1 00:00:00,000 --> 00:00:02,400 [SQUEAKING] 2 00:00:02,400 --> 00:00:03,360 [RUSTLING] 3 00:00:03,360 --> 00:00:05,760 [CLICKING] 4 00:00:10,570 --> 00:00:13,348 FRANK SCHILBACH: Welcome to the mid-term review of 14.13. 5 00:00:13,348 --> 00:00:14,890 Tonight, as I mentioned before, there 6 00:00:14,890 --> 00:00:18,500 will be three types of questions on the exam. 7 00:00:18,500 --> 00:00:21,050 There will be true/false/uncertain questions. 8 00:00:21,050 --> 00:00:23,800 These are questions where you're given a statement, 9 00:00:23,800 --> 00:00:26,440 and you can ascertain whether the statement is 10 00:00:26,440 --> 00:00:28,460 true or false or uncertain. 11 00:00:28,460 --> 00:00:30,250 True means essentially it's strictly true 12 00:00:30,250 --> 00:00:32,840 and it's always true, or false means it's strictly false, 13 00:00:32,840 --> 00:00:36,250 and there's a counterexample that's shows you it's false. 14 00:00:36,250 --> 00:00:41,300 And uncertain means like without having further information, 15 00:00:41,300 --> 00:00:43,220 you cannot actually answer the questions. 16 00:00:43,220 --> 00:00:45,640 So you cannot really make a statement whether that 17 00:00:45,640 --> 00:00:48,610 statement is true or false. 18 00:00:48,610 --> 00:00:52,480 In any of the ones you do, please always explain 19 00:00:52,480 --> 00:00:54,040 your answers carefully. 20 00:00:54,040 --> 00:00:56,685 So just writing true, false, or uncertain 21 00:00:56,685 --> 00:00:59,110 will almost surely or will surely not 22 00:00:59,110 --> 00:01:00,670 give you full credit. 23 00:01:00,670 --> 00:01:02,830 You have to explain your answers. 24 00:01:02,830 --> 00:01:06,450 It doesn't actually matter that much whether you get it right, 25 00:01:06,450 --> 00:01:08,950 whether the statement is true, or false, or uncertain, 26 00:01:08,950 --> 00:01:13,202 as long as your answer is actually reasonable and shows 27 00:01:13,202 --> 00:01:14,660 some understanding of the material. 28 00:01:17,360 --> 00:01:21,260 So the quality of your answer is quite important. 29 00:01:21,260 --> 00:01:23,290 We also want you to provide intuitions. 30 00:01:23,290 --> 00:01:27,340 We do not want you to just write some math and nothing else. 31 00:01:27,340 --> 00:01:31,120 You can use some math to sort of clarify 32 00:01:31,120 --> 00:01:33,460 your answer or the like, but you always 33 00:01:33,460 --> 00:01:36,610 need to provide a verbal explanation and some intuition 34 00:01:36,610 --> 00:01:40,970 and some explanation of why you think the answer might be 35 00:01:40,970 --> 00:01:42,830 or the statement might be true, false, 36 00:01:42,830 --> 00:01:45,478 or [AUDIO OUT] your answer. 37 00:01:45,478 --> 00:01:47,520 For multiple choice questions, you're essentially 38 00:01:47,520 --> 00:01:49,650 given different choices, and you just 39 00:01:49,650 --> 00:01:51,810 pick one answer with no further explanation needed. 40 00:01:51,810 --> 00:01:53,105 Very simple. 41 00:01:53,105 --> 00:01:55,230 And then there will be sort of pset-style questions 42 00:01:55,230 --> 00:01:58,120 that sort of you're familiar with, either from like psets 43 00:01:58,120 --> 00:02:01,080 or like previous exams. 44 00:02:01,080 --> 00:02:03,250 These will be quite similar to the psets questions. 45 00:02:03,250 --> 00:02:06,480 It will be sort of generally not on the harder side, 46 00:02:06,480 --> 00:02:08,740 just, you know, we only have so much time to do this. 47 00:02:08,740 --> 00:02:10,530 There will be some algebra involved. 48 00:02:10,530 --> 00:02:16,900 Again, please always explain your answers carefully. 49 00:02:16,900 --> 00:02:20,650 Now, what materials are you responsible for? 50 00:02:20,650 --> 00:02:23,380 You're responsible for the lectures up to and including 51 00:02:23,380 --> 00:02:26,710 lecture 12, which was the lecture on March 11, 52 00:02:26,710 --> 00:02:30,985 up to slide 67 of lectures 11, or lecture-- 53 00:02:30,985 --> 00:02:34,120 the notes for the lecture, the handout of the notes for 54 00:02:34,120 --> 00:02:35,930 lectures 11 through 13. 55 00:02:35,930 --> 00:02:39,340 You're also responsible for recitations 1 through 5, 56 00:02:39,340 --> 00:02:44,680 recitations 6, now, which the recitation that Will gave, 57 00:02:44,680 --> 00:02:47,200 which is quite similar to this one, and 7. 58 00:02:47,200 --> 00:02:49,120 These recitations are just reviews 59 00:02:49,120 --> 00:02:51,460 that might be helpful for some of you. 60 00:02:51,460 --> 00:02:54,090 But there's no new material that we use that [INAUDIBLE] 61 00:02:54,090 --> 00:02:54,590 right now. 62 00:02:54,590 --> 00:03:00,070 Psets 1 to 3 are sort of part of the deal. 63 00:03:00,070 --> 00:03:04,120 And readings, starred or non-starred readings 64 00:03:04,120 --> 00:03:07,300 cited in class are only relevant to the extent 65 00:03:07,300 --> 00:03:11,000 that they appear in lectures and/or recitation. 66 00:03:11,000 --> 00:03:13,810 But to say, like if I discuss a certain paper and the content 67 00:03:13,810 --> 00:03:16,060 of that paper, the content of the people that's 68 00:03:16,060 --> 00:03:19,750 discussed in the class is relevant to you. 69 00:03:19,750 --> 00:03:23,230 However, anything that's in the paper that 70 00:03:23,230 --> 00:03:26,170 doesn't appear in the lecture is not relevant, 71 00:03:26,170 --> 00:03:28,570 as in like, we're not going to ask you questions that 72 00:03:28,570 --> 00:03:31,510 sort of ask about some obscure details of the papers 73 00:03:31,510 --> 00:03:35,110 that you never heard about in class. 74 00:03:35,110 --> 00:03:36,600 How do you get ready for the exam? 75 00:03:36,600 --> 00:03:41,050 Well, you study, study that the lecture and presentation slides 76 00:03:41,050 --> 00:03:42,580 carefully. 77 00:03:42,580 --> 00:03:45,880 You should make sure you're familiar and comfortable 78 00:03:45,880 --> 00:03:48,190 with the psets and the solutions make 79 00:03:48,190 --> 00:03:51,730 you understand and are able to solve the psets on your own. 80 00:03:51,730 --> 00:03:56,650 A great resource practice is to do those psets and exams. 81 00:03:56,650 --> 00:04:01,558 Again, readings, starred and non-starred, are not required, 82 00:04:01,558 --> 00:04:03,850 but they may sort of help you deepen your understanding 83 00:04:03,850 --> 00:04:05,850 of the material, and maybe sometimes the lecture 84 00:04:05,850 --> 00:04:07,560 notes a little bit. 85 00:04:07,560 --> 00:04:10,090 [AUDIO OUT] dense and don't have that much detail. 86 00:04:10,090 --> 00:04:12,400 So you can sort of like try to consult them, and try 87 00:04:12,400 --> 00:04:14,380 to understand the material as best or better. 88 00:04:14,380 --> 00:04:21,208 But we won't ask you about the details of those readings 89 00:04:21,208 --> 00:04:25,246 beyond those details of was covered in class. 90 00:04:25,246 --> 00:04:27,735 So let me now sort of review the material overall and sort 91 00:04:27,735 --> 00:04:29,860 of give you a sense of what are the kinds of things 92 00:04:29,860 --> 00:04:31,180 that we want to know. 93 00:04:31,180 --> 00:04:33,250 To be clear, this is not exhaustive 94 00:04:33,250 --> 00:04:36,135 in the sense of like, there are some other materials 95 00:04:36,135 --> 00:04:38,260 in the lecture slides that I'm not mentioning here. 96 00:04:38,260 --> 00:04:39,460 What I'm trying to do here is just 97 00:04:39,460 --> 00:04:40,877 give you a sense for a lot of what 98 00:04:40,877 --> 00:04:43,970 the most important key things that you should know. 99 00:04:43,970 --> 00:04:45,905 This will cover perhaps like 80%, 100 00:04:45,905 --> 00:04:47,920 90% of the material that's actually 101 00:04:47,920 --> 00:04:49,570 potentially in the exam. 102 00:04:49,570 --> 00:04:52,570 Having said that, you know you should really 103 00:04:52,570 --> 00:04:54,130 go through the entire lecture slides, 104 00:04:54,130 --> 00:04:56,200 make sure you're sort of gathering everything 105 00:04:56,200 --> 00:04:58,650 in understanding what's going on there. 106 00:04:58,650 --> 00:04:59,370 OK. 107 00:04:59,370 --> 00:05:02,460 So the first thing we discuss, these time preferences 108 00:05:02,460 --> 00:05:05,623 in particular, the exponential discounting model. 109 00:05:05,623 --> 00:05:08,040 Here, you should know, what is the exponential discounting 110 00:05:08,040 --> 00:05:08,820 model? 111 00:05:08,820 --> 00:05:11,010 What is delta, the time preference parameter 112 00:05:11,010 --> 00:05:12,000 that model? 113 00:05:12,000 --> 00:05:13,210 What does it measure? 114 00:05:13,210 --> 00:05:15,900 How can we estimate it, assuming that an exponential discounting 115 00:05:15,900 --> 00:05:18,090 model is correct? 116 00:05:18,090 --> 00:05:20,100 Further, what are the main assumptions 117 00:05:20,100 --> 00:05:23,760 of exponential discounting model, 118 00:05:23,760 --> 00:05:26,430 and what evidence do we have against those assumptions? 119 00:05:26,430 --> 00:05:30,270 So lectures 3 and 4 discussed those assumptions 120 00:05:30,270 --> 00:05:35,220 and the evidence against these assumptions in detail. 121 00:05:35,220 --> 00:05:38,280 And after that, we discussed the quasi-hyperbolic discounting 122 00:05:38,280 --> 00:05:41,340 model, which is a relatively small modification 123 00:05:41,340 --> 00:05:43,440 of the exponential discounting model. 124 00:05:43,440 --> 00:05:45,390 Again, you know what is this model? 125 00:05:45,390 --> 00:05:48,220 How is it different from the exponential discounting model? 126 00:05:48,220 --> 00:05:50,160 And of course, the difference is that there's 127 00:05:50,160 --> 00:05:53,460 a present bias parameter of beta that 128 00:05:53,460 --> 00:05:59,940 is added to this model, which measures short-run discounting, 129 00:05:59,940 --> 00:06:02,310 as in like, how much weight you put on the present, 130 00:06:02,310 --> 00:06:04,920 versus everything else that's in the future. 131 00:06:04,920 --> 00:06:10,080 Now, what empirical evidence can the quasi-exponential, 132 00:06:10,080 --> 00:06:13,020 quasi-hyperbolic model explain better 133 00:06:13,020 --> 00:06:15,577 than the exponential discounting model, and why is that? 134 00:06:15,577 --> 00:06:17,410 So you should be familiar with, for example, 135 00:06:17,410 --> 00:06:19,618 things like preference reversals and then [INAUDIBLE] 136 00:06:19,618 --> 00:06:20,220 commitment. 137 00:06:20,220 --> 00:06:23,040 Why is that consistent with the quasi-hyperbolic model? 138 00:06:23,040 --> 00:06:25,275 And you know, why is it not consistent 139 00:06:25,275 --> 00:06:26,400 with the exponential model? 140 00:06:26,400 --> 00:06:27,375 Of course, that's what we discussed 141 00:06:27,375 --> 00:06:29,760 in the previous slide, is the somewhat the assumptions 142 00:06:29,760 --> 00:06:33,360 of the exponential discount model, where for example, one 143 00:06:33,360 --> 00:06:35,980 of the key assumptions are that there are no preference 144 00:06:35,980 --> 00:06:36,480 reversals. 145 00:06:36,480 --> 00:06:39,770 And we discussed that in detail. 146 00:06:39,770 --> 00:06:42,770 Then, of course you need to know in the course 147 00:06:42,770 --> 00:06:46,370 of hyperbolic discounting model, what is this sophistication? 148 00:06:46,370 --> 00:06:47,360 What is naivete? 149 00:06:47,360 --> 00:06:49,610 And what is partial naivete? 150 00:06:49,610 --> 00:06:51,200 So what does theta measure? 151 00:06:51,200 --> 00:06:52,730 What does theta hat measure? 152 00:06:52,730 --> 00:06:54,560 What is full sophistication, full naivete, 153 00:06:54,560 --> 00:06:56,470 and partial naivete? 154 00:06:56,470 --> 00:06:58,700 And answer questions such as like, 155 00:06:58,700 --> 00:07:01,520 does sophistication make people always better off? 156 00:07:01,520 --> 00:07:05,210 And why, or why not, if that's the case? 157 00:07:05,210 --> 00:07:09,290 Further, you should understand, what is demand for commitment? 158 00:07:09,290 --> 00:07:11,270 Who demands commitment and who doesn't? 159 00:07:11,270 --> 00:07:13,700 Or like are fully sophisticated or fully naive 160 00:07:13,700 --> 00:07:17,610 people, or partial naive people potentially demanding 161 00:07:17,610 --> 00:07:18,110 commitment? 162 00:07:18,110 --> 00:07:20,000 What are the conditions under which 163 00:07:20,000 --> 00:07:22,790 somebody demands commitment? 164 00:07:22,790 --> 00:07:29,000 And what kinds of people do not, or do demand commitment? 165 00:07:29,000 --> 00:07:32,030 And what kinds of people benefit from doing so? 166 00:07:32,030 --> 00:07:34,880 And in particular, can people be worse off 167 00:07:34,880 --> 00:07:36,995 from being offered a commitment device? 168 00:07:36,995 --> 00:07:40,650 And why is that, or why not? 169 00:07:40,650 --> 00:07:42,180 Next, we discussed in quite a bit 170 00:07:42,180 --> 00:07:44,700 of detail empirical applications from a range 171 00:07:44,700 --> 00:07:46,570 of different settings. 172 00:07:46,570 --> 00:07:48,930 This is lecture 5 and 6. 173 00:07:48,930 --> 00:07:51,120 So we want you to be familiar with 174 00:07:51,120 --> 00:07:55,437 those empirical applications. 175 00:07:55,437 --> 00:07:57,770 You should understand why the quasi-hyperbolic model can 176 00:07:57,770 --> 00:08:02,453 explain or cannot explain some of the empirical evidence 177 00:08:02,453 --> 00:08:04,370 better than the exponential discounting model. 178 00:08:07,010 --> 00:08:09,890 And I think you know , kind of like why do we think 179 00:08:09,890 --> 00:08:13,670 quasi-hyperbolic model is a good fit for some of the empirical 180 00:08:13,670 --> 00:08:17,290 examples that I have shown? 181 00:08:17,290 --> 00:08:20,200 Then you need to be able to solve problems. 182 00:08:20,200 --> 00:08:22,800 These are similar problems to the problems in the problem 183 00:08:22,800 --> 00:08:25,400 sets that you have seen before, problem sets 184 00:08:25,400 --> 00:08:28,310 for people either like exponential discounters, 185 00:08:28,310 --> 00:08:33,799 or like for beta delta equals, where beta equals 1, 186 00:08:33,799 --> 00:08:35,870 and delta being like either 1, or 0.95, 187 00:08:35,870 --> 00:08:39,394 or close to 1, because hyperbolic discounters, 188 00:08:39,394 --> 00:08:42,455 and again, fully naive, fully sophisticated, and partially 189 00:08:42,455 --> 00:08:43,490 naive agents. 190 00:08:43,490 --> 00:08:45,000 How does one solve such problems? 191 00:08:45,000 --> 00:08:46,760 Of course, we had plenty of practice 192 00:08:46,760 --> 00:08:49,970 already in the psets in the term examples, 193 00:08:49,970 --> 00:08:51,800 also it's in the finals. 194 00:08:51,800 --> 00:08:58,490 So you should use backwards or forwards 195 00:08:58,490 --> 00:09:02,660 induction or iteration, depending on the kinks. 196 00:09:02,660 --> 00:09:06,560 I sort of discuss this in slide 62 of lectures 3 and 4, 197 00:09:06,560 --> 00:09:09,500 and slide 37 of lectures 5 and 6. 198 00:09:09,500 --> 00:09:12,320 There's also a recitation that covers that in detail. 199 00:09:12,320 --> 00:09:17,260 So you should have plenty of practice of doing so. 200 00:09:17,260 --> 00:09:19,630 Now, let me now give you some examples 201 00:09:19,630 --> 00:09:22,840 of what are some examples true/false/uncertain questions, 202 00:09:22,840 --> 00:09:24,790 and how should you answer them? 203 00:09:24,790 --> 00:09:26,950 So let me sort of just read the statements for you, 204 00:09:26,950 --> 00:09:28,700 and you can think about this for a second. 205 00:09:28,700 --> 00:09:31,240 And then we're going to give you the answer. 206 00:09:31,240 --> 00:09:34,540 So consider individuals with beta delta preferences-- 207 00:09:34,540 --> 00:09:37,300 this is quasi-hyperbolic discounters-- 208 00:09:37,300 --> 00:09:40,510 who only differ by their present bias-- 209 00:09:40,510 --> 00:09:46,260 had like beta equals between 0 and 1. 210 00:09:46,260 --> 00:09:49,110 And suppose there's a commitment savings device available. 211 00:09:49,110 --> 00:09:51,300 Their willingness to pay for this commitment device 212 00:09:51,300 --> 00:09:55,270 strictly decreases in beta. 213 00:09:55,270 --> 00:09:59,530 Now, is that a true or false statement? 214 00:09:59,530 --> 00:10:04,330 The answer is this statement is false. 215 00:10:04,330 --> 00:10:05,490 Now, why is that? 216 00:10:05,490 --> 00:10:08,220 Well, there's, again, like if you just click false, 217 00:10:08,220 --> 00:10:10,080 that's not going to give you full credit, 218 00:10:10,080 --> 00:10:13,577 even if, in fact, the statement is false. 219 00:10:13,577 --> 00:10:15,660 What you need to do is you need to sort of provide 220 00:10:15,660 --> 00:10:19,140 some further explanation. 221 00:10:19,140 --> 00:10:22,110 Ideally, you'll provide us like a somewhat detailed 222 00:10:22,110 --> 00:10:26,250 explanation, so that you know if it's false, 223 00:10:26,250 --> 00:10:28,620 if you provide several examples, that's should better 224 00:10:28,620 --> 00:10:30,540 amount than just one. 225 00:10:30,540 --> 00:10:32,640 In this case, why is it false? 226 00:10:32,640 --> 00:10:35,050 Well, A, individuals might be naive, right? 227 00:10:35,050 --> 00:10:37,620 So in particular, if people are fully naive, 228 00:10:37,620 --> 00:10:40,380 they will not be willing to pay anything 229 00:10:40,380 --> 00:10:41,920 for commitment devices. 230 00:10:41,920 --> 00:10:44,740 So regardless to the beta, willingness to pay 231 00:10:44,740 --> 00:10:46,140 will always be 0. 232 00:10:46,140 --> 00:10:49,520 So it would surely not decrease in beta. 233 00:10:49,520 --> 00:10:52,100 Second, the commitment device may just not 234 00:10:52,100 --> 00:10:53,160 be effective at all. 235 00:10:53,160 --> 00:10:55,520 If the commitment device is useless, 236 00:10:55,520 --> 00:10:58,940 it doesn't matter what beta is, nobody will demand commitment 237 00:10:58,940 --> 00:11:00,290 anyway. 238 00:11:00,290 --> 00:11:02,330 And finally, even if individuals-- this 239 00:11:02,330 --> 00:11:04,130 is a little bit more trickier explanation, 240 00:11:04,130 --> 00:11:05,750 and I wouldn't necessarily expect 241 00:11:05,750 --> 00:11:08,510 you to know that answer-- but even 242 00:11:08,510 --> 00:11:10,430 if individuals are fully sophisticated 243 00:11:10,430 --> 00:11:13,400 and devices effective, the commitment device is effective, 244 00:11:13,400 --> 00:11:18,950 willingness to pay, it may not be strictly decreasing in beta. 245 00:11:18,950 --> 00:11:20,540 And here's sort of an example. 246 00:11:20,540 --> 00:11:23,720 Well, if beta equals 0 or beta equals 1, 247 00:11:23,720 --> 00:11:26,030 then individuals would be willing to pay 0 248 00:11:26,030 --> 00:11:27,590 for the commitment device, right? 249 00:11:27,590 --> 00:11:30,500 Because if beta equals 0, they don't care about the future. 250 00:11:30,500 --> 00:11:31,980 If beta equals 1, they just don't 251 00:11:31,980 --> 00:11:34,250 need any commitment devices. 252 00:11:34,250 --> 00:11:37,100 But in between, the willingness to pay 253 00:11:37,100 --> 00:11:40,640 might be positive for betas between 0 and 1. 254 00:11:40,640 --> 00:11:43,730 That essentially means that willingness to pay 255 00:11:43,730 --> 00:11:46,250 would be an inverse U in beta. 256 00:11:46,250 --> 00:11:47,990 Again, that's a bit of a tricky answer. 257 00:11:47,990 --> 00:11:50,750 Like the third bullet point, we wouldn't necessarily 258 00:11:50,750 --> 00:11:53,390 expect you to know that answer. 259 00:11:53,390 --> 00:11:58,580 But you should be able to say or to answer reason one, or reason 260 00:11:58,580 --> 00:12:02,380 two in this specific case. 261 00:12:02,380 --> 00:12:05,740 Second example statement-- fully sophisticated 262 00:12:05,740 --> 00:12:09,100 individuals can experience large welfare losses 263 00:12:09,100 --> 00:12:10,400 from their present bias. 264 00:12:13,700 --> 00:12:14,970 The answer is this is true. 265 00:12:14,970 --> 00:12:15,738 Why is that? 266 00:12:15,738 --> 00:12:17,530 I'll let you think about this for a second. 267 00:12:29,170 --> 00:12:31,840 Well, the answer is that awareness 268 00:12:31,840 --> 00:12:33,850 of present bias, that is, like sophistication, 269 00:12:33,850 --> 00:12:36,160 does not remove present bias. 270 00:12:36,160 --> 00:12:39,860 Even if people are sophisticated, 271 00:12:39,860 --> 00:12:41,650 the present bias is still there. 272 00:12:41,650 --> 00:12:43,690 There might be some commitment devices, 273 00:12:43,690 --> 00:12:48,850 some ways in which people are able to overcome the losses 274 00:12:48,850 --> 00:12:50,930 associated with present bias. 275 00:12:50,930 --> 00:12:54,100 But in the absence of commitment devices, 276 00:12:54,100 --> 00:12:56,080 people may still make suboptimal decisions. 277 00:12:56,080 --> 00:12:58,660 And some of these decisions might, in fact, 278 00:12:58,660 --> 00:13:00,490 induce large welfare needs. 279 00:13:00,490 --> 00:13:02,970 We discussed some examples in class, 280 00:13:02,970 --> 00:13:04,640 sort of numerical examples, where 281 00:13:04,640 --> 00:13:07,450 you could see that sophisticated people might actually 282 00:13:07,450 --> 00:13:11,170 be worse off than naive people, and could in fact 283 00:13:11,170 --> 00:13:14,210 suffer quite a bit from or similar to the large welfare 284 00:13:14,210 --> 00:13:14,710 losses. 285 00:13:14,710 --> 00:13:16,600 With the present bias, all of this 286 00:13:16,600 --> 00:13:19,360 is evaluated by the long-run [INAUDIBLE] ideas 287 00:13:19,360 --> 00:13:24,730 of like if somebody makes choices for the future where 288 00:13:24,730 --> 00:13:27,770 the beta is not relevant, compared 289 00:13:27,770 --> 00:13:30,150 to that kind of welfare criterion, 290 00:13:30,150 --> 00:13:34,650 people might be a lot worse off due to their present bias. 291 00:13:34,650 --> 00:13:36,820 OK, here's example number three. 292 00:13:36,820 --> 00:13:39,640 Present bias individuals always have positive demand 293 00:13:39,640 --> 00:13:42,320 for commitment devices. 294 00:13:42,320 --> 00:13:45,960 Again, the statement is false. 295 00:13:45,960 --> 00:13:46,550 Why is that? 296 00:13:46,550 --> 00:13:48,200 And we kind of discussed this already 297 00:13:48,200 --> 00:13:50,055 in the previous questions. 298 00:13:50,055 --> 00:13:51,425 Now, let me be very explicit. 299 00:13:51,425 --> 00:13:53,180 There's sort of three conditions that 300 00:13:53,180 --> 00:13:55,470 must be met for positive demand for commitment. 301 00:13:55,470 --> 00:13:57,610 I discussed this in class. 302 00:13:57,610 --> 00:13:59,570 We know that person must be present biased, 303 00:13:59,570 --> 00:14:02,533 or have some form of some self-control problems. 304 00:14:02,533 --> 00:14:03,950 Individuals must-- the person must 305 00:14:03,950 --> 00:14:07,170 be aware of the present bias, so they can't be fully naive. 306 00:14:07,170 --> 00:14:10,370 They could be partially naive, but it can't be fully naive. 307 00:14:10,370 --> 00:14:12,725 And the individual must perceive the commitment device 308 00:14:12,725 --> 00:14:16,293 as effective in helping overcome the self-control problem. 309 00:14:16,293 --> 00:14:18,710 That's to say, if somebody is offered a commitment device, 310 00:14:18,710 --> 00:14:20,660 and that commitment device is useless, 311 00:14:20,660 --> 00:14:23,570 well, there's not going to be any demand for commitment, 312 00:14:23,570 --> 00:14:26,240 particularly if the person perceives that the commitment 313 00:14:26,240 --> 00:14:28,887 device is useless. 314 00:14:28,887 --> 00:14:30,470 Notice that that person might actually 315 00:14:30,470 --> 00:14:32,590 perceive the commitment device to be effective, 316 00:14:32,590 --> 00:14:33,938 while in reality, it's not. 317 00:14:33,938 --> 00:14:36,230 So they might have some positive demand for commitment, 318 00:14:36,230 --> 00:14:38,930 even if the commitment device is in reality useless, 319 00:14:38,930 --> 00:14:41,360 because their beliefs are wrong, because of essentially 320 00:14:41,360 --> 00:14:44,690 some form of partial naivete. 321 00:14:44,690 --> 00:14:47,570 But it cannot be that the person perceives the commitment device 322 00:14:47,570 --> 00:14:50,120 as not effective, because then the person would just not 323 00:14:50,120 --> 00:14:52,110 demand it in the first place. 324 00:14:52,110 --> 00:14:54,920 Now, that makes now the statement false. 325 00:14:54,920 --> 00:14:57,170 When only the first condition of the three is met, 326 00:14:57,170 --> 00:15:00,050 as in like the person's only present-biased, 327 00:15:00,050 --> 00:15:03,500 then we cannot be sure that there will be positive demand 328 00:15:03,500 --> 00:15:06,740 for commitment. 329 00:15:06,740 --> 00:15:09,860 And so the statement, as had been said, 330 00:15:09,860 --> 00:15:13,040 is false, because we said sort of like always have 331 00:15:13,040 --> 00:15:14,840 positive demand for commitment. 332 00:15:14,840 --> 00:15:16,990 If it were without the always, then 333 00:15:16,990 --> 00:15:19,590 you could also answer uncertain, because then you would say, 334 00:15:19,590 --> 00:15:22,160 well, it depends on their naivete, 335 00:15:22,160 --> 00:15:26,390 or on the sophistication, or it depends on how effective 336 00:15:26,390 --> 00:15:27,830 the commitment device is. 337 00:15:27,830 --> 00:15:30,830 But the way the statement is written, it says like, always. 338 00:15:30,830 --> 00:15:32,360 And so now, you can easily come up 339 00:15:32,360 --> 00:15:36,170 with some counterexamples that show that the statement in fact 340 00:15:36,170 --> 00:15:36,850 is false. 341 00:15:39,790 --> 00:15:41,740 The second topic that we discussed 342 00:15:41,740 --> 00:15:44,780 after time preferences was risk preferences, 343 00:15:44,780 --> 00:15:47,180 in particular expected utility. 344 00:15:47,180 --> 00:15:49,952 So here, you know, you need to have 345 00:15:49,952 --> 00:15:53,520 a clear understanding of what is the expected utility model? 346 00:15:53,520 --> 00:15:54,750 What is risk aversion? 347 00:15:54,750 --> 00:15:56,760 Why are people risk averse? 348 00:15:56,760 --> 00:15:58,980 How is risk aversion specifically modeled 349 00:15:58,980 --> 00:16:01,290 in the expected utility model? 350 00:16:01,290 --> 00:16:04,840 What is the expected monetary value? 351 00:16:04,840 --> 00:16:08,040 And also things like what is concavity? 352 00:16:08,040 --> 00:16:09,720 What does concavity have to do-- what 353 00:16:09,720 --> 00:16:13,170 is the expected utility versus the expected monetary value? 354 00:16:13,170 --> 00:16:17,220 And what does the concavity imply for risk aversion 355 00:16:17,220 --> 00:16:20,510 in the expected utility model? 356 00:16:20,510 --> 00:16:23,030 Then next, how can we measure risk aversion 357 00:16:23,030 --> 00:16:25,110 within the expected utility model? 358 00:16:25,110 --> 00:16:27,110 In particular, we discussed three types 359 00:16:27,110 --> 00:16:29,040 of ways of measuring risk aversion. 360 00:16:29,040 --> 00:16:30,620 We discuss certainty equivalents. 361 00:16:30,620 --> 00:16:32,840 We discussed choices from gambles. 362 00:16:32,840 --> 00:16:35,570 We also discussed insurance choices. 363 00:16:35,570 --> 00:16:38,250 This is the Sydnor paper. 364 00:16:38,250 --> 00:16:41,670 And then, we discussed what is problematic about the estimates 365 00:16:41,670 --> 00:16:44,760 of risk aversion in the expected utility model. 366 00:16:44,760 --> 00:16:46,650 In particular, we discussed evidence 367 00:16:46,650 --> 00:16:50,360 that found that there tends to be substantial small-scale risk 368 00:16:50,360 --> 00:16:53,075 aversion, so when you give people small gambles, 369 00:16:53,075 --> 00:16:54,450 they tend to be quite risk averse 370 00:16:54,450 --> 00:16:58,680 or they tend to have a very quite high gamma. 371 00:16:58,680 --> 00:17:02,130 But we know also from large scale choices that the risk 372 00:17:02,130 --> 00:17:06,569 aversion cannot be actually that high when people make these 373 00:17:06,569 --> 00:17:07,349 choices. 374 00:17:07,349 --> 00:17:09,480 People leave the house every day and engage 375 00:17:09,480 --> 00:17:11,760 in quite a few risks. 376 00:17:11,760 --> 00:17:14,250 In the long run, they hold stocks and so on. 377 00:17:14,250 --> 00:17:17,430 This must mean that their long run 378 00:17:17,430 --> 00:17:20,400 using sort of long-run choices, that implies essentially 379 00:17:20,400 --> 00:17:22,199 a relatively low gamma. 380 00:17:22,199 --> 00:17:25,140 But since the expected utility model only has like one 381 00:17:25,140 --> 00:17:28,950 parameter, it cannot explain both of those features and sort 382 00:17:28,950 --> 00:17:30,870 of ignores this trouble. 383 00:17:30,870 --> 00:17:35,250 So essentially, if you try to match a small-scale risk 384 00:17:35,250 --> 00:17:38,650 aversion, then you need to have a very large gamma, high gamma. 385 00:17:38,650 --> 00:17:40,590 If you try to match large-scale risk aversion, 386 00:17:40,590 --> 00:17:42,030 you need to have a low gamma. 387 00:17:42,030 --> 00:17:45,030 And that sort of brings trouble, because you can't just 388 00:17:45,030 --> 00:17:46,770 explain both of those things. 389 00:17:46,770 --> 00:17:49,440 And sort of Matthew Rabin in Rabin & Thaler, 390 00:17:49,440 --> 00:17:53,940 as well as recitation 4 discuss this conundrum and this issue 391 00:17:53,940 --> 00:17:58,370 and in quite a bit of detail. 392 00:17:58,370 --> 00:18:02,310 Next, we discussed Kahneman and Tversky's 1979 prospect theory. 393 00:18:02,310 --> 00:18:04,260 This is a seminal paper. 394 00:18:04,260 --> 00:18:08,100 And if you'd remember just a few papers from this class, 395 00:18:08,100 --> 00:18:11,780 this is one of the papers that you should really know about. 396 00:18:11,780 --> 00:18:13,450 So what evidence in Kahneman and Tversky 397 00:18:13,450 --> 00:18:15,740 is inconsistent with expected utility? 398 00:18:15,740 --> 00:18:18,050 Well, in particular, sort of several things in there, 399 00:18:18,050 --> 00:18:19,910 but you discuss mostly one feature, which 400 00:18:19,910 --> 00:18:22,130 is risk aversion in the gain domain, 401 00:18:22,130 --> 00:18:27,480 and risk lovingness in the loss domain. 402 00:18:27,480 --> 00:18:29,620 And so, now, what are the most important points 403 00:18:29,620 --> 00:18:31,573 from Kahneman and Tversky's prospect theory-- 404 00:18:31,573 --> 00:18:33,240 this is sort of the proposed alternative 405 00:18:33,240 --> 00:18:34,520 to expected utilities. 406 00:18:34,520 --> 00:18:39,203 This is on slide 3 of 51 of lecture 9, 407 00:18:39,203 --> 00:18:41,370 where there's sort of three futures discussed there. 408 00:18:41,370 --> 00:18:45,000 One is like changes rather than levels are the arguments 409 00:18:45,000 --> 00:18:46,590 of the utility function. 410 00:18:46,590 --> 00:18:48,490 Then there's loss aversion. 411 00:18:48,490 --> 00:18:52,560 So there's a kink of the utility function around the reference 412 00:18:52,560 --> 00:18:53,520 point. 413 00:18:53,520 --> 00:18:57,120 And there's diminishing sensitivity, 414 00:18:57,120 --> 00:19:01,890 meaning the utility function is concave in the gain domain 415 00:19:01,890 --> 00:19:05,568 or [INAUDIBLE] above the reference point and is convex. 416 00:19:05,568 --> 00:19:07,860 And the last domain would be below the reference point. 417 00:19:11,140 --> 00:19:14,830 OK, so now, what does sort of this proposed alternative 418 00:19:14,830 --> 00:19:16,788 utility or value function look like? 419 00:19:16,788 --> 00:19:18,830 How does it incorporate the three features again? 420 00:19:18,830 --> 00:19:22,070 We sort of discussed this and this utility function 421 00:19:22,070 --> 00:19:26,380 in lecture 9 that talks about this in detail. 422 00:19:26,380 --> 00:19:29,273 Or the lecture talks about this utility function in detail. 423 00:19:29,273 --> 00:19:31,690 One key question here, that is, how is the reference point 424 00:19:31,690 --> 00:19:32,260 determined? 425 00:19:32,260 --> 00:19:34,490 What are some candidate reference points? 426 00:19:34,490 --> 00:19:36,430 So one candidate would be the status quo. 427 00:19:36,430 --> 00:19:39,658 Another reference point candidate would be expectation. 428 00:19:39,658 --> 00:19:41,200 But there could be also other things, 429 00:19:41,200 --> 00:19:43,420 such as goal and aspirations. 430 00:19:43,420 --> 00:19:46,120 Recitation 5 discuss this a little bit, 431 00:19:46,120 --> 00:19:49,650 not in too much detail. 432 00:19:49,650 --> 00:19:52,950 Then, we discussed empirical evidence, in particular, 433 00:19:52,950 --> 00:19:55,560 what empirical evidence of loss aversion do we have? 434 00:19:55,560 --> 00:19:57,610 We talked about small-scale gambles. 435 00:19:57,610 --> 00:20:00,390 We talked about the endowment effect, in particular. 436 00:20:00,390 --> 00:20:05,108 And then we discussed some applications in lecture 9. 437 00:20:05,108 --> 00:20:06,400 So what are these applications? 438 00:20:06,400 --> 00:20:09,100 We discuss labor supply, the housing market, stocks, 439 00:20:09,100 --> 00:20:11,010 marathon running, and golf. 440 00:20:11,010 --> 00:20:15,090 So you should be familiar with these empirical applications 441 00:20:15,090 --> 00:20:16,350 from lecture 9. 442 00:20:16,350 --> 00:20:18,030 You should in particular understand 443 00:20:18,030 --> 00:20:21,630 why reference-dependent preferences can explain 444 00:20:21,630 --> 00:20:24,390 some of the empirical evidence that 445 00:20:24,390 --> 00:20:26,680 is showed better than the expected utility models. 446 00:20:26,680 --> 00:20:29,290 You should understand, is some of the evidence 447 00:20:29,290 --> 00:20:31,770 that we see consistent with the expected utility model? 448 00:20:31,770 --> 00:20:33,380 And if not, why not? 449 00:20:33,380 --> 00:20:35,700 And why is the reference-dependent model-- 450 00:20:35,700 --> 00:20:38,700 in particular, which feature of the reference-dependent model 451 00:20:38,700 --> 00:20:39,600 can explain the loss. 452 00:20:39,600 --> 00:20:41,310 What are the features again? 453 00:20:41,310 --> 00:20:43,290 Changes rather than levels-- 454 00:20:43,290 --> 00:20:46,200 some reference point, reference-dependence, the loss 455 00:20:46,200 --> 00:20:49,195 aversion and diminished sensitivity, 456 00:20:49,195 --> 00:20:50,820 so which feature is actually important? 457 00:20:50,820 --> 00:20:54,450 Explaining things, we focused mostly on loss aversion. 458 00:20:54,450 --> 00:20:58,470 What is not relevant is the deal or no deal evidence 459 00:20:58,470 --> 00:21:00,027 and the paper by Pierce et al. 460 00:21:00,027 --> 00:21:02,110 There's a couple of slides at the end of lecture 9 461 00:21:02,110 --> 00:21:04,560 that I didn't really cover or didn't cover at all. 462 00:21:04,560 --> 00:21:08,915 So we're not going to ask any questions about that. 463 00:21:08,915 --> 00:21:10,290 So now, how do you solve problems 464 00:21:10,290 --> 00:21:11,832 with reference-dependent preferences? 465 00:21:11,832 --> 00:21:14,940 You can see problem set 3, question number 1 466 00:21:14,940 --> 00:21:16,590 had some questions about this. 467 00:21:16,590 --> 00:21:19,282 And again, there's additional psets and exam questions 468 00:21:19,282 --> 00:21:20,490 that you could practice with. 469 00:21:20,490 --> 00:21:24,140 So there are quite a few of those kinds of questions. 470 00:21:24,140 --> 00:21:27,320 Now, here is sort of some example of a multiple choice 471 00:21:27,320 --> 00:21:28,880 question, which is-- 472 00:21:28,880 --> 00:21:32,210 so Maddie wrote this question, so Maddie, in fact, appears. 473 00:21:32,210 --> 00:21:34,780 And the question is, Maddie is writing 474 00:21:34,780 --> 00:21:36,890 a problem set for 14.13. 475 00:21:36,890 --> 00:21:40,440 She gets utility u of q from the number of questions she writes. 476 00:21:40,440 --> 00:21:42,080 She has reference-dependent preferences 477 00:21:42,080 --> 00:21:45,953 around the goal of writing 10 questions, with 10 478 00:21:45,953 --> 00:21:47,870 as her reference points. 479 00:21:47,870 --> 00:21:51,500 If you normalize the utility of 10 questions to 0, 480 00:21:51,500 --> 00:21:55,345 which of the following would be consistent with loss aversion? 481 00:21:55,345 --> 00:21:56,720 I'll let you have a look at this. 482 00:21:56,720 --> 00:22:00,770 So there's option A, u of of 8 equals minus 2, u of 12 483 00:22:00,770 --> 00:22:02,660 equals 1. 484 00:22:02,660 --> 00:22:06,560 And number B, option B yields 8 equals minus 2, u of 12 485 00:22:06,560 --> 00:22:07,700 equals 2. 486 00:22:07,700 --> 00:22:13,820 And then C is, option C is u of 8 equals minus 1, and u of 12 487 00:22:13,820 --> 00:22:16,670 equals 2. 488 00:22:16,670 --> 00:22:20,810 So which of those answers is consistent with loss aversion? 489 00:22:20,810 --> 00:22:24,050 The answer is option A. Now, why is that? 490 00:22:24,050 --> 00:22:28,610 Well, loss aversion sort of implies that losses hurt more 491 00:22:28,610 --> 00:22:30,390 than gains help. 492 00:22:30,390 --> 00:22:32,780 So with preferences as in A, Maddie 493 00:22:32,780 --> 00:22:35,810 would have the utility cost of 2 from falling 494 00:22:35,810 --> 00:22:37,970 short of her goal of two questions, 495 00:22:37,970 --> 00:22:42,200 but only a gain of 1 util from exceeding her goal by two 496 00:22:42,200 --> 00:22:43,040 questions, right? 497 00:22:43,040 --> 00:22:49,050 So the goal of, the game of exceeding, 498 00:22:49,050 --> 00:22:54,980 of being by two questions above, so being two questions above 499 00:22:54,980 --> 00:22:57,050 gives her some gain of 1 util. 500 00:22:57,050 --> 00:22:59,240 But the utility costs of falling short with two 501 00:22:59,240 --> 00:23:00,828 is twice as large, which just tends 502 00:23:00,828 --> 00:23:02,495 to be kind of like the evidence that you 503 00:23:02,495 --> 00:23:04,040 would see in loss aversion. 504 00:23:04,040 --> 00:23:07,450 And the other two examples don't have that feature. 505 00:23:07,450 --> 00:23:11,480 So really only answer number A is correct. 506 00:23:15,200 --> 00:23:15,950 Second question. 507 00:23:15,950 --> 00:23:19,580 Maddie is walking home and passes a bakery unexpectedly. 508 00:23:19,580 --> 00:23:21,140 She decides to buy a pastry. 509 00:23:21,140 --> 00:23:23,060 For example, she looks at the pastry, 510 00:23:23,060 --> 00:23:24,650 and it looks really nice. 511 00:23:24,650 --> 00:23:27,740 Prior to purchasing the pastry, her maximum willingness 512 00:23:27,740 --> 00:23:31,010 to pay for the pastry was P0. 513 00:23:31,010 --> 00:23:35,030 Then she runs into Alan-- this is our previous, excellent TA-- 514 00:23:35,030 --> 00:23:37,640 who asks to buy the pastry from her. 515 00:23:37,640 --> 00:23:41,480 She offers him the lowest price she is willing to accept, P1. 516 00:23:41,480 --> 00:23:44,510 Which of the following comparisons between P0 and P1 517 00:23:44,510 --> 00:23:46,910 is consistent with an endowment effect, 518 00:23:46,910 --> 00:23:52,790 P0 larger than P1, P0 equals P1, or P0 smaller than P1? 519 00:23:57,660 --> 00:24:01,320 The answer is answer number C. Why is that? 520 00:24:01,320 --> 00:24:05,550 Well, the endowment effect says that people are-- 521 00:24:05,550 --> 00:24:10,590 when their willingness to pay, that is in this case P0, 522 00:24:10,590 --> 00:24:13,800 is smaller than their willingness to accept, 523 00:24:13,800 --> 00:24:14,860 when they accept. 524 00:24:14,860 --> 00:24:18,450 That is to say, being endowed with an item in this case, 525 00:24:18,450 --> 00:24:22,320 like a pastry, increases one's willingness to pay. 526 00:24:22,320 --> 00:24:25,380 That is to say, if somebody asks you to sell something 527 00:24:25,380 --> 00:24:27,960 that you own, you ask for more money 528 00:24:27,960 --> 00:24:30,690 than you're willing to pay in the first place 529 00:24:30,690 --> 00:24:32,830 when you don't own the item. 530 00:24:32,830 --> 00:24:36,810 And so the endowment effect will then predict that, 531 00:24:36,810 --> 00:24:42,150 or the endowment effect entails that now P1 is larger 532 00:24:42,150 --> 00:24:46,950 than P0, which is answer number C. So Maddie values 533 00:24:46,950 --> 00:24:50,010 the pastry more after she has bought it, 534 00:24:50,010 --> 00:24:51,520 compared to like prior to buying it. 535 00:24:54,610 --> 00:24:57,430 The third broad set of preferences to be discussed 536 00:24:57,430 --> 00:24:58,660 were social preferences. 537 00:24:58,660 --> 00:25:00,050 We didn't quite finish with this, 538 00:25:00,050 --> 00:25:02,470 so we're not going to cover everything, in particular, 539 00:25:02,470 --> 00:25:05,370 not lecture 13. 540 00:25:05,370 --> 00:25:08,340 And the estimation part of social preferences, 541 00:25:08,340 --> 00:25:13,140 which will apart from choices, which will be at pset 4. 542 00:25:13,140 --> 00:25:15,817 You should understand what social preferences are. 543 00:25:15,817 --> 00:25:17,400 You should also understand how can you 544 00:25:17,400 --> 00:25:19,920 measure social preferences? 545 00:25:19,920 --> 00:25:21,940 In lab games, we discussed at length 546 00:25:21,940 --> 00:25:25,290 the dictator game, the ultimatum game, and the trust game. 547 00:25:25,290 --> 00:25:27,910 You should also be broadly familiar, not in detail, 548 00:25:27,910 --> 00:25:29,580 but broadly familiar of what evidence 549 00:25:29,580 --> 00:25:33,740 do we typically find in dictator and ultimatum games? 550 00:25:33,740 --> 00:25:35,460 For instance, in dictator games, people 551 00:25:35,460 --> 00:25:40,020 tend to give something like 20% to 30% of their share. 552 00:25:40,020 --> 00:25:42,840 People tend to be quite nice in those games. 553 00:25:42,840 --> 00:25:46,305 Now, then we discussed then like given that evidence, 554 00:25:46,305 --> 00:25:50,720 so given that people look quite nice in these types of games, 555 00:25:50,720 --> 00:25:53,010 there's also some other evidence people give 556 00:25:53,010 --> 00:25:55,900 to charity or things like that. 557 00:25:55,900 --> 00:25:57,720 Do we think that is evidence of people 558 00:25:57,720 --> 00:26:02,410 being generally nice to others because of poor altruism? 559 00:26:02,410 --> 00:26:04,960 Or if not, why not? 560 00:26:04,960 --> 00:26:08,520 We discussed sort of three sets of evidence in particular. 561 00:26:08,520 --> 00:26:13,343 We discussed the costly exit or exit options in dictator games. 562 00:26:13,343 --> 00:26:15,760 So people essentially won't be able to leave the dictator. 563 00:26:15,760 --> 00:26:17,830 They rather sort of leave and give, 564 00:26:17,830 --> 00:26:19,300 and sort of keep the money. 565 00:26:19,300 --> 00:26:22,690 Or they're willing to pay some small amount of money 566 00:26:22,690 --> 00:26:24,160 to leave the dictator game and not 567 00:26:24,160 --> 00:26:26,500 having to face some other person, 568 00:26:26,500 --> 00:26:29,320 but it then feel compelled to give to others. 569 00:26:29,320 --> 00:26:33,880 There was the option of hiding behind the computer. 570 00:26:33,880 --> 00:26:35,620 If a computer gives you the option 571 00:26:35,620 --> 00:26:38,590 to hide behind the computer to be mean to others, 572 00:26:38,590 --> 00:26:41,560 you might take advantage of that and actually be meaner 573 00:26:41,560 --> 00:26:42,780 than you would be otherwise. 574 00:26:42,780 --> 00:26:46,240 So altruism, or people's giving tends to go down quite a bit 575 00:26:46,240 --> 00:26:49,960 when they're able to hide behind the computer. 576 00:26:49,960 --> 00:26:53,040 So both of those types of pieces of evidence 577 00:26:53,040 --> 00:26:56,670 are evidence of social image being important in giving. 578 00:26:56,670 --> 00:26:59,580 People care a lot about what others think, 579 00:26:59,580 --> 00:27:01,710 in particular what others think about them, 580 00:27:01,710 --> 00:27:03,850 and they don't want to upset others. 581 00:27:03,850 --> 00:27:09,420 And so if they're able to avoid those kind of situations, 582 00:27:09,420 --> 00:27:10,590 they might be able to-- 583 00:27:10,590 --> 00:27:13,230 they might want to do that, which would suggest that it's 584 00:27:13,230 --> 00:27:16,650 not really that they want others to do well in the sense 585 00:27:16,650 --> 00:27:19,830 that they really want others to have money or more money 586 00:27:19,830 --> 00:27:23,500 than before, but rather it's because of social pressure 587 00:27:23,500 --> 00:27:26,610 or social image concerns, people might give 588 00:27:26,610 --> 00:27:29,580 in dictator or ultimatum games. 589 00:27:29,580 --> 00:27:32,460 Moreover, there's some evidence of self-image 590 00:27:32,460 --> 00:27:37,000 about people caring about what they think about themselves. 591 00:27:37,000 --> 00:27:40,410 They want to think of themselves as being a good person. 592 00:27:40,410 --> 00:27:43,260 And so the evidence of the moral wiggle room 593 00:27:43,260 --> 00:27:48,270 seems to suggest that these concerns are quite important. 594 00:27:48,270 --> 00:27:52,890 In particular, people are engaging in some behavior 595 00:27:52,890 --> 00:27:57,180 where they delude themselves that they are in fact nice, 596 00:27:57,180 --> 00:27:59,610 when in reality, they're not. 597 00:27:59,610 --> 00:28:03,810 So that's discussed in detail in the lecture. 598 00:28:03,810 --> 00:28:05,435 Again, we will not ask, and so you 599 00:28:05,435 --> 00:28:07,980 should be familiar with this type of evidence. 600 00:28:07,980 --> 00:28:10,380 And you should be familiar why that sort of tells us 601 00:28:10,380 --> 00:28:14,100 that people are perhaps not as nice as you might have thought. 602 00:28:14,100 --> 00:28:17,010 They are just coming from like dictator or ultimatum games, 603 00:28:17,010 --> 00:28:19,440 or from like donations. 604 00:28:19,440 --> 00:28:21,540 We will not ask you about models that 605 00:28:21,540 --> 00:28:23,010 estimate social preferences. 606 00:28:23,010 --> 00:28:26,330 That will be in problem set 4, so don't worry about that. 607 00:28:26,330 --> 00:28:29,530 Sort of the stuff on [INAUDIBLE] and Rabin that's there, 608 00:28:29,530 --> 00:28:32,700 we will not ask you about that. 609 00:28:32,700 --> 00:28:35,910 OK, so then here's another example 610 00:28:35,910 --> 00:28:37,800 of a true/false/uncertain question. 611 00:28:37,800 --> 00:28:40,570 Statement, if a person gives 0 in a dictator game, 612 00:28:40,570 --> 00:28:44,430 this is evidence that this person is selfish. 613 00:28:44,430 --> 00:28:46,860 Now, the answer is false. 614 00:28:46,860 --> 00:28:48,490 I think if you answer it uncertain, 615 00:28:48,490 --> 00:28:50,310 that would also be fine. 616 00:28:50,310 --> 00:28:51,060 Why is that? 617 00:28:51,060 --> 00:28:55,230 Well, the person might give 0 to the other person 618 00:28:55,230 --> 00:28:58,745 in the dictator game and then donate the money to someone 619 00:28:58,745 --> 00:28:59,807 in greater need, right? 620 00:28:59,807 --> 00:29:01,890 And then that person is in fact really quite nice. 621 00:29:04,590 --> 00:29:06,242 Second, the person might be very poor 622 00:29:06,242 --> 00:29:07,950 relative to the other person in the game. 623 00:29:07,950 --> 00:29:10,650 So her marginal utility is just very high. 624 00:29:10,650 --> 00:29:13,200 So now, even if you have like equal weights 625 00:29:13,200 --> 00:29:17,460 to your own utility and the other person's utility, 626 00:29:17,460 --> 00:29:21,090 since the marginal utility of giving $10 to yourself compared 627 00:29:21,090 --> 00:29:23,820 to the other person is way higher than the other person's 628 00:29:23,820 --> 00:29:28,687 marginal utility, you would just give everything to yourself. 629 00:29:28,687 --> 00:29:30,770 And that doesn't mean you don't care about others. 630 00:29:30,770 --> 00:29:33,380 It just means like, in fact, even 631 00:29:33,380 --> 00:29:36,655 if you were sort of like a social climber, 632 00:29:36,655 --> 00:29:38,030 you would give it to that person, 633 00:29:38,030 --> 00:29:39,780 because that person was like a huge meany, 634 00:29:39,780 --> 00:29:42,682 or the rich person just might not really need that money. 635 00:29:42,682 --> 00:29:44,390 Looking at this more closely, actually, I 636 00:29:44,390 --> 00:29:47,310 think uncertain would be a better answer here, and not 637 00:29:47,310 --> 00:29:47,810 false. 638 00:29:47,810 --> 00:29:50,790 I just wrote this question fairly quickly. 639 00:29:50,790 --> 00:29:53,430 So if the question says-- 640 00:29:53,430 --> 00:29:55,020 if the question instead were to say, 641 00:29:55,020 --> 00:29:59,030 this is conclusive evidence that this person is selfish, 642 00:29:59,030 --> 00:30:01,340 then it would be false. 643 00:30:01,340 --> 00:30:03,200 Or this would be, this is clearly 644 00:30:03,200 --> 00:30:04,760 evidence where this person must be 645 00:30:04,760 --> 00:30:07,880 selfish, that would be false. 646 00:30:07,880 --> 00:30:10,670 The way it's written right now, it's rather sort of uncertain, 647 00:30:10,670 --> 00:30:12,180 because we just don't quite know. 648 00:30:12,180 --> 00:30:16,520 It could be that the person is really selfish. 649 00:30:16,520 --> 00:30:19,320 Or it could be that the person-- 650 00:30:19,320 --> 00:30:21,770 it could be this evidence that the person is selfish. 651 00:30:21,770 --> 00:30:26,680 But it could be the other two reasons that I just mentioned. 652 00:30:26,680 --> 00:30:30,640 OK, so now, finally, I'm going to give you a long question, 653 00:30:30,640 --> 00:30:34,120 or an example of a long question. 654 00:30:34,120 --> 00:30:36,890 And this is the question of laptop policies. 655 00:30:36,890 --> 00:30:40,530 We talked about a little bit earlier in class, 656 00:30:40,530 --> 00:30:42,200 in fact in the first lecture. 657 00:30:42,200 --> 00:30:44,560 And we had the first problem set about this. 658 00:30:44,560 --> 00:30:46,480 And this is sort of like an algebraic version 659 00:30:46,480 --> 00:30:49,020 of that kind of question. 660 00:30:49,020 --> 00:30:52,310 So assume the 14.13 students are present 661 00:30:52,310 --> 00:30:57,020 biased with beta small than 1 and delta equals 0. 662 00:30:57,020 --> 00:30:59,450 All students have the same beta smaller 663 00:30:59,450 --> 00:31:01,160 than 1 and delta equals 0. 664 00:31:01,160 --> 00:31:04,580 But they differ in the value they derive 665 00:31:04,580 --> 00:31:07,680 from using laptops in class. 666 00:31:07,680 --> 00:31:11,050 L is constant for each student from class to class, 667 00:31:11,050 --> 00:31:14,790 but uniformly distributed across students on the interval 668 00:31:14,790 --> 00:31:17,260 between 0 and 1. 669 00:31:17,260 --> 00:31:21,460 So some people have like a huge value 670 00:31:21,460 --> 00:31:25,255 of using the laptop in class. 671 00:31:25,255 --> 00:31:29,400 And others do not. 672 00:31:29,400 --> 00:31:32,100 Each lecture generates no immediate utility. 673 00:31:32,100 --> 00:31:35,160 So it's neither fun nor like annoying or the like. 674 00:31:35,160 --> 00:31:40,730 But it does give a benefit, a future benefit of V. 675 00:31:40,730 --> 00:31:42,260 So like you might learn something, 676 00:31:42,260 --> 00:31:45,110 or there's some things that might be valuable to you 677 00:31:45,110 --> 00:31:46,280 in the future. 678 00:31:46,280 --> 00:31:50,810 Using a laptop reduces the long-run benefit by D. 679 00:31:50,810 --> 00:31:53,730 This might be like distractions, in particular. 680 00:31:53,730 --> 00:31:56,570 So you might be distracted during class 681 00:31:56,570 --> 00:31:57,860 when you use a laptop. 682 00:31:57,860 --> 00:32:01,320 So the future benefits are diminished by that. 683 00:32:01,320 --> 00:32:03,410 Maybe he had said something really insightful, 684 00:32:03,410 --> 00:32:07,400 you didn't pay attention, and so now, the benefits 685 00:32:07,400 --> 00:32:11,540 is B minus D if you use a laptop. 686 00:32:11,540 --> 00:32:14,340 Both B and D are the same for all students. 687 00:32:14,340 --> 00:32:18,170 So there's no variation across students there. 688 00:32:18,170 --> 00:32:21,120 The only variation's coming from some students who really, 689 00:32:21,120 --> 00:32:24,080 really like using laptops for various reasons, perhaps 690 00:32:24,080 --> 00:32:27,560 because they like to surf the internet. 691 00:32:27,560 --> 00:32:31,940 Perhaps they're really in need of using laptops 692 00:32:31,940 --> 00:32:35,740 because that allows them to take proper notes. 693 00:32:35,740 --> 00:32:40,470 So in summary, a student uses a laptop in class-- 694 00:32:40,470 --> 00:32:42,810 sorry, a student that uses the laptop in class 695 00:32:42,810 --> 00:32:46,590 gets immediate utility L and future, undiscounted utility 696 00:32:46,590 --> 00:32:50,340 of B minus D. And a student who does not use a laptop 697 00:32:50,340 --> 00:32:55,390 gets immediate utility of 0 and future discounted utility of V. 698 00:32:55,390 --> 00:32:59,320 So and now the social planner is not present biased 699 00:32:59,320 --> 00:33:03,100 and seeks to maximize the utility of 14.13 students 700 00:33:03,100 --> 00:33:05,298 from his perspective, OK? 701 00:33:08,620 --> 00:33:10,680 First question. 702 00:33:10,680 --> 00:33:14,340 Show that students are just indifferent between using 703 00:33:14,340 --> 00:33:18,300 and not using a laptop in the current class 704 00:33:18,300 --> 00:33:21,960 if L equals beat times D. Explain 705 00:33:21,960 --> 00:33:24,270 why students with lower values of L-- 706 00:33:24,270 --> 00:33:27,150 so that's like L being lower than beta D-- 707 00:33:27,150 --> 00:33:29,070 don't use laptops in class, but students 708 00:33:29,070 --> 00:33:30,570 with higher values of L-- 709 00:33:30,570 --> 00:33:32,290 so L exceeds beta D-- 710 00:33:32,290 --> 00:33:35,450 do use laptops in class. 711 00:33:35,450 --> 00:33:37,310 OK, so now what we're going to do 712 00:33:37,310 --> 00:33:40,280 is we're going to write down the utilities from the two choices. 713 00:33:40,280 --> 00:33:44,660 That's essentially already given in the explanation, 714 00:33:44,660 --> 00:33:48,170 except for that we need to be careful of where the beta comes 715 00:33:48,170 --> 00:33:48,750 in, right? 716 00:33:48,750 --> 00:33:52,130 So if students use a laptop in class, 717 00:33:52,130 --> 00:33:55,220 they get the immediate benefit of L. And then in the future, 718 00:33:55,220 --> 00:33:58,360 they get what's discounted by beta, which is beta times B 719 00:33:58,360 --> 00:34:01,867 minus D, right, because both B and D are very far 720 00:34:01,867 --> 00:34:02,450 in the future. 721 00:34:02,450 --> 00:34:06,620 These are the benefits, or the diminished benefits, B minus D, 722 00:34:06,620 --> 00:34:08,210 that they get in the future. 723 00:34:08,210 --> 00:34:11,670 L is in the present, so there's no beta here. 724 00:34:11,670 --> 00:34:14,820 Instead, if someone uses no laptop, the students get 0. 725 00:34:14,820 --> 00:34:17,760 So there's no laptop benefits in the present. 726 00:34:17,760 --> 00:34:21,690 And the benefits in the future, the value 727 00:34:21,690 --> 00:34:24,360 of the lecture the future is undiminished, 728 00:34:24,360 --> 00:34:26,429 which means essentially, it's just V. 729 00:34:26,429 --> 00:34:32,570 So the person just gets 0 plus beta times V. 730 00:34:32,570 --> 00:34:37,219 Now, students who are indifferent, by definition, 731 00:34:37,219 --> 00:34:40,670 have the same utility of using laptops and using no laptop. 732 00:34:40,670 --> 00:34:44,929 So we can essentially just equate those two things. 733 00:34:44,929 --> 00:34:48,020 Notice that like the beta times B is always there. 734 00:34:48,020 --> 00:34:50,239 So it sort of essentially just cancels. 735 00:34:50,239 --> 00:34:53,389 And then what we get is the person that's indifferent, 736 00:34:53,389 --> 00:34:57,990 for that person, L equals beta times D. 737 00:34:57,990 --> 00:35:02,010 Now, students that choose not to use laptops 738 00:35:02,010 --> 00:35:05,490 will have low valuations L of using laptops, 739 00:35:05,490 --> 00:35:07,980 while students that choose to use the laptops 740 00:35:07,980 --> 00:35:10,215 have high L, right? 741 00:35:10,215 --> 00:35:13,110 If L is very large, then it exceeds beta times D. 742 00:35:13,110 --> 00:35:17,010 If L is very small, then L does not exceed, 743 00:35:17,010 --> 00:35:18,660 is smaller than beta times D. 744 00:35:18,660 --> 00:35:21,090 So given the indifference condition, 745 00:35:21,090 --> 00:35:24,780 we have essentially, as I just said, students that do not 746 00:35:24,780 --> 00:35:28,720 use the laptop, L is smaller than beta times D. And students 747 00:35:28,720 --> 00:35:35,570 that use the laptop, for them, L is larger than beta times D. 748 00:35:35,570 --> 00:35:37,000 OK. 749 00:35:37,000 --> 00:35:39,700 OK, now question number two. 750 00:35:39,700 --> 00:35:42,640 Now consider the policy that allows students to use laptops 751 00:35:42,640 --> 00:35:46,623 only if they sign up in advance to sit in a laptop section. 752 00:35:46,623 --> 00:35:48,790 This is a little bit different than we had in class, 753 00:35:48,790 --> 00:35:50,623 but it's a version, or something that I also 754 00:35:50,623 --> 00:35:52,780 considered, in fact, in previous years that that 755 00:35:52,780 --> 00:35:54,320 was what was used. 756 00:35:54,320 --> 00:35:58,480 Now, the question here is, why is L larger or equal than D, 757 00:35:58,480 --> 00:36:01,500 not L is larger equal than beta D, 758 00:36:01,500 --> 00:36:05,880 the threshold for opting in to the laptop section? 759 00:36:05,880 --> 00:36:06,420 OK? 760 00:36:06,420 --> 00:36:07,980 So now this is a choice for people 761 00:36:07,980 --> 00:36:10,590 to choose not for the present, when 762 00:36:10,590 --> 00:36:14,130 they come to class, whether they want to use a laptop right now. 763 00:36:14,130 --> 00:36:16,230 But instead, they sign up in advance 764 00:36:16,230 --> 00:36:18,740 to sit in a laptop section for the rest of the semester. 765 00:36:18,740 --> 00:36:20,490 So essentially, people are in the present, 766 00:36:20,490 --> 00:36:24,840 and they choose for the future whether they can use laptops 767 00:36:24,840 --> 00:36:26,880 in class. 768 00:36:26,880 --> 00:36:29,520 Now, the utility is now different. 769 00:36:29,520 --> 00:36:33,030 And the key difference is that the laptop benefits are now 770 00:36:33,030 --> 00:36:34,540 in the future, right? 771 00:36:34,540 --> 00:36:38,760 So the utility of using a laptop now again is 0 in the present. 772 00:36:38,760 --> 00:36:40,410 And this is-- again, we make choices 773 00:36:40,410 --> 00:36:41,733 in the present for the future. 774 00:36:41,733 --> 00:36:43,650 There's going to be no utility in the present. 775 00:36:43,650 --> 00:36:47,850 There's no lecture right now, no laptop benefits and the like. 776 00:36:47,850 --> 00:36:52,230 And in the future, everything is discounted by beta now, 777 00:36:52,230 --> 00:36:55,530 because the person is present biased potentially. 778 00:36:55,530 --> 00:36:59,420 So it's beta times L plus B minus D. 779 00:36:59,420 --> 00:37:02,780 The utility of not using a laptop is just 0 again. 780 00:37:02,780 --> 00:37:04,520 Like in the present, nothing happens. 781 00:37:04,520 --> 00:37:07,190 And then in the future, the benefits, as before, 782 00:37:07,190 --> 00:37:10,970 are beta times D. So the only thing that you prepare 783 00:37:10,970 --> 00:37:14,840 is choice, or these two options, the previous version, 784 00:37:14,840 --> 00:37:17,420 is the L now is discounted by beta. 785 00:37:17,420 --> 00:37:18,950 But before, it was not. 786 00:37:18,950 --> 00:37:20,840 Let me just go back here. 787 00:37:20,840 --> 00:37:23,840 So here, you see the L was like sort of-- 788 00:37:23,840 --> 00:37:27,980 here, you see the L not discounted by beta. 789 00:37:27,980 --> 00:37:30,490 This is when people make choices for the present. 790 00:37:30,490 --> 00:37:33,080 Instead, here now, the L is discounted by beta, 791 00:37:33,080 --> 00:37:34,750 because that choice is-- 792 00:37:34,750 --> 00:37:38,662 the laptop benefits are in the future. 793 00:37:38,662 --> 00:37:41,000 Now, we can do the same thing as before. 794 00:37:41,000 --> 00:37:43,087 Threshold for opting in is defined as like-- 795 00:37:43,087 --> 00:37:45,170 or you can say that's essentially the indifference 796 00:37:45,170 --> 00:37:45,690 conditions. 797 00:37:45,690 --> 00:37:50,270 You can just equalize the two, the utility of using the laptop 798 00:37:50,270 --> 00:37:51,830 or not using the laptop. 799 00:37:51,830 --> 00:37:54,770 And what you now get is that if L is larger than D, or larger 800 00:37:54,770 --> 00:37:57,870 equals than D, then the person uses the laptop. 801 00:37:57,870 --> 00:38:03,690 And otherwise, they do not. 802 00:38:03,690 --> 00:38:06,930 Now, why does the threshold now change from beta D to-- 803 00:38:06,930 --> 00:38:08,880 sorry, from beta times D to D? 804 00:38:08,880 --> 00:38:12,960 Well, because when laptop use can only happen in the future, 805 00:38:12,960 --> 00:38:16,500 all benefits and costs are discounted at the same rate, 806 00:38:16,500 --> 00:38:17,670 and that that rate is beta. 807 00:38:22,680 --> 00:38:24,990 OK, question number three. 808 00:38:24,990 --> 00:38:28,910 Assume there's no laptop policy at all. 809 00:38:28,910 --> 00:38:32,630 Show that if beta times D is smaller than L 810 00:38:32,630 --> 00:38:35,180 and smaller than D, the student engages 811 00:38:35,180 --> 00:38:37,130 in preference reversals. 812 00:38:37,130 --> 00:38:40,890 She prefers not to use the laptop in future classes 813 00:38:40,890 --> 00:38:43,010 but changes her mind while she's actually sitting 814 00:38:43,010 --> 00:38:44,370 in those future classes. 815 00:38:47,320 --> 00:38:49,785 That's sort of essentially the typical behavior 816 00:38:49,785 --> 00:38:50,590 of present bias. 817 00:38:50,590 --> 00:38:53,000 People that, when they're present biased, 818 00:38:53,000 --> 00:38:56,470 for at least some parameter of constellations, 819 00:38:56,470 --> 00:38:57,960 people engage in present bias. 820 00:38:57,960 --> 00:39:00,980 So let's do the math and see what comes out here. 821 00:39:00,980 --> 00:39:04,060 So when thinking about the future laptop use, 822 00:39:04,060 --> 00:39:06,430 this student's problem is identical to the problem 823 00:39:06,430 --> 00:39:09,340 in part B, sorry, in part 2. 824 00:39:09,340 --> 00:39:10,520 Why is that? 825 00:39:10,520 --> 00:39:15,170 Well, because she discounts time, 826 00:39:15,170 --> 00:39:17,547 both one and two periods in advance, by beta. 827 00:39:17,547 --> 00:39:19,255 Essentially, everything is in the future, 828 00:39:19,255 --> 00:39:21,970 so you just discount everything by beta. 829 00:39:21,970 --> 00:39:25,330 That's exactly kind of the choice that we just had. 830 00:39:25,330 --> 00:39:26,990 So when thinking about future class, 831 00:39:26,990 --> 00:39:30,070 where thinking about opting in into a laptop section, 832 00:39:30,070 --> 00:39:33,360 the person makes a choice for the future. 833 00:39:33,360 --> 00:39:38,660 So here, when she thinks about any future choices, what really 834 00:39:38,660 --> 00:39:43,310 matters is essentially the value of the laptop is in the future, 835 00:39:43,310 --> 00:39:47,660 I mean back in like part 2 that we just solved. 836 00:39:47,660 --> 00:39:51,950 Now, we know from part 2 that if L equals smaller than D, 837 00:39:51,950 --> 00:39:54,100 then she would like to not use the laptop, right? 838 00:39:54,100 --> 00:39:59,240 So we just solve for that in part number 2. 839 00:39:59,240 --> 00:40:04,010 But from part number 1, we know that if beta D is smaller 840 00:40:04,010 --> 00:40:07,280 than L, she will end up using the laptop when she's actually 841 00:40:07,280 --> 00:40:08,600 sitting in the future class. 842 00:40:08,600 --> 00:40:11,720 That is to say, if she has a choice in any given class 843 00:40:11,720 --> 00:40:14,150 and shows up in class, she will say, oh, 844 00:40:14,150 --> 00:40:15,770 like using the laptop would be great. 845 00:40:15,770 --> 00:40:19,040 The same would be true of phones, by the way. 846 00:40:19,040 --> 00:40:21,860 And when she has a choice in any given class that 847 00:40:21,860 --> 00:40:26,270 happens right now, if beta D equals smaller than L, 848 00:40:26,270 --> 00:40:29,060 she will end up using the laptop in class. 849 00:40:29,060 --> 00:40:33,560 So that implies essentially like a preference reversal using 850 00:40:33,560 --> 00:40:36,510 these parameter assumptions. 851 00:40:36,510 --> 00:40:39,800 She prefers not to use the laptop in future classes 852 00:40:39,800 --> 00:40:42,530 but switches her mind or changes her mind when she's actually 853 00:40:42,530 --> 00:40:44,565 sitting in those future classes. 854 00:40:47,840 --> 00:40:50,870 OK, question number 4. 855 00:40:50,870 --> 00:40:54,830 Explain why the fraction 1 minus beta D of the class 856 00:40:54,830 --> 00:40:58,820 uses a laptop in part 1, but fraction 1 minus D of the class 857 00:40:58,820 --> 00:41:02,150 uses the laptop in part 2. 858 00:41:02,150 --> 00:41:04,070 Why does a smaller share of the class 859 00:41:04,070 --> 00:41:05,750 use their laptop in part 2? 860 00:41:09,150 --> 00:41:12,480 All right, so now we're just essentially comparing 861 00:41:12,480 --> 00:41:13,740 part 1 and part 2. 862 00:41:13,740 --> 00:41:16,200 And I'm going to look at what fraction of people 863 00:41:16,200 --> 00:41:19,758 are actually using the laptop. 864 00:41:19,758 --> 00:41:21,300 So we can sort of do the math version 865 00:41:21,300 --> 00:41:23,920 and think about why that answer makes sense. 866 00:41:23,920 --> 00:41:25,950 So in part 1, a student uses the laptop 867 00:41:25,950 --> 00:41:36,570 if L is larger than beta times D. If F is CDF of L-- 868 00:41:36,570 --> 00:41:38,730 and then started as a puritan thing, 869 00:41:38,730 --> 00:41:42,920 but to define F as the CDF of L. And then given 870 00:41:42,920 --> 00:41:47,060 the uniform distribution, the probability 871 00:41:47,060 --> 00:41:54,850 of L being larger than beta D is 1 minus the CDF 1 minus 872 00:41:54,850 --> 00:41:57,190 F of beta delta, which is in this uniform. 873 00:41:57,190 --> 00:42:00,760 It's just 1 minus beta D. 874 00:42:00,760 --> 00:42:03,820 Now, likewise, in part 2, a student uses a laptop 875 00:42:03,820 --> 00:42:07,810 if L is larger than the D. So we have the probability of L 876 00:42:07,810 --> 00:42:10,800 being larger than D is 1 minus the CDF 877 00:42:10,800 --> 00:42:14,380 or the F of D, which is 1 minus D. 878 00:42:14,380 --> 00:42:17,350 So a smaller share uses the laptop in part 2, 879 00:42:17,350 --> 00:42:19,060 because the benefit of using a laptop 880 00:42:19,060 --> 00:42:22,910 is delayed and hence discounted by beta. 881 00:42:22,910 --> 00:42:24,500 So why is that? 882 00:42:24,500 --> 00:42:27,040 Well, essentially, think about it like this. 883 00:42:27,040 --> 00:42:30,550 If somebody has beta equals 1, which 884 00:42:30,550 --> 00:42:32,620 is kind of equivalent to like part 2, where 885 00:42:32,620 --> 00:42:36,280 people made choices for the future, if you use a laptop-- 886 00:42:36,280 --> 00:42:39,190 so a share of people will like the laptop, because-- 887 00:42:39,190 --> 00:42:42,460 or like to use the laptop when making choices for the future, 888 00:42:42,460 --> 00:42:44,680 not because of self-control problems or the like, 889 00:42:44,680 --> 00:42:47,200 but just because they find lots of really helpful in taking 890 00:42:47,200 --> 00:42:49,050 notes. 891 00:42:49,050 --> 00:42:53,100 Now, if then a person, in addition, is present biased 892 00:42:53,100 --> 00:42:55,020 and makes a choice for the present, 893 00:42:55,020 --> 00:42:58,950 that enhances the short-run benefits, 894 00:42:58,950 --> 00:43:01,830 because now the L is not discounted by beta anymore. 895 00:43:01,830 --> 00:43:07,710 And now, it's essentially, it's given sort of the benefit. 896 00:43:07,710 --> 00:43:10,260 It's in the present, and everything else beta 897 00:43:10,260 --> 00:43:11,560 is in the future. 898 00:43:11,560 --> 00:43:14,370 And so that means essentially that when making choices 899 00:43:14,370 --> 00:43:17,985 for the present, the present benefits, the L, 900 00:43:17,985 --> 00:43:21,030 gets more weight relative to everything else, which 901 00:43:21,030 --> 00:43:24,200 is against the weight of beta less than 1. 902 00:43:24,200 --> 00:43:29,010 So now, if you are already, when making choices for the future, 903 00:43:29,010 --> 00:43:31,440 chose the laptop anyway, that implies 904 00:43:31,440 --> 00:43:35,100 that you also choose the laptop for the present. 905 00:43:35,100 --> 00:43:38,010 Essentially, anybody who chooses the laptop for the future 906 00:43:38,010 --> 00:43:40,440 would also choose the laptop for the present. 907 00:43:40,440 --> 00:43:42,570 And now, there are some people essentially don't 908 00:43:42,570 --> 00:43:43,740 have like huge variations. 909 00:43:43,740 --> 00:43:45,765 They might not choose the laptop for the future, 910 00:43:45,765 --> 00:43:47,640 but they might choose it, they will choose it 911 00:43:47,640 --> 00:43:50,350 for the present because of their present bias. 912 00:43:50,350 --> 00:43:52,223 And therefore, then the fraction of people 913 00:43:52,223 --> 00:43:53,640 who choose for the present will be 914 00:43:53,640 --> 00:43:56,143 larger than the fraction of people 915 00:43:56,143 --> 00:43:57,810 who choose for the future, which we just 916 00:43:57,810 --> 00:44:02,010 showed using some algebra. 917 00:44:02,010 --> 00:44:04,620 OK, then finally, why would the social planner 918 00:44:04,620 --> 00:44:07,740 prefer the opt-in policy to both the the policy of allowing 919 00:44:07,740 --> 00:44:10,620 students to choose whether to use their laptops 920 00:44:10,620 --> 00:44:14,530 and to banning laptops altogether? 921 00:44:14,530 --> 00:44:18,980 So let's think about through this, the opt-in policy, 922 00:44:18,980 --> 00:44:20,975 what does that really entail? 923 00:44:20,975 --> 00:44:23,160 Well, the opt-in policy, as we said, 924 00:44:23,160 --> 00:44:27,110 is the planner is not present biased. 925 00:44:27,110 --> 00:44:30,230 So the planner would only want students 926 00:44:30,230 --> 00:44:33,230 with L being larger than D to use laptops. 927 00:44:33,230 --> 00:44:35,870 And so the opt-in policy, as we just showed above, 928 00:44:35,870 --> 00:44:36,650 achieves this. 929 00:44:36,650 --> 00:44:38,090 So that's great. 930 00:44:38,090 --> 00:44:40,370 I know like a free choice policy instead, 931 00:44:40,370 --> 00:44:44,240 students with beta times D is smaller than L, smaller than D, 932 00:44:44,240 --> 00:44:46,220 will suboptimally use their laptops, 933 00:44:46,220 --> 00:44:49,170 and the social planner does not like this. 934 00:44:49,170 --> 00:44:51,935 On the other hand, banning laptops altogether 935 00:44:51,935 --> 00:44:54,290 is suboptimal because welfare is gained 936 00:44:54,290 --> 00:44:57,500 by allowing the students with the highest valuations, with L 937 00:44:57,500 --> 00:44:58,670 equals-- 938 00:44:58,670 --> 00:45:01,670 sorry, L larger than D to use laptops, right? 939 00:45:01,670 --> 00:45:04,170 So banning laptops is not great because essentially, there's 940 00:45:04,170 --> 00:45:05,545 some people who really would love 941 00:45:05,545 --> 00:45:07,910 to use their laptops regardless of present bias. 942 00:45:07,910 --> 00:45:11,660 And not allowing that is not great. 943 00:45:11,660 --> 00:45:13,910 Free choice is not great, because essentially, 944 00:45:13,910 --> 00:45:16,250 once you let people choose any given day, 945 00:45:16,250 --> 00:45:18,440 temptation will sort of kick in, and some people 946 00:45:18,440 --> 00:45:20,210 will suboptimally use their laptops 947 00:45:20,210 --> 00:45:23,150 and just surf the internet all day, or all class, 948 00:45:23,150 --> 00:45:24,560 and not learn very much. 949 00:45:24,560 --> 00:45:29,330 And instead, the policy where people opt in for the future 950 00:45:29,330 --> 00:45:32,330 essentially achieves the objective 951 00:45:32,330 --> 00:45:36,740 of the social planner, who wants only students with L 952 00:45:36,740 --> 00:45:39,320 larger than D to use their laptops. 953 00:45:39,320 --> 00:45:42,530 And the social planner will be happy, and therefore 954 00:45:42,530 --> 00:45:45,860 prefer that policy over both free choice 955 00:45:45,860 --> 00:45:48,170 and over banning laptops. 956 00:45:52,650 --> 00:45:53,400 So that's the end. 957 00:45:53,400 --> 00:45:56,970 That's all I have to say about getting ready for the exam. 958 00:45:56,970 --> 00:45:59,310 And I think you should prepare well and try 959 00:45:59,310 --> 00:46:01,260 to look at the materials. 960 00:46:01,260 --> 00:46:04,530 Please ask any questions in case things aren't clear. 961 00:46:04,530 --> 00:46:15,200 Again, I have office hours on Friday, April 3 962 00:46:15,200 --> 00:46:16,940 at 4:30 to 6:00. 963 00:46:16,940 --> 00:46:18,530 I emailed you about that. 964 00:46:18,530 --> 00:46:22,280 [INAUDIBLE] also has office hours from 1:30 to 3:30. 965 00:46:22,280 --> 00:46:24,050 Again, that's in my email. 966 00:46:24,050 --> 00:46:26,600 If you have questions, please let us 967 00:46:26,600 --> 00:46:29,840 know in particular on Piazza or during office hours. 968 00:46:29,840 --> 00:46:33,380 But in any case, please do not worry too much about the exam. 969 00:46:33,380 --> 00:46:35,840 Try your best, and you will do great. 970 00:46:35,840 --> 00:46:38,480 But you know, even if you don't do great, you'll be fine. 971 00:46:38,480 --> 00:46:41,570 You will pass this class as long as you take the exam 972 00:46:41,570 --> 00:46:45,890 and write something that is remotely reasonable. 973 00:46:45,890 --> 00:46:46,640 Thank you so much. 974 00:46:46,640 --> 00:46:50,320 And I look forward to seeing you in class soon.