WEBVTT 00:02.380 --> 00:05.650 Professor Robert Shiller: I wanted to talk today 00:05.647 --> 00:07.887 about insurance, which is another risk 00:07.885 --> 00:10.965 management device that's traditionally separate from 00:10.971 --> 00:13.631 securities, which we talked about last 00:13.631 --> 00:16.771 time, but the underlying principles are the same. 00:16.770 --> 00:21.280 Before I begin, I want to just give some more 00:21.282 --> 00:25.492 thoughts about the diversification through 00:25.487 --> 00:30.817 securities and that will lead us into insurance. 00:30.820 --> 00:36.790 Let me just review the preceding lecture briefly for 00:36.793 --> 00:40.313 that purpose. What we did--the core 00:40.306 --> 00:44.906 theoretical framework that we had--was the mean variance 00:44.905 --> 00:49.835 theory, which led us to the capital asset pricing model. 00:49.840 --> 00:54.910 But the basic thing was that we had to--in order to use the 00:54.907 --> 00:59.707 framework--we had to start by producing estimates of the 00:59.712 --> 01:04.592 expected returns on each asset, we called those r, 01:04.591 --> 01:09.041 and the standard deviation of the return on each asset and the 01:09.036 --> 01:13.186 covariance between the returns of each pair of assets. 01:13.189 --> 01:16.699 Then, once we did that we could plug that into the formula that 01:16.696 --> 01:19.916 I gave you last time and get the standard deviation of the 01:19.919 --> 01:23.029 portfolio and the expected return on the portfolio. 01:23.030 --> 01:25.920 From then on, if you accept the analysis and 01:25.924 --> 01:29.224 the assumptions or the estimates that underlie it, 01:29.223 --> 01:33.063 then we pretty much know how to construct portfolios. 01:33.060 --> 01:39.160 The underlying estimates may not accord with your belief or 01:39.158 --> 01:43.258 your intuitive sense of common sense. 01:43.260 --> 01:47.100 The other thing that I mentioned last time was that 01:47.100 --> 01:51.790 there seems to be a really big difference between the expected 01:51.786 --> 01:55.776 return on the stock market and the expected return on 01:55.780 --> 01:59.890 short-term debt. We found an equity premium--or 01:59.885 --> 02:04.635 actually Jeremy Siegel's book gave an equity premium of 4% a 02:04.635 --> 02:07.385 year. Some people find that hard to 02:07.386 --> 02:10.416 believe. How can it be that one asset 02:10.415 --> 02:13.465 does 4% a year better than another? 02:13.469 --> 02:15.979 Some people say, well if that's the case I want 02:15.975 --> 02:18.585 to invest in nothing more than that one asset. 02:18.590 --> 02:22.250 Why should I take something that is underperforming? 02:22.250 --> 02:26.250 Jeremy Siegel goes on further to say that since the 02:26.250 --> 02:30.890 mid-nineteenth century we've never had a thirty-year period 02:30.889 --> 02:33.929 when stocks under performed bonds, 02:33.930 --> 02:38.910 so stocks are really--if anyone who has an investment horizon of 02:38.914 --> 02:43.064 thirty years--you'd think, why should I ever holds bonds. 02:43.060 --> 02:49.780 The numbers that Jeremy Siegel produces seem implausibly high 02:49.782 --> 02:52.362 for the stock market. 02:52.360 --> 02:58.200 What we call this is the--I want to emphasize it, 02:58.202 --> 03:04.412 I'll write this again--the equity premium puzzle. 03:04.409 --> 03:11.689 That term was actually coined by economists, 03:11.693 --> 03:19.933 Prescott and Mehra; it's now in general use. 03:19.930 --> 03:23.810 That is, it just seems that stocks so much outperform other 03:23.812 --> 03:26.562 investments. For Jeremy Siegel, 03:26.555 --> 03:32.565 in the latest edition of his book, the equity premium is 4% a 03:32.574 --> 03:38.104 year since 1802. That's almost--no that's more 03:38.097 --> 03:41.587 than 100--that's 206 years. 03:41.590 --> 03:44.140 Why would that be and can you believe that? 03:44.139 --> 03:50.629 One question that comes up is that maybe--this is for the U.S. 03:50.629 --> 03:53.999 data--and some people say, well, maybe, 03:53.996 --> 03:56.916 why are we looking at the U.S.? 03:56.920 --> 04:01.860 Because the U.S. is an arguably very successful 04:01.860 --> 04:05.550 country, so we have, potentially, 04:05.550 --> 04:10.510 a bias in--it's called a selection bias. 04:10.509 --> 04:16.139 If you pick as the country you study one of the most successful 04:16.139 --> 04:22.749 countries in the world, that doesn't inform you very 04:22.745 --> 04:31.875 well about what it is for a random country or for the U.S. 04:31.879 --> 04:33.949 going forward, there's something wrong. 04:33.950 --> 04:35.950 The U.S. has been successful in 04:35.948 --> 04:39.288 financial markets and it's being imitated by lots of countries. 04:39.290 --> 04:43.380 Financial markets similar to ours are being set up in many 04:43.375 --> 04:45.255 places. You wonder, you know, 04:45.259 --> 04:49.009 maybe they're over imitating; maybe we were just lucky or 04:49.014 --> 04:51.224 maybe it was because the U.S. 04:51.220 --> 04:54.050 was the first, in some ways, 04:54.052 --> 04:59.722 to develop some of these financial institutions--or one 04:59.716 --> 05:02.946 of the first. But now, when more and more 05:02.952 --> 05:05.912 countries start doing it, maybe it won't work so well. 05:05.910 --> 05:10.940 One way of investigating this is--to get around the selection 05:10.943 --> 05:14.553 bias--is to try to look at all countries. 05:14.550 --> 05:16.300 Let's not just look at the United States; 05:16.300 --> 05:21.480 let's look at every country of the world and let's see if they 05:21.477 --> 05:23.597 have an equity premium. 05:23.600 --> 05:29.090 There's a problem with that and the problem is that countries 05:29.094 --> 05:34.684 that are less successful don't keep data--that's a problem. 05:34.680 --> 05:38.180 Or they--sometimes they just shut down their stock markets at 05:38.181 --> 05:42.441 some point. This is since 1802--now how 05:42.439 --> 05:50.609 many countries do you think have uninterrupted stock market data 05:50.605 --> 05:53.505 since 1802? What do you think? 05:53.509 --> 05:58.989 Name another country that probably has it. 05:58.990 --> 06:00.690 What's that? England, UK? 06:00.689 --> 06:07.429 If you go onto the continent, though, they tended to be 06:07.429 --> 06:13.169 interrupted by World War I and World War II. 06:13.170 --> 06:15.050 What about Japan, do they have--do you think they 06:15.048 --> 06:15.868 have uninterrupted? 06:15.870 --> 06:20.540 They had a little bit of a problem around World War II and 06:20.538 --> 06:24.758 you can try to bridge the gap, but--anyway, 06:24.762 --> 06:31.442 there are people who have tried to sort this out. 06:31.439 --> 06:35.159 There's one, it's a book by Dimson, 06:35.164 --> 06:38.784 Marsh, & Staunton that--called 06:38.779 --> 06:43.489 Triumph of the Optimists--that Jeremy 06:43.489 --> 06:47.709 Siegel quotes. He has a table in the new, 06:47.707 --> 06:49.887 fourth edition of his book. 06:49.889 --> 06:57.059 Dimson, Marsh and Staunton look at the following countries: 06:57.057 --> 07:00.267 Belgium, Italy, Germany, 07:00.269 --> 07:02.209 France, Spain, Japan, Switzerland, 07:02.207 --> 07:03.907 Ireland, Denmark, Netherlands, 07:03.909 --> 07:06.359 UK, Canada, U.S., South Africa, 07:06.359 --> 07:08.259 Australia, and Sweden. 07:08.259 --> 07:11.039 Every one of them has a positive equity premium; 07:11.040 --> 07:12.310 although the U.S. 07:12.310 --> 07:15.690 is on the high side of them all, it's not the best. 07:15.689 --> 07:18.439 The country that has the highest equity premium--and 07:18.439 --> 07:20.649 that's for the whole twentieth century, 07:20.649 --> 07:25.489 they couldn't go back to 1802--the most successful 07:25.487 --> 07:29.927 country is Sweden and after that Australia. 07:29.930 --> 07:34.110 U.S. is not the most successful 07:34.106 --> 07:40.506 stock market although it's high on the list. 07:40.509 --> 07:45.469 Jeremy Siegel concludes that there's--that the equity--he 07:45.466 --> 07:49.356 said that these--that this book by Dimson, 07:49.360 --> 07:54.440 Marsh and Staunton puts to bed any concerns about selection 07:54.435 --> 07:59.765 bias and he claims that so many countries have shown an equity 07:59.772 --> 08:03.012 premium that we can be confident. 08:03.009 --> 08:06.059 His book is really very strong on the conclusions. 08:06.060 --> 08:08.360 The title of the book, Stocks for the Long 08:08.360 --> 08:11.180 Run--stocks always outperform other investments for 08:11.183 --> 08:14.323 the long run and he says it's not due to selection bias. 08:14.319 --> 08:17.559 You know, I kind of wonder, the list of countries that I 08:17.563 --> 08:19.453 just read to you, that Dimson, 08:19.449 --> 08:22.059 Marsh and Staunton studied, excludes some important 08:22.058 --> 08:23.308 countries, doesn't it? 08:23.310 --> 08:25.430 Who does it exclude? 08:25.430 --> 08:29.420 Well, it doesn't have India, Russia, and China in it, 08:29.416 --> 08:32.506 for example. At least Russia and China--do 08:32.511 --> 08:35.231 you know anything about their history? 08:35.230 --> 08:38.560 They have any stock market disruptions in the last one 08:38.563 --> 08:41.153 hundred years? That's kind of obvious. 08:41.149 --> 08:44.959 They had a communist revolution in both places, 08:44.957 --> 08:48.717 right? Russia and China are not 08:48.720 --> 08:56.470 mentioned by--or not studied by Dimson, Marsh and Staunton. 08:56.470 --> 08:58.050 Why not? Well, they can't get data, 08:58.047 --> 08:59.247 there wasn't a stock market. 08:59.250 --> 09:03.750 Well there actually was a stock market in Russia before 1918 and 09:03.751 --> 09:07.611 in China before 1949, so what happened to investors? 09:07.610 --> 09:12.280 If you were a Chinese investor in Chinese stocks in 1949, 09:12.284 --> 09:15.864 what happened? We know what happened. 09:15.860 --> 09:18.900 It went--that's that famous minus 100% return, 09:18.898 --> 09:21.328 right, which dominates everything. 09:21.330 --> 09:26.890 I think--what would Siegel say? 09:26.889 --> 09:31.039 He's really saying that this equity premium is enduring and 09:31.043 --> 09:32.693 we should believe it. 09:32.690 --> 09:36.260 I don't know, I think that--I think what 09:36.264 --> 09:40.574 Jeremy would say is, well you're looking--if you 09:40.572 --> 09:44.772 look at Russia and China, you're looking at political 09:44.771 --> 09:48.311 factors and I'm only looking at politically stable countries, 09:48.309 --> 09:51.329 so this whole thing is irrelevant. 09:51.330 --> 09:54.480 Really, we're not going to have a communist revolution in any of 09:54.480 --> 09:56.030 these advanced countries now. 09:56.029 --> 09:58.409 So Jeremy would say, forget that, 09:58.409 --> 10:02.939 it looks pretty sound that we have an equity premium so we can 10:02.944 --> 10:06.244 trust that. Well, he's a good friend of 10:06.236 --> 10:10.826 mine, but I think he may be overstating it a little bit; 10:10.830 --> 10:12.550 we have some disagreements. 10:12.549 --> 10:17.839 The thing that comes to my mind is that--I want to say before 10:17.840 --> 10:23.040 concluding this review of the last lecture--that is that the 10:23.042 --> 10:27.982 stock market is inherently political in any country. 10:27.980 --> 10:34.150 Politics have tremendous effects on the values in the 10:34.150 --> 10:40.910 stock market and that's because of--even if the government 10:40.913 --> 10:48.273 doesn't nationalize the stock market or confiscate assets, 10:48.270 --> 10:52.130 they tax them. Do you know we have, 10:52.127 --> 10:55.427 in the U.S., a corporate profits tax? 10:55.430 --> 11:03.850 11:03.850 --> 11:07.190 Well, it's not just in the U.S., essentially every--I don't 11:07.194 --> 11:09.044 know if there's any exception. 11:09.039 --> 11:14.079 There may not be an exception, but essentially every country 11:14.079 --> 11:19.119 has a corporate profits tax and then we also have a personal 11:19.118 --> 11:20.398 income tax. 11:20.400 --> 11:25.380 11:25.379 --> 11:29.789 The corporate profits tax goes after the profits that 11:29.785 --> 11:31.475 corporations make. 11:31.480 --> 11:35.730 The personal--it's taken from corporations before they pay out 11:35.726 --> 11:39.056 their dividends. The personal income tax is 11:39.059 --> 11:43.859 levied on individuals and these individuals have to pay it. 11:43.860 --> 11:47.380 The personal income tax is not simple; 11:47.379 --> 11:50.229 it's not just a flat rate on your income, it depends on the 11:50.233 --> 11:52.973 type of income. Dividend income or capital 11:52.973 --> 11:57.003 gains income in the stock market is taxed differently. 11:57.000 --> 12:00.340 The interesting thing is that through time, 12:00.342 --> 12:04.642 as political winds change, these taxes have changed and 12:04.640 --> 12:09.260 they've gone up to some very high levels in the past in the 12:09.256 --> 12:12.436 United States and other countries. 12:12.440 --> 12:15.210 I'm going to give some U.S. 12:15.210 --> 12:18.550 tax rates. The personal tax on 12:18.550 --> 12:26.250 dividends--of course it depends also on your tax bracket and 12:26.251 --> 12:29.291 your income; I'm going to talk about the 12:29.289 --> 12:30.129 highest tax bracket. 12:30.129 --> 12:35.889 In the U.S., it went over 90% in World War 12:35.887 --> 12:40.097 II and the succeeding years. 12:40.100 --> 12:45.640 The government was taking 90% of your dividend income. 12:45.640 --> 12:47.720 What is it today? 12:47.720 --> 12:48.510 Does anyone know? 12:48.509 --> 12:50.619 What's the tax rate of dividends today? 12:50.620 --> 12:56.820 It might be zero for some people, but it's actually--it 12:56.815 --> 13:03.925 is--the standard rate for people who have not negligible income 13:03.928 --> 13:09.398 is 15%. It's gone down from over 90% to 13:09.400 --> 13:11.270 15%. Why did it do that? 13:11.269 --> 13:15.879 Well, it's some kind of political change and the 13:15.876 --> 13:20.966 corporate–incidentally, at the beginning of the 13:20.973 --> 13:24.603 twentieth century you were right. 13:24.600 --> 13:28.270 Who said zero? We didn't even have income tax 13:28.269 --> 13:32.409 until 1913 when the Supreme Court allowed it, 13:32.409 --> 13:35.339 so it was zero, then it went up to 90%--or 13:35.341 --> 13:39.561 actually it was 94% at the peak--and it came down to 15%. 13:39.559 --> 13:42.079 That's a pretty big hit on the stock market. 13:42.080 --> 13:45.260 So, it wasn't just China that took the stock market. 13:45.259 --> 13:48.959 When we were taking 90% of dividends that was 90% of the 13:48.961 --> 13:51.991 stock market being taken by the government; 13:51.990 --> 13:54.270 but that's not all because we were also taxing the 13:54.271 --> 13:57.161 corporations. In the early post-war period, 13:57.163 --> 14:01.553 the corporate--now I'm going to talk--there's a distinction 14:01.551 --> 14:05.641 between the rate that they charge and the actual amount 14:05.636 --> 14:09.656 that they take. Most advanced countries of the 14:09.655 --> 14:14.155 world today have a corporate profits tax rate for large 14:14.164 --> 14:16.924 corporations of about a third. 14:16.919 --> 14:19.879 So they--a typical advanced country takes a third of the 14:19.879 --> 14:22.839 profits, the government takes a third of the profits. 14:22.840 --> 14:26.320 That's not the actual amount that they pay because the tax 14:26.315 --> 14:29.725 law is so complicated and there are so many loopholes. 14:29.730 --> 14:32.940 What I looked at--and I have this on the website, 14:32.939 --> 14:36.549 I have a chart showing corporate profits taxes paid, 14:36.549 --> 14:39.939 as a fraction of corporate profits for the United States 14:39.938 --> 14:43.048 since 1929. That has moved around a lot, 14:43.049 --> 14:48.159 but it got almost up to 60% in the post-World War II period and 14:48.164 --> 14:51.304 now it's down to less than a third. 14:51.300 --> 14:55.690 Why is it down? It's because they're changing 14:55.693 --> 14:59.013 enforcement of the taxes and changing amounts of loopholes, 14:59.009 --> 15:02.899 so most countries have a tax rate of about a third, 15:02.897 --> 15:07.017 but corporations are paying less than a third of their 15:07.017 --> 15:08.647 profits to taxes. 15:08.649 --> 15:12.779 If we want to look going forward at the equity premium, 15:12.784 --> 15:16.464 we have to know how much--what's the politics? 15:16.460 --> 15:18.520 And what's the government going to do in the future? 15:18.519 --> 15:21.709 They've moved these tax rates around a lot, 15:21.707 --> 15:26.187 so I think that it's very hard to be sure that we know going 15:26.185 --> 15:29.825 forward what are σ, the expected return, 15:29.825 --> 15:33.915 the standard deviation of returns and the covariance of 15:33.917 --> 15:35.657 returns--are really. 15:35.659 --> 15:38.969 We have a nice theoretical framework, but the application 15:38.974 --> 15:42.534 of the framework to the real data is hard and it ends up with 15:42.525 --> 15:47.515 politics underneath it all; that's just the real world. 15:47.519 --> 15:54.899 I want to say one more thing about the diversification and 15:54.899 --> 15:59.139 the mutual fund. Ideally, mutual funds are 15:59.139 --> 16:03.569 calculating r and σ and σ_12 and 16:03.570 --> 16:08.000 plugging in and finding what the optimal portfolio that you 16:08.002 --> 16:11.392 should hold is, then offering that to you in 16:11.388 --> 16:12.648 their mutual fund. 16:12.649 --> 16:15.819 Firms that do that are, in practice, 16:15.818 --> 16:21.248 however, the minority and most mutual funds have some gimmick 16:21.251 --> 16:26.231 or some special--they claim to be beating the market not 16:26.230 --> 16:29.490 forming the optimal portfolio. 16:29.490 --> 16:35.570 I also have up on the website some questions that I asked 16:35.569 --> 16:41.539 investors about what they think about picking stocks. 16:41.539 --> 16:44.089 Picking stocks means trying to find a stock that's going to do 16:44.085 --> 16:47.595 really well. What I found--the question I 16:47.597 --> 16:53.657 asked is, do you agree--which of the following is a correct 16:53.660 --> 16:59.410 answer to this statement: Trying to time the market, 16:59.409 --> 17:03.559 to get out before it goes down and in before it goes up is (a) 17:03.562 --> 17:08.772 a smart thing to do, or (b) not a smart thing to do? 17:08.769 --> 17:12.359 Most people think that it's not a smart thing to do--to try to 17:12.360 --> 17:16.420 time the market. Only 11% said yes to that. 17:16.420 --> 17:20.480 But then I asked another question, do you think trying to 17:20.480 --> 17:24.010 pick mutual funds, trying to find a mutual fund 17:24.009 --> 17:28.619 that will beat the market is a smart thing to do or not a smart 17:28.615 --> 17:31.765 thing to do? Most people think it's a smart 17:31.767 --> 17:34.527 thing to do. What I think the mutual fund 17:34.532 --> 17:38.032 industry has turned into, largely, is a stock picking 17:38.029 --> 17:41.659 industry, not a portfolio diversification industry. 17:41.660 --> 17:45.290 What most people are doing when they go into a mutual fund is 17:45.291 --> 17:48.441 they're trying to find smart people who will beat the 17:48.439 --> 17:51.889 market--who will pick those stocks that will do well. 17:51.890 --> 17:55.420 The mutual fund theory that we gave last time said, 17:55.416 --> 17:59.436 no the mutual fund is just supposed to be diversifying for 17:59.435 --> 18:02.015 you. In fact, the truth is somewhere 18:02.018 --> 18:06.258 in between. Most mutual funds are providing 18:06.260 --> 18:13.290 some diversification service and they're also trying to beat the 18:13.293 --> 18:16.633 market. Finally, I just want to say 18:16.629 --> 18:20.249 that I've been talking mainly about the U.S., 18:20.247 --> 18:25.017 but mutual funds have been growing in importance around the 18:25.015 --> 18:30.725 world. There was a recent paper by 18:30.730 --> 18:39.280 Khorana, Servaes, and Tufano, that looked at what 18:39.281 --> 18:48.371 it is that explains which countries have had rapidly 18:48.366 --> 18:54.776 growing mutual fund industries. 18:54.779 --> 18:58.769 They found, not surprisingly, that it tends to--mutual funds 18:58.770 --> 19:02.900 have been growing more rapidly in countries that have stronger 19:02.897 --> 19:05.397 securities laws and institutions, 19:05.400 --> 19:10.340 especially laws that protect individual shareholders rights. 19:10.339 --> 19:14.139 Also, mutual funds have been growing more in countries that 19:14.138 --> 19:18.198 have higher level of education and a higher level of wealth. 19:18.200 --> 19:22.210 They also grow more in countries that have 19:22.206 --> 19:27.876 institutional structures that encourage investing in mutual 19:27.875 --> 19:31.095 funds, such as pension plans. 19:31.099 --> 19:34.619 I think it's a trend around the world that we're going to see 19:34.619 --> 19:38.079 more and more mutual funds and I think it's a good thing. 19:38.079 --> 19:40.879 I think they will help us to diversify our risks. 19:40.880 --> 19:47.530 Anyway, I want then to move on to the topic of today's lecture, 19:47.528 --> 19:49.778 which is insurance. 19:49.779 --> 19:57.739 Insurance is the other side of the risk management institutions 19:57.735 --> 20:02.865 that we have. Insurance evolves separately 20:02.870 --> 20:06.440 from securities. It's long been a different 20:06.444 --> 20:09.134 industry, but the principles are the same. 20:09.130 --> 20:13.030 The principles of insurance--the fundamental 20:13.028 --> 20:17.288 powerhouse--is the principle of risk pooling. 20:17.289 --> 20:21.159 Insurers, just like mutual funds, are providing risk 20:21.163 --> 20:24.953 pooling for you. Risk pooling means they put a 20:24.948 --> 20:29.868 lot of people with independent or low-correlated risk into a 20:29.869 --> 20:33.789 pool and reduce the risk for the whole pool. 20:33.789 --> 20:39.509 They have to contend with something called moral hazard, 20:39.512 --> 20:45.442 which is the risk that people will be affected by the fact 20:45.443 --> 20:50.233 that they're insured and do something bad. 20:50.230 --> 20:53.950 The classic moral hazard problem is the problem that you 20:53.948 --> 20:58.138 give fire insurance on a house and someone burns down the house 20:58.140 --> 21:00.710 in order to collect the insurance. 21:00.710 --> 21:07.420 They also have to deal with selection bias. 21:07.420 --> 21:11.690 What this means in the insurance context is that if you 21:11.687 --> 21:16.347 offer insurance policies you will tend to attract people who 21:16.349 --> 21:19.699 are higher risk. If you offer life insurance, 21:19.700 --> 21:23.490 you have life tables which give you probabilities of dying at 21:23.493 --> 21:25.963 various ages, but that's for the general 21:25.958 --> 21:28.818 population. All the sick people will come 21:28.824 --> 21:32.664 to you to buy life insurance and they will turn out to have a 21:32.662 --> 21:35.862 higher death rate than the population at large. 21:35.859 --> 21:40.329 These are the problems of insurance that we have to deal 21:40.330 --> 21:44.010 with. I just want to review the 21:44.012 --> 21:47.252 mathematics of insurance. 21:47.250 --> 21:51.800 This is actually just, in part, just a review of what 21:51.799 --> 21:55.299 we talked about in the second lecture. 21:55.299 --> 22:03.749 In the ideal world, if you have independent risks, 22:03.745 --> 22:09.945 under the independence assumption, 22:09.950 --> 22:15.400 the probability distribution for the number of insurance 22:15.399 --> 22:21.149 contracts that you will have to pay on follows the binomial 22:21.145 --> 22:26.605 distribution. x is the number of 22:26.613 --> 22:35.093 accidents--let's say this is some accident insurance--the 22:35.089 --> 22:44.319 probability of having--n is the number of policies that 22:44.322 --> 22:49.932 you're writing. If you're going to have 22:49.934 --> 22:56.684 p as the probability of an accident, then the binomial 22:56.678 --> 23:02.968 distribution gives you the probability of having x 23:02.972 --> 23:08.032 accidents out of your n policies. 23:08.029 --> 23:19.779 We had that before, that is p^(x)(1- 23:19.775 --> 23:32.915 p)^((n-x)) n!/(x! 23:32.920 --> 23:41.350 (n - x)!) That is the binomial distribution and it 23:41.353 --> 23:48.723 allows you to calculate the probability of any number of 23:48.715 --> 23:58.395 accidents. The mean proportion--The mean 23:58.398 --> 24:08.358 of x/n, is equal to p. 24:08.359 --> 24:14.169 If x is the number of accidents, the mean number of 24:14.168 --> 24:20.688 accidents divided by your number of policies is given just by the 24:20.689 --> 24:27.039 probability, but the standard deviation of 24:27.041 --> 24:36.491 x/n is equal to the square root of p(1 24:36.485 --> 24:41.365 – p)/n. 24:41.369 --> 24:54.919 That is the--that gives you the mean and standard deviation of 24:54.922 --> 25:01.822 the proportion. To actually apply this it helps 25:01.821 --> 25:07.461 to go to something called the normal approximation to the 25:07.461 --> 25:14.031 binomial, because it's kind of difficult 25:14.031 --> 25:18.501 to compute this formula. 25:18.500 --> 25:24.030 There's an easier formula and you assume that the binomial 25:24.031 --> 25:28.881 distribution is really a normal distribution with a 25:28.883 --> 25:33.473 mean–I'm sorry, the proportion of accidents 25:33.467 --> 25:37.937 x/n follows a normal distribution with mean 25:37.943 --> 25:41.863 p and standard deviation given by this. 25:41.859 --> 25:48.779 That's the whole theory that I have here; 25:48.780 --> 25:52.770 it's simple. Maybe I should make a 25:52.766 --> 26:01.166 bigger--let me do it here, I can fit it in here I think. 26:01.170 --> 26:04.310 I'm going to draw an example. 26:04.309 --> 26:06.529 I've got it plotted out and it's on the website, 26:06.530 --> 26:08.090 but it's a very simple example. 26:08.089 --> 26:17.989 I have the case where p = .2, so the probability of an 26:17.990 --> 26:22.750 accident is 20%. This is a significantly high 26:22.748 --> 26:24.408 probability of accidents. 26:24.410 --> 26:27.620 Can you see this? 26:27.620 --> 26:31.270 Let's do this from 0 to .4. 26:31.269 --> 26:37.179 If you wrote only one policy, what's the probability 26:37.178 --> 26:41.348 distribution of x/n? 26:41.349 --> 26:43.639 Well, it has two possible values. 26:43.640 --> 26:45.680 It could be the one person doesn't have the accident or 26:45.683 --> 26:46.633 does have the accident. 26:46.630 --> 26:59.330 So, if n = 1 we have an 80% chance of no accident 26:59.330 --> 27:10.880 and--let's make this 1 not .4--then a 20% chance of 27:10.876 --> 27:16.876 x/n = 1. 27:16.880 --> 27:22.200 I'm plotting--this is the probability of various values of 27:22.195 --> 27:24.055 x/n. 27:24.059 --> 27:27.429 If n = 1, x/n can take on 27:27.431 --> 27:29.381 only two values: 0 or 1. 27:29.380 --> 27:33.340 It takes on the value 1 with the probability of 20% and the 27:33.340 --> 27:36.140 value of 0 with the probability of 80%. 27:36.140 --> 27:39.650 I didn't use the normal approximation there--that's 27:39.651 --> 27:42.321 obvious--I used the binomial itself. 27:42.319 --> 27:47.259 Let's go to the case where n = 100. 27:47.260 --> 27:56.760 You can see this okay? 27:56.759 --> 28:01.689 Then, if n = 100--now I'm going to label this x 28:01.688 --> 28:06.288 differently, I'm now going to show the normal bell-shaped 28:06.287 --> 28:10.227 curve and I'm going to do this from 0 to .4. 28:10.230 --> 28:12.800 Maybe I do have to make this bigger--can you see this all 28:12.798 --> 28:13.668 right back there? 28:13.670 --> 28:23.590 0 to .4, so .2 is here in the middle; 28:23.590 --> 28:29.370 that's .2. For n = --now I'm going 28:29.365 --> 28:31.895 to do n = 100. 28:31.900 --> 28:35.380 What is the mean? 28:35.380 --> 28:39.200 Well, the mean is always .2 no matter how many policies you 28:39.203 --> 28:42.833 write, so it's going to be--we're going to have a normal 28:42.830 --> 28:44.940 distribution centered on .2. 28:44.940 --> 28:47.990 This .2 is not very readable here. 28:47.990 --> 28:57.930 28:57.930 --> 28:59.450 What is the standard deviation? 28:59.450 --> 29:05.060 It's the square root of (.2 x .8)/100. 29:05.060 --> 29:17.550 So that's .16/100; so the standard deviation is 29:17.549 --> 29:22.919 .04. What does the curve look like? 29:22.920 --> 29:28.520 It's a bell-shaped curve that looks about like that. 29:28.520 --> 29:31.890 29:31.890 --> 29:37.130 I didn't draw that very well, let's do it again. 29:37.130 --> 29:39.160 I should be able to do nice bell-shaped curves, 29:39.158 --> 29:41.318 but it's harder than it looks standing up here; 29:41.319 --> 29:47.429 so that's your bell-shaped curve. 29:47.430 --> 29:51.010 With 100 policies, they can't really count that 29:51.007 --> 29:54.737 accurately on having 20% of the policies paid. 29:54.740 --> 29:59.590 There's still substantial insurance risk because the--it 29:59.588 --> 30:04.698 could easily be only 15% or it could be 25% of the policies 30:04.702 --> 30:06.732 that end up paying. 30:06.730 --> 30:10.430 So, the insurance company with one hundred policies would have 30:10.425 --> 30:11.935 still substantial risk. 30:11.940 --> 30:14.610 It's much better--the uncertainty about 30:14.611 --> 30:19.041 x/n is much lower for the case n = 100 than 30:19.039 --> 30:21.639 it was for the case n = 1, 30:21.640 --> 30:22.680 but it's still there. 30:22.680 --> 30:31.670 Now, I want to draw--what if we write 10,000 policies? 30:31.670 --> 30:35.160 What is the probability distribution for 30:35.163 --> 30:38.123 x/n in that case? 30:38.119 --> 30:43.229 Well, you can see you're going to be dividing not by a hundred 30:43.229 --> 30:48.419 here, but by a hundred times a hundred, so it's going to reduce 30:48.422 --> 30:50.602 this from .04 to .004. 30:50.599 --> 30:54.399 So the normal--the bell-shaped curve in that case for the 30:54.396 --> 30:58.256 x/n is going to look something like this. 30:58.260 --> 31:02.190 I've got it plotted here; it's even more steep than that. 31:02.190 --> 31:09.240 That's a bell-shaped curve, but it's centered exactly on 31:09.239 --> 31:11.839 .2. Number of policies doesn't 31:11.844 --> 31:16.154 affect the means but it affects that standard deviation, 31:16.150 --> 31:20.190 so it becomes very collapsed and this is the basic core idea 31:20.192 --> 31:23.272 of insurance. You have to be a big company to 31:23.266 --> 31:26.896 do it and if you have a big company you've exhausted the 31:26.895 --> 31:29.465 risk of--never goes away completely. 31:29.470 --> 31:36.010 If we did a--if we go another two decimals--if we did a 31:36.007 --> 31:41.087 million policies, then we would--this would 31:41.091 --> 31:46.661 almost just be a spike here at that point, 31:46.660 --> 31:48.780 so that's the concept of insurance. 31:48.779 --> 31:54.809 The idea really took root--the idea, the intuitive idea is that 31:54.807 --> 31:58.887 as you write a large number of policies, 31:58.890 --> 32:04.700 the fraction that will result in accidents becomes closer and 32:04.699 --> 32:08.959 closer to the probability of one accident. 32:08.960 --> 32:14.860 That's an idea that struck people intuitively at various 32:14.864 --> 32:19.914 times in history, but they didn't know how to do 32:19.910 --> 32:22.380 these calculations. 32:22.380 --> 32:28.580 Historians of probability have noted that--this basic idea is 32:28.579 --> 32:32.319 actually in Aristotle, the ancient Greek 32:32.323 --> 32:35.593 philosopher-scientist, and I'm going to quote 32:35.593 --> 32:39.663 Aristotle from his book, De Cielo. 32:39.660 --> 32:43.380 He says--he's talking very generally here but I think it's 32:43.376 --> 32:45.916 just--he's really talking about this. 32:45.920 --> 32:50.340 He says, "to succeed in many things or many times is 32:50.337 --> 32:52.767 'difficult.' For instance, 32:52.770 --> 32:57.750 to repeat the same throw of dice 10,000 times would be 32:57.745 --> 33:00.645 impossible; whereas, to make it once or 33:00.651 --> 33:02.421 twice is comparatively easy." 33:02.420 --> 33:04.070 That's Aristotle talking. 33:04.069 --> 33:06.889 It's exactly this theory, but he doesn't have the word 33:06.885 --> 33:09.325 probability, which hadn't been invented yet. 33:09.329 --> 33:12.659 So, he's using intuitive--he says difficult or easy--so he 33:12.663 --> 33:15.883 says it's difficult, meaning the probability is very 33:15.876 --> 33:18.356 low, or easy, meaning that probability is 33:18.362 --> 33:20.282 high. He had the idea but he didn't 33:20.277 --> 33:22.857 have the math. I think that may be the 33:22.856 --> 33:27.026 first-known statement of the binomial distribution. 33:27.029 --> 33:29.849 Well, it doesn't get very precise but it has the intuitive 33:29.845 --> 33:33.475 concept. The idea of using this theory 33:33.479 --> 33:37.589 for insurance, it's--the earliest-known 33:37.592 --> 33:44.412 statement of it is a--was in an anonymous letter written in 1609 33:44.412 --> 33:48.962 to Count Oldenburg; but that's not the person who 33:48.959 --> 33:52.589 said it, that was the person who received the letter. 33:52.589 --> 33:57.099 Anyway, the person wrote--he was talking about fires and was 33:57.098 --> 34:01.608 proposing that people should pay 1% of the value of the home 34:01.607 --> 34:05.887 every year into a fund and then the fund would be used to 34:05.886 --> 34:09.016 replace the home if it burned down. 34:09.019 --> 34:12.379 Quoting this anonymous person writing in 1609, 34:12.376 --> 34:16.176 "There is no doubt that it would be fully proved, 34:16.179 --> 34:20.289 if a calculation were made of the number of houses consumed by 34:20.288 --> 34:24.328 fire within a certain space in the course of thirty years, 34:24.329 --> 34:27.149 that the loss would not amount, by a good deal, 34:27.146 --> 34:30.266 to the sum that would be collected in that time." 34:30.269 --> 34:32.849 It's interesting that this person uses the word 34:32.853 --> 34:35.603 calculation; this person has the idea--this 34:35.596 --> 34:39.246 was--1600 was around the year when probability theory was 34:39.247 --> 34:42.047 invented. So, someone had the idea of 34:42.052 --> 34:46.322 going beyond the intuitive notion that Aristotle mentioned 34:46.323 --> 34:49.773 and moving to something that is calculable. 34:49.769 --> 34:52.339 That's when the insurance industry was really born. 34:52.340 --> 34:59.580 Insurance relies on this theory of risk pooling but it has to 34:59.579 --> 35:04.049 make it work. I stressed in the third lecture 35:04.045 --> 35:08.755 that like any--insurance, like any other risk management 35:08.757 --> 35:11.067 device, is an invention. 35:11.070 --> 35:18.840 Every risk management device relies on a design and the 35:18.843 --> 35:26.333 design is usually complex and has--it all has to work 35:26.328 --> 35:30.288 together. In order for a design to work 35:30.289 --> 35:33.459 well we have to have every component there. 35:33.460 --> 35:36.860 If one component is missing we may have a failure. 35:36.860 --> 35:44.300 All these components have to be compatible with each other and 35:44.303 --> 35:50.893 it has to work according to a plan, which ultimately is 35:50.893 --> 35:54.313 informed by this theory. 35:54.309 --> 36:07.579 Insurance as an invention has to have what things? 36:07.579 --> 36:11.599 It has to have a contract design; 36:11.599 --> 36:17.379 that would be a document, which is a contract between the 36:17.376 --> 36:21.086 insured and the insurance company. 36:21.090 --> 36:24.290 It specifies--what does it specify? 36:24.289 --> 36:30.889 It specifies what risks are covered, exclusions--some risks 36:30.891 --> 36:36.101 are not covered. Those exclusions are carefully 36:36.098 --> 36:41.888 designed in regard to moral hazard and selection bias. 36:41.889 --> 36:46.769 There has to be the mathematical model, 36:46.769 --> 36:54.089 which I just presented, but it may be more complicated. 36:54.090 --> 36:58.070 There has to be a form for the company. 36:58.070 --> 36:59.910 There could be a corporate form. 36:59.909 --> 37:05.119 There's the insurance company, could be a corporation and it 37:05.119 --> 37:09.969 could be either a non-profit corporation or a for-profit 37:09.974 --> 37:13.334 corporation owned by shareholders. 37:13.329 --> 37:17.419 Or the insurance company could be a mutual insurance company, 37:17.416 --> 37:21.636 in which case the insurance company is owned by the insured. 37:21.639 --> 37:26.909 Then you need, as well--you need government 37:26.907 --> 37:34.557 regulation because the insurance companies don't seem to exist 37:34.558 --> 37:38.348 without it. There have to be regulators 37:38.354 --> 37:42.754 that are at least verifying that the insurance company is doing 37:42.750 --> 37:45.090 what it says it's going to do. 37:45.090 --> 37:50.930 One thing that the government has to do is reassure that 37:50.933 --> 37:54.443 reserves' requirements are met. 37:54.440 --> 37:59.900 The company has to have enough money on hand to pay out in the 37:59.899 --> 38:01.509 case of default. 38:01.510 --> 38:07.330 38:07.329 --> 38:18.589 One way of classifying insurance companies is that they 38:18.587 --> 38:31.097 can exist as either multiline insurance companies or monoline 38:31.096 --> 38:36.096 insurance companies. 38:36.099 --> 38:41.919 A multiline insurance company insures many different kinds of 38:41.923 --> 38:46.683 things and doesn't confine itself to one thing. 38:46.679 --> 38:50.439 If a company were just a fire insurance company it would be a 38:50.435 --> 38:52.245 monoline insurance company. 38:52.250 --> 38:56.980 They're essentially more risky and more--regulators have to 38:56.981 --> 39:00.981 watch them more because they're standing at higher 39:00.978 --> 39:05.218 probability--a monoline--of some major disaster. 39:05.219 --> 39:13.619 In contrast, a multiline kind of insures 39:13.621 --> 39:17.761 itself. We hear a lot about monoline 39:17.761 --> 39:21.301 insurance companies in the newspapers today. 39:21.300 --> 39:24.100 I don't know if you were reading about the subprime 39:24.095 --> 39:26.215 crisis, but we are living in a time, 39:26.219 --> 39:30.189 at this moment, of a financial crisis called 39:30.189 --> 39:32.219 the subprime crisis. 39:32.219 --> 39:37.689 That naturally hits monoline insurance companies more than 39:37.688 --> 39:42.488 multiline insurance companies because they are more 39:42.486 --> 39:45.936 specialized and more vulnerable. 39:45.940 --> 39:50.780 I'll come back to this distinction in a minute. 39:50.780 --> 39:59.260 I want to talk about certain kinds of monoline insurance 39:59.260 --> 40:07.900 companies that--the biggest--one category is property and 40:07.895 --> 40:12.065 casualty. These are insurance--or 40:12.074 --> 40:17.654 P&C--these are insurance companies that insure the value 40:17.650 --> 40:21.810 of a home or a business or an automobile. 40:21.809 --> 40:27.499 Another kind of monoline insurance is health insurance 40:27.496 --> 40:33.716 company that merely insures people against health costs, 40:33.719 --> 40:39.419 and another important category is life insurance. 40:39.420 --> 40:48.130 Life insurance is--what it insures is a beneficiary against 40:48.126 --> 40:52.026 the death of an insured. 40:52.030 --> 40:57.630 The classic example--the most important example is families. 40:57.630 --> 41:02.610 You would buy life insurance on both the husband and the 41:02.609 --> 41:06.139 wife--you might do different amounts; 41:06.139 --> 41:09.329 it's done by families with young children to protect them 41:09.333 --> 41:13.043 against the economic cost of the death of one of their parents. 41:13.039 --> 41:19.219 You need insurance on both parents because both parents are 41:19.224 --> 41:24.134 contributing to the success of the children. 41:24.130 --> 41:27.190 These are big industries. 41:27.190 --> 41:38.030 The total assets of property and casualty in 2007 in the 41:38.025 --> 41:48.265 United States--and these are assets--the property and 41:48.270 --> 41:59.300 casualty was $1.4 trillion dollars and life insurance was 41:59.303 --> 42:04.823 $4.9 trillion dollars. 42:04.820 --> 42:09.240 Somehow I've missed getting the health, but that would be 42:09.237 --> 42:11.127 another big insurance. 42:11.130 --> 42:21.040 These are hugely important institutions in our society. 42:21.039 --> 42:24.859 What our--what does--let me go to property and casualty. 42:24.860 --> 42:27.910 What do they insure? 42:27.909 --> 42:33.139 Actually, the most important things that they insure are 42:33.138 --> 42:34.468 automobiles. 42:34.470 --> 42:38.150 42:38.150 --> 42:41.300 The total of the premia collected on automobiles is 42:41.301 --> 42:44.141 about--is much bigger, like five times bigger, 42:44.138 --> 42:47.288 than of the premia they collect on homeowners. 42:47.290 --> 42:52.510 42:52.510 --> 42:55.120 Automobile insurance is collision insurance; 42:55.119 --> 42:58.059 homeowners insurance is–it used to be called 42:58.056 --> 43:01.646 fire insurance but now they've extended it to include so many 43:01.651 --> 43:04.171 other risks. It depends on the particular 43:04.165 --> 43:05.605 policy--what it includes. 43:05.610 --> 43:09.180 It might include risks, as well, of personal liability 43:09.179 --> 43:13.019 if someone injures themselves on your property or risks of 43:13.019 --> 43:14.709 storms, of hurricanes, 43:14.714 --> 43:16.764 earthquakes, everything else; 43:16.760 --> 43:18.420 so we call it homeowners' insurance. 43:18.420 --> 43:22.710 Actually, the automobile insurance is more important than 43:22.713 --> 43:26.543 the homeowners' insurance, even though homes are so much 43:26.541 --> 43:29.151 more valuable, because--I think that's because 43:29.150 --> 43:32.920 cars move and they drive around and they bump into each other. 43:32.920 --> 43:41.280 Homes just stay where they are, so we have far fewer accidents, 43:41.277 --> 43:48.417 so they don't have to charge as high a premium for the 43:48.422 --> 43:53.142 insurance. These insurance contracts have 43:53.143 --> 43:58.623 come across gradually through time as we develop the theory 43:58.622 --> 44:04.382 and--I'll talk about the growth of insurance and about some of 44:04.384 --> 44:06.844 the components of it. 44:06.840 --> 44:24.170 44:24.170 --> 44:28.290 The real insurance industry, as I mentioned before, 44:28.288 --> 44:33.068 began in the 1600s with the invention of probability theory 44:33.065 --> 44:37.835 and with the invention of life tables for--the invention of 44:37.843 --> 44:42.373 actuarial science; but, it grew slowly. 44:42.369 --> 44:46.079 I think the reason that it grew slowly was that insurance is a 44:46.084 --> 44:47.854 very sophisticated concept. 44:47.849 --> 44:51.769 In order to explain it--I had to write down the binomial 44:51.770 --> 44:54.480 distribution to explain it properly. 44:54.480 --> 44:57.040 For most people, that's a difficult concept 44:57.036 --> 45:00.256 and--I think I may have mentioned some of this before, 45:00.261 --> 45:02.271 but let me give this history. 45:02.269 --> 45:09.089 Insurance was invented in the 1600s but it did not proliferate 45:09.087 --> 45:13.667 fast, it proliferated only very slowly. 45:13.670 --> 45:19.090 Some of the important figures--there had to be certain 45:19.092 --> 45:22.062 inventions to make it work. 45:22.059 --> 45:25.689 Again, I'm repeating a theme that is in my book, 45:25.692 --> 45:30.022 New Financial Order, that financial innovation and 45:30.021 --> 45:34.121 insurance innovation are successions of inventions and 45:34.117 --> 45:38.057 each invention propels the idea more forward. 45:38.059 --> 45:42.529 It's like we have laws of thermodynamics that underlie the 45:42.534 --> 45:46.074 use of engines, but you can't just go from the 45:46.066 --> 45:49.516 laws of thermodynamics to an automobile; 45:49.519 --> 45:51.679 there are a million steps along the way. 45:51.679 --> 45:55.489 If you look at the history of engines there are discrete 45:55.489 --> 45:59.509 advances when people were able to apply the theory more and 45:59.506 --> 46:01.256 more. Well, in insurance there's a 46:01.255 --> 46:02.375 similar list of inventions. 46:02.380 --> 46:12.480 Morris Robinson was head of Mutual Life of New York in the 46:12.478 --> 46:22.048 1840s and he got the idea of highly-paid life insurance 46:22.045 --> 46:27.275 salespeople. The idea was that it was 46:27.283 --> 46:31.193 difficult to sell people on insurance. 46:31.190 --> 46:33.590 Back in the 1840s, life insurance was very 46:33.594 --> 46:36.884 important because the average expectancy of life was only 46:36.878 --> 46:40.418 something like forty-five years, so that meant parents were 46:40.420 --> 46:41.630 dying left and right. 46:41.630 --> 46:45.750 What would be the probability that a married couple would 46:45.753 --> 46:49.443 live, both of them, to the time when their children 46:49.435 --> 46:53.825 were grown? Well, it was fairly low, right? 46:53.829 --> 46:58.619 If the average age of death was something like forty-five, 46:58.624 --> 47:03.174 maybe a fifty-fifty chance of--high risk--of one of the 47:03.166 --> 47:07.246 parents dying. You should think that people 47:07.249 --> 47:12.939 would really want life insurance but they were not buying it. 47:12.940 --> 47:14.220 Why was it? First of all, 47:14.220 --> 47:15.720 they were dying in such numbers; 47:15.719 --> 47:18.839 p was so high that it was expensive. 47:18.840 --> 47:22.300 They have to pay--the premium has to cover the costs, 47:22.301 --> 47:26.031 so it was tough to get someone to buy life insurance even 47:26.029 --> 47:30.489 though they really needed it--it was such a good idea for them. 47:30.489 --> 47:35.499 It was partly because there was a psychological resistance to 47:35.502 --> 47:38.262 it. This is still very much alive 47:38.264 --> 47:40.534 today. I was standing at the World 47:40.526 --> 47:44.166 Economic Forum at one of our lunch things and a young woman 47:44.170 --> 47:48.280 approached from Swiss Re, which is the Swiss Reinsurance 47:48.276 --> 47:52.966 Company, and she said she wanted my ideas on how to sell crop 47:52.970 --> 47:54.770 insurance in Africa. 47:54.769 --> 47:57.799 She said, we have it now at Swiss Re and, 47:57.800 --> 48:01.740 of course, The World Bank sponsors crop insurance for 48:01.740 --> 48:05.080 farmers. There are some very poor areas 48:05.075 --> 48:09.675 in Africa where farmers really run the risk--if their crop 48:09.683 --> 48:14.203 fails it could be really bad; they would be approaching 48:14.195 --> 48:17.095 starvation. So, wouldn't you think that 48:17.102 --> 48:21.142 farmers would want to buy crop insurance from this Swiss 48:21.142 --> 48:23.872 company? Looks like a good idea to me, 48:23.867 --> 48:26.637 but she says, we're having a lot of trouble 48:26.644 --> 48:30.244 selling it. She said, if you talk to the 48:30.238 --> 48:35.858 farmers in these rural areas what do you think they say when 48:35.861 --> 48:38.441 you offer? They say, I can't afford the 48:38.435 --> 48:40.525 insurance. Well, they're not thinking 48:40.527 --> 48:43.007 right. The whole idea of insurance is 48:43.014 --> 48:47.084 that you take from your good years and you move it into your 48:47.075 --> 48:50.925 bad years so that you make it through all your years. 48:50.929 --> 48:54.269 So, you're having a good year this year maybe--it looks that 48:54.271 --> 48:56.821 way now--and you think you can't afford it, 48:56.820 --> 48:59.220 but just think how bad it will be if it's a bad year. 48:59.219 --> 49:02.749 Then you really won't be able to afford to stay alive. 49:02.750 --> 49:07.490 But she says, they didn't; some of them respond but a 49:07.487 --> 49:09.327 great majority of them don't. 49:09.329 --> 49:12.659 I think it's because of a psychological aversion that 49:12.660 --> 49:15.350 people have to thinking about insurance; 49:15.350 --> 49:16.510 it's just unpleasant. 49:16.510 --> 49:19.970 Life insurance is actually an insurance against one of you 49:19.965 --> 49:22.155 dying; it's a very unpleasant topic. 49:22.159 --> 49:26.439 If someone comes to your home and says, I would like to sell 49:26.437 --> 49:27.957 you life insurance. 49:27.960 --> 49:31.090 You think, some other day, I don't want to talk to day 49:31.090 --> 49:33.630 about the probability of one of us dying. 49:33.630 --> 49:37.070 So, it was a tough sell. 49:37.070 --> 49:39.910 Morris Robinson, however, realized that some 49:39.909 --> 49:43.609 people are very talented salespeople and they're probably 49:43.606 --> 49:46.046 talented at other things as well. 49:46.050 --> 49:51.840 It may sound like a small improvement but he just had very 49:51.836 --> 49:57.826 highly-paid insurance salesmen and they were paid as long as 49:57.826 --> 50:00.766 people kept their policy. 50:00.769 --> 50:04.209 That motivated insurance salesmen to form long-term 50:04.207 --> 50:08.537 relationships with the families that he was insuring and to keep 50:08.538 --> 50:11.218 them from canceling their policies. 50:11.219 --> 50:13.839 He got talented, respected members of the 50:13.839 --> 50:17.829 community who people admired to become life insurance salesmen 50:17.834 --> 50:21.904 and he had to pay them enough so that they would stick with the 50:21.895 --> 50:24.775 job; then it finally worked. 50:24.780 --> 50:29.060 It may seem crazy but--it may seem like a modest innovation 50:29.062 --> 50:33.052 but it actually was--it was an important innovation. 50:33.050 --> 50:36.240 I don't know how you think of life insurance salespeople, 50:36.239 --> 50:38.859 but they have been pillars of the community. 50:38.860 --> 50:41.450 They are people that you--the community--trusts, 50:41.445 --> 50:44.575 that people are willing to let into their home and discuss 50:44.579 --> 50:49.309 things like death; that was an innovation that 50:49.313 --> 50:56.013 came in then. The other innovation was by 50:56.006 --> 51:05.016 Henry Hyde and he was at another insurance company, 51:05.018 --> 51:14.028 Equitable Life, and that was in the late 1800s. 51:14.030 --> 51:18.220 51:18.219 --> 51:22.249 What he invented was an insurance policy that had a cash 51:22.250 --> 51:26.570 value and that's another--it's a cash value on the insurance 51:26.573 --> 51:30.543 policy. That is, the insurance policy 51:30.542 --> 51:37.412 doesn't just insure you against death, it also builds value over 51:37.412 --> 51:40.722 the years. This invention was very 51:40.719 --> 51:44.989 important because it stopped people from canceling. 51:44.989 --> 51:48.949 The big problem life insurance had was people would buy it and 51:48.945 --> 51:53.025 they'd pay for several years and eventually they would think, 51:53.030 --> 51:55.480 well we didn't die, we're losing all this money, 51:55.484 --> 51:57.264 let's just cancel the insurance. 51:57.260 --> 52:00.810 It especially happens--the way life is in real families is: 52:00.809 --> 52:04.359 you start getting accustomed to a certain style of life; 52:04.360 --> 52:06.310 you start spending more and more money; 52:06.309 --> 52:09.569 there comes a time when you have a little crunch and you're 52:09.569 --> 52:12.529 a little short on money; you're casting your net out for 52:12.529 --> 52:14.319 some way to come up with some money; 52:14.320 --> 52:17.910 and canceling your life insurance was a good idea. 52:17.909 --> 52:23.529 It turns out that if you make them forfeit their cash value on 52:23.525 --> 52:26.835 canceling, then they won't cancel. 52:26.840 --> 52:31.570 Both of these were ideas that were copied all over the 52:31.566 --> 52:36.736 world--that's the way inventions are--so a lot of insurance 52:36.739 --> 52:39.949 policies today have cash values. 52:39.949 --> 52:47.779 The third thing I was going to report was that sociologist 52:47.780 --> 52:55.610 Viviana Zelizer wrote a book about life insurance sales in 52:55.610 --> 52:59.320 the nineteenth century. 52:59.320 --> 53:03.560 She found that there was a lot of resistance to the purchase of 53:03.559 --> 53:06.499 life insurance in the nineteenth century. 53:06.500 --> 53:09.890 She tried to study it and tried to figure out how it was that 53:09.885 --> 53:12.925 life insurance became more and more important over that 53:12.932 --> 53:15.852 century. One of her conclusions was that 53:15.850 --> 53:19.900 life insurance seemed to be opposed quite a bit by women, 53:19.896 --> 53:21.916 nineteenth century women. 53:21.920 --> 53:31.640 53:31.640 --> 53:33.180 Why didn't they like it? 53:33.179 --> 53:36.329 You would think that any rational woman in the nineteenth 53:36.332 --> 53:39.652 century would reflect on the fact that there's a significant 53:39.654 --> 53:42.984 probability that her husband will die of something while you 53:42.976 --> 53:44.436 still have children. 53:44.440 --> 53:47.590 Why wouldn't they want it? 53:47.590 --> 53:50.840 Well, what she found was that life insurance salespeople were 53:50.842 --> 53:54.152 going to families and trying to sell them on life insurance by 53:54.148 --> 53:57.778 explaining the concept and they would say something like I did. 53:57.780 --> 54:01.250 Maybe they didn't write down the binomial distribution, 54:01.252 --> 54:05.052 but they explained the idea of insurance and it didn't sound 54:05.046 --> 54:08.836 right to the typical nineteenth century American woman. 54:08.840 --> 54:11.580 I suppose it wasn't just America, it was a worldwide 54:11.582 --> 54:14.362 problem. The reaction that salespeople 54:14.357 --> 54:17.907 would get from women was, you're giving me some 54:17.905 --> 54:22.225 probabilities or something, you're asking me to--it sounds 54:22.231 --> 54:26.461 like you're asking me to play in some gambling thing where I win 54:26.455 --> 54:31.505 if my husband dies; it doesn't sound right to me. 54:31.510 --> 54:34.160 I think, in fact, a lot of women would say 54:34.164 --> 54:36.884 something like, I put my faith in God and I 54:36.882 --> 54:40.772 think I might bring down God's wrath if I did such a thing as 54:40.766 --> 54:43.676 to bet that my husband is going to die. 54:43.679 --> 54:48.039 So, she would refuse to take part in it; 54:48.039 --> 54:50.969 it doesn't sound right to her--I'll trust in prayer and 54:50.966 --> 54:54.516 other things. It didn't work and you couldn't 54:54.518 --> 54:57.928 sell them on it. What Zelizer reported was that 54:57.932 --> 55:01.442 some life insurance companies surmounted this problem by 55:01.437 --> 55:04.677 changing the pitch, by telling their salespeople, 55:04.677 --> 55:07.657 don't try to explain probability theory to these 55:07.656 --> 55:10.616 people. What you have to do is come 55:10.616 --> 55:12.336 across differently. 55:12.340 --> 55:16.850 The thing that she said they started doing was to pretend, 55:16.851 --> 55:21.121 in a way, that they were missionaries with a gospel and 55:21.124 --> 55:23.424 the gospel was insurance. 55:23.420 --> 55:28.120 They would tell these women, you know, if you buy life 55:28.124 --> 55:32.754 insurance on your husband, then should anything happen 55:32.754 --> 55:36.894 your husband can love and protect you from beyond the 55:36.885 --> 55:40.635 grave. That sounds good--it worked and 55:40.641 --> 55:46.631 these little things--they may seem like little things but they 55:46.626 --> 55:49.566 are technological advances. 55:49.570 --> 55:56.960 People in a profession over the years learn more and more about 55:56.959 --> 56:03.399 how to manage the public's expectations and get them to 56:03.395 --> 56:07.205 actually purchase insurance. 56:07.210 --> 56:12.150 I want to talk about government regulation of insurance because 56:12.148 --> 56:16.208 I said that was an important aspect of insurance. 56:16.210 --> 56:22.190 The regulators do many things, but insurance regulators, 56:22.192 --> 56:26.332 most of all, are concerned with capital 56:26.325 --> 56:27.735 adequacy. 56:27.740 --> 56:34.480 56:34.480 --> 56:37.000 Insurance now, in the United States, 56:36.996 --> 56:41.166 is regulated not by the federal government but by insurance 56:41.167 --> 56:44.257 regulators in each of the fifty states. 56:44.260 --> 56:47.380 It's very hard to summarize insurance regulation in the 56:47.384 --> 56:49.804 United States. Why is it regulated separately 56:49.804 --> 56:50.854 by the fifty states? 56:50.849 --> 56:53.799 That's because it started out that way; 56:53.800 --> 56:57.130 when the federal government wasn't involved in those sorts 56:57.134 --> 57:00.064 of things and somehow we never had the--in fact, 57:00.059 --> 57:02.089 it's talked about--we should--maybe we'll see it in 57:02.090 --> 57:03.350 the next ten or twenty years. 57:03.349 --> 57:06.959 It's talked about that there should be a federal government 57:06.961 --> 57:10.761 regulation of insurance but it's actually state regulators. 57:10.760 --> 57:13.360 It makes it--this is a handicap to the U.S. 57:13.360 --> 57:14.240 insurance industry. 57:14.239 --> 57:17.109 Other countries have it centralized but in the U.S. 57:17.110 --> 57:18.460 it's split up over fifty states. 57:18.460 --> 57:21.170 It's very hard to start an insurance company because you've 57:21.170 --> 57:23.740 got to meet the requirements of all fifty regulators. 57:23.739 --> 57:27.919 The most important thing that these do is: they specify what 57:27.921 --> 57:31.041 reserves insurance companies have to hold. 57:31.039 --> 57:34.949 So they--in other words, the insurance company--doesn't 57:34.954 --> 57:39.164 trust the insurance companies to do the calculations like I 57:39.158 --> 57:41.838 showed with the binomial theorem. 57:41.840 --> 57:45.730 They want to make sure that there's a significantly high 57:45.734 --> 57:49.984 probability--sufficiently high probability--that even if they 57:49.982 --> 57:54.372 get a bad draw and there are a lot policies that require paying 57:54.371 --> 57:56.881 out, that these reserves will 57:56.883 --> 58:00.743 satisfy. An insurance company must hold 58:00.737 --> 58:04.887 the reserves; it can hold more and that's 58:04.891 --> 58:07.861 called the statutory surplus. 58:07.860 --> 58:12.020 The reserves are an accounting entry--it's how much they are 58:12.020 --> 58:15.370 required to hold, but the companies will hold 58:15.372 --> 58:18.732 more than that, typically, in order to protect 58:18.727 --> 58:22.447 themselves--more than is required--and their policy 58:22.454 --> 58:26.484 holders--more than is required by the regulators. 58:26.480 --> 58:32.680 Occasionally, you get into problems with 58:32.681 --> 58:34.431 reserves. 58:34.430 --> 58:38.720 58:38.719 --> 58:44.649 Before I get to that let me just say--I want to just mention 58:44.653 --> 58:49.283 a few types of insurance that are important. 58:49.280 --> 58:53.660 Let's talk about life insurance--and we're talking 58:53.664 --> 58:57.964 about kinds of policies that are around today. 58:57.960 --> 59:02.460 The simplest form of life insurance is called "term 59:02.460 --> 59:06.340 insurance." This is insurance that you pay 59:06.335 --> 59:11.655 each year for insurance in that year and it does not build a 59:11.657 --> 59:15.767 cash value. I can buy insurance for myself 59:15.768 --> 59:21.468 this year and I can stop and if I--there's nothing gained--or 59:21.466 --> 59:25.926 next year I would just have to pay it again. 59:25.929 --> 59:32.319 "Whole life" is more complicated because it builds a 59:32.319 --> 59:38.709 cash value according to a schedule and there is both 59:38.708 --> 59:43.718 non-participating and participating. 59:43.719 --> 59:46.669 With non-participating--with participating, 59:46.666 --> 59:50.446 you are participating in the portfolio outcome that the 59:50.454 --> 59:53.124 insurance company is experiencing, 59:53.119 --> 59:55.989 so you have some uncertainty about your cash value. 59:55.989 --> 1:00:02.209 There's something called "variable life," which refers to 1:00:02.211 --> 1:00:08.211 a life insurance policy where the policyholder can make 1:00:08.210 --> 1:00:14.880 decisions about the investment of the money in the whole life 1:00:14.875 --> 1:00:17.975 policy. It's not just taking what's 1:00:17.976 --> 1:00:19.936 given by the insurance company. 1:00:19.940 --> 1:00:24.000 There's something called "universal life"--these are all 1:00:23.997 --> 1:00:26.207 explained in Fabozzi, et al. 1:00:26.210 --> 1:00:31.230 It's a whole life policy that gives the policyholder 1:00:31.227 --> 1:00:35.357 flexibility over the insurance premiums. 1:00:35.360 --> 1:00:40.010 You can pay more into the cash value in one year and less in 1:00:40.010 --> 1:00:43.240 another year, as long as you keep paying a 1:00:43.242 --> 1:00:48.722 minimum amount. There are also survivorship 1:00:48.721 --> 1:00:53.881 policies that will pay, for example, 1:00:53.881 --> 1:01:02.431 the second to die--there would be a policy if a husband and 1:01:02.433 --> 1:01:06.833 wife get it, then it pays out when the 1:01:06.825 --> 1:01:08.405 second of them dies. 1:01:08.409 --> 1:01:12.749 Also, about regulation, I want to mention the 1:01:12.751 --> 1:01:18.181 NAIC--it's a very important institution--that stands for 1:01:18.178 --> 1:01:23.308 National Association of Insurance Commissioners. 1:01:23.309 --> 1:01:27.049 This is something that's very important in the United States 1:01:27.051 --> 1:01:30.941 because, as I mentioned, insurance regulation is divided 1:01:30.935 --> 1:01:35.045 up over the fifty states and the problem that that creates 1:01:35.045 --> 1:01:38.715 is--it's a nightmare for insurance companies because 1:01:38.723 --> 1:01:42.043 every different state has different laws. 1:01:42.039 --> 1:01:46.539 Recognizing this problem, the insurance commissioners in 1:01:46.537 --> 1:01:51.277 the various states have decided to form an organization, 1:01:51.280 --> 1:01:54.040 that's the NAIC, and they hold regular 1:01:54.036 --> 1:01:57.066 conferences. At these conferences, 1:01:57.066 --> 1:02:02.646 they decide on a recommended or uniform insurance regulation 1:02:02.646 --> 1:02:07.466 policy and it acts a little bit like a Congress. 1:02:07.469 --> 1:02:11.269 They make laws but the laws are not binding on anyone, 1:02:11.271 --> 1:02:14.501 they're just suggested laws or regulations. 1:02:14.500 --> 1:02:19.340 It's an effort to try to get the complexity of fifty 1:02:19.338 --> 1:02:22.278 different regulators uniform. 1:02:22.280 --> 1:02:25.860 As I say, when the NAIC decides on some regulation, 1:02:25.860 --> 1:02:30.010 it's only a recommendation to the separate state regulatory 1:02:30.014 --> 1:02:32.174 commissions. But of course, 1:02:32.168 --> 1:02:36.418 it has a lot of force because all fifty states have met--the 1:02:36.418 --> 1:02:40.738 commissioners of fifty states have met and hashed it out. 1:02:40.739 --> 1:02:44.419 So, most states would essentially adopt what the NAIC 1:02:44.417 --> 1:02:48.377 says, otherwise--they're not going to figure it all--they 1:02:48.378 --> 1:02:52.408 can't claim to beat--it would be rational to do that. 1:02:52.409 --> 1:02:54.889 You wouldn't think that we would rethink all of these 1:02:54.892 --> 1:02:57.712 things ourselves and then have it different in our state. 1:02:57.710 --> 1:03:00.930 The NAIC is kind of a quasi-regulatory body--or a 1:03:00.926 --> 1:03:04.536 quasi-government--it's not government because it has no 1:03:04.544 --> 1:03:08.634 authority and yet it acts almost like a parliament where these 1:03:08.632 --> 1:03:12.652 people get together and decide on things and they end up as a 1:03:12.652 --> 1:03:14.062 force of law. 1:03:14.060 --> 1:03:22.810 1:03:22.809 --> 1:03:27.609 Another important–-there are some milestones in insurance 1:03:27.606 --> 1:03:29.536 that I want to mention. 1:03:29.539 --> 1:03:33.279 Then I'm going to come back to finally concluding with problems 1:03:33.284 --> 1:03:36.674 that we see. Of course, the problems are not 1:03:36.672 --> 1:03:40.302 damning; they're problems that reflect 1:03:40.298 --> 1:03:45.178 the progress yet needed to be made in insurance. 1:03:45.179 --> 1:04:02.139 I just want to mention the Gramm-Leach-Bliley Financial 1:04:02.137 --> 1:04:10.927 Modernization Act of 1999. 1:04:10.929 --> 1:04:16.989 What this did is it allowed banks to offer insurance or to 1:04:16.994 --> 1:04:22.744 ally and join in and merge with insurance companies. 1:04:22.739 --> 1:04:24.659 Before that, as I said before, 1:04:24.661 --> 1:04:28.041 insurance is really the same thing as securities. 1:04:28.039 --> 1:04:31.729 They're both based on risk management and pooling of risks. 1:04:31.730 --> 1:04:35.250 Diversification and pooling are really the same thing but we had 1:04:35.247 --> 1:04:37.087 a separate set of institutions. 1:04:37.090 --> 1:04:40.390 People never thought of insurance as the same as a bank, 1:04:40.389 --> 1:04:43.749 but since this is relatively recently--that was only nine 1:04:43.750 --> 1:04:48.970 years ago--at this time, it means that we are now seeing 1:04:48.967 --> 1:04:54.327 an expansion of our banking system in the U.S. 1:04:54.330 --> 1:04:58.760 to become insurance-related. 1:04:58.760 --> 1:05:00.350 It's different in other countries; 1:05:00.349 --> 1:05:03.889 in Europe, they've had universal banking, 1:05:03.889 --> 1:05:09.199 which allowed banks to offer insurance for a much longer time 1:05:09.198 --> 1:05:13.008 but in the U.S. it's kind of a recent 1:05:13.005 --> 1:05:15.855 innovation. I want to conclude with 1:05:15.856 --> 1:05:20.266 problems, which may sound kind of negative but this is--I don't 1:05:20.269 --> 1:05:22.119 shrink from negativism. 1:05:22.120 --> 1:05:25.190 It's not really negativism; it's talking about what should 1:05:25.190 --> 1:05:28.640 be done. I want to come back to--I told 1:05:28.639 --> 1:05:33.389 you that monoline was in the news a lot lately. 1:05:33.390 --> 1:05:36.540 So, what are we talking about? 1:05:36.539 --> 1:05:44.449 I don't know how much you read about the--right now we are 1:05:44.446 --> 1:05:52.626 going through a major financial crisis, which started in the 1:05:52.630 --> 1:05:56.140 U.S. but has spread over the world 1:05:56.139 --> 1:05:59.449 and this is called the subprime crisis. 1:05:59.450 --> 1:06:05.900 1:06:05.900 --> 1:06:10.310 Subprime refers to mortgages, which is not the subject of 1:06:10.314 --> 1:06:13.784 today's lecture, but a subprime mortgage is a 1:06:13.782 --> 1:06:18.592 mortgage issued to a borrower who is not considered prime--not 1:06:18.591 --> 1:06:21.611 a good risk. They are borrowers that are 1:06:21.610 --> 1:06:25.220 thought, by the various models, to be likely to fail to pay on 1:06:25.224 --> 1:06:28.014 their mortgage and to have to be foreclosed; 1:06:28.010 --> 1:06:31.300 often low income, but also they're people with 1:06:31.303 --> 1:06:33.063 poor credit histories. 1:06:33.059 --> 1:06:37.609 The crisis that we're in now--and this is very important, 1:06:37.609 --> 1:06:42.239 you're living through it and we'll see how it pans out. 1:06:42.239 --> 1:06:46.829 The subprime crisis is happening today because home 1:06:46.834 --> 1:06:51.524 prices are falling and with falling home prices more 1:06:51.521 --> 1:06:56.761 subprime borrowers are failing to pay their mortgage. 1:06:56.760 --> 1:07:01.950 Default rates are shooting up and valuation of securitized 1:07:01.947 --> 1:07:07.407 subprime mortgages have crashed and it's throwing turmoil all 1:07:07.407 --> 1:07:10.407 over the financial community. 1:07:10.409 --> 1:07:17.139 We're seeing what they call systemic effects--that's a very 1:07:17.143 --> 1:07:22.023 general term that goes beyond insurance. 1:07:22.019 --> 1:07:25.099 By systemic effects, I mean something that affects 1:07:25.098 --> 1:07:26.918 the whole financial system. 1:07:26.920 --> 1:07:30.040 When you do these calculations, which I just did for 1:07:30.037 --> 1:07:33.637 insurance--I was talking about insurance companies--assuming 1:07:33.644 --> 1:07:37.564 that risks are independent, everything's independent and 1:07:37.562 --> 1:07:39.732 we've got it all figured out. 1:07:39.730 --> 1:07:42.450 Underlying it there were other things besides just the 1:07:42.453 --> 1:07:45.083 calculation of their accident rates that had certain 1:07:45.075 --> 1:07:46.355 assumptions built in. 1:07:46.360 --> 1:07:50.720 The failure of the assumptions in many different industries can 1:07:50.721 --> 1:07:52.551 create systemic effects. 1:07:52.550 --> 1:07:53.480 So what happened? 1:07:53.480 --> 1:07:56.710 I'm going to talk about a particular line of monoline 1:07:56.708 --> 1:08:00.678 insurers that you may never have even heard of because they don't 1:08:00.681 --> 1:08:02.731 deal with the general public. 1:08:02.730 --> 1:08:12.140 These are municipal bond insurers. 1:08:12.139 --> 1:08:14.509 These are private companies--insurance 1:08:14.512 --> 1:08:17.972 companies--they're monoline because they look only at a 1:08:17.974 --> 1:08:21.954 certain class of risks and not all risks that they insure. 1:08:21.949 --> 1:08:24.359 They deal with, principally, 1:08:24.358 --> 1:08:29.978 state and local governments who are issuing bonds to raise money 1:08:29.979 --> 1:08:33.559 for their activities, like New Haven, 1:08:33.562 --> 1:08:37.082 for example, or any other state or local 1:08:37.077 --> 1:08:40.037 government. We refer to the bonds that they 1:08:40.042 --> 1:08:41.472 issue as municipal bonds. 1:08:41.470 --> 1:08:46.380 Now, the problem is that if you invest in bonds from some state 1:08:46.380 --> 1:08:50.420 or local government, they might not pay you back; 1:08:50.420 --> 1:08:52.890 they can just go bankrupt. 1:08:52.890 --> 1:08:56.780 Cities go bankrupt and they just can't pay, 1:08:56.776 --> 1:09:01.956 so you as a buyer of these bonds feel reluctant to invest 1:09:01.959 --> 1:09:05.619 in them. In order to make their bonds 1:09:05.620 --> 1:09:10.810 saleable to the general public, city and state governments 1:09:10.811 --> 1:09:16.001 go--or any region or local governments go to the municipal 1:09:16.003 --> 1:09:21.383 bond insurer companies and they have names--the big ones are 1:09:21.377 --> 1:09:26.107 MBIA, AMBAC, FGIC--that's not FDIC, 1:09:26.107 --> 1:09:32.627 it's not the Federal Deposit Insurance Corporation, 1:09:32.630 --> 1:09:35.810 it's FGIC--and there are others as well. 1:09:35.810 --> 1:09:40.260 What these do is they insure the bond--they insure the 1:09:40.255 --> 1:09:44.445 investor against the municipality failing to pay on 1:09:44.449 --> 1:09:48.629 the bond. Lot's of investors won't buy a 1:09:48.630 --> 1:09:52.340 municipal bond unless it's insured. 1:09:52.340 --> 1:09:56.710 I'm talking about systemic effects. 1:09:56.710 --> 1:10:00.480 So, these insurance companies have reserves on hand and they 1:10:00.483 --> 1:10:02.853 invest these reserves in something. 1:10:02.850 --> 1:10:04.780 What do they invest them in? 1:10:04.779 --> 1:10:08.649 Well, one of the things they've been investing them in are 1:10:08.653 --> 1:10:10.083 subprime mortgages. 1:10:10.079 --> 1:10:12.709 Now, you can see why I'm talking about systemic effects. 1:10:12.710 --> 1:10:14.990 The housing sector in the United States, 1:10:14.992 --> 1:10:17.452 you might say, is completely different from 1:10:17.451 --> 1:10:20.491 municipal bond insurers; but when the housing sector 1:10:20.488 --> 1:10:22.768 starts going down, people start defaulting on 1:10:22.766 --> 1:10:26.966 their mortgages. Then, the value of the subprime 1:10:26.973 --> 1:10:33.423 loans that the municipal bond insurers own starts to go down. 1:10:33.420 --> 1:10:37.590 Now, these guys have never failed to live up to their 1:10:37.586 --> 1:10:42.306 guarantees--they're doing just fine--but people are noticing 1:10:42.314 --> 1:10:45.604 that their portfolios are going down, 1:10:45.600 --> 1:10:50.410 so their surplus--statutory surplus--is going down, 1:10:50.411 --> 1:10:55.321 so we're starting to worry about these companies. 1:10:55.319 --> 1:11:03.499 Notably, it was January 18, a rating agency, 1:11:03.504 --> 1:11:14.164 Fitch, lowered AMBAC from AAA-rated to AA-rated--that was 1:11:14.162 --> 1:11:18.162 a big news story. 1:11:18.159 --> 1:11:19.749 You might not have caught it, right? 1:11:19.750 --> 1:11:23.090 It was a big news story for the people who work in the municipal 1:11:23.090 --> 1:11:24.680 government. You say, oh oh, 1:11:24.684 --> 1:11:27.404 this is bad news, if this is the only--if only 1:11:27.399 --> 1:11:31.589 one rating agency does that, okay, but--and it's only AMBAC 1:11:31.585 --> 1:11:35.595 that's been down-rated, but you starting wondering. 1:11:35.600 --> 1:11:38.420 So, reporters start calling the other rating agencies--Standard 1:11:38.415 --> 1:11:40.405 & Poor's and Moody's and others--and say, 1:11:40.412 --> 1:11:42.322 are you going to down-rate these guys? 1:11:42.319 --> 1:11:46.379 You read in the newspaper their interpretation of what's going 1:11:46.379 --> 1:11:50.239 to be said and so far the other rating agencies haven't; 1:11:50.240 --> 1:11:52.020 they're not saying anything. 1:11:52.020 --> 1:11:55.350 Now people are suspecting that these are going to get 1:11:55.347 --> 1:11:57.447 down-rated. Well, if they're down-rated, 1:11:57.446 --> 1:11:59.976 then nobody trusts them anymore as insurance companies. 1:11:59.979 --> 1:12:05.019 That means that the municipal governments will find it hard to 1:12:05.018 --> 1:12:08.568 issue bonds to continue their activities, 1:12:08.569 --> 1:12:12.429 so the municipal governments might have to shut down some 1:12:12.429 --> 1:12:14.909 activities, like fixing the roads, 1:12:14.909 --> 1:12:19.099 or bridges, or building schools--whatever the cities do 1:12:19.104 --> 1:12:23.384 with their municipal bond money that they raise from the 1:12:23.376 --> 1:12:26.266 municipal bonds. You can see how things are 1:12:26.268 --> 1:12:29.148 feeding through the system: it starts out in one thing, 1:12:29.149 --> 1:12:32.079 it goes to the insurance companies, and then it goes to 1:12:32.080 --> 1:12:33.600 the municipal governments. 1:12:33.600 --> 1:12:37.090 All that kind of thing might put this economy into a 1:12:37.089 --> 1:12:39.739 recession. It's not just municipal 1:12:39.744 --> 1:12:43.654 governments, but a lot of different aspects of our 1:12:43.653 --> 1:12:48.283 financial system are going to be touched by a crisis that's 1:12:48.279 --> 1:12:51.709 spreading from one segment to another. 1:12:51.710 --> 1:12:56.750 That's where this stands right now. 1:12:56.750 --> 1:13:00.580 The latest thing–well, now this is a crisis--a 1:13:00.580 --> 1:13:03.810 municipal bond crisis which is unfolding. 1:13:03.810 --> 1:13:07.650 The New York state regulators have been trying to get 1:13:07.651 --> 1:13:11.351 companies to subscribe more capital to the monoline 1:13:11.345 --> 1:13:15.845 insurance companies to bolster their capital so that they have 1:13:15.851 --> 1:13:19.661 more money to pay out; that will prevent any more 1:13:19.656 --> 1:13:21.456 lowering of their ratings. 1:13:21.460 --> 1:13:27.980 1:13:27.979 --> 1:13:32.559 The other thing that's happened now is that Warren Buffett said 1:13:32.558 --> 1:13:36.028 he wants to get into the municipal bond insuring 1:13:36.029 --> 1:13:38.339 business. Well, he's just a business 1:13:38.337 --> 1:13:39.267 person coming in. 1:13:39.270 --> 1:13:42.940 So, there will be new municipal bond insurers appearing that 1:13:42.939 --> 1:13:46.609 will take up slack and we'll be all right but the problem is 1:13:46.608 --> 1:13:49.218 that we may have a crisis for a while. 1:13:49.220 --> 1:13:52.400 The other thing that I want to talk about--and I guess I'm 1:13:52.400 --> 1:13:55.860 running out of time--was about management of disaster risks. 1:13:55.859 --> 1:14:00.199 We have another insurance crisis developing now because 1:14:00.200 --> 1:14:03.410 of–well, a very important source of 1:14:03.414 --> 1:14:08.404 insurance risk that's developing has to do with the rising rate 1:14:08.398 --> 1:14:13.218 of hurricane damage that we're observing in the east coast of 1:14:13.220 --> 1:14:15.310 the United States. 1:14:15.310 --> 1:14:20.620 Notably, we saw Hurricane Katrina that caused huge 1:14:20.620 --> 1:14:25.710 property damage in--a couple of years ago--that, 1:14:25.714 --> 1:14:30.704 kind of, tested insurance companies again. 1:14:30.699 --> 1:14:37.719 The risk is that global warming will make hurricanes more common 1:14:37.723 --> 1:14:43.523 and I guess I'll have to--I'll just conclude that. 1:14:43.520 --> 1:14:47.780 I'll talk a little bit about that--I've gotten through almost 1:14:47.783 --> 1:14:49.563 all of this--next time. 1:14:49.560 --> 1:14:53.500 The only other thing--then next period we're going to talk--on 1:14:53.497 --> 1:14:57.427 Friday we're meeting again and we're meeting again about--this 1:14:57.434 --> 1:15:01.114 time it's about efficient markets and I want to talk about 1:15:01.114 --> 1:15:04.794 the evidence for efficient markets and against it and that 1:15:04.793 --> 1:15:08.413 will lead you into your third problem set about efficient 1:15:08.408 --> 1:15:10.728 markets. That's not coming for a little 1:15:10.729 --> 1:15:10.999 while.