WEBVTT 00:00:00.000 --> 00:00:01.500 align:middle line:90% [SQUEAKING] 00:00:01.500 --> 00:00:04.000 align:middle line:90% [RUSTLING] 00:00:04.000 --> 00:00:05.500 align:middle line:90% [CLICKING] 00:00:05.500 --> 00:00:10.600 align:middle line:90% 00:00:10.600 --> 00:00:13.920 align:middle line:84% PROFESSOR: So what we're going to do today, chapter 21, 00:00:13.920 --> 00:00:17.162 align:middle line:84% is talking about how taxation affects labor supply. 00:00:17.162 --> 00:00:19.120 align:middle line:84% So we're going to talk in the next two lectures 00:00:19.120 --> 00:00:23.460 align:middle line:84% in particular about how taxation affects two behaviors. 00:00:23.460 --> 00:00:25.840 align:middle line:84% One is labor supply, and one is savings. 00:00:25.840 --> 00:00:27.920 align:middle line:90% So we'll just start. 00:00:27.920 --> 00:00:29.820 align:middle line:84% The base of taxation labor supply 00:00:29.820 --> 00:00:31.680 align:middle line:90% we've kind of done before. 00:00:31.680 --> 00:00:36.180 align:middle line:84% It's really just another version of what we did for welfare. 00:00:36.180 --> 00:00:39.380 align:middle line:84% But let's just remind ourselves of how it works. 00:00:39.380 --> 00:00:42.240 align:middle line:84% You've got a standard consumption leisure trade off. 00:00:42.240 --> 00:00:42.980 align:middle line:90% Here's leisure. 00:00:42.980 --> 00:00:44.860 align:middle line:90% Here's consumption. 00:00:44.860 --> 00:00:47.000 align:middle line:84% We've got a standard consumption leisure trade off. 00:00:47.000 --> 00:00:49.840 align:middle line:84% And imagine you have an individual whose wage rate is 00:00:49.840 --> 00:00:52.960 align:middle line:84% $15 an hour, during the Massachusetts minimum 00:00:52.960 --> 00:00:54.560 align:middle line:90% wage of $15 an hour. 00:00:54.560 --> 00:00:56.800 align:middle line:84% Then basically they have some trade 00:00:56.800 --> 00:01:00.160 align:middle line:84% off between consumption and leisure, 00:01:00.160 --> 00:01:03.680 align:middle line:84% where the slope of that line is minus 15. 00:01:03.680 --> 00:01:06.720 align:middle line:84% That is the rate at which they can trade off 00:01:06.720 --> 00:01:09.880 align:middle line:90% consumption and leisure. 00:01:09.880 --> 00:01:14.760 align:middle line:84% Alternatively, importantly, it is the price of leisure. 00:01:14.760 --> 00:01:16.880 align:middle line:90% Here's the key thing. 00:01:16.880 --> 00:01:18.960 align:middle line:90% We talked about this in 14.01. 00:01:18.960 --> 00:01:22.140 align:middle line:84% In economics, we don't know how to measure bads. 00:01:22.140 --> 00:01:23.420 align:middle line:90% We only know how to measure-- 00:01:23.420 --> 00:01:25.607 align:middle line:84% we don't know how to do optimization with bads. 00:01:25.607 --> 00:01:26.940 align:middle line:90% We have optimization with goods. 00:01:26.940 --> 00:01:28.607 align:middle line:84% So if you're choosing between two goods, 00:01:28.607 --> 00:01:32.140 align:middle line:84% I know how to do the optimization. 00:01:32.140 --> 00:01:33.920 align:middle line:84% If it's a good and a bad, it gets harder. 00:01:33.920 --> 00:01:35.840 align:middle line:84% So we do to come back to our basic framework 00:01:35.840 --> 00:01:37.960 align:middle line:84% as we turn all the bads into goods. 00:01:37.960 --> 00:01:40.600 align:middle line:90% Labor supply is a bad. 00:01:40.600 --> 00:01:44.700 align:middle line:84% Remember, regular people in the real world don't like working. 00:01:44.700 --> 00:01:47.160 align:middle line:84% It's something MIT students have a hard time understanding. 00:01:47.160 --> 00:01:50.900 align:middle line:84% But actually, leisure is a normal good. 00:01:50.900 --> 00:01:53.060 align:middle line:84% People would rather not work than work. 00:01:53.060 --> 00:01:55.080 align:middle line:90% That's our standard assumption. 00:01:55.080 --> 00:01:58.220 align:middle line:84% As a result, labor is a bad, and leisure is good. 00:01:58.220 --> 00:02:02.190 align:middle line:84% So what we do is we model consumption and leisure, 00:02:02.190 --> 00:02:08.590 align:middle line:84% and then we just create labor as the converse of leisure. 00:02:08.590 --> 00:02:12.350 align:middle line:84% So whenever we're talking about anything affecting labor 00:02:12.350 --> 00:02:14.710 align:middle line:84% supply, remember, the way we think about it is, first, 00:02:14.710 --> 00:02:16.190 align:middle line:90% how does it affect leisure? 00:02:16.190 --> 00:02:19.070 align:middle line:84% And then labor supply you solve as a residual. 00:02:19.070 --> 00:02:21.630 align:middle line:84% We're not modeling labor, modeling leisure. 00:02:21.630 --> 00:02:24.790 align:middle line:84% So essentially now we have a well-defined trade off 00:02:24.790 --> 00:02:26.730 align:middle line:84% of the kind we've been doing since 14.01, 00:02:26.730 --> 00:02:30.110 align:middle line:84% lecture 2, which is the trade off between two goods, 00:02:30.110 --> 00:02:32.790 align:middle line:84% consumption and leisure, and that the price at which you 00:02:32.790 --> 00:02:35.070 align:middle line:84% trade those off is the price of leisure. 00:02:35.070 --> 00:02:39.470 align:middle line:84% The price of leisure is the opportunity cost of-- 00:02:39.470 --> 00:02:42.310 align:middle line:84% the opportunity cost of leisure is basically 00:02:42.310 --> 00:02:46.510 align:middle line:84% that you could be working, earning $15 an hour. 00:02:46.510 --> 00:02:50.390 align:middle line:84% So you have some indifference curve. 00:02:50.390 --> 00:02:55.030 align:middle line:84% And there's some tangency at point A. In the book, 00:02:55.030 --> 00:03:00.670 align:middle line:84% we say, let's say, point A happens at 900 hours of leisure. 00:03:00.670 --> 00:03:03.470 align:middle line:84% And at 900 hours of leisure, that means you're working-- 00:03:03.470 --> 00:03:07.457 align:middle line:84% and let's say there's a total of 2,000 hours possible. 00:03:07.457 --> 00:03:09.790 align:middle line:84% Valerie, once again, can you send me an email that's not 00:03:09.790 --> 00:03:12.950 align:middle line:90% labeled in this graph, 21-1? 00:03:12.950 --> 00:03:14.170 align:middle line:90% I've got to add the x-- 00:03:14.170 --> 00:03:17.670 align:middle line:90% I got to add the x-axis labels. 00:03:17.670 --> 00:03:20.603 align:middle line:84% So you have 2,000 hours possible of leisure. 00:03:20.603 --> 00:03:23.270 align:middle line:84% You're going to take 900 hours, which means you're going to work 00:03:23.270 --> 00:03:24.870 align:middle line:90% 1,100 hours. 00:03:24.870 --> 00:03:30.670 align:middle line:84% If you work 1,100 hours, you make 16,500. 00:03:30.670 --> 00:03:33.430 align:middle line:84% So that is your initial allocation. 00:03:33.430 --> 00:03:36.950 align:middle line:84% You take 900 hours of leisure, which means you work 1,100 00:03:36.950 --> 00:03:42.970 align:middle line:90% hours, and you consume $16,500. 00:03:42.970 --> 00:03:44.430 align:middle line:90% That's your initial choice. 00:03:44.430 --> 00:03:50.550 align:middle line:84% Now, let's say the government rolls in and taxes your wages. 00:03:50.550 --> 00:03:52.890 align:middle line:84% Let's say the government particular rolls in and says, 00:03:52.890 --> 00:03:57.540 align:middle line:84% we are going to have a 33% tax on earnings. 00:03:57.540 --> 00:03:58.860 align:middle line:90% What does that mean? 00:03:58.860 --> 00:04:03.940 align:middle line:84% That means that the price of leisure has now fallen. 00:04:03.940 --> 00:04:05.480 align:middle line:84% The price of leisure has now fallen. 00:04:05.480 --> 00:04:07.160 align:middle line:84% If you take 2,000 hours of leisure, 00:04:07.160 --> 00:04:10.140 align:middle line:84% nothing's changed because you're not earning anything. 00:04:10.140 --> 00:04:13.700 align:middle line:84% But once you take less than 2,000 hours of leisure, 00:04:13.700 --> 00:04:15.880 align:middle line:84% it costs less to take that hour of leisure. 00:04:15.880 --> 00:04:19.060 align:middle line:84% It used to be an opportunity cost of $15 an hour 00:04:19.060 --> 00:04:20.640 align:middle line:90% to take that hour of leisure. 00:04:20.640 --> 00:04:25.320 align:middle line:84% Now, the opportunity cost is only $10.50. 00:04:25.320 --> 00:04:31.200 align:middle line:84% The slope of this was 15, minus 15 was the slope, minus 15. 00:04:31.200 --> 00:04:37.140 align:middle line:84% The slope of this is minus 10.50 because the 33% tax. 00:04:37.140 --> 00:04:41.300 align:middle line:84% So we've basically lowered the opportunity cost of leisure. 00:04:41.300 --> 00:04:44.580 align:middle line:84% What does that do to your decision 00:04:44.580 --> 00:04:47.120 align:middle line:90% of how much leisure to take? 00:04:47.120 --> 00:04:47.620 align:middle line:90% Yeah? 00:04:47.620 --> 00:04:48.420 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:04:48.420 --> 00:04:49.000 align:middle line:90% PROFESSOR: What? 00:04:49.000 --> 00:04:49.780 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:04:49.780 --> 00:04:50.988 align:middle line:90% PROFESSOR: Take more leisure. 00:04:50.988 --> 00:04:51.822 align:middle line:90% Why? 00:04:51.822 --> 00:04:52.780 align:middle line:90% AUDIENCE: It's cheaper. 00:04:52.780 --> 00:04:53.780 align:middle line:90% PROFESSOR: It's cheaper. 00:04:53.780 --> 00:04:55.987 align:middle line:90% And we call that what? 00:04:55.987 --> 00:04:57.320 align:middle line:90% AUDIENCE: The sunk cost fallacy. 00:04:57.320 --> 00:04:58.420 align:middle line:90% PROFESSOR: No. 00:04:58.420 --> 00:04:59.180 align:middle line:90% No. 00:04:59.180 --> 00:05:01.620 align:middle line:90% Good try, though. 00:05:01.620 --> 00:05:02.600 align:middle line:90% What do we call that? 00:05:02.600 --> 00:05:03.820 align:middle line:90% That's-- yeah. 00:05:03.820 --> 00:05:05.120 align:middle line:90% That's the substitution effect. 00:05:05.120 --> 00:05:06.537 align:middle line:84% But remember, there's two effects. 00:05:06.537 --> 00:05:08.140 align:middle line:90% What's the other one? 00:05:08.140 --> 00:05:12.580 align:middle line:84% So the substitution effect says that we 00:05:12.580 --> 00:05:14.117 align:middle line:90% should take more leisure. 00:05:14.117 --> 00:05:16.200 align:middle line:84% The price has fallen, so you should do more of it. 00:05:16.200 --> 00:05:17.580 align:middle line:90% That's unambiguous. 00:05:17.580 --> 00:05:20.620 align:middle line:84% But there's also an income effect. 00:05:20.620 --> 00:05:23.820 align:middle line:90% You are now effectively poorer. 00:05:23.820 --> 00:05:26.520 align:middle line:84% For any amount of work you do, you have less money. 00:05:26.520 --> 00:05:28.620 align:middle line:90% You're now effectively poorer. 00:05:28.620 --> 00:05:31.360 align:middle line:84% So that income effect says, when you're poor, 00:05:31.360 --> 00:05:33.620 align:middle line:84% you want to buy less of everything. 00:05:33.620 --> 00:05:37.680 align:middle line:84% One of the things you want to buy less of is leisure. 00:05:37.680 --> 00:05:40.140 align:middle line:90% So that says take less leisure. 00:05:40.140 --> 00:05:42.040 align:middle line:84% So substitution and income effects, 00:05:42.040 --> 00:05:44.180 align:middle line:84% while they work together in consumption, 00:05:44.180 --> 00:05:46.260 align:middle line:84% go against each other with factor supply, 00:05:46.260 --> 00:05:48.860 align:middle line:90% like labor supply. 00:05:48.860 --> 00:05:54.195 align:middle line:84% So actually it's ambiguous what a cut in wages will do. 00:05:54.195 --> 00:05:55.070 align:middle line:90% Think of it this way. 00:05:55.070 --> 00:05:57.730 align:middle line:84% Here's the way I like to-- think of an extreme example. 00:05:57.730 --> 00:06:00.570 align:middle line:84% Imagine you have a fixed income target. 00:06:00.570 --> 00:06:02.750 align:middle line:84% You want to work enough to earn x. 00:06:02.750 --> 00:06:04.250 align:middle line:84% Well, in that case, it's unambiguous 00:06:04.250 --> 00:06:05.667 align:middle line:84% that cutting your wages will cause 00:06:05.667 --> 00:06:07.550 align:middle line:90% you to work more, not less. 00:06:07.550 --> 00:06:09.610 align:middle line:84% You got to work more to get to your target. 00:06:09.610 --> 00:06:11.090 align:middle line:90% That's an extreme case. 00:06:11.090 --> 00:06:13.290 align:middle line:84% But in general, you can see it's ambiguous 00:06:13.290 --> 00:06:14.610 align:middle line:90% which way this will go. 00:06:14.610 --> 00:06:17.530 align:middle line:84% You could end up taking more leisure or less leisure, 00:06:17.530 --> 00:06:19.530 align:middle line:84% and it depends on whether income or substitution 00:06:19.530 --> 00:06:21.610 align:middle line:90% effects dominate. 00:06:21.610 --> 00:06:23.783 align:middle line:84% Valerie, is [? Ahmed ?] on the way? 00:06:23.783 --> 00:06:26.510 align:middle line:84% VALERIE: [INAUDIBLE] I think five minutes later. 00:06:26.510 --> 00:06:28.530 align:middle line:90% PROFESSOR: OK. 00:06:28.530 --> 00:06:33.050 align:middle line:84% So basically, the bottom line is, in this simple model, 00:06:33.050 --> 00:06:38.170 align:middle line:84% it is ambiguous whether taxing wages will 00:06:38.170 --> 00:06:40.230 align:middle line:84% lead to more leisure or less leisure 00:06:40.230 --> 00:06:44.170 align:middle line:84% and, therefore, more labor supply or less labor supply. 00:06:44.170 --> 00:06:49.290 align:middle line:84% Now, of course, this is a very simple model 00:06:49.290 --> 00:06:50.430 align:middle line:90% of the labor market. 00:06:50.430 --> 00:06:52.690 align:middle line:84% There's a lot of complications we don't have here. 00:06:52.690 --> 00:06:56.483 align:middle line:84% For example, overtime creates a kink in the budget constraint. 00:06:56.483 --> 00:06:58.150 align:middle line:84% When you work more than a certain hours, 00:06:58.150 --> 00:06:59.618 align:middle line:90% your wage changes. 00:06:59.618 --> 00:07:01.410 align:middle line:84% There's the fact that you don't necessarily 00:07:01.410 --> 00:07:04.115 align:middle line:84% get to choose your number of hours freely, 00:07:04.115 --> 00:07:05.490 align:middle line:84% which means you don't necessarily 00:07:05.490 --> 00:07:07.830 align:middle line:84% get to pick the optimal point on this distribution. 00:07:07.830 --> 00:07:09.630 align:middle line:90% There's complications like that. 00:07:09.630 --> 00:07:11.675 align:middle line:84% But for now, in this simple model, 00:07:11.675 --> 00:07:14.050 align:middle line:84% we've got just the trade off, the income and substitution 00:07:14.050 --> 00:07:16.170 align:middle line:90% effect trade off. 00:07:16.170 --> 00:07:18.750 align:middle line:90% So which dominates? 00:07:18.750 --> 00:07:21.362 align:middle line:84% How do we decide whether income or substitution 00:07:21.362 --> 00:07:22.070 align:middle line:90% effects dominate? 00:07:22.070 --> 00:07:23.850 align:middle line:90% Well, we go to the data. 00:07:23.850 --> 00:07:25.990 align:middle line:84% And we go to the data, and we go to the data. 00:07:25.990 --> 00:07:29.410 align:middle line:84% And there's essentially been four different approaches 00:07:29.410 --> 00:07:34.250 align:middle line:84% to trying to measure, how do wages affect labor supply? 00:07:34.250 --> 00:07:37.830 align:middle line:84% The first approach is a cross-sectional regression. 00:07:37.830 --> 00:07:39.630 align:middle line:84% So this is the old-school approach, 00:07:39.630 --> 00:07:41.290 align:middle line:90% which literally just regress. 00:07:41.290 --> 00:07:43.650 align:middle line:84% If my wages go up, do I work more? 00:07:43.650 --> 00:07:46.250 align:middle line:84% Or if my wages are higher, do I work more? 00:07:46.250 --> 00:07:49.560 align:middle line:84% That is the standard chapter 3 non-comparability problem. 00:07:49.560 --> 00:07:53.877 align:middle line:84% High-wage people probably may work harder. 00:07:53.877 --> 00:07:56.460 align:middle line:84% Some with a high wage may work harder because they're intense. 00:07:56.460 --> 00:07:58.835 align:middle line:84% That's why they have a high wage, not because it actually 00:07:58.835 --> 00:08:00.240 align:middle line:90% costs them to work harder. 00:08:00.240 --> 00:08:02.540 align:middle line:84% So there's a standard unobserved variable problem, 00:08:02.540 --> 00:08:05.400 align:middle line:84% so the cross-sectional thing doesn't really work. 00:08:05.400 --> 00:08:08.640 align:middle line:84% There was actually a real randomized controlled trial 00:08:08.640 --> 00:08:12.480 align:middle line:84% of this in the 1970s, something called the negative income tax 00:08:12.480 --> 00:08:13.262 align:middle line:90% experiment. 00:08:13.262 --> 00:08:15.720 align:middle line:84% What they did is they actually literally ran an experiment. 00:08:15.720 --> 00:08:18.520 align:middle line:84% They literally assigned different tax rates to people. 00:08:18.520 --> 00:08:21.880 align:middle line:84% And they looked at what happened to their labor supply. 00:08:21.880 --> 00:08:23.880 align:middle line:90% So that's the second approach. 00:08:23.880 --> 00:08:28.580 align:middle line:84% There's quasi-experimental approaches, quasi experimental. 00:08:28.580 --> 00:08:30.420 align:middle line:84% And here, there's two famous approaches. 00:08:30.420 --> 00:08:32.760 align:middle line:84% One is a diff-in-diff approach of the type 00:08:32.760 --> 00:08:34.679 align:middle line:90% we discussed in chapter 3. 00:08:34.679 --> 00:08:36.480 align:middle line:84% So in particular, imagine you have 00:08:36.480 --> 00:08:39.640 align:middle line:84% a tax reform that changes tax rates for one group and not 00:08:39.640 --> 00:08:40.740 align:middle line:90% another group. 00:08:40.740 --> 00:08:43.240 align:middle line:84% You can look what happens to labor supply of the group whose 00:08:43.240 --> 00:08:46.500 align:middle line:84% tax rates went up versus the group whose tax rates didn't, so 00:08:46.500 --> 00:08:48.000 align:middle line:84% a standard difference-in-differences 00:08:48.000 --> 00:08:51.080 align:middle line:84% analysis, the kind you should be familiar with now. 00:08:51.080 --> 00:08:54.800 align:middle line:84% Another approach would be to do a regression-- would 00:08:54.800 --> 00:08:59.600 align:middle line:84% be to do what's called a bunching approach. 00:08:59.600 --> 00:09:02.320 align:middle line:84% So the bunching approach, so DD, and then the other, 00:09:02.320 --> 00:09:04.540 align:middle line:84% it's a new method we haven't talked much about. 00:09:04.540 --> 00:09:06.260 align:middle line:90% But it's gaining a lot of favor. 00:09:06.260 --> 00:09:08.680 align:middle line:84% It's been really popularized by the economist Raj Chetty 00:09:08.680 --> 00:09:10.120 align:middle line:90% at Harvard. 00:09:10.120 --> 00:09:13.320 align:middle line:84% And the idea of the bunching approach is the following. 00:09:13.320 --> 00:09:17.520 align:middle line:84% Imagine there's a kink in a budget constraint. 00:09:17.520 --> 00:09:21.420 align:middle line:90% Basically, if people are-- 00:09:21.420 --> 00:09:23.800 align:middle line:90% so think of the extreme case. 00:09:23.800 --> 00:09:25.960 align:middle line:90% Think of the window tax case. 00:09:25.960 --> 00:09:28.280 align:middle line:84% Remember, the window tax case where basically, once you 00:09:28.280 --> 00:09:29.905 align:middle line:84% went above a certain number of windows, 00:09:29.905 --> 00:09:31.960 align:middle line:84% there's a huge marginal tax rate for going up. 00:09:31.960 --> 00:09:33.640 align:middle line:84% And we showed a lot of people had right 00:09:33.640 --> 00:09:35.200 align:middle line:90% below that number of windows. 00:09:35.200 --> 00:09:39.680 align:middle line:84% That's a bunching estimate, which is, if people are elastic, 00:09:39.680 --> 00:09:43.120 align:middle line:84% you'll see them all bunching right before there's 00:09:43.120 --> 00:09:47.270 align:middle line:84% a disincentive coming ahead, right at the notch or the kink. 00:09:47.270 --> 00:09:50.510 align:middle line:84% And that's the idea of this bunching estimator 00:09:50.510 --> 00:09:54.750 align:middle line:84% is you can infer elasticities from how much people respond 00:09:54.750 --> 00:09:57.230 align:middle line:90% to kinks in the budget set. 00:09:57.230 --> 00:10:00.630 align:middle line:84% If they don't respond much, then they're not very elastic. 00:10:00.630 --> 00:10:02.772 align:middle line:90% If they respond a ton, they are. 00:10:02.772 --> 00:10:05.230 align:middle line:84% So basically, you could have imagined the number of windows 00:10:05.230 --> 00:10:06.450 align:middle line:84% wouldn't have changed because people are like, 00:10:06.450 --> 00:10:07.250 align:middle line:90% I want to breathe. 00:10:07.250 --> 00:10:08.788 align:middle line:84% I don't care about the window tax. 00:10:08.788 --> 00:10:10.330 align:middle line:84% That would say it's not very elastic. 00:10:10.330 --> 00:10:12.580 align:middle line:84% The fact that it was bunching of the number of windows 00:10:12.580 --> 00:10:14.150 align:middle line:84% right at that limit suggests people 00:10:14.150 --> 00:10:15.870 align:middle line:84% were willing to sacrifice some breathing 00:10:15.870 --> 00:10:18.110 align:middle line:90% to avoid paying the window tax. 00:10:18.110 --> 00:10:22.730 align:middle line:84% And indeed, there's-- I can't draw the graphs well enough, 00:10:22.730 --> 00:10:24.330 align:middle line:90% but in the book-- yeah, sorry. 00:10:24.330 --> 00:10:25.390 align:middle line:90% Go ahead. 00:10:25.390 --> 00:10:28.306 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:10:28.306 --> 00:10:31.070 align:middle line:90% 00:10:31.070 --> 00:10:33.070 align:middle line:84% PROFESSOR: OK, describe [? the ?] [? bunching ?] 00:10:33.070 --> 00:10:33.450 align:middle line:90% already. 00:10:33.450 --> 00:10:34.117 align:middle line:90% Excellent point. 00:10:34.117 --> 00:10:34.950 align:middle line:90% Great, great, point. 00:10:34.950 --> 00:10:36.630 align:middle line:84% And I almost made the very mistake here. 00:10:36.630 --> 00:10:38.950 align:middle line:84% The [INAUDIBLE] bunching is the following. 00:10:38.950 --> 00:10:42.050 align:middle line:84% An RD estimate, a Regression Discontinuity Estimate-- 00:10:42.050 --> 00:10:42.610 align:middle line:90% thank you. 00:10:42.610 --> 00:10:44.830 align:middle line:84% Let's just pass them, pass each pass each row here. 00:10:44.830 --> 00:10:47.190 align:middle line:84% A regression discontinuity estimate 00:10:47.190 --> 00:10:49.750 align:middle line:84% is an estimate where essentially you 00:10:49.750 --> 00:10:53.810 align:middle line:84% say, on either side of a cutoff, there's different incentives. 00:10:53.810 --> 00:10:56.430 align:middle line:84% And you look how-- and an RD estimate is basically 00:10:56.430 --> 00:10:58.033 align:middle line:90% sort of a difference estimate. 00:10:58.033 --> 00:10:59.950 align:middle line:84% You look and ask, do people behave differently 00:10:59.950 --> 00:11:01.390 align:middle line:90% on either side of a cutoff? 00:11:01.390 --> 00:11:02.770 align:middle line:90% Here it's just a cutoff. 00:11:02.770 --> 00:11:04.990 align:middle line:84% It's at the cutoff to behave differently 00:11:04.990 --> 00:11:07.150 align:middle line:90% than on either side of it. 00:11:07.150 --> 00:11:09.250 align:middle line:84% So basically, so now you have the handout. 00:11:09.250 --> 00:11:15.670 align:middle line:84% So if you look at figure 21.3, this 00:11:15.670 --> 00:11:18.690 align:middle line:84% shows Chetty's work around a tax increase. 00:11:18.690 --> 00:11:21.110 align:middle line:84% So what he did was he looked at Denmark, 00:11:21.110 --> 00:11:25.390 align:middle line:84% where there was a big tax increase around 267,000 kroners. 00:11:25.390 --> 00:11:27.490 align:middle line:90% There was not an increase. 00:11:27.490 --> 00:11:28.970 align:middle line:90% There was a big notch. 00:11:28.970 --> 00:11:31.870 align:middle line:84% Basically your tax jumped at that point. 00:11:31.870 --> 00:11:36.250 align:middle line:84% So basically, he's not doing an RD. 00:11:36.250 --> 00:11:38.792 align:middle line:84% This is not something we see dramatic things on either side. 00:11:38.792 --> 00:11:40.750 align:middle line:84% It's that you see a dramatic thing at one point 00:11:40.750 --> 00:11:42.570 align:middle line:84% relative to what's on either side. 00:11:42.570 --> 00:11:44.480 align:middle line:90% It's a break from smoothness. 00:11:44.480 --> 00:11:46.913 align:middle line:84% It's not a discontinuity in the series. 00:11:46.913 --> 00:11:48.580 align:middle line:84% It's a break from a smooth series, where 00:11:48.580 --> 00:11:50.540 align:middle line:90% it kind of gets non-monotonic. 00:11:50.540 --> 00:11:54.220 align:middle line:84% So basically you see here, this shows essentially 00:11:54.220 --> 00:11:59.980 align:middle line:84% how many people are at each income level in Denmark. 00:11:59.980 --> 00:12:04.940 align:middle line:84% And what you see is, at 267, there's a little bit of a peak. 00:12:04.940 --> 00:12:09.140 align:middle line:84% So in other words, it's a smoothly declining function. 00:12:09.140 --> 00:12:11.320 align:middle line:84% But then there's a little peak at 267, 00:12:11.320 --> 00:12:14.800 align:middle line:84% which suggests that people are responding by bunching there, 00:12:14.800 --> 00:12:17.020 align:middle line:90% but it's not very big. 00:12:17.020 --> 00:12:18.360 align:middle line:90% That's for employed workers. 00:12:18.360 --> 00:12:20.900 align:middle line:84% Now look at self-employed workers. 00:12:20.900 --> 00:12:24.100 align:middle line:84% Self-employed workers, there's a massive peak 00:12:24.100 --> 00:12:27.460 align:middle line:90% at exactly that point. 00:12:27.460 --> 00:12:31.620 align:middle line:84% Self-employed workers are responding very, very highly 00:12:31.620 --> 00:12:34.700 align:middle line:90% to this incentive. 00:12:34.700 --> 00:12:38.140 align:middle line:90% Employed workers are not. 00:12:38.140 --> 00:12:39.600 align:middle line:90% What's the difference? 00:12:39.600 --> 00:12:40.100 align:middle line:90% Why? 00:12:40.100 --> 00:12:42.965 align:middle line:84% Why do you think there's such more elastic behavior 00:12:42.965 --> 00:12:44.340 align:middle line:84% among these self-employed workers 00:12:44.340 --> 00:12:45.480 align:middle line:90% and the employed workers? 00:12:45.480 --> 00:12:48.260 align:middle line:90% 00:12:48.260 --> 00:12:49.723 align:middle line:90% Yeah? 00:12:49.723 --> 00:12:52.125 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:12:52.125 --> 00:12:54.500 align:middle line:84% PROFESSOR: Yeah, self-employed workers are unconstrained. 00:12:54.500 --> 00:12:55.958 align:middle line:84% They can work as hard as they want. 00:12:55.958 --> 00:12:57.700 align:middle line:90% They can also cheat more. 00:12:57.700 --> 00:12:59.860 align:middle line:84% But certainly, whether through cheating or actually 00:12:59.860 --> 00:13:02.280 align:middle line:84% changing the labor supply, they can set their own hours. 00:13:02.280 --> 00:13:03.480 align:middle line:90% Employed workers can't. 00:13:03.480 --> 00:13:07.260 align:middle line:84% It's hard to target I'm going to go exactly 267,000. 00:13:07.260 --> 00:13:08.940 align:middle line:84% So that's why and what Chetty found 00:13:08.940 --> 00:13:11.660 align:middle line:84% was much more elastic response among the self-employed 00:13:11.660 --> 00:13:13.620 align:middle line:90% than the employed. 00:13:13.620 --> 00:13:15.580 align:middle line:90% So you have all these methods. 00:13:15.580 --> 00:13:20.100 align:middle line:84% These methods actually deliver a fairly consistent answer, 00:13:20.100 --> 00:13:24.060 align:middle line:84% which is, if you look back in the old days, say, 00:13:24.060 --> 00:13:29.220 align:middle line:84% 1980s and before, you could think of the world as defined 00:13:29.220 --> 00:13:33.540 align:middle line:84% as primary earners, e.g. men, and secondary earners, e.g. 00:13:33.540 --> 00:13:36.900 align:middle line:84% women, especially married couples, especially with kids. 00:13:36.900 --> 00:13:43.890 align:middle line:84% In that world, what they found was primary earners, e.g. men, 00:13:43.890 --> 00:13:47.290 align:middle line:90% were not at all elastic-- 00:13:47.290 --> 00:13:51.170 align:middle line:84% basically there was very inelastic labor supply-- 00:13:51.170 --> 00:13:56.110 align:middle line:84% but that secondary earners, e.g. moms, were fairly elastic. 00:13:56.110 --> 00:13:59.690 align:middle line:90% Elasticity is between 0.5 and 1. 00:13:59.690 --> 00:14:03.970 align:middle line:84% The argument for why comes back to early substitution effects. 00:14:03.970 --> 00:14:07.910 align:middle line:84% If you're a guy in 1975 and you don't work, 00:14:07.910 --> 00:14:10.090 align:middle line:90% there's nothing else to do. 00:14:10.090 --> 00:14:11.450 align:middle line:90% None of your friends are home. 00:14:11.450 --> 00:14:12.970 align:middle line:90% They're all working. 00:14:12.970 --> 00:14:15.597 align:middle line:84% Where if you're a married woman in 1975, you have a choice. 00:14:15.597 --> 00:14:16.430 align:middle line:90% Some women are home. 00:14:16.430 --> 00:14:17.590 align:middle line:90% Some women are working. 00:14:17.590 --> 00:14:19.243 align:middle line:84% So it's more plausible to be elastic. 00:14:19.243 --> 00:14:21.410 align:middle line:84% You wouldn't know what to do with yourself if you're 00:14:21.410 --> 00:14:22.590 align:middle line:90% a guy at home in 1975. 00:14:22.590 --> 00:14:24.150 align:middle line:90% There was no internet. 00:14:24.150 --> 00:14:25.470 align:middle line:90% None of your friends were home. 00:14:25.470 --> 00:14:29.370 align:middle line:84% There's only so much golf you can play, basically, 00:14:29.370 --> 00:14:31.248 align:middle line:84% whereas, for women, there was a choice. 00:14:31.248 --> 00:14:32.790 align:middle line:84% They could stay home, raise the kids. 00:14:32.790 --> 00:14:34.582 align:middle line:84% Some of their friends were home, et cetera. 00:14:34.582 --> 00:14:35.970 align:middle line:90% So it makes sense. 00:14:35.970 --> 00:14:37.370 align:middle line:90% But here's what's fascinating. 00:14:37.370 --> 00:14:41.562 align:middle line:84% Over time, we are losing the distinction 00:14:41.562 --> 00:14:43.770 align:middle line:84% in primary and secondary workers as women become more 00:14:43.770 --> 00:14:45.610 align:middle line:90% integrated in the labor force. 00:14:45.610 --> 00:14:47.730 align:middle line:90% We're losing that distinction. 00:14:47.730 --> 00:14:51.250 align:middle line:84% What's happening is men's labor supply remains inelastic, 00:14:51.250 --> 00:14:54.847 align:middle line:84% and women's labor supply is becoming more inelastic. 00:14:54.847 --> 00:14:56.930 align:middle line:84% So overall labor supply is becoming just much more 00:14:56.930 --> 00:14:59.370 align:middle line:84% inelastic, where everybody works. 00:14:59.370 --> 00:15:00.270 align:middle line:90% OK, yeah? 00:15:00.270 --> 00:15:01.410 align:middle line:84% AUDIENCE: Is there no [? significant ?] [? apart ?] 00:15:01.410 --> 00:15:02.667 align:middle line:90% from the men's side as well? 00:15:02.667 --> 00:15:03.250 align:middle line:90% PROFESSOR: No. 00:15:03.250 --> 00:15:05.250 align:middle line:90% This is what's fascinating. 00:15:05.250 --> 00:15:08.530 align:middle line:84% So what's happening is men now just work and women just work, 00:15:08.530 --> 00:15:10.290 align:middle line:90% and kids are in childcare. 00:15:10.290 --> 00:15:13.503 align:middle line:84% So the difference is, it's not that the burden of home care 00:15:13.503 --> 00:15:15.170 align:middle line:84% has shifted, where men are more flexible 00:15:15.170 --> 00:15:15.910 align:middle line:90% and they might stay home. 00:15:15.910 --> 00:15:17.350 align:middle line:84% Now, obviously, I know a stay-at-home dad. 00:15:17.350 --> 00:15:19.517 align:middle line:84% There are more stay-at-home dads than there used to. 00:15:19.517 --> 00:15:22.788 align:middle line:84% But by and large, it hasn't really changed for men. 00:15:22.788 --> 00:15:24.330 align:middle line:84% It's women are becoming more like men 00:15:24.330 --> 00:15:25.788 align:middle line:84% in terms of their working behavior, 00:15:25.788 --> 00:15:27.371 align:middle line:84% in terms of more like primary earners. 00:15:27.371 --> 00:15:29.996 align:middle line:84% So we're getting to a much more inelastic world of labor supply 00:15:29.996 --> 00:15:30.530 align:middle line:90% overall. 00:15:30.530 --> 00:15:32.470 align:middle line:84% And I would say, roughly speaking, 00:15:32.470 --> 00:15:34.330 align:middle line:84% you can view the labor supply margin 00:15:34.330 --> 00:15:36.870 align:middle line:84% as a fairly inelastic one in the US economy. 00:15:36.870 --> 00:15:40.100 align:middle line:90% 00:15:40.100 --> 00:15:45.690 align:middle line:84% Now, of course, that is if we measure labor supply 00:15:45.690 --> 00:15:48.490 align:middle line:90% by things like hours worked. 00:15:48.490 --> 00:15:51.270 align:middle line:84% But in fact, that may be the wrong thing to measure it by. 00:15:51.270 --> 00:15:54.158 align:middle line:84% In fact, it may be, in the new economy, 00:15:54.158 --> 00:15:56.450 align:middle line:84% that what's more important than how many hours you work 00:15:56.450 --> 00:16:01.450 align:middle line:84% is like things like how hard you work and what type of job 00:16:01.450 --> 00:16:02.330 align:middle line:90% you have. 00:16:02.330 --> 00:16:04.930 align:middle line:84% And you could imagine, taxes, while they might not 00:16:04.930 --> 00:16:07.050 align:middle line:84% affect your hours of work, could very much 00:16:07.050 --> 00:16:10.410 align:middle line:84% affect things like how hard you work to get a promotion 00:16:10.410 --> 00:16:13.452 align:middle line:90% or what type of job you take. 00:16:13.452 --> 00:16:15.410 align:middle line:84% If taxes are high, maybe you don't care as much 00:16:15.410 --> 00:16:17.330 align:middle line:84% about taking that high-wage job, working hard 00:16:17.330 --> 00:16:19.830 align:middle line:84% to get that promotion and get that high-wage job. 00:16:19.830 --> 00:16:21.850 align:middle line:84% But if taxes are low, maybe you do. 00:16:21.850 --> 00:16:23.290 align:middle line:84% So that's ultimately what we might 00:16:23.290 --> 00:16:27.330 align:middle line:84% care about is not hours of work but total labor earnings. 00:16:27.330 --> 00:16:30.290 align:middle line:84% And this is a foreshadowing of chapter 25, when 00:16:30.290 --> 00:16:32.790 align:middle line:84% we'll ultimately conclude that, really, in some sense, 00:16:32.790 --> 00:16:36.210 align:middle line:84% it's total income that we care about, not these components. 00:16:36.210 --> 00:16:38.200 align:middle line:90% So we're building up to that. 00:16:38.200 --> 00:16:40.960 align:middle line:84% But nonetheless, at least on the easily measurable aspects 00:16:40.960 --> 00:16:43.080 align:middle line:84% of labor supply, it doesn't look like there's 00:16:43.080 --> 00:16:45.840 align:middle line:90% much of an effect of taxes. 00:16:45.840 --> 00:16:47.840 align:middle line:90% Questions about that? 00:16:47.840 --> 00:16:49.000 align:middle line:90% All right. yeah, Steven? 00:16:49.000 --> 00:16:50.080 align:middle line:84% AUDIENCE: Are there any other countries 00:16:50.080 --> 00:16:52.040 align:middle line:84% where the labor supply is noticeably more 00:16:52.040 --> 00:16:53.260 align:middle line:90% or less than others? 00:16:53.260 --> 00:16:54.433 align:middle line:90% Or is that not [INAUDIBLE] 00:16:54.433 --> 00:16:55.600 align:middle line:90% PROFESSOR: Oh, I don't know. 00:16:55.600 --> 00:16:56.840 align:middle line:90% I don't know. 00:16:56.840 --> 00:16:59.420 align:middle line:84% So that's the basics of tax and labor supply. 00:16:59.420 --> 00:17:00.772 align:middle line:90% Pretty straightforward. 00:17:00.772 --> 00:17:02.480 align:middle line:84% What I want to do for the rest of chapter 00:17:02.480 --> 00:17:06.760 align:middle line:84% 1 and 21 is apply that talking about two interesting policy 00:17:06.760 --> 00:17:07.960 align:middle line:90% issues. 00:17:07.960 --> 00:17:11.480 align:middle line:84% The first is something we mentioned before, which is the 00:17:11.480 --> 00:17:16.560 align:middle line:84% earned income tax credit, the EITC. 00:17:16.560 --> 00:17:19.760 align:middle line:84% The earned income tax credit, we mentioned this in chapter 17, 00:17:19.760 --> 00:17:23.480 align:middle line:84% it's a tax credit that you get that essentially 00:17:23.480 --> 00:17:24.980 align:middle line:90% acts as a wage subsidy. 00:17:24.980 --> 00:17:27.563 align:middle line:84% It's a tax credit you get where, over some range, 00:17:27.563 --> 00:17:29.980 align:middle line:84% the harder you work, the more you get from the government. 00:17:29.980 --> 00:17:31.300 align:middle line:90% It's a negative tax. 00:17:31.300 --> 00:17:32.800 align:middle line:90% It's a subsidy. 00:17:32.800 --> 00:17:35.340 align:middle line:90% So let's describe how it works. 00:17:35.340 --> 00:17:38.680 align:middle line:84% First of all, the ITC has grown tremendously over time. 00:17:38.680 --> 00:17:45.040 align:middle line:84% If you look at figure 21.4, you see 00:17:45.040 --> 00:17:50.560 align:middle line:84% that the EITC has gone from being 00:17:50.560 --> 00:17:55.800 align:middle line:84% very small to now being about $70 billion a year in spending. 00:17:55.800 --> 00:17:57.800 align:middle line:90% The ITC has grown a lot in time. 00:17:57.800 --> 00:17:59.220 align:middle line:90% How does the ITC work. 00:17:59.220 --> 00:18:02.720 align:middle line:84% Well, the best thing to do is look at figure 21.5. 00:18:02.720 --> 00:18:05.560 align:middle line:84% On figure 21.5 on the x-axis is how much 00:18:05.560 --> 00:18:07.020 align:middle line:90% earned income a family has. 00:18:07.020 --> 00:18:09.187 align:middle line:84% It's irrelevant for things like dividends and stuff. 00:18:09.187 --> 00:18:11.800 align:middle line:84% It's just about wages, how much earned income you have. 00:18:11.800 --> 00:18:15.000 align:middle line:84% On the y-axis, how big your ITC is. 00:18:15.000 --> 00:18:17.560 align:middle line:84% The way it works is, starting from 0, 00:18:17.560 --> 00:18:21.320 align:middle line:84% for the first roughly $15,000 of earnings, 00:18:21.320 --> 00:18:25.080 align:middle line:84% for every dollar you earn, you get an extra $0.40 from 00:18:25.080 --> 00:18:25.860 align:middle line:90% the government. 00:18:25.860 --> 00:18:28.520 align:middle line:90% So it's a negative tax. 00:18:28.520 --> 00:18:34.140 align:middle line:84% Until you reach the maximum amount of $5,920, then you stop. 00:18:34.140 --> 00:18:36.830 align:middle line:90% You just get a check for $5,920. 00:18:36.830 --> 00:18:42.630 align:middle line:84% But starting again at 19,350, they start to reduce that check 00:18:42.630 --> 00:18:47.430 align:middle line:84% by a somewhat different slope, by $0.21, so that by 47,440, 00:18:47.430 --> 00:18:49.910 align:middle line:84% you get nothing from the government. 00:18:49.910 --> 00:18:52.830 align:middle line:84% So what this is a government subsidy 00:18:52.830 --> 00:18:55.530 align:middle line:84% where the level of the subsidy depends on your earnings. 00:18:55.530 --> 00:18:57.270 align:middle line:84% Essentially, you can look at this chart, 00:18:57.270 --> 00:18:58.330 align:middle line:90% say, what's my earnings? 00:18:58.330 --> 00:19:00.247 align:middle line:84% And that will tell you how big the subsidy is. 00:19:00.247 --> 00:19:02.910 align:middle line:90% 00:19:02.910 --> 00:19:05.430 align:middle line:84% That's how the ITC works, is basically 00:19:05.430 --> 00:19:11.930 align:middle line:84% a subsidy that depends on what you earn. 00:19:11.930 --> 00:19:14.510 align:middle line:90% 00:19:14.510 --> 00:19:16.065 align:middle line:90% Now, remember let's step back. 00:19:16.065 --> 00:19:17.190 align:middle line:90% What's the idea of the ITC? 00:19:17.190 --> 00:19:18.607 align:middle line:84% The idea of the ITC is we're going 00:19:18.607 --> 00:19:20.010 align:middle line:90% to distribute to poor people. 00:19:20.010 --> 00:19:21.510 align:middle line:84% We're going to do so in a way to try 00:19:21.510 --> 00:19:24.343 align:middle line:84% to fix the leaky bucket by paying them to work. 00:19:24.343 --> 00:19:25.010 align:middle line:90% That's the idea. 00:19:25.010 --> 00:19:26.370 align:middle line:90% I should have said this first. 00:19:26.370 --> 00:19:28.950 align:middle line:84% The idea is we've got this leaky bucket problem, this iron 00:19:28.950 --> 00:19:32.290 align:middle line:84% law of redistribution, where we have the G and the T. 00:19:32.290 --> 00:19:34.500 align:middle line:84% And we can't solve all our problems. 00:19:34.500 --> 00:19:36.750 align:middle line:84% One way to solve it is by actually, instead of taxing, 00:19:36.750 --> 00:19:38.190 align:middle line:90% let's promote work. 00:19:38.190 --> 00:19:40.490 align:middle line:84% But actually does the ITC promote work? 00:19:40.490 --> 00:19:41.630 align:middle line:90% It's ambiguous. 00:19:41.630 --> 00:19:44.523 align:middle line:84% To see this, let's go to figure 21.6, 00:19:44.523 --> 00:19:45.690 align:middle line:90% which is a bit hard to read. 00:19:45.690 --> 00:19:46.730 align:middle line:84% I couldn't draw this very clearly. 00:19:46.730 --> 00:19:48.605 align:middle line:84% But it's actually really interesting in terms 00:19:48.605 --> 00:19:51.670 align:middle line:84% of how it helps think about our intuition. 00:19:51.670 --> 00:19:55.590 align:middle line:84% What the ITC does to work varies and depends what kind of work 00:19:55.590 --> 00:19:56.090 align:middle line:90% you are. 00:19:56.090 --> 00:19:58.810 align:middle line:84% What this does is it takes a standard labor supply, 00:19:58.810 --> 00:20:01.812 align:middle line:84% standard budget constraint-- assume a wage of $20 an hour. 00:20:01.812 --> 00:20:03.270 align:middle line:84% It's a standard consumption leisure 00:20:03.270 --> 00:20:05.630 align:middle line:84% diagram with a wage $20 an hour-- 00:20:05.630 --> 00:20:09.670 align:middle line:84% and overlays an ITC on top of it. 00:20:09.670 --> 00:20:12.270 align:middle line:84% So the new budget constraint, which 00:20:12.270 --> 00:20:15.830 align:middle line:84% is the blue plus the green, is the new budget constraint 00:20:15.830 --> 00:20:18.150 align:middle line:90% with the ITC in place. 00:20:18.150 --> 00:20:20.010 align:middle line:84% And we have four different types of people. 00:20:20.010 --> 00:20:22.990 align:middle line:84% And we can ask, what does the ITC do to the labor supply? 00:20:22.990 --> 00:20:28.150 align:middle line:84% Well, for person A, it's unambiguous. 00:20:28.150 --> 00:20:30.910 align:middle line:84% There is only a substitution effect. 00:20:30.910 --> 00:20:31.633 align:middle line:90% Why? 00:20:31.633 --> 00:20:34.300 align:middle line:84% Because they have no income, so there can't be an income effect. 00:20:34.300 --> 00:20:36.133 align:middle line:84% I'm just saying, I'll pay if you go to work. 00:20:36.133 --> 00:20:38.180 align:middle line:84% If you don't go to work, you don't get anything. 00:20:38.180 --> 00:20:42.140 align:middle line:84% So it's unambiguous that it will induce increased labor 00:20:42.140 --> 00:20:45.940 align:middle line:90% supply among those not working. 00:20:45.940 --> 00:20:50.500 align:middle line:84% Person B is someone who's working and at $5,000 is below 00:20:50.500 --> 00:20:51.560 align:middle line:90% the first notch. 00:20:51.560 --> 00:20:53.760 align:middle line:84% So if you flip back to the previous diagram, 00:20:53.760 --> 00:20:57.837 align:middle line:84% they're on the upward sloping part of the diagram. 00:20:57.837 --> 00:20:58.920 align:middle line:90% What's the effect on them? 00:20:58.920 --> 00:21:01.140 align:middle line:90% Well, actually it's ambiguous. 00:21:01.140 --> 00:21:02.640 align:middle line:84% There's income substitution effects. 00:21:02.640 --> 00:21:08.700 align:middle line:84% On the one hand, by providing a subsidy to work, 00:21:08.700 --> 00:21:10.580 align:middle line:84% I'm raising the price of leisure, which 00:21:10.580 --> 00:21:12.360 align:middle line:90% means I'll take less of it. 00:21:12.360 --> 00:21:13.860 align:middle line:84% On the other hand, I'm richer, which 00:21:13.860 --> 00:21:15.500 align:middle line:90% means I'll take more leisure. 00:21:15.500 --> 00:21:18.020 align:middle line:90% So the effect is ambiguous. 00:21:18.020 --> 00:21:19.900 align:middle line:90% What about person C? 00:21:19.900 --> 00:21:23.920 align:middle line:90% Person C earns $15,000. 00:21:23.920 --> 00:21:26.740 align:middle line:84% They're on the flat part, if you go to the previous diagram. 00:21:26.740 --> 00:21:31.060 align:middle line:84% How does it affect their labor supply, and why? 00:21:31.060 --> 00:21:33.600 align:middle line:84% How does the ITC affect someone like person C, and why? 00:21:33.600 --> 00:21:37.940 align:middle line:90% 00:21:37.940 --> 00:21:39.240 align:middle line:90% See the answer to the diagram. 00:21:39.240 --> 00:21:40.657 align:middle line:84% So you've just got to explain why. 00:21:40.657 --> 00:21:44.980 align:middle line:90% 00:21:44.980 --> 00:21:46.660 align:middle line:90% Come on guys, you got this. 00:21:46.660 --> 00:21:48.360 align:middle line:84% There's nothing I've been talking about. 00:21:48.360 --> 00:21:50.780 align:middle line:90% Yeah? 00:21:50.780 --> 00:21:52.400 align:middle line:90% It's a purely income effect. 00:21:52.400 --> 00:21:53.900 align:middle line:90% There's no substitution. 00:21:53.900 --> 00:21:57.102 align:middle line:84% If you look at that segment, if you look at that segment, 00:21:57.102 --> 00:21:59.060 align:middle line:84% it's parallel to the original budget contraint. 00:21:59.060 --> 00:22:00.260 align:middle line:90% It's a little hard to see. 00:22:00.260 --> 00:22:01.240 align:middle line:90% There's no substitution effect. 00:22:01.240 --> 00:22:02.112 align:middle line:90% They're just richer. 00:22:02.112 --> 00:22:04.820 align:middle line:84% If you make someone richer, they work less, so the income effect. 00:22:04.820 --> 00:22:07.760 align:middle line:84% So a person C unambiguously they'll work less. 00:22:07.760 --> 00:22:10.000 align:middle line:84% I've just given them a check for $5,000. 00:22:10.000 --> 00:22:11.940 align:middle line:90% They're going to work less. 00:22:11.940 --> 00:22:13.580 align:middle line:90% What about person D? 00:22:13.580 --> 00:22:16.260 align:middle line:84% Well, person D, there's now both an income and a substitution 00:22:16.260 --> 00:22:18.260 align:middle line:84% effect causing them to work less. 00:22:18.260 --> 00:22:19.860 align:middle line:84% The substitution effect is now they're 00:22:19.860 --> 00:22:22.180 align:middle line:84% on the downward sloping parts, so they're taxed 00:22:22.180 --> 00:22:24.555 align:middle line:90% at $0.21 per dollar they earn. 00:22:24.555 --> 00:22:26.680 align:middle line:84% The income effect is they're still getting a check, 00:22:26.680 --> 00:22:27.900 align:middle line:84% so still richer than they would be if they 00:22:27.900 --> 00:22:29.210 align:middle line:90% weren't getting the check. 00:22:29.210 --> 00:22:32.810 align:middle line:84% So that will also cause them to work less. 00:22:32.810 --> 00:22:34.030 align:middle line:90% So look at this diagram. 00:22:34.030 --> 00:22:35.530 align:middle line:84% First of all, questions about what's 00:22:35.530 --> 00:22:37.930 align:middle line:90% happening with A, B, C and D? 00:22:37.930 --> 00:22:39.475 align:middle line:90% Look at this diagram. 00:22:39.475 --> 00:22:40.850 align:middle line:84% It doesn't necessarily seem clear 00:22:40.850 --> 00:22:42.850 align:middle line:84% that ITC's encouraging labor supply. 00:22:42.850 --> 00:22:45.710 align:middle line:84% Well, it does for A. But it's ambiguous for B, 00:22:45.710 --> 00:22:49.330 align:middle line:84% and it ambiguously deters labor supply for C and D. 00:22:49.330 --> 00:22:52.170 align:middle line:84% So it's a very open empirical question about whether the ITC 00:22:52.170 --> 00:22:53.530 align:middle line:90% will actually-- 00:22:53.530 --> 00:22:56.210 align:middle line:90% how it will change labor supply. 00:22:56.210 --> 00:22:58.890 align:middle line:84% Fortunately, the ITC has been fucked around 00:22:58.890 --> 00:23:01.230 align:middle line:84% with a lot over the last few decades, 00:23:01.230 --> 00:23:03.210 align:middle line:84% so you have lots quasi experiments 00:23:03.210 --> 00:23:05.070 align:middle line:90% to look at to understand it. 00:23:05.070 --> 00:23:10.050 align:middle line:84% So for example, if you go to Figure 21.7, 00:23:10.050 --> 00:23:12.170 align:middle line:84% this shows, for example, how the ITC changed 00:23:12.170 --> 00:23:15.070 align:middle line:90% after Tax Reform Act of 1986. 00:23:15.070 --> 00:23:18.650 align:middle line:84% And the Tax Act, this was when it was first introduced. 00:23:18.650 --> 00:23:22.250 align:middle line:84% The slope used to be minus 0.11 slope. 00:23:22.250 --> 00:23:23.830 align:middle line:84% On the upwards part, you see 0.11. 00:23:23.830 --> 00:23:25.370 align:middle line:90% Now it's 0.4. 00:23:25.370 --> 00:23:26.050 align:middle line:90% It went up. 00:23:26.050 --> 00:23:27.210 align:middle line:90% The slope went up. 00:23:27.210 --> 00:23:28.290 align:middle line:90% They raised the maximum. 00:23:28.290 --> 00:23:29.478 align:middle line:90% They change the phase out. 00:23:29.478 --> 00:23:31.770 align:middle line:84% So you could set up a standard difference in difference 00:23:31.770 --> 00:23:33.937 align:middle line:84% by looking at groups that were more or less affected 00:23:33.937 --> 00:23:35.090 align:middle line:90% by this change. 00:23:35.090 --> 00:23:41.450 align:middle line:84% Moreover, the ITC depends on the number of children you have. 00:23:41.450 --> 00:23:43.242 align:middle line:90% I wrote this-- let's see. 00:23:43.242 --> 00:23:44.950 align:middle line:84% What number of kids did I wrote this for? 00:23:44.950 --> 00:23:48.770 align:middle line:90% 00:23:48.770 --> 00:23:50.610 align:middle line:90% Single earner with two children. 00:23:50.610 --> 00:23:52.850 align:middle line:84% Well, if you have one child, your ITC 00:23:52.850 --> 00:23:54.650 align:middle line:84% is smaller than if you have two children, 00:23:54.650 --> 00:23:57.010 align:middle line:84% so that sets up another difference-in-difference 00:23:57.010 --> 00:23:57.810 align:middle line:90% approach. 00:23:57.810 --> 00:23:59.250 align:middle line:84% So there's lots quasi experiments 00:23:59.250 --> 00:24:01.750 align:middle line:84% you can do here, looking at, what happens to the ITC changes 00:24:01.750 --> 00:24:04.170 align:middle line:84% for women with one kid versus two kids, women 00:24:04.170 --> 00:24:05.910 align:middle line:90% with no kids versus some kids? 00:24:05.910 --> 00:24:08.490 align:middle line:84% Women, no kids get a very tiny ITC. 00:24:08.490 --> 00:24:11.490 align:middle line:84% By the way, it's generally women because generally married 00:24:11.490 --> 00:24:14.510 align:middle line:84% couples, men and married couples earn higher amounts. 00:24:14.510 --> 00:24:15.890 align:middle line:90% Well, actually, let me back up. 00:24:15.890 --> 00:24:17.530 align:middle line:90% Strike that last statement. 00:24:17.530 --> 00:24:20.810 align:middle line:84% I refer to women because that is the population that folks 00:24:20.810 --> 00:24:23.550 align:middle line:84% are thinking about because the big thing of the ITC was, 00:24:23.550 --> 00:24:25.600 align:middle line:90% can it get people off welfare? 00:24:25.600 --> 00:24:28.160 align:middle line:84% So the initial focus was, what effect does it have on women? 00:24:28.160 --> 00:24:31.720 align:middle line:84% And how does it compare with the disincentive effects of welfare? 00:24:31.720 --> 00:24:33.840 align:middle line:84% So that's where most of the literature 00:24:33.840 --> 00:24:36.720 align:middle line:84% is, is looking at single moms compared to, say, single women 00:24:36.720 --> 00:24:40.720 align:middle line:84% without kids who used to get no EITC and now get a tiny one 00:24:40.720 --> 00:24:43.058 align:middle line:90% or, say, single men. 00:24:43.058 --> 00:24:44.600 align:middle line:84% So they look at the effect of the ITC 00:24:44.600 --> 00:24:47.920 align:middle line:84% on those groups using changes in ITC generosity. 00:24:47.920 --> 00:24:49.300 align:middle line:90% And what did they find? 00:24:49.300 --> 00:24:51.880 align:middle line:90% Well, the effects were striking. 00:24:51.880 --> 00:24:55.260 align:middle line:90% Huge increases in working. 00:24:55.260 --> 00:24:59.640 align:middle line:84% So people type A responded dramatically, 00:24:59.640 --> 00:25:02.620 align:middle line:90% big increase in working. 00:25:02.620 --> 00:25:03.120 align:middle line:90% And 00:25:03.120 --> 00:25:05.180 align:middle line:84% No apparent other effects on groups B, C, 00:25:05.180 --> 00:25:09.160 align:middle line:84% and D. It didn't seem to deter work among groups B, C, and D. 00:25:09.160 --> 00:25:12.720 align:middle line:84% So in other words, the group where it unambiguously increases 00:25:12.720 --> 00:25:15.840 align:middle line:90% work, it did, large effects. 00:25:15.840 --> 00:25:19.480 align:middle line:84% But you didn't really see the offsetting reduced effects 00:25:19.480 --> 00:25:22.560 align:middle line:84% among those elsewhere in the budget constraint. 00:25:22.560 --> 00:25:25.120 align:middle line:84% Now, there's two possible explanations for why. 00:25:25.120 --> 00:25:26.700 align:middle line:84% One is differential elasticities. 00:25:26.700 --> 00:25:28.720 align:middle line:84% Maybe the elasticity of participation 00:25:28.720 --> 00:25:30.885 align:middle line:84% that is going 0 to 1 is very elastic, 00:25:30.885 --> 00:25:32.760 align:middle line:84% but the elasticity of how many hours you work 00:25:32.760 --> 00:25:34.880 align:middle line:84% isn't because you can't really control your hours. 00:25:34.880 --> 00:25:35.755 align:middle line:90% You work for the guy. 00:25:35.755 --> 00:25:37.800 align:middle line:90% He tells you how hard to work. 00:25:37.800 --> 00:25:39.520 align:middle line:84% So one explanation is just literally 00:25:39.520 --> 00:25:40.960 align:middle line:84% the one thing you clearly have a choice over 00:25:40.960 --> 00:25:42.180 align:middle line:90% is whether to work or not. 00:25:42.180 --> 00:25:44.320 align:middle line:84% How hard you work you don't have as much control. 00:25:44.320 --> 00:25:45.960 align:middle line:90% That's one explanation. 00:25:45.960 --> 00:25:48.702 align:middle line:84% The other explanation is that it's really pretty fucking 00:25:48.702 --> 00:25:50.160 align:middle line:84% hard to figure out these tax rates, 00:25:50.160 --> 00:25:51.525 align:middle line:90% to figure out this system. 00:25:51.525 --> 00:25:53.900 align:middle line:84% If I say to you, hey, if you go to work, you get a check. 00:25:53.900 --> 00:25:54.620 align:middle line:90% If you don't, you don't. 00:25:54.620 --> 00:25:55.135 align:middle line:90% That's easy. 00:25:55.135 --> 00:25:56.760 align:middle line:84% If I say, by the way, the marginal hour 00:25:56.760 --> 00:25:59.500 align:middle line:84% you will have [INAUDIBLE] marginal tax rate of minus $0.21 00:25:59.500 --> 00:26:01.560 align:middle line:84% if you work on this kink and not on this kink, 00:26:01.560 --> 00:26:04.000 align:middle line:84% it's like, I don't understand how that works. 00:26:04.000 --> 00:26:05.420 align:middle line:90% It's pretty hard to understand. 00:26:05.420 --> 00:26:07.760 align:middle line:84% So the other explanation for what's going on 00:26:07.760 --> 00:26:10.680 align:middle line:84% is that basically people recognize the incentive effect 00:26:10.680 --> 00:26:12.720 align:middle line:84% of getting a check if they go to work now 00:26:12.720 --> 00:26:14.898 align:middle line:84% but don't recognize the marginal tax rate. 00:26:14.898 --> 00:26:16.440 align:middle line:84% And indeed, there's a lot of evidence 00:26:16.440 --> 00:26:19.360 align:middle line:84% that people aren't responsive to marginal tax rates, at least 00:26:19.360 --> 00:26:20.920 align:middle line:90% in much of the range. 00:26:20.920 --> 00:26:23.750 align:middle line:84% And remember, as I talked before, I talk about people 00:26:23.750 --> 00:26:27.370 align:middle line:84% misunderstand the tax system and this may be an example of that. 00:26:27.370 --> 00:26:28.890 align:middle line:90% So not sure why it is. 00:26:28.890 --> 00:26:35.270 align:middle line:84% But the bottom line is that essentially the EITC 00:26:35.270 --> 00:26:43.630 align:middle line:84% seems to be, on net, create a large increase in labor supply. 00:26:43.630 --> 00:26:48.930 align:middle line:84% This is fascinating because this is unambiguously good news. 00:26:48.930 --> 00:26:53.190 align:middle line:84% It is a patch to Okun's bucket because we're clearly 00:26:53.190 --> 00:26:55.110 align:middle line:84% redistributing to low income people. 00:26:55.110 --> 00:26:58.150 align:middle line:84% No one with income more than $50,000 gets the ITC. 00:26:58.150 --> 00:26:59.670 align:middle line:90% That's pretty low income. 00:26:59.670 --> 00:27:01.630 align:middle line:84% And we're doing so in a way that raises, 00:27:01.630 --> 00:27:04.230 align:middle line:90% not lowering labor supply. 00:27:04.230 --> 00:27:08.430 align:middle line:84% So this is wonderful news, which is that basically we've 00:27:08.430 --> 00:27:13.230 align:middle line:84% discovered a tool for fighting poverty that actually does so 00:27:13.230 --> 00:27:17.670 align:middle line:84% without causing reduction in the size of social pie. 00:27:17.670 --> 00:27:21.230 align:middle line:84% Nonetheless, the ITC is still controversial, 00:27:21.230 --> 00:27:24.190 align:middle line:90% and that's for two reasons. 00:27:24.190 --> 00:27:26.670 align:middle line:84% One reason is, as I said, it can cause people 00:27:26.670 --> 00:27:28.910 align:middle line:90% to have negative tax burdens. 00:27:28.910 --> 00:27:32.030 align:middle line:84% And some people, as I said, irrationally 00:27:32.030 --> 00:27:34.310 align:middle line:84% feel the tax system should not be paying people money. 00:27:34.310 --> 00:27:35.810 align:middle line:84% Tax burden should never be negative. 00:27:35.810 --> 00:27:38.212 align:middle line:84% As a result, the ITC's a bad idea. 00:27:38.212 --> 00:27:39.670 align:middle line:84% The other reason it's controversial 00:27:39.670 --> 00:27:42.390 align:middle line:84% is because there's a decent amount of cheating. 00:27:42.390 --> 00:27:45.110 align:middle line:84% Basically, like with any part of the tax system, 00:27:45.110 --> 00:27:48.450 align:middle line:84% there are people who will lie about how much they make 00:27:48.450 --> 00:27:50.410 align:middle line:90% if it's not really recorded. 00:27:50.410 --> 00:27:52.970 align:middle line:84% it's lying to say you made more than what you made before. 00:27:52.970 --> 00:27:54.470 align:middle line:90% It's an odd kind of lying. 00:27:54.470 --> 00:27:56.885 align:middle line:84% But once again, people will do that. 00:27:56.885 --> 00:27:59.010 align:middle line:84% They might misreport other features of their taxes, 00:27:59.010 --> 00:28:00.593 align:middle line:84% and they might say kids live with them 00:28:00.593 --> 00:28:02.870 align:middle line:84% who don't to try to claim the credit, et cetera. 00:28:02.870 --> 00:28:05.870 align:middle line:84% Now, we're going to come back to this in chapter 25. 00:28:05.870 --> 00:28:07.310 align:middle line:84% The amount of cheating on the ITC 00:28:07.310 --> 00:28:09.560 align:middle line:84% is like a tiny fraction of all cheating done on taxes. 00:28:09.560 --> 00:28:10.935 align:middle line:84% Fundamentally, only the rich guys 00:28:10.935 --> 00:28:13.050 align:middle line:84% really do the big amount of cheating on taxes. 00:28:13.050 --> 00:28:16.070 align:middle line:84% But nonetheless, so I view these as thinly veiled excuses 00:28:16.070 --> 00:28:18.570 align:middle line:84% for people who really don't like redistributing to the poor, 00:28:18.570 --> 00:28:21.520 align:middle line:84% to be honest, but they are excuses that play in Washington, 00:28:21.520 --> 00:28:24.300 align:middle line:84% and that's one reason the ITC is probably even bigger than it is. 00:28:24.300 --> 00:28:27.240 align:middle line:84% There is a third weakness, which is important, 00:28:27.240 --> 00:28:30.140 align:middle line:84% and this comes back to chapter 17. 00:28:30.140 --> 00:28:32.160 align:middle line:84% You might say, well, gee, this is awesome. 00:28:32.160 --> 00:28:34.740 align:middle line:84% Let's get rid of welfare and replace-- let's forget 00:28:34.740 --> 00:28:35.720 align:middle line:90% these other problems. 00:28:35.720 --> 00:28:38.162 align:middle line:84% Let's imagine a rational system that honest debates. 00:28:38.162 --> 00:28:39.620 align:middle line:84% Let's say, let's get rid of welfare 00:28:39.620 --> 00:28:41.140 align:middle line:90% and just replace the ITC. 00:28:41.140 --> 00:28:43.775 align:middle line:84% What would the problem would that be? 00:28:43.775 --> 00:28:46.400 align:middle line:84% Just say, let's get rid of this whole labor distorting welfare, 00:28:46.400 --> 00:28:48.432 align:middle line:90% just having a big EITC system. 00:28:48.432 --> 00:28:49.640 align:middle line:90% What's the problem with that? 00:28:49.640 --> 00:28:52.980 align:middle line:90% 00:28:52.980 --> 00:28:53.520 align:middle line:90% Yeah? 00:28:53.520 --> 00:28:54.580 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:28:54.580 --> 00:28:57.020 align:middle line:84% PROFESSOR: Some people legitimately can't work. 00:28:57.020 --> 00:29:00.580 align:middle line:84% You still need-- there's a distortion from subsidizing 00:29:00.580 --> 00:29:05.060 align:middle line:84% non-work, but there's also a consumption-smoothing benefit. 00:29:05.060 --> 00:29:07.300 align:middle line:84% So an optimal system would never feature nothing 00:29:07.300 --> 00:29:09.900 align:middle line:84% for those who can't work, as long as there 00:29:09.900 --> 00:29:11.722 align:middle line:84% is some exogenous reason people can't work. 00:29:11.722 --> 00:29:13.180 align:middle line:84% You never want a system that leaves 00:29:13.180 --> 00:29:14.980 align:middle line:90% people that work with nothing. 00:29:14.980 --> 00:29:17.820 align:middle line:84% So the optimal has to be some combination of money 00:29:17.820 --> 00:29:19.820 align:middle line:84% to those who don't work but then incentives 00:29:19.820 --> 00:29:21.380 align:middle line:90% to work for those who can. 00:29:21.380 --> 00:29:23.860 align:middle line:84% What that combination looks like is complicated. 00:29:23.860 --> 00:29:25.540 align:middle line:90% All right. 00:29:25.540 --> 00:29:29.540 align:middle line:90% So that's the EITC. 00:29:29.540 --> 00:29:31.300 align:middle line:84% And it's a great example of how we 00:29:31.300 --> 00:29:34.940 align:middle line:84% think about applying our logic of income substitution effects 00:29:34.940 --> 00:29:38.300 align:middle line:90% to making policy. 00:29:38.300 --> 00:29:39.100 align:middle line:90% Yeah? 00:29:39.100 --> 00:29:40.600 align:middle line:84% AUDIENCE: How do people study things 00:29:40.600 --> 00:29:42.728 align:middle line:84% like cheating that go, by definition, unrecorded? 00:29:42.728 --> 00:29:43.520 align:middle line:90% PROFESSOR: Awesome. 00:29:43.520 --> 00:29:44.040 align:middle line:90% Great question. 00:29:44.040 --> 00:29:45.707 align:middle line:84% I'll talk about this more in chapter 25. 00:29:45.707 --> 00:29:48.340 align:middle line:84% But basically IRS does audits, and with the audits, 00:29:48.340 --> 00:29:50.540 align:middle line:84% they can actually compare people report to what 00:29:50.540 --> 00:29:51.760 align:middle line:90% you see in the audit data. 00:29:51.760 --> 00:29:54.540 align:middle line:90% 00:29:54.540 --> 00:29:55.993 align:middle line:84% AUDIENCE: Based on that sampling? 00:29:55.993 --> 00:29:56.660 align:middle line:90% PROFESSOR: Yeah. 00:29:56.660 --> 00:29:57.640 align:middle line:90% That's your random sample. 00:29:57.640 --> 00:29:59.140 align:middle line:84% They literally go and make your life miserable 00:29:59.140 --> 00:30:01.015 align:middle line:84% and collect all the data, and they compare it 00:30:01.015 --> 00:30:03.860 align:middle line:90% to what you actually declared. 00:30:03.860 --> 00:30:08.320 align:middle line:90% So that's one application. 00:30:08.320 --> 00:30:12.100 align:middle line:84% The other application I want to talk about is child care. 00:30:12.100 --> 00:30:14.200 align:middle line:84% We talked a little about child care in chapter 17. 00:30:14.200 --> 00:30:15.242 align:middle line:90% I want to return to that. 00:30:15.242 --> 00:30:21.580 align:middle line:84% Remember, that basically, today, fewer than of preschool 00:30:21.580 --> 00:30:24.700 align:middle line:84% aged children are cared for by a parent. 00:30:24.700 --> 00:30:26.840 align:middle line:84% 12.5% are cared for by their relatives, 00:30:26.840 --> 00:30:29.820 align:middle line:84% and 62% are in some form of child care. 00:30:29.820 --> 00:30:33.180 align:middle line:84% So almost 2/3 of kids today are preschool age are in child care. 00:30:33.180 --> 00:30:35.780 align:middle line:84% Child care is a very big deal, and it 00:30:35.780 --> 00:30:38.260 align:middle line:84% makes it possible for women to actually finally realize 00:30:38.260 --> 00:30:40.980 align:middle line:84% their potential and join the labor force in a way they 00:30:40.980 --> 00:30:44.460 align:middle line:90% couldn't in decades past. 00:30:44.460 --> 00:30:47.580 align:middle line:84% The interesting question is, how should we treat childcare 00:30:47.580 --> 00:30:48.880 align:middle line:90% for tax purposes? 00:30:48.880 --> 00:30:52.800 align:middle line:90% 00:30:52.800 --> 00:30:55.420 align:middle line:84% And here we're going to go back to chapter 18 00:30:55.420 --> 00:30:57.780 align:middle line:84% and thinking about Haig-Simons, think 00:30:57.780 --> 00:31:01.100 align:middle line:90% about Haig-Simons principles 00:31:01.100 --> 00:31:04.420 align:middle line:84% So basically if you think about it, 00:31:04.420 --> 00:31:09.700 align:middle line:84% when you deliver-- when my wife stayed home and took care 00:31:09.700 --> 00:31:14.420 align:middle line:84% of our kids, she was delivering something 00:31:14.420 --> 00:31:16.085 align:middle line:90% which expanded our resources. 00:31:16.085 --> 00:31:18.210 align:middle line:84% We were better off because she stayed home and took 00:31:18.210 --> 00:31:19.650 align:middle line:90% care of my kids. 00:31:19.650 --> 00:31:23.850 align:middle line:84% But if she'd gone to work, she would have been taxed. 00:31:23.850 --> 00:31:27.450 align:middle line:84% But the work she did staying home was not taxed. 00:31:27.450 --> 00:31:29.670 align:middle line:84% And that's a violation of Haig-Simons. 00:31:29.670 --> 00:31:32.515 align:middle line:84% Haig-Simons says you should tax everybody's potential to pay. 00:31:32.515 --> 00:31:33.390 align:middle line:90% Well, think about it. 00:31:33.390 --> 00:31:38.610 align:middle line:84% Let's say my wife stays home, and your wife goes to work. 00:31:38.610 --> 00:31:41.207 align:middle line:84% And they're both delivering the same value of goods. 00:31:41.207 --> 00:31:42.790 align:middle line:84% It's just my wife's still in the home, 00:31:42.790 --> 00:31:43.915 align:middle line:90% and yours is in the market. 00:31:43.915 --> 00:31:44.510 align:middle line:90% You get taxed. 00:31:44.510 --> 00:31:45.010 align:middle line:90% I don't. 00:31:45.010 --> 00:31:45.723 align:middle line:90% That's unfair. 00:31:45.723 --> 00:31:47.390 align:middle line:84% That violates the Haig-Simons principle. 00:31:47.390 --> 00:31:50.450 align:middle line:84% We should be taxed on our underlying resources, which 00:31:50.450 --> 00:31:52.870 align:middle line:84% are the same, yet we pay very different taxes. 00:31:52.870 --> 00:31:55.985 align:middle line:84% It's a violation of some concept of horizontal equity. 00:31:55.985 --> 00:31:57.110 align:middle line:90% So what do you do about it? 00:31:57.110 --> 00:31:59.090 align:middle line:84% Well, basically there's three approaches. 00:31:59.090 --> 00:32:00.850 align:middle line:90% So let's go to an example. 00:32:00.850 --> 00:32:07.010 align:middle line:84% Let's go to table 21.1 Go to table 21.1 00:32:07.010 --> 00:32:11.810 align:middle line:90% So imagine you have a couple. 00:32:11.810 --> 00:32:14.393 align:middle line:84% Imagine you have a couple where the man's going to go to work, 00:32:14.393 --> 00:32:16.518 align:middle line:84% and the woman is considering whether to go to work. 00:32:16.518 --> 00:32:18.450 align:middle line:84% The wife is considering whether to go to work. 00:32:18.450 --> 00:32:23.330 align:middle line:84% And imagine that, if she went to work, she'd earn $1,000, 00:32:23.330 --> 00:32:29.290 align:middle line:84% but she would pay $600 in child care costs. 00:32:29.290 --> 00:32:31.730 align:middle line:84% And let's assume markets are perfect. 00:32:31.730 --> 00:32:34.770 align:middle line:84% So $600 in child care costs is literally 00:32:34.770 --> 00:32:37.850 align:middle line:84% the inherent social value of caring for the kids. 00:32:37.850 --> 00:32:41.610 align:middle line:84% And the $1,000 is the inherent social value of her work. 00:32:41.610 --> 00:32:45.490 align:middle line:84% So what that means is she should work, full stop. 00:32:45.490 --> 00:32:48.650 align:middle line:84% The society for working is $1,000. 00:32:48.650 --> 00:32:50.318 align:middle line:84% The society of her kids being cared 00:32:50.318 --> 00:32:52.610 align:middle line:84% for-- ignoring externalities of different people caring 00:32:52.610 --> 00:32:53.970 align:middle line:90% for kids, put all that away. 00:32:53.970 --> 00:32:56.490 align:middle line:84% Simple model, where basically the market 00:32:56.490 --> 00:32:59.550 align:middle line:84% works well [INAUDIBLE] caring for those kids is $600. 00:32:59.550 --> 00:33:03.290 align:middle line:84% So the size social pie is higher if she works. 00:33:03.290 --> 00:33:09.250 align:middle line:84% However, if she goes to work in the 1 to 3 or 5th column, 00:33:09.250 --> 00:33:12.230 align:middle line:84% if she goes to work at a 50% tax rate-- because remember, 00:33:12.230 --> 00:33:14.040 align:middle line:84% her husband's already earning money, 00:33:14.040 --> 00:33:17.720 align:middle line:84% so the marginal tax rate on her, if you've got a high earning 00:33:17.720 --> 00:33:21.698 align:middle line:84% spouse, your next dollar is already in a top tax bracket. 00:33:21.698 --> 00:33:22.740 align:middle line:90% You don't get to benefit. 00:33:22.740 --> 00:33:23.780 align:middle line:84% Remember, we talked about marriage tax. 00:33:23.780 --> 00:33:25.660 align:middle line:84% You don't get to benefit from that low tax rate initially. 00:33:25.660 --> 00:33:28.160 align:middle line:84% You're already in the high tax bracket when you go to work. 00:33:28.160 --> 00:33:32.320 align:middle line:84% So the marginal dollar, let's say, is taxed at 50%. 00:33:32.320 --> 00:33:35.413 align:middle line:84% What that means is, if she goes to work and earns $1,000, 00:33:35.413 --> 00:33:37.880 align:middle line:90% she'll go $500 in taxes. 00:33:37.880 --> 00:33:42.800 align:middle line:84% So her the after tax value of her work will be $500. 00:33:42.800 --> 00:33:47.360 align:middle line:84% But if she stays home, she'll effectively make $600. 00:33:47.360 --> 00:33:50.120 align:middle line:90% So she won't go to work. 00:33:50.120 --> 00:33:52.960 align:middle line:84% So in other words, we have distorted the market. 00:33:52.960 --> 00:33:56.830 align:middle line:84% We have not made a social welfare-improving trade. 00:33:56.830 --> 00:33:59.080 align:middle line:84% The social welfare-improving trade is for her to work, 00:33:59.080 --> 00:34:01.120 align:middle line:90% but she doesn't. 00:34:01.120 --> 00:34:04.320 align:middle line:84% This is the distortion that arises from not taxing 00:34:04.320 --> 00:34:06.680 align:middle line:90% the true Haig-Simons base. 00:34:06.680 --> 00:34:10.960 align:middle line:90% We call this a tax wedge. 00:34:10.960 --> 00:34:16.400 align:middle line:84% A tax wedge exists whenever taxes 00:34:16.400 --> 00:34:21.620 align:middle line:84% impact the incentives to engage in two, if you will, 00:34:21.620 --> 00:34:23.960 align:middle line:84% economically similar activities or activities 00:34:23.960 --> 00:34:26.560 align:middle line:84% where you basically would like to compare them-- 00:34:26.560 --> 00:34:30.159 align:middle line:84% basically, if you have a comparison where there's 00:34:30.159 --> 00:34:31.864 align:middle line:84% a relative advantage to one-- let 00:34:31.864 --> 00:34:33.239 align:middle line:84% me think about this-- actually, I 00:34:33.239 --> 00:34:34.219 align:middle line:84% think I have the definition in the book. 00:34:34.219 --> 00:34:35.136 align:middle line:90% Let me get this right. 00:34:35.136 --> 00:34:39.159 align:middle line:90% 00:34:39.159 --> 00:34:42.878 align:middle line:84% Basically, tax wedges are basically 00:34:42.878 --> 00:34:44.920 align:middle line:84% the difference between the returns to an activity 00:34:44.920 --> 00:34:46.060 align:middle line:90% in different sectors. 00:34:46.060 --> 00:34:47.139 align:middle line:84% That's the way to think about it. 00:34:47.139 --> 00:34:49.681 align:middle line:84% You can deliver labor supply in the market sector or the home 00:34:49.681 --> 00:34:51.080 align:middle line:90% sector. 00:34:51.080 --> 00:34:54.600 align:middle line:84% If taxes change the relative returns to those activities, 00:34:54.600 --> 00:34:57.460 align:middle line:90% that creates a tax wedge. 00:34:57.460 --> 00:34:59.400 align:middle line:90% And here that's happened. 00:34:59.400 --> 00:35:00.940 align:middle line:90% Now, what can we do about it? 00:35:00.940 --> 00:35:02.720 align:middle line:90% Well, there's two answers. 00:35:02.720 --> 00:35:05.340 align:middle line:84% One is what Haig-Simons would have us do, which is say, look, 00:35:05.340 --> 00:35:07.360 align:middle line:90% let's measure her full income. 00:35:07.360 --> 00:35:10.650 align:middle line:84% Her full income is she makes $1,000 if she works. 00:35:10.650 --> 00:35:13.270 align:middle line:84% She makes $600 if she stays home. 00:35:13.270 --> 00:35:15.410 align:middle line:84% Why does she make $600 if she stays home? 00:35:15.410 --> 00:35:18.790 align:middle line:84% Because she's delivering $600 worth of child care. 00:35:18.790 --> 00:35:23.230 align:middle line:84% So in the Haig-Simons world, we would impute earnings to her 00:35:23.230 --> 00:35:24.370 align:middle line:90% if she stayed home. 00:35:24.370 --> 00:35:26.710 align:middle line:84% We'd say, if you go to work, you're wearing 1,000. 00:35:26.710 --> 00:35:30.350 align:middle line:90% If you stay home, you earn 600. 00:35:30.350 --> 00:35:32.450 align:middle line:84% And in either case, you'd then pay taxes. 00:35:32.450 --> 00:35:34.710 align:middle line:84% You'd pay 500 taxes if you worked, 00:35:34.710 --> 00:35:36.710 align:middle line:90% 300 tax if you stayed home. 00:35:36.710 --> 00:35:38.942 align:middle line:84% So the after-tax value of work would be 500. 00:35:38.942 --> 00:35:40.650 align:middle line:84% The after-tax value of home would be 300, 00:35:40.650 --> 00:35:42.270 align:middle line:90% and she'd go to work. 00:35:42.270 --> 00:35:44.150 align:middle line:84% She'd get the right answer because you've 00:35:44.150 --> 00:35:46.190 align:middle line:90% gotten rid of the tax wedge. 00:35:46.190 --> 00:35:48.430 align:middle line:84% You've gotten rid of the differential effect of taxes 00:35:48.430 --> 00:35:50.693 align:middle line:84% on that labor supply delivered in two different ways. 00:35:50.693 --> 00:35:53.110 align:middle line:84% It used to be you're taxed differently if you delivered it 00:35:53.110 --> 00:35:54.190 align:middle line:90% in market at home. 00:35:54.190 --> 00:35:56.750 align:middle line:90% Now you're taxed the same. 00:35:56.750 --> 00:35:58.430 align:middle line:90% That would be one way to do it. 00:35:58.430 --> 00:36:00.790 align:middle line:84% The other way to do it would be to allow 00:36:00.790 --> 00:36:06.310 align:middle line:84% her to deduct the cost of child care from her income. 00:36:06.310 --> 00:36:08.310 align:middle line:90% So she'd earn $1,000. 00:36:08.310 --> 00:36:10.470 align:middle line:84% She'd spend spent 600 in child care. 00:36:10.470 --> 00:36:13.470 align:middle line:84% But she would then get to deduct that 600 00:36:13.470 --> 00:36:17.870 align:middle line:84% so her net taxable income would only be 400. 00:36:17.870 --> 00:36:21.870 align:middle line:84% Therefore she would owe 200 in taxes, 00:36:21.870 --> 00:36:24.870 align:middle line:84% and she would make more money at work, $800, 00:36:24.870 --> 00:36:27.130 align:middle line:84% than staying home, which is $600. 00:36:27.130 --> 00:36:29.470 align:middle line:84% So she'd, once again, go to work. 00:36:29.470 --> 00:36:32.910 align:middle line:84% So two different approaches that end the distortion 00:36:32.910 --> 00:36:35.790 align:middle line:90% caused by this tax wedge. 00:36:35.790 --> 00:36:39.030 align:middle line:84% Here's where public finance gets really fun. 00:36:39.030 --> 00:36:41.470 align:middle line:84% How do I think about which one is better? 00:36:41.470 --> 00:36:43.667 align:middle line:84% First of all, let's forget feasibility. 00:36:43.667 --> 00:36:45.250 align:middle line:84% Let's forget feasibility for a second. 00:36:45.250 --> 00:36:47.320 align:middle line:84% Obviously, imputing income would be a nightmare politically, 00:36:47.320 --> 00:36:48.230 align:middle line:90% et cetera. 00:36:48.230 --> 00:36:49.810 align:middle line:90% Let's forget that. 00:36:49.810 --> 00:36:52.790 align:middle line:84% But how do I think about, between these two approaches, 00:36:52.790 --> 00:36:54.750 align:middle line:90% which one is better, and why? 00:36:54.750 --> 00:36:56.670 align:middle line:84% In a world where I could do either, 00:36:56.670 --> 00:36:58.550 align:middle line:90% there's an obvious answer. 00:36:58.550 --> 00:37:01.550 align:middle line:84% Which one would I prefer, and why? 00:37:01.550 --> 00:37:03.790 align:middle line:90% Yeah? 00:37:03.790 --> 00:37:04.850 align:middle line:90% You got to speak up. 00:37:04.850 --> 00:37:07.260 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:37:07.260 --> 00:37:08.480 align:middle line:90% PROFESSOR: Because of what? 00:37:08.480 --> 00:37:09.553 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:37:09.553 --> 00:37:11.720 align:middle line:84% PROFESSOR: No, I'm saying imagine where we could do. 00:37:11.720 --> 00:37:13.845 align:middle line:84% Both I'm getting rid of the administrative problem. 00:37:13.845 --> 00:37:18.580 align:middle line:84% Just on pure economics, which would you choose, and why? 00:37:18.580 --> 00:37:21.780 align:middle line:84% Appealing to the lessons from last lecture, from chapter 20, 00:37:21.780 --> 00:37:23.740 align:middle line:90% which would you choose, and why? 00:37:23.740 --> 00:37:24.700 align:middle line:90% Yeah? 00:37:24.700 --> 00:37:27.820 align:middle line:84% AUDIENCE: It seems like [INAUDIBLE] is overall cheaper 00:37:27.820 --> 00:37:30.565 align:middle line:84% to do from the government's perspective and does the same-- 00:37:30.565 --> 00:37:31.940 align:middle line:84% PROFESSOR: Well, remember, what's 00:37:31.940 --> 00:37:33.220 align:middle line:90% the key source of distortion? 00:37:33.220 --> 00:37:33.720 align:middle line:90% It's two. 00:37:33.720 --> 00:37:36.740 align:middle line:84% It's elasticities and tau squared. 00:37:36.740 --> 00:37:41.020 align:middle line:84% The smaller the tax base, the higher the tax rate has to be. 00:37:41.020 --> 00:37:44.142 align:middle line:84% So by shrinking the tax base in the second approach 00:37:44.142 --> 00:37:45.600 align:middle line:84% to raise a given amount of revenue, 00:37:45.600 --> 00:37:49.500 align:middle line:84% we're going to have to tax more highly, so it's less efficient. 00:37:49.500 --> 00:37:53.500 align:middle line:84% So it's unambiguous, from the tau squared perspective, 00:37:53.500 --> 00:37:55.820 align:middle line:84% you'd like the first approach because you 00:37:55.820 --> 00:37:58.300 align:middle line:84% get the same-- you remove the wedge, 00:37:58.300 --> 00:37:59.560 align:middle line:90% but you collect more revenues. 00:37:59.560 --> 00:38:01.700 align:middle line:84% And therefore total tax rates can be lower. 00:38:01.700 --> 00:38:04.380 align:middle line:90% People understand that? 00:38:04.380 --> 00:38:06.780 align:middle line:84% All else equal, we want taxes with broad bases 00:38:06.780 --> 00:38:07.900 align:middle line:90% and lower rates. 00:38:07.900 --> 00:38:12.040 align:middle line:84% That's how we maximize efficiency. 00:38:12.040 --> 00:38:12.540 align:middle line:90% Yeah? 00:38:12.540 --> 00:38:14.580 align:middle line:84% AUDIENCE: Why is it for [INAUDIBLE], 00:38:14.580 --> 00:38:17.590 align:middle line:90% the value [INAUDIBLE]? 00:38:17.590 --> 00:38:18.840 align:middle line:90% PROFESSOR: Why is it not what? 00:38:18.840 --> 00:38:21.420 align:middle line:84% AUDIENCE: Why is it not 900 [INAUDIBLE]? 00:38:21.420 --> 00:38:23.912 align:middle line:90% 00:38:23.912 --> 00:38:25.120 align:middle line:90% PROFESSOR: Why is it not 900? 00:38:25.120 --> 00:38:28.213 align:middle line:84% No, you're delivering $600 worth of child care. 00:38:28.213 --> 00:38:30.380 align:middle line:84% AUDIENCE: But with the earnings on top, or is that-- 00:38:30.380 --> 00:38:31.130 align:middle line:90% PROFESSOR: No, no. 00:38:31.130 --> 00:38:32.980 align:middle line:84% If you stay home, you're delivering 00:38:32.980 --> 00:38:36.220 align:middle line:90% 600 worth of child care. 00:38:36.220 --> 00:38:37.240 align:middle line:90% What's the extra 300? 00:38:37.240 --> 00:38:38.180 align:middle line:90% I don't get it. 00:38:38.180 --> 00:38:39.740 align:middle line:84% AUDIENCE: Because if you're getting 00:38:39.740 --> 00:38:42.660 align:middle line:84% earnings from the government in that case but also-- 00:38:42.660 --> 00:38:44.040 align:middle line:84% PROFESSOR: You're not getting earnings from the government. 00:38:44.040 --> 00:38:45.400 align:middle line:84% You mean the imputed case in the second row? 00:38:45.400 --> 00:38:47.420 align:middle line:84% You're not getting earnings from the government. 00:38:47.420 --> 00:38:51.040 align:middle line:84% Yeah, you're just getting-- it's you're earning $600. 00:38:51.040 --> 00:38:54.100 align:middle line:84% They're going to tax $300 away from you. 00:38:54.100 --> 00:38:55.600 align:middle line:84% Now, let's make it more interesting. 00:38:55.600 --> 00:38:56.933 align:middle line:90% Let's get rid of the second row. 00:38:56.933 --> 00:38:58.800 align:middle line:84% We're never going to do the second row. 00:38:58.800 --> 00:39:01.060 align:middle line:84% First row versus third row, Tell. 00:39:01.060 --> 00:39:04.492 align:middle line:84% me actually why this might-- now here's a really hard one. 00:39:04.492 --> 00:39:06.450 align:middle line:84% But once again, it all comes out of the formula 00:39:06.450 --> 00:39:08.130 align:middle line:90% for deadweight loss. 00:39:08.130 --> 00:39:11.050 align:middle line:84% Tell me why the first versus the third row 00:39:11.050 --> 00:39:14.665 align:middle line:84% is actually ambiguous which is a better one to do? 00:39:14.665 --> 00:39:16.290 align:middle line:84% It's actually ambiguous, and the answer 00:39:16.290 --> 00:39:20.330 align:middle line:84% comes directly from the formula for deadweight loss. 00:39:20.330 --> 00:39:23.850 align:middle line:90% Where's the ambiguity come from? 00:39:23.850 --> 00:39:27.450 align:middle line:84% We know the third row raises less money, 00:39:27.450 --> 00:39:29.710 align:middle line:84% so we know which way the tau squared effect goes. 00:39:29.710 --> 00:39:33.487 align:middle line:84% But what's the other effect that drives deadweight loss? 00:39:33.487 --> 00:39:35.570 align:middle line:84% What's the other thing that drives deadweight loss 00:39:35.570 --> 00:39:38.330 align:middle line:90% besides the tau squared term? 00:39:38.330 --> 00:39:41.850 align:middle line:90% Think about your formula. 00:39:41.850 --> 00:39:43.690 align:middle line:90% Elasticity. 00:39:43.690 --> 00:39:47.410 align:middle line:84% So why might the elasticity go the other way? 00:39:47.410 --> 00:39:49.490 align:middle line:90% Well, think about it. 00:39:49.490 --> 00:39:52.650 align:middle line:84% If you look at the last-- if you look at the first row, 00:39:52.650 --> 00:39:56.010 align:middle line:84% the relevant elasticity is the elasticity 00:39:56.010 --> 00:40:00.890 align:middle line:84% of women substituting between work and home. 00:40:00.890 --> 00:40:03.815 align:middle line:84% If you look at the last row, the relevant of elasticity, 00:40:03.815 --> 00:40:05.190 align:middle line:84% elasticity of the whole tax base, 00:40:05.190 --> 00:40:07.790 align:middle line:84% we're basically collecting less tax revenues. 00:40:07.790 --> 00:40:09.030 align:middle line:90% So let's simplify the world. 00:40:09.030 --> 00:40:11.810 align:middle line:84% Imagine the world consists of men working and women working. 00:40:11.810 --> 00:40:13.370 align:middle line:90% That's it. 00:40:13.370 --> 00:40:15.690 align:middle line:84% Well, in the first row, the thing you're distorting 00:40:15.690 --> 00:40:17.930 align:middle line:90% is a woman's decision to work. 00:40:17.930 --> 00:40:20.470 align:middle line:84% In the second row, you have to collect more money from both, 00:40:20.470 --> 00:40:23.330 align:middle line:84% so you're distorting both their decisions to work. 00:40:23.330 --> 00:40:26.110 align:middle line:84% Imagine, as is still true, less true than it used to be 00:40:26.110 --> 00:40:27.850 align:middle line:84% but is still true, women's labor supply 00:40:27.850 --> 00:40:30.850 align:middle line:84% is more elastic than men's labor supply, 00:40:30.850 --> 00:40:34.170 align:middle line:84% then the first row implies a more elastic decision. 00:40:34.170 --> 00:40:36.990 align:middle line:84% The decision for a woman to work or not is more elastic, 00:40:36.990 --> 00:40:39.410 align:middle line:84% the decision for all society to work or not because men 00:40:39.410 --> 00:40:41.890 align:middle line:84% have a lower elasticity than women. 00:40:41.890 --> 00:40:44.150 align:middle line:84% So actually the first versus third row is ambiguous. 00:40:44.150 --> 00:40:47.303 align:middle line:84% On the one hand, you raise less revenue than the third row. 00:40:47.303 --> 00:40:48.970 align:middle line:84% On the other, the first row, you distort 00:40:48.970 --> 00:40:51.450 align:middle line:84% a more elastic margin, which is women's decision 00:40:51.450 --> 00:40:56.170 align:middle line:84% to work or not work, and which dominates is unclear. 00:40:56.170 --> 00:40:57.497 align:middle line:90% Pretty mind blowing, I know. 00:40:57.497 --> 00:40:58.830 align:middle line:90% Take it home and think about it. 00:40:58.830 --> 00:41:00.250 align:middle line:90% Described clearly in the book. 00:41:00.250 --> 00:41:03.800 align:middle line:84% But this is how I want to think about this trade off. 00:41:03.800 --> 00:41:05.480 align:middle line:90% What do we do in the US? 00:41:05.480 --> 00:41:09.560 align:middle line:84% What do we do in the US is we compromise. 00:41:09.560 --> 00:41:14.920 align:middle line:84% We basically say that you get a tax credit for childcare, 00:41:14.920 --> 00:41:21.240 align:middle line:84% up to $3,000 for one child and $6,000 for two or more. 00:41:21.240 --> 00:41:23.400 align:middle line:84% And you get it as a percent of your income. 00:41:23.400 --> 00:41:27.520 align:middle line:84% So you can't like claim $6,000 and you only make $5,000. 00:41:27.520 --> 00:41:29.100 align:middle line:84% They say it's a percent of income. 00:41:29.100 --> 00:41:30.180 align:middle line:90% You can get a credit. 00:41:30.180 --> 00:41:33.600 align:middle line:84% They basically say, you can deduct a percent of your income 00:41:33.600 --> 00:41:37.600 align:middle line:84% up to the maximum of $6,000 for two kids, $3,000 for one kid. 00:41:37.600 --> 00:41:42.400 align:middle line:84% So we have a limited deduction is how we resolve it in reality. 00:41:42.400 --> 00:41:45.320 align:middle line:90% So that's how we think about-- 00:41:45.320 --> 00:41:50.000 align:middle line:84% how we operationalize the abstract concepts of chapter 20, 00:41:50.000 --> 00:41:52.040 align:middle line:84% which is thinking practice about thinking 00:41:52.040 --> 00:41:53.820 align:middle line:90% about what's more elastic. 00:41:53.820 --> 00:41:56.300 align:middle line:84% You want to avoid distorting things that are more elastic, 00:41:56.300 --> 00:41:58.240 align:middle line:90% like women's labor supply. 00:41:58.240 --> 00:42:00.440 align:middle line:84% But you also want to try to have as big a tax base 00:42:00.440 --> 00:42:02.680 align:middle line:84% as is possible, and trading those two off is how we 00:42:02.680 --> 00:42:05.520 align:middle line:90% think about optimal tax design. 00:42:05.520 --> 00:42:06.240 align:middle line:90% Yeah? 00:42:06.240 --> 00:42:07.840 align:middle line:84% AUDIENCE: How elastic-- are there 00:42:07.840 --> 00:42:11.320 align:middle line:84% any policies to which people's decision to have kids or not 00:42:11.320 --> 00:42:12.440 align:middle line:90% is elastic? 00:42:12.440 --> 00:42:14.720 align:middle line:84% PROFESSOR: So the decision to have kids is just like-- 00:42:14.720 --> 00:42:16.845 align:middle line:84% I think you asked the same question about marriage. 00:42:16.845 --> 00:42:19.420 align:middle line:84% Decision to have kids is totally inelastic to everything. 00:42:19.420 --> 00:42:20.660 align:middle line:90% The timing is elastic. 00:42:20.660 --> 00:42:22.240 align:middle line:84% I told you my story of my kid being 00:42:22.240 --> 00:42:24.280 align:middle line:90% born in January 2, my sad tale. 00:42:24.280 --> 00:42:26.263 align:middle line:84% But actually, this is a big issue 00:42:26.263 --> 00:42:28.180 align:middle line:84% because fertility is falling around the world. 00:42:28.180 --> 00:42:30.540 align:middle line:84% A number of countries are really worried about falling fertility. 00:42:30.540 --> 00:42:32.700 align:middle line:84% And there's been nothing they figure out how to do. 00:42:32.700 --> 00:42:34.980 align:middle line:84% Now, there's been tools which lower fertility. 00:42:34.980 --> 00:42:37.412 align:middle line:84% China's One Child Policy did lower fertility. 00:42:37.412 --> 00:42:39.620 align:middle line:84% But you can't really have-- mandate people have kids. 00:42:39.620 --> 00:42:41.320 align:middle line:84% You can mandate they can't have kids. 00:42:41.320 --> 00:42:44.160 align:middle line:84% It's hard to mandate they do have kids. 00:42:44.160 --> 00:42:46.200 align:middle line:84% And other than that, there's really been-- 00:42:46.200 --> 00:42:49.440 align:middle line:84% there's no evidence of any financial tool that basically 00:42:49.440 --> 00:42:50.780 align:middle line:90% promotes people to have kids. 00:42:50.780 --> 00:42:52.520 align:middle line:84% It really is a very inelastic margin 00:42:52.520 --> 00:42:53.733 align:middle line:90% with respect to incentives. 00:42:53.733 --> 00:42:55.400 align:middle line:84% AUDIENCE: Which, for a lot of economies, 00:42:55.400 --> 00:42:57.220 align:middle line:84% is a goal that they would like to achieve? 00:42:57.220 --> 00:42:58.040 align:middle line:84% PROFESSOR: Yeah, they'd like to achieve 00:42:58.040 --> 00:43:00.210 align:middle line:84% and they get big child bonuses and things like that, 00:43:00.210 --> 00:43:02.230 align:middle line:90% and none of it works. 00:43:02.230 --> 00:43:05.250 align:middle line:84% Now, once again, if we think about a tax break, 00:43:05.250 --> 00:43:07.070 align:middle line:84% we have the marginal for marginal parts. 00:43:07.070 --> 00:43:08.630 align:middle line:84% Marginally, it's a bad idea to give 00:43:08.630 --> 00:43:12.090 align:middle line:84% a child tax break because it does induce childbearing. 00:43:12.090 --> 00:43:13.750 align:middle line:84% But inframarginally, it may be you're 00:43:13.750 --> 00:43:15.458 align:middle line:84% attributing to exactly the people we want 00:43:15.458 --> 00:43:17.070 align:middle line:90% to have more money in society. 00:43:17.070 --> 00:43:19.870 align:middle line:84% So I talked about, from an efficiency perspective, 00:43:19.870 --> 00:43:22.110 align:middle line:84% the inframarginal piece is a waste. 00:43:22.110 --> 00:43:24.550 align:middle line:84% But from an equity perspective, maybe we just equitably 00:43:24.550 --> 00:43:26.050 align:middle line:84% want to reward people who have kids, 00:43:26.050 --> 00:43:29.230 align:middle line:84% just out of some notion of what's socially just. 00:43:29.230 --> 00:43:33.070 align:middle line:84% But it's not inducing more children. 00:43:33.070 --> 00:43:35.070 align:middle line:90% That's chapter 21. 00:43:35.070 --> 00:43:37.830 align:middle line:84% Now we're going to move on to chapter 22, which 00:43:37.830 --> 00:43:39.250 align:middle line:90% is a parallel analysis. 00:43:39.250 --> 00:43:41.510 align:middle line:84% We go through the basic theory, and then we'll 00:43:41.510 --> 00:43:43.790 align:middle line:90% talk about applications. 00:43:43.790 --> 00:43:47.030 align:middle line:84% The basic theory for the tax on-- 00:43:47.030 --> 00:43:50.430 align:middle line:84% now, instead of looking at the effect of taxes on labor supply, 00:43:50.430 --> 00:43:53.270 align:middle line:84% we're going to let the effect of taxes on savings. 00:43:53.270 --> 00:43:54.890 align:middle line:84% This is a little more complicated, 00:43:54.890 --> 00:43:56.990 align:middle line:90% but the logic is similar. 00:43:56.990 --> 00:43:58.630 align:middle line:84% The idea is when you save-- we're 00:43:58.630 --> 00:44:01.297 align:middle line:84% going to start in a simple world where the only way you can save 00:44:01.297 --> 00:44:04.750 align:middle line:84% is put money in the bank, simple world, money in the bank, 00:44:04.750 --> 00:44:07.970 align:middle line:84% and earn some interest rate R [INAUDIBLE] I'm sorry, 00:44:07.970 --> 00:44:09.370 align:middle line:90% I. What do you have here? 00:44:09.370 --> 00:44:11.406 align:middle line:90% R or I? 00:44:11.406 --> 00:44:13.590 align:middle line:90% R. 00:44:13.590 --> 00:44:16.830 align:middle line:84% We're going to start with a simple world of two periods. 00:44:16.830 --> 00:44:19.010 align:middle line:84% There's a working period and a retirement period. 00:44:19.010 --> 00:44:20.830 align:middle line:84% This is the classic overlapping generations 00:44:20.830 --> 00:44:21.970 align:middle line:90% model we talked about. 00:44:21.970 --> 00:44:23.910 align:middle line:90% We talked about Social Security. 00:44:23.910 --> 00:44:26.710 align:middle line:84% There's a working period and retirement period. 00:44:26.710 --> 00:44:30.710 align:middle line:84% And people are going to optimize their consumption 00:44:30.710 --> 00:44:32.750 align:middle line:90% bundle across time. 00:44:32.750 --> 00:44:37.150 align:middle line:84% We call this a model of intertemporal substitution. 00:44:37.150 --> 00:44:39.310 align:middle line:84% You're going to substitute your consumption 00:44:39.310 --> 00:44:42.250 align:middle line:84% across different periods, intertemporal substitution. 00:44:42.250 --> 00:44:44.270 align:middle line:84% In this simple case, there's two periods. 00:44:44.270 --> 00:44:46.330 align:middle line:84% Obviously, in reality, there's infinite periods. 00:44:46.330 --> 00:44:47.410 align:middle line:90% There's two periods. 00:44:47.410 --> 00:44:49.270 align:middle line:90% There's working life. 00:44:49.270 --> 00:44:51.510 align:middle line:90% That's the x-axis. 00:44:51.510 --> 00:44:53.290 align:middle line:90% And there's retirement. 00:44:53.290 --> 00:44:54.550 align:middle line:90% That's the y-axis. 00:44:54.550 --> 00:44:59.350 align:middle line:84% So on the x-axis, consumption while working, C super W, 00:44:59.350 --> 00:45:01.310 align:middle line:84% on the y-axis consumption while retired, 00:45:01.310 --> 00:45:04.550 align:middle line:84% C super R. And your question is simply, 00:45:04.550 --> 00:45:08.510 align:middle line:84% how am I going to allocate my consumption across these two? 00:45:08.510 --> 00:45:10.510 align:middle line:84% Furthermore, we're going to assume you only earn 00:45:10.510 --> 00:45:14.190 align:middle line:84% income in period W. So you're going 00:45:14.190 --> 00:45:16.750 align:middle line:84% to need some income Y that you've 00:45:16.750 --> 00:45:20.310 align:middle line:84% got to allocate across the period when you work 00:45:20.310 --> 00:45:23.510 align:middle line:84% and the period when you don't work. 00:45:23.510 --> 00:45:26.630 align:middle line:84% We're also going to say that you can choose 00:45:26.630 --> 00:45:29.190 align:middle line:90% to save some of your money. 00:45:29.190 --> 00:45:32.130 align:middle line:84% When you work, you can choose to save. 00:45:32.130 --> 00:45:39.430 align:middle line:84% So savings is Y minus C, Y minus C super W, 00:45:39.430 --> 00:45:42.190 align:middle line:84% whatever you earn minus what you consume 00:45:42.190 --> 00:45:45.350 align:middle line:90% while working is savings. 00:45:45.350 --> 00:45:50.550 align:middle line:84% The rate of return to savings is the interest rate R. 00:45:50.550 --> 00:45:58.540 align:middle line:84% So what we have here is the blue line in figure 22.1. 00:45:58.540 --> 00:46:02.740 align:middle line:84% You can either consume Y while you're working. 00:46:02.740 --> 00:46:05.620 align:middle line:84% That's the x-intercept of the blue line. 00:46:05.620 --> 00:46:08.780 align:middle line:84% Or you could save it all, consume nothing 00:46:08.780 --> 00:46:11.260 align:middle line:84% while you're working, and have Y times 1 00:46:11.260 --> 00:46:13.460 align:middle line:90% plus R when you're retired. 00:46:13.460 --> 00:46:15.100 align:middle line:84% So either consume it all when you're 00:46:15.100 --> 00:46:16.880 align:middle line:84% working [INAUDIBLE] while you're working, 00:46:16.880 --> 00:46:19.740 align:middle line:84% and nothing while you're retired or save it all, in which case 00:46:19.740 --> 00:46:21.740 align:middle line:84% you get Y times 1 plus R when you're retired 00:46:21.740 --> 00:46:24.980 align:middle line:84% and nothing when you're working or some combination in between. 00:46:24.980 --> 00:46:28.060 align:middle line:84% As a result, the slope of the budget constraint is minus 1 00:46:28.060 --> 00:46:30.020 align:middle line:90% plus R. 00:46:30.020 --> 00:46:35.048 align:middle line:84% In other words, the opportunity cost of consuming in period 1-- 00:46:35.048 --> 00:46:37.340 align:middle line:84% by the way, I'm going to refer to period 1 in working-- 00:46:37.340 --> 00:46:38.800 align:middle line:84% and I'm referring interchangeably. 00:46:38.800 --> 00:46:41.092 align:middle line:84% When I say period 1 or 2, I mean working and retirement 00:46:41.092 --> 00:46:42.217 align:middle line:90% I just slip back and forth. 00:46:42.217 --> 00:46:44.134 align:middle line:84% The bottom line is there's a two-period model. 00:46:44.134 --> 00:46:45.420 align:middle line:90% We've got before and after. 00:46:45.420 --> 00:46:51.540 align:middle line:84% The opportunity cost of consuming while you're working 00:46:51.540 --> 00:46:54.440 align:middle line:84% is that, for every dollar you consume while you're working, 00:46:54.440 --> 00:46:57.820 align:middle line:84% you're giving up 1 plus R dollars consumption when you're 00:46:57.820 --> 00:47:00.540 align:middle line:84% retired because you could have put that away 00:47:00.540 --> 00:47:02.500 align:middle line:90% and saved and earned interest. 00:47:02.500 --> 00:47:04.500 align:middle line:84% This is the key concept here with savings. 00:47:04.500 --> 00:47:06.420 align:middle line:90% It's a little easier with labor. 00:47:06.420 --> 00:47:08.060 align:middle line:84% Either you sit at the couch, or you 00:47:08.060 --> 00:47:09.685 align:middle line:84% go to work and earn W. It's easy to see 00:47:09.685 --> 00:47:11.120 align:middle line:90% W. It's the opportunity cost. 00:47:11.120 --> 00:47:12.460 align:middle line:90% But it's the same logic. 00:47:12.460 --> 00:47:14.660 align:middle line:84% The opportunity cost, if I don't eat it today, 00:47:14.660 --> 00:47:17.380 align:middle line:84% I put it in the bank, and it becomes 1 plus R dollars 00:47:17.380 --> 00:47:18.100 align:middle line:90% tomorrow. 00:47:18.100 --> 00:47:20.960 align:middle line:84% Or when I'm working [INAUDIBLE] dollars when I'm retired. 00:47:20.960 --> 00:47:24.460 align:middle line:84% So that's the opportunity cost of consumption today 00:47:24.460 --> 00:47:28.540 align:middle line:84% is the savings I could have had when I was retired. 00:47:28.540 --> 00:47:34.300 align:middle line:84% Now, imagine the government comes along and taxes interest. 00:47:34.300 --> 00:47:37.100 align:middle line:84% This is what we call capital taxation. 00:47:37.100 --> 00:47:38.640 align:middle line:90% It taxes interest. 00:47:38.640 --> 00:47:39.880 align:middle line:90% What does that do? 00:47:39.880 --> 00:47:44.460 align:middle line:84% That lowers the effective rate of return to saving. 00:47:44.460 --> 00:47:46.460 align:middle line:90% Is your sweatshirt on backwards? 00:47:46.460 --> 00:47:48.140 align:middle line:90% Cool. 00:47:48.140 --> 00:47:51.180 align:middle line:84% So that lowers the effective rate of return to savings. 00:47:51.180 --> 00:47:55.530 align:middle line:84% So basically what that does, it moves from the blue line 00:47:55.530 --> 00:47:57.810 align:middle line:90% to the red line. 00:47:57.810 --> 00:48:02.050 align:middle line:84% You now see the opportunity cost is no longer 1 plus R. It's now 00:48:02.050 --> 00:48:06.110 align:middle line:90% 1 plus R times 1 minus tau. 00:48:06.110 --> 00:48:07.110 align:middle line:90% It's a little confusing. 00:48:07.110 --> 00:48:07.985 align:middle line:90% Think of it this way. 00:48:07.985 --> 00:48:10.535 align:middle line:90% Imagine the tax rate was 100%. 00:48:10.535 --> 00:48:11.910 align:middle line:84% There'd be no slope to this line. 00:48:11.910 --> 00:48:13.330 align:middle line:90% It'd be a flat line. 00:48:13.330 --> 00:48:16.930 align:middle line:84% So basically it's reducing the rate of return 00:48:16.930 --> 00:48:20.410 align:middle line:90% by how much that is taxed. 00:48:20.410 --> 00:48:21.660 align:middle line:90% So the slope pivots downwards. 00:48:21.660 --> 00:48:24.118 align:middle line:84% Once again, if you consume everything while you're working, 00:48:24.118 --> 00:48:25.270 align:middle line:90% there's no change. 00:48:25.270 --> 00:48:29.610 align:middle line:84% You still have C super W equals Y. 00:48:29.610 --> 00:48:32.370 align:middle line:84% But now if you wait until-- if you consume nothing 00:48:32.370 --> 00:48:35.150 align:middle line:84% while you work and consume it all when you retire, 00:48:35.150 --> 00:48:38.930 align:middle line:84% you end up with less when you retire with Y times 1 00:48:38.930 --> 00:48:42.090 align:middle line:90% plus R times 1 minus tau. 00:48:42.090 --> 00:48:44.650 align:middle line:90% Questions about that? 00:48:44.650 --> 00:48:45.245 align:middle line:90% Yeah? 00:48:45.245 --> 00:48:46.870 align:middle line:84% AUDIENCE: [INAUDIBLE] a tax on savings? 00:48:46.870 --> 00:48:48.150 align:middle line:90% PROFESSOR: Tax on interest. 00:48:48.150 --> 00:48:49.197 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:48:49.197 --> 00:48:51.030 align:middle line:84% PROFESSOR: It's not on your stock of wealth. 00:48:51.030 --> 00:48:52.197 align:middle line:90% I'll come back to that next. 00:48:52.197 --> 00:48:54.730 align:middle line:84% It's a tax on the-- it's what's called a capital income 00:48:54.730 --> 00:49:00.570 align:middle line:84% tax, a tax on the flow of income you earn from your capital. 00:49:00.570 --> 00:49:02.490 align:middle line:90% Very important distinction. 00:49:02.490 --> 00:49:04.287 align:middle line:90% I call them savings taxes. 00:49:04.287 --> 00:49:06.370 align:middle line:84% But technically we're talking about capital income 00:49:06.370 --> 00:49:08.410 align:middle line:84% taxes, taxes on the flow of income 00:49:08.410 --> 00:49:10.950 align:middle line:90% from your capital holdings. 00:49:10.950 --> 00:49:13.430 align:middle line:84% Other questions about how this works? 00:49:13.430 --> 00:49:13.930 align:middle line:90% Yeah? 00:49:13.930 --> 00:49:15.770 align:middle line:84% AUDIENCE: That would be experienced as when 00:49:15.770 --> 00:49:17.070 align:middle line:90% you sell this, the difference? 00:49:17.070 --> 00:49:18.310 align:middle line:84% PROFESSOR: No, not when you sell it. 00:49:18.310 --> 00:49:19.310 align:middle line:90% This is a bank account. 00:49:19.310 --> 00:49:20.930 align:middle line:84% Every year you earn interest from the bank, 00:49:20.930 --> 00:49:22.305 align:middle line:84% and it's taxed by the government. 00:49:22.305 --> 00:49:25.530 align:middle line:84% Just like labor income, this is capital income. 00:49:25.530 --> 00:49:27.275 align:middle line:84% We'll come next time to assets you sell. 00:49:27.275 --> 00:49:28.150 align:middle line:90% That's capital gains. 00:49:28.150 --> 00:49:31.030 align:middle line:84% We'll come back to that next time. 00:49:31.030 --> 00:49:33.267 align:middle line:90% Other questions? 00:49:33.267 --> 00:49:35.350 align:middle line:84% This is hard, so don't be afraid to ask questions. 00:49:35.350 --> 00:49:36.770 align:middle line:90% This isn't clear. 00:49:36.770 --> 00:49:42.690 align:middle line:84% Now, what happens to your savings decisions 00:49:42.690 --> 00:49:48.290 align:middle line:90% when we tax interest? 00:49:48.290 --> 00:49:50.280 align:middle line:84% We know the answer is going to be ambiguous. 00:49:50.280 --> 00:49:53.960 align:middle line:84% I've given you a hint, so let's go and explore it in figure 00:49:53.960 --> 00:49:55.680 align:middle line:90% 22-1. 00:49:55.680 --> 00:49:57.900 align:middle line:84% And this is basically the same as for labor. 00:49:57.900 --> 00:50:00.160 align:middle line:90% Its just more complicated. 00:50:00.160 --> 00:50:02.120 align:middle line:90% Here is the key insight. 00:50:02.120 --> 00:50:03.560 align:middle line:90% Remember this. 00:50:03.560 --> 00:50:06.240 align:middle line:90% We are not modeling savings. 00:50:06.240 --> 00:50:08.740 align:middle line:84% We're modeling consumption in period 1. 00:50:08.740 --> 00:50:10.342 align:middle line:90% Remember, savings is a bad. 00:50:10.342 --> 00:50:11.800 align:middle line:84% Savings is something you have to do 00:50:11.800 --> 00:50:13.920 align:middle line:84% to make sure you can survive when you're retired. 00:50:13.920 --> 00:50:15.740 align:middle line:84% The good is consumption while working. 00:50:15.740 --> 00:50:17.120 align:middle line:90% We model this. 00:50:17.120 --> 00:50:18.340 align:middle line:90% This is fixed. 00:50:18.340 --> 00:50:21.040 align:middle line:90% And we compute S as a residual. 00:50:21.040 --> 00:50:24.280 align:middle line:84% So just like we don't model labor, we monitor leisure, 00:50:24.280 --> 00:50:25.620 align:middle line:90% here we don't model savings. 00:50:25.620 --> 00:50:28.760 align:middle line:84% We model consumption while working. 00:50:28.760 --> 00:50:35.400 align:middle line:84% So in the first case, we have a substitution effect and income 00:50:35.400 --> 00:50:35.900 align:middle line:90% effect. 00:50:35.900 --> 00:50:39.880 align:middle line:84% The substitution effect is the price of first period 00:50:39.880 --> 00:50:42.000 align:middle line:90% consumption has fallen. 00:50:42.000 --> 00:50:43.140 align:middle line:90% What do I mean by that? 00:50:43.140 --> 00:50:45.953 align:middle line:84% The opportunity cost of eating now 00:50:45.953 --> 00:50:48.620 align:middle line:84% has fallen because you don't get as much later if you don't eat. 00:50:48.620 --> 00:50:52.240 align:middle line:84% Once you think of tau as 100%, why not eat it all now? 00:50:52.240 --> 00:50:55.880 align:middle line:84% The opportunity cost of eating now has fallen. 00:50:55.880 --> 00:50:59.000 align:middle line:84% So as a result, you consume more today. 00:50:59.000 --> 00:51:01.960 align:middle line:90% That is a substitution effect. 00:51:01.960 --> 00:51:02.460 align:middle line:90% Yeah? 00:51:02.460 --> 00:51:07.680 align:middle line:84% AUDIENCE: [INAUDIBLE] that means you gained no interest but you 00:51:07.680 --> 00:51:08.220 align:middle line:90% still have-- 00:51:08.220 --> 00:51:10.720 align:middle line:84% PROFESSOR: You'd still want to split your income because you 00:51:10.720 --> 00:51:11.640 align:middle line:90% don't want to starve. 00:51:11.640 --> 00:51:14.100 align:middle line:84% AUDIENCE: So it's not like you have nothing in the future. 00:51:14.100 --> 00:51:14.683 align:middle line:90% PROFESSOR: No. 00:51:14.683 --> 00:51:15.540 align:middle line:90% No if tau is-- 00:51:15.540 --> 00:51:17.560 align:middle line:90% it's not negative 100. 00:51:17.560 --> 00:51:19.520 align:middle line:84% Yeah, you've gained nothing by saving. 00:51:19.520 --> 00:51:20.520 align:middle line:84% So it's not like you're necessarily 00:51:20.520 --> 00:51:21.770 align:middle line:90% going to consume it all today. 00:51:21.770 --> 00:51:24.320 align:middle line:84% It's just the incentive to wait has lowered. 00:51:24.320 --> 00:51:24.862 align:middle line:90% Good point. 00:51:24.862 --> 00:51:25.820 align:middle line:90% I wasn't clear on that. 00:51:25.820 --> 00:51:27.800 align:middle line:90% Thank you. 00:51:27.800 --> 00:51:31.280 align:middle line:84% So now on the other hand, you have an income effect. 00:51:31.280 --> 00:51:35.560 align:middle line:84% You are now poorer because every dollar you save yields 00:51:35.560 --> 00:51:39.120 align:middle line:90% less lifetime income. 00:51:39.120 --> 00:51:44.000 align:middle line:84% Since you're poorer, you want less of everything, including 00:51:44.000 --> 00:51:46.160 align:middle line:90% first period consumption. 00:51:46.160 --> 00:51:50.470 align:middle line:84% So that says you consume less in the first period. 00:51:50.470 --> 00:51:53.670 align:middle line:84% In the first case, the substitution effect dominates. 00:51:53.670 --> 00:51:56.110 align:middle line:84% In the second case, the income effect dominates. 00:51:56.110 --> 00:51:58.653 align:middle line:84% Once again, not implausible, in fact. 00:51:58.653 --> 00:52:00.070 align:middle line:84% Here I find it even more plausible 00:52:00.070 --> 00:52:01.550 align:middle line:90% than labor supply case. 00:52:01.550 --> 00:52:03.670 align:middle line:90% Why do we save? 00:52:03.670 --> 00:52:06.350 align:middle line:84% We save because we want to buy something. 00:52:06.350 --> 00:52:07.490 align:middle line:90% We have some goal. 00:52:07.490 --> 00:52:08.390 align:middle line:90% I want a car. 00:52:08.390 --> 00:52:10.590 align:middle line:84% I want to have this much when I retire. 00:52:10.590 --> 00:52:13.270 align:middle line:84% Well, in that model, income effects would dominate. 00:52:13.270 --> 00:52:15.910 align:middle line:84% In a model where I'm saving to meet a goal, 00:52:15.910 --> 00:52:19.595 align:middle line:84% then if the interest rate gets taxed, I'll have to save more. 00:52:19.595 --> 00:52:21.470 align:middle line:84% So it's not at all implausible income effects 00:52:21.470 --> 00:52:22.303 align:middle line:90% could dominate here. 00:52:22.303 --> 00:52:25.108 align:middle line:84% In fact, from my intuition, it might be more likely, 00:52:25.108 --> 00:52:26.650 align:middle line:84% which is, in most parts of the world, 00:52:26.650 --> 00:52:28.483 align:middle line:84% we think substitution effects dominate here. 00:52:28.483 --> 00:52:31.470 align:middle line:84% I think that presumption is more difficult, although we'll still 00:52:31.470 --> 00:52:33.392 align:middle line:90% assume it because we always do. 00:52:33.392 --> 00:52:34.850 align:middle line:84% But it's a little more challenging. 00:52:34.850 --> 00:52:37.670 align:middle line:90% 00:52:37.670 --> 00:52:39.640 align:middle line:84% But now the diagrams are complicated. 00:52:39.640 --> 00:52:40.890 align:middle line:90% Let's go back to the diagrams. 00:52:40.890 --> 00:52:42.230 align:middle line:84% They're complicated because let's look at literally 00:52:42.230 --> 00:52:43.350 align:middle line:90% what happens. 00:52:43.350 --> 00:52:49.590 align:middle line:84% In the first case, you were consuming C super W sub 1 00:52:49.590 --> 00:52:52.110 align:middle line:90% and saving S sub 1. 00:52:52.110 --> 00:52:57.270 align:middle line:84% Now you're consuming C super W sub 2 and saving S sub 2. 00:52:57.270 --> 00:53:01.230 align:middle line:84% Now, we're not modeling period 2 consumption period-- period R 00:53:01.230 --> 00:53:02.010 align:middle line:90% consumption. 00:53:02.010 --> 00:53:03.010 align:middle line:90% But it has implications. 00:53:03.010 --> 00:53:05.590 align:middle line:84% Look at what happens period R consumption. 00:53:05.590 --> 00:53:10.350 align:middle line:84% It falls from S1 times 1 plus R to S2 times 1 plus R times 00:53:10.350 --> 00:53:11.170 align:middle line:90% 1 minus tau. 00:53:11.170 --> 00:53:13.310 align:middle line:90% That is C2-- 00:53:13.310 --> 00:53:18.150 align:middle line:84% I'm sorry, CR falls for two reasons. 00:53:18.150 --> 00:53:21.550 align:middle line:84% First of all, it falls because you're saving less. 00:53:21.550 --> 00:53:23.590 align:middle line:84% Second of all, it falls because even the savings 00:53:23.590 --> 00:53:25.830 align:middle line:90% you have earns less interest. 00:53:25.830 --> 00:53:29.350 align:middle line:84% So that's why, look at the second panel. 00:53:29.350 --> 00:53:32.990 align:middle line:84% In the second panel, you save more because income effects 00:53:32.990 --> 00:53:34.650 align:middle line:90% dominate. 00:53:34.650 --> 00:53:36.990 align:middle line:90% You go from saving S1 to S2. 00:53:36.990 --> 00:53:39.690 align:middle line:84% But look what happens to retirement consumption. 00:53:39.690 --> 00:53:43.250 align:middle line:84% It still falls because there's two effects. 00:53:43.250 --> 00:53:47.420 align:middle line:84% You save more, but it earns a lower interest rate. 00:53:47.420 --> 00:53:52.060 align:middle line:84% This is why model CW-- and let the rest come out 00:53:52.060 --> 00:53:53.100 align:middle line:90% of the equations. 00:53:53.100 --> 00:53:55.220 align:middle line:84% Don't be fucking around with C2 because you're 00:53:55.220 --> 00:53:57.940 align:middle line:84% going to mess yourself up because here we 00:53:57.940 --> 00:54:01.317 align:middle line:84% have a case where, in one case, savings falls. 00:54:01.317 --> 00:54:02.400 align:middle line:90% In other case, it goes up. 00:54:02.400 --> 00:54:04.233 align:middle line:84% But in both cases, second period consumption 00:54:04.233 --> 00:54:06.300 align:middle line:84% falls because there's two effects. 00:54:06.300 --> 00:54:08.300 align:middle line:84% There's the effect of saving, and there's effect 00:54:08.300 --> 00:54:10.175 align:middle line:84% that literally the interest rate's gone down. 00:54:10.175 --> 00:54:11.940 align:middle line:84% So you get less for every dollar you save. 00:54:11.940 --> 00:54:13.605 align:middle line:84% So once again, make your life easy. 00:54:13.605 --> 00:54:15.480 align:middle line:84% This is the choice variable we want to model. 00:54:15.480 --> 00:54:19.200 align:middle line:84% Everything else just falls out of the budget constraint. 00:54:19.200 --> 00:54:20.465 align:middle line:90% Yeah? 00:54:20.465 --> 00:54:24.700 align:middle line:90% AUDIENCE: [INAUDIBLE] effect. 00:54:24.700 --> 00:54:28.300 align:middle line:84% PROFESSOR: So substitution effect is the interest rate, 00:54:28.300 --> 00:54:33.300 align:middle line:84% the after-tax interest, rate R times 1 minus tau 00:54:33.300 --> 00:54:37.528 align:middle line:84% is the price of working period consumption. 00:54:37.528 --> 00:54:39.320 align:middle line:84% Why is it the price of working consumption? 00:54:39.320 --> 00:54:41.900 align:middle line:84% Because the higher it is, the more costly 00:54:41.900 --> 00:54:43.740 align:middle line:90% it is to consume today. 00:54:43.740 --> 00:54:48.940 align:middle line:84% So when we raise taxes, that is a reduction in price. 00:54:48.940 --> 00:54:50.500 align:middle line:84% When you reduce the price of a good, 00:54:50.500 --> 00:54:53.340 align:middle line:84% the substitution effects consume more of it. 00:54:53.340 --> 00:54:55.560 align:middle line:84% So that would say more period 1 consumption. 00:54:55.560 --> 00:54:57.940 align:middle line:90% The income effect if it's-- 00:54:57.940 --> 00:54:59.260 align:middle line:90% so it says more of it. 00:54:59.260 --> 00:55:02.240 align:middle line:84% The income effect is, unlike with goods, which I'm buying, 00:55:02.240 --> 00:55:04.260 align:middle line:90% this is a factor I'm earning. 00:55:04.260 --> 00:55:07.660 align:middle line:90% Now I'm effectively poorer. 00:55:07.660 --> 00:55:08.220 align:middle line:90% Why? 00:55:08.220 --> 00:55:10.040 align:middle line:84% Because if R minus 1 minus T falls, 00:55:10.040 --> 00:55:12.760 align:middle line:84% I have less money in my lifetime budget constraint. 00:55:12.760 --> 00:55:14.280 align:middle line:90% We'll do the math in section. 00:55:14.280 --> 00:55:19.900 align:middle line:84% My lifetime budget constraint is basically is essentially going-- 00:55:19.900 --> 00:55:22.260 align:middle line:84% my lifetime total consumption is going 00:55:22.260 --> 00:55:24.500 align:middle line:90% to turn by how much I save. 00:55:24.500 --> 00:55:27.420 align:middle line:84% The more I save, the more my lifetime total consumption is. 00:55:27.420 --> 00:55:31.300 align:middle line:84% And so my eventually my resources fall. 00:55:31.300 --> 00:55:31.880 align:middle line:90% Yeah? 00:55:31.880 --> 00:55:34.588 align:middle line:84% AUDIENCE: When you're mentioning how people respond to this based 00:55:34.588 --> 00:55:36.500 align:middle line:84% on the [? income ?] distribution effects, 00:55:36.500 --> 00:55:41.450 align:middle line:84% how do we [INAUDIBLE] so I'm imagining a person is facing 00:55:41.450 --> 00:55:44.170 align:middle line:84% this decision, and the ways that they can adjust their income 00:55:44.170 --> 00:55:45.170 align:middle line:90% are-- 00:55:45.170 --> 00:55:47.170 align:middle line:84% or the ways that they can adjust are maybe 00:55:47.170 --> 00:55:49.710 align:middle line:84% extending the amount of time for which they are working-- 00:55:49.710 --> 00:55:52.657 align:middle line:84% PROFESSOR: No, no, but I'm taking Y fixed here. 00:55:52.657 --> 00:55:53.990 align:middle line:90% This is just a savings decision. 00:55:53.990 --> 00:55:54.490 align:middle line:90% Y is fixed. 00:55:54.490 --> 00:55:55.406 align:middle line:90% It's not about income. 00:55:55.406 --> 00:55:56.490 align:middle line:90% That was the last chapter. 00:55:56.490 --> 00:55:58.810 align:middle line:84% This is just about savings, just about how much 00:55:58.810 --> 00:56:00.530 align:middle line:90% you consume with income fixed. 00:56:00.530 --> 00:56:02.310 align:middle line:84% Now, you could imagine a world, say, 00:56:02.310 --> 00:56:05.690 align:middle line:84% on an exam, where we combine these two insights 00:56:05.690 --> 00:56:08.490 align:middle line:84% and allow both income and savings to change. 00:56:08.490 --> 00:56:10.290 align:middle line:84% But for now, we're holding it fixed. 00:56:10.290 --> 00:56:12.485 align:middle line:84% The last chapter is what drives this. 00:56:12.485 --> 00:56:14.110 align:middle line:84% This chapter is about what drives this. 00:56:14.110 --> 00:56:16.650 align:middle line:90% 00:56:16.650 --> 00:56:20.262 align:middle line:84% Now, that's one reason this is more complicated 00:56:20.262 --> 00:56:21.970 align:middle line:84% than labor supply, why I teach it second, 00:56:21.970 --> 00:56:23.770 align:middle line:84% is that graphs are a lot more complicated. 00:56:23.770 --> 00:56:27.450 align:middle line:84% It's also a lot more complicated because the intertemporal choice 00:56:27.450 --> 00:56:31.690 align:middle line:84% model is only one of a number of models that could plausibly 00:56:31.690 --> 00:56:34.010 align:middle line:90% describe savings behavior. 00:56:34.010 --> 00:56:36.810 align:middle line:84% Indeed, many people think it might 00:56:36.810 --> 00:56:41.042 align:middle line:84% be the third most important determinant of savings, 00:56:41.042 --> 00:56:43.250 align:middle line:84% that things like how you respond to the interest rate 00:56:43.250 --> 00:56:44.870 align:middle line:84% might be the third most important, 00:56:44.870 --> 00:56:50.410 align:middle line:84% because there's two other very valuable model of savings. 00:56:50.410 --> 00:56:51.590 align:middle line:90% So there's three models. 00:56:51.590 --> 00:56:54.290 align:middle line:84% There's the intertemporal substitution model, 00:56:54.290 --> 00:56:57.430 align:middle line:90% intertemp substitution model. 00:56:57.430 --> 00:57:00.050 align:middle line:84% That's the classic model we're working with 00:57:00.050 --> 00:57:01.210 align:middle line:90% based on interest rates. 00:57:01.210 --> 00:57:04.610 align:middle line:84% There's also a precautionary savings model. 00:57:04.610 --> 00:57:09.570 align:middle line:84% If you ask people why they save, the number one answer they give 00:57:09.570 --> 00:57:11.430 align:middle line:84% is they're worried about stuff going wrong, 00:57:11.430 --> 00:57:13.130 align:middle line:84% not the [? respond ?] to the interest rate [? 00:57:13.130 --> 00:57:15.547 align:middle line:84% and worry ?] about allocating the consumption over working 00:57:15.547 --> 00:57:16.047 align:middle line:90% retirement. 00:57:16.047 --> 00:57:18.547 align:middle line:84% The main reason is they want to have money in case something 00:57:18.547 --> 00:57:19.330 align:middle line:90% goes wrong. 00:57:19.330 --> 00:57:21.130 align:middle line:90% That's not in this model. 00:57:21.130 --> 00:57:23.850 align:middle line:84% That's what's called the precautionary savings model. 00:57:23.850 --> 00:57:29.730 align:middle line:84% The precautionary savings model would be basically-- 00:57:29.730 --> 00:57:32.170 align:middle line:84% the precautionary savings model would 00:57:32.170 --> 00:57:34.270 align:middle line:84% be one where basically-- think of it this way. 00:57:34.270 --> 00:57:35.353 align:middle line:90% Here's how I think about-- 00:57:35.353 --> 00:57:36.937 align:middle line:84% we're not going to make you write down 00:57:36.937 --> 00:57:39.200 align:middle line:84% the math, this, but here's how to operationalize it. 00:57:39.200 --> 00:57:42.560 align:middle line:84% Imagine a world where people face liquidity constraints. 00:57:42.560 --> 00:57:43.675 align:middle line:90% They can't borrow. 00:57:43.675 --> 00:57:46.300 align:middle line:84% If you could borrow, there's no need for precautionary savings. 00:57:46.300 --> 00:57:49.460 align:middle line:84% If you can costlessly borrow, then you wouldn't need to-- 00:57:49.460 --> 00:57:51.532 align:middle line:84% this is only a world with liquidity constraints. 00:57:51.532 --> 00:57:53.740 align:middle line:84% Otherwise, there's no need for precautionary savings. 00:57:53.740 --> 00:57:55.545 align:middle line:90% But imagine you can't borrow. 00:57:55.545 --> 00:57:56.920 align:middle line:84% Then you're going to want to make 00:57:56.920 --> 00:57:59.905 align:middle line:84% sure you have some savings in case stuff goes wrong. 00:57:59.905 --> 00:58:01.780 align:middle line:84% And so in a world with liquidity constraints, 00:58:01.780 --> 00:58:03.655 align:middle line:84% there will naturally be precautionary savings 00:58:03.655 --> 00:58:07.120 align:middle line:90% as a form of self-insurance. 00:58:07.120 --> 00:58:08.820 align:middle line:90% So that's an alternative model. 00:58:08.820 --> 00:58:11.577 align:middle line:84% And indeed, there's lots of evidence for this. 00:58:11.577 --> 00:58:13.160 align:middle line:84% First of all, there's lots of evidence 00:58:13.160 --> 00:58:16.280 align:middle line:84% that people who are in riskier jobs save more. 00:58:16.280 --> 00:58:19.520 align:middle line:84% Now, that's not super convincing because maybe they're different. 00:58:19.520 --> 00:58:22.720 align:middle line:84% Better evidence is there's evidence 00:58:22.720 --> 00:58:26.160 align:middle line:84% that when the generosity of social insurance goes up, 00:58:26.160 --> 00:58:27.880 align:middle line:90% people save less. 00:58:27.880 --> 00:58:30.120 align:middle line:84% Remember, we talked about this with Social Security. 00:58:30.120 --> 00:58:32.360 align:middle line:90% We see in other contexts too. 00:58:32.360 --> 00:58:36.120 align:middle line:84% For instance, in Taiwan, they put in national health insurance 00:58:36.120 --> 00:58:38.170 align:middle line:84% for some groups and not others, and the ones 00:58:38.170 --> 00:58:39.920 align:middle line:84% that got national health insurers suddenly 00:58:39.920 --> 00:58:42.120 align:middle line:90% started saving a lot less. 00:58:42.120 --> 00:58:44.800 align:middle line:84% Now, that's because they were saving against getting sick. 00:58:44.800 --> 00:58:46.920 align:middle line:84% I found a similar result here in America. 00:58:46.920 --> 00:58:50.320 align:middle line:84% When Medicaid expanded, people saved less and consume more 00:58:50.320 --> 00:58:53.840 align:middle line:84% because they didn't have to save as much against getting sick. 00:58:53.840 --> 00:58:56.800 align:middle line:84% So clearly the precautionary model has some validity. 00:58:56.800 --> 00:58:59.980 align:middle line:84% We'll come back next time to how this changes our conclusions. 00:58:59.980 --> 00:59:02.400 align:middle line:84% But that's one alternative model. 00:59:02.400 --> 00:59:05.000 align:middle line:84% The other alternative model that's very relevant 00:59:05.000 --> 00:59:12.680 align:middle line:84% is the self-control model, the self-control model, 00:59:12.680 --> 00:59:14.680 align:middle line:84% this model which says, it's not really 00:59:14.680 --> 00:59:18.140 align:middle line:90% an affirmative model of savings. 00:59:18.140 --> 00:59:19.680 align:middle line:84% It's kind of a Homer Simpson model 00:59:19.680 --> 00:59:23.820 align:middle line:84% of the world, where basically we just want to consume. 00:59:23.820 --> 00:59:26.532 align:middle line:84% We want to eat what's in front of us. 00:59:26.532 --> 00:59:28.240 align:middle line:84% And the only thing that allows us to save 00:59:28.240 --> 00:59:31.360 align:middle line:84% is imposing self-control to stop ourselves 00:59:31.360 --> 00:59:34.480 align:middle line:84% and that savings is really whatever the motivation. 00:59:34.480 --> 00:59:35.920 align:middle line:84% The driver of savings is how much 00:59:35.920 --> 00:59:37.480 align:middle line:84% self control do I have to make myself 00:59:37.480 --> 00:59:41.760 align:middle line:84% save for either precaution or retirement. 00:59:41.760 --> 00:59:44.800 align:middle line:84% Now, this is very much related to the behavioral economics 00:59:44.800 --> 00:59:46.380 align:middle line:84% insights we talked about in chapter 6 00:59:46.380 --> 00:59:48.320 align:middle line:84% when we talked about smoking, which 00:59:48.320 --> 00:59:50.792 align:middle line:84% is the notion of self-control problems. 00:59:50.792 --> 00:59:53.000 align:middle line:84% And there's lots of evidence that, in savings, people 00:59:53.000 --> 00:59:54.380 align:middle line:90% face self-control problems. 00:59:54.380 --> 00:59:55.720 align:middle line:90% The classic evidence? 00:59:55.720 --> 00:59:58.502 align:middle line:84% Existence of what we call Christmas clubs. 00:59:58.502 --> 00:59:59.960 align:middle line:84% What a Christmas club was the thing 00:59:59.960 --> 01:00:02.520 align:middle line:84% banks used to offer, where you could set aside money 01:00:02.520 --> 01:00:06.480 align:middle line:84% every week in a noninterest-bearing account 01:00:06.480 --> 01:00:08.400 align:middle line:84% or a tiny interest-bearing account 01:00:08.400 --> 01:00:09.785 align:middle line:90% to have money at Christmas time. 01:00:09.785 --> 01:00:11.660 align:middle line:84% Well, why the hell do people need to do that? 01:00:11.660 --> 01:00:13.700 align:middle line:84% They could just save it somewhere else, that money, 01:00:13.700 --> 01:00:14.325 align:middle line:90% Christmas time? 01:00:14.325 --> 01:00:16.840 align:middle line:84% Because they're afraid they wouldn't be able to. 01:00:16.840 --> 01:00:21.040 align:middle line:84% More modern evidence comes to the following fascinating 01:00:21.040 --> 01:00:28.680 align:middle line:84% confluence of facts, which is that most individuals have 01:00:28.680 --> 01:00:31.000 align:middle line:90% virtually nothing in the bank. 01:00:31.000 --> 01:00:34.350 align:middle line:84% Or many individuals have virtually nothing in the bank 01:00:34.350 --> 01:00:37.390 align:middle line:84% but lots of money in illiquid forms, 01:00:37.390 --> 01:00:41.350 align:middle line:84% like housing and pensions, forms they can't get at. 01:00:41.350 --> 01:00:43.270 align:middle line:84% Moreover, it goes further than this. 01:00:43.270 --> 01:00:45.290 align:middle line:84% The modal individual in America or many-- 01:00:45.290 --> 01:00:46.950 align:middle line:84% I don't know if it's the modal, but almost the majority 01:00:46.950 --> 01:00:48.992 align:middle line:84% of people in America have the following [? fee ?] 01:00:48.992 --> 01:00:50.870 align:middle line:90% combination of assets-- 01:00:50.870 --> 01:00:54.893 align:middle line:84% little money in the bank, money in their pensions and housing, 01:00:54.893 --> 01:00:56.310 align:middle line:84% and huge credit card debt on which 01:00:56.310 --> 01:00:58.830 align:middle line:90% they're paying 18% interest. 01:00:58.830 --> 01:01:00.790 align:middle line:90% Think about that for a second. 01:01:00.790 --> 01:01:03.097 align:middle line:90% That makes no sense. 01:01:03.097 --> 01:01:04.930 align:middle line:84% They've got money in the bank, first of all. 01:01:04.930 --> 01:01:07.110 align:middle line:84% They should at least spend the money in the bank. 01:01:07.110 --> 01:01:10.070 align:middle line:84% Moreover, they've got money in a pension, which is earning 6%, 01:01:10.070 --> 01:01:13.510 align:middle line:84% and they have credit card debt at 18% Why are they doing that? 01:01:13.510 --> 01:01:16.570 align:middle line:84% Because it's what we call a commitment device, 01:01:16.570 --> 01:01:18.470 align:middle line:84% like we talked about with cigarettes. 01:01:18.470 --> 01:01:20.705 align:middle line:84% They know that they can't control themselves 01:01:20.705 --> 01:01:22.330 align:middle line:84% from hitting their credit card maximum. 01:01:22.330 --> 01:01:23.985 align:middle line:90% That's just a state of being. 01:01:23.985 --> 01:01:26.110 align:middle line:84% So to set a commitment device, they put their money 01:01:26.110 --> 01:01:29.977 align:middle line:84% where they can't get it, in illiquid assets, 01:01:29.977 --> 01:01:32.310 align:middle line:84% even though they're sacrificing a massive rate of return 01:01:32.310 --> 01:01:33.412 align:middle line:90% differential. 01:01:33.412 --> 01:01:35.870 align:middle line:84% Essentially the price they're willing to pay for commitment 01:01:35.870 --> 01:01:38.470 align:middle line:84% is to hide their money away so they can't get it 01:01:38.470 --> 01:01:39.490 align:middle line:84% because they know, if they took it out, 01:01:39.490 --> 01:01:41.290 align:middle line:84% yeah, they pay their credit card bill off today. 01:01:41.290 --> 01:01:42.707 align:middle line:84% But then they go back up tomorrow, 01:01:42.707 --> 01:01:45.230 align:middle line:84% and they would have just blown through their money. 01:01:45.230 --> 01:01:47.670 align:middle line:84% They're still able to keep some money in the bank 01:01:47.670 --> 01:01:49.310 align:middle line:90% because they have-- 01:01:49.310 --> 01:01:52.953 align:middle line:84% they're able to keep some money in the bank because basically 01:01:52.953 --> 01:01:54.870 align:middle line:84% because of another model, self-control models, 01:01:54.870 --> 01:01:57.110 align:middle line:84% we like to call self-control/general behavioral 01:01:57.110 --> 01:01:58.537 align:middle line:90% models. 01:01:58.537 --> 01:02:00.870 align:middle line:84% Another important behavioral model that can explain this 01:02:00.870 --> 01:02:04.750 align:middle line:84% is a model of what we call mental accounting. 01:02:04.750 --> 01:02:06.610 align:middle line:84% The mental accounting model is fascinating. 01:02:06.610 --> 01:02:09.230 align:middle line:84% This says that people don't think about dollars 01:02:09.230 --> 01:02:11.230 align:middle line:84% fungibly, that dollars in different places 01:02:11.230 --> 01:02:12.430 align:middle line:90% mean different things. 01:02:12.430 --> 01:02:16.430 align:middle line:84% There's a famous story I told in 14.01 about the actors Dustin 01:02:16.430 --> 01:02:17.490 align:middle line:90% Hoffman and Gene Hackman. 01:02:17.490 --> 01:02:19.740 align:middle line:84% Famous old actors were roommates together when they're 01:02:19.740 --> 01:02:21.150 align:middle line:90% acting students and poor. 01:02:21.150 --> 01:02:24.150 align:middle line:84% And Gene Hackman knows Dustin Hoffman 01:02:24.150 --> 01:02:27.310 align:middle line:84% and all these bottles labeled like rent, money, et cetera-- 01:02:27.310 --> 01:02:28.350 align:middle line:90% rent, food, et cetera. 01:02:28.350 --> 01:02:30.310 align:middle line:90% And the food bottle was empty. 01:02:30.310 --> 01:02:33.560 align:middle line:84% So he went to Gene Hackman and said, can I borrow some money? 01:02:33.560 --> 01:02:35.882 align:middle line:84% And Gene Hackman said, yeah, but your rent bottle 01:02:35.882 --> 01:02:37.840 align:middle line:84% is totally full, and your movie bottle is full. 01:02:37.840 --> 01:02:39.298 align:middle line:84% He says, well, I can't touch those. 01:02:39.298 --> 01:02:40.740 align:middle line:90% Those are for renting movies. 01:02:40.740 --> 01:02:42.680 align:middle line:84% And that's a way people think about it. 01:02:42.680 --> 01:02:45.380 align:middle line:84% And that's how people are able to leave some money in the bank 01:02:45.380 --> 01:02:46.580 align:middle line:84% because at least they're like, I won't 01:02:46.580 --> 01:02:47.840 align:middle line:90% touch that for my credit card. 01:02:47.840 --> 01:02:49.757 align:middle line:84% But once it gets too big, they'll dip into it. 01:02:49.757 --> 01:02:52.260 align:middle line:90% 01:02:52.260 --> 01:02:57.140 align:middle line:84% And then maybe the best piece of evidence 01:02:57.140 --> 01:03:00.563 align:middle line:84% I find for these models is a fascinating experiment 01:03:00.563 --> 01:03:02.980 align:middle line:84% or product introduced by the Nobel Prize-winning economist 01:03:02.980 --> 01:03:05.100 align:middle line:90% Richard Thaler. 01:03:05.100 --> 01:03:09.260 align:middle line:84% He went to workers and said, I want to offer you 01:03:09.260 --> 01:03:11.700 align:middle line:90% the foreign savings plan. 01:03:11.700 --> 01:03:13.600 align:middle line:84% It's called the save more tomorrow plan. 01:03:13.600 --> 01:03:15.580 align:middle line:90% Here's how it works. 01:03:15.580 --> 01:03:18.660 align:middle line:84% Next time you get a raise, you will 01:03:18.660 --> 01:03:24.502 align:middle line:84% commit to saving 50% of that raise in a separate account. 01:03:24.502 --> 01:03:25.960 align:middle line:84% It doesn't cost you anything today. 01:03:25.960 --> 01:03:28.002 align:middle line:84% You don't have to give up your consumption today. 01:03:28.002 --> 01:03:30.377 align:middle line:84% It's just money you're going to get in the future anyway. 01:03:30.377 --> 01:03:31.900 align:middle line:84% Just decide to set half of it apart. 01:03:31.900 --> 01:03:34.680 align:middle line:84% Now, in a rational, non self-controlled model, 01:03:34.680 --> 01:03:36.020 align:middle line:90% I would say no. 01:03:36.020 --> 01:03:37.145 align:middle line:90% Why should I commit myself? 01:03:37.145 --> 01:03:39.478 align:middle line:84% What good does [INAUDIBLE] I want to save in the future? 01:03:39.478 --> 01:03:40.700 align:middle line:90% I'll save in the future. 01:03:40.700 --> 01:03:43.460 align:middle line:84% But the model where I know I have trouble saving money, 01:03:43.460 --> 01:03:45.780 align:middle line:84% I might find that attractive because it commits me 01:03:45.780 --> 01:03:49.740 align:middle line:84% at no cost today to having savings in the future. 01:03:49.740 --> 01:03:52.660 align:middle line:90% Well, he offered that plan. 01:03:52.660 --> 01:03:55.900 align:middle line:84% 80% of employees signed up for it, 01:03:55.900 --> 01:04:00.752 align:middle line:84% and their savings rate went from 4% to 14%, 01:04:00.752 --> 01:04:02.460 align:middle line:84% which is exactly consistent with the fact 01:04:02.460 --> 01:04:07.300 align:middle line:84% that people had self-control problems and they couldn't save. 01:04:07.300 --> 01:04:10.507 align:middle line:84% So clearly self-control is a major determinant of savings. 01:04:10.507 --> 01:04:12.340 align:middle line:84% And once again, we'll come back to what that 01:04:12.340 --> 01:04:17.370 align:middle line:90% means next time for policy. 01:04:17.370 --> 01:04:20.860 align:middle line:84% Now, like last time, I want to have 01:04:20.860 --> 01:04:25.685 align:middle line:84% two applications for thinking about how tax affects savings. 01:04:25.685 --> 01:04:27.560 align:middle line:84% Like in chapter 21, we have two applications. 01:04:27.560 --> 01:04:30.010 align:middle line:84% I'll cover one today and one next time. 01:04:30.010 --> 01:04:31.347 align:middle line:90% So basically step back. 01:04:31.347 --> 01:04:33.430 align:middle line:84% Just like the other slide, we have a basic theory, 01:04:33.430 --> 01:04:35.370 align:middle line:90% a basic traditional theory. 01:04:35.370 --> 01:04:38.023 align:middle line:84% Before, it was the model of allocation 01:04:38.023 --> 01:04:39.190 align:middle line:90% across consumption, leisure. 01:04:39.190 --> 01:04:41.090 align:middle line:90% It's the model of substitution. 01:04:41.090 --> 01:04:45.970 align:middle line:84% Unlike with labor supply, we have other very plausible models 01:04:45.970 --> 01:04:48.550 align:middle line:84% that may, in fact, be more important, 01:04:48.550 --> 01:04:51.970 align:middle line:84% those other models And that's the theory setup. 01:04:51.970 --> 01:04:55.650 align:middle line:84% Now let's go to two empirical applications. 01:04:55.650 --> 01:05:01.770 align:middle line:84% The first empirical application is talking about inflation. 01:05:01.770 --> 01:05:08.107 align:middle line:84% Now, in the chart I drew, I used R, the real interest rate. 01:05:08.107 --> 01:05:10.190 align:middle line:84% But in fact, banks don't pay a real interest rate. 01:05:10.190 --> 01:05:12.250 align:middle line:84% They pay a nominal interest rate. 01:05:12.250 --> 01:05:14.930 align:middle line:84% And we typically assume that the nominal interest 01:05:14.930 --> 01:05:18.970 align:middle line:84% rate is the real interest rate plus expected inflation. 01:05:18.970 --> 01:05:22.630 align:middle line:84% So basically if inflation is stable at 3%, 01:05:22.630 --> 01:05:24.130 align:middle line:84% then nominal interest rate will just 01:05:24.130 --> 01:05:25.630 align:middle line:90% be real interest rate plus 3%. 01:05:25.630 --> 01:05:28.770 align:middle line:84% If we expect inflation to go up, the nominal rate will go up. 01:05:28.770 --> 01:05:32.050 align:middle line:84% So that's our model of interest rate determination. 01:05:32.050 --> 01:05:40.730 align:middle line:84% It turns out that, in inflationary environments, 01:05:40.730 --> 01:05:45.090 align:middle line:84% capital taxes have a particularly bad feature. 01:05:45.090 --> 01:05:47.410 align:middle line:84% They have particularly bad feature. 01:05:47.410 --> 01:05:52.450 align:middle line:84% And that's because, when we tax interest, 01:05:52.450 --> 01:05:55.890 align:middle line:84% we tax nominal interest, not real interest. 01:05:55.890 --> 01:05:59.490 align:middle line:84% And this creates an additional wedge between consumption 01:05:59.490 --> 01:06:00.290 align:middle line:90% and savings. 01:06:00.290 --> 01:06:07.890 align:middle line:84% To see that, let's go to table 22.1. 01:06:07.890 --> 01:06:12.710 align:middle line:84% Imagine we're all Marshawn Lynch and all we eat is Skittles. 01:06:12.710 --> 01:06:13.210 align:middle line:90% Nothing? 01:06:13.210 --> 01:06:13.710 align:middle line:90% Nobody? 01:06:13.710 --> 01:06:15.090 align:middle line:90% No Marshawn Lynch fans here? 01:06:15.090 --> 01:06:16.250 align:middle line:90% OK. 01:06:16.250 --> 01:06:20.390 align:middle line:84% So basically, I want to consider three worlds. 01:06:20.390 --> 01:06:24.690 align:middle line:84% In the first world, there's no inflation, 01:06:24.690 --> 01:06:28.600 align:middle line:84% And there's no taxation of interest earnings. 01:06:28.600 --> 01:06:30.760 align:middle line:90% I'm going to save $100. 01:06:30.760 --> 01:06:32.680 align:middle line:90% I want to save $100. 01:06:32.680 --> 01:06:35.560 align:middle line:90% And the interest rate is 10% 01:06:35.560 --> 01:06:38.620 align:middle line:84% So in this world, in the first row, I saved $100 at 10%. 01:06:38.620 --> 01:06:41.640 align:middle line:90% I get $10. 01:06:41.640 --> 01:06:45.240 align:middle line:84% After tax of $110, skittles are $1, 01:06:45.240 --> 01:06:48.900 align:middle line:90% so I buy 110 bags of Skittles. 01:06:48.900 --> 01:06:51.065 align:middle line:90% People understand the setup? 01:06:51.065 --> 01:06:52.940 align:middle line:84% Now imagine the government comes in and says, 01:06:52.940 --> 01:06:55.960 align:middle line:84% no, I'm going to tax your interest at 50%. 01:06:55.960 --> 01:07:00.010 align:middle line:84% Well, now that $10 of savings only becomes 01:07:00.010 --> 01:07:02.560 align:middle line:90% $5 after tax return. 01:07:02.560 --> 01:07:04.840 align:middle line:84% That $10 of interest only becomes $5. 01:07:04.840 --> 01:07:08.400 align:middle line:84% So now I have $105 after tax, and I only 01:07:08.400 --> 01:07:10.640 align:middle line:90% buy 105 bags of Skittles. 01:07:10.640 --> 01:07:14.720 align:middle line:84% So what taxation has done is cause me to-- 01:07:14.720 --> 01:07:19.680 align:middle line:84% has distorted my consumption so that I can only consume 105 out 01:07:19.680 --> 01:07:22.680 align:middle line:84% of 110 bags of Skittles if I save, which may make me 01:07:22.680 --> 01:07:25.000 align:middle line:90% want to cause to save less. 01:07:25.000 --> 01:07:26.780 align:middle line:84% So it's distorted my consumption. 01:07:26.780 --> 01:07:28.300 align:middle line:84% That could have income and a substitution effect. 01:07:28.300 --> 01:07:29.383 align:middle line:90% We know which way it goes. 01:07:29.383 --> 01:07:33.280 align:middle line:84% But that illustrates the distortion of capital taxation. 01:07:33.280 --> 01:07:36.800 align:middle line:84% Now, imagine a world where inflation comes along, 01:07:36.800 --> 01:07:39.120 align:middle line:84% and the interest rate doesn't adjust. 01:07:39.120 --> 01:07:41.965 align:middle line:84% Imagine there's now 10% inflation, 01:07:41.965 --> 01:07:43.340 align:middle line:84% and interest rate doesn't adjust. 01:07:43.340 --> 01:07:44.660 align:middle line:90% It stays at 10%. 01:07:44.660 --> 01:07:48.520 align:middle line:84% This is an implausible world, but it's a useful middle step. 01:07:48.520 --> 01:07:51.280 align:middle line:90% Well, now I saved my 100. 01:07:51.280 --> 01:07:54.160 align:middle line:90% I earn $10. 01:07:54.160 --> 01:07:57.080 align:middle line:90% In row 3, I keep the 110. 01:07:57.080 --> 01:07:58.740 align:middle line:90% But there's 10% inflation. 01:07:58.740 --> 01:08:00.380 align:middle line:90% Skittles have gone to $1.10. 01:08:00.380 --> 01:08:03.480 align:middle line:84% So you only get 100 bags of Skittles. 01:08:03.480 --> 01:08:05.640 align:middle line:84% And with interest taxation, that 10, 01:08:05.640 --> 01:08:11.400 align:middle line:84% I only get $105, so only get 95.5 bags of Skittles. 01:08:11.400 --> 01:08:14.000 align:middle line:84% So basically in the first row, I end up 01:08:14.000 --> 01:08:17.479 align:middle line:84% with fewer bags of Skittles because inflation basically 01:08:17.479 --> 01:08:18.439 align:middle line:90% offset my interest. 01:08:18.439 --> 01:08:20.522 align:middle line:84% In the second row, I ended up with even fewer bags 01:08:20.522 --> 01:08:21.060 align:middle line:90% of Skittles. 01:08:21.060 --> 01:08:24.050 align:middle line:90% 01:08:24.050 --> 01:08:25.350 align:middle line:90% People clear on that? 01:08:25.350 --> 01:08:27.750 align:middle line:84% Now let's go to the realistic case. 01:08:27.750 --> 01:08:30.790 align:middle line:84% The realistic case is one where the real interest 01:08:30.790 --> 01:08:36.750 align:middle line:84% rate adjusts so that now the real interest rate is 20%. 01:08:36.750 --> 01:08:38.050 align:middle line:90% There's a 10% inflation. 01:08:38.050 --> 01:08:39.470 align:middle line:90% Interest rate's 20%. 01:08:39.470 --> 01:08:42.229 align:middle line:90% So look at the next to last row. 01:08:42.229 --> 01:08:44.510 align:middle line:90% Inflation is 10%. 01:08:44.510 --> 01:08:48.550 align:middle line:84% I'm sorry, it's 21% because it's multiplicative. 01:08:48.550 --> 01:08:50.590 align:middle line:90% I save $100. 01:08:50.590 --> 01:08:53.109 align:middle line:84% I end up with a nominal rate of 21% 01:08:53.109 --> 01:08:55.370 align:middle line:90% because it's not really plus pi. 01:08:55.370 --> 01:08:56.770 align:middle line:90% It's times 1 plus pi. 01:08:56.770 --> 01:08:58.729 align:middle line:90% So that's why it's 21%. 01:08:58.729 --> 01:09:01.630 align:middle line:84% But we can always round and just say plus. 01:09:01.630 --> 01:09:05.830 align:middle line:90% So I end up with $121. 01:09:05.830 --> 01:09:07.890 align:middle line:90% The price of Skittles is $1.10. 01:09:07.890 --> 01:09:08.950 align:middle line:90% So guess what? 01:09:08.950 --> 01:09:12.910 align:middle line:84% I'm the same as the first row, which is what should happen. 01:09:12.910 --> 01:09:15.750 align:middle line:84% Absent taxation, inflation should not 01:09:15.750 --> 01:09:18.350 align:middle line:84% affect how many Skittles I can buy. 01:09:18.350 --> 01:09:18.990 align:middle line:90% Why? 01:09:18.990 --> 01:09:23.189 align:middle line:84% Because the interest rate goes up to offset, goes up to offset. 01:09:23.189 --> 01:09:26.310 align:middle line:84% The price change, I'm indifferent. 01:09:26.310 --> 01:09:31.350 align:middle line:84% But now look what happens when there's taxation. 01:09:31.350 --> 01:09:33.830 align:middle line:84% There's a 50% tax rate on interest. 01:09:33.830 --> 01:09:36.109 align:middle line:90% My savings is 100. 01:09:36.109 --> 01:09:40.430 align:middle line:84% I earn $21, but half of that is taxed away. 01:09:40.430 --> 01:09:43.830 align:middle line:90% So after tax, I have 110.5. 01:09:43.830 --> 01:09:46.710 align:middle line:84% Skittles are $1.10, so I can only afford 100 bags 01:09:46.710 --> 01:09:49.430 align:middle line:90% of Skittles, plus half a bag. 01:09:49.430 --> 01:09:53.350 align:middle line:84% So look at the difference between rows 1 and 2 and rows 5 01:09:53.350 --> 01:09:55.550 align:middle line:90% and 6. 01:09:55.550 --> 01:09:59.705 align:middle line:84% In rows 1 and 2 without inflation, 01:09:59.705 --> 01:10:01.830 align:middle line:84% capital taxation has reduced the number of Skittles 01:10:01.830 --> 01:10:04.150 align:middle line:90% I can eat by 5%. 01:10:04.150 --> 01:10:07.613 align:middle line:84% In rows 5 and 6 with inflation, capital taxation 01:10:07.613 --> 01:10:09.030 align:middle line:84% has reduced the number of Skittles 01:10:09.030 --> 01:10:11.750 align:middle line:90% I can eat by almost 10%. 01:10:11.750 --> 01:10:13.230 align:middle line:90% What's going on? 01:10:13.230 --> 01:10:16.790 align:middle line:84% What's going on is there's another wedge. 01:10:16.790 --> 01:10:21.380 align:middle line:84% And the wedge here is that we tax nominal interest rates, not 01:10:21.380 --> 01:10:22.980 align:middle line:90% real interest rates. 01:10:22.980 --> 01:10:25.602 align:middle line:90% So when inflation comes along-- 01:10:25.602 --> 01:10:27.060 align:middle line:84% and think of it as an extreme case. 01:10:27.060 --> 01:10:29.620 align:middle line:84% If there's massive inflation, then 01:10:29.620 --> 01:10:31.580 align:middle line:84% even if real interest rates adjust, 01:10:31.580 --> 01:10:34.738 align:middle line:84% I'm still not going to save because basically I'm 01:10:34.738 --> 01:10:36.280 align:middle line:84% going to-- when interest rates go up, 01:10:36.280 --> 01:10:39.220 align:middle line:84% I'm going to lose a lot of that through taxation, 01:10:39.220 --> 01:10:41.820 align:middle line:84% even though, in fact, it should be like row 5. 01:10:41.820 --> 01:10:44.060 align:middle line:84% I should be indifferent to inflation. 01:10:44.060 --> 01:10:46.300 align:middle line:84% But I'm not because it's a wedge created by the fact 01:10:46.300 --> 01:10:50.180 align:middle line:84% that we only tax nominal interest, not real interest. 01:10:50.180 --> 01:10:52.980 align:middle line:84% Now, this point was made in the 1970s by the famous economist 01:10:52.980 --> 01:10:54.960 align:middle line:84% Martin Feldstein, really, if you will, 01:10:54.960 --> 01:10:57.580 align:middle line:84% the father of empirical public finance. 01:10:57.580 --> 01:11:00.040 align:middle line:90% This was a very important point. 01:11:00.040 --> 01:11:02.620 align:middle line:84% And then we forgot about it starting about the mid-1980s. 01:11:02.620 --> 01:11:03.120 align:middle line:90% Why? 01:11:03.120 --> 01:11:04.348 align:middle line:90% Because inflation went away. 01:11:04.348 --> 01:11:06.640 align:middle line:84% So every year I teach this and be like, yeah, whatever. 01:11:06.640 --> 01:11:07.560 align:middle line:90% There's no inflation. 01:11:07.560 --> 01:11:08.927 align:middle line:90% Well, guess what? 01:11:08.927 --> 01:11:09.760 align:middle line:90% There was inflation. 01:11:09.760 --> 01:11:11.600 align:middle line:84% And now you understand why something like this matters. 01:11:11.600 --> 01:11:13.780 align:middle line:84% Now, fortunately, we're back to low inflation. 01:11:13.780 --> 01:11:17.380 align:middle line:84% But you could see we were in a world with almost 10% inflation. 01:11:17.380 --> 01:11:22.660 align:middle line:84% And you could see in that world the fact that capital taxes are 01:11:22.660 --> 01:11:26.700 align:middle line:84% more distortionary in a high inflation environment 01:11:26.700 --> 01:11:31.260 align:middle line:84% because we tax nominal interest, not real interest. 01:11:31.260 --> 01:11:36.740 align:middle line:84% Now, in fact, we could fix this by we could just simply index 01:11:36.740 --> 01:11:38.140 align:middle line:90% capital taxation. 01:11:38.140 --> 01:11:40.120 align:middle line:84% Indeed, we've done this otherwise. 01:11:40.120 --> 01:11:44.980 align:middle line:84% When I was a kid, back in the day, in the 1970s, 01:11:44.980 --> 01:11:47.620 align:middle line:84% the whole tax system used to have this problem, 01:11:47.620 --> 01:11:52.500 align:middle line:84% which is that basically the tax brackets were fixed, 01:11:52.500 --> 01:11:53.560 align:middle line:90% but your income was not. 01:11:53.560 --> 01:11:56.260 align:middle line:84% So let's say inflation went up, and my boss gave me a raise. 01:11:56.260 --> 01:11:58.272 align:middle line:84% That could move me to a higher tax bracket. 01:11:58.272 --> 01:11:59.980 align:middle line:84% And I could pay more taxes, even though I 01:11:59.980 --> 01:12:01.780 align:middle line:90% didn't have any more resources. 01:12:01.780 --> 01:12:03.620 align:middle line:84% So imagine where prices went up 20%, 01:12:03.620 --> 01:12:07.300 align:middle line:84% as they were going up in the 1970s, or 15%. 01:12:07.300 --> 01:12:10.220 align:middle line:90% I get a 15% wage increase. 01:12:10.220 --> 01:12:12.580 align:middle line:90% I can't buy any more goods. 01:12:12.580 --> 01:12:15.660 align:middle line:84% But my taxes have gone up because the tax 01:12:15.660 --> 01:12:18.450 align:middle line:84% bracket was fixed, so I'm bumped into higher tax bracket. 01:12:18.450 --> 01:12:20.490 align:middle line:90% That was called bracket creep. 01:12:20.490 --> 01:12:22.970 align:middle line:90% We fixed that in 1979. 01:12:22.970 --> 01:12:24.990 align:middle line:90% We indexed the tax brackets. 01:12:24.990 --> 01:12:29.038 align:middle line:84% So the tax brackets themselves go up with inflation. 01:12:29.038 --> 01:12:31.330 align:middle line:84% We could do the same thing with capital taxation, which 01:12:31.330 --> 01:12:35.490 align:middle line:84% is we could index what you pay to what happens to inflation. 01:12:35.490 --> 01:12:36.850 align:middle line:90% We don't. 01:12:36.850 --> 01:12:42.650 align:middle line:84% Part of the reason we don't is that, who would that help? 01:12:42.650 --> 01:12:44.870 align:middle line:84% The wealthiest, the ones who save the most. 01:12:44.870 --> 01:12:45.810 align:middle line:90% So that'd be no. 01:12:45.810 --> 01:12:47.602 align:middle line:84% We do lots of things that help the wealthy, 01:12:47.602 --> 01:12:49.250 align:middle line:84% and maybe this Congress will do that. 01:12:49.250 --> 01:12:53.850 align:middle line:84% But traditionally, that's been viewed as a wonky fix that 01:12:53.850 --> 01:12:54.910 align:middle line:90% plays bad politically. 01:12:54.910 --> 01:12:56.730 align:middle line:90% So it's not really been done. 01:12:56.730 --> 01:13:03.970 align:middle line:84% But in principle, you could solve this by indexing taxation 01:13:03.970 --> 01:13:07.490 align:middle line:84% to account for inflation, just like we did with brackets. 01:13:07.490 --> 01:13:09.250 align:middle line:90% Questions about that? 01:13:09.250 --> 01:13:09.750 align:middle line:90% All right. 01:13:09.750 --> 01:13:12.680 align:middle line:90% If not, let's stop here. 01:13:12.680 --> 01:13:17.000 align:middle line:90%