:48.080 align:middle line:84% But we're very conservative about what we tell them. 00:30:48.080 --> 00:30:50.480 align:middle line:84% We don't tell them that that's bad. 00:30:50.480 --> 00:30:52.070 align:middle line:84% That's fine if they want to do that. 00:30:52.070 --> 00:30:53.930 align:middle line:84% It's the ones who are eating too much 00:30:53.930 --> 00:30:57.080 align:middle line:84% and not saving that we're more worried about. 00:30:57.080 --> 00:30:57.980 align:middle line:90% OK. 00:30:57.980 --> 00:30:59.990 align:middle line:84% So this looks like Thai villages, 00:30:59.990 --> 00:31:02.690 align:middle line:84% but the life cycle model is universal. 00:31:02.690 --> 00:31:09.200 align:middle line:84% And just to shift gears, this is Larry Kotlikoff's website. 00:31:09.200 --> 00:31:12.240 align:middle line:84% He's both a professor and an entrepreneur. 00:31:12.240 --> 00:31:17.540 align:middle line:84% So he is selling this product, this ESB planner. 00:31:17.540 --> 00:31:21.050 align:middle line:84% But he has a lot of great economic research on life cycle 00:31:21.050 --> 00:31:22.070 align:middle line:90% planning. 00:31:22.070 --> 00:31:25.970 align:middle line:84% And no one wants to splurge today and starve 00:31:25.970 --> 00:31:28.340 align:middle line:90% tomorrow, nor the opposite. 00:31:28.340 --> 00:31:32.850 align:middle line:84% Lifecycle consumption-smoothing balances our goals. 00:31:32.850 --> 00:31:34.930 align:middle line:84% Now, you would think, could everyone do that? 00:31:34.930 --> 00:31:36.650 align:middle line:90% The answer is no. 00:31:36.650 --> 00:31:40.430 align:middle line:84% There's only two companies I know of in the US that do this. 00:31:40.430 --> 00:31:41.780 align:middle line:90% And it's not Vanguard. 00:31:41.780 --> 00:31:43.280 align:middle line:90% It's not Fidelity. 00:31:43.280 --> 00:31:45.650 align:middle line:84% They're adopting kind of a rule of thumb. 00:31:45.650 --> 00:31:48.860 align:middle line:84% They ask you what's your current expenses? 00:31:48.860 --> 00:31:54.030 align:middle line:84% And then say 70% of that is target for retirement. 00:31:54.030 --> 00:31:57.810 align:middle line:84% And then how much do you have to save to make that goal? 00:31:57.810 --> 00:32:01.610 align:middle line:84% And so consumption could be low because you're saving a lot. 00:32:01.610 --> 00:32:04.550 align:middle line:84% And that can end up in a very imbalanced profile 00:32:04.550 --> 00:32:08.540 align:middle line:84% in which you're starving today in order to splurge tomorrow. 00:32:08.540 --> 00:32:13.020 align:middle line:84% Whereas, we're using this strictly concave utility, 00:32:13.020 --> 00:32:16.130 align:middle line:84% and so is Larry, to say we want to smooth over time. 00:32:16.130 --> 00:32:18.800 align:middle line:84% Just like we want to smooth over bad shocks, 00:32:18.800 --> 00:32:23.610 align:middle line:84% we want to smooth over time, as in the medieval village seed 00:32:23.610 --> 00:32:25.220 align:middle line:90% and so on. 00:32:25.220 --> 00:32:27.110 align:middle line:90% OK. 00:32:27.110 --> 00:32:31.940 align:middle line:84% So let me make a transition then to the other account, which 00:32:31.940 --> 00:32:32.660 align:middle line:90% is cash flow. 00:32:32.660 --> 00:32:35.000 align:middle line:84% And I'm going to do this by explaining a little bit 00:32:35.000 --> 00:32:38.330 align:middle line:84% about Bitcoin, and many of you probably know about this 00:32:38.330 --> 00:32:38.910 align:middle line:90% already. 00:32:38.910 --> 00:32:44.510 align:middle line:84% So every transection in Bitcoin is an encoded message sent out 00:32:44.510 --> 00:32:48.020 align:middle line:84% with a-- generated with a private/public key scheme, 00:32:48.020 --> 00:32:51.750 align:middle line:84% which has the advantage that it's known 00:32:51.750 --> 00:32:53.840 align:middle line:90% which node sent the message. 00:32:53.840 --> 00:32:55.940 align:middle line:84% And there's full commitment to the message. 00:32:55.940 --> 00:32:58.160 align:middle line:84% You can't go back and deny that you sent it. 00:32:58.160 --> 00:33:00.090 align:middle line:90% And it prevents fraud. 00:33:00.090 --> 00:33:02.990 align:middle line:84% Nobody can transact on your ID because nobody else 00:33:02.990 --> 00:33:06.440 align:middle line:84% has the private key, unless you get careless. 00:33:06.440 --> 00:33:10.430 align:middle line:84% So a transection in Bitcoin is the moving 00:33:10.430 --> 00:33:13.980 align:middle line:84% of the coin from one person to another. 00:33:13.980 --> 00:33:16.250 align:middle line:84% The message sent might be, I want 00:33:16.250 --> 00:33:19.580 align:middle line:84% to give the coin to person X. Arguably, 00:33:19.580 --> 00:33:23.000 align:middle line:84% because person X is a merchant and you're buying something 00:33:23.000 --> 00:33:24.440 align:middle line:90% or whatever. 00:33:24.440 --> 00:33:28.640 align:middle line:84% The various transactions are accumulated into a block, 00:33:28.640 --> 00:33:32.450 align:middle line:84% and the block is subject to a validation routine, which is 00:33:32.450 --> 00:33:35.600 align:middle line:90% mining and it's quite costly. 00:33:35.600 --> 00:33:37.560 align:middle line:90% I'll come back to that later. 00:33:37.560 --> 00:33:39.320 align:middle line:84% The point here is that Bitcoin is 00:33:39.320 --> 00:33:41.340 align:middle line:90% combining two different things. 00:33:41.340 --> 00:33:43.100 align:middle line:84% One is the statement of cash flow 00:33:43.100 --> 00:33:48.320 align:middle line:84% as the coin moves from out or in to the person in question. 00:33:48.320 --> 00:33:50.010 align:middle line:84% And the other is the balance sheet, 00:33:50.010 --> 00:33:55.250 align:middle line:84% which is at a moment in time how many coins do you own? 00:33:55.250 --> 00:33:58.940 align:middle line:84% So the accounting of the cash flow works for Bitcoin. 00:33:58.940 --> 00:34:02.360 align:middle line:84% It also works for paper currency in Thailand. 00:34:02.360 --> 00:34:04.745 align:middle line:84% So let me show you a statement of cash flow 00:34:04.745 --> 00:34:05.870 align:middle line:90% as an accounting statement. 00:34:05.870 --> 00:34:08.510 align:middle line:84% Remember, before, I showed you income and balance sheet. 00:34:08.510 --> 00:34:12.050 align:middle line:84% I haven't shown you the cash flow statement until now. 00:34:12.050 --> 00:34:13.520 align:middle line:90% It's pretty small print. 00:34:13.520 --> 00:34:15.830 align:middle line:84% Let me just give you the big picture here. 00:34:15.830 --> 00:34:17.940 align:middle line:84% We have cash flow from production, 00:34:17.940 --> 00:34:20.360 align:middle line:84% which you can imagine, if you sold 00:34:20.360 --> 00:34:23.600 align:middle line:84% something in got paper currency as a consequence, 00:34:23.600 --> 00:34:25.040 align:middle line:90% you'd have a plus. 00:34:25.040 --> 00:34:27.380 align:middle line:84% If you're buying fertilizer to grow the rice, 00:34:27.380 --> 00:34:29.639 align:middle line:90% you have a minus sign. 00:34:29.639 --> 00:34:33.050 align:middle line:84% So you have inflows and outflows associated with production. 00:34:33.050 --> 00:34:35.090 align:middle line:84% You can spend money on consumption. 00:34:35.090 --> 00:34:36.409 align:middle line:90% That's typically an outflow. 00:34:36.409 --> 00:34:38.870 align:middle line:84% You can spend money on investment, also 00:34:38.870 --> 00:34:40.800 align:middle line:90% typically an outflow. 00:34:40.800 --> 00:34:43.460 align:middle line:84% So we have cash flow from consumption and investment, 00:34:43.460 --> 00:34:46.469 align:middle line:84% and then we have cash flow from financing, 00:34:46.469 --> 00:34:49.580 align:middle line:84% which is, you can put cash in the bank as a deposit, 00:34:49.580 --> 00:34:51.050 align:middle line:90% or you could borrow cash. 00:34:51.050 --> 00:34:52.969 align:middle line:84% Hopefully, you get it as an inflow, 00:34:52.969 --> 00:34:54.420 align:middle line:90% through gifts and so on. 00:34:54.420 --> 00:34:59.780 align:middle line:84% So cash flow from production, consumption, and financing 00:34:59.780 --> 00:35:01.500 align:middle line:90% are the three categories. 00:35:01.500 --> 00:35:03.230 align:middle line:84% And this stuff up here is complicated. 00:35:03.230 --> 00:35:06.620 align:middle line:84% When I said I'd shown you the income statement, but not 00:35:06.620 --> 00:35:09.800 align:middle line:84% the cash flow statement, you may have been thinking, well, 00:35:09.800 --> 00:35:11.550 align:middle line:90% aren't they the same? 00:35:11.550 --> 00:35:14.870 align:middle line:84% And the answer is, well, they could be, but not necessarily. 00:35:14.870 --> 00:35:18.410 align:middle line:84% So if you're a business, and you sell something and the guy 00:35:18.410 --> 00:35:21.530 align:middle line:84% doesn't pay you, then he's going to pay later, 00:35:21.530 --> 00:35:25.100 align:middle line:84% that's an account receivable, but there's no cash. 00:35:25.100 --> 00:35:27.470 align:middle line:84% So the sale would count as revenue 00:35:27.470 --> 00:35:29.690 align:middle line:84% in the income statement using the accrued 00:35:29.690 --> 00:35:31.520 align:middle line:84% notion of accounting, but it would not 00:35:31.520 --> 00:35:33.090 align:middle line:90% count as a cash inflow. 00:35:33.090 --> 00:35:35.480 align:middle line:84% So oddly enough, this top stuff here 00:35:35.480 --> 00:35:37.490 align:middle line:84% is just starting with the income statement 00:35:37.490 --> 00:35:39.650 align:middle line:84% and making adjustments to it according 00:35:39.650 --> 00:35:42.140 align:middle line:84% to whether the movement of cash you 00:35:42.140 --> 00:35:44.550 align:middle line:84% have to adjust one way or the other. 00:35:44.550 --> 00:35:47.030 align:middle line:84% So, I mean, if you have an account receivable, 00:35:47.030 --> 00:35:48.860 align:middle line:84% and you didn't get the cash, then you 00:35:48.860 --> 00:35:51.260 align:middle line:84% have to subtract off from accrued income 00:35:51.260 --> 00:35:53.630 align:middle line:84% to get the cash flow because you're 00:35:53.630 --> 00:35:55.700 align:middle line:84% counting something that is not yet associated 00:35:55.700 --> 00:35:57.510 align:middle line:90% with the cash inflow. 00:35:57.510 --> 00:35:58.010 align:middle line:90% All right. 00:35:58.010 --> 00:36:00.590 align:middle line:84% Mainly, I'm just trying to give you the big picture here. 00:36:00.590 --> 00:36:02.730 align:middle line:84% This is the statement of cash flow. 00:36:02.730 --> 00:36:06.980 align:middle line:84% And it is true, and this is automatically true in Bitcoin, 00:36:06.980 --> 00:36:11.060 align:middle line:84% that you can check whether the change in cash holdings that 00:36:11.060 --> 00:36:14.060 align:middle line:84% come at the bottom line of the statement of cash flow 00:36:14.060 --> 00:36:17.720 align:middle line:84% are exactly equal or not to the change in cash 00:36:17.720 --> 00:36:19.650 align:middle line:84% that we have on the balance sheet. 00:36:19.650 --> 00:36:21.500 align:middle line:84% So you have cash in a moment in time, 00:36:21.500 --> 00:36:23.750 align:middle line:84% and then you have these inflows and outflows. 00:36:23.750 --> 00:36:26.210 align:middle line:84% You have cash in a future period. 00:36:26.210 --> 00:36:28.610 align:middle line:84% In each of those periods now and in the future, 00:36:28.610 --> 00:36:29.850 align:middle line:90% you have two balance sheets. 00:36:29.850 --> 00:36:31.850 align:middle line:84% Take the difference in the balance sheet, that's 00:36:31.850 --> 00:36:34.350 align:middle line:84% the change in the cash flow on the balance sheet, 00:36:34.350 --> 00:36:36.050 align:middle line:84% and it should be exactly equal to what's 00:36:36.050 --> 00:36:39.170 align:middle line:84% showing up here in the cash flow from the statement of cash 00:36:39.170 --> 00:36:40.230 align:middle line:90% flow. 00:36:40.230 --> 00:36:42.380 align:middle line:84% So we run a check on the Thai data. 00:36:42.380 --> 00:36:45.620 align:middle line:84% And, again, if we're making a mistake these are not equal, 00:36:45.620 --> 00:36:47.570 align:middle line:90% and that's a big guide to us. 00:36:47.570 --> 00:36:49.190 align:middle line:84% I should save this for later, but I 00:36:49.190 --> 00:36:53.570 align:middle line:84% will tell you that no top US survey does this. 00:36:53.570 --> 00:36:58.730 align:middle line:84% Not the Survey of Consumer Finances, not the PSID. 00:36:58.730 --> 00:37:01.130 align:middle line:84% They did not gather variables and check 00:37:01.130 --> 00:37:04.970 align:middle line:84% what they gathered against these comprehensive statements 00:37:04.970 --> 00:37:07.160 align:middle line:90% of integrated accounts. 00:37:07.160 --> 00:37:08.810 align:middle line:90% And Bitcoin is automatic. 00:37:08.810 --> 00:37:11.840 align:middle line:84% Because you either have it in your account and you spend it, 00:37:11.840 --> 00:37:15.210 align:middle line:84% or you don't spend it and it's still in the account. 00:37:15.210 --> 00:37:18.380 align:middle line:84% So that's kind of the unspoken advantage of Bitcoin. 00:37:18.380 --> 00:37:23.600 align:middle line:84% There is something that people talk about a lot on Bitcoin 00:37:23.600 --> 00:37:27.730 align:middle line:90% and that is consensus. 00:37:27.730 --> 00:37:29.700 align:middle line:84% So I keep giving this example in Bitcoin 00:37:29.700 --> 00:37:34.320 align:middle line:84% that Bob wants to transfer a coin to Alice. 00:37:34.320 --> 00:37:37.830 align:middle line:84% We have both Alice's accounts and Bob's account. 00:37:37.830 --> 00:37:41.880 align:middle line:84% And what Bob gives up is what Alice gets. 00:37:41.880 --> 00:37:45.840 align:middle line:84% So the accounts have to be consistent with each other 00:37:45.840 --> 00:37:47.610 align:middle line:90% across the households. 00:37:47.610 --> 00:37:49.620 align:middle line:90% And that is true in Bitcoin. 00:37:49.620 --> 00:37:52.080 align:middle line:84% But it is not-- well, and again, this 00:37:52.080 --> 00:37:54.480 align:middle line:84% is a bit technical for now, but the whole thing 00:37:54.480 --> 00:37:57.750 align:middle line:84% about the mining and the validation of the transactions 00:37:57.750 --> 00:38:03.300 align:middle line:84% in Bitcoin is to precisely be assured that Bob is not 00:38:03.300 --> 00:38:04.740 align:middle line:90% double-spending the coin. 00:38:04.740 --> 00:38:06.983 align:middle line:90% You can't spend it twice. 00:38:06.983 --> 00:38:08.400 align:middle line:84% And they kind of figured out a way 00:38:08.400 --> 00:38:11.230 align:middle line:84% to do that-- a very ingenious way to do that. 00:38:11.230 --> 00:38:15.450 align:middle line:84% But anyway, the concept of reconciliation and validation 00:38:15.450 --> 00:38:19.290 align:middle line:84% could be taken to accounting in the Thai villages, 00:38:19.290 --> 00:38:22.260 align:middle line:84% in the sense that we really did try to measure 00:38:22.260 --> 00:38:23.590 align:middle line:90% all the transactions. 00:38:23.590 --> 00:38:26.730 align:middle line:84% And when a household says that it bought something with cash, 00:38:26.730 --> 00:38:29.160 align:middle line:84% if the person they bought it from is in the village, 00:38:29.160 --> 00:38:31.260 align:middle line:84% then that person should get the cash, 00:38:31.260 --> 00:38:33.480 align:middle line:84% while the person in question is giving it up. 00:38:33.480 --> 00:38:37.350 align:middle line:84% So it should appear twice, both in the accounts of the person 00:38:37.350 --> 00:38:39.720 align:middle line:84% buying and the accounts of the person selling. 00:38:39.720 --> 00:38:42.780 align:middle line:84% So with one of your former classmates, 00:38:42.780 --> 00:38:46.770 align:middle line:84% who was a student in this class about two years ago, Chad, 00:38:46.770 --> 00:38:48.060 align:middle line:90% we set out-- 00:38:48.060 --> 00:38:50.760 align:middle line:84% he wanted to help on these projects. 00:38:50.760 --> 00:38:52.320 align:middle line:90% He was an RA. 00:38:52.320 --> 00:38:55.293 align:middle line:84% And I'm not pushing that hard, but if any of you 00:38:55.293 --> 00:38:56.460 align:middle line:90% are interested, let me know. 00:38:56.460 --> 00:38:58.830 align:middle line:84% There's a lot of interesting stuff to do. 00:38:58.830 --> 00:39:01.660 align:middle line:84% So Chad was very interested, so here's what he did. 00:39:01.660 --> 00:39:04.740 align:middle line:84% He looked literally to see whether what shows up 00:39:04.740 --> 00:39:07.200 align:middle line:84% in one person's account is also showing up 00:39:07.200 --> 00:39:09.240 align:middle line:90% in the other one or not. 00:39:09.240 --> 00:39:11.310 align:middle line:84% And these are some of the details 00:39:11.310 --> 00:39:14.190 align:middle line:84% of going through and doing that as an algorithm. 00:39:14.190 --> 00:39:15.480 align:middle line:90% He created these accounts. 00:39:15.480 --> 00:39:17.700 align:middle line:84% He checks to see if it's reported. 00:39:17.700 --> 00:39:19.980 align:middle line:90% If not, et cetera, et cetera. 00:39:19.980 --> 00:39:22.200 align:middle line:84% Now, right in the middle of this slide 00:39:22.200 --> 00:39:24.900 align:middle line:84% it jumps from Thailand and the thing 00:39:24.900 --> 00:39:28.860 align:middle line:84% that Chad was doing to a very different application, which is 00:39:28.860 --> 00:39:31.150 align:middle line:90% the stock market in Australia. 00:39:31.150 --> 00:39:34.050 align:middle line:84% So the stock market in Australia is 00:39:34.050 --> 00:39:38.160 align:middle line:84% about to run on distributed ledgers the way Bitcoin does. 00:39:38.160 --> 00:39:41.430 align:middle line:84% Not using Bitcoin, but using the ledgers. 00:39:41.430 --> 00:39:44.250 align:middle line:84% And so the idea is when someone owns the stock, 00:39:44.250 --> 00:39:46.020 align:middle line:84% and there's a buyer and a seller, 00:39:46.020 --> 00:39:50.730 align:middle line:84% the traditional legacy system is that they enter into a deal. 00:39:50.730 --> 00:39:54.900 align:middle line:84% And then only after a day, or maybe three, 00:39:54.900 --> 00:40:00.420 align:middle line:84% the process of reconciliation kicks in and then they settle. 00:40:00.420 --> 00:40:03.540 align:middle line:84% So the trade of the stocks is separate from the settlement 00:40:03.540 --> 00:40:04.500 align:middle line:90% of the stocks. 00:40:04.500 --> 00:40:09.410 align:middle line:84% And believe it or not, there's often huge discrepancies. 00:40:09.410 --> 00:40:11.570 align:middle line:84% For one thing, prices have moved. 00:40:11.570 --> 00:40:14.720 align:middle line:84% So a household could claim, I didn't buy it. 00:40:14.720 --> 00:40:17.210 align:middle line:84% And in the New York financial markets, 00:40:17.210 --> 00:40:20.300 align:middle line:84% there was an attempt to provide this in the repo market. 00:40:20.300 --> 00:40:23.030 align:middle line:84% Because every day in the US repo market 00:40:23.030 --> 00:40:25.460 align:middle line:84% with treasuries and so on, they spend two hours 00:40:25.460 --> 00:40:28.820 align:middle line:84% at the end of each day after the end of trading 00:40:28.820 --> 00:40:30.470 align:middle line:90% trying to reconcile accounts. 00:40:30.470 --> 00:40:35.690 align:middle line:84% State Street Bank here in Boston is a big third-party broker 00:40:35.690 --> 00:40:38.420 align:middle line:84% for people that hold accounts, and it does 00:40:38.420 --> 00:40:39.950 align:middle line:90% this kind of reconciliation. 00:40:39.950 --> 00:40:43.670 align:middle line:84% It's quite costly and quite time-consuming. 00:40:43.670 --> 00:40:47.840 align:middle line:84% But if you're entering these trades onto the ledgers, 00:40:47.840 --> 00:40:50.660 align:middle line:84% there's instantaneous reconciliation. 00:40:50.660 --> 00:40:53.400 align:middle line:84% And so they could settle right away if they wanted to. 00:40:53.400 --> 00:40:56.630 align:middle line:84% The decision to not pay for it is a credit decision. 00:40:56.630 --> 00:40:59.300 align:middle line:84% They're not forced to settle immediately. 00:40:59.300 --> 00:41:01.700 align:middle line:84% But in principle, trade and settlement 00:41:01.700 --> 00:41:04.850 align:middle line:84% can happen at exactly the same time all 00:41:04.850 --> 00:41:07.400 align:middle line:90% through the distributed ledgers. 00:41:07.400 --> 00:41:10.700 align:middle line:84% And the one in Australia is able to process 00:41:10.700 --> 00:41:13.730 align:middle line:90% 115 million daily trades. 00:41:13.730 --> 00:41:18.650 align:middle line:84% It's 6,300 trades per second continuously over five hours. 00:41:18.650 --> 00:41:22.370 align:middle line:84% That well exceeds the volume in Bitcoin. 00:41:22.370 --> 00:41:25.700 align:middle line:84% Bitcoin is not capable of doing this because of the mining. 00:41:25.700 --> 00:41:27.950 align:middle line:84% But these other ledger technologies 00:41:27.950 --> 00:41:30.750 align:middle line:84% are a lot faster and a lot cheaper. 00:41:30.750 --> 00:41:33.590 align:middle line:84% So let's come back to the village 00:41:33.590 --> 00:41:35.960 align:middle line:84% and think about how an individual household is 00:41:35.960 --> 00:41:38.150 align:middle line:90% managing its currency. 00:41:38.150 --> 00:41:40.040 align:middle line:90% Let me back up for a second. 00:41:40.040 --> 00:41:42.710 align:middle line:84% That seemed like an abrupt transition. 00:41:42.710 --> 00:41:46.700 align:middle line:84% I featured in lecture one that we will say a lot about Bitcoin 00:41:46.700 --> 00:41:50.480 align:middle line:84% and cryptocurrencies, and it's spread out 00:41:50.480 --> 00:41:51.590 align:middle line:90% throughout the lectures. 00:41:51.590 --> 00:41:54.650 align:middle line:84% So Bitcoin, blockchain, and distributed ledgers-- 00:41:54.650 --> 00:41:57.020 align:middle line:84% it's the first time it really came up. 00:41:57.020 --> 00:41:58.640 align:middle line:90% And I started with a ledger. 00:41:58.640 --> 00:42:00.920 align:middle line:84% So when we're talking about financial accounts, 00:42:00.920 --> 00:42:02.990 align:middle line:84% that's why I started to bring in Bitcoin. 00:42:02.990 --> 00:42:06.680 align:middle line:84% And we will revisit Bitcoin in other lectures 00:42:06.680 --> 00:42:09.890 align:middle line:84% when we talk about monetary economics and so on. 00:42:09.890 --> 00:42:10.400 align:middle line:90% OK. 00:42:10.400 --> 00:42:14.510 align:middle line:84% So now, back to paper currency management. 00:42:14.510 --> 00:42:17.660 align:middle line:84% I say paper currency just to constantly remind myself 00:42:17.660 --> 00:42:19.490 align:middle line:90% that cash is kind of ambiguous. 00:42:19.490 --> 00:42:22.550 align:middle line:84% When you talk about cash in the US you mean, currency, yeah. 00:42:22.550 --> 00:42:25.520 align:middle line:84% But also, demand deposits and other variables, 00:42:25.520 --> 00:42:27.500 align:middle line:90% like maybe even debit cards. 00:42:27.500 --> 00:42:29.810 align:middle line:84% But in Thailand, it's paper currency. 00:42:29.810 --> 00:42:30.560 align:middle line:90% They love it. 00:42:30.560 --> 00:42:32.030 align:middle line:90% That's what they use. 00:42:32.030 --> 00:42:33.900 align:middle line:90% We're on that momentarily. 00:42:33.900 --> 00:42:38.300 align:middle line:84% So I'm going to, in notation, describe a Miller/Orr model-- 00:42:38.300 --> 00:42:41.360 align:middle line:84% the management of cash for firms, 00:42:41.360 --> 00:42:43.490 align:middle line:84% for business firms, which do not have 00:42:43.490 --> 00:42:45.500 align:middle line:90% regular nor predictable flows. 00:42:45.500 --> 00:42:50.090 align:middle line:84% And the result solving the model is rather intuitive. 00:42:50.090 --> 00:42:51.950 align:middle line:90% There's going to be bounds. 00:42:51.950 --> 00:42:54.800 align:middle line:84% So cash holdings can drift around. 00:42:54.800 --> 00:42:57.230 align:middle line:84% But when they go above an upper bound, 00:42:57.230 --> 00:43:01.010 align:middle line:84% it's time to make an adjustment and deposit the paper 00:43:01.010 --> 00:43:02.600 align:middle line:90% in the bank, or lend it out. 00:43:02.600 --> 00:43:05.515 align:middle line:84% And likewise, when it goes below a lower bound, 00:43:05.515 --> 00:43:06.890 align:middle line:84% they're going to have to withdraw 00:43:06.890 --> 00:43:08.900 align:middle line:90% money or call in a loan. 00:43:08.900 --> 00:43:12.200 align:middle line:84% So sometimes that lower bound is zero, obviously. 00:43:12.200 --> 00:43:14.360 align:middle line:84% You can't run out of stock of something. 00:43:14.360 --> 00:43:15.710 align:middle line:90% It can't go negative. 00:43:15.710 --> 00:43:19.140 align:middle line:84% So that's going to be what the solution looks like. 00:43:19.140 --> 00:43:23.300 align:middle line:84% So let me introduce a little bit of notation, e being 00:43:23.300 --> 00:43:28.850 align:middle line:84% expenditure outflows of paper currency, which looks like c. 00:43:28.850 --> 00:43:31.520 align:middle line:84% But it's basically more than consumption. 00:43:31.520 --> 00:43:36.860 align:middle line:84% It's all expenses out, and then y being stuff coming in. 00:43:36.860 --> 00:43:37.730 align:middle line:90% OK. 00:43:37.730 --> 00:43:41.420 align:middle line:84% So expenditure is outflows minus inflows, which 00:43:41.420 --> 00:43:42.930 align:middle line:90% could be positive or negative. 00:43:42.930 --> 00:43:46.280 align:middle line:84% So we postulate a stochastic process 00:43:46.280 --> 00:43:50.990 align:middle line:84% for the cash flows, expenditure flows, and estimate 00:43:50.990 --> 00:43:52.370 align:middle line:90% in the Thai data. 00:43:52.370 --> 00:43:55.040 align:middle line:84% We postulate that mostly things are 00:43:55.040 --> 00:43:56.990 align:middle line:90% pretty smooth and predictable. 00:43:56.990 --> 00:43:59.540 align:middle line:84% Delta refers to the length of the time interval. 00:43:59.540 --> 00:44:04.370 align:middle line:84% And it's easier conceptually to think of delta as equal to 1. 00:44:04.370 --> 00:44:07.910 align:middle line:84% So with roughly probability 1/2, they're 00:44:07.910 --> 00:44:10.880 align:middle line:84% going to have an expenditure of c, 00:44:10.880 --> 00:44:14.660 align:middle line:84% but it could be a higher order of magnitude, sigma. 00:44:14.660 --> 00:44:18.950 align:middle line:84% Or with residual probability roughly 1/2, 00:44:18.950 --> 00:44:20.700 align:middle line:90% it's c minus sigma. 00:44:20.700 --> 00:44:24.290 align:middle line:84% So these are the ups and downs of expenditures-- 00:44:24.290 --> 00:44:25.970 align:middle line:90% probabilistic basis. 00:44:25.970 --> 00:44:28.850 align:middle line:84% How big or small depends on how big sigma is. 00:44:28.850 --> 00:44:31.490 align:middle line:84% And I keep saying roughly probability 1/2, 00:44:31.490 --> 00:44:35.930 align:middle line:84% because there are these zapper things that are very 00:44:35.930 --> 00:44:38.630 align:middle line:90% large inflows and outflows. 00:44:38.630 --> 00:44:41.070 align:middle line:90% So expenditure is an outflow. 00:44:41.070 --> 00:44:47.960 align:middle line:84% So positive zapper, Zp, is a very large expenditure that 00:44:47.960 --> 00:44:50.810 align:middle line:90% happens with probability kappa. 00:44:50.810 --> 00:44:53.030 align:middle line:90% Again, say delta equal 1. 00:44:53.030 --> 00:44:56.390 align:middle line:84% So this could be like 2% or 3% of the time, 00:44:56.390 --> 00:44:58.310 align:middle line:84% there is an outflow which is more 00:44:58.310 --> 00:45:00.590 align:middle line:84% than two standard deviations away 00:45:00.590 --> 00:45:03.560 align:middle line:90% from the mean positive outflow. 00:45:03.560 --> 00:45:05.340 align:middle line:90% Or it could be an inflow. 00:45:05.340 --> 00:45:07.880 align:middle line:84% So these are rare revenues coming 00:45:07.880 --> 00:45:12.210 align:middle line:84% in with different kappas and different Zs, which 00:45:12.210 --> 00:45:13.950 align:middle line:90% will be estimated in the data. 00:45:13.950 --> 00:45:17.670 align:middle line:84% And these shocks-- you take the probability 00:45:17.670 --> 00:45:20.280 align:middle line:84% of the positive outflow, the probability 00:45:20.280 --> 00:45:24.870 align:middle line:84% of the negative outflow add them together, delta, say, 00:45:24.870 --> 00:45:26.460 align:middle line:90% is still equal to 1. 00:45:26.460 --> 00:45:30.150 align:middle line:84% So with probability 1, something happens. 00:45:30.150 --> 00:45:33.990 align:middle line:84% With probability 1 minus the sum of those kappas, 00:45:33.990 --> 00:45:35.880 align:middle line:90% you're in this middle branch. 00:45:35.880 --> 00:45:39.960 align:middle line:84% And that's the 1/2 adjusted up or down. 00:45:39.960 --> 00:45:42.270 align:middle line:84% As always, I'm assuming you will interrupt me 00:45:42.270 --> 00:45:43.930 align:middle line:90% if things are not clear. 00:45:43.930 --> 00:45:44.670 align:middle line:90% OK. 00:45:44.670 --> 00:45:47.220 align:middle line:84% So let's think about how much cash they 00:45:47.220 --> 00:45:51.790 align:middle line:90% have in house that's paper. 00:45:51.790 --> 00:45:56.050 align:middle line:84% So they have a certain amount of cash in a moment in time. 00:45:56.050 --> 00:45:58.660 align:middle line:84% They have an expenditure outflow, which 00:45:58.660 --> 00:46:01.060 align:middle line:90% could be positive or negative. 00:46:01.060 --> 00:46:04.600 align:middle line:84% And they're at t, and we're going 00:46:04.600 --> 00:46:07.540 align:middle line:84% to try to do the accounting for how much money they would 00:46:07.540 --> 00:46:09.940 align:middle line:84% have at t plus delta, where again, delta 00:46:09.940 --> 00:46:11.920 align:middle line:90% could be equal to 1. 00:46:11.920 --> 00:46:16.740 align:middle line:84% Money at t less outflows, and then they could adjust. 00:46:16.740 --> 00:46:20.100 align:middle line:84% They could either withdraw money or deposit it. 00:46:20.100 --> 00:46:23.790 align:middle line:84% So these two things the w and d is endogenous. 00:46:23.790 --> 00:46:27.150 align:middle line:84% And the model is going to have to solve for w and d, 00:46:27.150 --> 00:46:30.360 align:middle line:84% when to do it, as I was saying, and how much to do. 00:46:30.360 --> 00:46:34.440 align:middle line:84% And that will result in this endogenously determined 00:46:34.440 --> 00:46:36.210 align:middle line:90% statement of cash flow. 00:46:36.210 --> 00:46:38.550 align:middle line:84% You can think about this like life cycle planning, 00:46:38.550 --> 00:46:41.730 align:middle line:84% except now it's not over 50 years. 00:46:41.730 --> 00:46:45.990 align:middle line:84% It may be a matter of a couple of weeks or 30 days. 00:46:45.990 --> 00:46:48.240 align:middle line:84% In the data, it's 30 days because that's 00:46:48.240 --> 00:46:50.200 align:middle line:90% when we get the data. 00:46:50.200 --> 00:46:51.730 align:middle line:84% They're trying to minimize costs. 00:46:51.730 --> 00:46:53.440 align:middle line:90% Now, there's two costs. 00:46:53.440 --> 00:46:57.140 align:middle line:84% One is kind of obvious, really, for economists, 00:46:57.140 --> 00:47:00.670 align:middle line:84% which is if you're holding cash and it doesn't earn interest, 00:47:00.670 --> 00:47:02.980 align:middle line:84% you've lost the interest you could have earned if you 00:47:02.980 --> 00:47:04.810 align:middle line:90% had deposited it in the bank. 00:47:04.810 --> 00:47:06.850 align:middle line:84% So that's a real cost to the household. 00:47:06.850 --> 00:47:09.440 align:middle line:84% That's the opportunity cost to the household. 00:47:09.440 --> 00:47:11.440 align:middle line:84% We're going to let the interest rate in the bank 00:47:11.440 --> 00:47:14.320 align:middle line:84% be capital R. If you hold too much cash, 00:47:14.320 --> 00:47:17.140 align:middle line:84% you're giving up a lot of foregone interest. 00:47:17.140 --> 00:47:18.990 align:middle line:90% That's a cost. 00:47:18.990 --> 00:47:21.560 align:middle line:84% There's also a physical adjustment cost 00:47:21.560 --> 00:47:25.040 align:middle line:84% to get to the bank, which you can think of as the hassle, 00:47:25.040 --> 00:47:27.350 align:middle line:84% or in these villages you've got to get in your truck 00:47:27.350 --> 00:47:30.530 align:middle line:84% and drive to the village, to the town, et cetera. 00:47:30.530 --> 00:47:32.930 align:middle line:84% So that should have had the notation here, 00:47:32.930 --> 00:47:34.370 align:middle line:90% but it's a little b. 00:47:34.370 --> 00:47:37.140 align:middle line:84% You'll see a little b momentarily. 00:47:37.140 --> 00:47:39.170 align:middle line:90% So the next slide-- don't panic. 00:47:39.170 --> 00:47:43.370 align:middle line:84% I'm going to walk you through the value function. 00:47:43.370 --> 00:47:45.800 align:middle line:84% It looks pretty intense at first. 00:47:45.800 --> 00:47:50.420 align:middle line:84% They're sitting here with a certain amount of money, m, 00:47:50.420 --> 00:47:52.580 align:middle line:84% and then deciding whether they want 00:47:52.580 --> 00:47:58.640 align:middle line:84% to stay put and carry m into the second period or adjust. 00:47:58.640 --> 00:48:03.410 align:middle line:84% Now, if m is already negative, they're in trouble. 00:48:03.410 --> 00:48:06.320 align:middle line:84% Can't go negative, so they have to adjust. 00:48:06.320 --> 00:48:09.410 align:middle line:84% And you get this cost, b, which is gone. 00:48:09.410 --> 00:48:11.450 align:middle line:84% This cost function is not utility. 00:48:11.450 --> 00:48:13.070 align:middle line:90% It's not net profits. 00:48:13.070 --> 00:48:17.430 align:middle line:84% It's a cost, and they're trying to minimize the cost. 00:48:17.430 --> 00:48:20.900 align:middle line:84% So this is enters positively because it's a cost. 00:48:20.900 --> 00:48:23.180 align:middle line:84% They've gone to all the costs, to getting to the bank, 00:48:23.180 --> 00:48:24.890 align:middle line:84% they get to make an adjustment, which 00:48:24.890 --> 00:48:27.780 align:middle line:84% means they can choose a new m prime going out. 00:48:27.780 --> 00:48:32.960 align:middle line:84% And so they'll try to minimize the value from that point 00:48:32.960 --> 00:48:36.020 align:middle line:90% on by that choice of m prime-- 00:48:36.020 --> 00:48:38.960 align:middle line:84% minimize the cost from that point on. 00:48:38.960 --> 00:48:41.270 align:middle line:84% Here, m is positive, but they still 00:48:41.270 --> 00:48:43.770 align:middle line:84% get to choose whether to adjust or not. 00:48:43.770 --> 00:48:46.790 align:middle line:84% And if they don't adjust, they keep it. 00:48:46.790 --> 00:48:48.710 align:middle line:84% They suffer the opportunity cost, 00:48:48.710 --> 00:48:50.360 align:middle line:90% the lost interest in the bank. 00:48:50.360 --> 00:48:52.980 align:middle line:84% And then they'll have their value function tomorrow, 00:48:52.980 --> 00:48:55.730 align:middle line:90% which is discounted by little r. 00:48:55.730 --> 00:48:57.920 align:middle line:90% That's like a discount rate. 00:48:57.920 --> 00:49:01.160 align:middle line:84% And things next period are stochastic. 00:49:01.160 --> 00:49:05.120 align:middle line:84% So you take expectations over the future e values 00:49:05.120 --> 00:49:06.990 align:middle line:84% following that stochastic process. 00:49:06.990 --> 00:49:10.700 align:middle line:84% So this is v of m minus e is all expectation, 00:49:10.700 --> 00:49:13.880 align:middle line:84% is all written down here in terms of these branches, which 00:49:13.880 --> 00:49:19.400 align:middle line:84% is positive zappers, negative zappers, ups and downs, roughly 00:49:19.400 --> 00:49:21.260 align:middle line:90% probability 1/2, et cetera. 00:49:21.260 --> 00:49:25.160 align:middle line:84% So this is the expected cost in the future 00:49:25.160 --> 00:49:29.280 align:middle line:84% if you don't make a monetary adjustment today. 00:49:29.280 --> 00:49:31.940 align:middle line:84% So this is the second or third time 00:49:31.940 --> 00:49:33.440 align:middle line:84% you've seen these value functions. 00:49:33.440 --> 00:49:36.020 align:middle line:84% I introduced them at the end of the medieval village 00:49:36.020 --> 00:49:37.730 align:middle line:90% dynamic programming problem. 00:49:37.730 --> 00:49:41.930 align:middle line:84% We then used that last class to talk about implications 00:49:41.930 --> 00:49:45.015 align:middle line:84% of risk sharing for production, with the planner's value 00:49:45.015 --> 00:49:45.515 align:middle line:90% function. 00:49:45.515 --> 00:49:46.920 align:middle line:90% This is the third time. 00:49:46.920 --> 00:49:51.980 align:middle line:84% This is the value function for a household minimizing its costs. 00:49:51.980 --> 00:49:53.970 align:middle line:84% More fun-- let me show you some pictures. 00:49:53.970 --> 00:49:57.050 align:middle line:90% So these are real live people. 00:49:57.050 --> 00:49:59.630 align:middle line:84% This is the cash flow of this house-- the cash 00:49:59.630 --> 00:50:01.760 align:middle line:90% holdings of this household. 00:50:01.760 --> 00:50:04.760 align:middle line:84% Up and up and up and up she goes. 00:50:04.760 --> 00:50:07.460 align:middle line:84% This one also had a trend, and then boom, 00:50:07.460 --> 00:50:12.050 align:middle line:84% down, and then to zero, and then back up again. 00:50:12.050 --> 00:50:13.760 align:middle line:90% You can't see it very well. 00:50:13.760 --> 00:50:14.330 align:middle line:90% I'm sorry. 00:50:14.330 --> 00:50:16.220 align:middle line:90% The pictures are too small. 00:50:16.220 --> 00:50:22.040 align:middle line:84% But what's plotted over here are net cash inflows or outflows, 00:50:22.040 --> 00:50:26.600 align:middle line:84% along with, say, a deposit when you have a net inflow, 00:50:26.600 --> 00:50:30.330 align:middle line:84% or a withdrawal when you have a net outflow. 00:50:30.330 --> 00:50:33.260 align:middle line:84% So there's three lines here, and sometimes 00:50:33.260 --> 00:50:35.870 align:middle line:84% things correspond the way I was just implying. 00:50:35.870 --> 00:50:39.440 align:middle line:84% You're short, so you withdraw, et cetera. 00:50:39.440 --> 00:50:41.210 align:middle line:84% But when-- you can't see much here, 00:50:41.210 --> 00:50:43.550 align:middle line:84% but when you integrate it out and look 00:50:43.550 --> 00:50:45.680 align:middle line:84% at the consequence for the levels 00:50:45.680 --> 00:50:50.120 align:middle line:84% using this flow equation, then you can see what 00:50:50.120 --> 00:50:51.942 align:middle line:90% their stocks are. 00:50:51.942 --> 00:50:53.150 align:middle line:90% It looks like a lot of money. 00:50:53.150 --> 00:50:56.150 align:middle line:90% It is going to be unbelievable-- 00:50:56.150 --> 00:50:58.010 align:middle line:84% unbelievable in the sense of three 00:50:58.010 --> 00:51:01.650 align:middle line:84% years research running to try to understand this pattern. 00:51:01.650 --> 00:51:03.950 align:middle line:84% So how much money does a typical household 00:51:03.950 --> 00:51:07.370 align:middle line:84% have-- currency, paper, squirreled away in the house? 00:51:07.370 --> 00:51:11.180 align:middle line:84% In the rural villages that we were looking at, 00:51:11.180 --> 00:51:15.050 align:middle line:84% roughly 32 months worth of consumption. 00:51:15.050 --> 00:51:16.670 align:middle line:90% What's in your wallet? 00:51:16.670 --> 00:51:18.800 align:middle line:90% You're going to say credit card. 00:51:18.800 --> 00:51:20.540 align:middle line:84% These guys-- it's not in a wallet. 00:51:20.540 --> 00:51:23.810 align:middle line:84% It's in the house or hidden in the backyard. 00:51:23.810 --> 00:51:27.800 align:middle line:84% Almost three years worth of expenditures in cash-- 00:51:27.800 --> 00:51:29.360 align:middle line:90% cash equivalent consumption. 00:51:29.360 --> 00:51:32.450 align:middle line:84% We measure the consumption for each household, 00:51:32.450 --> 00:51:34.160 align:middle line:84% and we see how much cash they have, 00:51:34.160 --> 00:51:35.960 align:middle line:90% which we measured in the data. 00:51:35.960 --> 00:51:38.630 align:middle line:90% 32 months-- it's unbelievable. 00:51:38.630 --> 00:51:41.510 align:middle line:84% And we also get deposits and withdrawals 00:51:41.510 --> 00:51:42.770 align:middle line:90% measured in the data. 00:51:42.770 --> 00:51:44.450 align:middle line:84% That's a lot of detail to absorb. 00:51:44.450 --> 00:51:48.230 align:middle line:84% Let me show you what's going on in the model. 00:51:48.230 --> 00:51:50.660 align:middle line:90% We estimate the cash flow. 00:51:50.660 --> 00:51:54.530 align:middle line:84% We recover all the parameters that we need, 00:51:54.530 --> 00:51:57.920 align:middle line:84% take a stand on the interest rate in the bank, 00:51:57.920 --> 00:52:01.900 align:middle line:84% and then look at how small or big the cost would be. 00:52:01.900 --> 00:52:05.060 align:middle line:90% b is varying down here. 00:52:05.060 --> 00:52:09.980 align:middle line:84% And the consequences for currency, especially but also 00:52:09.980 --> 00:52:12.360 align:middle line:90% withdrawals and deposits. 00:52:12.360 --> 00:52:16.100 align:middle line:84% So taking the reality to be the calibrated expenditure 00:52:16.100 --> 00:52:19.840 align:middle line:84% flows we see in the data, along with the few other reasonable 00:52:19.840 --> 00:52:22.660 align:middle line:84% assumptions, I assure you we can look 00:52:22.660 --> 00:52:25.150 align:middle line:84% at how much cash they should be holding relative 00:52:25.150 --> 00:52:26.140 align:middle line:90% to consumption. 00:52:26.140 --> 00:52:27.910 align:middle line:90% That's this blue line. 00:52:27.910 --> 00:52:32.690 align:middle line:84% And we can barely get it up to three. 00:52:32.690 --> 00:52:35.150 align:middle line:90% The number in the data is 32. 00:52:35.150 --> 00:52:38.840 align:middle line:84% I mean, you could imagine increasing those costs 00:52:38.840 --> 00:52:40.690 align:middle line:90% and decreasing the interest. 00:52:40.690 --> 00:52:45.800 align:middle line:84% And this is like asymptoting out to something flat. 00:52:45.800 --> 00:52:48.590 align:middle line:84% So you end up with something ridiculous. 00:52:48.590 --> 00:52:50.150 align:middle line:90% The model isn't working. 00:52:50.150 --> 00:52:53.275 align:middle line:84% You don't go to the bank and forego two years 00:52:53.275 --> 00:52:54.650 align:middle line:84% worth of consumption while you're 00:52:54.650 --> 00:52:57.110 align:middle line:90% doing your banking transection. 00:52:57.110 --> 00:52:59.540 align:middle line:84% It's just not a reasonable value. 00:52:59.540 --> 00:53:02.510 align:middle line:90% The sign makes sense. 00:53:02.510 --> 00:53:06.140 align:middle line:84% The higher the cost, b, of going to the bank 00:53:06.140 --> 00:53:09.750 align:middle line:84% to either withdraw or deposit, the less you're going to do it. 00:53:09.750 --> 00:53:12.830 align:middle line:84% So when you do it, the transactions are bigger. 00:53:12.830 --> 00:53:15.515 align:middle line:84% And meanwhile, you're holding more cash. 00:53:15.515 --> 00:53:19.700 align:middle line:84% If b goes to infinity, then it's all in cash. 00:53:19.700 --> 00:53:21.008 align:middle line:90% Forget the bank. 00:53:21.008 --> 00:53:22.550 align:middle line:84% And when you go rarely, then you have 00:53:22.550 --> 00:53:24.510 align:middle line:90% to make up a big difference. 00:53:24.510 --> 00:53:26.450 align:middle line:84% And this is the frequency, which again, 00:53:26.450 --> 00:53:29.540 align:middle line:84% as b goes up, the frequency of going to the bank 00:53:29.540 --> 00:53:31.010 align:middle line:90% is going down. 00:53:31.010 --> 00:53:36.710 align:middle line:84% Question-- so because we did that life cycle planner, 00:53:36.710 --> 00:53:40.760 align:middle line:84% we wanted to evaluate as a randomized controlled trial 00:53:40.760 --> 00:53:44.730 align:middle line:84% how well or poorly it was working, and in the process, 00:53:44.730 --> 00:53:48.170 align:middle line:84% we gathered monthly data for the urban household. 00:53:48.170 --> 00:53:50.330 align:middle line:84% So we're not in villages anymore. 00:53:50.330 --> 00:53:53.780 align:middle line:84% We're in towns or even city neighborhoods. 00:53:53.780 --> 00:53:57.110 align:middle line:84% In each of the four provinces, we basically 00:53:57.110 --> 00:53:59.910 align:middle line:84% replicated all the existing surveys. 00:53:59.910 --> 00:54:02.430 align:middle line:84% So the same questions are being asked. 00:54:02.430 --> 00:54:04.430 align:middle line:84% And this is kind of buried here, but if you 00:54:04.430 --> 00:54:08.420 align:middle line:84% look at the normalized cash in terms of consumption 00:54:08.420 --> 00:54:12.800 align:middle line:84% for the urban sample, and look over here, 00:54:12.800 --> 00:54:16.280 align:middle line:90% we got a number of 3.87-- 00:54:16.280 --> 00:54:17.420 align:middle line:90% much closer to 3. 00:54:17.420 --> 00:54:19.070 align:middle line:90% Nothing like the 37. 00:54:19.070 --> 00:54:21.890 align:middle line:84% Well, actually, I should be over here at the median. 00:54:21.890 --> 00:54:23.630 align:middle line:90% So it's 1.94. 00:54:23.630 --> 00:54:25.760 align:middle line:84% But when we do some adjustments in the data, 00:54:25.760 --> 00:54:29.200 align:middle line:84% I'm happy to tell you about them. 00:54:29.200 --> 00:54:31.660 align:middle line:84% That's within reach of the model. 00:54:31.660 --> 00:54:37.360 align:middle line:84% We calibrate the model again using the cash statistics-- 00:54:37.360 --> 00:54:40.240 align:middle line:84% the zapping, big shock, small shock, 00:54:40.240 --> 00:54:42.640 align:middle line:84% how often they happen, et cetera et cetera, 00:54:42.640 --> 00:54:44.350 align:middle line:90% and get a model prediction. 00:54:44.350 --> 00:54:48.310 align:middle line:84% And the implied value, m, depending 00:54:48.310 --> 00:54:50.740 align:middle line:84% on whether you use the mean or the median, 00:54:50.740 --> 00:54:55.330 align:middle line:84% is something like 2, which is close to what 00:54:55.330 --> 00:54:56.990 align:middle line:90% we're getting in the data. 00:54:56.990 --> 00:54:58.870 align:middle line:84% So now we're in kind of business, 00:54:58.870 --> 00:55:02.170 align:middle line:84% in the sense that a reasonable economic model is actually 00:55:02.170 --> 00:55:05.230 align:middle line:84% predicting pretty well how much cash they're holding. 00:55:05.230 --> 00:55:07.840 align:middle line:84% Whereas, in the rural data it was a disaster. 00:55:07.840 --> 00:55:08.950 align:middle line:90% We didn't get close. 00:55:08.950 --> 00:55:15.040 align:middle line:84% So now we switch into how can we understand the error 00:55:15.040 --> 00:55:18.580 align:middle line:84% that the households are making with their cash management? 00:55:18.580 --> 00:55:19.930 align:middle line:90% And the guess is-- 00:55:19.930 --> 00:55:21.190 align:middle line:90% several guesses. 00:55:21.190 --> 00:55:24.130 align:middle line:84% Some turned out to be correct and some not. 00:55:24.130 --> 00:55:26.950 align:middle line:84% So why would cash holdings be so much lower 00:55:26.950 --> 00:55:29.800 align:middle line:84% in the urban area relative to the rural area? 00:55:29.800 --> 00:55:32.680 align:middle line:84% Is it because households in the urban survey deposit 00:55:32.680 --> 00:55:34.990 align:middle line:90% more often in larger amounts? 00:55:34.990 --> 00:55:38.410 align:middle line:84% Or is it because they withdraw in the urban areas less 00:55:38.410 --> 00:55:40.790 align:middle line:90% and withdraw smaller amounts? 00:55:40.790 --> 00:55:44.890 align:middle line:84% So for deposits, it turns out households in the rural area 00:55:44.890 --> 00:55:49.020 align:middle line:84% deposit more and not less, and they deposit larger amounts. 00:55:49.020 --> 00:55:51.160 align:middle line:84% So this is exactly the opposite of what 00:55:51.160 --> 00:55:54.280 align:middle line:84% we would have predicted because the cash is going into the bank 00:55:54.280 --> 00:55:56.360 align:middle line:90% or into the local money market. 00:55:56.360 --> 00:55:58.480 align:middle line:84% However, for withdrawals, it works 00:55:58.480 --> 00:56:00.610 align:middle line:84% the way we would have conjectured, 00:56:00.610 --> 00:56:02.860 align:middle line:84% that when urban households put money in the bank 00:56:02.860 --> 00:56:04.720 align:middle line:84% they don't withdraw it very often, 00:56:04.720 --> 00:56:06.760 align:middle line:90% and they withdraw a lot less. 00:56:06.760 --> 00:56:09.940 align:middle line:84% It almost looks like they're putting their proxy 00:56:09.940 --> 00:56:11.710 align:middle line:90% for their salary into the bank. 00:56:11.710 --> 00:56:15.160 align:middle line:84% And any other money which they have, from business and so on, 00:56:15.160 --> 00:56:17.200 align:middle line:84% goes into the bank and stays there. 00:56:17.200 --> 00:56:19.960 align:middle line:84% They just withdraw what they need for consumption, 00:56:19.960 --> 00:56:22.990 align:middle line:84% taking a, quote unquote, salary equivalent out 00:56:22.990 --> 00:56:24.440 align:middle line:90% and leaving the rest in. 00:56:24.440 --> 00:56:28.420 align:middle line:84% So this is why their money value's a lot lower. 00:56:28.420 --> 00:56:32.290 align:middle line:84% There's still something to fix, which just like in life cycle 00:56:32.290 --> 00:56:36.790 align:middle line:84% planner, we could create from the research cash flow app 00:56:36.790 --> 00:56:40.430 align:middle line:84% and put it on the phone and tell households how to manage it. 00:56:40.430 --> 00:56:43.030 align:middle line:84% And, in fact, the question is, why are they 00:56:43.030 --> 00:56:44.410 align:middle line:90% having cash at all? 00:56:44.410 --> 00:56:46.690 align:middle line:84% Why don't they just put it on their cell phone? 00:56:46.690 --> 00:56:49.130 align:middle line:90% And that brings us to-- 00:56:49.130 --> 00:56:49.630 align:middle line:90% oh. 00:56:49.630 --> 00:56:53.560 align:middle line:84% And the costs of mismanagement of currency 00:56:53.560 --> 00:56:56.530 align:middle line:84% are very high-- something like 9 and 1/2% 00:56:56.530 --> 00:56:58.510 align:middle line:90% of monthly consumption. 00:56:58.510 --> 00:57:00.760 align:middle line:90% That's huge in economics. 00:57:00.760 --> 00:57:02.290 align:middle line:90% Say, exaggerate a bit-- 00:57:02.290 --> 00:57:04.630 align:middle line:90% 10, or call it 5 if you want. 00:57:04.630 --> 00:57:08.170 align:middle line:84% 5% of monthly consumption could be saved or used 00:57:08.170 --> 00:57:11.600 align:middle line:84% for consumption if they just managed their currency better. 00:57:11.600 --> 00:57:13.300 align:middle line:90% So one way to put this-- 00:57:13.300 --> 00:57:17.800 align:middle line:84% enormous cost of mismanagement of petty cash. 00:57:17.800 --> 00:57:20.650 align:middle line:84% Businesses get cash advisory services. 00:57:20.650 --> 00:57:22.340 align:middle line:90% They pay for it. 00:57:22.340 --> 00:57:24.670 align:middle line:84% Why should these poor households not have access 00:57:24.670 --> 00:57:26.950 align:middle line:90% to better management tools? 00:57:26.950 --> 00:57:27.490 align:middle line:90% OK. 00:57:27.490 --> 00:57:28.390 align:middle line:90% Sorry. 00:57:28.390 --> 00:57:29.680 align:middle line:90% That's my view. 00:57:29.680 --> 00:57:30.910 align:middle line:90% OK. 00:57:30.910 --> 00:57:32.450 align:middle line:90% Go to Sweden. 00:57:32.450 --> 00:57:34.750 align:middle line:84% How much paper currency are they using? 00:57:34.750 --> 00:57:35.890 align:middle line:90% Close to nothing. 00:57:35.890 --> 00:57:37.360 align:middle line:90% They're gone-- it's gone. 00:57:37.360 --> 00:57:43.090 align:middle line:84% This is the amount of currency relative to GDP ever-declining. 00:57:43.090 --> 00:57:46.450 align:middle line:84% This is the absolute level of cash, 00:57:46.450 --> 00:57:50.180 align:middle line:84% going down in an absolute not only relative basis. 00:57:50.180 --> 00:57:53.080 align:middle line:84% So I got this and this slide from the governor 00:57:53.080 --> 00:57:56.230 align:middle line:90% of the Riksbank in Sweden. 00:57:56.230 --> 00:57:58.820 align:middle line:84% So I don't know whether you will appreciate the humor here. 00:57:58.820 --> 00:58:00.940 align:middle line:84% But this is a picture of a stone. 00:58:00.940 --> 00:58:03.820 align:middle line:84% So economists and others have written books 00:58:03.820 --> 00:58:07.720 align:middle line:84% about the first coins, the first use of cash, 00:58:07.720 --> 00:58:10.090 align:middle line:84% and there were claims on stones that were buried 00:58:10.090 --> 00:58:11.440 align:middle line:90% at the bottom of the lake. 00:58:11.440 --> 00:58:14.620 align:middle line:84% So now in Sweden, they're transferring the stone around 00:58:14.620 --> 00:58:15.550 align:middle line:90% on their cell phone. 00:58:15.550 --> 00:58:16.900 align:middle line:90% Well, not really. 00:58:16.900 --> 00:58:18.940 align:middle line:90% It's Swish, basically. 00:58:18.940 --> 00:58:22.630 align:middle line:84% They have a bank account, and for no money at all 00:58:22.630 --> 00:58:26.810 align:middle line:84% households can transfer cash to each other, buy things, 00:58:26.810 --> 00:58:27.800 align:middle line:90% and so on. 00:58:27.800 --> 00:58:29.740 align:middle line:84% And it's replaced the paper currency. 00:58:29.740 --> 00:58:32.290 align:middle line:84% It's actually a bit of a concern for them, which 00:58:32.290 --> 00:58:34.900 align:middle line:84% is why they're thinking if there's no paper 00:58:34.900 --> 00:58:37.300 align:middle line:84% currency other than reserves, there's 00:58:37.300 --> 00:58:40.090 align:middle line:90% no role for the central bank. 00:58:40.090 --> 00:58:44.020 align:middle line:84% So they feel they should be providing a payment 00:58:44.020 --> 00:58:45.280 align:middle line:90% service as a backup. 00:58:45.280 --> 00:58:48.610 align:middle line:84% So they're going to create, almost surely, a central bank 00:58:48.610 --> 00:58:49.910 align:middle line:90% digital currency. 00:58:49.910 --> 00:58:52.240 align:middle line:84% So households will have digital accounts 00:58:52.240 --> 00:58:55.810 align:middle line:84% in the central bank, which is what South Africa is doing. 00:58:55.810 --> 00:58:58.810 align:middle line:84% China has a different version of it and so on. 00:58:58.810 --> 00:59:01.060 align:middle line:90% This happened in the Bahamas. 00:59:01.060 --> 00:59:03.580 align:middle line:84% They created something called the Sand Dollar. 00:59:03.580 --> 00:59:05.770 align:middle line:84% The virus is accelerating this problem 00:59:05.770 --> 00:59:08.050 align:middle line:84% because people need government aid, 00:59:08.050 --> 00:59:09.730 align:middle line:90% and they can't get the aid-- 00:59:09.730 --> 00:59:11.740 align:middle line:84% truck the paper currency over there. 00:59:11.740 --> 00:59:15.280 align:middle line:84% When this disastrous Hurricane Dorian destroyed an island 00:59:15.280 --> 00:59:18.590 align:middle line:84% in the Bahamas, they couldn't get emergency assistance. 00:59:18.590 --> 00:59:21.070 align:middle line:84% So these are things that are happening all over the world 00:59:21.070 --> 00:59:22.120 align:middle line:90% right now. 00:59:22.120 --> 00:59:24.610 align:middle line:90% This last slide is Kenya. 00:59:24.610 --> 00:59:27.490 align:middle line:84% And I mentioned they have two kinds of currency. 00:59:27.490 --> 00:59:30.750 align:middle line:84% One is Kenyan shillings-- used to be a British colony. 00:59:30.750 --> 00:59:34.350 align:middle line:84% And the other is e-credits at Safaricom, 00:59:34.350 --> 00:59:36.330 align:middle line:90% which is a cell phone provider. 00:59:36.330 --> 00:59:38.880 align:middle line:84% So what you do is you take your shillings, 00:59:38.880 --> 00:59:41.640 align:middle line:90% you go to this Safaricom agent. 00:59:41.640 --> 00:59:45.030 align:middle line:84% You give him your Fiat money, and you 00:59:45.030 --> 00:59:48.690 align:middle line:84% get credited cell accounts called M-Pesa. 00:59:48.690 --> 00:59:51.180 align:middle line:84% And then you can use that to give to your mom 00:59:51.180 --> 00:59:54.540 align:middle line:84% back in the village, who goes to the agent there 00:59:54.540 --> 00:59:56.940 align:middle line:84% and transfers it back out into shillings, 00:59:56.940 --> 00:59:58.960 align:middle line:90% or you can buy stuff. 00:59:58.960 --> 01:00:02.530 align:middle line:84% So here you see the use of the paper currency, 01:00:02.530 --> 01:00:06.840 align:middle line:84% but also this e-currency, which is created by Safaricom. 01:00:06.840 --> 01:00:09.300 align:middle line:84% And it has reached almost everybody. 01:00:09.300 --> 01:00:10.830 align:middle line:90% Some economies try to do this. 01:00:10.830 --> 01:00:12.330 align:middle line:90% We kind of tried in Thailand. 01:00:12.330 --> 01:00:14.340 align:middle line:90% I have not yet succeeded. 01:00:14.340 --> 01:00:17.760 align:middle line:84% In Kenya, even the lowest income quartile-- 01:00:17.760 --> 01:00:19.890 align:middle line:90% we need to update this graph. 01:00:19.890 --> 01:00:21.900 align:middle line:90% Tavneet Suri did this. 01:00:21.900 --> 01:00:26.250 align:middle line:84% Roughly, 85% of the low-income population, 01:00:26.250 --> 01:00:29.730 align:middle line:84% and a similar, higher number for those with no bank accounts, 01:00:29.730 --> 01:00:31.860 align:middle line:90% are using Safaricom. 01:00:31.860 --> 01:00:33.610 align:middle line:84% With no bank accounts-- in other words, 01:00:33.610 --> 01:00:37.020 align:middle line:84% you don't need the traditional legacy infrastructure. 01:00:37.020 --> 01:00:41.640 align:middle line:84% And this is, again, the allure, the so-called disintermediation 01:00:41.640 --> 01:00:44.340 align:middle line:84% that the fintechs are bringing to the world. 01:00:44.340 --> 01:00:47.310 align:middle line:84% Of course, the commercial bankers don't like it much. 01:00:47.310 --> 01:00:48.900 align:middle line:90% We talked about risk-sharing. 01:00:48.900 --> 01:00:53.700 align:middle line:84% And Tavneet and her co-author, Billy Jack, 01:00:53.700 --> 01:00:58.710 align:middle line:84% looked at how much consumption is being smooth in Kenya, 01:00:58.710 --> 01:01:00.840 align:middle line:84% comparing households who have not yet 01:01:00.840 --> 01:01:04.440 align:middle line:84% gotten early on the M-Pesa expansion versus those 01:01:04.440 --> 01:01:05.430 align:middle line:90% that did. 01:01:05.430 --> 01:01:09.640 align:middle line:84% And there's a lot more smoother consumption, 01:01:09.640 --> 01:01:11.070 align:middle line:84% which you see these common themes. 01:01:11.070 --> 01:01:12.330 align:middle line:90% We did risk-sharing. 01:01:12.330 --> 01:01:14.910 align:middle line:84% We did gift giving in Thai villages. 01:01:14.910 --> 01:01:20.370 align:middle line:84% Now people are giving gifts and transfers through M-Pesa. 01:01:20.370 --> 01:01:22.726 align:middle line:90% And so it has an impact. 01:01:22.726 --> 01:01:24.100 align:middle line:90% Well, what's the big deal? 01:01:24.100 --> 01:01:26.910 align:middle line:90% It's a big deal for Kenya. 01:01:26.910 --> 01:01:28.830 align:middle line:84% Sweden already has this infrastructure. 01:01:28.830 --> 01:01:30.800 align:middle line:90% It's not a big deal. 01:01:30.800 --> 01:01:33.590 align:middle line:84% But it makes a big deal for low-income countries. 01:01:33.590 --> 01:01:36.440 align:middle line:84% There are some qualifications, though. 01:01:36.440 --> 01:01:39.090 align:middle line:84% First of all, it's not on a distributed ledger. 01:01:39.090 --> 01:01:43.310 align:middle line:84% So you have to trust Safaricom to keep the accounts properly, 01:01:43.310 --> 01:01:45.030 align:middle line:90% and they have the cash. 01:01:45.030 --> 01:01:47.690 align:middle line:84% So they're supposed to put it in a commercial bank. 01:01:47.690 --> 01:01:50.010 align:middle line:90% You trust them to do that. 01:01:50.010 --> 01:01:53.570 align:middle line:84% So you have to rely on a third-party vendor in Kenya. 01:01:53.570 --> 01:01:55.550 align:middle line:84% Not only that, the money's in the bank, 01:01:55.550 --> 01:01:57.050 align:middle line:90% and Kenya is subject to runs. 01:01:57.050 --> 01:01:58.998 align:middle line:90% So how safe is that? 01:01:58.998 --> 01:02:01.040 align:middle line:84% You may think you're going to get the money back, 01:02:01.040 --> 01:02:02.780 align:middle line:90% but will you really? 01:02:02.780 --> 01:02:05.870 align:middle line:84% And in other economies, they don't have that kind of trust, 01:02:05.870 --> 01:02:08.370 align:middle line:90% so it doesn't take off. 01:02:08.370 --> 01:02:13.070 align:middle line:84% The other thing is those agents, these guys, 01:02:13.070 --> 01:02:14.950 align:middle line:90% they run out of stuff. 01:02:14.950 --> 01:02:16.870 align:middle line:84% So you go and take your shillings 01:02:16.870 --> 01:02:18.580 align:middle line:90% and say I want M-Pesa. 01:02:18.580 --> 01:02:21.160 align:middle line:90% And they can't create it. 01:02:21.160 --> 01:02:22.780 align:middle line:90% There's no net creation. 01:02:22.780 --> 01:02:24.340 align:middle line:84% It has to be a one-to-one transfer. 01:02:24.340 --> 01:02:26.510 align:middle line:84% They have to have it in their accounts already. 01:02:26.510 --> 01:02:28.180 align:middle line:84% And if don't have it in the account, 01:02:28.180 --> 01:02:30.610 align:middle line:84% they've got to say, well, wait, or come back tomorrow. 01:02:30.610 --> 01:02:32.830 align:middle line:84% Likewise, and perhaps more obviously 01:02:32.830 --> 01:02:36.880 align:middle line:84% for the currency withdrawal, say here's my cell account. 01:02:36.880 --> 01:02:37.990 align:middle line:90% I want to cash it out. 01:02:37.990 --> 01:02:40.165 align:middle line:84% Oh, I don't have that cash in my drawer. 01:02:40.165 --> 01:02:43.900 align:middle line:84% What Tavneet discovered is these agents are kind of linked 01:02:43.900 --> 01:02:46.420 align:middle line:84% to each other in an informal money market, 01:02:46.420 --> 01:02:48.280 align:middle line:84% giving gifts to each other, actually, 01:02:48.280 --> 01:02:49.870 align:middle line:84% and borrowing and lending in order 01:02:49.870 --> 01:02:52.720 align:middle line:84% to smooth the fluctuation, from smoothing at the household 01:02:52.720 --> 01:02:54.970 align:middle line:84% level, to smoothing at the wholesale level, 01:02:54.970 --> 01:02:58.540 align:middle line:90% and similar models would apply. 01:02:58.540 --> 01:03:00.760 align:middle line:84% So this brings us to the theme of building 01:03:00.760 --> 01:03:02.930 align:middle line:90% financial infrastructure. 01:03:02.930 --> 01:03:05.650 align:middle line:84% How can we not just get currency on ledgers 01:03:05.650 --> 01:03:08.080 align:middle line:84% and transfer it around in a cheap way? 01:03:08.080 --> 01:03:09.700 align:middle line:90% How can we build in the trust? 01:03:09.700 --> 01:03:12.460 align:middle line:84% And how can we deal with this shortage 01:03:12.460 --> 01:03:15.970 align:middle line:84% problem among over-the-counter vendors and agents and so on? 01:03:15.970 --> 01:03:18.010 align:middle line:84% And next lecture, starting next time, 01:03:18.010 --> 01:03:21.050 align:middle line:84% we're going to do mechanism design and private information, 01:03:21.050 --> 01:03:22.540 align:middle line:84% and I'll tell you how to design-- 01:03:22.540 --> 01:03:25.530 align:middle line:84% at some point-- how to design this system. 01:03:25.530 --> 01:03:26.220 align:middle line:90% All right. 01:03:26.220 --> 01:03:28.830 align:middle line:84% Went over by an extra minute today. 01:03:28.830 --> 01:03:30.990 align:middle line:90% Questions? 01:03:30.990 --> 01:03:32.710 align:middle line:90% All right. 01:03:32.710 --> 01:03:36.160 align:middle line:84% So otherwise, I'll see you next Thursday or in my office 01:03:36.160 --> 01:03:38.310 align:middle line:90% momentarily. 01:03:38.310 --> 01:03:51.386 align:middle line:90%