WEBVTT 00:00:00.000 --> 00:00:01.988 align:middle line:90% [SQUEAKING] 00:00:01.988 --> 00:00:04.473 align:middle line:90% [RUSTLING] 00:00:04.473 --> 00:00:07.455 align:middle line:90% [CLICKING] 00:00:07.455 --> 00:00:17.970 align:middle line:90% 00:00:17.970 --> 00:00:18.970 align:middle line:90% BEN OLKEN: Hi, everyone. 00:00:18.970 --> 00:00:21.143 align:middle line:90% Let's get started. 00:00:21.143 --> 00:00:23.310 align:middle line:84% So I think it's now me for the rest of the semester. 00:00:23.310 --> 00:00:27.290 align:middle line:84% So for whatever reason, yesterday I 00:00:27.290 --> 00:00:28.630 align:middle line:90% swapped one lecture last week. 00:00:28.630 --> 00:00:31.460 align:middle line:90% Now I won't do that this week. 00:00:31.460 --> 00:00:32.623 align:middle line:90% Now I'm back. 00:00:32.623 --> 00:00:34.040 align:middle line:84% So what I want to talk about today 00:00:34.040 --> 00:00:38.360 align:middle line:84% is land markets, including the paper you guys read for today. 00:00:38.360 --> 00:00:40.550 align:middle line:84% And broadly speaking, we're going to cover, I think, 00:00:40.550 --> 00:00:42.780 align:middle line:90% two broad themes in that market. 00:00:42.780 --> 00:00:47.940 align:middle line:84% So one is why the allocation of land may matter. 00:00:47.940 --> 00:00:50.150 align:middle line:84% So that's moral-hazard and sharecropping issues 00:00:50.150 --> 00:00:52.760 align:middle line:84% in particular, related to the paper you write today. 00:00:52.760 --> 00:00:56.210 align:middle line:84% But I'll start with some basic theory on this stuff. 00:00:56.210 --> 00:00:59.150 align:middle line:84% And then the other question is, how do we 00:00:59.150 --> 00:01:03.080 align:middle line:84% think about property rights and why 00:01:03.080 --> 00:01:05.090 align:middle line:84% do secure rights over property matter, 00:01:05.090 --> 00:01:07.040 align:middle line:84% how do we think about land titling and things 00:01:07.040 --> 00:01:08.150 align:middle line:84% like that, and the implications of those 00:01:08.150 --> 00:01:09.240 align:middle line:90% for investment decisions. 00:01:09.240 --> 00:01:12.180 align:middle line:84% So mostly, today is going to be moral hazard and sharecropping. 00:01:12.180 --> 00:01:13.610 align:middle line:90% But I guess today is Wednesday. 00:01:13.610 --> 00:01:17.850 align:middle line:84% So Monday we'll be primarily on investment decisions. 00:01:17.850 --> 00:01:24.350 align:middle line:90% 00:01:24.350 --> 00:01:29.480 align:middle line:84% So stepping way back, sharecropping 00:01:29.480 --> 00:01:33.050 align:middle line:84% is a very, very, very, very old subject. 00:01:33.050 --> 00:01:34.970 align:middle line:84% It's been around for forever, essentially, 00:01:34.970 --> 00:01:36.870 align:middle line:90% or for a very, very long time. 00:01:36.870 --> 00:01:40.620 align:middle line:84% And that's when laborers pay owners a share of the output. 00:01:40.620 --> 00:01:43.430 align:middle line:84% So sharecropping is a ubiquitous subject. 00:01:43.430 --> 00:01:45.930 align:middle line:84% And it has been thought about for a very long time, 00:01:45.930 --> 00:01:48.260 align:middle line:84% going back even to economists like Smith and Marshall 00:01:48.260 --> 00:01:50.840 align:middle line:84% were talking about this issue, and not only just talking 00:01:50.840 --> 00:01:53.000 align:middle line:84% about it, but recognizing that there 00:01:53.000 --> 00:01:54.770 align:middle line:84% may be an inefficiency from sharecropping. 00:01:54.770 --> 00:01:57.020 align:middle line:84% And the inefficiency is you don't get the full returns 00:01:57.020 --> 00:01:59.810 align:middle line:84% to your labor because some share of it is going to the landlord. 00:01:59.810 --> 00:02:02.360 align:middle line:84% So that could lead you to not work 00:02:02.360 --> 00:02:03.800 align:middle line:90% as hard as you would otherwise. 00:02:03.800 --> 00:02:06.450 align:middle line:90% That's the basic issue. 00:02:06.450 --> 00:02:09.590 align:middle line:84% So here's just a super-duper-duper simple, 00:02:09.590 --> 00:02:13.400 align:middle line:84% simple model to make this very clear. 00:02:13.400 --> 00:02:16.340 align:middle line:84% So imagine that, basically, there's some output. 00:02:16.340 --> 00:02:19.340 align:middle line:84% The output is of f of i, that's where that's some input. 00:02:19.340 --> 00:02:23.810 align:middle line:84% Tenant pays some cost, c, could be their own labor, whatever. 00:02:23.810 --> 00:02:27.278 align:middle line:84% Owner receives a share, alpha, of the output. 00:02:27.278 --> 00:02:28.820 align:middle line:84% So what's going to happen, the tenant 00:02:28.820 --> 00:02:32.360 align:middle line:90% is going to solve some problem. 00:02:32.360 --> 00:02:35.420 align:middle line:84% They get taxed, essentially, 1 minus alpha on the output. 00:02:35.420 --> 00:02:37.010 align:middle line:90% They pay the full costs. 00:02:37.010 --> 00:02:38.510 align:middle line:84% The first-order condition of that 00:02:38.510 --> 00:02:40.430 align:middle line:84% is just that the marginal product 00:02:40.430 --> 00:02:46.160 align:middle line:84% is equal to the marginal cost, the module of a tax rate. 00:02:46.160 --> 00:02:49.640 align:middle line:84% So if they're being taxed, that's going to be-- 00:02:49.640 --> 00:02:53.180 align:middle line:84% the marginal product is going to be greater than c. 00:02:53.180 --> 00:02:54.950 align:middle line:84% And therefore, if f is concave, they're 00:02:54.950 --> 00:02:57.320 align:middle line:84% going to put less input than they would otherwise. 00:02:57.320 --> 00:03:01.640 align:middle line:84% That's super-duper-duper basic tax-- 00:03:01.640 --> 00:03:03.530 align:middle line:84% if you tax things, people do less of them. 00:03:03.530 --> 00:03:06.290 align:middle line:90% 00:03:06.290 --> 00:03:12.200 align:middle line:84% So one thing that's worth noting, actually, 00:03:12.200 --> 00:03:14.510 align:middle line:90% that's very important is-- 00:03:14.510 --> 00:03:16.610 align:middle line:84% this is the key thing that's driving 00:03:16.610 --> 00:03:19.620 align:middle line:84% this distortion is that this is a tax on output, 00:03:19.620 --> 00:03:22.970 align:middle line:90% not a tax on profits. 00:03:22.970 --> 00:03:25.830 align:middle line:90% So if I wrote this exact same-- 00:03:25.830 --> 00:03:27.350 align:middle line:84% if I changed this so that instead 00:03:27.350 --> 00:03:30.440 align:middle line:84% of having a tax on the output, here, 00:03:30.440 --> 00:03:34.250 align:middle line:84% I taxed your profits, which is your f of i 00:03:34.250 --> 00:03:39.500 align:middle line:84% minus cl, then in that case, it wouldn't matter. 00:03:39.500 --> 00:03:45.380 align:middle line:84% You can just check that the solution, here, is the same 00:03:45.380 --> 00:03:46.800 align:middle line:90% as if there was no-- 00:03:46.800 --> 00:03:48.890 align:middle line:84% I didn't have to share the output here. 00:03:48.890 --> 00:03:51.800 align:middle line:84% And that's this-- general fact is that profit taxes are 00:03:51.800 --> 00:03:53.660 align:middle line:90% nondistortionary. 00:03:53.660 --> 00:03:55.827 align:middle line:84% And that's actually related to broader issues of why 00:03:55.827 --> 00:03:58.077 align:middle line:84% corporate income taxes, for example, taxes on profits, 00:03:58.077 --> 00:03:59.060 align:middle line:90% not taxes on revenues. 00:03:59.060 --> 00:04:03.440 align:middle line:84% It has to do with these kinds of ideas. 00:04:03.440 --> 00:04:06.440 align:middle line:84% Of course, the problem is that what we can observe 00:04:06.440 --> 00:04:09.410 align:middle line:90% is your output. 00:04:09.410 --> 00:04:10.820 align:middle line:84% Although as you saw on the paper, 00:04:10.820 --> 00:04:12.260 align:middle line:84% even observing the output is hard. 00:04:12.260 --> 00:04:14.000 align:middle line:84% But it's at least more plausible that you 00:04:14.000 --> 00:04:16.940 align:middle line:84% can observe the output, rather than observe your profits, 00:04:16.940 --> 00:04:20.480 align:middle line:84% especially when we think of an agricultural-- 00:04:20.480 --> 00:04:22.815 align:middle line:84% like a self-employed farmer or something, 00:04:22.815 --> 00:04:25.190 align:middle line:84% where a lot of their inputs are their own labor or things 00:04:25.190 --> 00:04:27.410 align:middle line:84% like that, that's very difficult to actually observe 00:04:27.410 --> 00:04:28.050 align:middle line:90% and measure. 00:04:28.050 --> 00:04:30.290 align:middle line:84% So the distortion comes from the fact 00:04:30.290 --> 00:04:35.010 align:middle line:84% that we're taxing revenue, not taxing profits. 00:04:35.010 --> 00:04:35.510 align:middle line:90% OK? 00:04:35.510 --> 00:04:37.340 align:middle line:90% Clear? 00:04:37.340 --> 00:04:42.260 align:middle line:84% All right, so the solution-- and actually, 00:04:42.260 --> 00:04:43.790 align:middle line:84% I think one of you-- at least one 00:04:43.790 --> 00:04:48.680 align:middle line:84% of you talked about this in your reading response, 00:04:48.680 --> 00:04:53.570 align:middle line:84% is the solution to this problem is a rental contract. 00:04:53.570 --> 00:04:56.090 align:middle line:84% So faced with this, what do they do? 00:04:56.090 --> 00:04:58.322 align:middle line:84% Rather than have this distortionary contract, what 00:04:58.322 --> 00:04:59.780 align:middle line:84% should happen is, the tenant should 00:04:59.780 --> 00:05:03.230 align:middle line:84% rent the land from the landlord for a fixed rent, r, 00:05:03.230 --> 00:05:05.670 align:middle line:90% and keep all the output. 00:05:05.670 --> 00:05:10.160 align:middle line:84% And so that contract would look like this. 00:05:10.160 --> 00:05:12.950 align:middle line:84% It would look-- the tenant would maximize i minus f-- 00:05:12.950 --> 00:05:16.430 align:middle line:84% sorry, maximize over the inputs, f of I minus cl minus r. 00:05:16.430 --> 00:05:18.660 align:middle line:84% The fact that this r-- r is a lump sum here. 00:05:18.660 --> 00:05:20.340 align:middle line:84% So it doesn't affect the decisions. 00:05:20.340 --> 00:05:23.690 align:middle line:90% So that gets us the first best. 00:05:23.690 --> 00:05:28.070 align:middle line:84% So in some sense, that's the benchmark solution 00:05:28.070 --> 00:05:29.870 align:middle line:84% to all these problems, is we should just-- 00:05:29.870 --> 00:05:34.440 align:middle line:84% the tenant should rent the rent the land for a fixed fee. 00:05:34.440 --> 00:05:36.260 align:middle line:90% So why does that not happen? 00:05:36.260 --> 00:05:39.440 align:middle line:90% 00:05:39.440 --> 00:05:42.595 align:middle line:84% There's a couple of reasons why this doesn't happen. 00:05:42.595 --> 00:05:43.970 align:middle line:84% And that's what I'm going to talk 00:05:43.970 --> 00:05:46.580 align:middle line:84% about for the next half an hour or so. 00:05:46.580 --> 00:05:49.910 align:middle line:90% So one comes from Stiglitz. 00:05:49.910 --> 00:05:54.440 align:middle line:84% And the idea is that there's a trade off between incentives 00:05:54.440 --> 00:05:57.020 align:middle line:90% and risk sharing. 00:05:57.020 --> 00:06:00.920 align:middle line:84% So the problem with this contract 00:06:00.920 --> 00:06:04.400 align:middle line:84% is that if we imagine that the output is risky, 00:06:04.400 --> 00:06:07.550 align:middle line:84% the tenant is bearing all the risk. 00:06:07.550 --> 00:06:11.630 align:middle line:84% And the landlord is getting a fixed fee regardless. 00:06:11.630 --> 00:06:15.560 align:middle line:84% You might imagine that the tenant is typically 00:06:15.560 --> 00:06:18.980 align:middle line:84% a poor farmer and the landlord might be a rich landlord. 00:06:18.980 --> 00:06:22.250 align:middle line:84% In that case, you might imagine that the rich landlord 00:06:22.250 --> 00:06:24.800 align:middle line:84% might be less risk averse than the poor tenant. 00:06:24.800 --> 00:06:27.230 align:middle line:84% But even, in fact, if they're equally risk averse, 00:06:27.230 --> 00:06:29.870 align:middle line:84% the optimal thing is not for one of them 00:06:29.870 --> 00:06:32.570 align:middle line:84% to bear all the risk and the other to be totally insured. 00:06:32.570 --> 00:06:36.590 align:middle line:84% The optimal thing is to share some of the-- 00:06:36.590 --> 00:06:38.542 align:middle line:90% share of the risk between them. 00:06:38.542 --> 00:06:40.250 align:middle line:84% And once you share the risk between them, 00:06:40.250 --> 00:06:43.730 align:middle line:84% then you end up in some kinds of contracts where you end up-- 00:06:43.730 --> 00:06:46.310 align:middle line:84% that look more like sharecropping contracts. 00:06:46.310 --> 00:06:51.450 align:middle line:84% And so that's the basic idea of this Stiglitz model, 00:06:51.450 --> 00:06:54.200 align:middle line:84% which is actually one of the original moral-hazard models, 00:06:54.200 --> 00:06:57.200 align:middle line:84% actually developed with sharecropping as an example, 00:06:57.200 --> 00:06:58.790 align:middle line:90% is to say, well, we'd like to-- 00:06:58.790 --> 00:07:03.380 align:middle line:84% we end up in these contracts that 00:07:03.380 --> 00:07:06.350 align:middle line:84% are productively inefficient because we 00:07:06.350 --> 00:07:08.510 align:middle line:90% want to do some risk sharing. 00:07:08.510 --> 00:07:12.715 align:middle line:84% So let me go through a little bit of that in a little-- go 00:07:12.715 --> 00:07:14.340 align:middle line:84% through that a little bit [INAUDIBLE].. 00:07:14.340 --> 00:07:15.650 align:middle line:90% Any questions? 00:07:15.650 --> 00:07:16.780 align:middle line:90% No? 00:07:16.780 --> 00:07:18.900 align:middle line:90% Clear? 00:07:18.900 --> 00:07:23.555 align:middle line:90% OK, so that's basically what I-- 00:07:23.555 --> 00:07:25.680 align:middle line:84% so just to be clear, the idea of this, that farming 00:07:25.680 --> 00:07:27.180 align:middle line:84% is going to be risky, so we're going 00:07:27.180 --> 00:07:29.610 align:middle line:84% to put some uncertainty on this, we 00:07:29.610 --> 00:07:32.820 align:middle line:84% know risk-averse agents want to be insured against this risk. 00:07:32.820 --> 00:07:35.370 align:middle line:84% But the problem is, if this effort choice 00:07:35.370 --> 00:07:41.530 align:middle line:84% is uncontractable, that leads to inefficiencies. 00:07:41.530 --> 00:07:44.310 align:middle line:90% So here is a very simple-- 00:07:44.310 --> 00:07:46.830 align:middle line:84% like the simplest possible kind of moral-- 00:07:46.830 --> 00:07:49.690 align:middle line:90% simplest version of this model. 00:07:49.690 --> 00:07:55.635 align:middle line:84% So let's consider a two-state version, high and low output. 00:07:55.635 --> 00:07:58.500 align:middle line:90% 00:07:58.500 --> 00:08:03.630 align:middle line:84% There's effort, which is denoted by e. 00:08:03.630 --> 00:08:06.060 align:middle line:84% The farmer chooses some effort, e. 00:08:06.060 --> 00:08:07.590 align:middle line:90% But the landlord doesn't see e. 00:08:07.590 --> 00:08:09.600 align:middle line:90% They only see output. 00:08:09.600 --> 00:08:10.555 align:middle line:90% So we can't write-- 00:08:10.555 --> 00:08:12.930 align:middle line:84% and the key point is we can't, therefore, write contracts 00:08:12.930 --> 00:08:15.278 align:middle line:90% over the level of effort, e. 00:08:15.278 --> 00:08:17.820 align:middle line:84% Effort's going to be costly to the tenant with some costs 1/2 00:08:17.820 --> 00:08:19.260 align:middle line:90% c squared. 00:08:19.260 --> 00:08:23.340 align:middle line:84% And the output is going to be stochastic, 00:08:23.340 --> 00:08:27.990 align:middle line:84% with the effort determining the probability of the high state. 00:08:27.990 --> 00:08:32.440 align:middle line:84% So with probability e, the output is high. 00:08:32.440 --> 00:08:38.159 align:middle line:84% And with probability 1 minus e, the output is going to be 0. 00:08:38.159 --> 00:08:41.387 align:middle line:84% And the way we think about these models is 00:08:41.387 --> 00:08:42.929 align:middle line:84% the farmer and the landlord are going 00:08:42.929 --> 00:08:45.210 align:middle line:84% to write a contract which is going 00:08:45.210 --> 00:08:51.060 align:middle line:84% to specify a payment to the farmer, from the landlord-- 00:08:51.060 --> 00:08:52.800 align:middle line:90% I think that's the way I did it. 00:08:52.800 --> 00:08:56.470 align:middle line:84% Doesn't matter which way you write the payment-- 00:08:56.470 --> 00:09:00.510 align:middle line:84% which is going to be a payment h, little h, if the output is 00:09:00.510 --> 00:09:04.087 align:middle line:84% big H, and a payment, little l, if the output is zero. 00:09:04.087 --> 00:09:06.420 align:middle line:84% So we're going to be interested in what those little h's 00:09:06.420 --> 00:09:07.665 align:middle line:90% and little l's are. 00:09:07.665 --> 00:09:10.360 align:middle line:90% OK, so it's set up clear? 00:09:10.360 --> 00:09:12.360 align:middle line:84% So just to be clear, the key thing in this model 00:09:12.360 --> 00:09:15.270 align:middle line:84% is that the effort choice is going 00:09:15.270 --> 00:09:19.800 align:middle line:84% to drive the probability of the high state or the low state. 00:09:19.800 --> 00:09:25.240 align:middle line:90% OK, so what's the first best? 00:09:25.240 --> 00:09:25.980 align:middle line:90% Yeah, Davey? 00:09:25.980 --> 00:09:27.897 align:middle line:84% AUDIENCE: The solution-- like here, the output 00:09:27.897 --> 00:09:29.600 align:middle line:90% is [INAUDIBLE]. 00:09:29.600 --> 00:09:31.010 align:middle line:90% BEN OLKEN: Output is observable. 00:09:31.010 --> 00:09:31.510 align:middle line:90% What's that? 00:09:31.510 --> 00:09:33.380 align:middle line:90% AUDIENCE: And contractable. 00:09:33.380 --> 00:09:34.880 align:middle line:90% BEN OLKEN: Yes, exactly. 00:09:34.880 --> 00:09:36.110 align:middle line:90% The output is observable. 00:09:36.110 --> 00:09:39.200 align:middle line:84% We're going to write contracts on output. 00:09:39.200 --> 00:09:42.770 align:middle line:84% But we can't write contracts on effort. 00:09:42.770 --> 00:09:45.770 align:middle line:84% So basically-- so the way this model is going to work is, 00:09:45.770 --> 00:09:47.960 align:middle line:84% the landlord is going to set up the contract-- 00:09:47.960 --> 00:09:50.660 align:middle line:84% here's the little h, here's the little l-- 00:09:50.660 --> 00:09:53.330 align:middle line:84% which are functions of what the realized output is. 00:09:53.330 --> 00:09:56.025 align:middle line:84% And then the tenant gets to observe h and l 00:09:56.025 --> 00:09:58.400 align:middle line:84% and make their decisions about how much effort to put in. 00:09:58.400 --> 00:10:00.860 align:middle line:84% And the landlord gets to set-- and the landlord, in turn, 00:10:00.860 --> 00:10:02.450 align:middle line:84% chooses the little h and little l, 00:10:02.450 --> 00:10:04.640 align:middle line:84% knowing how the tenant will respond and make 00:10:04.640 --> 00:10:07.970 align:middle line:90% their contract decisions. 00:10:07.970 --> 00:10:09.720 align:middle line:84% So this is a standard moral-hazard set up. 00:10:09.720 --> 00:10:11.345 align:middle line:84% But not everyone has seen these before. 00:10:11.345 --> 00:10:13.132 align:middle line:90% So I want to go through this. 00:10:13.132 --> 00:10:16.470 align:middle line:90% OK, clear? 00:10:16.470 --> 00:10:18.570 align:middle line:84% All right, so what's the first best? 00:10:18.570 --> 00:10:24.400 align:middle line:84% Well, the first best would be that-- 00:10:24.400 --> 00:10:28.337 align:middle line:84% so remember in the low state, we get 0, just by normalization. 00:10:28.337 --> 00:10:30.670 align:middle line:84% So that's going to make our-- just make the math easier. 00:10:30.670 --> 00:10:31.890 align:middle line:90% So what's the first best? 00:10:31.890 --> 00:10:37.380 align:middle line:84% Well, the social optimum is to choose the level of effort 00:10:37.380 --> 00:10:39.105 align:middle line:90% such that the return-- 00:10:39.105 --> 00:10:40.980 align:middle line:84% the expected return, which is the probability 00:10:40.980 --> 00:10:44.610 align:middle line:84% of getting the high return times the value of the high return, 00:10:44.610 --> 00:10:48.100 align:middle line:84% you balance that against the cost of effort. 00:10:48.100 --> 00:10:50.750 align:middle line:84% So we just take the derivative here, and e 00:10:50.750 --> 00:10:53.430 align:middle line:90% is equal to capital H over c. 00:10:53.430 --> 00:10:56.310 align:middle line:84% That's the optimal first-best level of effort. 00:10:56.310 --> 00:11:02.020 align:middle line:90% 00:11:02.020 --> 00:11:04.630 align:middle line:90% So that's the benchmark. 00:11:04.630 --> 00:11:06.130 align:middle line:84% The question is, are we going to get 00:11:06.130 --> 00:11:07.495 align:middle line:90% that level of effort or not? 00:11:07.495 --> 00:11:10.297 align:middle line:90% 00:11:10.297 --> 00:11:11.380 align:middle line:90% So what's going to happen? 00:11:11.380 --> 00:11:14.470 align:middle line:84% So the question is, can the land-- by the way, 00:11:14.470 --> 00:11:17.220 align:middle line:84% if I'm going too fast, you should let me-- 00:11:17.220 --> 00:11:18.128 align:middle line:90% slow me down. 00:11:18.128 --> 00:11:19.920 align:middle line:84% That's the problem with teaching on slides, 00:11:19.920 --> 00:11:21.090 align:middle line:84% teaching models on slides, is you 00:11:21.090 --> 00:11:22.757 align:middle line:84% tend to go faster than people can write. 00:11:22.757 --> 00:11:24.670 align:middle line:84% So if I'm going too fast, just slow me down. 00:11:24.670 --> 00:11:27.120 align:middle line:84% So can the landlord implement this first-best contract? 00:11:27.120 --> 00:11:28.690 align:middle line:90% What's going to actually happen? 00:11:28.690 --> 00:11:30.130 align:middle line:90% So how do we solve this? 00:11:30.130 --> 00:11:35.950 align:middle line:84% So the problem for the landlord is written as the following. 00:11:35.950 --> 00:11:39.930 align:middle line:84% So the landlord has two choices, what is 00:11:39.930 --> 00:11:43.650 align:middle line:90% little h and what is little l. 00:11:43.650 --> 00:11:46.230 align:middle line:84% And what is the landlord care about? 00:11:46.230 --> 00:11:50.190 align:middle line:84% Well, with probability e, which will 00:11:50.190 --> 00:11:55.200 align:middle line:84% be determined endogenously, the landlord-- the output is big H. 00:11:55.200 --> 00:11:57.617 align:middle line:84% And the landlord has to make a payment, little h. 00:11:57.617 --> 00:11:59.700 align:middle line:84% So that's the landlord's return in the high state, 00:11:59.700 --> 00:12:00.882 align:middle line:90% was probably 1 minus e. 00:12:00.882 --> 00:12:01.840 align:middle line:90% We're in the low state. 00:12:01.840 --> 00:12:04.947 align:middle line:84% And then the landlord has to pay l and gets no output. 00:12:04.947 --> 00:12:06.780 align:middle line:84% So that's the landlord's objective function. 00:12:06.780 --> 00:12:07.980 align:middle line:84% By the way, this is all going to be risk neutral, 00:12:07.980 --> 00:12:09.452 align:middle line:90% as you'll see in a second-- 00:12:09.452 --> 00:12:11.160 align:middle line:84% but we can introduce risk aversion later. 00:12:11.160 --> 00:12:17.010 align:middle line:84% But-- so OK, so that's the landlord's problem-- 00:12:17.010 --> 00:12:19.423 align:middle line:84% and has two constraints, the incentive-compatibility 00:12:19.423 --> 00:12:21.090 align:middle line:84% constraint in the individual-rationality 00:12:21.090 --> 00:12:23.010 align:middle line:90% constraint. 00:12:23.010 --> 00:12:24.750 align:middle line:84% The incentive-compatibility constraint 00:12:24.750 --> 00:12:29.010 align:middle line:84% says the landlord is going to solve this problem, knowing 00:12:29.010 --> 00:12:34.200 align:middle line:84% that the tenant will in turn take the contract as fixed, 00:12:34.200 --> 00:12:39.130 align:middle line:84% and choose her level of effort accordingly. 00:12:39.130 --> 00:12:47.110 align:middle line:84% So the tenant, here, is going to solve for a level-- 00:12:47.110 --> 00:12:48.660 align:middle line:84% so the level of effort, e, is going 00:12:48.660 --> 00:12:51.090 align:middle line:84% to be chosen by the tenant, to solve the tenant's 00:12:51.090 --> 00:12:54.150 align:middle line:84% own objective function, which is, well, with probability e, 00:12:54.150 --> 00:12:57.510 align:middle line:90% I get payment, little h. 00:12:57.510 --> 00:12:59.910 align:middle line:84% With probability 1 minus e, I get payment, little l. 00:12:59.910 --> 00:13:03.600 align:middle line:84% And I have to pay my cost of effort. 00:13:03.600 --> 00:13:06.300 align:middle line:84% And there's a third constraint, which 00:13:06.300 --> 00:13:08.592 align:middle line:84% is the individual-rationality constraint. 00:13:08.592 --> 00:13:10.800 align:middle line:84% And the way to think about the individual-rationality 00:13:10.800 --> 00:13:15.090 align:middle line:84% constraint is that the tenant has some outside option, w bar. 00:13:15.090 --> 00:13:19.950 align:middle line:84% And the tenant's net utility has to be at least equal to w bar, 00:13:19.950 --> 00:13:21.900 align:middle line:90% to be low bar. 00:13:21.900 --> 00:13:25.080 align:middle line:84% And that's to say, the tenant could not take this contract, 00:13:25.080 --> 00:13:28.860 align:middle line:84% and could walk away and just get a job as a day laborer 00:13:28.860 --> 00:13:30.240 align:middle line:90% at some wage rate, w. 00:13:30.240 --> 00:13:32.790 align:middle line:84% So the net utility from the tenant-- 00:13:32.790 --> 00:13:37.680 align:middle line:84% so that's this term, here, evaluated at the chosen e, 00:13:37.680 --> 00:13:43.250 align:middle line:84% has to be at least as high as her outside option. 00:13:43.250 --> 00:13:45.815 align:middle line:90% So this is the problem. 00:13:45.815 --> 00:13:46.315 align:middle line:90% Clear? 00:13:46.315 --> 00:13:49.090 align:middle line:90% 00:13:49.090 --> 00:13:51.790 align:middle line:90% OK. 00:13:51.790 --> 00:13:54.190 align:middle line:84% And those of you have taken contract theory already, 00:13:54.190 --> 00:13:55.555 align:middle line:90% this should be very familiar. 00:13:55.555 --> 00:13:59.860 align:middle line:90% 00:13:59.860 --> 00:14:02.750 align:middle line:90% So how do we solve this problem? 00:14:02.750 --> 00:14:05.740 align:middle line:84% Well, we begin by solving for the-- taking 00:14:05.740 --> 00:14:11.300 align:middle line:84% the farmer's solution, taking the contract-- the ic is given. 00:14:11.300 --> 00:14:16.570 align:middle line:84% So we basically say, let's first solve the farmer's problem. 00:14:16.570 --> 00:14:19.540 align:middle line:84% Given the little h and little l, what is the farmer going to do? 00:14:19.540 --> 00:14:25.030 align:middle line:84% And then we'll plug that back in to the landlord's problem. 00:14:25.030 --> 00:14:26.590 align:middle line:90% So what's the farmer's decision? 00:14:26.590 --> 00:14:28.180 align:middle line:84% Well, the farmer is going to solve-- 00:14:28.180 --> 00:14:31.150 align:middle line:84% is going to maximize the same with this objective function 00:14:31.150 --> 00:14:32.890 align:middle line:90% over here. 00:14:32.890 --> 00:14:34.598 align:middle line:84% Remember she gets payment, little h, 00:14:34.598 --> 00:14:36.140 align:middle line:84% with probability e, payment little l, 00:14:36.140 --> 00:14:37.265 align:middle line:90% with probability 1 minus e. 00:14:37.265 --> 00:14:38.480 align:middle line:90% She pays the cost of effort. 00:14:38.480 --> 00:14:41.540 align:middle line:84% So her effort decision, as you can see-- 00:14:41.540 --> 00:14:44.560 align:middle line:84% so we can just take the derivative of this e, 00:14:44.560 --> 00:14:49.960 align:middle line:84% the difference is, she's going to get h minus-- 00:14:49.960 --> 00:14:53.170 align:middle line:84% what additional e does, is it increases the probability she 00:14:53.170 --> 00:14:56.305 align:middle line:84% gets h minus l, the difference in her-- 00:14:56.305 --> 00:14:58.180 align:middle line:84% the difference of the high state is different 00:14:58.180 --> 00:14:59.500 align:middle line:84% than her payment in the state minus the difference 00:14:59.500 --> 00:15:01.690 align:middle line:84% in the payment in the low state, and just 00:15:01.690 --> 00:15:04.160 align:middle line:84% is equal to the marginal cost of effort. 00:15:04.160 --> 00:15:07.180 align:middle line:90% So her choice is going to be-- 00:15:07.180 --> 00:15:12.640 align:middle line:84% e star is going to be equal little h minus little l over c. 00:15:12.640 --> 00:15:15.160 align:middle line:84% So what you can immediately see here 00:15:15.160 --> 00:15:16.725 align:middle line:84% is, this is a model where, if there's 00:15:16.725 --> 00:15:18.850 align:middle line:84% no difference between the payment in the high state 00:15:18.850 --> 00:15:20.650 align:middle line:84% and the low state, if she's fully insured, 00:15:20.650 --> 00:15:22.510 align:middle line:90% she puts in no effort. 00:15:22.510 --> 00:15:26.590 align:middle line:84% So there's going to have to be some-- 00:15:26.590 --> 00:15:28.780 align:middle line:84% this is a model where, if there's full insurance, 00:15:28.780 --> 00:15:31.010 align:middle line:90% there's going to be no work. 00:15:31.010 --> 00:15:35.860 align:middle line:84% So they have to provide some incentives in order 00:15:35.860 --> 00:15:39.910 align:middle line:84% to get her to put in some effort, otherwise why would 00:15:39.910 --> 00:15:43.250 align:middle line:84% you do anything, because there's some cost of effort. 00:15:43.250 --> 00:15:43.750 align:middle line:90% Clear? 00:15:43.750 --> 00:15:46.870 align:middle line:90% 00:15:46.870 --> 00:15:53.710 align:middle line:84% All right, so now, the landlord-- 00:15:53.710 --> 00:15:56.600 align:middle line:84% so we can also see-- we can all see immediately, by the way, 00:15:56.600 --> 00:15:58.460 align:middle line:90% that to get the first best-- 00:15:58.460 --> 00:16:00.910 align:middle line:84% remember, the first-best effort-- 00:16:00.910 --> 00:16:03.400 align:middle line:84% doo-doo-doo-doo-- first best effort is we'd like 00:16:03.400 --> 00:16:05.005 align:middle line:90% the landlord to-- 00:16:05.005 --> 00:16:07.210 align:middle line:84% I'm sorry-- we'd like the tenant to put 00:16:07.210 --> 00:16:10.000 align:middle line:90% an effort equal to big h over c. 00:16:10.000 --> 00:16:11.600 align:middle line:90% So what does that mean? 00:16:11.600 --> 00:16:15.148 align:middle line:84% Well, the contract-- any contract 00:16:15.148 --> 00:16:16.690 align:middle line:84% that's going to get the tenant to put 00:16:16.690 --> 00:16:19.390 align:middle line:84% in the full amount of effort, that's 00:16:19.390 --> 00:16:23.830 align:middle line:84% going to imply that little h minus little l is 00:16:23.830 --> 00:16:28.300 align:middle line:84% equal to big H. Which is another way of saying it, 00:16:28.300 --> 00:16:30.640 align:middle line:84% is that the gap between what she gets in the high state 00:16:30.640 --> 00:16:32.110 align:middle line:84% and what she gets in the low state 00:16:32.110 --> 00:16:35.440 align:middle line:84% has to be equal to the social return of being 00:16:35.440 --> 00:16:38.520 align:middle line:84% in the high state versus the lowest state. 00:16:38.520 --> 00:16:41.790 align:middle line:84% That's the only way to get her to put in full effort. 00:16:41.790 --> 00:16:44.490 align:middle line:84% So you see that by saying, look, if we 00:16:44.490 --> 00:16:46.320 align:middle line:84% want this to equal to big h over c, 00:16:46.320 --> 00:16:48.880 align:middle line:84% is we have to make the difference between little h 00:16:48.880 --> 00:16:52.170 align:middle line:90% little l equal to big H. 00:16:52.170 --> 00:16:54.208 align:middle line:90% So that is rental. 00:16:54.208 --> 00:16:54.750 align:middle line:90% What is that? 00:16:54.750 --> 00:16:56.998 align:middle line:90% That's a rental contract. 00:16:56.998 --> 00:16:58.540 align:middle line:84% We've just derived a rental contract. 00:16:58.540 --> 00:16:59.790 align:middle line:90% Why is that a rental contract? 00:16:59.790 --> 00:17:03.990 align:middle line:84% Because the tenant will make some fixed payment 00:17:03.990 --> 00:17:05.710 align:middle line:90% to the landlord-- 00:17:05.710 --> 00:17:08.130 align:middle line:84% another way of viewing it, any contract 00:17:08.130 --> 00:17:10.560 align:middle line:84% where the difference between little h and little l 00:17:10.560 --> 00:17:16.079 align:middle line:84% is equal to big H is a contract where she makes some fixed 00:17:16.079 --> 00:17:17.917 align:middle line:90% payment to the landlord. 00:17:17.917 --> 00:17:19.500 align:middle line:84% Or maybe goes the other way, depending 00:17:19.500 --> 00:17:21.569 align:middle line:84% on what her incentive for that-- what her IR constraint is. 00:17:21.569 --> 00:17:23.944 align:middle line:84% There's some fixed payment that goes one way or the other 00:17:23.944 --> 00:17:28.130 align:middle line:84% and then she keeps all the output that is produced. 00:17:28.130 --> 00:17:30.760 align:middle line:90% So that's a rental contract. 00:17:30.760 --> 00:17:32.310 align:middle line:84% So we've just derived the result, 00:17:32.310 --> 00:17:35.130 align:middle line:84% that basically-- another way of saying it is, in this model, 00:17:35.130 --> 00:17:38.040 align:middle line:84% the only way to get the first-best effort 00:17:38.040 --> 00:17:42.870 align:middle line:84% is to have a rental contract, which is basically 00:17:42.870 --> 00:17:44.120 align:middle line:90% what I just said on the slide. 00:17:44.120 --> 00:17:44.890 align:middle line:90% So why is that? 00:17:44.890 --> 00:17:46.840 align:middle line:84% We need the worker to face the socially-optimal return 00:17:46.840 --> 00:17:47.390 align:middle line:90% to effort. 00:17:47.390 --> 00:17:48.550 align:middle line:90% So we need e stars-- 00:17:48.550 --> 00:17:52.330 align:middle line:84% we need this, her choice of effort, 00:17:52.330 --> 00:17:54.140 align:middle line:84% to be equal to the socially-optimal level. 00:17:54.140 --> 00:17:56.390 align:middle line:84% The only way to do that is to set h minus l 00:17:56.390 --> 00:17:58.390 align:middle line:84% equals to big H, which basically means she keeps 00:17:58.390 --> 00:18:00.175 align:middle line:90% all the return on the market. 00:18:00.175 --> 00:18:02.700 align:middle line:90% 00:18:02.700 --> 00:18:08.120 align:middle line:90% Now in this contract-- 00:18:08.120 --> 00:18:11.552 align:middle line:84% now we've pinned down little h minus little l. 00:18:11.552 --> 00:18:13.010 align:middle line:84% But we haven't pinned down-- but we 00:18:13.010 --> 00:18:16.250 align:middle line:84% can slide the level of those things up and down. 00:18:16.250 --> 00:18:18.590 align:middle line:90% That's the rent. 00:18:18.590 --> 00:18:20.210 align:middle line:84% And so how do we choose the rent? 00:18:20.210 --> 00:18:22.760 align:middle line:84% We choose the rent so that the farmer 00:18:22.760 --> 00:18:27.230 align:middle line:84% is going to obtain w lower bar and expectation. 00:18:27.230 --> 00:18:30.250 align:middle line:90% That's the IR constraint. 00:18:30.250 --> 00:18:32.270 align:middle line:84% So the farmer's utility-- remember the farmer-- 00:18:32.270 --> 00:18:33.520 align:middle line:90% what was the farmer's utility? 00:18:33.520 --> 00:18:36.940 align:middle line:90% The farmer's utility was-- 00:18:36.940 --> 00:18:40.600 align:middle line:84% so now, in the first-best contract, with probability e, 00:18:40.600 --> 00:18:41.665 align:middle line:90% the high state happens. 00:18:41.665 --> 00:18:45.400 align:middle line:90% 00:18:45.400 --> 00:18:47.290 align:middle line:90% And she gets h. 00:18:47.290 --> 00:18:50.500 align:middle line:90% The rent gets paid either way. 00:18:50.500 --> 00:18:52.090 align:middle line:84% No matter what, you pay the rent. 00:18:52.090 --> 00:18:54.250 align:middle line:84% And you have to put in the cost of effort. 00:18:54.250 --> 00:18:57.040 align:middle line:84% And we know, by the way, that the effort, here, 00:18:57.040 --> 00:18:59.290 align:middle line:84% is going to be equal to big H over c. 00:18:59.290 --> 00:19:02.620 align:middle line:90% So we can sub in big H over c. 00:19:02.620 --> 00:19:06.880 align:middle line:84% So substituting in big H over c, here, her utility 00:19:06.880 --> 00:19:12.100 align:middle line:84% is going to be H square over c, minus 1/2 H square over c. 00:19:12.100 --> 00:19:14.140 align:middle line:84% So we get half an H square over-- 00:19:14.140 --> 00:19:18.820 align:middle line:84% sorry-- yeah, so we get a 1/2 an H 00:19:18.820 --> 00:19:22.390 align:middle line:90% square over c, minus the rent. 00:19:22.390 --> 00:19:24.820 align:middle line:84% So that's her utility under this contract, her expected 00:19:24.820 --> 00:19:26.380 align:middle line:90% utility under this contract. 00:19:26.380 --> 00:19:28.420 align:middle line:84% And that has to be equal to the outside option. 00:19:28.420 --> 00:19:31.930 align:middle line:84% So that we're going to set that equal to w, lower bar. 00:19:31.930 --> 00:19:32.988 align:middle line:90% And so that's the rent. 00:19:32.988 --> 00:19:34.030 align:middle line:90% That determines the rent. 00:19:34.030 --> 00:19:37.270 align:middle line:84% So the rent is going to be the level where she's exactly 00:19:37.270 --> 00:19:40.525 align:middle line:90% in different ex-ante. 00:19:40.525 --> 00:19:42.150 align:middle line:84% And that's a general feature of the way 00:19:42.150 --> 00:19:43.830 align:middle line:84% that you solve a lot of these problems, 00:19:43.830 --> 00:19:45.622 align:middle line:84% is you figure out-- you first kind of solve 00:19:45.622 --> 00:19:47.910 align:middle line:84% for what the ic constraint and what 00:19:47.910 --> 00:19:52.057 align:middle line:84% is the difference between the-- you solve the problem to say, 00:19:52.057 --> 00:19:54.390 align:middle line:84% what's the difference between the highest and the lowest 00:19:54.390 --> 00:19:54.960 align:middle line:90% payoff. 00:19:54.960 --> 00:19:57.220 align:middle line:84% And then you slide the whole thing up or down 00:19:57.220 --> 00:19:59.550 align:middle line:84% so that the agent is exactly indifferent 00:19:59.550 --> 00:20:03.930 align:middle line:84% to their outside option in their IR constraint. 00:20:03.930 --> 00:20:05.580 align:middle line:84% So that's the general way you solve 00:20:05.580 --> 00:20:08.350 align:middle line:90% these moral-hazard problems. 00:20:08.350 --> 00:20:09.210 align:middle line:90% So that's the rent. 00:20:09.210 --> 00:20:12.150 align:middle line:90% That pins down the rent. 00:20:12.150 --> 00:20:13.580 align:middle line:90% OK. 00:20:13.580 --> 00:20:14.080 align:middle line:90% Clear? 00:20:14.080 --> 00:20:17.200 align:middle line:90% 00:20:17.200 --> 00:20:20.050 align:middle line:84% OK, so that's the final contract. 00:20:20.050 --> 00:20:22.810 align:middle line:84% Final contract, now-- so now we know both the little h 00:20:22.810 --> 00:20:24.190 align:middle line:90% and little l. 00:20:24.190 --> 00:20:28.870 align:middle line:84% So little h-- sorry, so our r, here, was 00:20:28.870 --> 00:20:33.250 align:middle line:90% h square over 2c minus w bar. 00:20:33.250 --> 00:20:34.630 align:middle line:90% So we can substitute in. 00:20:34.630 --> 00:20:43.860 align:middle line:84% So we get-- so the payment is 1 minus the r 00:20:43.860 --> 00:20:45.110 align:middle line:90% because it goes the other way. 00:20:45.110 --> 00:20:49.720 align:middle line:90% So this is minus r. 00:20:49.720 --> 00:20:54.490 align:middle line:84% That happens in either state of the world. 00:20:54.490 --> 00:20:59.180 align:middle line:84% And the difference is she gets to keep the output. 00:20:59.180 --> 00:21:01.270 align:middle line:84% So that's just another way of writing-- 00:21:01.270 --> 00:21:06.273 align:middle line:84% there's a rent equal to H square over 2 c, minus w bar. 00:21:06.273 --> 00:21:06.940 align:middle line:90% That's the rent. 00:21:06.940 --> 00:21:09.490 align:middle line:84% And then she keeps the return, the h. 00:21:09.490 --> 00:21:11.140 align:middle line:84% That's just the way of expressing it 00:21:11.140 --> 00:21:12.445 align:middle line:90% in this contract space. 00:21:12.445 --> 00:21:15.930 align:middle line:90% 00:21:15.930 --> 00:21:17.490 align:middle line:90% Is that all clear so far? 00:21:17.490 --> 00:21:20.500 align:middle line:90% 00:21:20.500 --> 00:21:21.000 align:middle line:90% All right. 00:21:21.000 --> 00:21:23.650 align:middle line:90% 00:21:23.650 --> 00:21:27.100 align:middle line:84% So this is a nice contract because the farmer 00:21:27.100 --> 00:21:29.320 align:middle line:90% is going to-- 00:21:29.320 --> 00:21:31.780 align:middle line:90% the farmer gets no rents. 00:21:31.780 --> 00:21:33.760 align:middle line:84% It's nice from the perspective of the landlord. 00:21:33.760 --> 00:21:34.690 align:middle line:90% The landlord likes his contract. 00:21:34.690 --> 00:21:36.160 align:middle line:90% The farmer is getting no rents. 00:21:36.160 --> 00:21:40.150 align:middle line:84% She's getting exactly equal to her outside option 00:21:40.150 --> 00:21:41.660 align:middle line:90% in expectation. 00:21:41.660 --> 00:21:43.570 align:middle line:84% So the landlord is capturing all the surplus. 00:21:43.570 --> 00:21:45.280 align:middle line:90% And she puts it in full effort. 00:21:45.280 --> 00:21:46.420 align:middle line:90% So that's great. 00:21:46.420 --> 00:21:48.220 align:middle line:90% Landlord loves this contract. 00:21:48.220 --> 00:21:51.685 align:middle line:84% Problem is, there's two challenges with this contract. 00:21:51.685 --> 00:21:55.540 align:middle line:90% 00:21:55.540 --> 00:21:57.160 align:middle line:84% The first challenge with this contract 00:21:57.160 --> 00:21:58.885 align:middle line:84% is the farmer is bearing all of the risk. 00:21:58.885 --> 00:22:02.670 align:middle line:90% 00:22:02.670 --> 00:22:06.690 align:middle line:90% So that's the first problem. 00:22:06.690 --> 00:22:12.020 align:middle line:84% Second problem is, the farmer, with positive probability, 00:22:12.020 --> 00:22:17.330 align:middle line:84% is going to earn this lower-- this l. 00:22:17.330 --> 00:22:18.590 align:middle line:90% And that may be negative. 00:22:18.590 --> 00:22:21.320 align:middle line:90% 00:22:21.320 --> 00:22:25.580 align:middle line:90% So she may get less than 0. 00:22:25.580 --> 00:22:31.418 align:middle line:84% So what happens if she has no money and can't pay? 00:22:31.418 --> 00:22:32.960 align:middle line:84% So then we have a second issue, which 00:22:32.960 --> 00:22:34.585 align:middle line:84% is what we can think of what are called 00:22:34.585 --> 00:22:39.440 align:middle line:84% limited-liability contracts, where you're capped at 0 00:22:39.440 --> 00:22:40.430 align:middle line:90% from below. 00:22:40.430 --> 00:22:42.590 align:middle line:90% You can't ever go below zero. 00:22:42.590 --> 00:22:46.160 align:middle line:84% If you have no money, you can't pay the rent. 00:22:46.160 --> 00:22:50.870 align:middle line:84% And so this one is like a risk-- an insurance issue. 00:22:50.870 --> 00:22:53.120 align:middle line:84% And this is a limited-liability issue. 00:22:53.120 --> 00:22:56.090 align:middle line:84% And both of these can give rise to what 00:22:56.090 --> 00:22:58.580 align:middle line:84% looks like sharecropping arrangements. 00:22:58.580 --> 00:23:00.290 align:middle line:84% But the implications of them are going 00:23:00.290 --> 00:23:01.663 align:middle line:90% to be a little bit different. 00:23:01.663 --> 00:23:04.080 align:middle line:84% So that's what I want to go through next, is to say, well, 00:23:04.080 --> 00:23:05.580 align:middle line:84% how do we think about each of those? 00:23:05.580 --> 00:23:09.290 align:middle line:84% And how could each of those different reasons 00:23:09.290 --> 00:23:14.110 align:middle line:84% lead us to a sharecropping arrangement? 00:23:14.110 --> 00:23:16.340 align:middle line:90% Questions? 00:23:16.340 --> 00:23:16.840 align:middle line:90% All clear? 00:23:16.840 --> 00:23:17.270 align:middle line:90% Yeah? 00:23:17.270 --> 00:23:18.937 align:middle line:84% AUDIENCE: In practice, limited liability 00:23:18.937 --> 00:23:21.104 align:middle line:90% just means that I can't go-- 00:23:21.104 --> 00:23:22.096 align:middle line:90% I can only take-- 00:23:22.096 --> 00:23:26.560 align:middle line:84% I can't force her to take out [INAUDIBLE]?? 00:23:26.560 --> 00:23:28.690 align:middle line:90% BEN OLKEN: Yeah, so correct. 00:23:28.690 --> 00:23:32.200 align:middle line:84% So when so when might limited liability be a thing 00:23:32.200 --> 00:23:33.200 align:middle line:90% or might not be a thing? 00:23:33.200 --> 00:23:35.020 align:middle line:90% So yes, right. 00:23:35.020 --> 00:23:37.817 align:middle line:90% So you can't-- so you're right. 00:23:37.817 --> 00:23:40.150 align:middle line:84% It means we can't force you to go take out a loan to pay 00:23:40.150 --> 00:23:41.545 align:middle line:90% the rent, for example. 00:23:41.545 --> 00:23:46.330 align:middle line:90% 00:23:46.330 --> 00:23:50.270 align:middle line:84% Although even in debt contracts, why might that be? 00:23:50.270 --> 00:23:53.560 align:middle line:84% Well, you could just leave, for example. 00:23:53.560 --> 00:23:57.310 align:middle line:90% It can be hard to-- 00:23:57.310 --> 00:23:58.870 align:middle line:90% yeah, so that's basically the-- 00:23:58.870 --> 00:24:01.450 align:middle line:84% and we'll talk about those contracts, limited liability, 00:24:01.450 --> 00:24:03.250 align:middle line:84% also, when we get to the credit lectures 00:24:03.250 --> 00:24:05.770 align:middle line:84% at the end of the semester, where 00:24:05.770 --> 00:24:09.920 align:middle line:84% a lot of these same issues are going to come up. 00:24:09.920 --> 00:24:10.460 align:middle line:90% Yeah. 00:24:10.460 --> 00:24:12.200 align:middle line:90% But you could leave. 00:24:12.200 --> 00:24:13.580 align:middle line:90% You could be hard to enforce. 00:24:13.580 --> 00:24:16.430 align:middle line:90% There's a variety of reasons. 00:24:16.430 --> 00:24:17.630 align:middle line:90% Or you might-- right. 00:24:17.630 --> 00:24:19.005 align:middle line:84% So you might have asked her that. 00:24:19.005 --> 00:24:20.620 align:middle line:90% Yeah. 00:24:20.620 --> 00:24:21.460 align:middle line:90% Other questions? 00:24:21.460 --> 00:24:24.460 align:middle line:90% 00:24:24.460 --> 00:24:26.970 align:middle line:90% So let's explore both of these. 00:24:26.970 --> 00:24:30.380 align:middle line:84% So what happens when the farmer is risk averse? 00:24:30.380 --> 00:24:38.550 align:middle line:84% So now let's assume that the landlord is risk neutral, 00:24:38.550 --> 00:24:44.340 align:middle line:84% but the farmer has some utility function, 00:24:44.340 --> 00:24:48.510 align:middle line:84% which is concave over consumption in both states. 00:24:48.510 --> 00:24:51.858 align:middle line:84% And actually, you can do this with, as I said, 00:24:51.858 --> 00:24:53.400 align:middle line:84% with both parties being race neutral. 00:24:53.400 --> 00:24:56.700 align:middle line:84% Actually, it doesn't-- just the math gets more complicated. 00:24:56.700 --> 00:25:00.570 align:middle line:84% Actually, all these moral-hazard models, actually, the math 00:25:00.570 --> 00:25:02.718 align:middle line:90% can get a little unpleasant. 00:25:02.718 --> 00:25:03.510 align:middle line:90% Ed, you're nodding. 00:25:03.510 --> 00:25:04.860 align:middle line:90% I can see you. 00:25:04.860 --> 00:25:08.550 align:middle line:84% They're surprisingly like not pretty models for-- 00:25:08.550 --> 00:25:10.555 align:middle line:90% I think there's nothing-- 00:25:10.555 --> 00:25:12.180 align:middle line:84% I don't have a good reason why that is. 00:25:12.180 --> 00:25:13.950 align:middle line:90% But that's true. 00:25:13.950 --> 00:25:15.390 align:middle line:90% So OK. 00:25:15.390 --> 00:25:17.010 align:middle line:84% And if anyone wants any proof of that, 00:25:17.010 --> 00:25:18.600 align:middle line:90% go back and read Stiglitz 74. 00:25:18.600 --> 00:25:23.900 align:middle line:90% And you'll see it's not pretty. 00:25:23.900 --> 00:25:24.400 align:middle line:90% OK. 00:25:24.400 --> 00:25:26.040 align:middle line:84% I think it's just like carrying aroun-- solving 00:25:26.040 --> 00:25:27.600 align:middle line:84% all these contracts, with all these concave utility 00:25:27.600 --> 00:25:29.267 align:middle line:84% functions, just gets a little difficult. 00:25:29.267 --> 00:25:35.310 align:middle line:90% Anyway, so now what happens? 00:25:35.310 --> 00:25:39.010 align:middle line:84% But the key result, actually, is very easy to see. 00:25:39.010 --> 00:25:45.580 align:middle line:84% So this is exactly the same as before. 00:25:45.580 --> 00:25:49.120 align:middle line:90% Now the farmer has-- 00:25:49.120 --> 00:25:50.710 align:middle line:90% this is concave. 00:25:50.710 --> 00:25:52.960 align:middle line:84% We're maximizing e times u of h and 1 00:25:52.960 --> 00:25:56.815 align:middle line:84% minus e times u of l, as opposed to just h and l directly. 00:25:56.815 --> 00:25:59.510 align:middle line:90% 00:25:59.510 --> 00:26:02.220 align:middle line:84% And so one thing to point out is, 00:26:02.220 --> 00:26:06.580 align:middle line:84% if the landlord implemented the contract from before, 00:26:06.580 --> 00:26:08.050 align:middle line:84% the farmer's utility is going to be 00:26:08.050 --> 00:26:12.010 align:middle line:90% strictly less than w lower bar. 00:26:12.010 --> 00:26:12.580 align:middle line:90% OK? 00:26:12.580 --> 00:26:14.770 align:middle line:90% Because of the four-- 00:26:14.770 --> 00:26:17.560 align:middle line:90% 00:26:17.560 --> 00:26:19.540 align:middle line:84% I mean, this is actually not totally precise. 00:26:19.540 --> 00:26:22.430 align:middle line:84% Oh, sorry, this is precise, the u of w lower bar. 00:26:22.430 --> 00:26:23.530 align:middle line:90% Exactly right. 00:26:23.530 --> 00:26:25.210 align:middle line:90% That's precise. 00:26:25.210 --> 00:26:26.680 align:middle line:90% And why is that? 00:26:26.680 --> 00:26:29.950 align:middle line:90% That's just from concavity. 00:26:29.950 --> 00:26:33.490 align:middle line:90% That's just because the-- 00:26:33.490 --> 00:26:43.540 align:middle line:84% so u of w, lower bar is the utility of the expected return. 00:26:43.540 --> 00:26:45.280 align:middle line:84% And what it means to be risk averse, 00:26:45.280 --> 00:26:47.620 align:middle line:84% or what concave utility functions give you, 00:26:47.620 --> 00:26:50.350 align:middle line:90% is that the utility from-- 00:26:50.350 --> 00:26:52.600 align:middle line:84% the expected utility is less than the utility 00:26:52.600 --> 00:26:55.360 align:middle line:90% of the certainty equivalent. 00:26:55.360 --> 00:26:59.150 align:middle line:84% That's just what we get from concavity. 00:26:59.150 --> 00:27:01.450 align:middle line:90% So that means that-- 00:27:01.450 --> 00:27:04.030 align:middle line:84% so what does that mean practically? 00:27:04.030 --> 00:27:09.070 align:middle line:84% It means, practically, if we took the contract from before, 00:27:09.070 --> 00:27:12.920 align:middle line:84% the worker is now below her IR constraint. 00:27:12.920 --> 00:27:16.420 align:middle line:84% So we've got to do something, otherwise you'll walk away. 00:27:16.420 --> 00:27:17.590 align:middle line:90% And what can we do? 00:27:17.590 --> 00:27:19.480 align:middle line:90% Well, we could lower the rent. 00:27:19.480 --> 00:27:21.723 align:middle line:90% We could raise the fixed fee. 00:27:21.723 --> 00:27:23.140 align:middle line:84% Or we could provide some insurance 00:27:23.140 --> 00:27:29.140 align:middle line:84% and reduce the gap between h minus l. 00:27:29.140 --> 00:27:33.120 align:middle line:84% And the question is, which should the landlord do? 00:27:33.120 --> 00:27:36.540 align:middle line:84% Should they reduce -- the should they do it on the fixed-fee 00:27:36.540 --> 00:27:40.410 align:middle line:84% margin or should they do some of it on the other margin? 00:27:40.410 --> 00:27:44.532 align:middle line:84% And one simple way to see that they should do at least some 00:27:44.532 --> 00:27:45.990 align:middle line:84% of it, at least a little bit of it, 00:27:45.990 --> 00:27:53.250 align:middle line:84% on reducing the h minus l gap, is as follows. 00:27:53.250 --> 00:27:55.980 align:middle line:84% So it's basically just this argument 00:27:55.980 --> 00:28:00.480 align:middle line:84% here, which is starting from the first, best level of effort, 00:28:00.480 --> 00:28:06.600 align:middle line:84% the efficiency loss from reducing h minus l just 00:28:06.600 --> 00:28:09.690 align:middle line:90% a little bit, is second order. 00:28:09.690 --> 00:28:13.230 align:middle line:84% But the utilities-- because we maximized that, 00:28:13.230 --> 00:28:15.120 align:middle line:84% so the loss there is second order. 00:28:15.120 --> 00:28:17.580 align:middle line:84% But the utility gain for the agent 00:28:17.580 --> 00:28:21.040 align:middle line:90% is going to be first order. 00:28:21.040 --> 00:28:23.560 align:middle line:84% So that's a simple argument, which 00:28:23.560 --> 00:28:27.400 align:middle line:84% says, the solution to this problem 00:28:27.400 --> 00:28:29.860 align:middle line:84% can't have h minus l be equal to the full-- 00:28:29.860 --> 00:28:32.380 align:middle line:90% 00:28:32.380 --> 00:28:35.470 align:middle line:84% equal to cap H. We're going to do at least a little bit 00:28:35.470 --> 00:28:39.173 align:middle line:84% of this reduction on the dimension of providing 00:28:39.173 --> 00:28:40.840 align:middle line:84% a little bit of insurance because that's 00:28:40.840 --> 00:28:45.310 align:middle line:84% second order, in terms of the total kind of productivity 00:28:45.310 --> 00:28:48.280 align:middle line:84% loss, but first order gain in terms of the insurance value. 00:28:48.280 --> 00:28:49.330 align:middle line:90% Yeah, Erin? 00:28:49.330 --> 00:28:52.330 align:middle line:84% AUDIENCE: Is that true when the landlord is also [INAUDIBLE]?? 00:28:52.330 --> 00:28:54.760 align:middle line:84% BEN OLKEN: Yes, because another way of thinking about it 00:28:54.760 --> 00:28:56.530 align:middle line:90% is the landlord is bearing-- 00:28:56.530 --> 00:29:00.678 align:middle line:90% 00:29:00.678 --> 00:29:01.720 align:middle line:90% let me think for one sec. 00:29:01.720 --> 00:29:02.640 align:middle line:90% Yes, it's true. 00:29:02.640 --> 00:29:04.177 align:middle line:90% But let me just double-check. 00:29:04.177 --> 00:29:12.823 align:middle line:90% 00:29:12.823 --> 00:29:15.240 align:middle line:84% Yeah, I'm pretty sure it's true because it's going to be-- 00:29:15.240 --> 00:29:15.840 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:29:15.840 --> 00:29:16.500 align:middle line:84% BEN OLKEN: Because I think it's going 00:29:16.500 --> 00:29:18.120 align:middle line:84% to be a similar kind of argument that the landlord-- 00:29:18.120 --> 00:29:20.070 align:middle line:84% I was trying to think of whether you need any other conditions. 00:29:20.070 --> 00:29:22.278 align:middle line:84% I don't think so because I think that the landlord is 00:29:22.278 --> 00:29:24.800 align:middle line:90% going to be totally, fully-- 00:29:24.800 --> 00:29:25.800 align:middle line:90% it's a similar argument. 00:29:25.800 --> 00:29:30.690 align:middle line:84% Going from fully insured to an epsilon bit of risk 00:29:30.690 --> 00:29:33.712 align:middle line:84% is going to be less costly than reducing the risk a little bit. 00:29:33.712 --> 00:29:34.920 align:middle line:90% I'm pretty sure that's right. 00:29:34.920 --> 00:29:39.030 align:middle line:90% 00:29:39.030 --> 00:29:41.530 align:middle line:84% So on the other hand-- so that says, 00:29:41.530 --> 00:29:45.840 align:middle line:84% look, the optimal contract is not going to be the first best. 00:29:45.840 --> 00:29:49.080 align:middle line:84% It's going to have at least a little bit of a sharing 00:29:49.080 --> 00:29:51.740 align:middle line:90% feature. 00:29:51.740 --> 00:29:55.430 align:middle line:84% On the other hand, we also know that we can't fully 00:29:55.430 --> 00:29:57.980 align:middle line:84% insure the agent because then the agent won't work. 00:29:57.980 --> 00:30:00.500 align:middle line:90% 00:30:00.500 --> 00:30:06.190 align:middle line:84% So this kind of argument says that the optimal contract is 00:30:06.190 --> 00:30:10.967 align:middle line:84% going to look something like a share contract, where there's-- 00:30:10.967 --> 00:30:12.550 align:middle line:84% maybe with a fixed fee, but it's going 00:30:12.550 --> 00:30:15.760 align:middle line:84% to have some feature where the share 00:30:15.760 --> 00:30:17.620 align:middle line:84% of the output that the agent keeps 00:30:17.620 --> 00:30:20.030 align:middle line:90% is less than the full output. 00:30:20.030 --> 00:30:25.150 align:middle line:84% So some amount of the output is shared with the landlord, 00:30:25.150 --> 00:30:30.760 align:middle line:84% even though that has a productivity loss associated 00:30:30.760 --> 00:30:32.712 align:middle line:90% with doing that. 00:30:32.712 --> 00:30:35.170 align:middle line:84% But it's not going to go all the way to the other dimension 00:30:35.170 --> 00:30:37.570 align:middle line:90% of full insurance either. 00:30:37.570 --> 00:30:40.660 align:middle line:90% So that's, in some sense-- 00:30:40.660 --> 00:30:43.480 align:middle line:90% this is the key insight. 00:30:43.480 --> 00:30:45.760 align:middle line:90% 00:30:45.760 --> 00:30:47.260 align:middle line:84% And this, as I said, is the argument 00:30:47.260 --> 00:30:48.820 align:middle line:84% for sharecropping given by Stiglitz, 00:30:48.820 --> 00:30:51.320 align:middle line:84% which is that landlords and peasants actually prefer-- 00:30:51.320 --> 00:30:55.960 align:middle line:84% so Stiglitz's argument is actually like-- 00:30:55.960 --> 00:30:59.140 align:middle line:84% it's almost revolutionary, in a sense. 00:30:59.140 --> 00:31:01.550 align:middle line:84% It's to say, look-- not revolutionary 00:31:01.550 --> 00:31:03.550 align:middle line:84% in the sense of like, let's have the revolution, 00:31:03.550 --> 00:31:07.160 align:middle line:84% but revolutionary, almost in the opposite sense, 00:31:07.160 --> 00:31:09.340 align:middle line:84% actually, to say, look, you might 00:31:09.340 --> 00:31:13.000 align:middle line:84% have thought that this sharecropping thing was 00:31:13.000 --> 00:31:15.145 align:middle line:90% this total exploitative thing. 00:31:15.145 --> 00:31:17.680 align:middle line:84% But if only the tenants could get out from under it, 00:31:17.680 --> 00:31:20.223 align:middle line:90% they would be much happier. 00:31:20.223 --> 00:31:21.640 align:middle line:84% But actually, he says, no, no, no. 00:31:21.640 --> 00:31:26.320 align:middle line:84% Actually, the sharecropping is preferred by the tenants, 00:31:26.320 --> 00:31:29.710 align:middle line:84% in some sense, to the full rental contract because they're 00:31:29.710 --> 00:31:31.960 align:middle line:90% getting some insurance value. 00:31:31.960 --> 00:31:34.900 align:middle line:84% And otherwise-- because life is really risky. 00:31:34.900 --> 00:31:38.050 align:middle line:84% And so it's a way of providing-- some combination of providing 00:31:38.050 --> 00:31:42.175 align:middle line:90% insurance and incentives. 00:31:42.175 --> 00:31:43.070 align:middle line:90% AUDIENCE: Ben. 00:31:43.070 --> 00:31:43.980 align:middle line:90% BEN OLKEN: Yes? 00:31:43.980 --> 00:31:45.891 align:middle line:84% AUDIENCE: Half the rents in this model, 00:31:45.891 --> 00:31:47.274 align:middle line:84% because you don't want to include 00:31:47.274 --> 00:31:49.774 align:middle line:84% landlords to extract all the surplus or something like that. 00:31:49.774 --> 00:31:52.300 align:middle line:84% Does that change efficiency at all? 00:31:52.300 --> 00:31:54.630 align:middle line:84% BEN OLKEN: It doesn't change efficiency. 00:31:54.630 --> 00:31:56.330 align:middle line:90% No, sorry. 00:31:56.330 --> 00:31:58.380 align:middle line:84% So what happens in this-- let me think for a sec. 00:31:58.380 --> 00:32:00.950 align:middle line:84% So you're saying, what happens if you put in a restriction 00:32:00.950 --> 00:32:03.745 align:middle line:84% saying they have to give more rents? 00:32:03.745 --> 00:32:04.370 align:middle line:90% AUDIENCE: Yeah. 00:32:04.370 --> 00:32:08.100 align:middle line:90% 00:32:08.100 --> 00:32:11.430 align:middle line:84% BEN OLKEN: Let me think for a sec. 00:32:11.430 --> 00:32:15.877 align:middle line:84% So if I increase the rents, will I-- 00:32:15.877 --> 00:32:17.835 align:middle line:84% I don't know the answer off the top of my head. 00:32:17.835 --> 00:32:18.377 align:middle line:90% AUDIENCE: OK. 00:32:18.377 --> 00:32:23.280 align:middle line:84% BEN OLKEN: So if you're saying, if I increase the rents, 00:32:23.280 --> 00:32:27.335 align:middle line:84% will the landlord compensate somehow? 00:32:27.335 --> 00:32:27.960 align:middle line:90% AUDIENCE: Yeah. 00:32:27.960 --> 00:32:29.022 align:middle line:84% I'm just trying to understand what-- 00:32:29.022 --> 00:32:30.105 align:middle line:90% I mean, I understand why-- 00:32:30.105 --> 00:32:32.720 align:middle line:84% BEN OLKEN: So I think, actually, in this model-- 00:32:32.720 --> 00:32:35.768 align:middle line:90% no, bum, bum, bum, bum, bum. 00:32:35.768 --> 00:32:52.970 align:middle line:90% 00:32:52.970 --> 00:32:53.970 align:middle line:90% So I'm not totally sure. 00:32:53.970 --> 00:32:56.010 align:middle line:84% And the reason I'm not totally sure is that-- 00:32:56.010 --> 00:32:57.718 align:middle line:84% in general, the way we solve these models 00:32:57.718 --> 00:32:59.668 align:middle line:84% is, we figure out what the gap is, 00:32:59.668 --> 00:33:01.710 align:middle line:84% and then slide up and down to hit the rent floor. 00:33:01.710 --> 00:33:03.000 align:middle line:84% And that's just essentially, what 00:33:03.000 --> 00:33:04.890 align:middle line:84% is the impact of raising the outside wage, essentially. 00:33:04.890 --> 00:33:06.600 align:middle line:84% And the reason I'm not 100% sure is it also 00:33:06.600 --> 00:33:07.683 align:middle line:90% affects the risk aversion. 00:33:07.683 --> 00:33:09.640 align:middle line:90% And that's what's confusing. 00:33:09.640 --> 00:33:13.560 align:middle line:84% That's why I'm not giving you a quick answer, because-- 00:33:13.560 --> 00:33:16.192 align:middle line:84% in a concave model, if I move the overall payment up, 00:33:16.192 --> 00:33:17.400 align:middle line:90% it makes me less risk averse. 00:33:17.400 --> 00:33:19.080 align:middle line:84% And so I'm not 100% sure, actually. 00:33:19.080 --> 00:33:22.320 align:middle line:84% Does anyone else know the answer to that question? 00:33:22.320 --> 00:33:25.108 align:middle line:84% That'd be a good problem-set question. 00:33:25.108 --> 00:33:25.650 align:middle line:90% I don't know. 00:33:25.650 --> 00:33:26.852 align:middle line:90% Sorry. 00:33:26.852 --> 00:33:27.810 align:middle line:90% That's a good question. 00:33:27.810 --> 00:33:29.190 align:middle line:90% Yeah, Kyle? 00:33:29.190 --> 00:33:32.760 align:middle line:84% AUDIENCE: It's just meant to-- you can't write this directly. 00:33:32.760 --> 00:33:34.774 align:middle line:84% It's like [INAUDIBLE] tax, right? 00:33:34.774 --> 00:33:35.355 align:middle line:90% [INAUDIBLE]? 00:33:35.355 --> 00:33:36.230 align:middle line:90% BEN OLKEN: I'm sorry? 00:33:36.230 --> 00:33:38.605 align:middle line:84% AUDIENCE: Like you said, you couldn't write this contract 00:33:38.605 --> 00:33:42.033 align:middle line:84% just like the alpha model that we saw [INAUDIBLE].. 00:33:42.033 --> 00:33:42.950 align:middle line:90% Or is it [? exact-- ?] 00:33:42.950 --> 00:33:43.520 align:middle line:90% BEN OLKEN: Yes, you could. 00:33:43.520 --> 00:33:44.090 align:middle line:90% Sure. 00:33:44.090 --> 00:33:44.760 align:middle line:90% AUDIENCE: OK. 00:33:44.760 --> 00:33:50.420 align:middle line:84% BEN OLKEN: Well, so that model didn't have any uncertainty. 00:33:50.420 --> 00:33:55.740 align:middle line:84% So what is different in this model versus that model? 00:33:55.740 --> 00:33:58.497 align:middle line:84% So that model that had no uncertainty, is the key thing. 00:33:58.497 --> 00:34:00.080 align:middle line:84% And the other thing which is important 00:34:00.080 --> 00:34:02.600 align:middle line:84% is that the effort choice, here, affects 00:34:02.600 --> 00:34:03.890 align:middle line:90% the probability of the-- 00:34:03.890 --> 00:34:05.820 align:middle line:90% is related to the uncertainty. 00:34:05.820 --> 00:34:09.080 align:middle line:84% So it is a bit of a different model, in that sense. 00:34:09.080 --> 00:34:14.659 align:middle line:84% But conceptually, yes, you can think of this as a bit of an-- 00:34:14.659 --> 00:34:16.595 align:middle line:84% the solution looks like an output [? tax. ?] 00:34:16.595 --> 00:34:18.100 align:middle line:84% AUDIENCE: Well, it's just the percentage of output. 00:34:18.100 --> 00:34:18.650 align:middle line:90% BEN OLKEN: I'm sorry? 00:34:18.650 --> 00:34:19.639 align:middle line:84% AUDIENCE: It's just the percentage 00:34:19.639 --> 00:34:20.510 align:middle line:90% of output that they-- 00:34:20.510 --> 00:34:22.052 align:middle line:84% BEN OLKEN: Well, this is-- so, sorry. 00:34:22.052 --> 00:34:26.600 align:middle line:90% So in the continuous case-- 00:34:26.600 --> 00:34:28.580 align:middle line:90% so you can write-- 00:34:28.580 --> 00:34:34.520 align:middle line:84% there obviously are continuous moral-hazard models, 00:34:34.520 --> 00:34:37.820 align:middle line:84% which are even more difficult to work with. 00:34:37.820 --> 00:34:38.913 align:middle line:90% But you can solve them. 00:34:38.913 --> 00:34:40.580 align:middle line:84% And if you take a contract theory class, 00:34:40.580 --> 00:34:42.210 align:middle line:90% you will solve them. 00:34:42.210 --> 00:34:46.302 align:middle line:84% And in those models, yes, I think it will look like some-- 00:34:46.302 --> 00:34:48.469 align:middle line:84% you'll get some payment as a function of the output. 00:34:48.469 --> 00:34:51.889 align:middle line:84% And it'll look like an output tax. 00:34:51.889 --> 00:34:54.407 align:middle line:84% Whether it will be a linear output tax, 00:34:54.407 --> 00:34:56.449 align:middle line:84% I don't know necessarily that that's always true. 00:34:56.449 --> 00:35:04.170 align:middle line:90% 00:35:04.170 --> 00:35:11.490 align:middle line:84% Although Bengt Holmstrom has a well-known paper in contract 00:35:11.490 --> 00:35:17.490 align:middle line:84% theory that shows under what circumstances linear incentives 00:35:17.490 --> 00:35:18.550 align:middle line:90% are optimal. 00:35:18.550 --> 00:35:21.135 align:middle line:84% So under-- I think the-- and I'm not 00:35:21.135 --> 00:35:23.760 align:middle line:84% remembering the details of these models off the top of my head. 00:35:23.760 --> 00:35:26.730 align:middle line:84% But my understanding is the general class 00:35:26.730 --> 00:35:30.420 align:middle line:84% of-- in the general answer, it will not necessarily 00:35:30.420 --> 00:35:32.830 align:middle line:90% be a linear-incentive function. 00:35:32.830 --> 00:35:35.340 align:middle line:84% However, what Bengt shows is that 00:35:35.340 --> 00:35:38.370 align:middle line:84% under some things under some various perturbations, where 00:35:38.370 --> 00:35:40.290 align:middle line:84% you can do some gaming or things like that, 00:35:40.290 --> 00:35:41.970 align:middle line:84% linear contracts often do emerge. 00:35:41.970 --> 00:35:44.740 align:middle line:90% That's my summary of that. 00:35:44.740 --> 00:35:45.670 align:middle line:90% And you're nodding. 00:35:45.670 --> 00:35:49.410 align:middle line:84% Is-- anyone else want to comment on that? 00:35:49.410 --> 00:35:51.230 align:middle line:90% All right. 00:35:51.230 --> 00:35:52.940 align:middle line:90% Other comments? 00:35:52.940 --> 00:35:55.620 align:middle line:84% So I think-- I don't if we still have 00:35:55.620 --> 00:35:56.870 align:middle line:90% an example in the problem set. 00:35:56.870 --> 00:35:58.700 align:middle line:90% We should make sure-- we don't? 00:35:58.700 --> 00:35:59.390 align:middle line:90% That's gone? 00:35:59.390 --> 00:36:01.348 align:middle line:84% AUDIENCE: There's no problem set on this topic. 00:36:01.348 --> 00:36:04.370 align:middle line:84% BEN OLKEN: OK, so if you would like 00:36:04.370 --> 00:36:06.440 align:middle line:84% to work through an example of this, I do 00:36:06.440 --> 00:36:09.260 align:middle line:84% think we have a previous problem set, that we have put together 00:36:09.260 --> 00:36:12.200 align:middle line:84% on this problem, that has an example of how 00:36:12.200 --> 00:36:15.710 align:middle line:84% to work through these moral-hazard models. 00:36:15.710 --> 00:36:16.430 align:middle line:90% Yes, Davey? 00:36:16.430 --> 00:36:19.700 align:middle line:84% AUDIENCE: Just with up here, the agents have sufficient 00:36:19.700 --> 00:36:21.075 align:middle line:84% constraints [? t minus, right? ?] 00:36:21.075 --> 00:36:21.742 align:middle line:90% BEN OLKEN: Yeah. 00:36:21.742 --> 00:36:23.840 align:middle line:84% AUDIENCE: It's no [? service ?] [? to ?] [? h. ?] 00:36:23.840 --> 00:36:24.710 align:middle line:90% BEN OLKEN: Right. 00:36:24.710 --> 00:36:25.550 align:middle line:90% Yep. 00:36:25.550 --> 00:36:27.200 align:middle line:84% They're still going to slide at that-- 00:36:27.200 --> 00:36:29.060 align:middle line:84% it's still going to be the case of the IR contraint 00:36:29.060 --> 00:36:29.977 align:middle line:90% is still going to buy. 00:36:29.977 --> 00:36:31.970 align:middle line:84% There's going to be no surplus here. 00:36:31.970 --> 00:36:32.470 align:middle line:90% Yeah? 00:36:32.470 --> 00:36:35.000 align:middle line:90% 00:36:35.000 --> 00:36:36.156 align:middle line:90% Comment? 00:36:36.156 --> 00:36:43.140 align:middle line:84% AUDIENCE: You're seeing-- you've seen these kind of contracts 00:36:43.140 --> 00:36:43.910 align:middle line:90% in-- 00:36:43.910 --> 00:36:46.590 align:middle line:90% for instance, the tax got cut. 00:36:46.590 --> 00:36:49.040 align:middle line:84% You pay a fixed rent and then you 00:36:49.040 --> 00:36:52.290 align:middle line:84% get everything-- say the first best is-- in fact, in the taxi 00:36:52.290 --> 00:36:54.110 align:middle line:90% business, we see the first pass. 00:36:54.110 --> 00:36:56.260 align:middle line:90% BEN OLKEN: We see both. 00:36:56.260 --> 00:36:56.930 align:middle line:90% But yes. 00:36:56.930 --> 00:36:59.462 align:middle line:84% With Uber, you see a sharecropping contract. 00:36:59.462 --> 00:37:00.920 align:middle line:84% They don't call it "sharecropping," 00:37:00.920 --> 00:37:06.320 align:middle line:84% but you see a share contract, where it looks-- 00:37:06.320 --> 00:37:07.970 align:middle line:90% where Uber takes 30% of-- 00:37:07.970 --> 00:37:10.520 align:middle line:84% or whatever the percentage is of the thing. 00:37:10.520 --> 00:37:11.540 align:middle line:90% And they get a fixed-- 00:37:11.540 --> 00:37:12.170 align:middle line:90% some share. 00:37:12.170 --> 00:37:14.920 align:middle line:90% 00:37:14.920 --> 00:37:17.470 align:middle line:84% AUDIENCE: Is there a reason why, say, like supr-- 00:37:17.470 --> 00:37:17.880 align:middle line:90% BEN OLKEN: But-- sorry. 00:37:17.880 --> 00:37:19.005 align:middle line:90% But sorry, one other thing. 00:37:19.005 --> 00:37:20.800 align:middle line:84% But in the Uber contract, it's observable. 00:37:20.800 --> 00:37:22.270 align:middle line:90% And I think it may depend-- 00:37:22.270 --> 00:37:25.680 align:middle line:84% I think-- I don't know the full economics of the taxicab 00:37:25.680 --> 00:37:26.180 align:middle line:90% literature. 00:37:26.180 --> 00:37:28.930 align:middle line:84% But I think it probably depends on to what degree-- 00:37:28.930 --> 00:37:31.870 align:middle line:84% certainly, to what degree effort is observable 00:37:31.870 --> 00:37:34.870 align:middle line:84% would affect what these contracts look like. 00:37:34.870 --> 00:37:37.750 align:middle line:84% And Davey is nodding since Davey is actively 00:37:37.750 --> 00:37:39.760 align:middle line:84% looking at exactly these issues of observability 00:37:39.760 --> 00:37:47.050 align:middle line:84% and the endogenous contract choice in Cote d'Ivoire 00:37:47.050 --> 00:37:47.762 align:middle line:90% at the moment. 00:37:47.762 --> 00:37:48.970 align:middle line:90% And we just lost my computer. 00:37:48.970 --> 00:37:52.660 align:middle line:84% Sorry, go ahead while I get my computer back. 00:37:52.660 --> 00:37:53.160 align:middle line:90% Yeah? 00:37:53.160 --> 00:37:54.203 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:37:54.203 --> 00:37:54.870 align:middle line:90% BEN OLKEN: What? 00:37:54.870 --> 00:37:55.960 align:middle line:90% Sorry. 00:37:55.960 --> 00:37:59.560 align:middle line:84% AUDIENCE: I was going to ask, in what kind of industries will 00:37:59.560 --> 00:38:02.020 align:middle line:84% expect to see, say, [INAUDIBLE]---- 00:38:02.020 --> 00:38:04.280 align:middle line:84% by the first of this kind of contract where you-- 00:38:04.280 --> 00:38:06.155 align:middle line:84% BEN OLKEN: So you would see-- so for example, 00:38:06.155 --> 00:38:08.780 align:middle line:84% if you had no observability of output, 00:38:08.780 --> 00:38:12.820 align:middle line:84% then you would see rental contract. 00:38:12.820 --> 00:38:16.000 align:middle line:84% So this contract requires the principal 00:38:16.000 --> 00:38:19.210 align:middle line:84% to be able to condition a payment on the output, 00:38:19.210 --> 00:38:21.950 align:middle line:90% for example. 00:38:21.950 --> 00:38:24.380 align:middle line:84% Or you also might think that taxi-cab-type drivers may not 00:38:24.380 --> 00:38:25.310 align:middle line:90% be that risk averse. 00:38:25.310 --> 00:38:28.790 align:middle line:84% So if the uncertainty is day by day-- 00:38:28.790 --> 00:38:30.590 align:middle line:90% this is over a whole season. 00:38:30.590 --> 00:38:32.348 align:middle line:90% It's one giant realization. 00:38:32.348 --> 00:38:33.890 align:middle line:84% You're really risk averse about that. 00:38:33.890 --> 00:38:34.840 align:middle line:90% Some days it rains. 00:38:34.840 --> 00:38:36.290 align:middle line:90% Sometimes it doesn't rain. 00:38:36.290 --> 00:38:38.040 align:middle line:84% I may not be right risk averse about that. 00:38:38.040 --> 00:38:39.620 align:middle line:84% I might be able to smooth over days. 00:38:39.620 --> 00:38:41.578 align:middle line:84% So that's another reason it might be different. 00:38:41.578 --> 00:38:44.060 align:middle line:90% 00:38:44.060 --> 00:38:45.890 align:middle line:90% Other questions, comments? 00:38:45.890 --> 00:38:46.880 align:middle line:90% Yeah, Shera. 00:38:46.880 --> 00:38:48.560 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:38:48.560 --> 00:38:49.560 align:middle line:90% BEN OLKEN: Sorry, Shera. 00:38:49.560 --> 00:38:50.935 align:middle line:84% You're going to have to speak up. 00:38:50.935 --> 00:38:52.300 align:middle line:90% I can't really-- 00:38:52.300 --> 00:38:54.236 align:middle line:84% AUDIENCE: When the farmers are risk averse, 00:38:54.236 --> 00:38:59.368 align:middle line:84% it's first best [INAUDIBLE] even if the effort was perfect, 00:38:59.368 --> 00:38:59.910 align:middle line:90% [INAUDIBLE]-- 00:38:59.910 --> 00:39:00.868 align:middle line:90% BEN OLKEN: Yeah, sorry. 00:39:00.868 --> 00:39:03.340 align:middle line:84% So when I say first best, I meant first in terms of output. 00:39:03.340 --> 00:39:08.710 align:middle line:84% But yes, the socially-optimal contract includes sharing risk. 00:39:08.710 --> 00:39:12.475 align:middle line:84% But it does not get the first best-level of output. 00:39:12.475 --> 00:39:15.770 align:middle line:84% AUDIENCE: So here, [INAUDIBLE] doesn't have 00:39:15.770 --> 00:39:17.817 align:middle line:90% the first-best [INAUDIBLE]. 00:39:17.817 --> 00:39:18.900 align:middle line:90% BEN OLKEN: Doesn't-- what? 00:39:18.900 --> 00:39:19.480 align:middle line:90% Sorry? 00:39:19.480 --> 00:39:21.570 align:middle line:84% AUDIENCE: [INAUDIBLE] hasn't [? combined-- ?] 00:39:21.570 --> 00:39:23.550 align:middle line:84% does the fact that can't observe effort, 00:39:23.550 --> 00:39:25.223 align:middle line:84% does that change the [INAUDIBLE]?? 00:39:25.223 --> 00:39:27.140 align:middle line:84% Because it seems like it does, like first-best 00:39:27.140 --> 00:39:30.320 align:middle line:84% are still coinciding with [? same best. ?] 00:39:30.320 --> 00:39:31.620 align:middle line:90% BEN OLKEN: No, no, no, so-- 00:39:31.620 --> 00:39:32.120 align:middle line:90% no, no, no. 00:39:32.120 --> 00:39:35.065 align:middle line:84% So if we could observe effort, what 00:39:35.065 --> 00:39:36.440 align:middle line:84% would be the first-best contract? 00:39:36.440 --> 00:39:38.600 align:middle line:84% First-best contract would be we write a contract 00:39:38.600 --> 00:39:39.410 align:middle line:90% based on effort. 00:39:39.410 --> 00:39:41.210 align:middle line:84% And we would observe your-- but we'd 00:39:41.210 --> 00:39:42.500 align:middle line:90% provide you full insurance. 00:39:42.500 --> 00:39:46.325 align:middle line:90% 00:39:46.325 --> 00:39:47.950 align:middle line:84% So the first-best contract would have-- 00:39:47.950 --> 00:39:50.830 align:middle line:84% if we could implement-- if we could contract an effort, 00:39:50.830 --> 00:39:55.300 align:middle line:84% we would say put in e equals big H over c. 00:39:55.300 --> 00:39:58.330 align:middle line:84% And I'll pay you h equals l equals the expected-- 00:39:58.330 --> 00:40:01.502 align:middle line:90% equals w lower bar, basically. 00:40:01.502 --> 00:40:03.335 align:middle line:84% Yeah, that would be the first-best contract. 00:40:03.335 --> 00:40:07.110 align:middle line:90% 00:40:07.110 --> 00:40:10.080 align:middle line:90% Other questions? 00:40:10.080 --> 00:40:12.740 align:middle line:90% OK. 00:40:12.740 --> 00:40:16.940 align:middle line:84% So that's the risk-aversion version. 00:40:16.940 --> 00:40:20.990 align:middle line:84% So what about the limited-liability version? 00:40:20.990 --> 00:40:31.790 align:middle line:84% So let's go back, now, to risk neutrality for a second. 00:40:31.790 --> 00:40:34.400 align:middle line:84% But let's impose limited liability. 00:40:34.400 --> 00:40:38.620 align:middle line:84% So let's assume that we can't impose l than 0. 00:40:38.620 --> 00:40:41.797 align:middle line:90% 00:40:41.797 --> 00:40:43.630 align:middle line:84% So now what happens to the optimal contract? 00:40:43.630 --> 00:40:47.800 align:middle line:84% Well, now the landlord has lost a degree of freedom. 00:40:47.800 --> 00:40:49.567 align:middle line:84% The landlord, before, could choose 00:40:49.567 --> 00:40:50.650 align:middle line:90% the little h and little l. 00:40:50.650 --> 00:40:52.450 align:middle line:84% Now they can now they have this problem 00:40:52.450 --> 00:40:55.900 align:middle line:90% that little l can't go below 0. 00:40:55.900 --> 00:40:58.900 align:middle line:90% 00:40:58.900 --> 00:41:02.050 align:middle line:84% So in the first-best kind of contract, we had this. 00:41:02.050 --> 00:41:04.690 align:middle line:84% What happens-- suppose we're in the world, where 00:41:04.690 --> 00:41:07.288 align:middle line:84% this thing is-- if this thing is positive, then great. 00:41:07.288 --> 00:41:08.080 align:middle line:90% We have no problem. 00:41:08.080 --> 00:41:10.960 align:middle line:90% 00:41:10.960 --> 00:41:14.110 align:middle line:84% But if this thing is negative, then we 00:41:14.110 --> 00:41:16.330 align:middle line:84% can't write that contract anymore 00:41:16.330 --> 00:41:19.180 align:middle line:84% because in the low state, they just won't pay. 00:41:19.180 --> 00:41:22.690 align:middle line:84% Another way of saying is, they won't pay the rent. 00:41:22.690 --> 00:41:24.020 align:middle line:90% So what do we do? 00:41:24.020 --> 00:41:26.440 align:middle line:90% What does the landlord do? 00:41:26.440 --> 00:41:32.290 align:middle line:84% So instead, little l has to be equal to 0 00:41:32.290 --> 00:41:37.240 align:middle line:84% because it can't be-- it's bounded by a limited liability. 00:41:37.240 --> 00:41:40.090 align:middle line:90% So what happens to h? 00:41:40.090 --> 00:41:44.060 align:middle line:84% So now, where we know the optimal choice of effort. 00:41:44.060 --> 00:41:45.680 align:middle line:84% So the landlord is going to say, well, 00:41:45.680 --> 00:41:48.580 align:middle line:84% I'm just going to choose-- all I'm ever going to get, 00:41:48.580 --> 00:41:50.770 align:middle line:84% if the landlord-- is a payment in the high state. 00:41:50.770 --> 00:41:54.580 align:middle line:84% I'm never going to get any money in the low state. 00:41:54.580 --> 00:41:57.677 align:middle line:84% So my-- but I know, still, the agent 00:41:57.677 --> 00:41:59.260 align:middle line:84% is only going to be incentivized based 00:41:59.260 --> 00:42:02.740 align:middle line:84% on the difference between the little h and little l. 00:42:02.740 --> 00:42:06.400 align:middle line:84% So now I'm stuck solving this contract, which 00:42:06.400 --> 00:42:09.080 align:middle line:84% is, I'm going to maximize my return, which 00:42:09.080 --> 00:42:11.080 align:middle line:84% is the probability of the high state, times what 00:42:11.080 --> 00:42:12.288 align:middle line:90% I get paid in the high state. 00:42:12.288 --> 00:42:14.380 align:middle line:84% And the low state is just a wash. 00:42:14.380 --> 00:42:16.330 align:middle line:84% I don't get anything in a low state. 00:42:16.330 --> 00:42:20.830 align:middle line:84% So that's this problem over here. 00:42:20.830 --> 00:42:23.200 align:middle line:90% This is h over c. 00:42:23.200 --> 00:42:27.490 align:middle line:84% We know that the tenant is going to choose h minus l over c. 00:42:27.490 --> 00:42:28.690 align:middle line:90% But their l is equal to 0. 00:42:28.690 --> 00:42:31.810 align:middle line:84% So the tenant's going to choose a little h over c times 00:42:31.810 --> 00:42:33.190 align:middle line:90% big H over c. 00:42:33.190 --> 00:42:39.400 align:middle line:84% And so the optimal choice, here, for the effort, 00:42:39.400 --> 00:42:45.790 align:middle line:84% for the landlord, is to choose little h equal to big H over 2. 00:42:45.790 --> 00:42:49.172 align:middle line:84% And therefore, the optimal amount of effort 00:42:49.172 --> 00:42:50.380 align:middle line:90% is going to be big H over 2c. 00:42:50.380 --> 00:42:52.900 align:middle line:90% 00:42:52.900 --> 00:42:53.950 align:middle line:90% Everyone see that? 00:42:53.950 --> 00:42:55.030 align:middle line:90% All right. 00:42:55.030 --> 00:43:04.005 align:middle line:84% So again, we have less than the first-best amount of effort. 00:43:04.005 --> 00:43:05.880 align:middle line:84% First-best amount of effort was big H over c. 00:43:05.880 --> 00:43:07.530 align:middle line:90% Now we only have big H over 2c. 00:43:07.530 --> 00:43:09.885 align:middle line:84% So we have half of the optimal amount of effort. 00:43:09.885 --> 00:43:13.890 align:middle line:90% 00:43:13.890 --> 00:43:16.870 align:middle line:84% So we're also in a world, here, where-- 00:43:16.870 --> 00:43:17.370 align:middle line:90% Sorry. 00:43:17.370 --> 00:43:19.560 align:middle line:84% And-- so that's the point number one. 00:43:19.560 --> 00:43:22.410 align:middle line:84% And point number two, this is sharecropping contracts 00:43:22.410 --> 00:43:27.153 align:middle line:84% because in the high state, you keep half, 00:43:27.153 --> 00:43:29.070 align:middle line:84% I get-- the lower gets half, tenant gets half. 00:43:29.070 --> 00:43:31.230 align:middle line:84% In the low state, nobody gets anything. 00:43:31.230 --> 00:43:32.700 align:middle line:84% So this looks like a perfect share 00:43:32.700 --> 00:43:34.080 align:middle line:84% cont-- in this simple example, it's 00:43:34.080 --> 00:43:35.955 align:middle line:84% a share contract with the share equal to 1/2. 00:43:35.955 --> 00:43:44.270 align:middle line:90% 00:43:44.270 --> 00:43:46.950 align:middle line:84% So now, what happens if we have an IR constraint here? 00:43:46.950 --> 00:43:50.182 align:middle line:84% Well, so what's the farmer's utility under this contract? 00:43:50.182 --> 00:43:52.640 align:middle line:84% Well, we can solve the farmer's utility under the contract. 00:43:52.640 --> 00:43:56.840 align:middle line:84% Farmer's utility the contract is their choice of eff-- 00:43:56.840 --> 00:43:57.590 align:middle line:90% we can just solve. 00:43:57.590 --> 00:43:59.923 align:middle line:84% Their choice of effort is going to be a little h over c. 00:43:59.923 --> 00:44:02.630 align:middle line:90% 00:44:02.630 --> 00:44:06.530 align:middle line:84% And they're going to get to keep payment h, in the state. 00:44:06.530 --> 00:44:08.130 align:middle line:90% Plus they have a cost of effort. 00:44:08.130 --> 00:44:11.480 align:middle line:84% So if you solve this out, it's 1/8 big H square 00:44:11.480 --> 00:44:12.560 align:middle line:90% over c, whatever. 00:44:12.560 --> 00:44:13.970 align:middle line:90% It's all thought out. 00:44:13.970 --> 00:44:16.560 align:middle line:84% The key point, though, is the following, 00:44:16.560 --> 00:44:20.520 align:middle line:84% which is that's the landlord's utility. 00:44:20.520 --> 00:44:24.360 align:middle line:84% I'm sorry, that's the tenant's utility under this. 00:44:24.360 --> 00:44:28.140 align:middle line:84% If the tenant is getting at least w lower 00:44:28.140 --> 00:44:33.630 align:middle line:84% bar, then under-- you see with utility, great, we're done. 00:44:33.630 --> 00:44:35.910 align:middle line:90% Fine. 00:44:35.910 --> 00:44:42.920 align:middle line:84% Now in this case, the tenant also is getting rents. 00:44:42.920 --> 00:44:46.700 align:middle line:84% If this thing is exactly equal to the w lower bar, fine. 00:44:46.700 --> 00:44:49.190 align:middle line:84% But that's like a knife-edge case. 00:44:49.190 --> 00:44:52.790 align:middle line:84% Generically, you may have parameter values 00:44:52.790 --> 00:44:57.320 align:middle line:84% where you have to pay the tenant above her outside option 00:44:57.320 --> 00:45:00.347 align:middle line:84% in order to get her to work the optimal amount. 00:45:00.347 --> 00:45:01.430 align:middle line:90% And what's different here? 00:45:01.430 --> 00:45:03.410 align:middle line:84% The difference is in the other thing, 00:45:03.410 --> 00:45:06.045 align:middle line:84% you could just slide the whole contract up and down, 00:45:06.045 --> 00:45:07.670 align:middle line:84% arbitrarily, so that the expected value 00:45:07.670 --> 00:45:08.780 align:middle line:90% was equal to w lower bar. 00:45:08.780 --> 00:45:10.730 align:middle line:84% But now you can't because you have 00:45:10.730 --> 00:45:14.420 align:middle line:84% this limited-liability constraint from below. 00:45:14.420 --> 00:45:18.110 align:middle line:90% So she may get rents here. 00:45:18.110 --> 00:45:20.870 align:middle line:84% And also, the contract doesn't depend on w lower bar anymore, 00:45:20.870 --> 00:45:24.620 align:middle line:90% in this piece of it. 00:45:24.620 --> 00:45:27.730 align:middle line:84% As long as this thing is well above the lower bar, 00:45:27.730 --> 00:45:29.390 align:middle line:84% we move the lower the bar up and down, 00:45:29.390 --> 00:45:32.450 align:middle line:84% the counter doesn't change, whereas in the more higher one, 00:45:32.450 --> 00:45:35.060 align:middle line:84% it was always moving around so that she got w lower bar 00:45:35.060 --> 00:45:36.273 align:middle line:90% and expectation. 00:45:36.273 --> 00:45:40.440 align:middle line:90% 00:45:40.440 --> 00:45:42.630 align:middle line:84% Now on the other hand, if this thing 00:45:42.630 --> 00:45:47.363 align:middle line:84% is less than the lower bar, now the tenant 00:45:47.363 --> 00:45:48.780 align:middle line:84% has to actually-- the landlord has 00:45:48.780 --> 00:45:52.170 align:middle line:84% to increase the share that the tenant gets 00:45:52.170 --> 00:45:54.630 align:middle line:84% to keep in the high state, so that she gets at least 00:45:54.630 --> 00:45:58.240 align:middle line:90% her outside-option expectation. 00:45:58.240 --> 00:46:01.020 align:middle line:84% So what will happen, in that case, we'll just pick h, 00:46:01.020 --> 00:46:03.450 align:middle line:84% such that the expectation in the high state 00:46:03.450 --> 00:46:05.220 align:middle line:84% is just exactly equal to the-- sorry, 00:46:05.220 --> 00:46:06.640 align:middle line:84% the expected value of the contract 00:46:06.640 --> 00:46:08.182 align:middle line:84% is exactly equal to the outside wage. 00:46:08.182 --> 00:46:10.980 align:middle line:90% 00:46:10.980 --> 00:46:14.130 align:middle line:84% And note that, in this case, increasing w lower bar 00:46:14.130 --> 00:46:18.203 align:middle line:84% actually increases effort because the contract becomes-- 00:46:18.203 --> 00:46:19.620 align:middle line:84% as her outside option gets better, 00:46:19.620 --> 00:46:21.162 align:middle line:84% the contract becomes more attractive. 00:46:21.162 --> 00:46:25.495 align:middle line:90% 00:46:25.495 --> 00:46:25.995 align:middle line:90% Clear? 00:46:25.995 --> 00:46:29.090 align:middle line:90% 00:46:29.090 --> 00:46:37.440 align:middle line:84% So both-- key point here is both moral-hazard insurance 00:46:37.440 --> 00:46:45.450 align:middle line:84% rationales and limited-liability rationales, 00:46:45.450 --> 00:46:49.115 align:middle line:84% both generate what look like sharecropping contracts. 00:46:49.115 --> 00:46:50.490 align:middle line:84% But there are very different kind 00:46:50.490 --> 00:46:51.900 align:middle line:90% of stories for what's going on. 00:46:51.900 --> 00:46:55.060 align:middle line:84% And they have different implications. 00:46:55.060 --> 00:46:58.120 align:middle line:84% And in particular-- so both of these models 00:46:58.120 --> 00:47:01.117 align:middle line:84% generate the fact that sharecropping emerges, 00:47:01.117 --> 00:47:02.575 align:middle line:84% effort is less than the first best. 00:47:02.575 --> 00:47:05.320 align:middle line:90% 00:47:05.320 --> 00:47:08.200 align:middle line:84% But one thing that's really, in particular, kind of interesting 00:47:08.200 --> 00:47:13.120 align:middle line:84% here, is they differ in their implications for land reform. 00:47:13.120 --> 00:47:16.930 align:middle line:84% So as an aside, I feel-- and I'll talk about this 00:47:16.930 --> 00:47:17.900 align:middle line:90% in the next slide-- 00:47:17.900 --> 00:47:23.020 align:middle line:84% I feel like land reform is, in my opinion, 00:47:23.020 --> 00:47:26.980 align:middle line:90% a hugely understudied topic. 00:47:26.980 --> 00:47:31.630 align:middle line:84% I feel like one of the salient issues of the 20th century 00:47:31.630 --> 00:47:33.790 align:middle line:90% was should we do land reform. 00:47:33.790 --> 00:47:36.778 align:middle line:84% In some sense, that you can think of like the whole-- 00:47:36.778 --> 00:47:39.070 align:middle line:84% you could think of the whole communist revolutions were 00:47:39.070 --> 00:47:40.420 align:middle line:84% really about this question of, should we 00:47:40.420 --> 00:47:41.260 align:middle line:90% do land reform or not? 00:47:41.260 --> 00:47:42.250 align:middle line:84% And they wrote many other things, too, 00:47:42.250 --> 00:47:44.590 align:middle line:84% but should we take away land from the landlords 00:47:44.590 --> 00:47:45.740 align:middle line:90% and give it to the farmers? 00:47:45.740 --> 00:47:49.090 align:middle line:84% It was a huge question in the world in the 20th century. 00:47:49.090 --> 00:47:50.677 align:middle line:84% And I feel like we don't have great-- 00:47:50.677 --> 00:47:52.510 align:middle line:84% I think the level of evidence we have is not 00:47:52.510 --> 00:47:59.110 align:middle line:84% commensurate with the importance of the question, in my view. 00:47:59.110 --> 00:48:02.050 align:middle line:84% But in any case, these models differ a lot, 00:48:02.050 --> 00:48:05.860 align:middle line:84% in terms of the implications of land reform. 00:48:05.860 --> 00:48:08.420 align:middle line:90% So what do I mean by that? 00:48:08.420 --> 00:48:15.760 align:middle line:84% So in the risk-aversion model, we do a land reform. 00:48:15.760 --> 00:48:18.160 align:middle line:84% We say, OK, now, tenant, you're the owner of this land. 00:48:18.160 --> 00:48:21.580 align:middle line:90% 00:48:21.580 --> 00:48:23.920 align:middle line:84% That does make you wealthier and makes you 00:48:23.920 --> 00:48:24.930 align:middle line:90% maybe less risk averse. 00:48:24.930 --> 00:48:26.110 align:middle line:84% But on the other hand, your risk is all kind of 00:48:26.110 --> 00:48:28.250 align:middle line:84% tied up in your-- your wealth all tied up in the land. 00:48:28.250 --> 00:48:29.560 align:middle line:84% So it's not easy-- that's not clear you 00:48:29.560 --> 00:48:31.748 align:middle line:84% can get consumption-smoothing benefits from that. 00:48:31.748 --> 00:48:33.040 align:middle line:90% Maybe you can if you borrow it. 00:48:33.040 --> 00:48:34.930 align:middle line:90% Maybe you can't. 00:48:34.930 --> 00:48:39.790 align:middle line:84% But you're still-- would want some insurance. 00:48:39.790 --> 00:48:43.730 align:middle line:84% And so you might, as a tenant, say, 00:48:43.730 --> 00:48:49.720 align:middle line:84% well, I don't want to be in this position of being the owner 00:48:49.720 --> 00:48:51.040 align:middle line:90% and bearing all this risk. 00:48:51.040 --> 00:48:53.297 align:middle line:84% I might prefer, actually, to be in a contract 00:48:53.297 --> 00:48:54.880 align:middle line:84% where I'm sharing some of the benefits 00:48:54.880 --> 00:48:55.930 align:middle line:90% and some of the costs-- 00:48:55.930 --> 00:48:58.420 align:middle line:84% or sorry, sharing some of the returns in return 00:48:58.420 --> 00:48:59.740 align:middle line:90% for some kind of insurance. 00:48:59.740 --> 00:49:02.607 align:middle line:84% So you could imagine, in a risk-aversion model, 00:49:02.607 --> 00:49:04.690 align:middle line:84% that some kind of these sharecropping arrangements 00:49:04.690 --> 00:49:09.473 align:middle line:90% might re-emerge endogenously. 00:49:09.473 --> 00:49:11.140 align:middle line:84% You can imagine, for example, a landlord 00:49:11.140 --> 00:49:14.448 align:middle line:84% might make you an offer and say, I'll buy the land from you. 00:49:14.448 --> 00:49:16.240 align:middle line:84% And then I'll provide you a share contract. 00:49:16.240 --> 00:49:19.000 align:middle line:90% And that might be better-- 00:49:19.000 --> 00:49:21.010 align:middle line:84% including the payment they get upfront 00:49:21.010 --> 00:49:22.270 align:middle line:90% for the value of the land-- 00:49:22.270 --> 00:49:23.170 align:middle line:84% they might be better off for that 00:49:23.170 --> 00:49:25.250 align:middle line:84% because then they're getting some insurance values later. 00:49:25.250 --> 00:49:26.333 align:middle line:90% And there's a trade there. 00:49:26.333 --> 00:49:30.430 align:middle line:84% But because insurance is helpful, 00:49:30.430 --> 00:49:33.760 align:middle line:90% they might emerge endogenously. 00:49:33.760 --> 00:49:36.820 align:middle line:90% Under limited liability, no. 00:49:36.820 --> 00:49:39.550 align:middle line:84% Under a limited-liability model, the only reason 00:49:39.550 --> 00:49:42.430 align:middle line:84% we had the sharecropping was we couldn't make this payment 00:49:42.430 --> 00:49:45.070 align:middle line:84% from the tenant to the landlord in the low state. 00:49:45.070 --> 00:49:48.610 align:middle line:90% But if I'm the owner, I'm good. 00:49:48.610 --> 00:49:50.117 align:middle line:90% I don't pay anybody. 00:49:50.117 --> 00:49:51.700 align:middle line:84% And I'm not risk averse in that model, 00:49:51.700 --> 00:49:52.880 align:middle line:90% in the risk-neutral model. 00:49:52.880 --> 00:49:58.240 align:middle line:84% So which of these models is important would 00:49:58.240 --> 00:49:59.800 align:middle line:84% have implications for what you would 00:49:59.800 --> 00:50:08.330 align:middle line:84% do in the context of a [? land owner. ?] OK, comments, 00:50:08.330 --> 00:50:11.020 align:middle line:90% questions? 00:50:11.020 --> 00:50:11.760 align:middle line:90% Yeah, Kyle? 00:50:11.760 --> 00:50:13.860 align:middle line:84% AUDIENCE: Do we know about that question, 00:50:13.860 --> 00:50:17.745 align:middle line:84% whether it's good to give the land back to the [INAUDIBLE]?? 00:50:17.745 --> 00:50:20.120 align:middle line:84% BEN OLKEN: I'm going to talk about that in a few minutes. 00:50:20.120 --> 00:50:21.140 align:middle line:84% I'm not going to talk about it much. 00:50:21.140 --> 00:50:22.848 align:middle line:84% But I'm going to talk about a little bit. 00:50:22.848 --> 00:50:24.900 align:middle line:90% 00:50:24.900 --> 00:50:26.870 align:middle line:90% There are a couple of papers-- 00:50:26.870 --> 00:50:30.978 align:middle line:84% I'm still waiting for the great land-reform paper. 00:50:30.978 --> 00:50:33.270 align:middle line:84% There are some really nice papers that I'll talk about. 00:50:33.270 --> 00:50:36.980 align:middle line:84% But I feel like they have their various challenges. 00:50:36.980 --> 00:50:39.600 align:middle line:84% I'll talk about a few of them in a couple of slides. 00:50:39.600 --> 00:50:40.250 align:middle line:90% Yeah, Becka? 00:50:40.250 --> 00:50:43.890 align:middle line:84% AUDIENCE: If both effects are present at the same time, does 00:50:43.890 --> 00:50:46.915 align:middle line:84% this just default to the [INAUDIBLE]?? 00:50:46.915 --> 00:50:47.790 align:middle line:90% BEN OLKEN: I'm sorry. 00:50:47.790 --> 00:50:48.623 align:middle line:90% I didn't understand. 00:50:48.623 --> 00:50:50.960 align:middle line:84% AUDIENCE: If both effects are present at the same time. 00:50:50.960 --> 00:50:52.668 align:middle line:84% BEN OLKEN: No, I think he would do some-- 00:50:52.668 --> 00:50:54.790 align:middle line:90% 00:50:54.790 --> 00:50:55.290 align:middle line:90% hold on. 00:50:55.290 --> 00:50:55.880 align:middle line:90% Let me think. 00:50:55.880 --> 00:50:56.480 align:middle line:90% Let me think. 00:50:56.480 --> 00:50:59.510 align:middle line:90% 00:50:59.510 --> 00:51:03.920 align:middle line:84% Yeah, it's got to because if you're capped at-- 00:51:03.920 --> 00:51:06.455 align:middle line:84% the whole limited-liability thing-- 00:51:06.455 --> 00:51:11.430 align:middle line:90% 00:51:11.430 --> 00:51:18.493 align:middle line:84% if you add risk aversion to the limited-liability thing, 00:51:18.493 --> 00:51:20.160 align:middle line:84% yeah, you're still going have a problem. 00:51:20.160 --> 00:51:22.950 align:middle line:84% You can't pay the rent in the low state. 00:51:22.950 --> 00:51:24.180 align:middle line:90% So yeah, I think it's-- 00:51:24.180 --> 00:51:27.360 align:middle line:84% I think if the thing is binding, then 00:51:27.360 --> 00:51:30.090 align:middle line:84% that limits your ability to trade off a fixed 00:51:30.090 --> 00:51:31.600 align:middle line:90% fee for insurance value. 00:51:31.600 --> 00:51:36.830 align:middle line:90% So, yes, I think so. 00:51:36.830 --> 00:51:37.700 align:middle line:90% Other questions? 00:51:37.700 --> 00:51:40.310 align:middle line:90% 00:51:40.310 --> 00:51:46.420 align:middle line:84% OK, so now, that brings us to the paper you guys wrote 00:51:46.420 --> 00:51:54.130 align:middle line:84% for today, which is, do we actually think that 00:51:54.130 --> 00:52:00.820 align:middle line:84% they're-- so one implication of these models is that there is 00:52:00.820 --> 00:52:06.130 align:middle line:84% some-- that the incentive-- that the contract is going to affect 00:52:06.130 --> 00:52:09.887 align:middle line:90% kind of the effort, choice. 00:52:09.887 --> 00:52:11.470 align:middle line:84% And you don't actually need the full-- 00:52:11.470 --> 00:52:14.027 align:middle line:84% that's true under any of these models that we talked about. 00:52:14.027 --> 00:52:16.360 align:middle line:84% You can go back to the very simple one, the two-line one 00:52:16.360 --> 00:52:17.950 align:middle line:90% I put up in the very beginning. 00:52:17.950 --> 00:52:22.030 align:middle line:84% But that's a key parameter that we need to understand all this. 00:52:22.030 --> 00:52:26.650 align:middle line:84% So before you actually got through an experiment, 00:52:26.650 --> 00:52:28.480 align:middle line:84% before we had an experiment, if you didn't 00:52:28.480 --> 00:52:30.563 align:middle line:84% have an experiment, which is what we talked about, 00:52:30.563 --> 00:52:34.990 align:middle line:84% how would you how would you go about estimating this thing? 00:52:34.990 --> 00:52:36.850 align:middle line:84% What would you want to estimate this? 00:52:36.850 --> 00:52:44.218 align:middle line:90% 00:52:44.218 --> 00:52:46.010 align:middle line:84% What kind of regression do you want to run? 00:52:46.010 --> 00:52:54.880 align:middle line:90% 00:52:54.880 --> 00:52:55.870 align:middle line:90% Somebody? 00:52:55.870 --> 00:52:56.645 align:middle line:90% Yeah, Paulo? 00:52:56.645 --> 00:52:58.270 align:middle line:84% AUDIENCE: You may want something that's 00:52:58.270 --> 00:53:02.050 align:middle line:84% like output based, on what share of output 00:53:02.050 --> 00:53:05.916 align:middle line:90% do they have if they had-- 00:53:05.916 --> 00:53:07.782 align:middle line:84% can't tell the individual farmers-- 00:53:07.782 --> 00:53:08.740 align:middle line:90% BEN OLKEN: What's that? 00:53:08.740 --> 00:53:10.157 align:middle line:84% AUDIENCE: So you want a regression 00:53:10.157 --> 00:53:14.483 align:middle line:84% where your output variable is the output of the farm. 00:53:14.483 --> 00:53:15.900 align:middle line:84% And [INAUDIBLE] only have an input 00:53:15.900 --> 00:53:19.890 align:middle line:84% on one of your regressors, what share of output or profits 00:53:19.890 --> 00:53:22.772 align:middle line:84% they are getting, essentially, the terms of the contract. 00:53:22.772 --> 00:53:23.480 align:middle line:90% BEN OLKEN: Sorry. 00:53:23.480 --> 00:53:23.840 align:middle line:90% I'm sorry. 00:53:23.840 --> 00:53:25.048 align:middle line:90% I don't fully-- say it again. 00:53:25.048 --> 00:53:27.470 align:middle line:84% So you-- yes, you won't put output on the left-hand side. 00:53:27.470 --> 00:53:28.940 align:middle line:84% What you want put the right-hand side? 00:53:28.940 --> 00:53:30.710 align:middle line:84% AUDIENCE: Basically terms of the contract? 00:53:30.710 --> 00:53:31.460 align:middle line:90% BEN OLKEN: Right, exactly. 00:53:31.460 --> 00:53:31.960 align:middle line:90% Right. 00:53:31.960 --> 00:53:33.440 align:middle line:90% The same thing you would do-- 00:53:33.440 --> 00:53:35.480 align:middle line:84% you'd like to regress what is the output on what 00:53:35.480 --> 00:53:36.730 align:middle line:90% are the terms of the contract. 00:53:36.730 --> 00:53:39.410 align:middle line:90% 00:53:39.410 --> 00:53:42.770 align:middle line:84% But the problem is, of course, that those-- 00:53:42.770 --> 00:53:44.880 align:middle line:84% those farmers are going to be different. 00:53:44.880 --> 00:53:48.560 align:middle line:84% So one thing to do, actually, would 00:53:48.560 --> 00:53:50.090 align:middle line:90% be if you had multiple farmer-- 00:53:50.090 --> 00:53:52.792 align:middle line:84% if you had farmers who had multiple plots-- 00:53:52.792 --> 00:53:54.500 align:middle line:84% so the basic problem with that regression 00:53:54.500 --> 00:53:56.780 align:middle line:84% is we have farm-- there might be farmer characteristics that 00:53:56.780 --> 00:53:57.740 align:middle line:90% might be different. 00:53:57.740 --> 00:53:59.740 align:middle line:84% Farmers who enter into those different contracts 00:53:59.740 --> 00:54:01.140 align:middle line:90% might be different people. 00:54:01.140 --> 00:54:02.770 align:middle line:84% So one option would be to use farmers 00:54:02.770 --> 00:54:06.230 align:middle line:84% who have multiple plots, some share crop and some owned, 00:54:06.230 --> 00:54:09.060 align:middle line:84% and run this regression with farmer-fixed effects, and say, 00:54:09.060 --> 00:54:15.110 align:middle line:84% are they putting in more effort on the owned plots 00:54:15.110 --> 00:54:17.250 align:middle line:84% compared to the shared plots for example, 00:54:17.250 --> 00:54:19.790 align:middle line:90% or more inputs, or whatever? 00:54:19.790 --> 00:54:22.530 align:middle line:84% So would that be a good regression or a bad regression? 00:54:22.530 --> 00:54:25.350 align:middle line:84% So that regression would basically-- sorry, Erica. 00:54:25.350 --> 00:54:25.850 align:middle line:90% Go ahead. 00:54:25.850 --> 00:54:27.017 align:middle line:90% You can answer the question. 00:54:27.017 --> 00:54:28.730 align:middle line:84% AUDIENCE: Oh, could there be an issue 00:54:28.730 --> 00:54:33.090 align:middle line:84% with some qualities of the land, that cause it to be 00:54:33.090 --> 00:54:33.590 align:middle line:90% [INAUDIBLE]? 00:54:33.590 --> 00:54:35.360 align:middle line:84% BEN OLKEN: Absolutely, yes, so there 00:54:35.360 --> 00:54:37.580 align:middle line:84% can be a problem with quality of the land. 00:54:37.580 --> 00:54:39.560 align:middle line:84% So it may be that different land is 00:54:39.560 --> 00:54:41.930 align:middle line:84% owned versus rented, and therefore, that could 00:54:41.930 --> 00:54:42.930 align:middle line:90% be related to the issue. 00:54:42.930 --> 00:54:45.890 align:middle line:84% And so in fact, the regression you could run 00:54:45.890 --> 00:54:50.150 align:middle line:84% is some output as a function of whether it's owned or not, 00:54:50.150 --> 00:54:53.630 align:middle line:90% with farmer-fixed effects. 00:54:53.630 --> 00:54:56.210 align:middle line:84% And this could be either inputs or outputs. 00:54:56.210 --> 00:54:59.420 align:middle line:84% So the problem is exactly this one, that you mentioned, 00:54:59.420 --> 00:55:01.110 align:middle line:90% of quality. 00:55:01.110 --> 00:55:02.900 align:middle line:90% So this is from an older paper. 00:55:02.900 --> 00:55:08.930 align:middle line:84% And you can basically see that they have-- 00:55:08.930 --> 00:55:11.983 align:middle line:84% looking among the owned plot-- this doesn't have the fixed 00:55:11.983 --> 00:55:13.400 align:middle line:84% effects but I think the same thing 00:55:13.400 --> 00:55:15.358 align:middle line:84% goes through the fixed effects-- if you compare 00:55:15.358 --> 00:55:19.850 align:middle line:84% the owned plots, the sharecrop plots and the fixed-rent plots, 00:55:19.850 --> 00:55:26.690 align:middle line:84% the owned plots look a lot better, in terms of their, 00:55:26.690 --> 00:55:31.607 align:middle line:84% say, average value, than the sharecropped plots. 00:55:31.607 --> 00:55:33.440 align:middle line:84% Actually, in this case, the fixed-rent ones, 00:55:33.440 --> 00:55:37.910 align:middle line:84% I think, look the worst of all, at least 00:55:37.910 --> 00:55:39.120 align:middle line:90% in some of the villages. 00:55:39.120 --> 00:55:42.950 align:middle line:84% So the point, though, is that you may have different-- 00:55:42.950 --> 00:55:44.700 align:middle line:84% and by the way, this is too small to read. 00:55:44.700 --> 00:55:46.290 align:middle line:90% That's why you have the handout. 00:55:46.290 --> 00:55:53.190 align:middle line:84% So that's why this regression may be challenging. 00:55:53.190 --> 00:55:56.390 align:middle line:84% And so the original approach is, before we had this RCT, 00:55:56.390 --> 00:55:58.680 align:middle line:84% in the paper that we used to teach here, was to say, 00:55:58.680 --> 00:56:03.290 align:middle line:84% let's control for really detailed plot characteristics. 00:56:03.290 --> 00:56:07.670 align:middle line:84% And the problem here is that there 00:56:07.670 --> 00:56:09.800 align:middle line:84% may be unobservable characteristics of the plot, 00:56:09.800 --> 00:56:12.370 align:middle line:84% beyond the measurable plot characteristics. 00:56:12.370 --> 00:56:15.380 align:middle line:84% It's really hard to-- if you end up 00:56:15.380 --> 00:56:19.620 align:middle line:84% writing an agricultural paper, you'll go into the weeds, 00:56:19.620 --> 00:56:20.360 align:middle line:90% so to speak. 00:56:20.360 --> 00:56:23.975 align:middle line:84% And you will find an incredible amount of-- 00:56:23.975 --> 00:56:26.900 align:middle line:90% 00:56:26.900 --> 00:56:30.470 align:middle line:84% how you describe a plot characteristics and soil 00:56:30.470 --> 00:56:35.310 align:middle line:84% quality and irrigation, water, quality and this, that, 00:56:35.310 --> 00:56:39.090 align:middle line:84% and the other thing, it's many, many, many dimensional. 00:56:39.090 --> 00:56:42.410 align:middle line:84% And so that approach is really hard to get that right. 00:56:42.410 --> 00:56:45.920 align:middle line:84% So what would be cool would be to have 00:56:45.920 --> 00:56:49.102 align:middle line:84% plot-fixed effects in this regression, 00:56:49.102 --> 00:56:50.310 align:middle line:90% like maybe some plots change. 00:56:50.310 --> 00:56:50.900 align:middle line:84% But then, of course, the question 00:56:50.900 --> 00:56:51.900 align:middle line:90% is why they're changing. 00:56:51.900 --> 00:56:54.422 align:middle line:84% So all of this, basically-- actually, I 00:56:54.422 --> 00:56:56.630 align:middle line:84% can't find a paper that does that because [INAUDIBLE] 00:56:56.630 --> 00:56:57.350 align:middle line:90% contracts don't change. 00:56:57.350 --> 00:56:59.090 align:middle line:84% And we don't know why they don't change. 00:56:59.090 --> 00:57:02.340 align:middle line:90% 00:57:02.340 --> 00:57:04.040 align:middle line:84% So that all leads us to the paper you 00:57:04.040 --> 00:57:06.745 align:middle line:84% guys read for today, which is, I think, why this paper-- 00:57:06.745 --> 00:57:08.870 align:middle line:84% even though, as some of you guys noted, the setting 00:57:08.870 --> 00:57:11.800 align:middle line:84% is a little unusual, it's these kind of-- 00:57:11.800 --> 00:57:19.820 align:middle line:84% whoops-- new farmers who are interested in farming, as part 00:57:19.820 --> 00:57:20.840 align:middle line:90% of these women's groups. 00:57:20.840 --> 00:57:23.390 align:middle line:84% And so the external validity of this 00:57:23.390 --> 00:57:25.760 align:middle line:84% may not necessarily be 100% there. 00:57:25.760 --> 00:57:31.010 align:middle line:84% The reason this paper got so much attention is it 00:57:31.010 --> 00:57:33.410 align:middle line:84% is the cleanest paper that we know of, 00:57:33.410 --> 00:57:36.500 align:middle line:90% that actually tests this. 00:57:36.500 --> 00:57:41.790 align:middle line:84% So as you guys saw, what are they going to do? 00:57:41.790 --> 00:57:45.440 align:middle line:84% They're going to work with tenant farmers in Uganda, 00:57:45.440 --> 00:57:50.010 align:middle line:84% randomize them to receive either 50% of their output-- 00:57:50.010 --> 00:57:51.710 align:middle line:84% so after they recruit them, so it's 00:57:51.710 --> 00:57:53.168 align:middle line:84% going to be surprise randomization. 00:57:53.168 --> 00:57:55.010 align:middle line:90% So there's no selection issues. 00:57:55.010 --> 00:57:58.670 align:middle line:84% Have either 50-- so the baseline is 50% of output. 00:57:58.670 --> 00:58:01.220 align:middle line:84% Or you can get to keep 75% of your output, 00:58:01.220 --> 00:58:05.240 align:middle line:84% or 50% of your output plus an exogenous cash transfer, 00:58:05.240 --> 00:58:08.990 align:middle line:84% calibrated to be equal to 25% of output and expectation, 00:58:08.990 --> 00:58:12.930 align:middle line:84% either is a fixed-fee or a risky contract. 00:58:12.930 --> 00:58:14.360 align:middle line:90% So that's the basic design. 00:58:14.360 --> 00:58:17.160 align:middle line:90% 00:58:17.160 --> 00:58:19.363 align:middle line:90% So what do you guys-- 00:58:19.363 --> 00:58:21.530 align:middle line:84% I'm curious to know what did you guys think of this. 00:58:21.530 --> 00:58:23.670 align:middle line:84% You like this-- you like this paper 00:58:23.670 --> 00:58:24.920 align:middle line:90% or you have concerns about it? 00:58:24.920 --> 00:58:25.700 align:middle line:90% What do you think? 00:58:25.700 --> 00:58:31.350 align:middle line:90% 00:58:31.350 --> 00:58:32.520 align:middle line:90% Yeah, Erin. 00:58:32.520 --> 00:58:35.168 align:middle line:84% AUDIENCE: So one thing I had was with the-- 00:58:35.168 --> 00:58:36.960 align:middle line:84% because they're trying to test for the fact 00:58:36.960 --> 00:58:39.840 align:middle line:84% that these incentive effects do explain what they observe, 00:58:39.840 --> 00:58:42.640 align:middle line:90% not income or risk effect. 00:58:42.640 --> 00:58:47.460 align:middle line:84% And so they argue that their risky, exogenous cash transfer 00:58:47.460 --> 00:58:50.290 align:middle line:84% is a way of making sure that these risk effects don't exist. 00:58:50.290 --> 00:58:53.190 align:middle line:84% What I-- I know this thing that's-- they draw really 00:58:53.190 --> 00:58:57.030 align:middle line:84% heavily on this more [? commercial ?] kind of model, 00:58:57.030 --> 00:59:01.020 align:middle line:84% but that in that model, what you see is that the probability 00:59:01.020 --> 00:59:05.850 align:middle line:84% of having a high output is dependent on your inputs, 00:59:05.850 --> 00:59:08.511 align:middle line:90% whereas if their's-- 00:59:08.511 --> 00:59:14.310 align:middle line:84% for the risky, exogenous case-- and it's not-- 00:59:14.310 --> 00:59:19.410 align:middle line:84% it's-- whether or not get that additional 25%, 00:59:19.410 --> 00:59:20.452 align:middle line:90% in the risky case, it's-- 00:59:20.452 --> 00:59:21.743 align:middle line:90% BEN OLKEN: Is exogenous, right? 00:59:21.743 --> 00:59:24.300 align:middle line:84% AUDIENCE: Yeah and that was [INAUDIBLE] versus in the 75% 00:59:24.300 --> 00:59:25.410 align:middle line:90% of output, that-- 00:59:25.410 --> 00:59:27.910 align:middle line:90% actually 25% was what we get. 00:59:27.910 --> 00:59:28.840 align:middle line:90% And so I felt that-- 00:59:28.840 --> 00:59:30.640 align:middle line:84% BEN OLKEN: So what's the concern, there? 00:59:30.640 --> 00:59:33.060 align:middle line:90% AUDIENCE: Well, given that-- 00:59:33.060 --> 00:59:36.630 align:middle line:84% if this additional 25% depends on your inputs, 00:59:36.630 --> 00:59:39.820 align:middle line:84% your inputs are going to change, versus if it is not. 00:59:39.820 --> 00:59:43.850 align:middle line:84% And so I don't understand how, then, you 00:59:43.850 --> 00:59:47.718 align:middle line:84% can say that risk does not account for the [? seasons. ?] 00:59:47.718 --> 00:59:49.260 align:middle line:84% BEN OLKEN: So does anyone else have-- 00:59:49.260 --> 00:59:50.700 align:middle line:90% anyone want to respond to that? 00:59:50.700 --> 00:59:54.030 align:middle line:90% 00:59:54.030 --> 00:59:55.110 align:middle line:90% Any idea, any thoughts? 00:59:55.110 --> 01:00:00.720 align:middle line:90% 01:00:00.720 --> 01:00:02.130 align:middle line:90% So-- no? 01:00:02.130 --> 01:00:03.840 align:middle line:90% Does anyone-- yeah? 01:00:03.840 --> 01:00:06.540 align:middle line:84% AUDIENCE: I'm thinking of the big st-- 01:00:06.540 --> 01:00:10.140 align:middle line:84% like the 50% was an exogenous cash transfer, 01:00:10.140 --> 01:00:11.850 align:middle line:90% where they get the fixed amount. 01:00:11.850 --> 01:00:13.703 align:middle line:90% That insures you completely, no? 01:00:13.703 --> 01:00:14.870 align:middle line:90% BEN OLKEN: That is the what? 01:00:14.870 --> 01:00:16.750 align:middle line:84% AUDIENCE: It insures you completely. 01:00:16.750 --> 01:00:21.020 align:middle line:84% Let's just say insurance, so maybe 35% [INAUDIBLE].. 01:00:21.020 --> 01:00:22.912 align:middle line:84% BEN OLKEN: Yeah, so the point of this one 01:00:22.912 --> 01:00:23.870 align:middle line:90% is not about insurance. 01:00:23.870 --> 01:00:26.360 align:middle line:84% The point of this is the income effects. 01:00:26.360 --> 01:00:29.760 align:middle line:84% So the point of this one is to say, well, 01:00:29.760 --> 01:00:31.970 align:middle line:84% maybe they're doing more stuff because they're 01:00:31.970 --> 01:00:33.860 align:middle line:90% richer in expectation. 01:00:33.860 --> 01:00:36.545 align:middle line:84% Maybe-- somehow maybe they're not quite credit constrained, 01:00:36.545 --> 01:00:38.420 align:middle line:84% but maybe they're sort of credit constrained. 01:00:38.420 --> 01:00:39.500 align:middle line:84% Maybe they only feel like they can 01:00:39.500 --> 01:00:42.080 align:middle line:84% afford to buy fertilizer if they're going to get a higher 01:00:42.080 --> 01:00:43.070 align:middle line:90% payment at the end. 01:00:43.070 --> 01:00:47.430 align:middle line:84% So that's what this one is for, is the level. 01:00:47.430 --> 01:00:49.070 align:middle line:84% Well, I think I think Erin's concern, 01:00:49.070 --> 01:00:51.570 align:middle line:84% if I understand it correctly, Erin, correct me if I'm wrong, 01:00:51.570 --> 01:00:57.020 align:middle line:84% was that the issue with this one is that the payment, here, 01:00:57.020 --> 01:00:58.940 align:middle line:90% doesn't depend on your inputs. 01:00:58.940 --> 01:01:04.250 align:middle line:84% But-- and so it doesn't exactly mirror the riskiness you would 01:01:04.250 --> 01:01:10.580 align:middle line:84% have in this contract because you would-- 01:01:10.580 --> 01:01:13.807 align:middle line:84% yeah, because here, you would also be putting in more inputs. 01:01:13.807 --> 01:01:15.890 align:middle line:84% And therefore your net returns would be different. 01:01:15.890 --> 01:01:16.682 align:middle line:90% Is that your point? 01:01:16.682 --> 01:01:17.270 align:middle line:90% Yeah. 01:01:17.270 --> 01:01:20.090 align:middle line:84% So I think the way-- at least the way I think about this is 01:01:20.090 --> 01:01:23.460 align:middle line:84% to say, this is actually testing a different issue, 01:01:23.460 --> 01:01:27.785 align:middle line:84% which is to say, look, let's suppose that the-- 01:01:27.785 --> 01:01:28.740 align:middle line:90% there's two thoughts. 01:01:28.740 --> 01:01:29.615 align:middle line:90% The first is to say-- 01:01:29.615 --> 01:01:32.150 align:middle line:90% 01:01:32.150 --> 01:01:39.350 align:middle line:84% so first order-- this contract gives you a lot-- 01:01:39.350 --> 01:01:40.648 align:middle line:90% sorry, pluses and minuses. 01:01:40.648 --> 01:01:42.440 align:middle line:84% I think what they're trying to say is, yes, 01:01:42.440 --> 01:01:45.840 align:middle line:84% this contract is just generally more risky than the other one. 01:01:45.840 --> 01:01:48.020 align:middle line:84% And so if you thought that what was happening 01:01:48.020 --> 01:01:49.760 align:middle line:84% is I was changing my behavior because I'm 01:01:49.760 --> 01:01:52.580 align:middle line:84% exposed to more risk, in some general sense, 01:01:52.580 --> 01:01:55.310 align:middle line:84% you would see things that look more similar between this 01:01:55.310 --> 01:01:56.220 align:middle line:90% and this. 01:01:56.220 --> 01:02:00.397 align:middle line:84% So I agree with you that it is different. 01:02:00.397 --> 01:02:02.480 align:middle line:84% It's not exactly the same in a couple of respects. 01:02:02.480 --> 01:02:09.330 align:middle line:84% The first is, this risky contract is actually not 01:02:09.330 --> 01:02:12.455 align:middle line:84% as risky as this risky contract because this one, 01:02:12.455 --> 01:02:13.830 align:middle line:84% I think, if I remember correctly, 01:02:13.830 --> 01:02:16.230 align:middle line:84% is exogenous risk, where this one is-- 01:02:16.230 --> 01:02:20.160 align:middle line:84% sorry, independent risk, whereas this one is correlated 01:02:20.160 --> 01:02:21.690 align:middle line:90% with the rest of your risk. 01:02:21.690 --> 01:02:23.940 align:middle line:84% So actually, to me, the first-order issue is that one, 01:02:23.940 --> 01:02:25.050 align:middle line:90% which is like-- 01:02:25.050 --> 01:02:29.987 align:middle line:84% this one is-- and I don't think-- this one-- 01:02:29.987 --> 01:02:31.320 align:middle line:90% someone correct me if I'm wrong. 01:02:31.320 --> 01:02:33.420 align:middle line:84% This is just an exogenous, risky payment, right? 01:02:33.420 --> 01:02:35.880 align:middle line:90% Yeah, it's like a random draw. 01:02:35.880 --> 01:02:38.220 align:middle line:84% So one thing that might have been even slightly better 01:02:38.220 --> 01:02:39.300 align:middle line:84% to deal with-- this is not your issue, 01:02:39.300 --> 01:02:40.440 align:middle line:84% but to deal with this other issue-- 01:02:40.440 --> 01:02:41.815 align:middle line:84% might have been to say, well, I'm 01:02:41.815 --> 01:02:46.140 align:middle line:84% going to tie the risky contract to, say, average returns 01:02:46.140 --> 01:02:51.060 align:middle line:84% in the neighborhood, in your area, so at least it 01:02:51.060 --> 01:02:54.030 align:middle line:84% becomes a correlated shock, not an independent shock, 01:02:54.030 --> 01:02:58.290 align:middle line:84% because otherwise, this is less risky than this 01:02:58.290 --> 01:03:00.405 align:middle line:90% because it's independent. 01:03:00.405 --> 01:03:03.460 align:middle line:90% You see what I'm saying? 01:03:03.460 --> 01:03:04.950 align:middle line:90% It's not a correlated shock. 01:03:04.950 --> 01:03:07.950 align:middle line:84% To your point, I don't know of a good way-- 01:03:07.950 --> 01:03:10.500 align:middle line:84% I see your point that the net risk-- your point 01:03:10.500 --> 01:03:13.260 align:middle line:84% is if I write down the utility, under the things, 01:03:13.260 --> 01:03:15.240 align:middle line:84% that the net risk is the same because I haven't 01:03:15.240 --> 01:03:16.532 align:middle line:90% put in these additional inputs. 01:03:16.532 --> 01:03:18.810 align:middle line:84% I don't know of a good way to deal with that, though. 01:03:18.810 --> 01:03:19.320 align:middle line:90% Do you? 01:03:19.320 --> 01:03:19.862 align:middle line:90% AUDIENCE: No. 01:03:19.862 --> 01:03:20.763 align:middle line:90% [INAUDIBLE] 01:03:20.763 --> 01:03:24.130 align:middle line:90% 01:03:24.130 --> 01:03:26.110 align:middle line:84% BEN OLKEN: What are some other concerns-- what 01:03:26.110 --> 01:03:28.193 align:middle line:84% other concerns did you guys have about this paper? 01:03:28.193 --> 01:03:32.717 align:middle line:90% 01:03:32.717 --> 01:03:35.300 align:middle line:84% I know in general, actually, I haven't read all your comments. 01:03:35.300 --> 01:03:36.508 align:middle line:90% You all pretty much liked it. 01:03:36.508 --> 01:03:38.480 align:middle line:84% But were there other concerns that you guys 01:03:38.480 --> 01:03:41.145 align:middle line:90% had, looking over it? 01:03:41.145 --> 01:03:43.437 align:middle line:84% Or things you would have wanted them to do differently? 01:03:43.437 --> 01:03:49.960 align:middle line:90% 01:03:49.960 --> 01:03:50.460 align:middle line:90% Yeah, Paulo? 01:03:50.460 --> 01:03:50.765 align:middle line:90% 01:03:50.765 --> 01:03:53.098 align:middle line:84% AUDIENCE: I think that probably one of the things that I 01:03:53.098 --> 01:03:55.850 align:middle line:84% was thinking about was, in making this experiment ex 01:03:55.850 --> 01:03:57.860 align:middle line:90% ante, and what-- 01:03:57.860 --> 01:03:59.360 align:middle line:84% the outcome that you're going to get 01:03:59.360 --> 01:04:01.580 align:middle line:84% is either something along the lines of, 01:04:01.580 --> 01:04:05.342 align:middle line:84% OK, it's mainly income effects or it's mainly 01:04:05.342 --> 01:04:06.800 align:middle line:84% incentive and moral-hazard effects, 01:04:06.800 --> 01:04:09.420 align:middle line:84% or some combination of the two, or neither of them. 01:04:09.420 --> 01:04:11.210 align:middle line:84% And so I think for me, just in terms 01:04:11.210 --> 01:04:14.420 align:middle line:84% of policy applications, what is one supposed 01:04:14.420 --> 01:04:16.322 align:middle line:90% to get from this paper? 01:04:16.322 --> 01:04:17.780 align:middle line:84% Had it been income effects, I think 01:04:17.780 --> 01:04:22.040 align:middle line:84% there's a very clear answer of, just give these tenants money 01:04:22.040 --> 01:04:23.790 align:middle line:84% and then you can be more productive. 01:04:23.790 --> 01:04:27.115 align:middle line:84% But given that the answer was this, mainly, a moral hazard-- 01:04:27.115 --> 01:04:28.490 align:middle line:84% BEN OLKEN: There's moral hazard-- 01:04:28.490 --> 01:04:30.650 align:middle line:84% AUDIENCE: Yeah, so what is going to be 01:04:30.650 --> 01:04:33.980 align:middle line:84% the reform that a government or something does, as a result? 01:04:33.980 --> 01:04:36.298 align:middle line:84% It seems very hard to just tell everyone 01:04:36.298 --> 01:04:38.090 align:middle line:84% to change the terms of contracts if they're 01:04:38.090 --> 01:04:39.620 align:middle line:90% tenants or things like that. 01:04:39.620 --> 01:04:41.200 align:middle line:84% BEN OLKEN: That's a good question. 01:04:41.200 --> 01:04:42.960 align:middle line:90% Are there comments to that? 01:04:42.960 --> 01:04:45.620 align:middle line:84% Did anyone else-- is there an obvious policy implication 01:04:45.620 --> 01:04:46.532 align:middle line:90% from this? 01:04:46.532 --> 01:04:49.190 align:middle line:90% 01:04:49.190 --> 01:04:49.690 align:middle line:90% Yeah? 01:04:49.690 --> 01:04:51.565 align:middle line:84% AUDIENCE: One thing that I was thinking about 01:04:51.565 --> 01:04:54.040 align:middle line:84% was, instead of just saying, oh, I'm 01:04:54.040 --> 01:04:57.460 align:middle line:84% going to take 50% of your output or only 25% of your output, 01:04:57.460 --> 01:05:01.030 align:middle line:84% if you could have some staggered design in your contract, where 01:05:01.030 --> 01:05:04.750 align:middle line:84% you would say, I'll initially take 50%, 01:05:04.750 --> 01:05:08.200 align:middle line:84% but if you make above this amount in your output, 01:05:08.200 --> 01:05:09.310 align:middle line:90% I'm going to take less. 01:05:09.310 --> 01:05:11.020 align:middle line:84% So then you have an incentive to keep 01:05:11.020 --> 01:05:12.860 align:middle line:90% producing more as you go along. 01:05:12.860 --> 01:05:15.580 align:middle line:84% And I would be curious what that would actually look like, 01:05:15.580 --> 01:05:17.110 align:middle line:90% implemented, if that would-- 01:05:17.110 --> 01:05:18.700 align:middle line:84% BEN OLKEN: Yeah, so this gets back to this whole question of 01:05:18.700 --> 01:05:20.420 align:middle line:90% are linear contracts optimal. 01:05:20.420 --> 01:05:23.350 align:middle line:84% So I'm going to not go into that in a lot of detail, 01:05:23.350 --> 01:05:26.600 align:middle line:84% and refer you to a contract theory class for this question. 01:05:26.600 --> 01:05:29.060 align:middle line:84% But I think there's a couple of issues. 01:05:29.060 --> 01:05:30.070 align:middle line:90% The first is-- 01:05:30.070 --> 01:05:32.680 align:middle line:84% I think the general issue with non-linear contracts is you 01:05:32.680 --> 01:05:34.930 align:middle line:90% want to put the-- 01:05:34.930 --> 01:05:39.250 align:middle line:84% you want to put the part where you give people 01:05:39.250 --> 01:05:43.000 align:middle line:84% the most incentive at the point where they're most elastic. 01:05:43.000 --> 01:05:47.110 align:middle line:84% And figuring that out, with heterogeneity across people, 01:05:47.110 --> 01:05:49.060 align:middle line:90% can be tricky. 01:05:49.060 --> 01:05:49.850 align:middle line:90% So it might be-- 01:05:49.850 --> 01:05:51.850 align:middle line:84% and if I put it in the wrong place, then I can-- 01:05:51.850 --> 01:05:54.250 align:middle line:90% so you want to say, I want to-- 01:05:54.250 --> 01:05:58.060 align:middle line:84% that can be a tricky set of issues. 01:05:58.060 --> 01:06:00.340 align:middle line:84% And the other thing is, you can also get gaming-- 01:06:00.340 --> 01:06:02.410 align:middle line:84% if you can do things across periods, 01:06:02.410 --> 01:06:03.628 align:middle line:90% you can sometimes get gaming. 01:06:03.628 --> 01:06:05.920 align:middle line:84% I don't know if that's really an issue in sharecropping 01:06:05.920 --> 01:06:06.430 align:middle line:90% contracts. 01:06:06.430 --> 01:06:10.387 align:middle line:84% And actually, I don't know of a good-- 01:06:10.387 --> 01:06:11.470 align:middle line:90% I'm sure there is a paper. 01:06:11.470 --> 01:06:13.512 align:middle line:84% I just can't think of it off the top of my head-- 01:06:13.512 --> 01:06:17.620 align:middle line:84% that has taken some of the Holmstrom linear-contracts 01:06:17.620 --> 01:06:21.108 align:middle line:84% ideas and said, do these gaming or other issues 01:06:21.108 --> 01:06:22.900 align:middle line:84% apply to this [? sharecropping contract. ?] 01:06:22.900 --> 01:06:26.110 align:middle line:84% I don't know the answer to that question. 01:06:26.110 --> 01:06:26.908 align:middle line:90% Yeah, Wesley? 01:06:26.908 --> 01:06:29.200 align:middle line:84% AUDIENCE: I only followed the policy division in things 01:06:29.200 --> 01:06:31.930 align:middle line:84% because what seems a little bit strange about this particular 01:06:31.930 --> 01:06:36.640 align:middle line:84% setting, these would be how sharecropping contracts 01:06:36.640 --> 01:06:40.475 align:middle line:84% actually work in practice, is that [INAUDIBLE] does not-- 01:06:40.475 --> 01:06:42.850 align:middle line:84% [INAUDIBLE] does not care if it's maximizing his profits, 01:06:42.850 --> 01:06:44.933 align:middle line:84% if he's is getting as much out these sharecroppers 01:06:44.933 --> 01:06:46.230 align:middle line:90% as it possibly can. 01:06:46.230 --> 01:06:51.180 align:middle line:84% And so even if the result is that if you 01:06:51.180 --> 01:06:54.090 align:middle line:84% were to increase the portion of the output 01:06:54.090 --> 01:06:56.160 align:middle line:84% that your share properties can keep, 01:06:56.160 --> 01:06:59.370 align:middle line:84% there's not really a way to impose that 01:06:59.370 --> 01:07:01.890 align:middle line:84% on landlords who are looking to extract as much as they can 01:07:01.890 --> 01:07:05.230 align:middle line:84% out of their contract, unless, similar to what 01:07:05.230 --> 01:07:07.480 align:middle line:84% you were saying, if there was like some income effect, 01:07:07.480 --> 01:07:08.400 align:middle line:90% where a governm-- 01:07:08.400 --> 01:07:10.150 align:middle line:84% or if your landlord is the government, who 01:07:10.150 --> 01:07:12.130 align:middle line:84% doesn't particularly care about taking a loss 01:07:12.130 --> 01:07:15.060 align:middle line:84% and is just looking to increase the output on the land. 01:07:15.060 --> 01:07:17.228 align:middle line:90% BEN OLKEN: Yeah, so I think-- 01:07:17.228 --> 01:07:19.020 align:middle line:84% my own take on this is, I don't necessarily 01:07:19.020 --> 01:07:21.395 align:middle line:84% think there's an obvious policy implication of this paper 01:07:21.395 --> 01:07:24.630 align:middle line:84% because it doesn't actually tell us 01:07:24.630 --> 01:07:29.100 align:middle line:84% if it's moral haz-- if it's the insurance kind of-- insurance 01:07:29.100 --> 01:07:29.760 align:middle line:90% issue. 01:07:29.760 --> 01:07:32.427 align:middle line:84% It doesn't tell us-- it tells us there's some inefficiency going 01:07:32.427 --> 01:07:34.320 align:middle line:90% on, in terms of the-- 01:07:34.320 --> 01:07:37.170 align:middle line:84% it says that there's some inefficiency going on, 01:07:37.170 --> 01:07:39.840 align:middle line:84% where the sharecropping is reducing effort. 01:07:39.840 --> 01:07:41.123 align:middle line:90% But it doesn't say-- 01:07:41.123 --> 01:07:42.540 align:middle line:84% it doesn't address the question we 01:07:42.540 --> 01:07:45.082 align:middle line:84% talked about before, about why is this emerging endogenously. 01:07:45.082 --> 01:07:47.580 align:middle line:84% And it seems that we know that in order 01:07:47.580 --> 01:07:49.028 align:middle line:90% to figure out the policy answer. 01:07:49.028 --> 01:07:51.570 align:middle line:84% So I think-- to me, this is an important piece of the puzzle. 01:07:51.570 --> 01:07:54.330 align:middle line:84% It says, first point is, yes, these contracts 01:07:54.330 --> 01:07:56.560 align:middle line:90% have efficiency consequences. 01:07:56.560 --> 01:07:58.018 align:middle line:84% So that's actually really important 01:07:58.018 --> 01:08:00.185 align:middle line:84% because if they didn't have efficiency consequences, 01:08:00.185 --> 01:08:02.527 align:middle line:84% we would just not worry about this whole set of issues, 01:08:02.527 --> 01:08:04.110 align:middle line:84% from an aggregate-wealth perspective-- 01:08:04.110 --> 01:08:06.300 align:middle line:84% aggregate-productivity perspective. 01:08:06.300 --> 01:08:07.938 align:middle line:84% The fact that it does say that there's 01:08:07.938 --> 01:08:10.230 align:middle line:84% like efficiency consequences says to me that, well, OK, 01:08:10.230 --> 01:08:13.930 align:middle line:84% now we want to figure out what's driving it. 01:08:13.930 --> 01:08:18.569 align:middle line:84% And if you really thought that the issue was about insurance 01:08:18.569 --> 01:08:21.000 align:middle line:84% value, is there a way of providing 01:08:21.000 --> 01:08:25.710 align:middle line:84% insurance that doesn't have the same moral-hazard consequences? 01:08:25.710 --> 01:08:29.160 align:middle line:84% And there is this a small, but growing literature 01:08:29.160 --> 01:08:31.620 align:middle line:84% on agricultural insurance issues, which-- 01:08:31.620 --> 01:08:33.600 align:middle line:84% and I'll tell you, actually, the main challenge 01:08:33.600 --> 01:08:36.029 align:middle line:84% with agricultural insurance is, I 01:08:36.029 --> 01:08:38.670 align:middle line:84% don't want to write-- because of these moral-hazard issues, 01:08:38.670 --> 01:08:41.439 align:middle line:84% I can't insure your output, otherwise you don't work. 01:08:41.439 --> 01:08:43.680 align:middle line:84% So what I do is, I write insurance 01:08:43.680 --> 01:08:46.380 align:middle line:84% based on your predicted output, which is essentially, 01:08:46.380 --> 01:08:48.550 align:middle line:84% I write insurance based on the weather, for example. 01:08:48.550 --> 01:08:49.633 align:middle line:90% So actually, you can buy-- 01:08:49.633 --> 01:08:53.020 align:middle line:84% in the US or in other places, you can buy weather insurance, 01:08:53.020 --> 01:08:53.520 align:middle line:90% essentially. 01:08:53.520 --> 01:08:54.930 align:middle line:84% So if it's a drought-- you can buy insurance 01:08:54.930 --> 01:08:55.680 align:middle line:90% against a drought. 01:08:55.680 --> 01:08:58.262 align:middle line:84% That's an exogenous event that's going to reduce your output. 01:08:58.262 --> 01:09:00.220 align:middle line:84% So we can ensure you, to some degree, that way. 01:09:00.220 --> 01:09:02.609 align:middle line:84% The problem with those insurance contracts 01:09:02.609 --> 01:09:06.630 align:middle line:84% is the thing we're insuring is not 01:09:06.630 --> 01:09:08.319 align:middle line:84% perfectly correlated with your shock. 01:09:08.319 --> 01:09:10.710 align:middle line:84% And so they don't provide as good insurance 01:09:10.710 --> 01:09:11.550 align:middle line:90% as you might like. 01:09:11.550 --> 01:09:14.399 align:middle line:84% And that issue is what's called basis risk, 01:09:14.399 --> 01:09:16.439 align:middle line:84% that the thing we're insuring is not perfectly 01:09:16.439 --> 01:09:17.819 align:middle line:90% correlated with your shocks. 01:09:17.819 --> 01:09:19.319 align:middle line:84% And so if you're interested in that, 01:09:19.319 --> 01:09:21.736 align:middle line:84% there are some papers that have tried to think about that. 01:09:21.736 --> 01:09:26.819 align:middle line:84% But if you thought that the issue was risk aversion, then 01:09:26.819 --> 01:09:29.439 align:middle line:84% you would be heading in this insurance direction. 01:09:29.439 --> 01:09:31.439 align:middle line:84% And if you got the issue with limited liability, 01:09:31.439 --> 01:09:33.314 align:middle line:84% then you might head in a different direction, 01:09:33.314 --> 01:09:35.700 align:middle line:84% of can we basically make some transfers to people so that 01:09:35.700 --> 01:09:36.210 align:middle line:90% they're-- 01:09:36.210 --> 01:09:39.060 align:middle line:84% don't have this limited-liability constraint. 01:09:39.060 --> 01:09:41.939 align:middle line:84% Or can we write-- can we help them facilitate them writing-- 01:09:41.939 --> 01:09:44.606 align:middle line:84% we might be much-- if we thought limited liability was an issue, 01:09:44.606 --> 01:09:47.370 align:middle line:84% we might be much more interested in writing long-term things 01:09:47.370 --> 01:09:48.330 align:middle line:84% that look more like debt contracts, 01:09:48.330 --> 01:09:50.880 align:middle line:84% for example, and facilitating that thing, which we might not 01:09:50.880 --> 01:09:53.970 align:middle line:84% be interested if other things are important. 01:09:53.970 --> 01:09:55.263 align:middle line:90% Yeah. 01:09:55.263 --> 01:09:55.930 align:middle line:90% Oh, [INAUDIBLE]. 01:09:55.930 --> 01:09:56.430 align:middle line:90% I'm sorry. 01:09:56.430 --> 01:09:57.090 align:middle line:90% Go ahead. 01:09:57.090 --> 01:10:00.970 align:middle line:84% AUDIENCE: I was wondering, is there a benefit to-- 01:10:00.970 --> 01:10:05.170 align:middle line:84% if you wanted to do some staggered levels 01:10:05.170 --> 01:10:10.540 align:middle line:84% and incentives, from getting a panel, if that was possible, 01:10:10.540 --> 01:10:13.562 align:middle line:84% as opposed to a sequence of cross-- 01:10:13.562 --> 01:10:15.020 align:middle line:84% BEN OLKEN: Repeated cross sections. 01:10:15.020 --> 01:10:16.700 align:middle line:84% AUDIENCE: --cross sections, if you're 01:10:16.700 --> 01:10:20.090 align:middle line:84% trying to back out what certain types of workers might be, 01:10:20.090 --> 01:10:24.580 align:middle line:84% and how those elasticities might differ between groups, where, 01:10:24.580 --> 01:10:28.580 align:middle line:84% from the perspective of the-- for policy implications, does 01:10:28.580 --> 01:10:29.800 align:middle line:90% that not even matter? 01:10:29.800 --> 01:10:33.480 align:middle line:90% 01:10:33.480 --> 01:10:35.580 align:middle line:90% BEN OLKEN: So I don't think-- 01:10:35.580 --> 01:10:38.520 align:middle line:84% so you're saying, if we had a panel, 01:10:38.520 --> 01:10:41.350 align:middle line:84% and we get repeated observations of each. 01:10:41.350 --> 01:10:41.850 align:middle line:90% So-- 01:10:41.850 --> 01:10:43.975 align:middle line:84% AUDIENCE: You, changing your contract [INAUDIBLE].. 01:10:43.975 --> 01:10:47.580 align:middle line:84% BEN OLKEN: Yeah, so I think it would be-- 01:10:47.580 --> 01:10:51.450 align:middle line:84% yes, so I think if you basically found that everybody 01:10:51.450 --> 01:10:55.335 align:middle line:84% was the same, if there was no heterogeneity in people's 01:10:55.335 --> 01:10:57.460 align:middle line:84% responsiveness at different points in the contract, 01:10:57.460 --> 01:10:59.010 align:middle line:84% then you might be more willing to do 01:10:59.010 --> 01:11:02.340 align:middle line:84% some of those nonlinear, [INAUDIBLE] contracts. 01:11:02.340 --> 01:11:03.660 align:middle line:90% So, yes, in that sense, yes. 01:11:03.660 --> 01:11:06.268 align:middle line:84% But my strong suspicion, you would 01:11:06.268 --> 01:11:07.810 align:middle line:84% do that, you would find heterogeneity 01:11:07.810 --> 01:11:09.030 align:middle line:84% because there's heterogeneity in most things in life. 01:11:09.030 --> 01:11:11.890 align:middle line:84% And then-- so I'm not sure, as a practical matter, 01:11:11.890 --> 01:11:14.030 align:middle line:90% it would make a huge difference. 01:11:14.030 --> 01:11:14.650 align:middle line:90% Yeah, Sean. 01:11:14.650 --> 01:11:15.760 align:middle line:84% AUDIENCE: Is the right way to think about this paper 01:11:15.760 --> 01:11:17.230 align:middle line:84% that, in the framework of the model 01:11:17.230 --> 01:11:20.900 align:middle line:84% we just looked at, that is basically estimating the high c 01:11:20.900 --> 01:11:22.930 align:middle line:84% constraint, or that [INAUDIBLE] problem? 01:11:22.930 --> 01:11:23.780 align:middle line:90% BEN OLKEN: Mm-hmm. 01:11:23.780 --> 01:11:25.970 align:middle line:84% AUDIENCE: But then what is optimal would depend on 01:11:25.970 --> 01:11:27.010 align:middle line:90% a bunch other things-- 01:11:27.010 --> 01:11:29.410 align:middle line:90% BEN OLKEN: Yep, yep, yep, yes. 01:11:29.410 --> 01:11:30.170 align:middle line:90% Yes, I like that. 01:11:30.170 --> 01:11:30.670 align:middle line:90% Exactly. 01:11:30.670 --> 01:11:32.842 align:middle line:84% It's exactly saying, we're randomly 01:11:32.842 --> 01:11:34.300 align:middle line:84% varying in the contract and looking 01:11:34.300 --> 01:11:35.480 align:middle line:90% at how the agent responds. 01:11:35.480 --> 01:11:36.190 align:middle line:84% AUDIENCE: And it's just telling us 01:11:36.190 --> 01:11:37.815 align:middle line:84% that the fact that the agent responded, 01:11:37.815 --> 01:11:40.350 align:middle line:84% [INAUDIBLE] the [? ic ?] is [INAUDIBLE].. 01:11:40.350 --> 01:11:41.430 align:middle line:90% BEN OLKEN: Is what? 01:11:41.430 --> 01:11:42.390 align:middle line:84% AUDIENCE: It's telling us that that-- 01:11:42.390 --> 01:11:42.855 align:middle line:90% BEN OLKEN: It's important. 01:11:42.855 --> 01:11:43.550 align:middle line:90% AUDIENCE: It's important. 01:11:43.550 --> 01:11:44.050 align:middle line:90% OK. 01:11:44.050 --> 01:11:45.000 align:middle line:90% BEN OLKEN: Yes, yes. 01:11:45.000 --> 01:11:49.200 align:middle line:90% 01:11:49.200 --> 01:11:51.750 align:middle line:90% Other comments? 01:11:51.750 --> 01:11:52.420 align:middle line:90% Yeah, Ray. 01:11:52.420 --> 01:11:54.378 align:middle line:84% AUDIENCE: You knew there existed heterogeneity. 01:11:54.378 --> 01:11:57.400 align:middle line:84% But then the person who's giving the contract 01:11:57.400 --> 01:12:00.700 align:middle line:84% could just administer some tests to you, 01:12:00.700 --> 01:12:05.260 align:middle line:84% by varying your contract at the start of more 01:12:05.260 --> 01:12:06.860 align:middle line:90% sharecropping relationship. 01:12:06.860 --> 01:12:09.623 align:middle line:84% So then you will [? set ?] your time. 01:12:09.623 --> 01:12:10.290 align:middle line:90% BEN OLKEN: Yeah. 01:12:10.290 --> 01:12:11.640 align:middle line:84% AUDIENCE: And [? could ?] you have complicated 01:12:11.640 --> 01:12:12.270 align:middle line:90% [INAUDIBLE] contracts? 01:12:12.270 --> 01:12:12.900 align:middle line:90% BEN OLKEN: Oh. 01:12:12.900 --> 01:12:14.483 align:middle line:84% BEN OLKEN: Well, yes, except you would 01:12:14.483 --> 01:12:17.070 align:middle line:84% need to find incentive-compatible ways 01:12:17.070 --> 01:12:18.580 align:middle line:90% of getting that information. 01:12:18.580 --> 01:12:22.110 align:middle line:84% So I'm playing a game with my landlord 01:12:22.110 --> 01:12:23.820 align:middle line:84% and I know that the answers to my game 01:12:23.820 --> 01:12:27.510 align:middle line:84% are going to affect my contract, may affect how I play the game. 01:12:27.510 --> 01:12:30.390 align:middle line:90% 01:12:30.390 --> 01:12:31.290 align:middle line:90% But I do-- 01:12:31.290 --> 01:12:31.930 align:middle line:90% I think you're onto something. 01:12:31.930 --> 01:12:33.347 align:middle line:84% This question is like, can we find 01:12:33.347 --> 01:12:35.940 align:middle line:84% incentive-compatible mechanisms for getting 01:12:35.940 --> 01:12:36.857 align:middle line:90% like better contracts? 01:12:36.857 --> 01:12:38.190 align:middle line:90% I don't know the answer to that. 01:12:38.190 --> 01:12:38.940 align:middle line:90% But maybe, yes. 01:12:38.940 --> 01:12:39.940 align:middle line:90% I don't know the answer. 01:12:39.940 --> 01:12:44.300 align:middle line:84% AUDIENCE: I would imagine if it was long-term enough, 01:12:44.300 --> 01:12:47.300 align:middle line:90% [INAUDIBLE] that-- 01:12:47.300 --> 01:12:49.940 align:middle line:84% I don't think people don't [INAUDIBLE].. 01:12:49.940 --> 01:12:51.890 align:middle line:90% BEN OLKEN: Yeah, so let me-- 01:12:51.890 --> 01:12:54.898 align:middle line:84% I keep referring you to contract theory classes. 01:12:54.898 --> 01:12:56.690 align:middle line:84% No, in all seriousness, these questions are 01:12:56.690 --> 01:12:59.690 align:middle line:84% like, how do I design contracts and how do I 01:12:59.690 --> 01:13:03.530 align:middle line:84% get you to reveal information that are going to be-- 01:13:03.530 --> 01:13:05.825 align:middle line:84% how do I design an incentive-compatible mechanism 01:13:05.825 --> 01:13:08.450 align:middle line:84% that's going to have you reveal information that's going to let 01:13:08.450 --> 01:13:09.710 align:middle line:90% me give you a better contract? 01:13:09.710 --> 01:13:11.692 align:middle line:84% That-- thinking about these things 01:13:11.692 --> 01:13:14.150 align:middle line:84% these models in more-- if you're interested in these models 01:13:14.150 --> 01:13:16.910 align:middle line:84% in more detail, that's where we explore them. 01:13:16.910 --> 01:13:20.090 align:middle line:90% 01:13:20.090 --> 01:13:21.942 align:middle line:90% OK. 01:13:21.942 --> 01:13:23.900 align:middle line:84% And I'm not a contract theorist, although I did 01:13:23.900 --> 01:13:25.192 align:middle line:90% take that class many years ago. 01:13:25.192 --> 01:13:28.242 align:middle line:84% So I apologize for not having all of the answers 01:13:28.242 --> 01:13:29.200 align:middle line:90% off the top of my head. 01:13:29.200 --> 01:13:31.100 align:middle line:90% But I have some of them. 01:13:31.100 --> 01:13:35.430 align:middle line:90% 01:13:35.430 --> 01:13:39.850 align:middle line:84% All right, a few other things I wanted 01:13:39.850 --> 01:13:48.260 align:middle line:84% to say about the paper, first point is-- 01:13:48.260 --> 01:13:51.390 align:middle line:90% 01:13:51.390 --> 01:13:55.382 align:middle line:84% so this is the 75% contract versus the 50% contract. 01:13:55.382 --> 01:13:57.340 align:middle line:84% And nothing else really makes a big difference. 01:13:57.340 --> 01:13:59.970 align:middle line:84% So also, going back to, I think, Aaron, your comment of, 01:13:59.970 --> 01:14:01.610 align:middle line:90% is the risky thing-- 01:14:01.610 --> 01:14:05.490 align:middle line:84% it may not be exactly the same, but it's certainly not the case 01:14:05.490 --> 01:14:07.800 align:middle line:84% that this contract looks-- these results look 01:14:07.800 --> 01:14:10.770 align:middle line:90% anything like those results. 01:14:10.770 --> 01:14:12.750 align:middle line:84% Plus, also, you actually might imagine 01:14:12.750 --> 01:14:13.950 align:middle line:90% it would go the other way. 01:14:13.950 --> 01:14:16.770 align:middle line:90% I'm making the world-- 01:14:16.770 --> 01:14:19.718 align:middle line:84% I might be-- I'm making my self more risky-- 01:14:19.718 --> 01:14:21.510 align:middle line:84% you might actually think if anything-- this 01:14:21.510 --> 01:14:22.677 align:middle line:90% could be opposite direction. 01:14:22.677 --> 01:14:25.650 align:middle line:84% You actually might make me more conservative, in some sense. 01:14:25.650 --> 01:14:27.630 align:middle line:90% But anyway, it's certainly-- 01:14:27.630 --> 01:14:29.880 align:middle line:84% to your point, I agree with-- you're right. 01:14:29.880 --> 01:14:33.490 align:middle line:84% But I think also, it's a little bit-- 01:14:33.490 --> 01:14:35.490 align:middle line:84% not necessarily-- it may not be perfectly right. 01:14:35.490 --> 01:14:37.710 align:middle line:84% But to first order, it's not-- like I'd look at this 01:14:37.710 --> 01:14:39.752 align:middle line:84% and I'd say, oh, it doesn't look like it's really 01:14:39.752 --> 01:14:41.970 align:middle line:84% the riskiness that's resulting in these effects. 01:14:41.970 --> 01:14:43.421 align:middle line:90% They're just so different. 01:14:43.421 --> 01:14:47.110 align:middle line:90% 01:14:47.110 --> 01:14:49.660 align:middle line:84% Several of you picked up, in your response notes, 01:14:49.660 --> 01:14:52.310 align:middle line:84% this question of randomization inference. 01:14:52.310 --> 01:14:56.800 align:middle line:84% So how many of you have seen randomization inference before? 01:14:56.800 --> 01:14:57.610 align:middle line:90% Yes? 01:14:57.610 --> 01:14:58.400 align:middle line:90% All of you? 01:14:58.400 --> 01:14:58.900 align:middle line:90% Great. 01:14:58.900 --> 01:14:59.275 align:middle line:90% AUDIENCE: [INAUDIBLE] 01:14:59.275 --> 01:14:59.650 align:middle line:90% BEN OLKEN: What? 01:14:59.650 --> 01:14:59.970 align:middle line:90% Sorry? 01:14:59.970 --> 01:15:01.120 align:middle line:84% AUDIENCE: [INAUDIBLE] recitation [? order. ?] 01:15:01.120 --> 01:15:03.620 align:middle line:84% BEN OLKEN: Great, so I won't go into that-- any more detail. 01:15:03.620 --> 01:15:06.500 align:middle line:90% 01:15:06.500 --> 01:15:07.210 align:middle line:90% Great. 01:15:07.210 --> 01:15:09.240 align:middle line:90% Awesome. 01:15:09.240 --> 01:15:09.740 align:middle line:90% OK. 01:15:09.740 --> 01:15:19.820 align:middle line:90% 01:15:19.820 --> 01:15:26.600 align:middle line:84% You see some effects on the intensive margin of capital 01:15:26.600 --> 01:15:28.925 align:middle line:90% usage of various inputs. 01:15:28.925 --> 01:15:30.300 align:middle line:84% So it does look like some input-- 01:15:30.300 --> 01:15:31.355 align:middle line:90% so I do use more inputs. 01:15:31.355 --> 01:15:34.670 align:middle line:90% 01:15:34.670 --> 01:15:41.750 align:middle line:84% You also see some impacts on unpaid labor, 01:15:41.750 --> 01:15:42.860 align:middle line:90% like family labor. 01:15:42.860 --> 01:15:44.750 align:middle line:84% Some of you had questions of what this is. 01:15:44.750 --> 01:15:45.890 align:middle line:84% I think the idea is, we have lots 01:15:45.890 --> 01:15:48.470 align:middle line:84% of people in the household who also are working on the farm. 01:15:48.470 --> 01:15:50.000 align:middle line:90% And I'll talk about-- 01:15:50.000 --> 01:15:54.740 align:middle line:84% I'm going to talk a lot about unpaid versus paid 01:15:54.740 --> 01:15:57.583 align:middle line:90% labor in the labor section. 01:15:57.583 --> 01:15:59.000 align:middle line:84% So let me not answer the questions 01:15:59.000 --> 01:16:01.125 align:middle line:84% of what do we mean about paid and unpaid labor now. 01:16:01.125 --> 01:16:04.700 align:middle line:84% But let me defer that to the-- there's a paper by Dan Benjamin 01:16:04.700 --> 01:16:07.185 align:middle line:84% that I'm going to talk about, in a lot of detail, 01:16:07.185 --> 01:16:09.560 align:middle line:84% in the first labor lecture, which is going to think about 01:16:09.560 --> 01:16:11.150 align:middle line:84% this issue, like why do we have-- 01:16:11.150 --> 01:16:13.130 align:middle line:84% why is there a difference, actually, between-- 01:16:13.130 --> 01:16:15.360 align:middle line:84% why might this respond and this not respond? 01:16:15.360 --> 01:16:17.360 align:middle line:84% There's an enormous literature on that question. 01:16:17.360 --> 01:16:18.735 align:middle line:84% And I'm going to talk about that. 01:16:18.735 --> 01:16:21.230 align:middle line:84% I'm just going to push that over to the land lecture. 01:16:21.230 --> 01:16:23.240 align:middle line:90% But we do see that. 01:16:23.240 --> 01:16:26.930 align:middle line:84% And the third thing we see is that we see 01:16:26.930 --> 01:16:28.700 align:middle line:90% a change in the kind of crop. 01:16:28.700 --> 01:16:31.250 align:middle line:90% So we see maize goes up. 01:16:31.250 --> 01:16:32.415 align:middle line:90% And tomatoes go up. 01:16:32.415 --> 01:16:34.040 align:middle line:84% And I guess peanuts, I think, can go up 01:16:34.040 --> 01:16:35.290 align:middle line:90% in some of the specifications. 01:16:35.290 --> 01:16:38.570 align:middle line:84% And they argue that beans are a safer crop in this context. 01:16:38.570 --> 01:16:40.070 align:middle line:84% The other ones are more risky crops. 01:16:40.070 --> 01:16:42.560 align:middle line:84% And so I take these more high-risk, high-return kind 01:16:42.560 --> 01:16:46.100 align:middle line:90% of crops, which is not obvious. 01:16:46.100 --> 01:16:48.290 align:middle line:84% But one thing I just wanted to point out 01:16:48.290 --> 01:16:54.570 align:middle line:84% is that the moral-hazard model had 01:16:54.570 --> 01:16:57.000 align:middle line:84% this unobservable, or at least noncontractible 01:16:57.000 --> 01:16:59.890 align:middle line:90% effort thing going on. 01:16:59.890 --> 01:17:04.230 align:middle line:84% And I wanted to ask what you thought, in terms of, 01:17:04.230 --> 01:17:08.250 align:middle line:84% if you think about the results, is it 01:17:08.250 --> 01:17:10.080 align:middle line:84% contract-- is the thing is the thing that's 01:17:10.080 --> 01:17:12.608 align:middle line:84% changing contractible or noncontractible? 01:17:12.608 --> 01:17:22.100 align:middle line:90% 01:17:22.100 --> 01:17:24.485 align:middle line:84% You're trying to map the results to the theory. 01:17:24.485 --> 01:17:32.010 align:middle line:90% 01:17:32.010 --> 01:17:33.890 align:middle line:90% What do you think? 01:17:33.890 --> 01:17:34.670 align:middle line:90% Yeah, Sondé. 01:17:34.670 --> 01:17:37.760 align:middle line:84% AUDIENCE: I think they're both [INAUDIBLE].. 01:17:37.760 --> 01:17:38.930 align:middle line:90% BEN OLKEN: It seems like it. 01:17:38.930 --> 01:17:44.090 align:middle line:84% It seems kind of crop is pretty darn observable. 01:17:44.090 --> 01:17:46.730 align:middle line:84% Are you planting like corn or are you planting peanuts 01:17:46.730 --> 01:17:48.210 align:middle line:90% or are you planting beans? 01:17:48.210 --> 01:17:51.320 align:middle line:84% That's pretty easy to figure out. 01:17:51.320 --> 01:17:52.040 align:middle line:90% Yeah. 01:17:52.040 --> 01:17:54.630 align:middle line:84% How about some of the other stuff. 01:17:54.630 --> 01:17:56.880 align:middle line:84% AUDIENCE: Maybe the amount of fertilizer [INAUDIBLE].. 01:17:56.880 --> 01:17:58.547 align:middle line:84% BEN OLKEN: Yeah, maybe those are harder. 01:17:58.547 --> 01:18:00.710 align:middle line:84% So to me, in some sense, one of the things 01:18:00.710 --> 01:18:02.460 align:middle line:84% that I think is a little surprising, here, 01:18:02.460 --> 01:18:06.240 align:middle line:84% is that at least-- and they do this decomposition thing. 01:18:06.240 --> 01:18:10.950 align:middle line:90% And about half of it is-- 01:18:10.950 --> 01:18:13.050 align:middle line:90% actually, sorry, sorry. 01:18:13.050 --> 01:18:14.670 align:middle line:84% What's observable, what's unobervable, 01:18:14.670 --> 01:18:15.690 align:middle line:90% let me not answer that. 01:18:15.690 --> 01:18:18.305 align:middle line:84% But they say, what are the obvious-- the inputs 01:18:18.305 --> 01:18:20.430 align:middle line:84% the production function, like land, labor, capital, 01:18:20.430 --> 01:18:21.013 align:middle line:90% is about half. 01:18:21.013 --> 01:18:22.890 align:middle line:84% And then crop choice is the other half. 01:18:22.890 --> 01:18:24.900 align:middle line:84% To me, actually, crop choice seems 01:18:24.900 --> 01:18:28.000 align:middle line:84% kind of the most observable and contractible of all of this. 01:18:28.000 --> 01:18:31.142 align:middle line:84% So in some sense, that's a bit of a puzzle, actually. 01:18:31.142 --> 01:18:32.850 align:middle line:84% Why can't we solve some of these problems 01:18:32.850 --> 01:18:35.070 align:middle line:90% by contracting on crop choice? 01:18:35.070 --> 01:18:37.950 align:middle line:84% From the first model, I think it's an interesting question. 01:18:37.950 --> 01:18:43.380 align:middle line:90% 01:18:43.380 --> 01:18:44.880 align:middle line:90% Other comments? 01:18:44.880 --> 01:18:47.730 align:middle line:84% Any other comments on that paper before I move on? 01:18:47.730 --> 01:18:48.926 align:middle line:90% Yeah, Davey? 01:18:48.926 --> 01:18:50.550 align:middle line:84% AUDIENCE: It's because [INAUDIBLE].. 01:18:50.550 --> 01:18:50.730 align:middle line:90% BEN OLKEN: What? 01:18:50.730 --> 01:18:51.570 align:middle line:90% Sorry? 01:18:51.570 --> 01:18:53.370 align:middle line:84% AUDIENCE: Just because to observe it. 01:18:53.370 --> 01:18:55.470 align:middle line:84% So all of this can be observed in some ways, 01:18:55.470 --> 01:18:56.700 align:middle line:90% but just the [? customers. ?] 01:18:56.700 --> 01:18:57.240 align:middle line:90% BEN OLKEN: Well, I don't know. 01:18:57.240 --> 01:18:59.240 align:middle line:84% These sharecropping contracts, sometimes they're 01:18:59.240 --> 01:19:04.140 align:middle line:84% literally paid in kind, literally-- certainly, 01:19:04.140 --> 01:19:06.060 align:middle line:84% I've seen examples where you literally 01:19:06.060 --> 01:19:09.527 align:middle line:90% deliver x bushels of rice. 01:19:09.527 --> 01:19:11.610 align:middle line:84% And then it's pretty observable what the crop was. 01:19:11.610 --> 01:19:14.222 align:middle line:90% 01:19:14.222 --> 01:19:15.680 align:middle line:84% I think in general, though, we want 01:19:15.680 --> 01:19:18.263 align:middle line:84% to monitor what the output is in order to figure out the half. 01:19:18.263 --> 01:19:19.820 align:middle line:84% So I think that they got permission. 01:19:19.820 --> 01:19:20.630 align:middle line:90% Yeah, Kaj. 01:19:20.630 --> 01:19:22.130 align:middle line:84% AUDIENCE: I think it would have been 01:19:22.130 --> 01:19:24.493 align:middle line:84% nice to measure his preferences just to [INAUDIBLE].. 01:19:24.493 --> 01:19:26.660 align:middle line:84% BEN OLKEN: I think that would have been interesting. 01:19:26.660 --> 01:19:29.007 align:middle line:90% In some sense, the paper-- 01:19:29.007 --> 01:19:30.590 align:middle line:84% going to the previous point, the paper 01:19:30.590 --> 01:19:33.547 align:middle line:84% is really about the responsiveness to the contract. 01:19:33.547 --> 01:19:35.630 align:middle line:84% It's about the ic constraint, which doesn't really 01:19:35.630 --> 01:19:38.130 align:middle line:84% depend on your risk preferences as much. 01:19:38.130 --> 01:19:38.630 align:middle line:90% Right? 01:19:38.630 --> 01:19:42.200 align:middle line:84% So if it was about-- if this was a paper about ex-ante, what 01:19:42.200 --> 01:19:44.840 align:middle line:84% would be the optimal contract, then for sure, 01:19:44.840 --> 01:19:48.020 align:middle line:90% that seems really important. 01:19:48.020 --> 01:19:50.120 align:middle line:84% It's not obvious-- it's not as obvious to me 01:19:50.120 --> 01:19:52.190 align:middle line:84% how important that is, given this is about how 01:19:52.190 --> 01:19:53.190 align:middle line:90% the agent is responding. 01:19:53.190 --> 01:19:54.200 align:middle line:90% You see what I'm saying? 01:19:54.200 --> 01:19:54.620 align:middle line:90% Yeah? 01:19:54.620 --> 01:19:56.745 align:middle line:84% AUDIENCE: Would it be the agent that's risk averse? 01:19:56.745 --> 01:19:58.750 align:middle line:84% What if they try to allocate [INAUDIBLE]?? 01:19:58.750 --> 01:20:00.130 align:middle line:90% BEN OLKEN: The agent [? lied? ?] 01:20:00.130 --> 01:20:01.797 align:middle line:84% AUDIENCE: Like the agent is the problem. 01:20:01.797 --> 01:20:02.890 align:middle line:90% BEN OLKEN: Yeah? 01:20:02.890 --> 01:20:05.560 align:middle line:84% It's relevant vis-a-vis they're doing it the other way around, 01:20:05.560 --> 01:20:08.823 align:middle line:84% by making the contract, obviously, more risky. 01:20:08.823 --> 01:20:09.490 align:middle line:90% So you're right. 01:20:09.490 --> 01:20:11.590 align:middle line:84% Yes, this contract does change the amount of risk 01:20:11.590 --> 01:20:12.760 align:middle line:90% they're exposed to. 01:20:12.760 --> 01:20:15.040 align:middle line:90% But-- so I do-- 01:20:15.040 --> 01:20:16.430 align:middle line:90% it is kind of relevant. 01:20:16.430 --> 01:20:18.888 align:middle line:84% But if we think back to why did we care about risk aversion 01:20:18.888 --> 01:20:20.888 align:middle line:84% so much in all the models we went through today, 01:20:20.888 --> 01:20:22.900 align:middle line:84% it was actually-- it was more of vis-a-vis the-- 01:20:22.900 --> 01:20:25.310 align:middle line:84% ex-ante contracting choice, not about the ic constraint. 01:20:25.310 --> 01:20:28.210 align:middle line:84% And I liked whoever framed as a saying, 01:20:28.210 --> 01:20:29.990 align:middle line:84% this is a paper about the ic constraint. 01:20:29.990 --> 01:20:31.240 align:middle line:90% And I think that that's right. 01:20:31.240 --> 01:20:37.730 align:middle line:84% And risk aversion, yes, it does matter there, but much less. 01:20:37.730 --> 01:20:38.660 align:middle line:90% OK, other questions? 01:20:38.660 --> 01:20:41.090 align:middle line:84% Then we're basically out of time. 01:20:41.090 --> 01:20:42.620 align:middle line:90% OK, so we'll stop here. 01:20:42.620 --> 01:20:44.690 align:middle line:84% And I'll pick up, on Monday, with the second half 01:20:44.690 --> 01:20:47.810 align:middle line:84% of this stuff, which is all going to be talking about-- 01:20:47.810 --> 01:20:51.170 align:middle line:84% primarily going to talk about land tenure. 01:20:51.170 --> 01:20:54.800 align:middle line:84% And I'll talk a little bit about some of the land-reform papers, 01:20:54.800 --> 01:20:55.550 align:middle line:90% just very briefly. 01:20:55.550 --> 01:20:57.470 align:middle line:84% And then mostly, I'm going to talk about security, property 01:20:57.470 --> 01:20:59.595 align:middle line:84% rights, and investment decisions, things like that. 01:20:59.595 --> 01:21:02.440 align:middle line:90% OK, I'll see you all on Monday. 01:21:02.440 --> 01:21:21.000 align:middle line:90%