1 00:00:00,000 --> 00:00:01,455 [SQUEAKING] 2 00:00:01,455 --> 00:00:04,365 [RUSTLING] 3 00:00:04,365 --> 00:00:09,357 [CLICKING] 4 00:00:09,357 --> 00:00:10,940 JONATHAN GRUBER: So today, we're going 5 00:00:10,940 --> 00:00:16,200 to talk about social insurance. 6 00:00:16,200 --> 00:00:20,990 So why do we have this thing called "social insurance?" 7 00:00:20,990 --> 00:00:24,740 Let's first talk about what social insurance is, and then 8 00:00:24,740 --> 00:00:27,650 ask why we have it. 9 00:00:27,650 --> 00:00:31,040 So basically, social insurance is 10 00:00:31,040 --> 00:00:34,700 government-provided insurance programs. 11 00:00:34,700 --> 00:00:36,950 This is the largest single category 12 00:00:36,950 --> 00:00:39,680 of government expenditure in the US today, 13 00:00:39,680 --> 00:00:42,770 is government-provided insurance programs. 14 00:00:42,770 --> 00:00:44,142 Now, why do we have these? 15 00:00:44,142 --> 00:00:46,100 You might say, well, we know why we have these. 16 00:00:46,100 --> 00:00:47,870 We learned about uncertainty. 17 00:00:47,870 --> 00:00:51,310 We already talked about how people dislike uncertainty, 18 00:00:51,310 --> 00:00:52,970 and about how as a result, insurance 19 00:00:52,970 --> 00:00:55,070 is big business in America. 20 00:00:55,070 --> 00:00:58,640 Private insurance for health, for auto, 21 00:00:58,640 --> 00:01:00,650 for life, for property and casualty 22 00:01:00,650 --> 00:01:03,380 adds up to about $1.5 trillion every year. 23 00:01:06,020 --> 00:01:09,090 So we already have big business of private insurance. 24 00:01:09,090 --> 00:01:11,220 So why does the government need to get involved? 25 00:01:11,220 --> 00:01:13,280 I mean, after all, people want insurance 26 00:01:13,280 --> 00:01:14,730 if they're risk averse. 27 00:01:14,730 --> 00:01:17,450 We talked about insurance markets can work. 28 00:01:17,450 --> 00:01:20,970 And as always in economics, the question is, what's 29 00:01:20,970 --> 00:01:22,260 wrong with the private market? 30 00:01:22,260 --> 00:01:25,605 What is the market failure that might generate interest 31 00:01:25,605 --> 00:01:28,280 in government being involved? 32 00:01:28,280 --> 00:01:32,218 And the market failure in the context of insurance 33 00:01:32,218 --> 00:01:33,760 is a different kind of market failure 34 00:01:33,760 --> 00:01:34,520 than we've talked about. 35 00:01:34,520 --> 00:01:35,937 We've talked about market failures 36 00:01:35,937 --> 00:01:39,050 like imperfect competition as a market failure. 37 00:01:39,050 --> 00:01:41,473 We talked about externalities as a market failure. 38 00:01:41,473 --> 00:01:43,890 The new kind of market failure we want to talk about today 39 00:01:43,890 --> 00:01:47,690 is what we call "information asymmetry." 40 00:01:53,750 --> 00:02:00,770 Information asymmetry, basically which 41 00:02:00,770 --> 00:02:03,980 is the difference in the information 42 00:02:03,980 --> 00:02:09,479 available to sellers and to buyers in a given market. 43 00:02:09,479 --> 00:02:12,150 So far in this course, we sort of assumed full information. 44 00:02:12,150 --> 00:02:15,220 We've assumed everybody knows everything. 45 00:02:15,220 --> 00:02:18,340 We weaken that assumption a little bit with uncertainty, 46 00:02:18,340 --> 00:02:20,530 and say that people don't know whether they're 47 00:02:20,530 --> 00:02:22,322 going to be sick or healthy, but they still 48 00:02:22,322 --> 00:02:23,740 knew the probabilities. 49 00:02:23,740 --> 00:02:26,025 Now we weaken it further by saying not only is 50 00:02:26,025 --> 00:02:28,600 the information imperfect, but different parties 51 00:02:28,600 --> 00:02:30,700 in a transaction might have different levels 52 00:02:30,700 --> 00:02:32,960 of information. 53 00:02:32,960 --> 00:02:37,390 And that's going to turn out to cause market failure. 54 00:02:37,390 --> 00:02:41,123 That information asymmetry will cause market failure. 55 00:02:41,123 --> 00:02:42,790 Now, the math here is quite hard, harder 56 00:02:42,790 --> 00:02:44,450 than we do in this class. 57 00:02:44,450 --> 00:02:47,710 So we'll just sort of do this by an example or two. 58 00:02:47,710 --> 00:02:50,500 And the best example to start with 59 00:02:50,500 --> 00:02:54,790 is the so-called "lemons problem" 60 00:02:54,790 --> 00:02:57,760 that was laid out by the Nobel Prize-winning economist George 61 00:02:57,760 --> 00:03:02,590 Akerlof in 1970, Nobel Prize-winning economist 62 00:03:02,590 --> 00:03:05,560 and husband of former Fed Chairman Janet Yellen-- 63 00:03:05,560 --> 00:03:08,150 quite a power couple. 64 00:03:08,150 --> 00:03:11,660 And here's what Akerlof laid out-- he said, 65 00:03:11,660 --> 00:03:13,330 let's look at the market for used cars. 66 00:03:13,330 --> 00:03:15,710 This is the market for used cars as of 1970. 67 00:03:15,710 --> 00:03:17,380 There's no CarFax. 68 00:03:17,380 --> 00:03:19,120 There was none of this information. 69 00:03:19,120 --> 00:03:20,340 In 1970, when you went to buy a used car, 70 00:03:20,340 --> 00:03:21,965 you sort of went, and kicked the tires, 71 00:03:21,965 --> 00:03:24,580 and decided if you're going to buy it. 72 00:03:24,580 --> 00:03:29,170 So this is a classic case of an information asymmetry, 73 00:03:29,170 --> 00:03:33,760 in that someone selling a car knows what's wrong with it, 74 00:03:33,760 --> 00:03:36,530 whereas the person buying the car doesn't. 75 00:03:36,530 --> 00:03:40,300 I'll go sort of kick the tires and hope for the best. 76 00:03:40,300 --> 00:03:47,030 So basically, in particular, sellers of cars 77 00:03:47,030 --> 00:03:49,250 might be selling them because they're not good. 78 00:03:49,250 --> 00:03:51,230 After all, why sell a car? 79 00:03:51,230 --> 00:03:53,375 Maybe because it's what's called a "lemon." 80 00:03:53,375 --> 00:03:57,080 A lemon is something which is a poorly performing product, 81 00:03:57,080 --> 00:04:01,280 in this case a car that's got something wrong with it. 82 00:04:01,280 --> 00:04:03,470 So when you go to buy a car, you're worried. 83 00:04:03,470 --> 00:04:05,840 You want to buy a car in the used car market, 84 00:04:05,840 --> 00:04:06,860 but you're worried. 85 00:04:06,860 --> 00:04:08,390 Why is someone selling this car? 86 00:04:08,390 --> 00:04:10,000 You should be. 87 00:04:10,000 --> 00:04:11,440 Why is someone selling this car? 88 00:04:11,440 --> 00:04:14,420 If it's really in good shape, why would they be selling it? 89 00:04:14,420 --> 00:04:17,300 Therefore, as a result, Akerlof argued, 90 00:04:17,300 --> 00:04:19,970 there might be a collapse of the entire used car market. 91 00:04:19,970 --> 00:04:21,470 There might not be transactions that 92 00:04:21,470 --> 00:04:24,440 happen that can make both parties better off. 93 00:04:24,440 --> 00:04:29,210 Remember, a market failure is whenever the private market 94 00:04:29,210 --> 00:04:31,600 fails to maximize welfare. 95 00:04:31,600 --> 00:04:33,930 What that means is a market failure 96 00:04:33,930 --> 00:04:37,150 arises whatever the private market does not 97 00:04:37,150 --> 00:04:41,530 deliver all transactions that make buyer and seller better 98 00:04:41,530 --> 00:04:44,020 off. 99 00:04:44,020 --> 00:04:46,810 And so let's look at an example of this. 100 00:04:46,810 --> 00:04:49,900 Suppose that I have a 10-year-old car, 101 00:04:49,900 --> 00:04:51,880 and I keep it in pristine shape. 102 00:04:51,880 --> 00:04:53,130 Hint-- that's not true for me. 103 00:04:53,130 --> 00:04:55,255 I'm terrible at cars, but imagine I was someone who 104 00:04:55,255 --> 00:04:56,380 wasn't. 105 00:04:56,380 --> 00:04:58,570 I kept my 10-year-old car in pristine shape. 106 00:04:58,570 --> 00:05:04,720 And let's say that a 10-year-old car in pristine shape is-- 107 00:05:04,720 --> 00:05:07,580 and let's say I'm trying to sell this car. 108 00:05:07,580 --> 00:05:10,670 And let's say that I would happily 109 00:05:10,670 --> 00:05:15,410 take $5,000 for this 10-year-old car that's in pristine shape. 110 00:05:15,410 --> 00:05:22,220 So I would be willing to sell at $5 K. And let's say that you-- 111 00:05:22,220 --> 00:05:25,190 let's say Patricia-- needs that used car, 112 00:05:25,190 --> 00:05:32,830 and she is willing to buy a car that's in good shape for $6 K. 113 00:05:32,830 --> 00:05:40,360 So my willingness to provide, willingness to supply, is $5 K. 114 00:05:40,360 --> 00:05:43,600 Her willingness to pay is $6 K. 115 00:05:43,600 --> 00:05:46,700 So that is a transaction that should happen. 116 00:05:46,700 --> 00:05:50,030 Given the quality of my car, given that's in good shape, 117 00:05:50,030 --> 00:05:52,190 she is willing to pay $1,000 more than I'm 118 00:05:52,190 --> 00:05:53,660 willing to sell it for. 119 00:05:53,660 --> 00:05:55,210 So that transaction should happen, 120 00:05:55,210 --> 00:05:57,670 and it would be welfare maximizing. 121 00:05:57,670 --> 00:06:01,360 But let's say that most 10-year-old cars are not 122 00:06:01,360 --> 00:06:03,340 in good shape. 123 00:06:03,340 --> 00:06:07,270 Most 10-year-old cars, in fact, are in kind of crappy shape. 124 00:06:07,270 --> 00:06:11,020 And in fact, for the typical 10-year-old car, to get it up 125 00:06:11,020 --> 00:06:12,610 and running well, you'd have to throw 126 00:06:12,610 --> 00:06:15,810 $2,000 in once you bought it. 127 00:06:15,810 --> 00:06:17,580 And Patricia knows this. 128 00:06:17,580 --> 00:06:19,950 She knows that for the typical 10-year-old car, 129 00:06:19,950 --> 00:06:22,840 she would have to put $2,000 in. 130 00:06:22,840 --> 00:06:25,620 So her willingness to pay is not $6 K, 131 00:06:25,620 --> 00:06:30,512 it's $4 K for an average 10-year-old car. 132 00:06:30,512 --> 00:06:31,970 Now I say to Patricia, well, that's 133 00:06:31,970 --> 00:06:33,440 an average 10-year-old car, but I 134 00:06:33,440 --> 00:06:35,130 have a perfect 10-year-old car. 135 00:06:35,130 --> 00:06:36,680 You don't need to put $2 K into it. 136 00:06:36,680 --> 00:06:38,443 It's good to go. 137 00:06:38,443 --> 00:06:40,610 So why don't we split the difference and pay $5.5 K. 138 00:06:40,610 --> 00:06:41,740 She says, no way. 139 00:06:41,740 --> 00:06:43,430 You're a damn liar. 140 00:06:43,430 --> 00:06:46,130 I have no way of knowing your car is better than average. 141 00:06:46,130 --> 00:06:47,840 All I know is the average 10-year-old car 142 00:06:47,840 --> 00:06:49,410 needs $2,000 of work. 143 00:06:49,410 --> 00:06:52,940 So I'm not going to pay more than $4,000 for your car. 144 00:06:52,940 --> 00:06:56,230 As a result, Patricia doesn't buy my car. 145 00:06:56,230 --> 00:06:58,750 And a transaction that would have made both parties better 146 00:06:58,750 --> 00:07:00,790 off does not happen. 147 00:07:00,790 --> 00:07:02,925 A transaction where there was full information, 148 00:07:02,925 --> 00:07:04,300 like we have much more of today-- 149 00:07:04,300 --> 00:07:06,370 we can get the entire record of the car, 150 00:07:06,370 --> 00:07:09,220 all the crashes its been in, how much care it's taken care of-- 151 00:07:09,220 --> 00:07:12,220 that problem would go away, should go away, 152 00:07:12,220 --> 00:07:14,410 because now, Patricia could look at my CarFax 153 00:07:14,410 --> 00:07:16,900 and note this, in fact, is a pristine car. 154 00:07:16,900 --> 00:07:19,760 And she's more willing to pay the $6,000 for it. 155 00:07:19,760 --> 00:07:23,110 But the bottom line is, in this world of 1970, 156 00:07:23,110 --> 00:07:27,010 this was a market failure, because a transaction that 157 00:07:27,010 --> 00:07:30,040 made both parties better off did not happen because 158 00:07:30,040 --> 00:07:33,860 of imperfect information. 159 00:07:33,860 --> 00:07:35,980 The buyer was perfectly happy to buy. 160 00:07:35,980 --> 00:07:38,180 Patricia is perfectly happy to buy my car, 161 00:07:38,180 --> 00:07:40,880 but because I had information that she didn't and she 162 00:07:40,880 --> 00:07:44,480 was just suspicious that I was lying, as a result, 163 00:07:44,480 --> 00:07:46,930 that transaction didn't happen. 164 00:07:46,930 --> 00:07:49,810 Now, questions about that? 165 00:07:49,810 --> 00:07:52,920 People understand it's a market failure. 166 00:07:52,920 --> 00:07:55,220 Now we come to insurance. 167 00:07:55,220 --> 00:07:57,790 The story is flipped. 168 00:07:57,790 --> 00:08:01,383 Now it's not the seller that has the information. 169 00:08:01,383 --> 00:08:03,050 It's the buyer that has the information. 170 00:08:03,050 --> 00:08:05,600 In particular, when you buy insurance, 171 00:08:05,600 --> 00:08:08,030 you know how healthy you are. 172 00:08:08,030 --> 00:08:09,770 You know your genetic history. 173 00:08:09,770 --> 00:08:11,665 You know whether you're a risk taker. 174 00:08:11,665 --> 00:08:13,790 You know whether you're around a lot of snotty kids 175 00:08:13,790 --> 00:08:15,147 who might get you sick. 176 00:08:15,147 --> 00:08:16,730 You know a lot of stuff about yourself 177 00:08:16,730 --> 00:08:19,640 that the insurer doesn't know. 178 00:08:19,640 --> 00:08:23,000 As a result, the information asymmetry is flipped. 179 00:08:23,000 --> 00:08:25,942 With insurance, the insurer is worried 180 00:08:25,942 --> 00:08:27,650 that when you come looking for insurance, 181 00:08:27,650 --> 00:08:28,910 they're worried you're looking for insurance because you're 182 00:08:28,910 --> 00:08:29,923 sick. 183 00:08:29,923 --> 00:08:31,340 You might be looking for insurance 184 00:08:31,340 --> 00:08:33,919 because you're risk averse and that's great for insurers. 185 00:08:33,919 --> 00:08:37,039 We talked about insurance, how there's essentially a game. 186 00:08:37,039 --> 00:08:39,650 I'm willing to pay a risk premium so insurers can 187 00:08:39,650 --> 00:08:41,002 make money by selling to me. 188 00:08:41,002 --> 00:08:43,419 But what if I'm not coming to you because I'm risk averse? 189 00:08:43,419 --> 00:08:45,669 What if I'm coming to you because I'm a huge skydiving 190 00:08:45,669 --> 00:08:46,190 fan? 191 00:08:46,190 --> 00:08:47,712 And you don't know that. 192 00:08:47,712 --> 00:08:49,420 You might be afraid to sell me insurance, 193 00:08:49,420 --> 00:08:52,890 because you might lose money on me. 194 00:08:52,890 --> 00:08:55,530 So let's work out another example to show this. 195 00:08:55,530 --> 00:08:58,560 Imagine you graduate, and you decide 196 00:08:58,560 --> 00:09:01,290 a great business model is to offer health 197 00:09:01,290 --> 00:09:03,630 insurance to recent MIT grads. 198 00:09:03,630 --> 00:09:06,270 You say, look, we're a bunch of kind of careful nerds. 199 00:09:06,270 --> 00:09:08,190 We're likely not going to go skydiving. 200 00:09:08,190 --> 00:09:10,380 We're just going to sit at our desks and work. 201 00:09:10,380 --> 00:09:12,870 Maybe there'll be carpal tunnel risk, but other than that, 202 00:09:12,870 --> 00:09:14,012 we're a pretty safe bunch. 203 00:09:14,012 --> 00:09:16,470 So I'm going to offer health insurance to recent MIT grads, 204 00:09:16,470 --> 00:09:19,000 because they're a healthy group. 205 00:09:19,000 --> 00:09:25,020 And let's say suppose that of every 100 MIT grads, 206 00:09:25,020 --> 00:09:31,350 90 are healthy, and 10 are sickly. 207 00:09:31,350 --> 00:09:33,385 Let's just suppose you know those facts. 208 00:09:33,385 --> 00:09:34,260 You know those facts. 209 00:09:34,260 --> 00:09:35,970 You've collected the data to know 210 00:09:35,970 --> 00:09:38,970 that on average, of every 100 MIT grads, 90 are healthy, 211 00:09:38,970 --> 00:09:39,750 10 are sickly. 212 00:09:39,750 --> 00:09:43,860 You don't know which are which, but you know the proportions. 213 00:09:43,860 --> 00:09:48,660 And let's say that with a healthy person 214 00:09:48,660 --> 00:09:51,930 over the next year, there's a 10% chance 215 00:09:51,930 --> 00:09:53,850 that they will need-- 216 00:09:53,850 --> 00:10:05,220 that they will incur a $10,000 charge, and a 90% chance 217 00:10:05,220 --> 00:10:08,510 that they'll have zero costs. 218 00:10:08,510 --> 00:10:12,390 So there's a 10% chance of a $10,000 cost, 90% chance 219 00:10:12,390 --> 00:10:15,380 they'll have a zero cost. 220 00:10:15,380 --> 00:10:23,020 So your expected cost for insuring this person is $1,000. 221 00:10:23,020 --> 00:10:31,460 You expect someone like that will cost you $1,000. 222 00:10:31,460 --> 00:10:37,400 Now suppose for the sickly guy, there's a 50% chance 223 00:10:37,400 --> 00:10:42,170 that they'll cost you $10 K, and a 50% chance that they'll 224 00:10:42,170 --> 00:10:43,880 cost you 0. 225 00:10:43,880 --> 00:10:47,210 So your expected costs for them, the expected costs 226 00:10:47,210 --> 00:10:52,898 for this person, is $5,000. 227 00:10:52,898 --> 00:10:54,190 Do people understand the setup? 228 00:10:54,190 --> 00:10:55,840 There's two types. 229 00:10:55,840 --> 00:11:01,040 I know these facts, but I don't know who's who. 230 00:11:01,040 --> 00:11:03,550 I just know these facts, because I'm good at math. 231 00:11:03,550 --> 00:11:05,480 I've done all the actuarial calculations. 232 00:11:08,240 --> 00:11:12,972 Now, if everyone buys in, now I'm going to set my price. 233 00:11:12,972 --> 00:11:14,430 What am I doing to do to the price? 234 00:11:14,430 --> 00:11:18,630 I'm going to say, look, if everyone buys health insurance, 235 00:11:18,630 --> 00:11:24,360 then I've got my expected cost is 0.9 times 236 00:11:24,360 --> 00:11:29,430 1,000 plus 0.1 times 5,000. 237 00:11:29,430 --> 00:11:33,800 My expected cost is 1,400, so I expect 238 00:11:33,800 --> 00:11:35,400 to have to spend $1,400 a year. 239 00:11:35,400 --> 00:11:38,310 In fact, on average, I'll spend $1,400 a year. 240 00:11:38,310 --> 00:11:41,130 With large enough samples, I can predict that with certainty-- 241 00:11:41,130 --> 00:11:45,810 that if everyone buys insurance, I'll spend $1,400 a year. 242 00:11:45,810 --> 00:11:48,060 So let's say you're risk neutral, because you're rich, 243 00:11:48,060 --> 00:11:49,140 and so forth. 244 00:11:49,140 --> 00:11:51,420 This is sort of risk neutral for you. 245 00:11:51,420 --> 00:11:53,310 So you say, look, I'll just charge-- 246 00:11:53,310 --> 00:11:58,330 I'll set a premium of $1,500, and I'll make $100 per person. 247 00:11:58,330 --> 00:12:00,930 If 100 will buy, that's $10,000 profit. 248 00:12:00,930 --> 00:12:03,340 That's pretty good. 249 00:12:03,340 --> 00:12:05,540 There are 1,000 kids in the graduating class. 250 00:12:05,540 --> 00:12:07,850 If 1,000 kids buy, and I make $100 profit, 251 00:12:07,850 --> 00:12:09,560 then that's $100,000. 252 00:12:09,560 --> 00:12:12,554 That's pretty good money for a year. 253 00:12:12,554 --> 00:12:19,240 Now, what is wrong with this calculation? 254 00:12:19,240 --> 00:12:23,590 What, in fact, will happen if you sell insurance for $1,500? 255 00:12:23,590 --> 00:12:24,290 Yeah. 256 00:12:24,290 --> 00:12:25,540 AUDIENCE: We'll buy insurance. 257 00:12:25,540 --> 00:12:27,802 JONATHAN GRUBER: Well let's go step by step. 258 00:12:27,802 --> 00:12:28,760 What about sick people? 259 00:12:28,760 --> 00:12:30,610 If you sell $1,500, what will they do? 260 00:12:30,610 --> 00:12:31,590 AUDIENCE: They'll buy. 261 00:12:31,590 --> 00:12:33,673 JONATHAN GRUBER: Yes, so if you set up for $1,500, 262 00:12:33,673 --> 00:12:37,150 you are certainly going to sell to all the sick. 263 00:12:37,150 --> 00:12:39,550 So if you sell for $1,500, you'll 264 00:12:39,550 --> 00:12:42,030 certainly sell to all the sick. 265 00:12:42,030 --> 00:12:44,110 What about the healthy? 266 00:12:44,110 --> 00:12:47,100 What will determine whether or not they buy? 267 00:12:47,100 --> 00:12:47,710 Yeah. 268 00:12:47,710 --> 00:12:50,228 AUDIENCE: When the price is smaller 269 00:12:50,228 --> 00:12:52,590 than the expected amount they'd have to pay on their own 270 00:12:52,590 --> 00:12:53,307 [INAUDIBLE]. 271 00:12:53,307 --> 00:12:54,890 JONATHAN GRUBER: Not quite, not quite. 272 00:12:54,890 --> 00:12:56,015 There's another piece, too. 273 00:12:56,015 --> 00:12:56,690 Don't forget. 274 00:12:56,690 --> 00:12:57,190 What else? 275 00:12:57,190 --> 00:12:58,100 What else? 276 00:12:58,100 --> 00:13:00,080 It's not just the expected cost. 277 00:13:00,080 --> 00:13:00,663 What else? 278 00:13:00,663 --> 00:13:01,830 AUDIENCE: The risk aversion. 279 00:13:01,830 --> 00:13:03,350 JONATHAN GRUBER: The risk aversion-- remember, there's 280 00:13:03,350 --> 00:13:05,320 a risk premium that they'll pay. 281 00:13:05,320 --> 00:13:07,280 So whether that healthy person will buy 282 00:13:07,280 --> 00:13:10,970 or not, if it's just expected cost, then they wouldn't buy-- 283 00:13:10,970 --> 00:13:14,000 the $1,000 expected cost, buy the $1,500. 284 00:13:14,000 --> 00:13:16,190 But some might be risk averse and buy. 285 00:13:16,190 --> 00:13:19,100 So let's just say half the guys are risk averse, 286 00:13:19,100 --> 00:13:21,860 and they're willing to pay $1,500 for $1,000 expected 287 00:13:21,860 --> 00:13:23,320 costs in half are. 288 00:13:23,320 --> 00:13:25,670 So let's say you end up selling to all the sick 289 00:13:25,670 --> 00:13:27,980 and half the healthy. 290 00:13:27,980 --> 00:13:29,900 So how much money do you make? 291 00:13:29,900 --> 00:13:35,220 Well, you sell the 60 people at $1,500 each, 292 00:13:35,220 --> 00:13:41,040 so your revenues are $75,000. 293 00:13:41,040 --> 00:13:44,940 You sell to 50 healthy and 10 sick, $75,000 revenues. 294 00:13:44,940 --> 00:13:46,870 What are your costs? 295 00:13:46,870 --> 00:13:51,600 Your costs are-- you have 10 people that are cost you 296 00:13:51,600 --> 00:13:57,470 $5,000, so it's $50,000, plus 50 people 297 00:13:57,470 --> 00:14:02,920 who are going to cost you $1,000, plus another $50,000 298 00:14:02,920 --> 00:14:06,660 equals $100,000. 299 00:14:06,660 --> 00:14:09,910 And you've lost money. 300 00:14:09,910 --> 00:14:15,090 You priced at above the expected total, and lost money. 301 00:14:15,090 --> 00:14:17,190 Why did you lose money? 302 00:14:17,190 --> 00:14:22,170 You lost money because of the problem of adverse selection-- 303 00:14:26,960 --> 00:14:30,320 the problem of we call "adverse selection." 304 00:14:30,320 --> 00:14:32,300 Adverse selection is the problem that, 305 00:14:32,300 --> 00:14:36,730 due to information asymmetries, only the worse risks 306 00:14:36,730 --> 00:14:39,680 will participate in the market. 307 00:14:39,680 --> 00:14:41,720 And that will cause people selling in the market 308 00:14:41,720 --> 00:14:44,180 to lose money, or likewise here, the concern 309 00:14:44,180 --> 00:14:47,970 is that only the worst cars will participate in the market. 310 00:14:47,970 --> 00:14:50,220 And so if people buy cars, we'll be worse off. 311 00:14:50,220 --> 00:14:50,720 Yeah. 312 00:14:50,720 --> 00:14:53,457 AUDIENCE: Wouldn't you make $90,000? 313 00:14:53,457 --> 00:14:54,290 JONATHAN GRUBER: No. 314 00:14:54,290 --> 00:14:59,720 You sell to-- you sell $1,500 each, and you sell to-- 315 00:14:59,720 --> 00:15:00,470 yes, you're right. 316 00:15:00,470 --> 00:15:02,510 You make $90,000, my bad. 317 00:15:02,510 --> 00:15:05,240 Yes, $1,500 each times 60 people is $90,000. 318 00:15:05,240 --> 00:15:07,190 You still lose money, but not as much. 319 00:15:07,190 --> 00:15:09,110 Right. 320 00:15:09,110 --> 00:15:13,410 Now, you might say, look, you're not losing that much. 321 00:15:13,410 --> 00:15:16,640 Your solution-- just raise the price. 322 00:15:16,640 --> 00:15:20,080 What if you said, fine, let's just raise the price, 323 00:15:20,080 --> 00:15:24,460 and let's charge $2,000 a person. 324 00:15:24,460 --> 00:15:26,920 Then that would cover, because $2,000 a person, 325 00:15:26,920 --> 00:15:31,000 60 will make $120,000, the costs are $100,000. 326 00:15:31,000 --> 00:15:32,620 You'd be golden, right? 327 00:15:32,620 --> 00:15:33,295 Yeah. 328 00:15:33,295 --> 00:15:34,600 AUDIENCE: [INAUDIBLE] 329 00:15:34,600 --> 00:15:35,350 JONATHAN GRUBER: Who would you lose? 330 00:15:35,350 --> 00:15:36,070 AUDIENCE: The healthy. 331 00:15:36,070 --> 00:15:37,320 JONATHAN GRUBER: Not the sick. 332 00:15:37,320 --> 00:15:38,860 The sick are delighted by $10,000. 333 00:15:38,860 --> 00:15:41,860 But once you raise the price, more healthy people drop out, 334 00:15:41,860 --> 00:15:44,760 because it's higher than their risk premium. 335 00:15:44,760 --> 00:15:46,427 So what happens is by raising the price, 336 00:15:46,427 --> 00:15:48,218 you're not necessarily going to make money. 337 00:15:48,218 --> 00:15:50,210 It depends on how many healthy people drop out. 338 00:15:50,210 --> 00:15:57,173 So for example, imagine that you raise it to $2,000, 339 00:15:57,173 --> 00:15:59,090 but now the number of healthy people that buys 340 00:15:59,090 --> 00:16:02,030 drops to 20 from 50. 341 00:16:02,030 --> 00:16:04,170 Then you lose money again. 342 00:16:04,170 --> 00:16:05,835 So the point is, you can't actually 343 00:16:05,835 --> 00:16:07,710 solve this problem just by raising the price, 344 00:16:07,710 --> 00:16:11,230 because there's this what we call "death spiral." 345 00:16:11,230 --> 00:16:13,350 This is a term called "death spiral," which 346 00:16:13,350 --> 00:16:15,840 is as you raise the price, you chase out 347 00:16:15,840 --> 00:16:17,640 the healthier people, which means 348 00:16:17,640 --> 00:16:19,110 you have to raise the price more, 349 00:16:19,110 --> 00:16:21,480 which shakes out even more healthy people. 350 00:16:21,480 --> 00:16:22,980 And you end up in this death spiral. 351 00:16:26,840 --> 00:16:30,810 So that is the problem of adverse selection. 352 00:16:30,810 --> 00:16:34,160 And that leads you to say, you know what, I'm not 353 00:16:34,160 --> 00:16:36,110 going to offer this product. 354 00:16:36,110 --> 00:16:39,030 I can't make money on it, because if I 355 00:16:39,030 --> 00:16:42,977 set the price, whatever price I set, I'm going to lose money. 356 00:16:42,977 --> 00:16:44,560 So I'm not going to offer the product. 357 00:16:44,560 --> 00:16:46,770 Therefore, the market has failed. 358 00:16:46,770 --> 00:16:48,570 A market that might have existed-- 359 00:16:48,570 --> 00:16:51,973 on average, this was a market that made people better off, 360 00:16:51,973 --> 00:16:54,390 but the market that might have existed doesn't does exist. 361 00:16:54,390 --> 00:16:55,394 Yeah. 362 00:16:55,394 --> 00:16:57,830 AUDIENCE: With the death spiral, wouldn't it 363 00:16:57,830 --> 00:17:01,980 converge like something equivalent [INAUDIBLE] 364 00:17:01,980 --> 00:17:08,603 market forecasting is [INAUDIBLE] still [INAUDIBLE].. 365 00:17:08,603 --> 00:17:10,770 JONATHAN GRUBER: Right, so if you leave this alone-- 366 00:17:10,770 --> 00:17:12,190 it's an excellent point-- 367 00:17:12,190 --> 00:17:13,766 what should the new equilibrium be? 368 00:17:13,766 --> 00:17:15,599 The price now should potentially have chased 369 00:17:15,599 --> 00:17:17,869 all the healthy people out. 370 00:17:17,869 --> 00:17:19,805 And then you price, but you'd have 371 00:17:19,805 --> 00:17:20,930 to price it, then, at what? 372 00:17:20,930 --> 00:17:22,097 AUDIENCE: $5,000 [INAUDIBLE] 373 00:17:22,097 --> 00:17:23,770 JONATHAN GRUBER: $5,000 plus something. 374 00:17:23,770 --> 00:17:25,839 So as long as sick people are risk averse, 375 00:17:25,839 --> 00:17:28,180 you could still make money. 376 00:17:28,180 --> 00:17:31,480 You could still make money if you sold, say, $5,500 377 00:17:31,480 --> 00:17:33,130 with even modest risk aversion. 378 00:17:33,130 --> 00:17:36,141 Why is that still a market failure? 379 00:17:36,141 --> 00:17:39,968 AUDIENCE: [INAUDIBLE] 380 00:17:39,968 --> 00:17:42,510 JONATHAN GRUBER: Because there's all these healthy people who 381 00:17:42,510 --> 00:17:45,000 now can't get health insurance. 382 00:17:45,000 --> 00:17:47,062 So yes, it doesn't mean the market collapse-- 383 00:17:47,062 --> 00:17:48,520 market failures doesn't necessarily 384 00:17:48,520 --> 00:17:50,520 mean market collapse. 385 00:17:50,520 --> 00:17:52,770 It means a reduction in welfare, because transactions 386 00:17:52,770 --> 00:17:55,140 that might make some people better off aren't happening. 387 00:17:55,140 --> 00:17:57,270 Here, you might be able to offer insurance 388 00:17:57,270 --> 00:17:58,470 for healthy people that makes them better off, 389 00:17:58,470 --> 00:17:59,220 but you're not. 390 00:17:59,220 --> 00:18:02,000 You're only offering insurance for the sick. 391 00:18:02,000 --> 00:18:05,400 So it's a market failure, because healthy people 392 00:18:05,400 --> 00:18:08,970 who might want the insurance end up being kept out of the market 393 00:18:08,970 --> 00:18:12,710 by adverse selection. 394 00:18:12,710 --> 00:18:15,450 Questions about that? 395 00:18:15,450 --> 00:18:17,460 And that is the fundamental market failure 396 00:18:17,460 --> 00:18:18,700 we face in insurance markets. 397 00:18:18,700 --> 00:18:21,090 That's why we think private insurance markets will not 398 00:18:21,090 --> 00:18:22,290 function well. 399 00:18:22,290 --> 00:18:24,690 Because private insurers-- in some sense 400 00:18:24,690 --> 00:18:27,150 the fundamental problem is that you're 401 00:18:27,150 --> 00:18:31,810 setting one price for multiple products. 402 00:18:31,810 --> 00:18:33,690 A great case of adverse selection 403 00:18:33,690 --> 00:18:35,830 is going to buy fruit at the beginning 404 00:18:35,830 --> 00:18:37,650 versus the end of the day. 405 00:18:37,650 --> 00:18:39,090 What's the difference? 406 00:18:39,090 --> 00:18:41,978 Buy fruit at the beginning or the end of the day? 407 00:18:41,978 --> 00:18:43,770 You guys probably don't buy a lot of fruit, 408 00:18:43,770 --> 00:18:44,550 but try to think about it. 409 00:18:44,550 --> 00:18:45,225 Yeah. 410 00:18:45,225 --> 00:18:46,965 AUDIENCE: Normally, better is at the beginning of the day. 411 00:18:46,965 --> 00:18:47,400 Then you have-- 412 00:18:47,400 --> 00:18:48,390 JONATHAN GRUBER: And at the end of the day, 413 00:18:48,390 --> 00:18:49,840 in particular, what's left? 414 00:18:49,840 --> 00:18:50,715 AUDIENCE: [INAUDIBLE] 415 00:18:50,715 --> 00:18:52,257 JONATHAN GRUBER: All the shitty fruit 416 00:18:52,257 --> 00:18:53,800 is left, because you set one price. 417 00:18:53,800 --> 00:18:57,000 You didn't say good apples $1.80, shitty apples $1.40. 418 00:18:57,000 --> 00:18:59,420 You said apples, $1.70. 419 00:18:59,420 --> 00:19:01,220 So people come, and they buy apples. 420 00:19:01,220 --> 00:19:02,552 They go and they feel it. 421 00:19:02,552 --> 00:19:03,260 They feel around. 422 00:19:03,260 --> 00:19:04,260 They find the good ones. 423 00:19:04,260 --> 00:19:06,790 The ones that left are crap. 424 00:19:06,790 --> 00:19:09,070 And that is the adverse selection problem. 425 00:19:09,070 --> 00:19:13,090 Now, with apples, the market still exists. 426 00:19:13,090 --> 00:19:13,750 Why? 427 00:19:13,750 --> 00:19:17,050 Because they charge so much they can live with a few bad apples 428 00:19:17,050 --> 00:19:18,490 being at the bottom, so to speak-- 429 00:19:18,490 --> 00:19:20,560 a few bad apples being at the bottom. 430 00:19:20,560 --> 00:19:24,310 With health insurance, if I get one bad risk-- 431 00:19:24,310 --> 00:19:26,860 someone, say, who's really, really sick 432 00:19:26,860 --> 00:19:28,300 and costs $1, million-- 433 00:19:28,300 --> 00:19:30,750 I go out of business. 434 00:19:30,750 --> 00:19:33,770 So adverse selection may not destroy markets. 435 00:19:33,770 --> 00:19:36,010 It doesn't destroy the apple market, 436 00:19:36,010 --> 00:19:38,110 but it can destroy or significantly impede 437 00:19:38,110 --> 00:19:40,840 insurance markets. 438 00:19:40,840 --> 00:19:42,480 Questions about that? 439 00:19:42,480 --> 00:19:43,020 Yeah. 440 00:19:43,020 --> 00:19:43,240 Manny. 441 00:19:43,240 --> 00:19:45,198 AUDIENCE: Is there some way insurance companies 442 00:19:45,198 --> 00:19:50,155 that hassle hospitals, like lower the prices or they give 443 00:19:50,155 --> 00:19:51,530 them better discounts so they can 444 00:19:51,530 --> 00:19:53,430 increase the price of people-- 445 00:19:53,430 --> 00:19:54,350 JONATHAN GRUBER: Well, that's a separate issue. 446 00:19:54,350 --> 00:19:55,400 We'll talk about that next lecture 447 00:19:55,400 --> 00:19:56,300 when we talk about health care. 448 00:19:56,300 --> 00:19:58,340 So that's separate, about the cost of health care. 449 00:19:58,340 --> 00:20:00,090 This is the reason why insurance companies 450 00:20:00,090 --> 00:20:03,230 make you fill out a lot of forms before you go in. 451 00:20:03,230 --> 00:20:05,660 So it's for this reason-- 452 00:20:05,660 --> 00:20:08,958 insurance companies are not powerless against this problem. 453 00:20:08,958 --> 00:20:11,250 They could try to collect as much information about you 454 00:20:11,250 --> 00:20:12,540 as they can. 455 00:20:12,540 --> 00:20:14,530 As I get more and more [INAUDIBLE] 456 00:20:14,530 --> 00:20:16,860 can learn more and more who's healthy, who's sick, 457 00:20:16,860 --> 00:20:18,110 then I can solve this problem. 458 00:20:21,290 --> 00:20:24,582 AUDIENCE: So your familiar with those 459 00:20:24,582 --> 00:20:26,790 home kits that [INAUDIBLE] and 23andme will send you. 460 00:20:26,790 --> 00:20:29,050 JONATHAN GRUBER: Yeah, 23andme. 461 00:20:29,050 --> 00:20:29,925 Yeah. 462 00:20:29,925 --> 00:20:34,290 AUDIENCE: So is it possible that at some point 463 00:20:34,290 --> 00:20:36,390 in the foreseeable future, those are going 464 00:20:36,390 --> 00:20:39,938 to become part of [INAUDIBLE]? 465 00:20:39,938 --> 00:20:42,580 Those are going to become part of how insurance is determined, 466 00:20:42,580 --> 00:20:46,340 like if it's in your DNA, get some condition when 467 00:20:46,340 --> 00:20:48,840 you get old that we can say you have a preexisting condition 468 00:20:48,840 --> 00:20:50,360 now that hasn't manifested yet? 469 00:20:50,360 --> 00:20:52,110 JONATHAN GRUBER: So this is a great point. 470 00:20:52,110 --> 00:20:53,735 I was going to talk about it next time. 471 00:20:53,735 --> 00:20:56,790 I'll talk about it now, which is in some sense, 472 00:20:56,790 --> 00:20:59,850 we are eventually moving to a point 473 00:20:59,850 --> 00:21:01,960 where there'll be no adverse selection. 474 00:21:01,960 --> 00:21:03,710 Now you might say, on the one hand, that's 475 00:21:03,710 --> 00:21:05,330 because ultimately, we'll know everything about you 476 00:21:05,330 --> 00:21:06,950 from the moment you're born. 477 00:21:06,950 --> 00:21:08,180 We know your genes. 478 00:21:08,180 --> 00:21:10,700 We won't know if you're a skydiver, but we'll know-- 479 00:21:10,700 --> 00:21:13,210 we'll probably know genes that determine risk taking. 480 00:21:13,210 --> 00:21:17,100 And we'll charge more for people who like taking risks. 481 00:21:17,100 --> 00:21:23,580 So the good news is that then, I can make the market work. 482 00:21:23,580 --> 00:21:28,854 The bad news is in that world, how would I set my insurance? 483 00:21:28,854 --> 00:21:30,625 AUDIENCE: [INAUDIBLE] 484 00:21:30,625 --> 00:21:32,000 JONATHAN GRUBER: I would charge-- 485 00:21:32,000 --> 00:21:33,458 because what I would do is I'd say, 486 00:21:33,458 --> 00:21:37,340 your genes say you're healthy, so I want $1,100 from you. 487 00:21:37,340 --> 00:21:41,040 Your genes say you're sick, so I want $6,000 from you. 488 00:21:41,040 --> 00:21:45,050 So essentially, insurance wouldn't exist anymore. 489 00:21:45,050 --> 00:21:46,050 There'd be no insurance. 490 00:21:46,050 --> 00:21:47,310 What is insurance? 491 00:21:47,310 --> 00:21:50,310 Insurance is pooling people with different probabilities 492 00:21:50,310 --> 00:21:53,430 of adverse events, and letting us all benefit from the fact 493 00:21:53,430 --> 00:21:55,920 that if it happens to us, at least we're protected. 494 00:21:55,920 --> 00:21:58,170 Well, if you charge me my expected cost, 495 00:21:58,170 --> 00:21:59,993 I'm no longer protected. 496 00:21:59,993 --> 00:22:02,160 So here's the example that makes it perfectly clear, 497 00:22:02,160 --> 00:22:03,330 one of the most famous examples. 498 00:22:03,330 --> 00:22:05,913 Ken Arrow was one of the great economists of the 20th century, 499 00:22:05,913 --> 00:22:06,960 died recently. 500 00:22:06,960 --> 00:22:08,442 He had his famous islands example. 501 00:22:08,442 --> 00:22:10,650 Ken Arrow's island example is the following-- imagine 502 00:22:10,650 --> 00:22:13,300 there's two islands somewhere in the South Pacific 503 00:22:13,300 --> 00:22:15,870 that are very small, that's got one farmer on each. 504 00:22:15,870 --> 00:22:18,760 And the farmers know a hurricane is coming 505 00:22:18,760 --> 00:22:20,760 and it's going to wipe out one of their islands, 506 00:22:20,760 --> 00:22:22,710 but they don't know which. 507 00:22:22,710 --> 00:22:24,663 They just know one's getting wiped out. 508 00:22:24,663 --> 00:22:25,830 What will they naturally do? 509 00:22:25,830 --> 00:22:28,020 They'll naturally get together and say, look, 510 00:22:28,020 --> 00:22:30,270 islands get wiped out, but let's insure each other. 511 00:22:30,270 --> 00:22:32,180 If I get wiped out, you give me a bunch of your crop. 512 00:22:32,180 --> 00:22:34,472 If you get wiped out, I'll give you a bunch of my crop. 513 00:22:34,472 --> 00:22:36,120 That will improve both our welfare, 514 00:22:36,120 --> 00:22:37,590 because getting wiped out is going to zero. 515 00:22:37,590 --> 00:22:38,090 You die. 516 00:22:38,090 --> 00:22:39,600 That's a terrible outcome. 517 00:22:39,600 --> 00:22:42,810 So that will improve our welfare if we insure each other. 518 00:22:42,810 --> 00:22:45,480 Now let's say a weather service comes along, and provides 519 00:22:45,480 --> 00:22:47,497 information, and tells you that farmer 520 00:22:47,497 --> 00:22:49,830 A's island is getting wiped out and farmer B's island is 521 00:22:49,830 --> 00:22:51,530 going to be fine. 522 00:22:51,530 --> 00:22:54,020 What has happened to welfare? 523 00:22:54,020 --> 00:22:55,650 It's gotten worse. 524 00:22:55,650 --> 00:22:56,267 Why? 525 00:22:56,267 --> 00:22:57,892 Because farmer A goes to farmer B, says 526 00:22:57,892 --> 00:22:59,517 it turns out I'm going to be wiped out. 527 00:22:59,517 --> 00:23:00,760 Farmer B says, well, see you. 528 00:23:00,760 --> 00:23:03,130 I'm just going to keep consuming my high level. 529 00:23:03,130 --> 00:23:04,950 Farmer B is somewhat better off. 530 00:23:04,950 --> 00:23:07,850 Farmer A is dead. 531 00:23:07,850 --> 00:23:10,850 Total social welfare has fallen because the concavity, 532 00:23:10,850 --> 00:23:12,830 because diminishing market utility. 533 00:23:12,830 --> 00:23:16,280 More information has made us worse. 534 00:23:16,280 --> 00:23:17,773 We say more is better in economics. 535 00:23:17,773 --> 00:23:19,940 Once you get into topics like, this you realize more 536 00:23:19,940 --> 00:23:21,090 is not always better. 537 00:23:21,090 --> 00:23:22,370 More is worse. 538 00:23:22,370 --> 00:23:24,560 More information has destroyed the insurance market 539 00:23:24,560 --> 00:23:26,220 that might function. 540 00:23:26,220 --> 00:23:28,370 So in fact, this issue I'm talking about 541 00:23:28,370 --> 00:23:31,010 is becoming paramount as we move more and more 542 00:23:31,010 --> 00:23:33,585 towards perfect information environment. 543 00:23:33,585 --> 00:23:35,960 So the kind of government policies I'm talking about next 544 00:23:35,960 --> 00:23:39,368 become critical as you move towards that environment. 545 00:23:39,368 --> 00:23:40,910 But first, I want to make sure we all 546 00:23:40,910 --> 00:23:43,100 understand why the private markets failed, 547 00:23:43,100 --> 00:23:46,040 why it's a failure. 548 00:23:46,040 --> 00:23:48,810 Now, what can the government do about this? 549 00:23:48,810 --> 00:23:51,710 What are some potential government solutions? 550 00:23:51,710 --> 00:23:54,210 And we've tried all of these in the US and around the world. 551 00:23:54,210 --> 00:23:55,585 Let's talk about three categories 552 00:23:55,585 --> 00:23:57,150 of government solutions. 553 00:23:57,150 --> 00:23:59,780 The first is subsidization. 554 00:24:02,810 --> 00:24:05,090 The government could subsidize the purchase 555 00:24:05,090 --> 00:24:06,290 of health insurance. 556 00:24:06,290 --> 00:24:09,500 So for example, what if the US government said 557 00:24:09,500 --> 00:24:13,730 to all the MIT grads, I'm going to give you a $400 558 00:24:13,730 --> 00:24:17,750 tax credit that you could have-- 559 00:24:17,750 --> 00:24:22,953 or $500 tax credit if you buy health insurance. 560 00:24:22,953 --> 00:24:24,370 Well, if there's a $500 tax credit 561 00:24:24,370 --> 00:24:28,410 if I buy health insurance, and I charge $1,500, 562 00:24:28,410 --> 00:24:30,730 then what's the effective price now to the healthy guy? 563 00:24:30,730 --> 00:24:31,647 AUDIENCE: [INAUDIBLE]. 564 00:24:31,647 --> 00:24:33,192 JONATHAN GRUBER: $1,000, so he buys. 565 00:24:33,192 --> 00:24:34,984 Even if he's risk neutral, he buys, as long 566 00:24:34,984 --> 00:24:36,950 as he's a tiny bit risk averse. 567 00:24:36,950 --> 00:24:38,250 So I do sell to everyone. 568 00:24:38,250 --> 00:24:40,332 I make my money. 569 00:24:40,332 --> 00:24:41,790 So one way to solve this problem is 570 00:24:41,790 --> 00:24:44,970 to basically pay the healthy people to get into the market. 571 00:24:44,970 --> 00:24:46,260 They can't just give money to the healthy people. 572 00:24:46,260 --> 00:24:47,370 You've got to give it to everyone, 573 00:24:47,370 --> 00:24:49,170 because you can't tell who's healthy. 574 00:24:49,170 --> 00:24:52,050 But if we give everyone a tax credit, 575 00:24:52,050 --> 00:24:54,210 then we could bring everyone to the market 576 00:24:54,210 --> 00:24:56,020 and solve this problem. 577 00:24:56,020 --> 00:24:58,170 Well, in fact, we do this in America. 578 00:24:58,170 --> 00:25:01,230 It's actually perhaps the largest hidden government 579 00:25:01,230 --> 00:25:03,700 expenditure in our country, which 580 00:25:03,700 --> 00:25:12,510 is the tax subsidy to employer-sponsored insurance, 581 00:25:12,510 --> 00:25:14,338 employer health insurance. 582 00:25:17,783 --> 00:25:19,700 The tax subsidy on employer health insurance-- 583 00:25:19,700 --> 00:25:20,880 what do I mean by that? 584 00:25:20,880 --> 00:25:24,530 What I mean is the following-- when MIT pays me in wages, 585 00:25:24,530 --> 00:25:26,810 I am taxed on that, like the taxation we talked 586 00:25:26,810 --> 00:25:28,810 about a couple lectures ago. 587 00:25:28,810 --> 00:25:30,940 When MIT pays me in health insurance, 588 00:25:30,940 --> 00:25:33,092 I am not taxed on that. 589 00:25:33,092 --> 00:25:34,050 So what does that mean? 590 00:25:34,050 --> 00:25:35,425 If MIT comes to me, and they say, 591 00:25:35,425 --> 00:25:39,050 would you like $1,000 raise or $1,000 orthodontic benefits 592 00:25:39,050 --> 00:25:40,370 for your daughter? 593 00:25:40,370 --> 00:25:42,560 I say, well, $1,000 raise in today's tax rates, 594 00:25:42,560 --> 00:25:44,432 I'm going to take home about $550. 595 00:25:44,432 --> 00:25:45,890 If you add up all the tax I'll pay, 596 00:25:45,890 --> 00:25:48,942 then I'll take home about $550. 597 00:25:48,942 --> 00:25:50,900 $1,000 of orthodontic benefits for my daughter, 598 00:25:50,900 --> 00:25:52,430 I get the whole $1,000. 599 00:25:52,430 --> 00:25:53,015 So why not? 600 00:25:53,015 --> 00:25:54,140 So I got these cool braces. 601 00:25:54,140 --> 00:25:55,310 They spin and change color. 602 00:25:55,310 --> 00:25:57,852 And every two weeks, she's in for a different kind of braces. 603 00:25:57,852 --> 00:26:00,480 It's great, because it's free. 604 00:26:00,480 --> 00:26:04,440 So we do subsidize health insurance in America. 605 00:26:04,440 --> 00:26:06,145 And this amounts to-- this program 606 00:26:06,145 --> 00:26:10,230 that I just talked about amounts to almost $300 billion 607 00:26:10,230 --> 00:26:11,580 per year. 608 00:26:11,580 --> 00:26:14,370 We spend almost $300 billion per year 609 00:26:14,370 --> 00:26:19,050 giving a tax break to people to buy health insurance. 610 00:26:19,050 --> 00:26:21,390 So that's one tactic we take to try 611 00:26:21,390 --> 00:26:24,030 to solve this problem to get healthy people into the market. 612 00:26:24,030 --> 00:26:26,450 That's approach one. 613 00:26:26,450 --> 00:26:29,240 A second approach one can use to try to get people 614 00:26:29,240 --> 00:26:31,130 into the market is a mandate. 615 00:26:33,660 --> 00:26:36,120 Suppose I just pass a law that says everyone 616 00:26:36,120 --> 00:26:38,010 has to buy health insurance. 617 00:26:38,010 --> 00:26:40,080 Then I've solved the problem. 618 00:26:40,080 --> 00:26:43,500 I know what my expected costs are if everyone has to buy. 619 00:26:43,500 --> 00:26:45,240 I know my expected costs are $1,400, 620 00:26:45,240 --> 00:26:48,970 so I know I can make money at $1,500. 621 00:26:48,970 --> 00:26:51,230 That's easier at one level. 622 00:26:51,230 --> 00:26:52,550 I don't have to spend-- 623 00:26:52,550 --> 00:26:53,800 $30 billion is a lot of money. 624 00:26:53,800 --> 00:26:55,872 This cost me $0. 625 00:26:55,872 --> 00:26:57,080 It's harder on another level. 626 00:26:57,080 --> 00:26:57,830 Why? 627 00:26:57,830 --> 00:26:59,760 What's the problem with that solution? 628 00:26:59,760 --> 00:27:00,847 Yeah. 629 00:27:00,847 --> 00:27:02,722 AUDIENCE: Not having the money for insurance. 630 00:27:02,722 --> 00:27:03,190 [INTERPOSING VOICES] 631 00:27:03,190 --> 00:27:05,150 JONATHAN GRUBER: Well, it may not have it. 632 00:27:05,150 --> 00:27:05,750 That's right. 633 00:27:05,750 --> 00:27:07,160 What else? 634 00:27:07,160 --> 00:27:07,680 Yeah. 635 00:27:07,680 --> 00:27:08,750 AUDIENCE: He may not want it. 636 00:27:08,750 --> 00:27:11,208 JONATHAN GRUBER: The healthy people are going to be pissed. 637 00:27:11,208 --> 00:27:13,280 They're like, look, if I had chosen not-- 638 00:27:13,280 --> 00:27:14,720 you're going to basically-- 639 00:27:14,720 --> 00:27:16,640 the mandate only has an effect if it 640 00:27:16,640 --> 00:27:18,060 changes people's behavior. 641 00:27:18,060 --> 00:27:19,100 But changing people's behavior means 642 00:27:19,100 --> 00:27:20,600 you're making them do something they 643 00:27:20,600 --> 00:27:22,130 didn't want to do beforehand. 644 00:27:22,130 --> 00:27:24,940 So the problem with that the problem with this is you 645 00:27:24,940 --> 00:27:25,980 spend a lot of money. 646 00:27:25,980 --> 00:27:28,030 The problem this is you piss off healthy people. 647 00:27:31,240 --> 00:27:34,030 The third approach we could do-- 648 00:27:34,030 --> 00:27:35,880 there's lots of examples of a mandate. 649 00:27:35,880 --> 00:27:37,150 Obviously, we know about the health insurance 650 00:27:37,150 --> 00:27:39,370 mandate that was originally part of Obamacare. 651 00:27:39,370 --> 00:27:41,050 But that's not the biggest example. 652 00:27:41,050 --> 00:27:42,970 The biggest example in the US is what's 653 00:27:42,970 --> 00:27:47,380 called "Workers' Comp Insurance," which is insurance 654 00:27:47,380 --> 00:27:49,750 that you have for on-the-job injuries. 655 00:27:49,750 --> 00:27:53,020 If you get hurt at work, your employer pays money so that you 656 00:27:53,020 --> 00:27:55,193 get reimbursed when you're-- 657 00:27:55,193 --> 00:27:57,360 it pays your medical bills when you get hurt at work 658 00:27:57,360 --> 00:28:00,400 and gives you partial replacement of your wages. 659 00:28:00,400 --> 00:28:02,770 That is mandated insurance on all employers 660 00:28:02,770 --> 00:28:05,830 in America, except in Texas. 661 00:28:05,830 --> 00:28:06,880 Texas, they can choose. 662 00:28:06,880 --> 00:28:08,791 Every other state, it's mandated. 663 00:28:08,791 --> 00:28:11,168 Mandated insurance for every employer in America. 664 00:28:11,168 --> 00:28:12,460 They have to buy Workers' Comp. 665 00:28:12,460 --> 00:28:13,543 So we've examples of that. 666 00:28:13,543 --> 00:28:15,700 And that's an $80 billion a year program. 667 00:28:15,700 --> 00:28:17,660 That's a big deal. 668 00:28:17,660 --> 00:28:21,560 Finally, we can just provide the insurance. 669 00:28:24,480 --> 00:28:27,000 That's actually the most common thing we do in America. 670 00:28:27,000 --> 00:28:30,240 Social Security is our program that provides insurance 671 00:28:30,240 --> 00:28:34,560 for the elderly, for the costs for survival after retirement. 672 00:28:34,560 --> 00:28:39,150 Medicare is insurance for the elderly we provide. 673 00:28:39,150 --> 00:28:41,280 Unemployment insurance is insurance 674 00:28:41,280 --> 00:28:43,470 we provide against losing your job. 675 00:28:43,470 --> 00:28:45,840 Disability insurance is insurance 676 00:28:45,840 --> 00:28:49,840 we provide against having a career-ending disability. 677 00:28:49,840 --> 00:28:51,590 So this is actually the most common thing. 678 00:28:51,590 --> 00:28:55,180 Indeed, provision of social insurance in America 679 00:28:55,180 --> 00:28:56,380 costs almost-- 680 00:28:56,380 --> 00:29:00,210 costs more than private insurance. 681 00:29:00,210 --> 00:29:02,710 So we spend about a trillion and a half on private insurance 682 00:29:02,710 --> 00:29:03,730 in America. 683 00:29:03,730 --> 00:29:06,340 Social insurance is probably about $1.7 to $2 trillion, 684 00:29:06,340 --> 00:29:08,710 depending how you measure it. 685 00:29:08,710 --> 00:29:10,410 So actually, this is the biggest thing 686 00:29:10,410 --> 00:29:14,320 we do is we just provide insurance, 687 00:29:14,320 --> 00:29:18,300 and that is a very large solution. 688 00:29:18,300 --> 00:29:21,570 Now once again, what's the problem with this? 689 00:29:21,570 --> 00:29:24,310 You don't make the healthy people unhappy, 690 00:29:24,310 --> 00:29:26,240 because everyone, you just give it to them. 691 00:29:26,240 --> 00:29:28,580 The problem is you have to spend money on this. 692 00:29:28,580 --> 00:29:31,760 This is $1.7 trillion in taxes we've got to raise every year. 693 00:29:31,760 --> 00:29:34,460 That's non-trivial. 694 00:29:34,460 --> 00:29:36,230 So basically, each of these solutions 695 00:29:36,230 --> 00:29:38,570 has potential problems. 696 00:29:38,570 --> 00:29:41,540 So the adverse selection problem will cause the private market 697 00:29:41,540 --> 00:29:42,540 to fail. 698 00:29:42,540 --> 00:29:44,350 There are potential government solutions, 699 00:29:44,350 --> 00:29:45,470 but they each have limitations. 700 00:29:45,470 --> 00:29:47,840 This one's pretty expensive, this one's super expensive, 701 00:29:47,840 --> 00:29:50,730 this one pisses off healthy people. 702 00:29:50,730 --> 00:29:53,595 Now, you'll note the middle one, the pissed-off healthy people 703 00:29:53,595 --> 00:29:54,640 is kind of subtle. 704 00:29:54,640 --> 00:29:56,515 You don't see a lot of healthy people railing 705 00:29:56,515 --> 00:29:58,310 against mandated [INAUDIBLE] Workers' Comp 706 00:29:58,310 --> 00:30:00,435 like they did against the health insurance mandate, 707 00:30:00,435 --> 00:30:01,830 because people don't know. 708 00:30:01,830 --> 00:30:04,320 So in some sense, this one's a little bit subtle, 709 00:30:04,320 --> 00:30:06,690 because people have to know. 710 00:30:06,690 --> 00:30:08,942 Basically, it's sort of crazy that I'm 711 00:30:08,942 --> 00:30:10,900 paying tax that I'm never getting hurt at work. 712 00:30:10,900 --> 00:30:11,810 Am I going to have-- 713 00:30:11,810 --> 00:30:14,718 what am I going to do, like slip at my desk or something? 714 00:30:14,718 --> 00:30:17,260 Never going to get hurt at work, but I pay taxes all the time 715 00:30:17,260 --> 00:30:19,860 just in case someone else at MIT gets hurt. 716 00:30:19,860 --> 00:30:22,590 Some of the janitorial staff has a risk of being hurt. 717 00:30:22,590 --> 00:30:25,290 I'm paying taxes in case the janitor gets hurt. 718 00:30:25,290 --> 00:30:28,620 I should be upset about that, but I'm not. 719 00:30:28,620 --> 00:30:30,810 And in some sense, it's about what people know, 720 00:30:30,810 --> 00:30:32,910 what they don't. 721 00:30:32,910 --> 00:30:38,150 So that is the basic argument for social insurance. 722 00:30:38,150 --> 00:30:39,830 But when we provide social insurance 723 00:30:39,830 --> 00:30:42,380 despite all these problems, we enter 724 00:30:42,380 --> 00:30:45,770 into a fundamental trade-off, which 725 00:30:45,770 --> 00:30:47,208 is, let's decide we've determined 726 00:30:47,208 --> 00:30:48,500 some optimal government policy. 727 00:30:48,500 --> 00:30:50,300 Let's decide that the markets failed, 728 00:30:50,300 --> 00:30:51,640 so we're going to do one of these things 729 00:30:51,640 --> 00:30:54,223 or are some combination of these things and solve the problem. 730 00:30:54,223 --> 00:30:59,570 The problem is that when you insure people for risks 731 00:30:59,570 --> 00:31:02,350 you create a new problem called "moral hazard." 732 00:31:05,930 --> 00:31:15,570 Moral hazard is basically the adverse behavior that 733 00:31:15,570 --> 00:31:18,950 is encouraged by insurance. 734 00:31:18,950 --> 00:31:24,180 When you insure people, you encourage adverse behavior. 735 00:31:24,180 --> 00:31:26,480 So the classic example of this is-- 736 00:31:26,480 --> 00:31:29,510 if I have health insurance, I ride my bike less carefully, 737 00:31:29,510 --> 00:31:30,870 because if I get in a crash-- 738 00:31:30,870 --> 00:31:31,970 I'm not crazy. 739 00:31:31,970 --> 00:31:33,020 I certainly don't want to get in a crash, 740 00:31:33,020 --> 00:31:34,130 but I'm a little bit less careful 741 00:31:34,130 --> 00:31:36,470 because I know I'm insured in case I get in a crash. 742 00:31:36,470 --> 00:31:39,380 If I have fire insurance, I don't buy a fire extinguisher 743 00:31:39,380 --> 00:31:41,270 for my house, because if it burns down, 744 00:31:41,270 --> 00:31:43,580 I'm just going to get the money back anyway. 745 00:31:43,580 --> 00:31:46,880 Or if workers have insurance against losing a job that 746 00:31:46,880 --> 00:31:49,340 pays them when out of work, they might search less hard 747 00:31:49,340 --> 00:31:51,320 for a new job. 748 00:31:51,320 --> 00:31:53,570 Basically, if I lose my job and I got nothing, 749 00:31:53,570 --> 00:31:54,930 I'm going to work my ass off to get a new job. 750 00:31:54,930 --> 00:31:56,388 If I lose my job and the government 751 00:31:56,388 --> 00:31:59,212 says, well, for 26 weeks, we'll give you half your salary 752 00:31:59,212 --> 00:32:01,670 while you look for a job, I'll be a little bit less rushed. 753 00:32:04,690 --> 00:32:07,530 And there's lots of evidence that moral hazard is a problem. 754 00:32:07,530 --> 00:32:09,220 I comes with two types of evidence. 755 00:32:09,220 --> 00:32:12,060 The first type evidence is fun anecdotes. 756 00:32:12,060 --> 00:32:14,380 So the great effect that-- workers' compensation, let's 757 00:32:14,380 --> 00:32:14,880 take that. 758 00:32:14,880 --> 00:32:17,255 Workers' compensation, it's a program [INAUDIBLE] needed. 759 00:32:17,255 --> 00:32:18,630 Lots of people get hurt at work. 760 00:32:18,630 --> 00:32:20,925 I don't, but lots of people do get hurt at work. 761 00:32:20,925 --> 00:32:23,920 And so it's a sensible social insurance program. 762 00:32:23,920 --> 00:32:26,460 The problem is it has a huge moral hazard component. 763 00:32:26,460 --> 00:32:28,200 And there's fun examples of this, 764 00:32:28,200 --> 00:32:31,410 like the prison guard in Massachusetts 765 00:32:31,410 --> 00:32:33,270 who claimed he got hurt on the job, 766 00:32:33,270 --> 00:32:36,870 collected $82,000 in benefits, while the whole time running 767 00:32:36,870 --> 00:32:38,842 a karate school and teaching students karate. 768 00:32:38,842 --> 00:32:40,800 And finally someone noticed online this guy who 769 00:32:40,800 --> 00:32:42,900 couldn't work was running a karate school 770 00:32:42,900 --> 00:32:45,090 and doing karate kicks and stuff online. 771 00:32:45,090 --> 00:32:47,520 So there's all sorts of fun examples about that. 772 00:32:47,520 --> 00:32:48,960 But more convincing for economists 773 00:32:48,960 --> 00:32:50,970 is statistical evidence. 774 00:32:50,970 --> 00:32:53,070 And the statistical evidence is clear 775 00:32:53,070 --> 00:32:56,160 that moral hazard is a big problem. 776 00:32:56,160 --> 00:32:59,090 For example, if you raise the benefits people get 777 00:32:59,090 --> 00:33:02,135 under workers' comp, suddenly they become injured more often 778 00:33:02,135 --> 00:33:03,800 and stay out of work longer. 779 00:33:03,800 --> 00:33:06,530 There's no reason-- injuries should be because you got hurt. 780 00:33:06,530 --> 00:33:08,030 So how can it be that suddenly, when 781 00:33:08,030 --> 00:33:09,613 a state raises its benefits, suddenly, 782 00:33:09,613 --> 00:33:10,580 there's more injuries? 783 00:33:10,580 --> 00:33:12,760 The answer is moral hazard. 784 00:33:12,760 --> 00:33:14,800 When states raise their unemployment insurance 785 00:33:14,800 --> 00:33:17,350 benefits, more people leave their jobs and they 786 00:33:17,350 --> 00:33:19,260 stay unemployed longer-- 787 00:33:19,260 --> 00:33:21,300 moral hazard. 788 00:33:21,300 --> 00:33:24,060 So the moral hazard problem is real. 789 00:33:24,060 --> 00:33:26,263 It's an inherent trade-off, actually not just 790 00:33:26,263 --> 00:33:27,180 with public insurance. 791 00:33:27,180 --> 00:33:29,160 Private insurance, too-- anytime you 792 00:33:29,160 --> 00:33:31,558 insure people and something bad happens, 793 00:33:31,558 --> 00:33:33,100 you're providing less of an incentive 794 00:33:33,100 --> 00:33:38,740 for them to try themselves to avoid that bad thing happening. 795 00:33:38,740 --> 00:33:41,390 So moral hazard is a real problem 796 00:33:41,390 --> 00:33:43,060 and it's essentially the trade-off. 797 00:33:43,060 --> 00:33:47,087 On the one hand, we talked about why people like insurance. 798 00:33:47,087 --> 00:33:49,420 We talked about why people like private insurers because 799 00:33:49,420 --> 00:33:50,290 of risk aversion. 800 00:33:50,290 --> 00:33:52,510 We talked about why government intervention insurance 801 00:33:52,510 --> 00:33:54,130 in markets is necessary. 802 00:33:54,130 --> 00:33:56,080 But that comes with the trade-off, which 803 00:33:56,080 --> 00:33:58,450 is the more insurance you provide, 804 00:33:58,450 --> 00:34:01,130 the less people take care of themselves. 805 00:34:01,130 --> 00:34:03,290 And that's the trade-off. 806 00:34:03,290 --> 00:34:06,090 Now, why do we care? 807 00:34:06,090 --> 00:34:08,590 Let's just sit back and say, that's an interesting economics 808 00:34:08,590 --> 00:34:09,715 concept, but why do I care? 809 00:34:09,715 --> 00:34:12,489 Why do I care if someone stays out of work longer, fakes 810 00:34:12,489 --> 00:34:14,920 an injury, or whatever? 811 00:34:14,920 --> 00:34:16,300 Why do I care about this? 812 00:34:16,300 --> 00:34:17,790 Why is this a problem? 813 00:34:17,790 --> 00:34:19,960 It's a problem for two reasons. 814 00:34:19,960 --> 00:34:22,650 There's two costs to moral hazard. 815 00:34:22,650 --> 00:34:31,981 The first cost to moral hazard is lower efficiency. 816 00:34:35,040 --> 00:34:36,540 And the best way to see this is just 817 00:34:36,540 --> 00:34:39,239 to think about the economics of the consumption/leisure 818 00:34:39,239 --> 00:34:41,960 trade-off. 819 00:34:41,960 --> 00:34:44,630 Think about how I make my decision of how hard to work. 820 00:34:47,510 --> 00:34:51,050 Basically, if there's no insurance, no social insurance, 821 00:34:51,050 --> 00:34:53,790 no workers' comp, no unemployment insurance, 822 00:34:53,790 --> 00:34:58,410 how do I choose how hard to work? 823 00:34:58,410 --> 00:35:01,350 How do I choose how hard to work? 824 00:35:01,350 --> 00:35:02,150 How do I do that? 825 00:35:02,150 --> 00:35:05,050 What's the trade-off I consider in deciding how hard to work? 826 00:35:07,890 --> 00:35:08,543 Yeah. 827 00:35:08,543 --> 00:35:09,450 AUDIENCE: Consumption versus leisure. 828 00:35:09,450 --> 00:35:11,283 JONATHAN GRUBER: Consumption versus leisure, 829 00:35:11,283 --> 00:35:14,220 in particular, I will trade them off 830 00:35:14,220 --> 00:35:17,440 until the marginal value of the next hour of leisure-- 831 00:35:17,440 --> 00:35:18,870 marginal value of leisure-- 832 00:35:18,870 --> 00:35:21,750 equals the wage. 833 00:35:21,750 --> 00:35:24,620 Because the marginal value of leisure is above the wage, 834 00:35:24,620 --> 00:35:25,620 I should work less hard. 835 00:35:25,620 --> 00:35:27,630 That means I'd rather be at home. 836 00:35:27,630 --> 00:35:29,757 If the marginal value of leisure is below the wage, 837 00:35:29,757 --> 00:35:31,590 that means I'm just wasting my time at home. 838 00:35:31,590 --> 00:35:33,243 I should work harder. 839 00:35:33,243 --> 00:35:35,160 So I'll continue to trade off work and leisure 840 00:35:35,160 --> 00:35:37,050 till the next hour of leisure makes 841 00:35:37,050 --> 00:35:40,590 me just as happy as the next hour of working. 842 00:35:40,590 --> 00:35:43,330 And that is the efficient outcome. 843 00:35:43,330 --> 00:35:45,010 That is the socially efficient outcome, 844 00:35:45,010 --> 00:35:47,050 because leisure is not a social bad. 845 00:35:47,050 --> 00:35:49,540 There's nothing wrong with leisure. 846 00:35:49,540 --> 00:35:51,100 People value leisure. 847 00:35:51,100 --> 00:35:53,290 They should get to trade off the leisure versus what 848 00:35:53,290 --> 00:35:55,623 they get from working till they choose the right amount. 849 00:35:55,623 --> 00:35:57,930 That's what makes society best off. 850 00:35:57,930 --> 00:36:02,370 Now, what happens if I say, if you sit at home, 851 00:36:02,370 --> 00:36:06,460 you're also going to get a check from the government? 852 00:36:06,460 --> 00:36:08,530 Now, what's my new equation? 853 00:36:08,530 --> 00:36:11,370 Now, if I work, I still get the wage. 854 00:36:11,370 --> 00:36:13,110 But what happens if I sit at home? 855 00:36:13,110 --> 00:36:17,100 I get the marginal value of leisure plus the government 856 00:36:17,100 --> 00:36:19,770 check. 857 00:36:19,770 --> 00:36:24,538 So now I sit at home until this equation is true, 858 00:36:24,538 --> 00:36:26,830 which means that I sit at home until the marginal value 859 00:36:26,830 --> 00:36:31,310 of leisure equals the wage minus the government transfer. 860 00:36:34,380 --> 00:36:36,810 That means the marginal value of leisure 861 00:36:36,810 --> 00:36:40,980 will be lower than it would be without the government 862 00:36:40,980 --> 00:36:42,600 transfer, which means I work what? 863 00:36:42,600 --> 00:36:44,490 More hard or less hard? 864 00:36:44,490 --> 00:36:47,157 If the marginal value of leisure is forced down, 865 00:36:47,157 --> 00:36:48,240 that means I'm doing what? 866 00:36:51,583 --> 00:36:52,666 Someone raised their hand. 867 00:36:52,666 --> 00:36:52,893 Yeah. 868 00:36:52,893 --> 00:36:53,600 AUDIENCE: More leisure. 869 00:36:53,600 --> 00:36:55,250 JONATHAN GRUBER: More leisure, because remember, 870 00:36:55,250 --> 00:36:57,292 there's diminishing marginal value of everything, 871 00:36:57,292 --> 00:37:00,050 so I'm taking more leisure, less work. 872 00:37:00,050 --> 00:37:02,450 So the government is causing me to work less 873 00:37:02,450 --> 00:37:04,790 by essentially saying, look, I'm going to reward 874 00:37:04,790 --> 00:37:07,830 you more for staying at home. 875 00:37:07,830 --> 00:37:09,120 What does that do? 876 00:37:09,120 --> 00:37:13,050 That means that people work less than is socially optimal. 877 00:37:13,050 --> 00:37:15,990 This is the social optimum. 878 00:37:15,990 --> 00:37:18,210 This means people are taking more leisure 879 00:37:18,210 --> 00:37:20,760 and working less than is socially optimal. 880 00:37:20,760 --> 00:37:23,910 When people work less, that shifts in the supply curve 881 00:37:23,910 --> 00:37:26,590 and creates a deadweight loss. 882 00:37:26,590 --> 00:37:30,040 Social welfare has fallen. 883 00:37:30,040 --> 00:37:32,260 Let me remind you, it's not falling because people 884 00:37:32,260 --> 00:37:34,090 take some time off. 885 00:37:34,090 --> 00:37:36,580 Many people on the conservative side of the spectrum 886 00:37:36,580 --> 00:37:38,620 will act as if work is a virtue. 887 00:37:38,620 --> 00:37:41,380 Work is not a virtue. 888 00:37:41,380 --> 00:37:44,432 The optimal solution is to work until your value of working 889 00:37:44,432 --> 00:37:45,640 equals your value of leisure. 890 00:37:45,640 --> 00:37:48,520 If you're someone who has a job that you hate 891 00:37:48,520 --> 00:37:51,590 and doesn't pay well, and you love watching TV, 892 00:37:51,590 --> 00:37:52,660 you should work less. 893 00:37:52,660 --> 00:37:55,270 That's what's optimal for society. 894 00:37:55,270 --> 00:37:58,360 But you shouldn't work even less because the government's 895 00:37:58,360 --> 00:38:00,150 paying you to stay home. 896 00:38:00,150 --> 00:38:02,530 That reduces efficiency. 897 00:38:02,530 --> 00:38:05,238 So that's a problem of moral hazard is it lowers efficiency. 898 00:38:05,238 --> 00:38:06,780 There is a second problem, of course, 899 00:38:06,780 --> 00:38:10,840 of moral hazard, which is if you work less, 900 00:38:10,840 --> 00:38:14,140 then we have to tax people who do work more to pay 901 00:38:14,140 --> 00:38:15,590 for these programs. 902 00:38:15,590 --> 00:38:21,260 So it raises taxation, raises the required tax revenues, 903 00:38:21,260 --> 00:38:23,380 raises the tax revenues required. 904 00:38:23,380 --> 00:38:25,290 Because if you're sitting at home more, 905 00:38:25,290 --> 00:38:28,020 I've got to make more money to pay for you to sit at home. 906 00:38:28,020 --> 00:38:31,220 And we know taxation also causes deadweight loss. 907 00:38:31,220 --> 00:38:33,230 So it's a double win. 908 00:38:33,230 --> 00:38:35,157 I cause you to stay at home and I 909 00:38:35,157 --> 00:38:37,615 cause other people to have to pay more taxes to pay for you 910 00:38:37,615 --> 00:38:41,630 to sit at home, which causes them to work less, too. 911 00:38:41,630 --> 00:38:44,170 There's a second round effect. 912 00:38:44,170 --> 00:38:48,680 As a result, moral hazard causes inefficiency in society. 913 00:38:48,680 --> 00:38:50,090 And that is the trade-off. 914 00:38:50,090 --> 00:38:52,090 Once again, I told you, this course is annoying. 915 00:38:52,090 --> 00:38:53,423 We don't give you right answers. 916 00:38:53,423 --> 00:38:54,830 We just tell you trade-offs. 917 00:38:54,830 --> 00:38:59,300 The trade-off here is we need programs like unemployment 918 00:38:59,300 --> 00:39:01,133 insurance, because otherwise-- 919 00:39:01,133 --> 00:39:03,050 let's take the case of unemployment insurance. 920 00:39:03,050 --> 00:39:04,467 We'll go through it one more time. 921 00:39:04,467 --> 00:39:06,710 Imagine there's no government unemployment insurance, 922 00:39:06,710 --> 00:39:07,835 and you said, that's great. 923 00:39:07,835 --> 00:39:10,010 I'll offer private unemployment insurance. 924 00:39:10,010 --> 00:39:11,610 Well, that's not going to work. 925 00:39:11,610 --> 00:39:12,200 Why? 926 00:39:12,200 --> 00:39:13,610 Because people know way more than 927 00:39:13,610 --> 00:39:16,270 you do about whether they're going to lose their job. 928 00:39:16,270 --> 00:39:18,370 If you tried to offer private insurance, 929 00:39:18,370 --> 00:39:21,690 you'd lose your shirt because of adverse selection. 930 00:39:21,690 --> 00:39:24,050 So absent government-provided unemployment insurance, 931 00:39:24,050 --> 00:39:25,758 there would be no unemployment insurance. 932 00:39:25,758 --> 00:39:26,880 And that would be bad. 933 00:39:26,880 --> 00:39:28,850 That would mean people would be subject to a risk that 934 00:39:28,850 --> 00:39:30,433 would drive their consumption to zero. 935 00:39:30,433 --> 00:39:32,677 Remember, most Americans have no savings. 936 00:39:32,677 --> 00:39:34,510 That mean Americans would be subject to risk 937 00:39:34,510 --> 00:39:37,580 where if they lost their job, they would starve. 938 00:39:37,580 --> 00:39:39,570 That's a very bad outcome. 939 00:39:39,570 --> 00:39:42,380 So it is socially valuable to insure 940 00:39:42,380 --> 00:39:44,460 against unemployment risk. 941 00:39:44,460 --> 00:39:48,240 The private market can't do it because of adverse selection. 942 00:39:48,240 --> 00:39:51,660 Therefore, there is a compelling case for government 943 00:39:51,660 --> 00:39:53,370 unemployment insurance. 944 00:39:53,370 --> 00:39:55,595 But with government-provided unemployment insurance, 945 00:39:55,595 --> 00:39:57,220 that causes people to sit at home extra 946 00:39:57,220 --> 00:39:58,710 and not work as hard. 947 00:39:58,710 --> 00:40:00,800 And that's the sort of chain of logic 948 00:40:00,800 --> 00:40:03,430 which teaches you the trade-off. 949 00:40:03,430 --> 00:40:06,190 What this says is optimal social insurance-- 950 00:40:06,190 --> 00:40:07,940 that in these markets, we're going 951 00:40:07,940 --> 00:40:12,270 to want some social insurance, but not too much. 952 00:40:12,270 --> 00:40:14,210 We're going to want enough to protect people 953 00:40:14,210 --> 00:40:16,702 against starving, but not so much that it 954 00:40:16,702 --> 00:40:17,910 causes people to sit at home. 955 00:40:17,910 --> 00:40:19,487 So for example, if I told you I'm 956 00:40:19,487 --> 00:40:21,570 going to set up an unemployment insurance program, 957 00:40:21,570 --> 00:40:24,170 and the way it's going to work is if you lose your job, 958 00:40:24,170 --> 00:40:26,900 I'm going to pay your entire wage for as long 959 00:40:26,900 --> 00:40:30,380 as you need until you find a new job, that would not 960 00:40:30,380 --> 00:40:32,650 be a good idea. 961 00:40:32,650 --> 00:40:34,990 That would cause a huge amount of moral hazard. 962 00:40:34,990 --> 00:40:38,080 And remember, compare that program to one 963 00:40:38,080 --> 00:40:42,340 where I'll pay you 50% of your wage till you find a new job. 964 00:40:42,340 --> 00:40:46,183 Well, 50%, going from 0% to 50%, 0% 965 00:40:46,183 --> 00:40:48,850 of your wage to 50% of your wage is a huge consumption smoothing 966 00:40:48,850 --> 00:40:49,350 benefit. 967 00:40:49,350 --> 00:40:51,550 You go from starving to being able to eat decently. 968 00:40:51,550 --> 00:40:54,640 50% to 100% is an increase, but not as much. 969 00:40:54,640 --> 00:40:57,803 But 50% to 100% has a huge moral hazard effect. 970 00:40:57,803 --> 00:40:59,720 So you want something more towards the middle, 971 00:40:59,720 --> 00:41:02,030 where you're getting people away from starving, but not 972 00:41:02,030 --> 00:41:05,450 so much that they don't work. 973 00:41:05,450 --> 00:41:06,710 So that's the trade-off. 974 00:41:06,710 --> 00:41:09,480 So let's talk about that trade in practice. 975 00:41:09,480 --> 00:41:12,230 Let's talk about the US Social Security program. 976 00:41:17,480 --> 00:41:20,900 The Social Security program in the US-- 977 00:41:20,900 --> 00:41:24,590 Social Security is our biggest single social insurance program 978 00:41:24,590 --> 00:41:25,970 in the US. 979 00:41:25,970 --> 00:41:31,310 Currently, the Social Security program is about $800 billion 980 00:41:31,310 --> 00:41:32,420 per year. 981 00:41:32,420 --> 00:41:33,300 That's real money. 982 00:41:33,300 --> 00:41:38,640 That's even more than Jeff Bezos has, $800 billion a year. 983 00:41:38,640 --> 00:41:41,537 That's more than he's worth, every year. 984 00:41:41,537 --> 00:41:42,620 What does this program do? 985 00:41:42,620 --> 00:41:45,230 What this program does, in a nutshell, 986 00:41:45,230 --> 00:41:50,000 is it insures you against the income loss you're 987 00:41:50,000 --> 00:41:52,790 going to face when you retire. 988 00:41:52,790 --> 00:41:54,830 When people retire, they suddenly 989 00:41:54,830 --> 00:41:58,510 go from having a lot of income to having no income. 990 00:41:58,510 --> 00:42:00,260 And basically, the idea of Social Security 991 00:42:00,260 --> 00:42:03,060 is to make sure you don't starve when you're old. 992 00:42:03,060 --> 00:42:06,090 So the way it works is you pay a tax. 993 00:42:06,090 --> 00:42:08,700 And if you ever see a line on your pay stub that says FICA, 994 00:42:08,700 --> 00:42:10,200 that's what this is for. 995 00:42:10,200 --> 00:42:12,400 You pay a FICA tax. 996 00:42:12,400 --> 00:42:16,563 It is 12.4% of payroll, half on you, half on your employer. 997 00:42:16,563 --> 00:42:18,730 But it doesn't matter that half is on your employer, 998 00:42:18,730 --> 00:42:20,647 because we learned two lectures ago it doesn't 999 00:42:20,647 --> 00:42:22,300 matter who pays the tax. 1000 00:42:22,300 --> 00:42:23,850 It's a 12.4% tax. 1001 00:42:23,850 --> 00:42:27,010 That's what matters on you. 1002 00:42:27,010 --> 00:42:29,500 That money then provides that when 1003 00:42:29,500 --> 00:42:32,920 you retire, starting at age 62, you get a check 1004 00:42:32,920 --> 00:42:34,060 from the government. 1005 00:42:34,060 --> 00:42:35,890 And you get a check from the government 1006 00:42:35,890 --> 00:42:37,288 that lasts until you die. 1007 00:42:37,288 --> 00:42:38,830 The check from the government you get 1008 00:42:38,830 --> 00:42:41,470 is what's called an "annuity." 1009 00:42:41,470 --> 00:42:42,670 An annuity is a payment. 1010 00:42:42,670 --> 00:42:45,080 Annuities are the opposite of life insurance. 1011 00:42:45,080 --> 00:42:47,890 Life insurance is money that your family gets when you die. 1012 00:42:47,890 --> 00:42:51,890 Annuity is a regular payment you get until you die. 1013 00:42:51,890 --> 00:42:54,500 The way it works, you pay 12 and 1/2 percent of your income 1014 00:42:54,500 --> 00:42:56,750 all the way through your working life till we turn 62, 1015 00:42:56,750 --> 00:42:59,300 or you can collect it later. 1016 00:42:59,300 --> 00:43:02,870 You then get a payment for the rest of your life. 1017 00:43:02,870 --> 00:43:07,310 That payment is typically about half 1018 00:43:07,310 --> 00:43:10,540 of what you made when you were working, 1019 00:43:10,540 --> 00:43:13,930 but it's very progressive, in the sense 1020 00:43:13,930 --> 00:43:17,305 that for someone who's very poor, 1021 00:43:17,305 --> 00:43:19,430 it would probably more than half of what they made. 1022 00:43:19,430 --> 00:43:21,097 For someone who's rich, it would be much 1023 00:43:21,097 --> 00:43:22,420 less than half what they made. 1024 00:43:22,420 --> 00:43:23,710 It's a progressive payment. 1025 00:43:23,710 --> 00:43:24,910 Everyone gets it. 1026 00:43:24,910 --> 00:43:26,320 Everyone gets Social Security. 1027 00:43:26,320 --> 00:43:28,570 But how much you get from it depends on your income. 1028 00:43:28,570 --> 00:43:30,112 The poorer you are, the more generous 1029 00:43:30,112 --> 00:43:34,260 it is relatively when you retire. 1030 00:43:34,260 --> 00:43:36,170 Now the-- yeah. 1031 00:43:36,170 --> 00:43:39,402 AUDIENCE: Is it possible as life expectancies get larger, 1032 00:43:39,402 --> 00:43:42,923 it's going to be harder to have Social Security 1033 00:43:42,923 --> 00:43:44,090 because if you're going to-- 1034 00:43:44,090 --> 00:43:45,440 JONATHAN GRUBER: That's a huge problem. 1035 00:43:45,440 --> 00:43:47,900 It sounds like you should definitely be enrolled in 1441. 1036 00:43:47,900 --> 00:43:50,190 That's a whole half a lecture we spend on that. 1037 00:43:50,190 --> 00:43:52,357 I don't have time to talk about it here, but clearly 1038 00:43:52,357 --> 00:43:53,510 this is a huge, huge-- 1039 00:43:53,510 --> 00:43:55,250 so just to give you a couple numbers just 1040 00:43:55,250 --> 00:43:56,430 to keep you up at night. 1041 00:43:56,430 --> 00:43:58,430 We all know, we're all talking about the deficit 1042 00:43:58,430 --> 00:43:59,370 is $500 billion. 1043 00:43:59,370 --> 00:44:00,590 It's a big deal. 1044 00:44:00,590 --> 00:44:04,400 If you ask how much has America promised 1045 00:44:04,400 --> 00:44:08,960 to pay to our senior citizens over the foreseeable future 1046 00:44:08,960 --> 00:44:11,060 minus how much we'll collected taxes, 1047 00:44:11,060 --> 00:44:14,880 we are currently, as a nation, $75 trillion in debt. 1048 00:44:14,880 --> 00:44:17,880 And it's because of the aging society and things like that. 1049 00:44:17,880 --> 00:44:19,713 We've got big problems coming down the road. 1050 00:44:19,713 --> 00:44:21,440 We can talk about that another time. 1051 00:44:21,440 --> 00:44:23,900 But let's focus on the program itself at a point in time 1052 00:44:23,900 --> 00:44:26,150 right now. 1053 00:44:26,150 --> 00:44:30,020 So basically, we see here the moral hazard trade-off. 1054 00:44:30,020 --> 00:44:31,850 On the one hand, we don't want people 1055 00:44:31,850 --> 00:44:33,230 to starve when they're old. 1056 00:44:33,230 --> 00:44:35,840 On the other hand, if I pay you once you're retired, 1057 00:44:35,840 --> 00:44:38,180 that could cause you to retire. 1058 00:44:38,180 --> 00:44:40,190 If I say, once you're retired you're 1059 00:44:40,190 --> 00:44:42,485 going to get a check, 50% of your wage, you might say, 1060 00:44:42,485 --> 00:44:44,360 50% of the wage isn't that much, but I really 1061 00:44:44,360 --> 00:44:45,152 don't like working. 1062 00:44:45,152 --> 00:44:48,860 I'd rather just hit the links at 50% of my wage. 1063 00:44:48,860 --> 00:44:51,710 So that's the trade-off. 1064 00:44:51,710 --> 00:44:54,773 Now, how do we think about evaluating that trade-off? 1065 00:44:54,773 --> 00:44:56,690 Evaluating that trade-off, different countries 1066 00:44:56,690 --> 00:44:58,640 think about it differently. 1067 00:44:58,640 --> 00:45:01,580 In the US, we think about it in what I would say 1068 00:45:01,580 --> 00:45:05,420 is a fairly rational way, which is let's consider your decision 1069 00:45:05,420 --> 00:45:08,658 to retire at 62 versus 63. 1070 00:45:08,658 --> 00:45:10,200 The way it works in the United States 1071 00:45:10,200 --> 00:45:15,060 is we say, look, if you work one year more, since it's 1072 00:45:15,060 --> 00:45:17,797 an annuity, you will get one less year off payment. 1073 00:45:17,797 --> 00:45:19,380 If you start at one year later, you're 1074 00:45:19,380 --> 00:45:22,080 going to die the same time, you get one year less. 1075 00:45:22,080 --> 00:45:24,420 So what we do is we pay you more every month. 1076 00:45:24,420 --> 00:45:26,520 Indeed, for every year you delay, 1077 00:45:26,520 --> 00:45:30,190 you get 6.7% more every month, reflecting the trade-off 1078 00:45:30,190 --> 00:45:33,080 that you're going to get it for a shorter period of time. 1079 00:45:33,080 --> 00:45:35,650 And that turns out to be roughly fair. 1080 00:45:35,650 --> 00:45:37,780 Given the expected life of Americans, 1081 00:45:37,780 --> 00:45:40,270 that's a roughly fair trade-off. 1082 00:45:40,270 --> 00:45:43,310 Every year you delay gets 6.7% more. 1083 00:45:43,310 --> 00:45:46,540 In Europe, they don't have this. 1084 00:45:46,540 --> 00:45:50,300 So every year delay, you just get less money before you die. 1085 00:45:50,300 --> 00:45:53,040 So let's take the example of the Netherlands. 1086 00:45:53,040 --> 00:45:56,730 In the Netherlands, you can retire at 55 1087 00:45:56,730 --> 00:46:00,760 with a benefit that is 90% of what you made. 1088 00:46:00,760 --> 00:46:04,450 So if you earned $30,000, you can retire at 55 1089 00:46:04,450 --> 00:46:07,510 with a benefit of $27,000. 1090 00:46:07,510 --> 00:46:11,530 And if you decide to work instead, 1091 00:46:11,530 --> 00:46:13,310 you just forego that $27,000. 1092 00:46:13,310 --> 00:46:14,980 There's no bump up of your benefits. 1093 00:46:14,980 --> 00:46:17,938 That's just one less year of $27,000 you get. 1094 00:46:17,938 --> 00:46:19,980 So what that means, if you're in the Netherlands, 1095 00:46:19,980 --> 00:46:23,010 your choice is work and get 30,000 or stay home 1096 00:46:23,010 --> 00:46:25,080 and get 27,000. 1097 00:46:25,080 --> 00:46:28,200 In other words, it's sort of like a 90% tax. 1098 00:46:28,200 --> 00:46:31,110 Think about it-- by working relative to staying home, 1099 00:46:31,110 --> 00:46:34,050 I'm only keeping 3,000 of the 30,000 I made. 1100 00:46:34,050 --> 00:46:36,690 It's basically like a 90% tax. 1101 00:46:36,690 --> 00:46:39,180 But that's not all. 1102 00:46:39,180 --> 00:46:41,130 How do they pay for this program? 1103 00:46:41,130 --> 00:46:41,838 They tax people. 1104 00:46:41,838 --> 00:46:43,380 You can't tax people who are at home. 1105 00:46:43,380 --> 00:46:45,090 You've got to tax workers. 1106 00:46:45,090 --> 00:46:49,530 So if you work, you also have to pay a 45% tax 1107 00:46:49,530 --> 00:46:52,710 to pay for this program and lots of other things. 1108 00:46:52,710 --> 00:46:56,140 45% tax is a high tax rate for everything. 1109 00:46:56,140 --> 00:47:01,260 What that means is if you stay at home, you get 27,000. 1110 00:47:01,260 --> 00:47:03,780 If you go to work, you get 30,000 times 1111 00:47:03,780 --> 00:47:10,120 1.55, or about 18,000. 1112 00:47:10,120 --> 00:47:12,430 That's 1.6, It's 16, 5. 1113 00:47:12,430 --> 00:47:17,710 So your choice is stay at home and get 27, work and get 16, 5. 1114 00:47:17,710 --> 00:47:18,460 Guess what? 1115 00:47:18,460 --> 00:47:20,410 No one works. 1116 00:47:20,410 --> 00:47:23,812 No one over 55 in the Netherlands works, like zero. 1117 00:47:23,812 --> 00:47:26,020 And they might work on the black market and ways they 1118 00:47:26,020 --> 00:47:27,360 don't report to the government. 1119 00:47:27,360 --> 00:47:29,700 But basically, they just sort of sit around coffee shops 1120 00:47:29,700 --> 00:47:32,640 and spend their retirement money. 1121 00:47:32,640 --> 00:47:35,280 So here's a case where they've made a very different decision 1122 00:47:35,280 --> 00:47:37,110 about how to make this trade-off, which 1123 00:47:37,110 --> 00:47:41,250 is it's a pretty sweet life for the elderly in the Netherlands, 1124 00:47:41,250 --> 00:47:44,190 but no one's working over age 55. 1125 00:47:44,190 --> 00:47:46,440 And that's a different way to resolve this trade-off. 1126 00:47:46,440 --> 00:47:51,270 So basically, this illustrates different design features 1127 00:47:51,270 --> 00:47:52,255 of the program. 1128 00:47:52,255 --> 00:47:53,880 What makes the Netherlands program have 1129 00:47:53,880 --> 00:47:57,030 much more moral hazard than the US is the benefits are higher 1130 00:47:57,030 --> 00:48:00,050 and they don't increase your benefits if you work more. 1131 00:48:00,050 --> 00:48:03,230 So essentially, these are little kind of tweaky details 1132 00:48:03,230 --> 00:48:04,980 that turn out to matter enormously for how 1133 00:48:04,980 --> 00:48:06,957 we think about the program. 1134 00:48:06,957 --> 00:48:08,540 Now, I hope you find that interesting. 1135 00:48:08,540 --> 00:48:10,100 That was a lot to put in one lecture. 1136 00:48:10,100 --> 00:48:11,850 Like I said, if you find this interesting, 1137 00:48:11,850 --> 00:48:14,880 there's the whole third of a semester in my class 14.41. 1138 00:48:14,880 --> 00:48:15,690 So take that. 1139 00:48:15,690 --> 00:48:18,070 We can learn a lot more about it.