WEBVTT 00:03.660 --> 00:05.290 Professor Robert Shiller: I want to talk today about 00:05.293 --> 00:06.283 the efficient markets hypothesis. 00:06.280 --> 00:10.820 00:10.820 --> 00:14.140 Let me just first say, last lecture was about 00:14.142 --> 00:18.672 insurance and I was telling you about the theory of insurance 00:18.672 --> 00:23.202 and how it has evolved over the years and how it has produced 00:23.202 --> 00:25.092 some real benefits. 00:25.090 --> 00:34.000 00:34.000 --> 00:36.600 Is that better? It says, "Mike volume." 00:36.600 --> 00:43.070 00:43.070 --> 00:47.630 I wanted to just tie this in--the advantages of insurance 00:47.625 --> 00:52.665 that we have to some big events that occurred and that will, 00:52.670 --> 00:57.100 I think, point out the strengths and weaknesses of our 00:57.104 --> 01:00.784 institutions. We had a terrible hurricane a 01:00.782 --> 01:03.792 couple of years ago in Los Angeles; 01:03.789 --> 01:11.199 Hurricane Katrina damaged the city of--I'm sorry did I say Los 01:11.204 --> 01:14.954 Angeles? You have to stop me when I say 01:14.948 --> 01:17.728 things that are obviously wrong. 01:17.730 --> 01:22.000 My mind lapses sometimes--New Orleans--and Los Angeles doesn't 01:22.000 --> 01:25.290 have to worry about hurricanes as far as I know, 01:25.290 --> 01:27.880 unless there's some major change. 01:27.879 --> 01:31.609 In New Orleans there was a Hurricane Katrina; 01:31.610 --> 01:36.190 it broke the levies that were surrounding the city and caused 01:36.189 --> 01:38.249 the flooding of the city. 01:38.250 --> 01:44.570 What saved the people of the city, mostly? 01:44.569 --> 01:49.429 I would say it was actually the insurance institutions because 01:49.425 --> 01:54.515 the city was heavily damaged but homes were generally insured. 01:54.519 --> 01:58.949 There were some conflicts when this huge disaster came. 01:58.950 --> 02:03.910 Some people had wind insurance and some people had flood 02:03.908 --> 02:08.508 insurance and it became difficult whether this was a 02:08.505 --> 02:13.195 wind or a flood problem, because the wind caused the 02:13.197 --> 02:15.107 flood. So, if you had only wind 02:15.110 --> 02:16.630 insurance are you covered? 02:16.629 --> 02:20.219 There was a lot bickering and arguments afterwards but I think 02:20.215 --> 02:21.445 it worked out well. 02:21.449 --> 02:26.409 There were surveys of customer satisfaction after the event and 02:26.409 --> 02:30.169 I think, generally, people were happy with their 02:30.169 --> 02:32.089 insurance companies. 02:32.090 --> 02:36.470 Of course, there were some that were not, who may have found out 02:36.465 --> 02:39.365 that they weren't covered; but on the whole, 02:39.369 --> 02:40.949 the experience worked well. 02:40.949 --> 02:47.869 The other thing I want to say about the last lecture is that 02:47.869 --> 02:53.209 as financial progress moves on, the distinction between 02:53.214 --> 02:57.194 insurance and other forms of risk management may get blurred. 02:57.190 --> 03:01.440 One very interesting thing that's been happening is that we 03:01.436 --> 03:05.826 are starting to see development of another institution called 03:05.829 --> 03:09.609 the catastrophe bond, which is another way that 03:09.606 --> 03:14.046 people have for protecting themselves against catastrophes 03:14.045 --> 03:16.065 and it's not insurance. 03:16.069 --> 03:21.389 A catastrophe bond is a bond that the issuer doesn't have to 03:21.392 --> 03:24.552 pay off if there's a catastrophe. 03:24.550 --> 03:29.710 You could have hurricanes--the City of New Orleans could raise 03:29.705 --> 03:34.515 money with catastrophe bonds that they have to pay back if 03:34.523 --> 03:39.603 there's no hurricane but they don't have to pay back if there 03:39.595 --> 03:43.135 is a hurricane. Or it could be some mixture: 03:43.137 --> 03:46.417 they'd pay back part of it if there is a hurricane. 03:46.420 --> 03:48.070 That's like insurance, isn't it? 03:48.069 --> 03:51.609 But it doesn't operate through an insurance company, 03:51.607 --> 03:54.517 it operates through a securities market. 03:54.520 --> 03:58.490 A good example of that is a couple of years ago the 03:58.490 --> 04:02.620 Government of Mexico issued catastrophe bonds against 04:02.619 --> 04:05.989 earthquakes. Mexico City was hit by a 04:05.989 --> 04:11.589 terrible earthquake about twenty years ago and it's vulnerable to 04:11.587 --> 04:15.677 being hit again. So, what does Mexico do about 04:15.677 --> 04:18.147 this? Mexico could wait until there 04:18.153 --> 04:22.243 is a hurricane (sic) and hope that there's some international 04:22.236 --> 04:25.226 relief effort, but that's not a very good way 04:25.231 --> 04:27.771 to proceed. We want to arrange it in 04:27.774 --> 04:30.564 advance. What Mexico did was issue cat 04:30.558 --> 04:35.268 bonds that have to be repaid in the absence of a hurricane and 04:35.267 --> 04:39.047 have a lower repayment if there is--I'm sorry, 04:39.050 --> 04:42.950 I'm not on my best form today--earthquake. 04:42.949 --> 04:46.639 Mexico City does not have to worry about hurricanes either. 04:46.639 --> 04:52.779 Every area is different and they have their own individual 04:52.780 --> 04:56.820 characteristics. Right now, the insurance 04:56.824 --> 05:01.474 industry is a bit challenged because--in terms of some 05:01.471 --> 05:07.171 risks--because the risks seem to be changing through time and, 05:07.170 --> 05:10.570 notably, it looks like hurricane risk is increasing. 05:10.569 --> 05:14.329 So people who are insured in--it seems to be increasing 05:14.334 --> 05:18.104 because of global warming, although I don't know if all 05:18.099 --> 05:20.469 scientists are agreed on that. 05:20.470 --> 05:24.340 If you live in a coastal area of Florida it does appear that 05:24.335 --> 05:26.885 your risk is increasing through time. 05:26.889 --> 05:31.709 So insurance companies want to raise your rates and this is a 05:31.710 --> 05:34.040 huge issue down in Florida. 05:34.040 --> 05:38.870 Well, the government has kind of taken over for the time 05:38.866 --> 05:42.636 being--insurance in Florida--because we have 05:42.639 --> 05:45.379 problems. People are having problems 05:45.379 --> 05:47.839 paying the increased insurance premium. 05:47.839 --> 05:53.869 I don't think we've figured out finally how insurance will 05:53.872 --> 05:58.002 ultimately look in a matter of years. 05:58.000 --> 06:01.930 But, I think that the important thing is that it's protecting 06:01.925 --> 06:04.315 against us already, maybe imperfectly, 06:04.320 --> 06:07.950 but it's already protecting us against some of our worst fears, 06:07.950 --> 06:10.650 like hurricanes and earthquakes. 06:10.649 --> 06:14.039 I think the system is evolving and we're getting new 06:14.037 --> 06:18.287 developments like cat bonds that are changing the way we're doing 06:18.289 --> 06:20.269 things. In the future, 06:20.270 --> 06:25.810 I think these will develop more and make us even better able to 06:25.808 --> 06:28.308 handle catastrophe risks. 06:28.310 --> 06:30.220 Anyway, that's the last lecture. 06:30.220 --> 06:34.210 Today I wanted to talk about--back to securities 06:34.214 --> 06:37.024 markets or actually, more general, 06:37.019 --> 06:42.079 asset markets. I want to talk today about the 06:42.083 --> 06:48.443 efficient markets hypothesis, which is a very important 06:48.443 --> 06:55.043 intellectual construct that has guided a lot of theory in 06:55.038 --> 06:58.518 finance. I want to talk first about the 06:58.524 --> 07:02.304 history of the hypothesis--I haven't defined it yet for you 07:02.303 --> 07:06.153 but maybe you already have heard of this--but history of the 07:06.147 --> 07:08.437 hypothesis, the arguments for it, 07:08.438 --> 07:10.518 and then the arguments against it. 07:10.519 --> 07:16.479 I want to talk about technical analysis and empirical evidence 07:16.476 --> 07:21.646 in the literature--about technical analysis--and other 07:21.651 --> 07:27.121 schools of thought that doubt the market efficiency, 07:27.120 --> 07:29.670 talk somewhat about behavioral economics. 07:29.670 --> 07:33.080 Then, finally, we have a homework assignment; 07:33.080 --> 07:34.220 actually, it's coming up. 07:34.220 --> 07:38.720 I'll start talking about it now so you--it will be an assignment 07:38.718 --> 07:42.998 for you to try to forecast the stock market using statistical 07:43.003 --> 07:46.853 methods. You don't have to look at the 07:46.851 --> 07:49.821 screen yet; I'll come back to that. 07:49.819 --> 07:57.109 What is the efficient markets hypothesis? 07:57.110 --> 08:03.110 The term actually is a fairly recent origin--that is, 08:03.106 --> 08:09.676 a few decades ago--but the idea goes back much further. 08:09.680 --> 08:17.430 The idea is that in asset markets that have good 08:17.429 --> 08:27.319 regulations and market makers and developed markets that have 08:27.321 --> 08:35.731 a lot of depth and liquidity--in these markets, 08:35.730 --> 08:42.400 the prices that you see are perfect indicators of true 08:42.402 --> 08:44.942 value. In other words, 08:44.937 --> 08:51.147 efficient market says that the market efficiently incorporates 08:51.149 --> 08:57.159 all information and the prices are like the best information 08:57.157 --> 09:00.617 about the value of something. 09:00.620 --> 09:03.800 In other words, efficient markets hypothesis 09:03.797 --> 09:07.047 tells you: trust markets, don't trust people, 09:07.049 --> 09:12.149 trust markets. I've been trying to find out 09:12.151 --> 09:14.981 who said that first. 09:14.980 --> 09:20.150 The earliest statement of the efficient markets hypothesis, 09:20.149 --> 09:24.429 although it doesn't call it the efficient markets 09:24.427 --> 09:28.967 hypothesis--the earliest statement that I could find 09:28.972 --> 09:34.142 comes from a book written in 1889 by George Gibson and it's 09:34.141 --> 09:38.421 called The Stock Exchanges of London, 09:38.420 --> 09:40.610 Paris, and New York. 09:40.610 --> 09:45.890 I'll quote him, he said, "When shares become 09:45.889 --> 09:49.739 publicly-known in an open market, 09:49.740 --> 09:53.560 the value, which they acquire there, may be regarded as the 09:53.559 --> 09:57.049 judgment of the best intelligence concerning them." 09:57.049 --> 10:00.069 I was kind of interested to see this book. 10:00.070 --> 10:03.770 I found it in the Mudd Library, actually by accident, 10:03.765 --> 10:06.675 looking through old stock market books. 10:06.679 --> 10:13.659 The book had some interesting observations. 10:13.659 --> 10:16.949 One observation, which took my interest, 10:16.954 --> 10:21.684 was that he points out that in this modern electronic age 10:21.683 --> 10:25.913 information speeds around the globe at the speed of 10:25.907 --> 10:29.367 electricity--or the speed of light. 10:29.370 --> 10:32.260 When I first read that I thought for a minute, 10:32.258 --> 10:33.668 wait a minute, 1889? 10:33.669 --> 10:36.009 That sounds like 1989 or something. 10:36.010 --> 10:41.860 10:41.860 --> 10:42.650 Just a moment. 10:42.650 --> 10:46.110 10:46.110 --> 10:47.920 We found the microphone. 10:47.920 --> 10:49.170 Oh, it was right there? 10:49.170 --> 10:57.540 10:57.540 --> 10:59.880 Now I am free from my tether. 10:59.880 --> 11:05.370 11:05.370 --> 11:09.950 In 1889, they had already invented the telegraph. 11:09.950 --> 11:12.060 In fact, that goes back decades earlier. 11:12.059 --> 11:16.829 The telephone was starting to appear, so it really was true 11:16.825 --> 11:20.765 that information would speed around the globe. 11:20.769 --> 11:25.659 Information that became publicly-known would be entered 11:25.659 --> 11:29.189 into market price almost immediately. 11:29.190 --> 11:34.690 The conclusion that Gibson had was that there's no hope in 11:34.685 --> 11:40.465 trying to beat the price or beat the market because the price 11:40.471 --> 11:44.811 already has all of the information in it. 11:44.809 --> 11:47.869 Let me elaborate on that theme a little bit. 11:47.870 --> 11:52.270 Actually, it goes back before telegraph, there was a famous 11:52.273 --> 11:55.543 story of Mr. Reuters, who had an information 11:55.537 --> 11:59.027 service before the telegraph was invented. 11:59.029 --> 12:04.099 He wanted to help his clients get the information first so 12:04.101 --> 12:09.441 that they could trade on it and he had the idea of using what 12:09.439 --> 12:12.019 you call carrier pigeons. 12:12.019 --> 12:15.249 What you do is you get these birds and you raise them--you 12:15.245 --> 12:17.335 know what I'm talking about, right? 12:17.340 --> 12:20.100 You raise them in one place and then they're going to want to go 12:20.100 --> 12:22.100 back there. Then you take them away in a 12:22.096 --> 12:24.846 cage to another city and when you need to get a message 12:24.847 --> 12:27.847 across, you tie the message to the pigeon's foot and release 12:27.852 --> 12:30.432 it. Then it will fly back and it 12:30.433 --> 12:34.473 will beat any other method of message transmission. 12:34.470 --> 12:38.010 Incidentally, Reuter's information service is 12:38.007 --> 12:42.587 still in business today and now they use computers the way 12:42.590 --> 12:46.110 everyone else does, but the whole principle 12:46.107 --> 12:48.787 precedes the invention of computers. 12:48.789 --> 12:52.629 The idea is that the only way you can beat the market is to 12:52.628 --> 12:55.208 get information that nobody else has. 12:55.210 --> 12:59.710 The way it works today is that the--we can't actually improve 12:59.713 --> 13:03.923 on the speed from 1889 because they were already going at 13:03.916 --> 13:08.116 virtually the speed of light with their information; 13:08.120 --> 13:11.650 but, we can improve on our access to it. 13:11.649 --> 13:15.979 Now, many people have beepers or something like this in their 13:15.983 --> 13:20.393 pocket that wakes up and tells them when news is announced. 13:20.389 --> 13:22.739 What happens where there's news about a stock? 13:22.740 --> 13:27.160 Let's say it's a drug company that just makes an announcement 13:27.156 --> 13:31.566 that it has a new drug--let's say good news--or it has gotten 13:31.572 --> 13:34.592 FDA approval to market some new drug. 13:34.590 --> 13:39.250 Well, it would put that out over the network of information 13:39.246 --> 13:44.136 and some people would have their things beep on them and alert 13:44.144 --> 13:45.834 them immediately. 13:45.830 --> 13:49.660 There are analysts who try to keep up with news about stocks. 13:49.659 --> 13:54.629 So, these analysts then would jump to action when they hear 13:54.632 --> 13:58.532 news like that. That's because they know that 13:58.527 --> 14:03.237 markets move really fast to important new information and 14:03.235 --> 14:06.005 they've got to be there first. 14:06.009 --> 14:09.479 What happens when the drug company announces that they have 14:09.483 --> 14:11.163 some news--important news? 14:11.159 --> 14:14.009 Well, the guys who are with their beepers immediately spring 14:14.013 --> 14:16.483 to action and try to figure out what the news is. 14:16.480 --> 14:19.860 Within seconds they're trading because you know you've got to 14:19.861 --> 14:23.091 be there first; otherwise, you can't benefit 14:23.086 --> 14:25.206 from the news. What happens? 14:25.210 --> 14:27.310 They say they've announced--they've got approval 14:27.314 --> 14:28.214 for this new drug. 14:28.210 --> 14:32.060 Then maybe do a quick call to their drug company expert and 14:32.059 --> 14:35.779 say, quick how much should I change the valuation of this 14:35.776 --> 14:38.146 stock? The guy will give a quick 14:38.152 --> 14:41.032 guess--this is now twenty seconds after the 14:41.028 --> 14:44.378 announcement--and then immediately you place a big 14:44.384 --> 14:47.744 trade for a million shares or plus or minus. 14:47.740 --> 14:51.500 Then the guy calls back thirty seconds later and says, 14:51.500 --> 14:54.410 "Oh no, I wasn't exactly right on that. 14:54.409 --> 14:57.429 I've had thirty more seconds to think about this, 14:57.433 --> 14:59.263 so let me change it again." 14:59.259 --> 15:02.519 Over the next few minutes, the price--a lot of people are 15:02.517 --> 15:05.947 trading like that so the price is jumping around rapidly. 15:05.950 --> 15:08.970 Then, after maybe five minutes, these people have had time to 15:08.965 --> 15:12.175 assimilate it and think about it and check their thinking and the 15:12.181 --> 15:13.791 price starts to settle down. 15:13.789 --> 15:17.069 Maybe I'm exaggerating how fast it settles down. 15:17.070 --> 15:21.080 Maybe an hour later you have a committee meeting and the 15:21.079 --> 15:25.529 experts are arguing about what this really means and trying to 15:25.527 --> 15:29.607 assimilate other information and coordinate with it; 15:29.610 --> 15:32.610 but, after two hours it started to really settle down. 15:32.610 --> 15:34.460 Suppose you then, the next day, 15:34.458 --> 15:37.658 read in The Wall Street Journal about this new 15:37.662 --> 15:40.222 announcement. Do you think you have any 15:40.216 --> 15:43.046 chance of beating the market by trading on it? 15:43.049 --> 15:47.099 I mean, you're like twenty-four hours late, but I hear people 15:47.096 --> 15:49.896 tell me--I hear, "I read in Business Week 15:49.896 --> 15:51.576 that there was a new announcement, 15:51.584 --> 15:53.124 so I'm thinking of buying." 15:53.120 --> 15:55.860 I say, "Well, Business Week--that 15:55.860 --> 15:58.530 information is probably a week old." 15:58.529 --> 16:02.269 Even other people will talk about trading on information 16:02.274 --> 16:05.544 that's years old, so you kind of think that maybe 16:07.720 --> 16:09.810 First of all, you're not a drug company 16:09.812 --> 16:12.072 expert or whatever it is that's needed. 16:12.070 --> 16:16.190 Secondly, you don't know the math--you don't know how to 16:16.189 --> 16:18.959 calculate present values, probably. 16:18.960 --> 16:22.870 Thirdly, you're a month late. 16:22.870 --> 16:27.520 You get the impression that a lot of people shouldn't be 16:27.521 --> 16:29.891 trying to beat the market. 16:29.889 --> 16:33.799 You might say, to a first approximation, 16:33.803 --> 16:38.623 the market has it all right so don't even try. 16:38.620 --> 16:44.130 The efficient markets hypothesis is a hypothesis that 16:44.128 --> 16:49.318 one should respect financial markets very much. 16:49.320 --> 16:54.600 Your textbook by Fabozzi, et al. 16:54.600 --> 16:58.110 mentions--I looked it up in the index to see what they say about 16:58.111 --> 16:59.841 efficient markets hypothesis. 16:59.840 --> 17:03.510 They define it. I'm quoting the textbook 17:03.505 --> 17:08.305 Fabozzi, "Publicly-available, relevant information about the 17:08.311 --> 17:12.551 issuers will lead to correct pricing of freely-traded 17:12.548 --> 17:16.538 securities in properly-functioning markets." 17:16.539 --> 17:20.429 That's their definition of the efficient markets hypothesis. 17:20.430 --> 17:24.760 They didn't say it was right; they just said that's the 17:24.761 --> 17:27.571 hypothesis. What Fabozzi, et al. 17:27.569 --> 17:33.259 said is that the hypothesis has informed a lot of regulation. 17:33.259 --> 17:36.899 The Securities and Exchange Commission and other agencies 17:36.900 --> 17:40.730 that regulate financial markets have shown some faith in the 17:40.734 --> 17:42.884 efficient markets hypothesis. 17:42.880 --> 17:47.010 Therefore, they feel that their--maybe their primary 17:47.014 --> 17:52.124 mission is to regulate the flow of information to make sure that 17:52.122 --> 17:56.662 it's an even playing field so that everyone has access to 17:56.662 --> 17:59.502 information at the same time. 17:59.500 --> 18:01.910 For example, the Securities and Exchange 18:01.907 --> 18:05.117 Commission requires that when a corporation publishes 18:05.116 --> 18:08.816 information that's relevant to the value of their stock, 18:08.819 --> 18:13.069 they have to put it out to everyone at once or there's 18:13.072 --> 18:15.482 rules about what that means. 18:15.480 --> 18:18.690 Typically, they'll have a webcast or something like that 18:18.688 --> 18:21.898 and it's announced in advance, so everyone who is really 18:21.897 --> 18:23.587 interested can listen in. 18:23.589 --> 18:28.099 I don't find a whole lot of enthusiasm in the Fabozzi book 18:28.098 --> 18:32.208 for the efficient markets hypothesis and maybe that's 18:32.211 --> 18:36.011 because it's not exactly right, which is my view. 18:36.010 --> 18:38.910 It's a half truth; I'll come back to that. 18:38.910 --> 18:41.960 I want to quote another best-selling textbook, 18:41.960 --> 18:45.010 not your own, but there's another textbook, 18:45.009 --> 18:48.449 Brealey and Myers, which is a textbook of 18:48.450 --> 18:50.170 corporate finance. 18:50.170 --> 18:53.910 They are much more enthusiastic about efficient markets 18:53.909 --> 18:56.729 hypothesis. At the end of their textbook, 18:56.731 --> 19:00.641 they have a concluding chapter and the concluding chapter is 19:00.640 --> 19:04.680 built around what they call the "seven most important ideas in 19:04.682 --> 19:07.492 finance." One of those seven ideas to 19:07.494 --> 19:09.384 them is efficient markets. 19:09.380 --> 19:13.010 They don't call it a hypothesis, they just say 19:13.005 --> 19:17.915 "efficient markets," which they define as the theory that--I'm 19:17.920 --> 19:20.900 quoting them, "Security prices accurately 19:20.897 --> 19:24.167 reflect available information and respond rapidly to new 19:24.172 --> 19:27.032 information as soon as it becomes available." 19:27.029 --> 19:29.829 Then they have a little qualifier--I think this is 19:29.832 --> 19:32.462 interesting--they say, "Don't misunderstand the 19:32.463 --> 19:35.793 efficient market idea, it doesn't say that there are 19:35.786 --> 19:36.986 no taxes or costs. 19:36.990 --> 19:40.390 It doesn't say that there aren't some clever people and 19:40.392 --> 19:41.592 some stupid ones. 19:41.589 --> 19:45.259 It merely implies that competition in capital markets 19:45.260 --> 19:48.630 is very tough. There are no money machines and 19:48.630 --> 19:53.110 security prices reflect the true underlying value of assets." 19:53.109 --> 19:58.929 That's a pretty enthusiastic endorsement of efficient 19:58.930 --> 20:02.680 markets. I said I have some doubts about 20:02.680 --> 20:05.810 it. I guess I don't--I guess what I 20:05.807 --> 20:10.927 don't like is their concluding statement, "Security prices 20:10.933 --> 20:15.253 reflect the true underlying value of assets." 20:15.250 --> 20:18.850 I don't think that's really true but I guess I agree, 20:18.851 --> 20:22.111 it's tough to make money reliably and quickly in 20:22.106 --> 20:23.626 financial markets. 20:23.630 --> 20:26.390 So, if that's what efficient markets means, 20:26.388 --> 20:27.438 they're right. 20:27.440 --> 20:32.710 20:32.710 --> 20:36.650 "Efficient markets" is not so easy to define. 20:36.650 --> 20:43.270 We can go back to--the term really goes back to 1967 and it 20:43.268 --> 20:49.548 was Professor Harry Roberts at University of Chicago who 20:49.545 --> 20:56.045 defined three different efficient markets hypotheses. 20:56.049 --> 21:05.149 There's the weak form, the semi-strong form, 21:05.153 --> 21:10.873 and then the strong form. 21:10.870 --> 21:14.540 21:14.539 --> 21:17.849 The weak form--these differ only in terms of the amount of 21:17.850 --> 21:21.450 information that is assumed to be efficiently incorporated into 21:21.451 --> 21:25.561 prices. The weak form says that 21:25.559 --> 21:34.159 information of past prices is already in the--incorporated 21:34.158 --> 21:39.588 into price, but only past prices. 21:39.589 --> 21:42.779 What it means "only" is that you can't predict stock prices 21:42.779 --> 21:43.989 by noting that, say, 21:43.990 --> 21:46.310 if it goes up today, it'll probably go up tomorrow 21:46.308 --> 21:48.298 or it goes up today, it'll probably go down 21:48.295 --> 21:50.815 tomorrow. That would be relying only on 21:50.822 --> 21:53.592 past prices. Harry Roberts felt most 21:53.593 --> 21:57.843 confident that this form of the hypothesis was good, 21:57.835 --> 22:01.815 so he called it the weak form; it's the least criticizable 22:01.819 --> 22:03.059 form of efficient markets. 22:03.059 --> 22:09.299 Semi-strong form says that market prices incorporate all 22:09.298 --> 22:11.678 public information. 22:11.680 --> 22:16.270 Anything that's known to the public is already incorporated 22:16.274 --> 22:19.844 into price, so don't bother to trade on it. 22:19.839 --> 22:25.879 The strong form says all information, whether public or 22:25.884 --> 22:29.694 not, is incorporated into price. 22:29.690 --> 22:33.570 This is a really strong--it's the least likely to be true 22:33.566 --> 22:37.926 because every time you increase the information set--what strong 22:37.927 --> 22:41.247 form says is that no information is private, 22:41.250 --> 22:43.150 really, it all gets out into price. 22:43.150 --> 22:46.640 Companies keep secrets, so that's not public 22:46.635 --> 22:49.495 information. The Securities and Exchange 22:49.497 --> 22:52.887 Commission insists that companies keep secrets because 22:52.888 --> 22:56.598 they have to disseminate information in an orderly way. 22:56.599 --> 23:00.849 So, there has to be a secret until a certain hour in which 23:00.845 --> 23:03.075 it's announced to everybody. 23:03.079 --> 23:06.449 But the strong form efficient markets hypothesis is cynical 23:06.448 --> 23:09.408 about that and says, you know, nobody keeps secrets, 23:09.410 --> 23:10.630 it all leaks out. 23:10.630 --> 23:13.960 I think that when we refer to the efficient markets 23:13.963 --> 23:17.633 hypothesis, it's the semi-strong form that we're usually 23:17.630 --> 23:21.430 referring to because the strong form is a bit strong. 23:21.430 --> 23:28.200 23:28.200 --> 23:31.420 The definitions of efficient markets that I gave you are 23:31.424 --> 23:33.364 intuitive but not very precise. 23:33.359 --> 23:37.249 Then you have to ask, well what does it mean to say 23:37.251 --> 23:40.521 that price incorporates all information? 23:40.519 --> 23:43.109 What does it mean to incorporate information? 23:43.109 --> 23:47.729 Unfortunately, there's not one answer to that 23:47.732 --> 23:53.302 question, so I'm going to give the simplest answer. 23:53.299 --> 23:56.139 What does efficient markets mean? 23:56.140 --> 24:07.180 I might--I'll have to--this is the simplest version but it's 24:07.178 --> 24:14.848 often the one that is referred to most. 24:14.849 --> 24:19.309 That is that price is the expected value--the expected 24:19.305 --> 24:23.755 present value--of future dividends paid on a stock. 24:23.760 --> 24:32.840 24:32.839 --> 24:35.589 The efficient markets hypothesis says, 24:35.594 --> 24:39.844 the true value of a stock comes from the dividends that it 24:39.837 --> 24:44.297 pays--that's a cash flow that is valued by the market and the 24:44.304 --> 24:47.514 market values it as the present value of the 24:47.505 --> 24:50.925 optimally-forecasted future dividends. 24:50.930 --> 24:56.650 The theory that's most often referred to is the simple--I've 24:56.648 --> 25:01.978 already talked in the second lecture about present value 25:01.979 --> 25:05.029 formulas; we had a growing-perpetuity 25:05.034 --> 25:08.904 model. Remember that I said that the 25:08.897 --> 25:15.337 present value--the present discounted value of a growing 25:15.339 --> 25:21.079 perpetuity that pays an amount D--well, 25:21.079 --> 25:31.049 if it pays De^(gt) or D_0e^(gt) is 25:31.047 --> 25:35.597 the dividend, so it's growing at rate, 25:35.604 --> 25:39.464 g. Then, the formula we had for 25:39.458 --> 25:43.158 the present value was D/(r-g). 25:43.160 --> 25:44.950 Remember that? 25:44.950 --> 25:48.880 25:48.880 --> 25:51.490 This assumes, of course, that the growth of 25:51.486 --> 25:54.396 dividends has to be less than the discount rate, 25:54.403 --> 25:57.813 r. The simplest version of the 25:57.809 --> 26:03.109 efficient markets hypothesis says, price is equal to the 26:03.105 --> 26:08.785 dividend all over the discount rate minus the growth rate of 26:08.785 --> 26:11.165 dividends, where g, 26:11.165 --> 26:14.595 the growth rate of dividends, is an optimally-forecasted 26:14.595 --> 26:16.275 growth rate of dividend. 26:16.279 --> 26:24.519 That gives us a value--a model for the price. 26:24.519 --> 26:27.749 Another way of writing, more generally--if I don't 26:27.754 --> 26:31.914 assume the constant growth rate of dividends--is to write just a 26:31.913 --> 26:33.633 present value formula. 26:33.630 --> 26:38.000 It's another less strong form of writing down the efficient 26:38.001 --> 26:42.371 markets hypothesis because it doesn't say how dividends are 26:42.372 --> 26:47.382 thought to grow. But you can write: 26:47.377 --> 26:58.647 price is equal to the summation of E(D_t+k)]/(1 + 26:58.646 --> 27:06.946 r)^(k); k = 1 to infinity. 27:06.950 --> 27:10.240 This is another--that's just the present value formula. 27:10.240 --> 27:16.850 I have the price--I'm sorry, this should be expectation at 27:16.851 --> 27:23.581 time, t--the price at t is the expectation of 27:23.578 --> 27:28.098 the dividend, at time t plus k, 27:28.103 --> 27:32.713 discounted by a discount factor r. That's just the 27:32.714 --> 27:37.824 present value formula where I've substituted an expectation for 27:37.821 --> 27:39.881 the future dividend. 27:39.880 --> 27:46.740 That's the efficient markets theory in this incarnation. 27:46.740 --> 27:49.370 There are other ways to envision the efficient 27:49.373 --> 27:52.473 markets--what it means--but let's consider this simple 27:52.474 --> 27:56.734 story. What this means then is that 27:56.731 --> 28:04.081 the price is a forecast of future dividends to be paid on 28:04.075 --> 28:08.175 the stock. This would be--of the present 28:08.180 --> 28:12.980 value of future dividends to be paid on the stock--and this 28:12.980 --> 28:17.580 means then that price, relative to dividend, 28:17.575 --> 28:24.425 is related to expected future growth rates of dividends. 28:24.430 --> 28:29.200 If you expect dividends to grow a lot, if g is high, 28:29.201 --> 28:32.471 then price will be high, relative to dividends, 28:32.471 --> 28:34.941 because this is subtracted off the denominator. 28:34.940 --> 28:38.860 It makes the denominator smaller and it makes the price 28:38.857 --> 28:40.467 higher. On the other hand, 28:40.465 --> 28:43.115 if you expect dividends to do poorly in the future, 28:43.121 --> 28:45.671 then price will be low relative to dividends; 28:45.670 --> 28:48.720 that's what the efficient markets hypothesis would say. 28:48.720 --> 28:55.770 I would give you an example of that. 28:55.769 --> 28:58.669 I talked last year about a company that I read in 28:58.671 --> 29:01.871 Business Week--that I read about in Business 29:01.874 --> 29:04.484 Week, so this is a year old story in 29:04.483 --> 29:05.823 Business Week. 29:05.819 --> 29:10.299 There was a company called First Federal Financial, 29:10.304 --> 29:14.254 which was a company that issues mortgages. 29:14.250 --> 29:18.200 This was in January of 2007, actually the Business 29:18.197 --> 29:21.307 Week story was in December of 2006, 29:21.309 --> 29:25.759 but I was still reacting to it a year ago, in January of 2007. 29:25.760 --> 29:30.610 29:30.609 --> 29:36.419 The Business Week story--this is First Federal 29:36.418 --> 29:43.008 Financial--was referring to the fact that the price-dividend 29:43.009 --> 29:49.599 ratio--actually they talked about price-earnings ratio, 29:49.599 --> 29:56.539 but the price-earnings ratio for this company was very low. 29:56.539 --> 30:09.589 It was only 8.5 in December 2006. 30:09.589 --> 30:13.519 Typically, price-earnings ratios of the companies are very 30:13.518 --> 30:17.238 high, much higher than that--typically like fifteen. 30:17.240 --> 30:21.240 The price of First Federal Financial, relative to its 30:21.235 --> 30:24.765 earnings, was very low, so some people might be 30:24.769 --> 30:28.149 inclined to think well that looks cheap. 30:28.150 --> 30:32.040 I can buy--by buying First Federal Financial, 30:32.035 --> 30:37.505 I can buy the stock at a low price relative to its earnings. 30:37.509 --> 30:41.129 But, if you believe in efficient markets you wouldn't 30:41.132 --> 30:44.962 think that this is any reason to buy the company because 30:44.963 --> 30:49.353 efficient markets would say that if the price is low relative to 30:49.351 --> 30:53.871 either dividends or earnings, it must mean that people think 30:53.867 --> 30:57.467 that bad things are going to come to the dividends or 30:57.468 --> 30:59.458 earnings. In other words, 30:59.463 --> 31:04.123 the First Federal Financial has a low expected growth rate of 31:04.116 --> 31:06.206 dividends in the future. 31:06.210 --> 31:10.560 So, I was interested in this particular story because 31:10.555 --> 31:15.645 Business Week wrote an article about them and noted the 31:15.654 --> 31:18.834 low price-earnings ratio and said, 31:18.830 --> 31:20.470 what does this mean? 31:20.470 --> 31:25.690 What Business Week presented a year ago was an 31:25.686 --> 31:30.196 argument why g was likely to be low. 31:30.200 --> 31:34.200 Yes? Student: [inaudible] 31:34.201 --> 31:40.161 Professor Robert Shiller: If they don't pay 31:40.157 --> 31:45.017 dividends you can't use this formula. 31:45.019 --> 31:48.829 I'm not sure what the dividends were at First Federal Financial. 31:48.829 --> 31:52.009 I only know the price-earnings ratio at that time. 31:52.009 --> 31:56.629 You're right, you cannot use this formula if 31:56.631 --> 32:02.871 they're not paying a dividend today because this formula--I 32:02.865 --> 32:05.975 erased it, but it was up here--assumes 32:05.982 --> 32:08.612 that dividends are following a growth path. 32:08.609 --> 32:12.109 If they're not paying a dividend, then we're very 32:12.109 --> 32:15.609 clearly--that's not an appropriate assumption. 32:15.609 --> 32:19.539 Student: So how would the efficient market theory 32:19.544 --> 32:23.554 explain prices of companies that do not pay dividends? 32:23.549 --> 32:24.119 Professor Robert Shiller: Right. 32:24.119 --> 32:26.519 So for–incidentally, if you know, 32:26.517 --> 32:30.047 Berkshire Hathaway is the company Warren Buffett owns and 32:30.051 --> 32:33.401 it's a very famous company because it's done extremely 32:33.395 --> 32:35.585 well. Warren Buffett is regarded as, 32:35.585 --> 32:37.625 by many people, as the financial genius. 32:37.630 --> 32:41.270 If Berkshire Hathaway is not paying a dividend, 32:41.265 --> 32:44.105 we have to revert to this formula. 32:44.109 --> 32:48.279 The efficient market theory would say, well they're going to 32:48.276 --> 32:52.016 pay a dividend eventually, so the price reflects these 32:52.019 --> 32:54.569 future terms. If you spell this out, 32:54.572 --> 32:58.202 this is the expected dividend next year plus the dividend in 32:58.197 --> 33:01.947 two years--this is divided by (1 + r)^(2)--the expected 33:01.946 --> 33:05.196 dividend in three years divided by (1 + r)^(3), 33:05.203 --> 33:08.673 etc. This theory would say that the 33:08.665 --> 33:13.765 dividend--that Berkshire Hathaway has value only because 33:13.772 --> 33:19.532 they--investors expect them to pay dividends in the future. 33:19.529 --> 33:23.389 That sounds right because if Berkshire Hathaway is never 33:23.390 --> 33:26.760 going to pay dividends, why would you hold it? 33:26.759 --> 33:29.129 You might say, I'll hold it because I could 33:29.133 --> 33:31.453 sell to someone else at a higher price. 33:31.450 --> 33:34.870 But then you say, well why would anyone else buy 33:34.866 --> 33:37.086 it? Look, if they're never going to 33:37.089 --> 33:39.069 pay a dividend, what good is it? 33:39.069 --> 33:43.829 It's just a piece of paper, unless I can sell to a greater 33:43.834 --> 33:46.314 fool. But anyone who buys it is 33:46.305 --> 33:50.445 either--would be buying it either on the assumption that 33:50.454 --> 33:54.604 there's some greater fool coming or they would be a fool 33:54.603 --> 33:57.793 themselves. This theory says that the 33:57.792 --> 34:02.632 value--if Warren Buffett--of course, he can't even say this, 34:02.630 --> 34:04.940 but let's somehow say the company could say, 34:04.936 --> 34:06.596 we will never pay a dividend. 34:06.599 --> 34:10.089 This company is going to give itself away to charity someday 34:10.094 --> 34:12.764 and you won't get a penny as a stockholder. 34:12.760 --> 34:15.890 Well, if that happened the price should be zero. 34:15.889 --> 34:18.059 It would convert into a non-profit, like Yale 34:18.059 --> 34:20.439 University. So what's the price of a share 34:20.440 --> 34:21.600 in Yale University? 34:21.599 --> 34:24.839 I mean, it's undefined--I its guess zero, right? 34:24.840 --> 34:27.660 Because Yale is not paying--I'll write a piece of 34:27.659 --> 34:30.009 paper for you and say this is a share. 34:30.010 --> 34:32.820 On my authority, this is a share in Yale 34:32.822 --> 34:35.842 University. It might just as well be that 34:35.841 --> 34:40.001 because it also says in my fine print, you will never get a 34:40.004 --> 34:41.444 dividend on this. 34:41.440 --> 34:43.420 So what's the point, right? 34:43.420 --> 34:47.200 Incidentally, Microsoft for many years never 34:47.198 --> 34:50.608 paid a dividend. It's often common for young 34:50.610 --> 34:54.350 companies not to pay dividends; but they did start paying a 34:54.348 --> 34:57.238 dividend. The whole theory--efficient 34:57.241 --> 35:02.551 markets theory--says that that's what people are looking for. 35:02.550 --> 35:06.320 That's why the value of a company is related to its 35:06.319 --> 35:09.149 activities; otherwise, if the company never 35:09.153 --> 35:11.843 paid a dividend, then what would you care what 35:11.844 --> 35:13.284 the company is doing? 35:13.280 --> 35:16.230 You only care about it because someday they're going to give 35:16.230 --> 35:19.920 you money. A lot of investors forget that; 35:21.579 --> 35:25.359 They think that somehow stocks generate capital gains--prices 35:25.361 --> 35:29.211 go up--but you have to realize and efficient markets theory is 35:29.207 --> 35:33.047 saying this: prices only go up because there's new information 35:33.052 --> 35:34.882 about future dividends. 35:34.880 --> 35:39.730 First Financial anyway--this is an efficient markets 35:39.734 --> 35:45.644 story--First Financial had a low P/E, so people were wondering, 35:45.635 --> 35:47.725 is this a bargain? 35:47.730 --> 35:51.220 This is a cheap stock--P/E means price-earnings ratio. 35:51.219 --> 35:55.909 The Business Week article pointed out that 40% of 35:55.905 --> 36:00.675 First Federal had been sold short--40% of their stock had 36:00.676 --> 36:03.706 been sold short; this is a very high level of 36:03.706 --> 36:05.506 shortage. You know what means? 36:05.510 --> 36:09.590 That means a lot of investors said, "I don't like First 36:09.591 --> 36:13.371 Federal Financial, I don't want to invest in it. 36:13.369 --> 36:15.519 Worse than that, I want to go and short them." 36:15.519 --> 36:18.739 That means you borrow shares and sell them and hope that the 36:18.738 --> 36:21.468 price goes down. When you have 40% of the shares 36:21.471 --> 36:25.031 sold short that means that there were a lot of people who didn't 36:25.032 --> 36:27.182 believe in First Federal Financial. 36:27.179 --> 36:29.639 Business Week, in its article, 36:29.640 --> 36:33.540 pointed out that this--First Federal Financial was a small 36:33.535 --> 36:37.835 mortgage lender--this is before the mortgage crisis that we have 36:37.841 --> 36:40.781 started--in Santa Monica, California. 36:40.780 --> 36:45.150 And it was particularly innovative, in a sense, 36:45.153 --> 36:47.763 in its lending. Notably, 80%, 36:47.761 --> 36:51.201 according to the Business Week article, 36:51.195 --> 36:55.615 80% of the mortgages it's issued were no-doc mortgages. 36:55.619 --> 36:57.669 Do you know what a no-doc mortgage is? 36:57.670 --> 37:00.050 It's something that appeared recently in the housing frenzy. 37:00.050 --> 37:03.310 A no-doc mortgage is a mortgage where you walk in and say, 37:03.313 --> 37:05.893 "I want to borrow money to buy this house." 37:05.889 --> 37:08.259 The company says, "Fine we'll give you a 37:08.261 --> 37:10.971 mortgage. We won't even ask you to have 37:10.966 --> 37:14.896 your employer send a letter saying that you have a job. 37:14.900 --> 37:17.760 We won't even ask you to prove that you own anything or 37:17.757 --> 37:19.977 have--we'll just give you the mortgage." 37:19.980 --> 37:24.620 That's considered by many people risky behavior but it was 37:24.624 --> 37:27.154 done during the housing boom. 37:27.150 --> 37:31.110 Also, they issued an unusually high proportion of what are 37:31.112 --> 37:34.522 called "option ARMs." These are mortgages that are 37:34.519 --> 37:39.039 adjustable-rate, but the person doesn't have to 37:39.043 --> 37:42.493 pay the full payment every time. 37:42.489 --> 37:45.759 You have the option of paying and if you don't want to pay, 37:45.761 --> 37:47.511 you can delay it for a while. 37:47.510 --> 37:50.210 These are also controversial because it thought they would 37:50.207 --> 37:52.097 attract borrowers who were unreliable. 37:52.099 --> 37:56.419 Borrowers who thought that they could afford the house because 37:56.416 --> 37:59.046 they didn't have to pay now; but of course, 37:59.053 --> 38:01.183 it's all going to come later and then they might default. 38:01.179 --> 38:05.579 Business Week thought that the low price-earnings 38:05.576 --> 38:10.366 ratio was because the market expected earnings to go down and 38:10.371 --> 38:12.371 dividends to go down. 38:12.369 --> 38:14.829 A year later, that is this morning, 38:14.832 --> 38:18.962 I wanted to see what First Federal Financial was doing. 38:18.960 --> 38:21.420 Actually, it's still in business; 38:21.420 --> 38:22.880 everything is all right. 38:22.880 --> 38:32.550 It has positive earnings but the price was $70 in beginning 38:32.550 --> 38:40.720 of 2007 and now it's down to about $40 in 2008. 38:40.719 --> 38:44.739 This is a testimony to efficient markets. 38:44.739 --> 38:48.489 The market was expecting the price-earnings ratio--they gave 38:48.485 --> 38:51.845 it a low P/E because they were worried about--they had 38:51.850 --> 38:55.340 information that there was something going badly in this 38:55.342 --> 38:58.112 company and indeed, they were right. 38:58.110 --> 39:01.610 It's not as bad as you might have thought with the 39:01.609 --> 39:04.679 Business Week story but it is bad. 39:04.679 --> 39:08.929 I guess the lesson--are you following what I'm talking--what 39:08.933 --> 39:11.893 happened was they did get into trouble. 39:11.889 --> 39:20.719 The low P/E was indeed a forecaster of bad performance 39:20.717 --> 39:22.047 later. 39:22.050 --> 39:29.290 39:29.289 --> 39:33.129 Now, I guess this isn't exactly--efficient markets 39:33.132 --> 39:36.902 wouldn't--in this case, the people who sold First 39:36.897 --> 39:41.757 Federal Financial short made a lot of money when the price went 39:41.759 --> 39:44.719 down so much. That's not really consistent 39:44.722 --> 39:47.772 with efficient markets because it makes it sound like--as a 39:47.773 --> 39:50.313 matter of fact, I could have--Reading the 39:50.313 --> 39:53.823 Business Week article at the beginning of 2007, 39:53.820 --> 39:58.180 I could have called my broker and I could have told my broker, 39:58.179 --> 40:01.789 short–please, I want to short First Federal 40:01.789 --> 40:03.929 Financial. I could have done that and it 40:03.931 --> 40:06.571 looks like I would have made a lot of money by shorting it 40:06.568 --> 40:08.278 because the price went down a lot. 40:08.280 --> 40:13.220 Efficient markets theory would have to say that that was an 40:13.215 --> 40:16.615 anomaly--that that was just good luck. 40:16.619 --> 40:19.739 Those people who shorted First Federal Financial did make a lot 40:19.736 --> 40:21.896 of money but, hey, they were just lucky this 40:21.897 --> 40:23.327 time. In other words, 40:23.328 --> 40:27.238 efficient markets theory would say that the price was already 40:27.243 --> 40:30.243 down as far as it would--basically as far as it 40:30.244 --> 40:33.384 would go in 2007 and that earnings would fall and 40:33.375 --> 40:37.415 dividends would fall later and that would explain why the price 40:37.420 --> 40:40.170 was low. But, it wouldn't allow you to 40:40.167 --> 40:41.237 predict the price. 40:41.239 --> 40:47.809 This anomaly of--the fact that it worked out so well is really 40:47.812 --> 40:53.202 not--to short investors--is really not testimony to 40:53.199 --> 40:55.569 efficient markets. 40:55.570 --> 41:01.960 Efficient markets--in some sense I think I'm sympathetic to 41:01.964 --> 41:04.174 efficient markets. 41:04.170 --> 41:05.980 You have to be sympathetic to some extent. 41:05.980 --> 41:09.190 If I were trying--I actually did not short First Federal 41:09.188 --> 41:12.398 Financial when I read this last year and it was because, 41:12.396 --> 41:14.966 I guess, I had my doubt; I had my feeling. 41:14.969 --> 41:18.459 You read a Business Week article and it's--everyone in 41:18.461 --> 41:21.901 the world knows it now that a lot people doubt First Federal 41:21.895 --> 41:24.285 Financial. If I come in late as a short 41:24.290 --> 41:26.040 seller, am I going to do well? 41:26.039 --> 41:29.099 I start to doubt it myself because there are so many people 41:29.103 --> 41:32.483 looking at it and so many people who are more knowledgeable about 41:32.483 --> 41:34.493 First Federal Financial than I am; 41:34.489 --> 41:36.829 so, maybe I won't try shorting it. 41:36.829 --> 41:39.269 That's what efficient markets is all about. 41:39.270 --> 41:44.680 41:44.679 --> 41:51.889 The efficient markets theory became very popular in finance 41:51.894 --> 41:59.114 around 1970 and it became a prominent theory in finance. 41:59.110 --> 42:04.940 It has a particular incarnation that I wanted to emphasize 42:04.938 --> 42:07.288 called "random walk." 42:07.289 --> 42:12.719 The random walk theory, which follows loosely from this 42:12.722 --> 42:17.952 formula, though I have to qualify it--but let me talk 42:17.953 --> 42:22.383 about random walk as a theory in itself. 42:22.380 --> 42:29.470 The random walk theory says that under efficient markets, 42:29.473 --> 42:35.813 stock prices and other speculative asset prices are 42:35.806 --> 42:41.106 random walks. That term goes back to Karl 42:41.110 --> 42:46.710 Pearson in an article in Nature in 1905. 42:46.710 --> 42:51.530 This idea is about a hundred years old. 42:51.530 --> 42:55.970 What Pearson wrote about was–well, 42:55.974 --> 42:59.854 the example he gave was a drunk. 42:59.849 --> 43:04.009 Imagine a person who is so drunk that every step that 43:04.014 --> 43:08.344 person takes is random and independent of the preceding 43:08.339 --> 43:11.719 step. Suppose we--this is a 1905 43:11.715 --> 43:18.045 story--suppose we had a lamppost and we had a drunk standing at 43:18.048 --> 43:21.928 the lamppost--I've drawn a picture, 43:21.929 --> 43:23.859 that's a lamp post and that's a drunk. 43:23.860 --> 43:28.040 This person is so intoxicated that steps are completely random 43:28.038 --> 43:32.288 and your objective is to predict where will this person be in a 43:32.286 --> 43:34.086 minute, in ten minutes, 43:34.090 --> 43:35.490 and twenty minutes. 43:35.489 --> 43:40.859 Well, Pearson wrote about this and said that the optimal 43:40.861 --> 43:47.011 prediction is to assume that the person in ten minutes--the best 43:47.013 --> 43:50.923 forecast is the person will be here. 43:50.920 --> 43:53.750 In twenty minutes, what's the best forecast? 43:53.750 --> 43:55.740 The person will be here. 43:55.739 --> 43:58.519 Of course, they probably won't because they're randomly 43:58.521 --> 44:01.611 staggering around but the point is that if it's a true random 44:01.610 --> 44:02.950 walk, there's no bias. 44:02.949 --> 44:05.709 It's equally likely to go this way or this way. 44:05.710 --> 44:09.520 The most likely place for the person is right where that 44:09.521 --> 44:12.601 person is now. So Pearson and other people 44:12.596 --> 44:16.456 following him thought that speculative prices are like 44:16.457 --> 44:18.907 that. That sounds like the markets 44:18.909 --> 44:22.179 are crazy--they're drunk--but they're not drunk. 44:22.179 --> 44:25.789 It's because they respond only to new information. 44:25.789 --> 44:29.549 New information is, by its essence, 44:29.550 --> 44:33.470 unforecastable. So, it has to look like the 44:33.466 --> 44:36.556 market is driven by a drunk when, in fact, 44:36.558 --> 44:40.178 it's very precise and responding optimally to new 44:40.177 --> 44:43.187 information. That's one of the paradoxes 44:43.190 --> 44:44.600 that confuses people. 44:44.599 --> 44:49.359 Statisticians developed the theory of a random walk and let 44:49.358 --> 44:51.408 me define that for you. 44:51.409 --> 44:55.839 A random walk occurs when you have a variable, 44:55.843 --> 45:00.573 x, at time, t, equal to x at 45:00.573 --> 45:05.803 time (t – 1) + e_t where 45:05.795 --> 45:11.505 e_t is noise--unforecastable noise. 45:11.510 --> 45:17.190 This is the random walk; x would be here--how far 45:17.185 --> 45:20.305 the person is from the lamppost. 45:20.309 --> 45:23.709 Here's 0, let's say, and x is the distance 45:23.712 --> 45:25.132 from the lamppost. 45:25.130 --> 45:28.100 At every time interval we have x_t which is 45:28.101 --> 45:30.671 how far the drunk has deviated from the lamppost. 45:30.670 --> 45:33.930 This is the random step; this is where the drunk was 45:33.932 --> 45:37.072 last period and this is where the drunk is this period. 45:37.070 --> 45:40.650 So that's a random walk. 45:40.650 --> 45:44.610 In the heyday of efficient markets theory, 45:44.613 --> 45:50.223 people said that efficient markets is working very well. 45:50.219 --> 45:54.809 In other words, the random walk hypothesis 45:54.805 --> 45:58.715 describes stock prices very well. 45:58.719 --> 46:03.909 Let me go to my spreadsheet, which I have up here, 46:03.914 --> 46:09.644 and I'll put this spreadsheet up on the web for you. 46:09.639 --> 46:14.699 What I have here is shown--there are two lines 46:14.697 --> 46:20.987 shown, one is a blue line, which is the--the blue line is 46:20.990 --> 46:26.950 the actual Standard & Poor's composite stock price 46:26.946 --> 46:30.426 index going back to 1871. 46:30.429 --> 46:32.419 This is a series that I–well, 46:32.416 --> 46:35.036 actually I got this series from Standard & 46:35.044 --> 46:37.794 Poor's and I emphasize it a lot in my book. 46:37.789 --> 46:42.919 It's a 130-year long stock price series; 46:42.920 --> 46:45.310 it's monthly for the United States. 46:45.309 --> 46:50.149 It's now called the S&P 500 because, starting in 1957, 46:50.150 --> 46:53.890 Standard & Poor's kind of reorganized the 46:53.887 --> 46:57.367 index and they kept it at 500 stocks. 46:57.369 --> 47:00.939 This is the second most famous stock price index after the Dow 47:00.939 --> 47:02.519 Jones Industrial Average. 47:02.519 --> 47:06.339 I think it's better than the Dow Jones Industrial Average 47:06.339 --> 47:10.359 because the Dow has only 30 stocks and this has 500 and it's 47:10.363 --> 47:13.163 representative of most of the market. 47:13.160 --> 47:17.530 That's history; that blue line there is history. 47:17.530 --> 47:22.740 What I did on Excel was I generated a random walk, 47:22.742 --> 47:28.382 because there's a random number generator on Excel. 47:28.380 --> 47:31.680 I used the random number generator and plugged it into 47:31.681 --> 47:33.861 this formula: x_t = 47:33.861 --> 47:37.221 x_t-1 + e_t That pink 47:37.224 --> 47:39.284 line is a true random walk. 47:39.280 --> 47:42.310 It was generated by a random number generator. 47:42.309 --> 47:47.249 The point here is--don't they look kind of similar? 47:47.250 --> 47:52.550 When you look at--when people look at stock prices they get 47:52.549 --> 47:54.559 a--it's an illusion. 47:54.559 --> 47:55.929 This is a psychological illusion. 47:55.929 --> 47:59.589 They get the sense that there are bull markets when the market 47:59.590 --> 48:03.310 is going up and there are bear markets when the market is going 48:03.309 --> 48:04.809 down. There must be some force 48:04.809 --> 48:05.889 pushing it up for a while. 48:05.889 --> 48:09.299 You can see, for example--I'm looking at the 48:09.297 --> 48:13.017 blue line--there was a bull market in the 1920s, 48:13.021 --> 48:16.351 a famous bull market, the Roaring 20s. 48:16.349 --> 48:19.719 You would say, that just can't be a random 48:19.715 --> 48:21.575 walk; it was just going up all the 48:21.582 --> 48:24.642 time. Then, this is the 1929 peak and 48:24.636 --> 48:27.946 there's the crash from 1929 to 1932; 48:27.950 --> 48:30.920 a big crash. That can't be a random walk, 48:30.916 --> 48:33.056 right? That's what people think. 48:33.060 --> 48:34.320 They have this intuitive idea. 48:34.320 --> 48:37.400 But, if you look, this random walk seems to do 48:37.404 --> 48:40.294 the same thing. The pink line has a--look 48:40.289 --> 48:44.209 there's a nice bull market there--there's another crash. 48:44.210 --> 48:48.400 Look at this whole period here; this was really strong and then 48:48.402 --> 48:50.252 it leveled out. We had bad--in fact, 48:50.246 --> 48:52.166 they kind of look similar, don't they? 48:52.170 --> 48:54.880 This is pure chance because that random walk that I 48:54.880 --> 48:56.670 generated was pure random walk. 48:56.670 --> 49:01.340 There's the bull--the actual bull market of the 1950s and 60s 49:01.341 --> 49:05.391 and my random walk comes out pretty close to that. 49:05.389 --> 49:09.749 The random walk theory says that people are operating under 49:09.748 --> 49:12.578 illusions; that there really are no trends 49:12.578 --> 49:15.128 or there's no way to predict the market. 49:15.130 --> 49:17.390 It is just completely unforecastable. 49:17.389 --> 49:20.889 Now incidentally, a nice thing about Excel is I 49:20.889 --> 49:25.529 can generate a new random walk for you on the spot by pressing 49:25.529 --> 49:28.419 F9--if you know that key on Excel. 49:28.420 --> 49:31.500 So, I will press F9 and history is going to stay the same--the 49:31.496 --> 49:34.616 blue line is going to stay the same--but the pink line is going 49:34.623 --> 49:36.733 to change. I can generate whole 49:36.732 --> 49:40.662 hundred-year history--pseudo histories--with a press of a 49:40.658 --> 49:43.798 button. Let's see if this works, 49:43.801 --> 49:47.261 we just got--what is that saying? 49:47.260 --> 49:51.080 I put an upward trend to the random walk there because I 49:51.077 --> 49:54.407 think there's an up trend to the stock market. 49:54.409 --> 49:56.609 It's a little different than this, I just added a time trend; 49:56.610 --> 49:59.640 otherwise, it's exactly the random walk formula that I used. 49:59.639 --> 50:04.389 Okay, that looks pretty much like history too except we would 50:04.389 --> 50:08.109 be in a bear market for quite a while, right? 50:08.110 --> 50:11.680 We have quite a bear market in the pink line but I can correct 50:11.680 --> 50:13.320 that by pressing F9 again. 50:13.320 --> 50:20.270 This is bad luck, the stock market crashed in 50:20.269 --> 50:28.209 the--it went off the chart here; that's bad luck but I can try 50:28.209 --> 50:29.979 again. Look at that one, 50:29.979 --> 50:31.399 isn't that a beauty? 50:31.400 --> 50:38.140 That almost--it really looks like the market we've had. 50:38.140 --> 50:43.060 50:43.059 --> 50:44.989 So, I can just go--it's just amazing how fast this thing 50:44.989 --> 50:45.409 generates. 50:45.410 --> 50:49.480 50:49.480 --> 50:52.410 I should stop. Tell me when one looks really 50:52.411 --> 50:54.191 interesting. That's an interesting one. 50:54.190 --> 50:57.610 That's--if that was the history that we randomly had, 50:57.608 --> 51:01.288 Jeremy Siegel would write not just Stocks for the Long 51:01.291 --> 51:04.101 Run. He would be jubilant if the 51:04.104 --> 51:08.834 returns on stocks--he would say for a hundred years--this is 130 51:08.827 --> 51:13.327 years--for 130 years stocks have wonderfully outperformed the 51:13.326 --> 51:16.246 market--outperformed other things. 51:16.250 --> 51:18.770 That is looking pretty good, isn't it? 51:18.769 --> 51:22.349 Now, I just press the button and get--I'm not getting many 51:22.349 --> 51:23.759 bad ones. That probably has an uptrend in 51:23.757 --> 51:23.847 it. 51:23.850 --> 51:29.180 51:29.180 --> 51:30.760 They all look similar, right? 51:30.760 --> 51:34.950 The other thing I wanted to talk about was another thing 51:34.945 --> 51:38.365 called an AR-1, which is a different story. 51:38.369 --> 51:45.519 The AR-1--this is the random walk, here. 51:45.520 --> 51:49.520 51:49.519 --> 51:52.719 The other idea is that--let me do it intuitively first. 51:52.719 --> 51:57.669 I'm going to tie an elastic cord around the ankle of the 51:57.670 --> 52:00.820 drunk and tie it to the lamppost. 52:00.820 --> 52:04.110 You see my elastic cord? 52:04.110 --> 52:07.520 It's very loose at the beginning so the drunk can move 52:07.518 --> 52:10.218 freely but if it starts to get stretchy, 52:10.219 --> 52:12.789 when the drunk makes it all the way over here, 52:12.790 --> 52:15.990 then its really hard for--it's pulling the drunk back. 52:15.989 --> 52:21.989 That means it biases the drunk back towards the starting point. 52:21.989 --> 52:40.079 That's called an AR-1 or a first order auto-regressive 52:40.079 --> 52:44.409 model. That means that there's 52:44.405 --> 52:47.315 something--the random walk has no center; 52:47.320 --> 52:48.610 it just has a starting point. 52:48.610 --> 52:51.640 It doesn't tend to go back anywhere, but here we have a 52:51.635 --> 52:53.885 center. What do I call that? 52:53.889 --> 52:57.669 I'll call it--x-bar is this point here, 52:57.670 --> 52:59.350 not necessarily 0. 52:59.349 --> 53:03.669 It says that (x_t - 53:03.673 --> 53:08.993 x-bar) = x-bar + ρ 53:08.985 --> 53:14.295 (x_t _– 1 - 53:14.297 --> 53:19.607 x-bar) + e_t. 53:19.610 --> 53:28.530 ρ lies between minus one and positive one; 53:28.530 --> 53:29.870 usually it's positive. 53:29.870 --> 53:33.750 You see that? So ρ--the smaller 53:33.754 --> 53:36.034 ρ is, the tighter the elastic cord 53:36.027 --> 53:38.827 is. It gets pulled back by 1- 53:38.826 --> 53:44.086 ρ of the way back to the x-bar in every time 53:44.088 --> 53:46.388 period. But then there's new noise that 53:46.392 --> 53:49.002 adds onto it. If there were no new noise, 53:49.002 --> 53:52.802 then the drunk would just be gradually pulled back; 53:52.800 --> 54:00.960 but there's new noise, so the drunk is not so 54:00.955 --> 54:05.235 predictable. That's another model of the 54:05.243 --> 54:08.633 stock market but it's not an efficient markets model. 54:08.630 --> 54:11.820 The random walk is the efficient markets model that 54:11.819 --> 54:15.389 says, you can't profit by trading because you just cannot 54:15.392 --> 54:17.372 predict the changing price. 54:17.369 --> 54:19.679 But with the first order auto-regressive, 54:19.682 --> 54:23.212 you can predict the change of price, at least a little bit. 54:23.210 --> 54:28.140 Well when it's--what goes up comes down. 54:28.139 --> 54:31.649 If the drunk is over here, it tends to get pulled back 54:31.646 --> 54:33.316 here; if the drunk is over here, 54:33.320 --> 54:34.810 it tends to get pulled back up. 54:34.809 --> 54:37.559 So, that's the random walk model. 54:37.559 --> 54:41.029 I programmed into this program an AR-1. 54:41.030 --> 54:43.710 I go down here and click on it. 54:43.710 --> 54:47.850 That looks very different, doesn't it? 54:47.849 --> 54:50.239 Maybe I didn't get the parameters exactly right, 54:50.239 --> 54:53.009 but let me just hit F9; I'll get it so it looks better. 54:53.010 --> 54:55.270 Well, it looks a little different; 54:55.269 --> 54:57.349 you can see a difference in this? 54:57.349 --> 55:01.799 That's because I chose ρ = .95 for that 55:01.799 --> 55:07.049 diagram so it's kind of getting tugged back to the trend. 55:07.050 --> 55:10.220 Sometimes it's hard to tell. 55:10.219 --> 55:13.819 I thought--can you tell the difference between a random walk 55:13.816 --> 55:17.206 and an AR-1? It seems to be coming back 55:17.212 --> 55:20.442 faster, so that looks different. 55:20.440 --> 55:26.450 Well, with ρ = .95 I guess it does look noticeably 55:26.445 --> 55:31.265 different because it's coming back pretty fast. 55:31.269 --> 55:34.429 I have 130 years of observations and it's coming 55:34.425 --> 55:38.315 back 5% of the way every year, so in five years it's 25% of 55:38.318 --> 55:41.488 the way back. I've got the cord too tight so 55:41.488 --> 55:43.448 it's not fooling you, right? 55:43.449 --> 55:47.339 You can see a difference between this AR-1 and the true 55:47.339 --> 55:50.189 stock market. The stock--so this is revealing 55:50.189 --> 55:52.929 some truth to the efficient markets hypothesis. 55:52.929 --> 55:55.439 What if I made ρ = .99? 55:55.440 --> 55:58.650 Then I think--I should have done that instead of .95; 55:58.650 --> 56:00.790 then it gets harder for you to see the difference. 56:00.789 --> 56:03.969 If I make ρ = 1 in this--what happens if I plug 56:03.971 --> 56:06.221 ρ = 1 into this expression? 56:06.219 --> 56:09.679 It comes back to the random walk because the x-bars 56:09.678 --> 56:13.138 drop out--because you've got 1*x-bar and you've got 56:13.137 --> 56:16.777 minus x-bar over here so they drop out of the equation 56:16.778 --> 56:18.718 and you come back up here. 56:18.719 --> 56:24.079 A random walk is just an AR-1 with a coefficient--of 56:24.078 --> 56:26.178 ρof 1. 56:26.179 --> 56:32.619 Now, the point is then that we can easily see that the stock 56:32.619 --> 56:36.879 market is not strongly mean-reverting or 56:36.876 --> 56:40.606 trend-reverting. It's somewhat--it may be 56:40.607 --> 56:44.617 somewhat trend-revering; it's very hard to tell whether 56:44.619 --> 56:47.099 its ρ is 1 or .99. 56:47.099 --> 56:51.679 So, that's the point of my little analysis here, 56:51.676 --> 56:54.496 of random walk versus AR-1. 56:54.500 --> 57:00.610 Now, the homework--the problem set that you will have next 57:00.605 --> 57:06.595 time--now you still haven't turned in your second problem 57:06.603 --> 57:09.333 set, but I want to work ahead now to 57:09.333 --> 57:12.523 the third one. What I want you to do is to try 57:12.516 --> 57:14.716 forecasting the stock market. 57:14.719 --> 57:19.129 So, let me show you what I did here and I want you to do 57:19.128 --> 57:21.968 something. You could use this spreadsheet 57:21.970 --> 57:25.230 but I'm also encouraging you to find your own data. 57:25.230 --> 57:31.820 Here's my spreadsheet, which has stock prices back to 57:31.822 --> 57:34.532 1871. Can you see all this? 57:34.530 --> 57:37.450 I have this on my website all the time; 57:37.449 --> 57:41.779 I've had it on my website how long? 57:41.780 --> 57:43.970 It must be twenty years. 57:43.969 --> 57:52.779 I give it away because I'm the person who updates S&P data 57:52.783 --> 57:56.543 for 1871 to the present. 57:56.539 --> 57:59.289 I now have a relationship with S&P because they're 57:59.286 --> 58:01.926 publishing our indices, but they still don't provide 58:01.928 --> 58:04.258 updates of this hundred year-long series. 58:04.260 --> 58:08.930 So, here it is all the way back to 1871. 58:08.930 --> 58:10.270 What do I have here? 58:10.269 --> 58:14.079 I have in column one is the Standard & 58:14.076 --> 58:19.366 Poor composite stock price index--today called the S&P 58:19.367 --> 58:23.037 500. This is the dividend on the 58:23.039 --> 58:26.129 S&P 500. They've been paying a dividend 58:26.127 --> 58:28.477 consistently; it has never missed a year. 58:28.480 --> 58:30.860 Maybe some companies have missed a year, 58:30.856 --> 58:33.896 like Berkshire Hathaway or Microsoft, but the whole 58:33.904 --> 58:36.224 aggregate has never missed a year. 58:36.219 --> 58:39.159 This is the earnings report on the company. 58:39.159 --> 58:44.589 It's all per share so this is per share earnings. 58:44.590 --> 58:49.640 I have here the Consumer Price Index in this column and then 58:49.635 --> 58:54.595 I've got long-term interest rates and I can convert it into 58:54.595 --> 58:59.465 a real price by dividing by the Consumer Price Index. 58:59.469 --> 59:02.739 This is the change in the real price. 59:02.739 --> 59:06.469 I don't--it's between this year and the next year. 59:06.469 --> 59:09.769 What I wanted to do is--this is your homework assignment and 59:09.773 --> 59:13.303 maybe I'll come back to this or maybe your T.A.'s will come back 59:13.302 --> 59:16.332 to it as well. I want you to try to forecast 59:16.333 --> 59:17.593 the stock market. 59:17.590 --> 59:22.280 What that means--remember what you're trying to forecast. 59:22.280 --> 59:25.830 You're trying to forecast e; 59:25.829 --> 59:28.529 you're not trying to forecast x. 59:28.530 --> 59:30.480 We already know x, it's easy to forecast; 59:30.480 --> 59:32.910 it's close to what it was last period. 59:32.909 --> 59:36.229 The hard thing is to predict where the next change will be. 59:36.230 --> 59:41.250 So, I had to generate a column of data here, 59:41.252 --> 59:45.342 which is the change in the price. 59:45.340 --> 59:53.930 You see, up here it says J10 minus J9 is the change in the 59:53.933 --> 59:56.483 price; that's what you want to try to 59:56.482 --> 1:00:00.282 forecast. I've created this column going 1:00:00.277 --> 1:00:06.407 all the way back to 1871, showing for each year how much 1:00:06.413 --> 1:00:11.103 the S&P composite index, in real terms, 1:00:11.099 --> 1:00:13.769 changed. I did it in logs. 1:00:13.769 --> 1:00:15.949 I took the change in the logs of real price. 1:00:15.949 --> 1:00:18.229 That's essentially the percentage change. 1:00:18.230 --> 1:00:20.780 That is going to be hard to forecast. 1:00:20.780 --> 1:00:23.940 Obviously, it's going to be hard to forecast because there's 1:00:23.937 --> 1:00:26.557 some truth to the efficient markets hypothesis. 1:00:26.559 --> 1:00:29.829 If that were easy to forecast, you could have been rich over 1:00:29.833 --> 1:00:32.333 this period--somebody could have been rich. 1:00:32.330 --> 1:00:33.530 It can't be that easy. 1:00:33.530 --> 1:00:36.570 Nonetheless, the problem set for you is to 1:00:36.572 --> 1:00:39.542 try to do that--to try to forecast it. 1:00:39.539 --> 1:00:43.469 What you have to do is go to regression analysis. 1:00:43.469 --> 1:00:46.639 You don't have to use Excel but that's just what I'm suggesting 1:00:46.636 --> 1:00:48.676 here because that's the easiest thing. 1:00:48.679 --> 1:00:56.989 To run a regression, you go up to Tools and--let me 1:00:56.993 --> 1:01:05.313 see--Data Analysis is down here--and then you go to 1:01:05.306 --> 1:01:10.836 Regression. I guess I have to say "okay." 1:01:10.840 --> 1:01:16.010 Then it asks you to fill in your "x range" and your 1:01:16.014 --> 1:01:17.744 "y range." 1:01:17.739 --> 1:01:26.259 The x-variable is the--remember the regression 1:01:26.257 --> 1:01:36.087 model is--it's y = α +βx + some error 1:01:36.085 --> 1:01:39.495 term, e or u--I'll call 1:01:39.501 --> 1:01:41.631 it u so we don't get it mixed up. 1:01:41.630 --> 1:01:45.830 It's asking you to say where your y-variable is and 1:01:45.829 --> 1:01:49.809 I've got that that my y variable is in column K, 1:01:49.806 --> 1:01:52.676 because that's what I've generated. 1:01:52.680 --> 1:01:56.470 That's the change--where is K? 1:01:56.470 --> 1:02:02.390 I've lost it; it's not column K, 1:02:02.391 --> 1:02:11.741 it's--I'm trying to find it here--it's column I, 1:02:11.742 --> 1:02:17.622 isn't it? That's the change in price. 1:02:17.619 --> 1:02:20.359 Then the x-variable--I can pick some variable to 1:02:20.361 --> 1:02:22.831 forecast. What I did is I just took 1:02:22.834 --> 1:02:27.204 column A, which is time--it's just the year--and I ran the 1:02:27.198 --> 1:02:31.178 regression since 1950 and I got the results here. 1:02:31.179 --> 1:02:39.439 I'm just going to show you the results from what regression I 1:02:39.443 --> 1:02:42.633 ran. That's how Excel prints out 1:02:42.625 --> 1:02:44.365 regression results. 1:02:44.369 --> 1:02:53.129 The α, which is the intercept, is shown here--so the 1:02:53.131 --> 1:02:56.261 intercept was .05. 1:02:56.260 --> 1:03:06.240 The βis the coefficient of the variable is -2 times 1:03:06.244 --> 1:03:10.364 10^(-5). You can see now I've struck out; 1:03:10.360 --> 1:03:14.820 it doesn't like time as a forecaster--it's a very small 1:03:14.822 --> 1:03:18.122 coefficient. The t-statistic is a 1:03:18.118 --> 1:03:21.438 measure of statistical significance and the 1:03:21.444 --> 1:03:25.484 t-statistic should probably be over two for a 1:03:25.483 --> 1:03:27.863 statistical significance. 1:03:27.860 --> 1:03:31.790 I've got insignificant t-statistics for both 1:03:31.792 --> 1:03:33.682 α and for β. 1:03:33.679 --> 1:03:38.779 The R^(2) is a measure of what fraction of the dependent 1:03:38.779 --> 1:03:43.229 variable y's variance I have predicted. 1:03:43.230 --> 1:03:51.280 It comes out with a fraction of .000147--this is a not a 1:03:51.276 --> 1:03:54.916 success. This is not a get rich quick 1:03:54.923 --> 1:03:59.373 story because I'm explaining 1/10,000th of the variance of 1:03:59.369 --> 1:04:03.129 the stock market; so I struck out. 1:04:03.130 --> 1:04:07.280 I wasn't really trying that hard, I just regressed the 1:04:07.280 --> 1:04:11.500 returns on time. The problem set for you is to 1:04:11.495 --> 1:04:16.425 think about seeing if you can forecast the market. 1:04:16.429 --> 1:04:21.219 You don't have to use my data set, you can use others and you 1:04:21.223 --> 1:04:23.623 can go online and find them. 1:04:23.619 --> 1:04:26.489 For example, finance.yahoo.com has a lot of 1:04:26.489 --> 1:04:29.769 indices, but you can find them on other sites. 1:04:29.769 --> 1:04:35.099 If you find some other data, you can try to see if you can 1:04:35.104 --> 1:04:37.744 beat me. I'm not setting up a very high 1:04:37.736 --> 1:04:38.976 standard for success. 1:04:38.980 --> 1:04:42.360 If it were this bad, efficient markets--well, 1:04:42.359 --> 1:04:45.969 this looks good for efficient markets theory. 1:04:45.970 --> 1:04:58.510 1:04:58.510 --> 1:05:01.220 What I thought you might do for the problem set is to try to 1:05:01.215 --> 1:05:03.595 think creatively about how to forecast the market. 1:05:03.599 --> 1:05:06.889 Let me say this: I'm not--I'm going to have to 1:05:06.892 --> 1:05:11.062 continue with my doubts about efficient markets in another 1:05:11.062 --> 1:05:14.892 lecture, but I think that the story has 1:05:14.890 --> 1:05:17.940 a good element of truth to it. 1:05:17.940 --> 1:05:22.250 It has to be hard to beat the market but it's oversold as 1:05:22.253 --> 1:05:27.033 well, especially in academic circles--I think us professors, 1:05:27.030 --> 1:05:32.140 who are poor are--it's kind of a wishful thinking bias. 1:05:32.139 --> 1:05:37.319 We're not making much money and you can't anyway and so we kind 1:05:37.322 --> 1:05:41.672 of tend to overstate efficient markets hypothesis. 1:05:41.670 --> 1:05:44.500 On the other hand, relative to your expectations 1:05:47.984 --> 1:05:50.574 overstate efficient markets hypothesis. 1:05:50.570 --> 1:05:54.210 It seems to be a lifecycle effect where young people think, 1:05:54.211 --> 1:05:58.231 I can surely predict the market and then they get beaten down. 1:05:58.230 --> 1:06:02.640 Brad Barber and Terry O'Dean were professors at UC--at 1:06:02.640 --> 1:06:06.970 different campuses of California--teamed up with some 1:06:06.968 --> 1:06:11.378 economists from Taiwan and looked at data of--they got 1:06:11.378 --> 1:06:15.958 really good data from Taiwan about day traders and their 1:06:15.955 --> 1:06:19.295 actual returns. Day traders are people who 1:06:19.302 --> 1:06:22.612 trade everyday in the markets and they found that there was a 1:06:22.607 --> 1:06:24.257 really predictable pattern. 1:06:24.260 --> 1:06:28.840 The young people--they start in as a day trader and they quickly 1:06:28.841 --> 1:06:31.531 lose everything; they lose badly because they're 1:06:31.527 --> 1:06:34.147 trading too much and they really can't predict the market. 1:06:34.150 --> 1:06:37.500 There's like 1% of them, though, who seem like they can 1:06:37.503 --> 1:06:39.183 actually beat the market. 1:06:39.179 --> 1:06:41.769 This looks like really good for efficient markets. 1:06:41.769 --> 1:06:48.009 They found that there are some Taiwanese people who know how to 1:06:48.010 --> 1:06:52.440 beat the market--1% survives and stays in. 1:06:52.440 --> 1:06:54.470 Is that contrary to efficient markets? 1:06:54.469 --> 1:06:57.849 Well, it does seem contrary because they found that a small 1:06:57.854 --> 1:07:01.594 number of people did find some forecasting rule and succeeded. 1:07:01.590 --> 1:07:04.850 On the other hand, none of the--hardly any of them 1:07:04.851 --> 1:07:08.711 got really rich and so it's very rare--Warren Buffett is an 1:07:08.712 --> 1:07:10.512 extremely rare outcome. 1:07:10.510 --> 1:07:13.540 So it's--I guess, when I talk about efficient 1:07:13.536 --> 1:07:17.526 markets I want to help prevent you from suffering under any 1:07:17.525 --> 1:07:20.615 delusions about your forecasting ability. 1:07:20.619 --> 1:07:23.759 I don't mean that Warren Buffett can't do it or that you 1:07:23.757 --> 1:07:27.177 can't do it if you develop yourself into a Warren Buffett. 1:07:27.179 --> 1:07:30.409 The next problem set, number three, 1:07:30.412 --> 1:07:36.122 asks you to think creatively about how you would forecast the 1:07:36.116 --> 1:07:41.816 market and to take a stab at it by running a regression. 1:07:41.820 --> 1:07:45.930 I expect you all to fail--or almost all of you to fail. 1:07:45.929 --> 1:07:48.829 Your grade will not depend on your success in forecasting; 1:07:48.829 --> 1:07:52.149 it may even depend inversely because if you show a big 1:07:52.145 --> 1:07:55.765 success in forecasting, your teaching assistant will 1:07:55.767 --> 1:07:59.947 look at it very carefully and try to find some mistake you 1:07:59.954 --> 1:08:02.294 made because, if you do succeed, 1:08:02.293 --> 1:08:03.913 it's probably a mistake. 1:08:03.909 --> 1:08:06.639 On the other hand, I don't want to tell you that 1:08:06.644 --> 1:08:09.444 you can't because, as I said--we're going to come 1:08:09.436 --> 1:08:12.686 back to this--I have doubts about efficient markets and I 1:08:12.694 --> 1:08:16.014 think that you just might be able to do it if you're smart 1:08:16.010 --> 1:08:17.000 about it.