WEBVTT 00:01.810 --> 00:05.630 Professor Robert Shiller: This is the second 00:05.626 --> 00:08.676 of two lectures from Lawrence Summers. 00:08.680 --> 00:11.920 Let me just say, again, this is the Okun Lecture 00:11.918 --> 00:16.258 Series created by an anonymous donor in honor of Arthur Okun. 00:16.260 --> 00:19.850 At the last lecture, I was very pleased to hear the 00:19.847 --> 00:23.787 number of fond memories that Larry Summers has of Arthur 00:23.794 --> 00:26.224 Okun. At today's lecture, 00:26.222 --> 00:31.762 I think some people felt that we didn't have enough time for 00:31.755 --> 00:36.345 question and answer, so we plan to allocate a half 00:36.351 --> 00:39.401 hour to that. Does that sound good? 00:39.400 --> 00:46.530 Larry will wrap up just before the hour and we'll have a good 00:46.534 --> 00:48.084 discussion. 00:48.080 --> 01:01.100 01:01.100 --> 01:05.730 Professor Larry Summers: Judging by the discussions I had 01:05.725 --> 01:09.025 an opportunity to have at lunch and dinner, 01:09.030 --> 01:15.770 I would suggest a proposal for you, Bob, with respect to next 01:15.774 --> 01:18.364 year's Okun lecturer. 01:18.360 --> 01:22.150 Invite the Okun lecturer to give his first lecture, 01:22.146 --> 01:26.686 then let him have dinner and let him have lunch with the Yale 01:26.691 --> 01:30.461 faculty; then, let him ponder what he's 01:30.459 --> 01:35.799 been told for three weeks and invite him to come back and 01:35.797 --> 01:41.327 encourage him to emphasize what he was told rather than any 01:41.326 --> 01:43.706 ideas he might have. 01:43.709 --> 01:48.679 You would get substantially better Okun lectures, 01:48.683 --> 01:54.593 I suspect, as a consequence, if the feedback that I got is 01:54.589 --> 01:58.799 any guide. I want to thank everybody for 01:58.795 --> 02:04.305 their hospitality and having seen a certain number of these 02:04.305 --> 02:07.815 kinds of lecture series over time. 02:07.819 --> 02:11.539 When I was President of Harvard, my deepest nightmare, 02:11.539 --> 02:14.279 one of my deepest nightmares--some of my 02:14.276 --> 02:17.856 nightmares actually came true--but some of my lesser 02:17.856 --> 02:21.576 nightmares as President of Harvard were the sequential 02:21.575 --> 02:27.645 lecture series, where at the third lecture no 02:27.651 --> 02:30.321 one came. As a lecturer, 02:30.323 --> 02:35.263 I was concerned the judgment of those who had seen the product 02:35.260 --> 02:40.360 is of more interest than the judgment of those who have not. 02:40.360 --> 02:45.210 So, the fact that the room looks a little bit like the room 02:45.208 --> 02:48.548 looked yesterday, I take as at least mild 02:48.551 --> 02:52.231 encouragement and validation to proceed. 02:52.229 --> 02:59.089 Yesterday, I talked about two types of recessions. 02:59.090 --> 03:03.380 What I referred to as disinflation recessions, 03:03.375 --> 03:07.275 in which the Fed stepped on the brakes, 03:07.280 --> 03:11.060 the economy slowed, the Feds stepped off the 03:11.062 --> 03:14.232 brakes, the economy reaccelerated. 03:14.229 --> 03:19.809 And, what was my larger focus, financial system breakdown 03:19.807 --> 03:24.287 recessions, in which in one way or another, 03:24.289 --> 03:31.799 bubbles burst or banks face liquidity problems or the supply 03:31.798 --> 03:37.778 of finance dried up and the economy suffered. 03:37.780 --> 03:43.260 I suggested that the latter category of recession was the 03:43.261 --> 03:46.591 historical pattern of recession. 03:46.590 --> 03:55.410 That there were some signs that the more relevant problem for 03:55.407 --> 04:04.667 the United States going forward might well be the latter kind of 04:04.666 --> 04:07.896 recession as well. 04:07.900 --> 04:13.790 I implied a thought I will develop in the course of these 04:13.787 --> 04:19.777 remarks--that the incipient recession in the United States 04:19.780 --> 04:25.250 is almost certainly of the latter financial breakdown 04:25.247 --> 04:30.607 variety rather than the disinflation variety. 04:30.610 --> 04:38.160 We have had instructive, for me, conversations at both 04:38.163 --> 04:46.863 dinner and lunch on the question of how one thinks about these 04:46.857 --> 04:55.547 two categories of recession in the context of an ISLM model or 04:55.550 --> 05:04.240 in the context of a simple--some kind of simple--macroeconomic 05:04.244 --> 05:08.444 model. The correct answer to that 05:08.442 --> 05:12.592 question, I think, is that it's pretty unambiguous 05:12.594 --> 05:17.854 that a disinflation recession is best thought of as the Federal 05:17.848 --> 05:23.648 Reserve tightening money, which represents some kind of 05:23.652 --> 05:30.182 leftward movement of the LM curve--that output declines and 05:30.178 --> 05:36.818 interest rates rise and that that's the proximate shock that 05:36.816 --> 05:40.976 is bringing about the recession. 05:40.980 --> 05:46.110 That it turns out to be a matter of semantic 05:46.114 --> 05:53.284 ambiguity--how to think about a financial crisis recession in 05:53.278 --> 05:59.008 the context of a simple macroeconomic model. 05:59.009 --> 06:04.589 Everybody agrees on the movement of three variables in 06:04.589 --> 06:06.589 such a recession. 06:06.590 --> 06:13.020 Output goes down; q goes down; 06:13.019 --> 06:19.129 and short-term interest rates go down. 06:19.129 --> 06:23.239 One instinct, which is probably not the first 06:23.238 --> 06:26.598 instinct in the City of New Haven, 06:26.600 --> 06:31.940 is to think of that recession in a traditional ISLM curve, 06:31.938 --> 06:37.368 with the IS curve and the LM curve being plotted in a space 06:37.370 --> 06:42.150 determined by the level of output and the short term 06:42.146 --> 06:46.586 interest rate. In that formulation, 06:46.590 --> 06:50.060 it is clearly an IS shock. 06:50.060 --> 06:57.090 The level of output declines; the level of interest rates 06:57.085 --> 07:02.775 decline as well. An alternative formulation and 07:02.781 --> 07:08.911 a formulation that is probably more in the spirit of Yale 07:08.906 --> 07:14.916 economics is to think of the ISLM diagram as having been 07:14.921 --> 07:21.701 drawn in a space that focuses on output and the required return 07:21.702 --> 07:27.392 to capital or focuses on output and q. 07:27.389 --> 07:31.609 In a formulation of that kind, there hasn't been a shock to 07:31.612 --> 07:35.332 the relationship between investment and the price of 07:35.326 --> 07:38.326 capital. There's been a shock to the 07:38.331 --> 07:43.151 price of capital at any given level of the short-term interest 07:43.148 --> 07:46.468 rate. And so, one thinks of it as 07:46.467 --> 07:50.247 part of the dynamics of the LM curve. 07:50.250 --> 07:57.230 However one views that question of the IS curve versus the LM 07:57.231 --> 08:04.561 curve, the operative distinction between the two kinds of shocks 08:04.561 --> 08:11.631 is that in a disinflation shock, interest rates are up and that 08:11.625 --> 08:13.825 is why output is down. 08:13.829 --> 08:20.419 In a financial breakdown shock, output is down and that is why 08:20.423 --> 08:26.803 policy interest rates or safe interest rates are reduced. 08:26.800 --> 08:31.280 From that perspective, it's very clear what the 08:31.276 --> 08:33.316 current episode is. 08:33.320 --> 08:37.650 As expectations of future output have declined, 08:37.648 --> 08:44.588 as asset prices have declined, one is seeing very substantial 08:44.591 --> 08:52.981 declines in Treasury bond yields and in the Federal Funds Rate. 08:52.980 --> 08:56.580 For those who are aficionados of such things, 08:56.579 --> 09:00.259 I will mention that for a period of a week, 09:00.259 --> 09:05.009 the five-year TIPS yield was significantly--it was 09:05.012 --> 09:10.442 negative--at one point reaching negative twenty-two basis 09:10.443 --> 09:14.633 points. There was a brief three-hour 09:14.625 --> 09:20.295 period in Japan when the one-month Treasury Bill was 09:20.298 --> 09:26.308 yielding not two basis points, not one basis point, 09:26.314 --> 09:33.114 but literally zero basis points as a consequence of various 09:33.111 --> 09:40.141 machinations in the money market that had the character of an 09:40.143 --> 09:43.663 extreme flight to safety. 09:43.659 --> 09:49.059 If one thinks about the duration and the policy response 09:49.057 --> 09:53.957 to recessions of these two kinds, they are usefully 09:53.963 --> 09:58.243 distinguished. The policy response and the 09:58.242 --> 10:03.172 dynamics of disinflation recessions are not profoundly 10:03.174 --> 10:06.374 complicated. The Fed steps on the brakes; 10:06.370 --> 10:10.220 depending on how hard the Fed steps on the brakes, 10:10.223 --> 10:13.373 the recession's duration is uncertain. 10:13.370 --> 10:16.760 At a certain point, when it is satisfied that 10:16.759 --> 10:20.999 sufficient disinflation has taken place or when it feels 10:20.997 --> 10:24.307 that other imperatives take precedence, 10:24.309 --> 10:27.919 the Fed removes its foot from the brakes, allows short-term 10:27.924 --> 10:32.524 interest rates to decline, and eventually the economy 10:32.515 --> 10:35.195 expands. There's no great mystery about 10:35.202 --> 10:37.542 it. The length and depth of the 10:37.535 --> 10:41.275 recession is closely tied to the strength, force, 10:41.279 --> 10:45.179 and length of the Fed tightening and the speed with 10:45.179 --> 10:48.409 which it eases. Hence, there was much more 10:48.414 --> 10:51.704 disinflation to be done in 1980-82 and the recession 10:51.698 --> 10:54.658 interest rates were pushed up much further; 10:54.659 --> 10:59.309 the recession was that much more severe. 10:59.309 --> 11:10.469 The matter is much more complex when one turns to financial 11:10.473 --> 11:15.673 overextension recessions. 11:15.669 --> 11:19.829 Thinking about the limited sample of data, 11:19.827 --> 11:23.577 it falls into two broad categories. 11:23.580 --> 11:28.520 The first is the recessions that didn't happen or happened 11:28.517 --> 11:31.977 only barely. There was a pretty exciting 11:31.980 --> 11:36.480 financial crisis in 1987 and nothing really happened. 11:36.480 --> 11:40.560 There was a pretty exciting financial crisis for those 11:40.558 --> 11:42.018 involved in 1998. 11:42.019 --> 11:44.839 In neither case, if I gave you the data on 11:44.835 --> 11:46.895 unemployment, GNP, consumption, 11:46.896 --> 11:50.946 any real variable and I didn't give you the time scale and I 11:50.948 --> 11:54.998 said--What year did we have a 20% stock market crash? 11:55.000 --> 11:58.740 What year did we have a really dramatic set of dislocations in 11:58.743 --> 12:01.753 financial markets as a major hedge fund failed? 12:01.750 --> 12:04.210 You wouldn't be able to see it. 12:04.210 --> 12:08.730 Even the 2001 recession, which was associated with the 12:08.728 --> 12:14.098 bursting of the NASDAQ bubble, is a pretty minimal recession. 12:14.100 --> 12:19.210 In retrospect, with the data revised, 12:19.206 --> 12:26.436 there is no two-quarter period when GDP declined. 12:26.440 --> 12:29.820 There are two quarters in which GDP declined, 12:29.820 --> 12:33.810 but they're separated by a quarter of modest positive 12:33.814 --> 12:37.714 growth. It is not the case that all 12:37.707 --> 12:43.007 financial crises, or even most financial crises, 12:43.013 --> 12:46.403 lead to economic downturns. 12:46.399 --> 12:53.179 Some have a natural explanation of that. 12:53.179 --> 12:58.529 Essentially, they explain that the financial 12:58.534 --> 13:06.504 sector is only one small part of the economy--that the decline in 13:06.504 --> 13:13.234 the financial sector of the economy is not of that much 13:13.228 --> 13:17.968 consequence. With reasonable policy 13:17.967 --> 13:24.777 responses, either with monetary policy or with fiscal policy, 13:24.780 --> 13:31.820 and automatic stabilizers any loss in demand can be offset. 13:31.820 --> 13:38.350 Not a typical calculation of this kind would say, 13:38.351 --> 13:45.021 suppose the stock market lost 20% of its value. 13:45.019 --> 13:51.779 If the value of the stock market is on the order of the 13:51.775 --> 13:57.275 value of GNP, that's a loss of 20% of GNP. 13:57.279 --> 14:03.489 If the marginal propensity to consume out of wealth is 4%, 14:03.490 --> 14:09.590 the loss is 8/10% of GNP--an amount that an economy would 14:09.591 --> 14:14.321 rather not lose, but hardly the difference 14:14.323 --> 14:20.253 between economic success and economic failure and something 14:20.247 --> 14:26.577 that's potentially offset if there is some policy response. 14:26.580 --> 14:33.730 There is data and evidence for the view that a significant 14:33.731 --> 14:41.261 amount of financial disruption can take place with relatively 14:41.259 --> 14:46.779 little consequence for the real economy. 14:46.779 --> 14:58.489 The difficulty is that there are other data points as well as 14:58.490 --> 15:03.370 1987 and 1998 and 2001. 15:03.370 --> 15:08.130 There are data points that come from the United States in the 15:08.125 --> 15:10.675 aftermath of 1929, while those, 15:10.684 --> 15:15.324 perhaps, are in so different in institutional environment, 15:15.315 --> 15:19.535 with such egregious errors as to not be relevant. 15:19.539 --> 15:25.869 There are data points from Japan during the 1990s. 15:25.870 --> 15:33.640 There are any number of data points from emerging markets in 15:33.639 --> 15:41.139 which the reality of financial problems created very large 15:41.144 --> 15:43.914 losses in output. 15:43.909 --> 15:48.589 It's natural to ask, why is the response to 15:48.585 --> 15:54.255 financial crises so heterogeneous with some episodes 15:54.261 --> 16:00.721 in which they seem to have very little consequence and some 16:00.717 --> 16:06.837 episodes in which they seem to have so very much? 16:06.840 --> 16:11.420 I think the natural explanation, though it's 16:11.419 --> 16:16.849 difficult to prove in any kind of conclusive way, 16:16.850 --> 16:22.300 is an idea that runs through the work of many economists and 16:22.296 --> 16:28.016 probably in recent decades most strongly through Ben Bernanke's 16:28.019 --> 16:31.119 work, suggesting that financial 16:31.116 --> 16:36.286 intermediation capital is a substantial contributor to the 16:36.285 --> 16:38.275 production process. 16:38.279 --> 16:45.999 When it is destroyed, suddenly there are substantial 16:46.004 --> 16:54.944 losses, both to aggregate supply and to aggregate demand. 16:54.940 --> 16:59.700 I made this point--tried to make this point--graphic in a 16:59.700 --> 17:04.200 conversation with one of your colleagues by asking the 17:04.204 --> 17:07.704 question, suppose one considered the 17:07.704 --> 17:10.564 following macroeconomic shock. 17:10.559 --> 17:16.839 For the next six months, it will not be possible to make 17:16.837 --> 17:22.427 and complete a phone call in the United States. 17:22.430 --> 17:28.180 What would be likely to happen to the GDP of the United States 17:28.183 --> 17:33.753 and what would be likely to happen to the performance of the 17:33.749 --> 17:35.729 American economy? 17:35.730 --> 17:39.260 It's actually sort of an interesting question and I don't 17:39.263 --> 17:42.233 have a clue exactly what the right answer is. 17:42.230 --> 17:48.220 My strong suspicion is that it would really be quite a bad 17:48.220 --> 17:52.740 event for the performance of the economy. 17:52.740 --> 17:58.160 The question of whether the telephone lines would transmit 17:58.159 --> 18:03.669 bits would obviously be very important in--or it would only 18:03.674 --> 18:09.194 stop working for the human voice--would be a very important 18:09.188 --> 18:12.298 determinant. In either case, 18:12.302 --> 18:18.192 it's hard to believe that there would not be substantial loss of 18:18.193 --> 18:23.153 output and significant loss of employment as well. 18:23.150 --> 18:31.730 There would be many methods for seeking to estimate the value of 18:31.729 --> 18:35.069 the losses, but I would argue that a 18:35.069 --> 18:38.859 particularly poor method would be to ask, what's the market 18:38.863 --> 18:41.483 value of all the telephone companies? 18:41.480 --> 18:44.700 Let's say the market value of all the telephone companies is a 18:44.697 --> 18:45.697 trillion dollars. 18:45.700 --> 18:49.130 If they're not going to earn anything for a year, 18:49.129 --> 18:53.629 then their market value will go down by $400 billion dollars. 18:53.630 --> 18:57.990 So, we'll multiply $400 billion dollars by a marginal propensity 18:57.993 --> 19:01.803 to demand--marginal propensity to consume--and we'll add 19:01.803 --> 19:05.963 something for the fact--and we'll call that lost demand. 19:05.960 --> 19:09.660 We'll say, there needs to be some increased demand because we 19:09.661 --> 19:13.001 need to fix the telephones; so, that'll be some increased 19:12.999 --> 19:16.629 investment. That would, it seems to me, 19:16.633 --> 19:20.073 be the wrong mode of analysis. 19:20.069 --> 19:23.959 In some way, one would want to capture the 19:23.961 --> 19:29.091 idea that the production potential of the economy would 19:29.087 --> 19:34.017 be lower because the opportunities to generate output 19:34.022 --> 19:39.622 out of a given amount of capital and a given amount of labor 19:39.622 --> 19:45.132 would be impossible without the intercommunication that the 19:45.127 --> 19:48.447 telephone makes possible. 19:48.450 --> 19:51.590 The results, I would suggest, 19:51.594 --> 19:58.454 of that telephone shock would be lost output and would also be 19:58.445 --> 20:05.515 lost output without substantial disinflation benefit because the 20:05.520 --> 20:12.370 lost output would not come in a way that created lots of goods 20:12.371 --> 20:19.111 hanging over markets that led to declines in prices. 20:19.109 --> 20:23.069 The benefit, in terms of disinflation, 20:23.069 --> 20:29.279 of a downturn induced by the inability to make a phone call 20:29.275 --> 20:35.155 would be much less than the benefit of a similar loss of 20:35.161 --> 20:41.581 output achieved through the more standard tools of fiscal and 20:41.581 --> 20:46.671 monetary policy. I emphasize this last point 20:46.669 --> 20:52.079 because it connects with the observation with which I began 20:52.081 --> 20:57.591 yesterday--that declines in output associated with financial 20:57.587 --> 21:00.927 crisis recession, if avoidable, 21:00.930 --> 21:06.390 probably create a welfare gain of the kind that Art Okun 21:06.388 --> 21:12.238 focused so much on and of the kind that Jim Tobin emphasized 21:12.244 --> 21:17.804 when he spoke of all those Harberger triangles filling an 21:17.802 --> 21:20.372 Okun gap. And [they] 21:20.373 --> 21:25.883 are not susceptible to the critique that what you lose on 21:25.883 --> 21:31.393 the swings you gain on the swerves that's associated with 21:31.392 --> 21:37.002 the more conventional view of cyclical fluctuations. 21:37.000 --> 21:43.860 It should be clear, from what I have said, 21:43.860 --> 21:47.040 where I am going. 21:47.039 --> 21:52.449 My suggestion is really, to echo Ben Bernanke's 21:52.454 --> 21:59.404 suggestion, that the destruction of financial intermediation 21:59.400 --> 22:06.820 capital should be thought of as reducing the economy's potential 22:06.816 --> 22:12.186 to generate output and, in that process, 22:12.192 --> 22:18.942 making people poorer and reducing the level of demand as 22:18.938 --> 22:23.688 well. When that process kicks in to a 22:23.689 --> 22:28.669 substantial extent--that, I would suggest, 22:28.674 --> 22:35.004 is the occasion on which financial crises prove to be 22:34.997 --> 22:41.807 exciting events outside of the precincts of Manhattan and 22:41.806 --> 22:47.006 Chicago. The crucial question for 22:47.011 --> 22:53.351 economic policy in responding to these crises, 22:53.352 --> 22:59.182 I would suggest, is preventing that kind of 22:59.175 --> 23:06.315 destruction of intermediation capital from taking place on a 23:06.315 --> 23:12.695 large scale. That brings me to the policy 23:12.700 --> 23:15.640 task that the U.S. 23:15.638 --> 23:22.328 authorities face at the current moment. 23:22.329 --> 23:32.729 I'd like to define the problem as managing and containing three 23:32.730 --> 23:38.770 separate vicious cycle mechanisms. 23:38.769 --> 23:43.159 The first is the one that I spoke about at length 23:43.159 --> 23:48.189 yesterday--what one might call the liquidation cycle. 23:48.190 --> 23:51.370 The tendency, more than a tendency, 23:51.370 --> 23:55.580 in the market for fixed-income instruments, 23:55.579 --> 24:02.459 particularly mortgage-backed securities, to some degree in 24:02.461 --> 24:09.221 the market for housing for declines in value to give rise 24:09.222 --> 24:16.102 to selling pressures that create further declines in value 24:16.103 --> 24:20.453 reinforcing the vicious cycle. 24:20.450 --> 24:26.020 The second vicious cycle, which builds on that, 24:26.017 --> 24:31.097 is what is traditionally referred to as the 24:31.100 --> 24:34.610 credit-accelerator cycle. 24:34.609 --> 24:40.349 A deteriorating financial economy leads to less lending; 24:40.349 --> 24:45.439 leads to a deteriorating real economy; 24:45.440 --> 24:51.900 leads to less capacity to pay back debt; 24:51.900 --> 24:58.200 leads to declining values of financial instruments; 24:58.200 --> 25:06.150 leads to further reductions in lending and the cycle continues. 25:06.150 --> 25:10.840 The third cycle--the third potentially destabilizing 25:10.841 --> 25:15.721 mechanism is the one that's perhaps most familiar--the 25:15.717 --> 25:20.587 Keynesian mechanism in which reduced spending leads to 25:20.592 --> 25:25.562 reduced income leads to reduced spending and on. 25:25.559 --> 25:31.079 Containing these mutually reinforcing cycles is, 25:31.082 --> 25:36.372 I would suggest now, the central challenge for 25:36.369 --> 25:41.799 economic policy. I am not in a position to make 25:41.803 --> 25:48.283 authoritative estimates of some of the relevant magnitudes and I 25:48.277 --> 25:54.027 don't want my estimates to be given more seriousness than 25:54.032 --> 25:59.812 their crudity deserves, which is why I'm not projecting 25:59.813 --> 26:01.503 them on a screen. 26:01.500 --> 26:07.610 Let me give you some round numbers that are quite close to 26:07.609 --> 26:13.609 the judgment of the various investment banks and informed 26:13.612 --> 26:18.652 observers who have studied these questions. 26:18.650 --> 26:26.540 The equity capital of leveraged financial institutions in the 26:26.537 --> 26:33.107 United States totals about two trillion dollars. 26:33.109 --> 26:35.349 It totaled a little more than two trillion dollars a few 26:35.350 --> 26:38.110 months ago; it probably totals a little 26:38.114 --> 26:41.264 less than two trillion dollars today. 26:41.259 --> 26:49.499 That two trillion dollars of capital supported about twenty 26:49.497 --> 26:54.607 trillion dollars of asset holding. 26:54.609 --> 26:59.389 In other words, those institutions were levered 26:59.389 --> 27:04.289 ten-to-one; some were levered much more 27:04.292 --> 27:09.682 than ten-to-one. Fannie Mae or Freddie Mac are, 27:09.680 --> 27:13.460 for example, levered thirty-to-one. 27:13.460 --> 27:20.270 A bank with an 8% capital requirement is levered 27:20.271 --> 27:26.001 twelve-to-one; others--hedge funds pursuing 27:25.997 --> 27:32.447 certain strategies--are levered considerably less than 27:32.446 --> 27:37.266 ten-to-one. On average, the leverage works 27:37.269 --> 27:42.879 out to about ten-to-one--two trillion dollars of equity 27:42.883 --> 27:47.773 capital, twenty trillion dollars of lending. 27:47.769 --> 27:54.659 The losses that have taken place over the last nine months 27:54.657 --> 28:01.057 and the projected losses based on reasonable models of 28:01.062 --> 28:07.712 fundamentals--these are not marked-to-market losses that 28:07.708 --> 28:13.748 build in what might be market overreactions. 28:13.750 --> 28:17.280 These are estimates of what will happen to mortgage 28:17.278 --> 28:20.028 securities based on assumptions that, 28:20.029 --> 28:23.339 for example, house prices will decline by 28:23.343 --> 28:26.163 another 15% from current levels. 28:26.160 --> 28:35.020 Estimates of those total losses are about a trillion dollars. 28:35.019 --> 28:38.569 That trillion-dollar figure sounds very scary, 28:38.567 --> 28:42.427 relative to two trillion dollars of capital in the 28:42.430 --> 28:44.480 intermediation sector. 28:44.480 --> 28:51.170 But, good news--about half those losses are outside the 28:51.173 --> 28:55.143 financial intermediary sector. 28:55.140 --> 29:00.590 They're Bill Brainard's pension; they're the State of Nebraska's 29:00.592 --> 29:03.502 pension fund; they're the Yale endowment; 29:03.500 --> 29:08.860 they're life insurance companies, and the like. 29:08.859 --> 29:20.649 So, about half are held by levered intermediaries. 29:20.650 --> 29:25.920 And more good news--the levered intermediaries are mostly, 29:25.922 --> 29:29.442 sooner or later, in one way or another, 29:29.437 --> 29:33.967 going to get tax deductions of various kinds. 29:33.970 --> 29:41.670 When all is said and done, the losses of capital to the 29:41.666 --> 29:50.356 intermediation sector represent about $300 billion dollars. 29:50.359 --> 29:57.179 $300 billion dollars of lost capital has been offset so far 29:57.176 --> 30:04.576 by about $150 billion dollars of capital that has been raised in 30:04.581 --> 30:10.341 Citigroup's deal with Abu Dhabi and the like. 30:10.339 --> 30:15.319 Lost capital, which can be estimated with 30:15.318 --> 30:22.288 some degree of accuracy compared to the next step in this 30:22.288 --> 30:29.238 calculation, represents about $150 billion 30:29.240 --> 30:34.010 dollars. $150 billion dollars of lost 30:34.008 --> 30:40.578 capital at ten-to-one leverage means a trillion and a half 30:40.579 --> 30:47.379 dollars of lost intermediation capacity or about 7.5% of the 30:47.381 --> 30:54.531 financial asset holding that was previously taking place. 30:54.530 --> 31:01.120 Now, here comes a wild card; what should happen to the 31:01.119 --> 31:07.939 leverage of intermediaries in the face of the events of the 31:07.938 --> 31:10.288 last nine months? 31:10.289 --> 31:14.419 Well, first of all, they probably should have 31:14.421 --> 31:19.211 learned from the last nine months that the degree of 31:19.211 --> 31:24.661 leverage with which they were operating nine months ago was 31:24.657 --> 31:29.157 excessive. Second, they almost certainly 31:29.157 --> 31:35.457 will judge that the volatility associated with the assets they 31:35.463 --> 31:40.113 hold and the degree of uncertainty about their 31:40.114 --> 31:45.494 portfolios is today very substantially greater, 31:45.490 --> 31:50.180 not just than they thought it was nine months ago, 31:50.178 --> 31:54.578 but than it objectively was nine months ago. 31:54.579 --> 32:00.839 Third, the magnitude of tail risk of existential events, 32:00.835 --> 32:07.315 which will cause the solvency of their institution to come 32:07.318 --> 32:14.758 into fundamental doubt, has surely increased as well. 32:14.759 --> 32:22.709 I don't think we have a sound basis for estimating how much 32:22.713 --> 32:29.573 prudent institutions will reduce their leverage. 32:29.569 --> 32:34.099 The leverage choices they may make will, of course, 32:34.100 --> 32:37.540 be very much influenced, in some cases, 32:37.543 --> 32:41.983 by the way in which the government applies capital 32:41.982 --> 32:47.112 rules. If one makes the conservative, 32:47.108 --> 32:54.508 but in my view not absurd, assumption that leverage will 32:54.512 --> 33:01.662 be reduced by 10%, 10% leverage--a 10% reduction 33:01.662 --> 33:10.752 in leverage on a $20 trillion base of asset holding represents 33:10.753 --> 33:18.953 the loss of another $2 trillion in lending capacity. 33:18.950 --> 33:28.500 Add the $1.5 trillion and the $2 trillion and you are looking 33:28.503 --> 33:33.533 at the loss of, on the order of, 33:33.525 --> 33:42.345 $3.5 trillion or on the order of 15% of the previous financial 33:42.352 --> 33:46.262 intermediation capacity. 33:46.259 --> 33:53.269 It is not unreasonable at all to suppose that this is a shock 33:53.271 --> 34:00.521 that is of substantial magnitude relative to the functioning of 34:00.516 --> 34:07.176 the financial system and to the extent that the economy is 34:07.178 --> 34:14.538 dependent on the ability to get credit to the functioning of the 34:14.540 --> 34:20.440 economy. At root, it is the challenge of 34:20.443 --> 34:26.953 policy to address this shock in this environment. 34:26.949 --> 34:35.909 What are the tools that are available to policymakers? 34:35.909 --> 34:42.659 The first imperative of a policy and one that has been 34:42.659 --> 34:49.409 pursued in this crisis since the beginning of 2008, 34:49.409 --> 34:56.489 albeit in my judgment somewhat belatedly, is to seek to 34:56.489 --> 35:00.159 maintain aggregate demand. 35:00.159 --> 35:05.379 Whatever else is happening, an environment of this kind is 35:05.381 --> 35:10.421 likely to be a disinflationary environment over time. 35:10.420 --> 35:15.450 It is likely to be an environment in which if spending 35:15.454 --> 35:21.254 is reduced, asset values will decline, further compounding the 35:21.250 --> 35:24.100 degree of economic damage. 35:24.099 --> 35:33.019 How do policymakers seek to maintain aggregate demand? 35:33.019 --> 35:38.169 Before the events of the last eighteen months, 35:38.170 --> 35:42.520 if asked to comment on this subject, 35:42.519 --> 35:48.169 I would have said the dominant tool of short-run stabilization 35:48.165 --> 35:53.065 policy should be monetary policy not fiscal policy. 35:53.070 --> 35:57.110 I would have explained that fiscal policy, 35:57.106 --> 36:03.106 except in the shortest of runs, was likely to be offset by the 36:03.111 --> 36:07.901 Federal Reserve. So, it would end up affecting 36:07.897 --> 36:12.947 the composition of output, not the level of output. 36:12.949 --> 36:16.729 I would further have explained that it was much more 36:16.732 --> 36:21.182 destructive of long-run values to try to vary spending or the 36:21.183 --> 36:25.413 tax structure in a short-run way in order to stabilize the 36:25.411 --> 36:29.351 economy; that was best accomplished 36:29.350 --> 36:32.140 through monetary policy. 36:32.139 --> 36:41.579 I continue to regard that as a reasonable judgment about policy 36:41.579 --> 36:46.349 in normal times. It seems to me that, 36:46.350 --> 36:52.140 while it is appropriate in this environment to seek to use 36:52.137 --> 36:56.297 monetary policy as a tool of stimulus, 36:56.300 --> 37:00.900 there are several considerations that suggest that 37:00.903 --> 37:07.013 it should not be the sole tool of stimulus in this environment. 37:07.010 --> 37:13.390 First, there is much more than the normal degree of uncertainty 37:13.388 --> 37:19.968 about what the impact of changes in interest rates will be on the 37:19.973 --> 37:23.063 level of aggregate demand. 37:23.059 --> 37:27.499 Imagine, for example, an intermediary that is 37:27.500 --> 37:29.720 capital-constrained. 37:29.719 --> 37:35.159 It only has a billion dollars of capital and it therefore can 37:35.159 --> 37:38.059 only lend ten billion dollars. 37:38.059 --> 37:41.759 Imagine now that the rate of interest is brought down; 37:41.760 --> 37:46.860 it can still only lend ten billion dollars and the rate at 37:46.858 --> 37:52.488 which it lends will still be the point at which the demand curve 37:52.494 --> 37:56.434 for its loans is at ten billion dollars. 37:56.429 --> 38:00.299 The only result of bringing down the interest rate will be a 38:00.302 --> 38:02.602 reduction in its cost of funding, 38:02.599 --> 38:07.009 which over time will allow it to rebuild its capital, 38:07.014 --> 38:11.094 but that is a very slow process--a very different 38:11.089 --> 38:15.499 situation than the more normal situation in which the 38:15.503 --> 38:18.733 intermediary is not constrained. 38:18.730 --> 38:21.330 Is this the case for most intermediaries? 38:21.330 --> 38:22.640 It's very difficult to know. 38:22.639 --> 38:27.919 Most intermediaries are not at their capital constraint, 38:27.920 --> 38:33.780 but they probably are planning for some given cushion relative 38:33.777 --> 38:36.847 to their capital constraint. 38:36.849 --> 38:39.849 In this environment, it is likely that the 38:39.847 --> 38:44.157 multiplier for monetary policy is less and it is surely more 38:44.161 --> 38:47.671 uncertain and that is an inhibition to the use of 38:47.671 --> 38:52.031 monetary policy. Second inhibition to the use of 38:52.025 --> 38:56.715 monetary policy as a sole tool is: most of the time, 38:56.719 --> 39:01.199 with respect to most economic policies, we may not know what 39:01.198 --> 39:05.068 the multiplier associated with the instrument is, 39:05.070 --> 39:11.010 but we know where we have set the policy dial. 39:11.010 --> 39:16.790 Through all of modern economic experience, until nine months 39:16.794 --> 39:22.094 ago, it would be completely impossible to use any known 39:22.089 --> 39:27.579 econometric technique to say anything about the impact of 39:27.580 --> 39:33.400 changes in the Fed Funds Rate, as distinct from changes in 39:33.401 --> 39:37.561 LIBOR, on the behavior of the aggregate economy. 39:37.559 --> 39:41.779 They were perfectly multi-linear as the spread 39:41.779 --> 39:46.279 between them was small and little fluctuating. 39:46.280 --> 39:51.360 In the last nine months, a substantial spread has 39:51.361 --> 39:57.501 emerged between the Fed Funds Rate--the rate at which banks 39:57.501 --> 40:02.371 lend to each other overnight--and the Interbank 40:02.370 --> 40:07.240 rate that prevails over longer horizons. 40:07.239 --> 40:12.429 For a variety of reasons that I don't have time to get into, 40:12.431 --> 40:17.891 it is almost certainly the case that large parts of this spread 40:17.887 --> 40:23.077 reflect the credit risk that banks experience when they lend 40:23.079 --> 40:27.269 to each other. Probably, the right indicator 40:27.270 --> 40:31.210 for monetary policy is not the Fed Funds Rate, 40:31.210 --> 40:34.230 but is LIBOR, in which case there has been 40:34.234 --> 40:38.664 rather less easing of monetary policy than looking at the Fed 40:38.660 --> 40:43.340 Funds Rate would suggest; but, one cannot make that 40:43.336 --> 40:46.376 decision in a definitive way. 40:46.380 --> 40:51.140 So, easing monetary policy as the Fed has been doing for some 40:51.142 --> 40:55.672 time and as the Fed did today in lowering the rate another 40:55.667 --> 41:00.507 twenty-five basis points is almost certainly appropriate. 41:00.510 --> 41:06.010 But it is almost certainly not appropriate to rely only on a 41:06.010 --> 41:11.330 single instrument in this context because one is uncertain 41:11.325 --> 41:16.355 about how much of that instrument one is applying, 41:16.360 --> 41:21.670 because one is uncertain about what the multiplier of that 41:21.665 --> 41:25.385 instrument is, and because in the current 41:25.388 --> 41:31.068 context there are reasons to be concerned about the pernicious 41:31.066 --> 41:35.996 side effects of excessively low interest rates. 41:36.000 --> 41:42.140 Two stand out. The first--while in general I 41:42.136 --> 41:49.506 am very much with the Keynesian consensus that the inflation 41:49.513 --> 41:54.643 process depends upon aggregate demand, 41:54.639 --> 41:58.399 in the current context, lower interest rates may well 41:58.404 --> 42:02.534 contribute directly to the inflation process through their 42:02.531 --> 42:06.661 impact on the dollar and through their impact on commodity 42:06.658 --> 42:10.578 prices. The second--while I give less 42:10.579 --> 42:17.049 weight than most to critiques of a policy environment that blame 42:17.052 --> 42:23.322 our current problems on the fact that interest rates were kept 42:23.320 --> 42:29.280 low for a long time responding to the last recession, 42:29.280 --> 42:34.120 it seems to me reasonable to give some weight to the concern 42:34.117 --> 42:38.957 that excessively low interest rates set the stage for future 42:38.956 --> 42:41.316 bubbles. I, therefore, 42:41.322 --> 42:47.702 think it's appropriate that the Congress acted--and I suspect 42:47.703 --> 42:53.453 the Congress won't need to act again--to provide fiscal 42:53.445 --> 42:56.525 stimulus to the economy. 42:56.530 --> 43:02.020 In advocating fiscal stimulus just after the turn of the year, 43:02.017 --> 43:06.787 I urged that the stimulus be timely because experience 43:06.785 --> 43:12.085 suggests that the difficulty of lags and implementation have 43:12.092 --> 43:17.762 traditionally been serious with respect to fiscal policy. 43:17.760 --> 43:23.790 That it be targeted--designed to generate as much spending as 43:23.785 --> 43:28.885 possible as quickly as possible, which meant that tax breaks 43:28.888 --> 43:32.518 should be channeled to those with lower incomes rather than 43:32.516 --> 43:34.326 those with higher incomes. 43:34.330 --> 43:39.030 That recommendation if adopted; means that taxes--means that 43:39.033 --> 43:43.203 benefit programs would go to people like food stamps and 43:43.204 --> 43:46.014 unemployment insurance recipients, 43:46.010 --> 43:47.640 who were likely to have a very high-margin propensity to 43:47.643 --> 43:53.243 consume. The political process was 43:53.237 --> 43:57.897 unwilling to do that. 43:57.900 --> 44:03.340 And that the fiscal stimulus must be temporary in light of 44:03.341 --> 44:09.071 the seriousness of the country's long-run fiscal problems. 44:09.070 --> 44:18.810 Macroeconomic policy in this environment can help to maintain 44:18.814 --> 44:25.754 aggregate demand, but it does not address the 44:25.751 --> 44:33.881 disruption in the productive mechanism associated with the 44:33.881 --> 44:43.011 decline in the financial sectors capacity for intermediation. 44:43.010 --> 44:47.980 There are a number of tools of policy directed at the 44:47.976 --> 44:51.506 maintenance of financial stability. 44:51.510 --> 44:56.800 One has received overwhelming emphasis to date and it is 44:56.804 --> 45:00.664 entirely appropriate, but I would suggest 45:00.655 --> 45:03.635 presumptively insufficient. 45:03.639 --> 45:11.059 That is, the Fed, through a dizzying array of new 45:11.061 --> 45:19.811 facilities and ad hoc actions, has made clear that it stands 45:19.807 --> 45:27.927 ready against good collateral to provide liquidity to financial 45:27.925 --> 45:33.755 institutions. Now, go back to the 45:33.759 --> 45:43.009 calculations that I went through a moment ago. 45:43.010 --> 45:50.800 I emphasized the loss of capital and I represented the 45:50.800 --> 45:59.770 desired reduction in levels of leverage associated with a more 45:59.766 --> 46:02.996 risky environment. 46:03.000 --> 46:08.400 I did not say anything in doing those calculations about the 46:08.398 --> 46:13.338 difficulty of attracting funds, about the difficulty of 46:13.339 --> 46:19.079 borrowing funds. I did not raise the concern 46:19.082 --> 46:23.432 about bank runs and the like. 46:23.429 --> 46:33.029 The provision of liquidity by the Fed is necessary; 46:33.030 --> 46:38.680 without it, in my judgment, it is very likely that we would 46:38.683 --> 46:44.243 have seen problematic runs on a number of institutions. 46:44.239 --> 46:49.469 I believe the Fed behaved in a broadly correct way in 46:49.467 --> 46:53.787 responding to the events at Bear Stearns. 46:53.789 --> 46:58.579 It is certainly possible that if the Fed had not acted, 46:58.581 --> 47:03.371 the losses to Bear Stearns shareholders would have been 47:03.372 --> 47:05.822 total. There would have been 47:05.822 --> 47:10.062 substantial disruption in the bond markets for a period of 47:10.062 --> 47:12.842 five days. A valuable lesson would have 47:12.844 --> 47:16.544 been taught and the matter--and the world would have moved 47:16.537 --> 47:19.897 forward. It is certainly possible that 47:19.897 --> 47:21.537 that is the case. 47:21.539 --> 47:25.319 It seems to me, one doesn't need to assign a 47:25.324 --> 47:30.344 very high probability to the alternative scenario in which 47:30.341 --> 47:35.271 runs on Bear Stearns cascaded to other institutions, 47:35.269 --> 47:39.839 which then cascaded to other institutions and led to 47:39.840 --> 47:45.310 substantial destruction of many institutions that were thought 47:45.308 --> 47:47.278 previously viable. 47:47.280 --> 47:51.870 I don't think one has to assign a very great probability to the 47:51.867 --> 47:55.787 latter event to believe the rescue of Bear Stearns was 47:55.789 --> 48:00.279 worthwhile. I, myself, would assign a much 48:00.281 --> 48:07.161 greater than 50% probability to a major systemic-risk incident 48:07.157 --> 48:13.467 if the damage at Bear Stearns had not been contained. 48:13.469 --> 48:17.469 In any event, while noting the necessity of 48:17.467 --> 48:23.457 all of that, I want to emphasize its insufficiency--that it does 48:23.463 --> 48:28.793 nothing about either the reduction in capital and it does 48:28.793 --> 48:35.363 very little about the reduction in desired degrees of leverage, 48:35.360 --> 48:40.640 that I would suggest are at the root of the problem. 48:40.639 --> 48:46.959 That brings me to what is, I think, the second crucial 48:46.961 --> 48:53.821 tool of financial stability, which is the most important 48:53.818 --> 49:00.008 financial policy issue in the current environment. 49:00.010 --> 49:01.960 Unfortunately, by its nature, 49:01.958 --> 49:05.788 it is difficult to tell the vigor with which it is being 49:05.785 --> 49:10.215 pursued. That is the infusion of new 49:10.218 --> 49:15.148 capital into financial institutions. 49:15.150 --> 49:21.400 Notice that if financial institutions raised their level 49:21.395 --> 49:28.205 capital, then at any given level of leverage they are able to 49:28.208 --> 49:35.018 raise their degree of lending and you are able to reinstitute 49:35.021 --> 49:41.041 the intermediation capacity that has been lost. 49:41.039 --> 49:45.609 There is a crucial un-internalized externality in 49:45.612 --> 49:49.902 our current system of financial regulation. 49:49.900 --> 49:57.060 Institutions are regulated and their soundness is judged on the 49:57.060 --> 50:03.530 basis of a capital ratio--on the basis of their degree of 50:03.528 --> 50:08.358 leverage. They are pressured rationally 50:08.362 --> 50:15.602 enough not to have deposits that are short-term liabilities that 50:15.600 --> 50:21.690 are too large relative to their degree of capital. 50:21.690 --> 50:27.680 Notice that an institution can acquire--can reduce its--can 50:27.680 --> 50:33.570 improve its capital ratio or reduce its leverage in one of 50:33.567 --> 50:38.497 two ways. It can reduce its leverage by 50:38.504 --> 50:45.824 shrinking its balance sheet or it can reduce its leverage by 50:45.816 --> 50:50.026 expanding its level of capital. 50:50.030 --> 50:54.080 If the institution is indifferent between those two 50:54.083 --> 50:58.953 choices, society is clearly not indifferent between those two 50:58.947 --> 51:03.787 choices. It has a strong preference for 51:03.792 --> 51:07.022 the issuance of capital. 51:07.019 --> 51:11.379 Or to put essentially the same point in a different way, 51:11.383 --> 51:14.163 if an institution raises capital, 51:14.159 --> 51:19.049 it dilutes its shareholders and a substantial part of the 51:19.053 --> 51:24.383 benefit of that extra capital goes to the holders of its risky 51:24.383 --> 51:28.693 debt, whose claims have now become 51:28.687 --> 51:32.437 more secure. As a consequence, 51:32.435 --> 51:37.535 the private incentive for capital raising, 51:37.538 --> 51:44.008 even with a capital rule operating as regulation, 51:44.010 --> 51:50.820 is substantially less than the social incentive. 51:50.820 --> 51:57.120 One hopes that the regulatory system is internalizing this 51:57.118 --> 52:02.198 externality--is placing substantial pressure on 52:02.201 --> 52:07.471 institutions, who after all are enormously 52:07.472 --> 52:14.862 dependent on their relationship with the financial authorities 52:14.861 --> 52:21.041 to raise more capital than they otherwise would. 52:21.039 --> 52:31.089 It is very difficult to know from the outside to what extent 52:31.091 --> 52:35.011 that is taking place. 52:35.010 --> 52:38.860 A prudent regulator who was forcing an institution to raise 52:38.858 --> 52:43.038 more capital would conspire with that institution to not have it 52:43.038 --> 52:47.218 be known that the regulator was forcing the institution to raise 52:47.219 --> 52:51.959 more capital, lest the health of the 52:51.959 --> 52:57.019 institution come into question. 52:57.019 --> 53:00.489 Certainly there have been a number of favorable developments 53:00.490 --> 53:03.550 involving the raising of capital in recent months, 53:03.550 --> 53:11.070 but in my judgment and the kind of numbers I presented suggest 53:11.073 --> 53:15.023 that we have a long way to go. 53:15.019 --> 53:19.309 I would suggest, of particular importance in 53:19.305 --> 53:23.485 this regard is the capital situation of the 53:23.492 --> 53:27.182 government-sponsored enterprises, 53:27.179 --> 53:32.389 which are levered thirty-to-one and which have a crucial role in 53:32.393 --> 53:36.783 this environment in supporting the mortgage market. 53:36.780 --> 53:40.550 To date, they have been encouraged to buy more mortgages 53:40.554 --> 53:44.474 through a reduction in their capital requirement and there 53:44.465 --> 53:48.025 have been rather diffuse statements about it as their 53:48.034 --> 53:52.224 hope and intention to raise more capital in the future. 53:52.219 --> 53:54.189 We've been moving towards the future; 53:54.190 --> 53:58.320 they have not yet raised the capital. 53:58.320 --> 54:03.060 Some rather skeptical statements about the need to 54:03.058 --> 54:08.468 raise capital have come from their management with as yet 54:08.473 --> 54:12.543 relatively little criticism from the public 54:12.535 --> 54:21.005 authorities--liquidity, raising capital. 54:21.010 --> 54:26.770 The third broad policy question involves more direct 54:26.765 --> 54:29.695 intervention in markets. 54:29.699 --> 54:33.119 In my judgment, any attempt to support the 54:33.124 --> 54:36.804 housing market, in general, at current levels 54:36.799 --> 54:38.719 would be misguided. 54:38.719 --> 54:43.319 There is little reason to believe that house prices have 54:43.323 --> 54:48.093 yet fallen to fundamental levels, let alone fallen through 54:48.094 --> 54:50.024 fundamental levels. 54:50.019 --> 54:54.009 The lesson of the Japanese experience is that price keeping 54:54.014 --> 54:58.014 operations by governments delay processes of adjustment. 54:58.010 --> 55:02.600 Why would anyone want to buy something whose price is being 55:02.602 --> 55:06.962 supported by the government since it's likely that there 55:06.958 --> 55:11.308 will be further downwards adjustment rather than upwards 55:11.313 --> 55:16.733 adjustment? Where there is a much stronger 55:16.726 --> 55:24.446 case is for policy to support the mortgage-backed security 55:24.453 --> 55:32.053 market where prices reflect expectations of terribly high 55:32.045 --> 55:35.565 default performances. 55:35.570 --> 55:40.250 Some mortgage-backed securities on some classes of subprime debt 55:40.250 --> 55:44.780 are building in the expectation that more than half the houses 55:44.781 --> 55:50.371 will be foreclosed; estimates that go way beyond 55:50.373 --> 55:57.033 what even fairly pessimistic forecasts suggest. 55:57.030 --> 56:02.160 By stealth, a rather substantial program of support 56:02.160 --> 56:08.010 for the mortgage-backed securities market is under way. 56:08.010 --> 56:10.250 Key elements include the actions of the 56:10.246 --> 56:13.776 government-sponsored enterprises that I referred to that will 56:13.778 --> 56:17.368 enable $180 billion of increased purchasing of mortgage-backed 56:17.369 --> 56:21.779 securities. A similar change on the part of 56:21.777 --> 56:28.147 the federal home loan banks and if, with that thirty-to-one 56:28.151 --> 56:32.621 leverage ratio, the GSE's are caused to raise 56:32.623 --> 56:37.013 significantly more capital, there is the prospect of 56:37.010 --> 56:41.480 significantly further support for the mortgage-backed 56:41.481 --> 56:46.191 securities market, whose greater health would 56:46.186 --> 56:51.846 impact both the availability of credit to potential new 56:51.847 --> 56:57.087 homeowners and the availability--and the health--of 56:57.090 --> 57:04.010 the financial institutions that are holding their securities. 57:04.010 --> 57:10.530 I think--and I will not dwell on the point for reasons of 57:10.533 --> 57:17.173 time--that there is also a strong case for policy measures 57:17.174 --> 57:23.354 directed at strategic intervention in near foreclosure 57:23.348 --> 57:29.868 situations to promote the writing down of mortgages as an 57:29.872 --> 57:33.952 alternative to foreclosure. 57:33.949 --> 57:38.959 Given that, available estimates suggest that the losses 57:38.964 --> 57:44.814 associated with the foreclosure process itself can easily add up 57:44.814 --> 57:50.854 to more than half the value--the residual value--of the home. 57:50.849 --> 57:56.359 Finally, I would highlight the international dimension of 57:56.360 --> 58:00.100 policy and would stress two aspects. 58:00.099 --> 58:07.059 The global dynamic here is that two things are happening. 58:07.059 --> 58:12.069 First, U.S. aggregate demand is falling. 58:12.070 --> 58:17.010 Second, the fall in the dollar associated with all the various 58:17.010 --> 58:22.190 ramifications of these financial events are switching expenditure 58:22.193 --> 58:26.813 from the rest of the world towards the United States. 58:26.809 --> 58:30.909 Third, some of the forces operating in the United States 58:30.909 --> 58:35.309 are also operating to reduce aggregate demand in the rest of 58:35.306 --> 58:38.756 the world. A crucial priority for policy 58:38.760 --> 58:43.420 coordination in that context has to be the stimulation of 58:43.416 --> 58:47.486 aggregate demand in the rest of the world and, 58:47.489 --> 58:52.679 probably, the achievement of a more balanced pattern of dollar 58:52.675 --> 58:55.335 depreciation. To date, the dollar has 58:55.344 --> 58:58.984 depreciated disproportionately against a very small number of 58:58.981 --> 59:01.641 currencies, principally the Euro, 59:01.643 --> 59:06.823 that stand for countries that do not have problems like those 59:06.823 --> 59:11.743 of the United States and that allow their exchange rate to 59:11.744 --> 59:14.504 fluctuate; a more balanced path would be 59:14.504 --> 59:17.704 desirable. Second priority for policy 59:17.696 --> 59:23.676 coordination is with respect to the very large quantities of 59:23.680 --> 59:28.650 liquidity that are now managed by governments. 59:28.650 --> 59:31.310 They represent, potentially, 59:31.308 --> 59:35.838 an enormously important resource in the current 59:35.838 --> 59:41.348 context--pools of patient capital that can be managed for 59:41.352 --> 59:46.672 the long term and that could provide liquidity to badly 59:46.669 --> 59:55.829 distorted fixed-income markets, but have not yet done so on a 59:55.832 --> 59:59.322 substantial scale. 59:59.320 --> 1:00:04.860 These sets of measures represent, I believe, 1:00:04.860 --> 1:00:10.530 a policy approach that offers a very good, 1:00:10.530 --> 1:00:15.590 though not certain, chance of containing the 1:00:15.594 --> 1:00:17.954 current situation. 1:00:17.949 --> 1:00:23.139 The risks, it seems to me, are much more on the side of it 1:00:23.142 --> 1:00:28.332 being pursued with too little vigor than on the side of it 1:00:28.334 --> 1:00:31.254 being pursued with too much. 1:00:31.250 --> 1:00:35.460 In particular, the imperative of the rapid 1:00:35.459 --> 1:00:41.209 recapitalization of financial institutions stands out. 1:00:41.210 --> 1:00:47.190 I will for reasons of time leave a discussion that I had 1:00:47.190 --> 1:00:53.390 intended, raising questions about the regulatory structure 1:00:53.389 --> 1:00:59.809 and changes that might make crises of these kind less likely 1:00:59.805 --> 1:01:06.545 in the future, to the question period. 1:01:06.550 --> 1:01:11.970 But, I will just make five telegraphic observations. 1:01:11.969 --> 1:01:18.529 First, it is the irony of financial crisis that the very 1:01:18.531 --> 1:01:25.211 forces that caused it are the opposite--are also the very 1:01:25.212 --> 1:01:30.822 forces that are necessary to get out of it. 1:01:30.820 --> 1:01:35.450 If the world's problem eighteen months ago was that there was 1:01:35.452 --> 1:01:38.312 too much greed and too little fear, 1:01:38.309 --> 1:01:44.599 the world's problem today is that there is too much fear and 1:01:44.595 --> 1:01:46.615 too little greed. 1:01:46.619 --> 1:01:51.189 The policies that would have been highly constructive 1:01:51.194 --> 1:01:55.774 eighteen months ago in inhibiting undesirable lending 1:01:55.769 --> 1:01:59.379 are potentially counterproductive today in 1:01:59.376 --> 1:02:02.716 inhibiting constructive lending. 1:02:02.719 --> 1:02:09.289 Second, there's an overwhelming case for stronger consumer 1:02:09.290 --> 1:02:13.440 regulation in all financial areas. 1:02:13.440 --> 1:02:18.390 If one studies carefully the extent to which subprime 1:02:18.391 --> 1:02:23.921 mortgages at egregious interest rates were taken on to help 1:02:23.915 --> 1:02:30.005 people get out of credit card debt at super egregious rates, 1:02:30.010 --> 1:02:34.490 one has a sense of the greater complexity of the problem and 1:02:34.491 --> 1:02:39.201 the need to view it from an integrated consumer perspective. 1:02:39.199 --> 1:02:46.509 Third, the government should be paid for providing the safety 1:02:46.509 --> 1:02:54.189 net and the modalities that are necessary to do that--need to be 1:02:54.185 --> 1:03:01.365 much more carefully designed so that if everything works out 1:03:01.373 --> 1:03:05.023 splendidly in, for example, 1:03:05.022 --> 1:03:10.482 the Bear Stearns transaction, the government would be a 1:03:10.479 --> 1:03:15.629 beneficiary of the success since it surely will be a 1:03:15.633 --> 1:03:21.803 beneficiary--surely will bear the burden if everything else is 1:03:21.797 --> 1:03:25.257 lost. Fourth, the dominant--the 1:03:25.260 --> 1:03:30.570 single most important area for thought in thinking about the 1:03:30.570 --> 1:03:34.800 future of financial regulation is the problem of 1:03:34.800 --> 1:03:40.620 pro-cyclicality. Prudent behavior by individual 1:03:40.621 --> 1:03:49.011 institutions involves shrinkage of balance sheets in bad times. 1:03:49.010 --> 1:03:55.090 Everyone's effort to shrink balance sheets in bad times has 1:03:55.087 --> 1:04:00.847 potentially catastrophic consequences for the system. 1:04:00.849 --> 1:04:06.599 How to regulate to avoid pro-cyclicality is a fourth 1:04:06.601 --> 1:04:08.971 crucial imperative. 1:04:08.969 --> 1:04:16.269 Fifth, these issues should be addressed slowly, 1:04:16.269 --> 1:04:22.149 carefully, and deliberately; as I suggested, 1:04:22.152 --> 1:04:26.722 the problems of getting us out of a situation of this kind are 1:04:26.718 --> 1:04:31.278 in many ways the opposite of the problems of preventing future 1:04:31.283 --> 1:04:32.933 financial crises. 1:04:32.929 --> 1:04:36.909 It will be tempting, but would be unfortunate, 1:04:36.906 --> 1:04:41.936 to address the concerns through whatever officials do when 1:04:41.942 --> 1:04:47.732 they're not sure what to do, which is reorganize. 1:04:47.730 --> 1:04:54.320 Our central challenges involve the questions of how we define 1:04:54.317 --> 1:04:58.157 capital, how we regulate capital, 1:04:58.159 --> 1:05:02.559 for whom we regulate capital, how we avoid pro-cyclicality. 1:05:02.559 --> 1:05:06.319 Until we know the answers to those questions, 1:05:06.324 --> 1:05:11.204 it will be difficult to think intelligently about what the 1:05:11.201 --> 1:05:15.481 right bureaucratic structure of regulation is. 1:05:15.480 --> 1:05:22.140 All of this is to suggest to the students here that the 1:05:22.136 --> 1:05:29.286 macroeconomics of the early twenty-first century will be no 1:05:29.285 --> 1:05:35.565 less important than the macroeconomics of the latter 1:05:35.572 --> 1:05:40.012 part of the twentieth century. 1:05:40.010 --> 1:05:44.600 That macroeconomics has a good chance of being much more 1:05:44.604 --> 1:05:48.614 centrally involved with financial issues than the 1:05:48.614 --> 1:05:53.214 macroeconomics of the late twentieth century--or perhaps 1:05:53.209 --> 1:05:59.069 fairness requires me to say, the macroeconomics practiced 1:05:59.067 --> 1:06:04.667 outside of New Haven in the late twentieth century. 1:06:04.670 --> 1:06:15.080 The many, many imprecisions and speculations in these remarks, 1:06:15.078 --> 1:06:20.468 I apologize for, but offer as a modest and 1:06:20.465 --> 1:06:25.745 inadequate defense that they point up that there's a great 1:06:25.746 --> 1:06:31.396 deal of very valuable research to be done that will contribute 1:06:31.397 --> 1:06:35.007 to our understanding of economies. 1:06:35.010 --> 1:06:37.710 Also, I believe, to Art Okun's vision, 1:06:37.712 --> 1:06:40.932 of harnessing that understanding to help make 1:06:40.926 --> 1:06:45.086 better and more secure lives for our fellow citizens. 1:06:45.090 --> 1:06:46.190 Thank you very much. 1:06:46.190 --> 1:06:56.160 1:06:56.159 --> 1:06:59.699 I went over my time, but I've got some time for 1:06:59.697 --> 1:07:02.417 questions. Jeff? 1:07:02.420 --> 1:07:05.650 Student: Larry, do you think that Secretary 1:07:05.654 --> 1:07:09.554 Paulson's recommendations are appropriate to the problem? 1:07:09.550 --> 1:07:11.800 Or do they that seem to be a dusted off failed answer to a 1:07:11.796 --> 1:07:12.896 different set of problems. 1:07:12.900 --> 1:07:13.950 Secretary Paulson's recommendations last year 1:07:13.954 --> 1:07:15.254 addressed a presumed crisis overregulation diminishing. 1:07:15.249 --> 1:07:16.689 You seem to suggest that the real problems are misperceptions 1:07:16.688 --> 1:07:17.528 of risk and not mistaken regulation. 1:07:17.527 --> 1:07:18.937 Will the Secretary's repurposed proposals die on the vine or 1:07:18.941 --> 1:07:20.621 will they still get political traction despite your warning? 1:07:20.620 --> 1:07:24.120 Sorry about the mixed metaphor! 1:07:24.119 --> 1:07:27.859 Professor Larry Summers: Get traction on the vine, 1:07:27.861 --> 1:07:32.071 I guess. There's certain--I'm going to 1:07:32.073 --> 1:07:39.533 not associate myself with all the words you used in describing 1:07:39.526 --> 1:07:44.976 the Secretary's proposals, though I can't understand 1:07:44.977 --> 1:07:48.027 why--I can understand why you use them. 1:07:48.030 --> 1:07:55.990 I think that it is almost always a good idea to bet on 1:07:55.992 --> 1:08:00.802 inertia in anything political. 1:08:00.800 --> 1:08:07.000 I suspect that in the remainder of this election year it is very 1:08:07.000 --> 1:08:12.410 unlikely that anything will get done that doesn't appear 1:08:12.414 --> 1:08:17.044 necessary to respond to current difficulty. 1:08:17.039 --> 1:08:19.949 And usually, when there's a new 1:08:19.952 --> 1:08:25.002 administration or a new secretary, the problem is not 1:08:25.000 --> 1:08:31.020 excessive fealty to the visions of the previous secretary, 1:08:31.020 --> 1:08:33.350 the problem is a desperate effort at artificial product 1:08:33.352 --> 1:08:35.952 differentiation. There were a number of issues I 1:08:35.949 --> 1:08:39.709 found when I was at the Treasury where there was a certain way to 1:08:39.705 --> 1:08:42.165 do it that was kind of a good idea and, 1:08:42.170 --> 1:08:45.020 actually, our predecessors had more or less figured it out. 1:08:45.020 --> 1:08:48.470 It was my experience that we never decided to say, 1:08:48.467 --> 1:08:52.267 let's really work to get Secretary Brady's good idea in 1:08:52.266 --> 1:08:54.956 place. We always found a few points of 1:08:54.961 --> 1:08:58.621 minor difference and found a variety of ways of relabeling 1:08:58.624 --> 1:09:00.254 it. That was when we thought the 1:09:00.254 --> 1:09:01.904 ideas were good and he had it right. 1:09:01.899 --> 1:09:07.009 In the cases where we didn't, it was much more dramatic. 1:09:07.010 --> 1:09:14.830 So, I would be surprised if the ideas contained in that report 1:09:14.834 --> 1:09:21.894 prove to be a dominant starting point for discussion. 1:09:21.890 --> 1:09:25.520 To be sure, I do want to say one thing. 1:09:25.520 --> 1:09:30.880 There is one thing that that report was getting at and there 1:09:30.875 --> 1:09:36.405 is one aspect of our system that I think we need to think very 1:09:36.411 --> 1:09:39.771 carefully about and it goes deep. 1:09:39.770 --> 1:09:44.340 That is, do we believe in competitive regulation of 1:09:44.343 --> 1:09:46.633 financial institutions? 1:09:46.630 --> 1:09:50.990 We have a banking system and it has for a long time been a 1:09:50.992 --> 1:09:55.662 fairly sacred principle that we have a so-called "dual banking 1:09:55.661 --> 1:09:57.211 system." The idea is, 1:09:57.211 --> 1:10:00.001 well you get to choose your regulator and therefore no 1:09:59.999 --> 1:10:03.049 regulator can be too stupid because if they're unreasonable 1:10:03.050 --> 1:10:04.930 or stupid, then people will choose a 1:10:04.932 --> 1:10:05.852 different regulator. 1:10:05.850 --> 1:10:09.080 Well, that's one way of thinking about it. 1:10:09.079 --> 1:10:11.639 Another is that if regulators like to have more jurisdiction 1:10:11.638 --> 1:10:16.418 and they like to do more stuff, competitive regulation is a 1:10:16.423 --> 1:10:21.023 prescription for limited regulation. 1:10:21.020 --> 1:10:25.190 Implicit in Secretary Paulson's--sort of, 1:10:25.187 --> 1:10:30.707 in the language one might expect from a Republican, 1:10:30.710 --> 1:10:36.850 emphasis on efficiency and the avoidance of duplication was a 1:10:36.851 --> 1:10:41.151 moving away from competitive regulation. 1:10:41.149 --> 1:10:46.599 I suspect that maybe quite an important idea. 1:10:46.600 --> 1:10:49.990 T,N? Student: It is all very 1:10:49.993 --> 1:10:53.473 well to say in a general way that it is good to avoid 1:10:53.465 --> 1:10:57.135 duplication and to move away from competitive regulation 1:10:57.137 --> 1:11:00.677 without going to where the trade-off lies in different 1:11:00.676 --> 1:11:03.076 contexts. After all there is a lot of 1:11:03.079 --> 1:11:04.679 experience to draw upon. 1:11:04.682 --> 1:11:06.822 For example U.S., India, the E.U. 1:11:06.818 --> 1:11:10.618 and many others have more than one regulaory authority for 1:11:10.624 --> 1:11:14.634 different parts of the financial sector sector while the U.K. 1:11:14.630 --> 1:11:17.500 has just one. All the time the sector itself 1:11:17.501 --> 1:11:21.241 expanding with new agents, poroducts etc entering all the 1:11:21.240 --> 1:11:24.580 time, in part as a response to the system in place. 1:11:24.579 --> 1:11:28.449 Based on your experience and analytical skills where do you 1:11:28.451 --> 1:11:31.591 come out on this issue? 1:11:31.590 --> 1:11:38.250 Professor Larry Summers: The question said--I'm going to 1:11:38.251 --> 1:11:40.831 rephrase the question. 1:11:40.829 --> 1:11:42.769 The question was, okay Larry that's good, 1:11:42.774 --> 1:11:45.064 but don't duck all the interesting questions. 1:11:45.060 --> 1:11:47.230 There's a lot of experience in a lot of places. 1:11:47.229 --> 1:11:49.279 In particular, England has a single financial 1:11:49.280 --> 1:11:51.020 regulator. If you had to guess right now, 1:11:51.024 --> 1:11:52.214 what's the best way to do it? 1:11:52.210 --> 1:12:00.020 Is that a fair rephrase of the question, T.N.? 1:12:00.020 --> 1:12:08.280 I think that the three things I would emphasize are: 1:12:08.284 --> 1:12:16.394 competitive regulation is probably a mistake if you 1:12:16.386 --> 1:12:20.596 believe in regulation. 1:12:20.600 --> 1:12:27.320 The central bank does need to have a role in financial 1:12:27.323 --> 1:12:34.943 regulation of major institutions because if it doesn't have a 1:12:34.936 --> 1:12:41.656 role in the regulation of financial institutions, 1:12:41.659 --> 1:12:46.779 it's not going to know things that it should know and it's 1:12:46.784 --> 1:12:51.014 going to end up living with the consequences. 1:12:51.010 --> 1:12:53.600 So, the central bank should be crucially involved in the 1:12:53.595 --> 1:12:55.705 regulation of major financial institutions. 1:12:55.710 --> 1:13:01.060 That--it's a--if you believe in regulations that are pursuing 1:13:01.055 --> 1:13:05.775 objectives that are not immediately in the interest of 1:13:05.777 --> 1:13:09.877 financial institutions--If you believe in those 1:13:09.876 --> 1:13:13.436 regulations--consumer regulations, 1:13:13.439 --> 1:13:17.529 community regulations and the like--If you believe in those 1:13:17.525 --> 1:13:21.395 regulations and you want them carried out with vigor, 1:13:21.399 --> 1:13:25.069 you need to give them to their own regulator, 1:13:25.073 --> 1:13:30.003 who is not responsible for the health of the institutions in 1:13:30.000 --> 1:13:33.340 the way that the Federal Reserve is. 1:13:33.340 --> 1:13:37.190 Separation of consumer regulation, elimination of 1:13:37.193 --> 1:13:40.893 competitive regulation, and regulation of major 1:13:40.887 --> 1:13:45.537 financial institutions by an entity--by a single entity--in 1:13:45.543 --> 1:13:49.643 which the central bank has a role would be the three 1:13:49.638 --> 1:13:52.848 principles that I would suggest. 1:13:52.850 --> 1:13:56.770 John? Student: I would like to 1:13:56.765 --> 1:13:59.955 make a comment and ask two questions. 1:13:59.960 --> 1:14:03.760 My comment is, like everyone I'm sure, 1:14:03.760 --> 1:14:07.560 we all enjoyed your talk very much. 1:14:07.560 --> 1:14:09.560 Professor Larry Summers: The comment is he enjoyed the 1:14:09.557 --> 1:14:11.877 talk. That tends to--that's the kind 1:14:11.881 --> 1:14:15.921 of thing you say before the devastating question and I knew 1:14:15.924 --> 1:14:20.044 John when we were in graduate school together and nothing in 1:14:20.037 --> 1:14:23.867 my four years of going to graduate school with him makes 1:14:23.871 --> 1:14:26.731 it less likely that after he said, 1:14:26.729 --> 1:14:29.519 he really enjoyed the talk, that the devastating question 1:14:29.522 --> 1:14:30.522 was going to come. 1:14:30.520 --> 1:14:33.280 So I am braced. Student: It took me five 1:14:33.284 --> 1:14:34.374 years of graduate school. 1:14:34.369 --> 1:14:38.499 But anyway, your story of the leverage cycle, 1:14:38.496 --> 1:14:42.806 as I called it, where people leveraged too much 1:14:42.811 --> 1:14:48.531 in good times and then in bad times just when you want them to 1:14:48.533 --> 1:14:51.523 leverage, the leverage drops in the 1:14:51.519 --> 1:14:54.549 system and there's not enough--there's too little 1:14:54.552 --> 1:14:57.272 leverage to get you out of the doldrums. 1:14:57.270 --> 1:15:00.770 I think it's a very important story. 1:15:00.770 --> 1:15:02.740 So, that was my comment, but the question though is-- 1:15:02.735 --> 1:15:04.965 Professor Larry Summers: I thought at least I got the 1:15:04.966 --> 1:15:06.496 comment, which was that he really liked 1:15:06.500 --> 1:15:08.340 the talk; that didn't even count at all. 1:15:08.340 --> 1:15:09.940 Okay go right ahead. 1:15:09.939 --> 1:15:12.309 Student: The question is, you don't seem to have 1:15:12.310 --> 1:15:14.110 follow through to the end of the story, 1:15:14.109 --> 1:15:17.339 which is that if you think that the leverage is not being set by 1:15:17.342 --> 1:15:20.272 supply and demand properly--too much in good times and not 1:15:20.267 --> 1:15:23.287 enough in bad times--then I would have thought--maybe you're 1:15:23.294 --> 1:15:25.454 saving this for the question period, 1:15:25.449 --> 1:15:28.659 so I'm giving you a chance to say it--then you would think 1:15:28.655 --> 1:15:31.295 that regulators ought to intervene to change the 1:15:31.298 --> 1:15:32.818 leverage. So for example, 1:15:32.820 --> 1:15:35.980 in good times they can prevent people from leveraging so much 1:15:35.982 --> 1:15:39.202 and in bad times--it's a little more difficult to see what you 1:15:39.197 --> 1:15:41.837 can do, but particularly what I want 1:15:41.839 --> 1:15:45.789 you to talk about--how can the Fed intervene in a way to 1:15:45.786 --> 1:15:47.576 increase the leverage? 1:15:47.579 --> 1:15:51.269 For example, should it directly lend against 1:15:51.265 --> 1:15:53.845 collateral. But, they'll lend more money 1:15:53.850 --> 1:15:56.530 against collateral than the market is willing to do, 1:15:56.529 --> 1:15:58.419 not lend at lower interest rates. 1:15:58.420 --> 1:16:01.510 But, just staying on your mortgage derivative, 1:16:01.507 --> 1:16:04.317 which the market will only lend $50 on, 1:16:04.319 --> 1:16:06.579 the Fed will take that as collateral, lend it the same 1:16:06.583 --> 1:16:09.843 interest rate the market will, but lend $60 on it instead of 1:16:09.835 --> 1:16:10.945 $50. Do you think the Fed should do 1:16:10.950 --> 1:16:11.900 that and if it has been doing that? 1:16:11.900 --> 1:16:15.250 That's the first question. 1:16:15.250 --> 1:16:18.920 The second question is, you talked about the housing 1:16:18.924 --> 1:16:23.324 crisis and there's 50% of houses and people may be [Inaudible] 1:16:23.319 --> 1:16:27.639 subprime borrowers--that's 2.5 million of them that have been 1:16:27.642 --> 1:16:30.022 tossed out of their houses. 1:16:30.020 --> 1:16:33.300 You said at the same time that you think innovation sometimes 1:16:33.296 --> 1:16:35.476 prolongs a problem and makes it worse. 1:16:35.479 --> 1:16:39.339 But getting rid of this $180 billion that you spoke of is not 1:16:39.343 --> 1:16:43.013 going to make much of a dent in 2.5 million families being 1:16:43.013 --> 1:16:45.013 tossed out of their houses. 1:16:45.010 --> 1:16:47.040 So, do you think there should be some more, 1:16:47.044 --> 1:16:48.114 bigger intervention? 1:16:48.109 --> 1:16:50.969 One that I would suggest is the people doing the lending are the 1:16:50.967 --> 1:16:53.727 ones who are going to lose all the money when these people are 1:16:53.734 --> 1:16:55.144 thrown out of their houses. 1:16:55.140 --> 1:16:58.810 They're the ones with the biggest interest to seeing 1:16:58.807 --> 1:17:01.837 something done. They're paralyzed because the 1:17:01.839 --> 1:17:05.149 bondholders are split up so completely that they can't 1:17:05.146 --> 1:17:07.756 coordinate. They're remodifying the terms 1:17:07.761 --> 1:17:11.491 of the loan and no single banker would let 2.5 million families 1:17:11.485 --> 1:17:15.205 be thrown out on the streets and have their houses sold for 50% 1:17:15.208 --> 1:17:18.168 of the loan. Can you think of a way of 1:17:18.174 --> 1:17:23.234 coordinating these people--these lenders--who are uncoordinated, 1:17:23.229 --> 1:17:25.769 so that they themselves, without the government having 1:17:25.766 --> 1:17:31.026 to put up any money, can help solve the problem? 1:17:31.029 --> 1:17:36.029 Professor Larry Summers: There are two characteristically 1:17:36.030 --> 1:17:40.070 good questions. On the first question, 1:17:40.072 --> 1:17:47.662 which was about pro-cyclical leverage--as I tried to explain, 1:17:47.659 --> 1:17:52.189 I'm not sure the focus on leverage is right rather than 1:17:52.194 --> 1:17:54.634 the focus being on lending. 1:17:54.630 --> 1:17:56.910 Those are quite different things. 1:17:56.909 --> 1:18:01.579 I actually think that the world is more risky; 1:18:01.579 --> 1:18:05.959 the uncertainty surrounding what's going to happen to loans 1:18:05.956 --> 1:18:09.796 is greater than it was a year--certainly than people 1:18:09.804 --> 1:18:13.204 perceived it to be a year and a half ago. 1:18:13.199 --> 1:18:18.299 So, I'm not sure institutions should have much higher leverage 1:18:18.300 --> 1:18:21.980 today than they did a year and a half ago. 1:18:21.979 --> 1:18:27.299 That's why I put emphasis on what I think is the right thing. 1:18:27.300 --> 1:18:30.030 I'm not sure they should have more leverage; 1:18:30.029 --> 1:18:33.589 they should have more lending and there's a way of getting to 1:18:33.590 --> 1:18:37.330 more lending without having more leverage, which is raising more 1:18:37.328 --> 1:18:41.268 capital. That's why I emphasized the 1:18:41.272 --> 1:18:47.262 importance of pressuring institutions to raise more 1:18:47.257 --> 1:18:52.917 capital and that's why, without having any terribly 1:18:52.918 --> 1:18:58.678 compelling way to concretize it, I think in our thinking about 1:18:58.675 --> 1:19:04.425 regulation we need to get a bit away from the capital ratio as 1:19:04.432 --> 1:19:10.852 the central variable since what we want is capital raising now, 1:19:10.850 --> 1:19:15.230 not balance sheet shrinking. 1:19:15.229 --> 1:19:20.519 An emphasis on the capital ratio treats those two as 1:19:20.518 --> 1:19:25.598 equivalent when they're not in fact equivalent. 1:19:25.600 --> 1:19:30.670 The one suggestion that has been made, and I have no idea 1:19:30.666 --> 1:19:35.006 whether it actually works out in a viable way, 1:19:35.010 --> 1:19:43.110 is to force institutions to issue some kind of security that 1:19:43.108 --> 1:19:51.478 is a reverse convertible that becomes--that converts from debt 1:19:51.480 --> 1:19:56.010 to equity in difficult times. 1:19:56.010 --> 1:20:00.960 It's hard to see why that security is going to feel a lot 1:20:00.962 --> 1:20:05.032 like debt when you buy it in the good times, 1:20:05.029 --> 1:20:07.879 so I'm not sure whether that works out, but that's the type 1:20:07.882 --> 1:20:09.802 of thing that one has to think about. 1:20:09.800 --> 1:20:19.510 I would resist the idea of framing the question through the 1:20:19.513 --> 1:20:26.193 leverage ratio. Should the government try in 1:20:26.188 --> 1:20:34.388 good times to require levels of capital or limits on leverage 1:20:34.390 --> 1:20:42.320 that go beyond what appears to be prudential for individual 1:20:42.320 --> 1:20:50.250 institutions in the name of protection of the system? 1:20:50.250 --> 1:20:56.740 I'm probably one big financial mess away from being in favor of 1:20:56.737 --> 1:21:00.827 that. Not quite there because I think 1:21:00.831 --> 1:21:07.451 you've got to be very careful about the adverse effects if you 1:21:07.446 --> 1:21:10.696 don't do it very carefully. 1:21:10.699 --> 1:21:13.649 Once you're talking about a punitive capital requirement 1:21:13.646 --> 1:21:16.856 that's not related to your institution's prudential needs, 1:21:16.859 --> 1:21:19.559 the incentive to avoid it is very strong. 1:21:19.560 --> 1:21:22.660 So, if you don't do it for every institution, 1:21:22.664 --> 1:21:26.764 the action is going to migrate away and if you do do it for 1:21:26.757 --> 1:21:30.727 every institution, you're getting to a pretty 1:21:30.725 --> 1:21:34.625 dirigiste place, in which the government is 1:21:34.626 --> 1:21:40.286 regulating in a lot of places where it's probably not going to 1:21:40.293 --> 1:21:43.083 be so good at regulating. 1:21:43.079 --> 1:21:47.529 Your other question was about the foreclosure and mortgage 1:21:47.529 --> 1:21:50.799 problem. I think one of the easiest 1:21:50.800 --> 1:21:55.110 judgments to make, in the sense that there's a 1:21:55.112 --> 1:22:00.762 measure available that has a reasonable chance of doing good 1:22:00.764 --> 1:22:04.824 and has, I think, roughly no chance of 1:22:04.822 --> 1:22:10.062 doing bad, is the legislation that's been introduced by 1:22:10.058 --> 1:22:15.678 Senator McCaskill and others to remove legal liability from 1:22:15.682 --> 1:22:21.602 servicers for any decisions they make to renegotiate the total 1:22:21.597 --> 1:22:24.697 value of these mortgages. 1:22:24.699 --> 1:22:27.579 I think there's considerable doubt as to how much 1:22:27.579 --> 1:22:31.179 activity--how much positive activity--that would generate, 1:22:31.180 --> 1:22:38.010 but there's no downside and potentially some upside. 1:22:38.010 --> 1:22:42.080 There's also another interesting problem in this 1:22:42.079 --> 1:22:45.629 area, which is many subprime mortgages, 1:22:45.630 --> 1:22:51.030 probably more than a third, have second liens against them 1:22:51.026 --> 1:22:55.946 that are serviced by somebody other than the original 1:22:55.949 --> 1:22:59.479 originator. Anybody's good plan for working 1:22:59.483 --> 1:23:03.503 these things through where you reduce the value of the house 1:23:03.504 --> 1:23:06.444 is, in effect, going to wipe out that second 1:23:06.435 --> 1:23:09.425 mortgage. But, it is going to require the 1:23:09.426 --> 1:23:13.566 consent of that second mortgage and that second mortgage doesn't 1:23:13.565 --> 1:23:16.845 have any very strong incentive to give consent. 1:23:16.850 --> 1:23:22.720 There's a long and dubiously honorable financial tradition of 1:23:22.721 --> 1:23:25.561 being paid to be a holdout. 1:23:25.560 --> 1:23:31.090 How one handles that is a potentially serious problem, 1:23:31.085 --> 1:23:36.295 not least because it's financially and economically 1:23:36.299 --> 1:23:39.009 advantageous to do so. 1:23:39.010 --> 1:23:43.420 People who are good at being holdouts are likely to acquire a 1:23:43.420 --> 1:23:47.390 substantial number of these second mortgages and so the 1:23:47.389 --> 1:23:52.019 holdout rights will be exercised more skillfully than they would 1:23:52.020 --> 1:23:54.520 be by their current holders. 1:23:54.520 --> 1:24:01.610 I guess my reaction is--to your question is--can the capital 1:24:01.607 --> 1:24:04.007 markets handle it? 1:24:04.010 --> 1:24:08.470 Yes, at a price and the price will be that de facto in 1:24:08.468 --> 1:24:09.878 this environment. 1:24:09.880 --> 1:24:13.940 The demand curve for equity slopes downward, 1:24:13.944 --> 1:24:18.014 so if a lot of equity needs to be raised, 1:24:18.010 --> 1:24:25.280 the price will be lower than if a small amount of equity needs 1:24:25.281 --> 1:24:29.961 to be raised. So, shareowners of these 1:24:29.964 --> 1:24:37.444 institutions are not wrong to be hoping to reduce the amount of 1:24:37.441 --> 1:24:41.061 capital they have to raise. 1:24:41.060 --> 1:24:48.290 That is rational behavior on their part, but it may not be 1:24:48.287 --> 1:24:54.497 desirable from the point of view of the system. 1:24:54.500 --> 1:24:56.060 Student: Then the question would be, 1:24:56.060 --> 1:24:58.250 I guess, where would they have to take capital from in order 1:24:58.251 --> 1:24:59.441 for lenders to do something? 1:24:59.439 --> 1:25:03.409 Professor Larry Summers: I don't think we have a--I 1:25:03.406 --> 1:25:07.296 think, with the trillions of dollars sitting in sovereign 1:25:07.303 --> 1:25:11.483 wealth funds with a very large volume of assets that are held 1:25:11.479 --> 1:25:14.509 by pension funds, and endowments, 1:25:14.510 --> 1:25:20.230 and life insurance companies, I don't think one needs to fear 1:25:20.232 --> 1:25:26.052 that if $200 billion more in capital was infused into U.S. 1:25:26.050 --> 1:25:29.710 financial institutions that somehow there would be a 1:25:29.707 --> 1:25:34.007 generalized--relative prices would adjust in various ways. 1:25:34.010 --> 1:25:39.620 But, I think the least of the concerns is that somehow we'd 1:25:39.616 --> 1:25:45.026 have a general shortage of capital for other purposes. 1:25:45.029 --> 1:25:49.699 I think there is a--this is a tricky subject, 1:25:49.701 --> 1:25:55.011 in the sense that everybody's got the same idea, 1:25:55.010 --> 1:25:59.490 which is that they want to be the last dollar of equity in 1:25:59.485 --> 1:26:04.115 because you want to be in after the dilutions rather than in 1:26:04.117 --> 1:26:06.077 before the dilutions. 1:26:06.079 --> 1:26:11.399 So, there's the concern that, when you speak of large needs 1:26:11.400 --> 1:26:17.180 for capital, that the incentive to be a partial solution may not 1:26:17.178 --> 1:26:20.158 be that great. On the other hand, 1:26:20.160 --> 1:26:24.230 the strategy of sort of deception--we only need a little 1:26:24.226 --> 1:26:28.876 and then we'll be done--also is probably not a strategy that can 1:26:28.883 --> 1:26:33.543 be carried on very successfully for a long period of time. 1:26:33.539 --> 1:26:37.699 That's why somewhat more comprehensive, 1:26:37.704 --> 1:26:43.294 somewhat larger approaches to capital raising would, 1:26:43.293 --> 1:26:46.913 it seems to me, be desirable. 1:26:46.910 --> 1:26:50.070 Gus? Student: You were 1:26:50.065 --> 1:26:54.955 mentioning that additions to the agreed short-term stimulus 1:26:54.961 --> 1:26:57.241 package might be warranted. 1:26:57.240 --> 1:27:01.630 What would you include in such a second package? 1:27:01.630 --> 1:27:07.570 Professor Larry Summers: I suspect there should be a 1:27:07.567 --> 1:27:12.987 contingency plan that is in place to give more rebates 1:27:12.993 --> 1:27:18.833 during the tax filing season next spring that somebody can 1:27:18.828 --> 1:27:25.278 decide to pull the trigger on or not to pull the trigger on late 1:27:25.277 --> 1:27:29.497 next fall, as we see where the economy is. 1:27:29.500 --> 1:27:34.860 There should be some supplementation of unemployment 1:27:34.864 --> 1:27:40.334 insurance and food stamps and there should be--almost 1:27:40.333 --> 1:27:46.333 certainly be--some effort to reduce the pro-cyclicality of 1:27:46.329 --> 1:27:51.899 the state and local sector, which because of balanced 1:27:51.895 --> 1:27:55.915 budgets and less tax collections, spending of all 1:27:55.917 --> 1:27:58.177 kinds is being cut back. 1:27:58.180 --> 1:28:01.790 I say "almost certainly" because there is a bit of a 1:28:01.787 --> 1:28:06.237 problem in the history of fiscal federalism in the United States 1:28:06.243 --> 1:28:11.213 in that whenever times are good, governors cut taxes and 1:28:11.214 --> 1:28:16.914 whenever times are bad, the federal government provides 1:28:16.908 --> 1:28:21.018 support to respond to the emergency. 1:28:21.020 --> 1:28:24.650 That's not a terribly healthy set of arrangements and there 1:28:24.650 --> 1:28:28.030 was a lot of tax cutting going on in the good times. 1:28:28.029 --> 1:28:30.339 But on balance, I think there are a variety of 1:28:30.343 --> 1:28:32.613 mechanisms, including Medicaid reimbursement, 1:28:32.605 --> 1:28:34.795 for doing it; but, doing something for state 1:28:34.795 --> 1:28:37.115 and local governments would be the other piece I would 1:28:37.119 --> 1:28:38.859 emphasize. Yeah? 1:28:38.859 --> 1:28:42.889 Student: What do you think is the main cause of the 1:28:42.888 --> 1:28:54.128 housing crisis? Why did it happen? 1:28:54.130 --> 1:29:00.070 Professor Larry Summers: I think most of positive 1:29:00.067 --> 1:29:05.787 feedback mechanisms that I talked about yesterday were 1:29:05.788 --> 1:29:11.828 operative both going up and going down in the housing and 1:29:11.833 --> 1:29:15.183 mortgage security market. 1:29:15.180 --> 1:29:20.290 It's not that you can't understand why people made these 1:29:20.287 --> 1:29:23.657 loans. The truth is that in the period 1:29:23.664 --> 1:29:28.514 from the late 1990s through 2006, no matter how crummy the 1:29:28.510 --> 1:29:33.780 person--no matter how crummy the credit of the person--you lent 1:29:33.779 --> 1:29:37.769 money to, you did not get defaulted on. 1:29:37.770 --> 1:29:41.310 The reason was that if he bought a house, 1:29:41.311 --> 1:29:46.531 the house went up in value and he was able to borrow more to 1:29:46.534 --> 1:29:50.174 pay you back. So, the lesson that was learned 1:29:50.168 --> 1:29:53.938 was that lending money to people, even people who didn't 1:29:53.944 --> 1:29:56.764 look like they'd be very good credits, 1:29:56.760 --> 1:30:01.010 was a safe thing to do. 1:30:01.010 --> 1:30:03.870 Now, ex post, we can see what the error in 1:30:03.867 --> 1:30:07.137 the inference was--that it presumed the continued strong 1:30:07.140 --> 1:30:11.010 behavior of house prices--but that wasn't completely obvious. 1:30:11.010 --> 1:30:15.570 That was less completely obvious at the time and then you 1:30:15.570 --> 1:30:17.770 had various other things. 1:30:17.770 --> 1:30:21.220 People made the investments and they--if you were in the 1:30:21.223 --> 1:30:23.863 mortgage business, you bought mortgages; 1:30:23.860 --> 1:30:27.190 you succeeded; your mortgages went up in value 1:30:27.185 --> 1:30:28.875 as risk premiums went down. 1:30:28.880 --> 1:30:32.200 Then, you had more capital; you got paid on the basis of 1:30:32.204 --> 1:30:33.904 how much you invested. 1:30:33.899 --> 1:30:39.009 With more capital, you decided to invest more. 1:30:39.010 --> 1:30:42.120 So, you had all the kind of positive feedback mechanisms I 1:30:42.122 --> 1:30:45.512 described on the upside and then you more or less had all those 1:30:45.507 --> 1:30:47.197 mechanisms on the downside. 1:30:47.199 --> 1:31:19.509 Student: But isn't that an issue for regulators? 1:31:19.510 --> 1:31:23.610 Professor Larry Summers: Those of an anti--those of a 1:31:23.605 --> 1:31:28.365 very strong, if you like, Chicago School bent make the 1:31:28.367 --> 1:31:33.727 point, so let's see who did well in this situation; 1:31:33.729 --> 1:31:35.589 let's see who did badly, let's see who did reasonably 1:31:35.593 --> 1:31:36.493 well in this situation. 1:31:36.489 --> 1:31:38.449 Hedge funds--nobody much regulates them; 1:31:38.450 --> 1:31:41.990 they seem to have done okay. 1:31:41.989 --> 1:31:46.049 Individuals--some of them got to--some of them got in trouble 1:31:46.048 --> 1:31:49.568 but some of them got loans on terms that weren't very 1:31:49.565 --> 1:31:51.735 attractive; that wasn't so bad. 1:31:51.740 --> 1:31:54.960 Who really made a huge mess? 1:31:54.960 --> 1:31:57.610 Well, the big money center banks made a huge mess. 1:31:57.609 --> 1:31:59.449 Who's most extensively regulated? 1:31:59.449 --> 1:32:02.049 Well, the big money center banks are most extensively 1:32:02.050 --> 1:32:04.830 regulated. I get it; we need more 1:32:04.825 --> 1:32:10.455 regulation--is this sort of Chicago argument and it should 1:32:10.461 --> 1:32:15.901 give one--I don't think it's the end of the subject. 1:32:15.899 --> 1:32:21.699 I think, it probably doesn't point in the right direction at 1:32:21.703 --> 1:32:26.133 the end of the day, but it's not a point to be 1:32:26.129 --> 1:32:30.139 taken lightly. Since I like the spirit of your 1:32:30.141 --> 1:32:34.121 observation, here's another observation in the spirit of 1:32:34.119 --> 1:32:37.289 yours. So, all these banks lost a lot 1:32:37.289 --> 1:32:40.389 of money; who made money? 1:32:40.390 --> 1:32:46.450 Well, a ton of the gain that is opposite the banks loss is a 1:32:46.453 --> 1:32:52.213 bunch of relatively poor people who sold their houses for 1:32:52.208 --> 1:32:58.268 substantially more than they otherwise would have because of 1:32:58.271 --> 1:33:03.921 the bad credit that the financial sector was prepared to 1:33:03.924 --> 1:33:07.794 give. I think the thing you have to 1:33:07.792 --> 1:33:12.952 say is that there are a fair number of people who are having 1:33:12.945 --> 1:33:18.355 to explain to their children why a policeman is knocking at the 1:33:18.360 --> 1:33:23.250 door and telling them they have to leave within six hours 1:33:23.251 --> 1:33:28.491 because they signed documents that they didn't understand and 1:33:28.491 --> 1:33:33.301 that weren't explained to them--that locked them into an 1:33:33.295 --> 1:33:38.765 arrangement that was, for them, going to be highly 1:33:38.767 --> 1:33:41.427 problematic. You can call it predatory 1:33:41.426 --> 1:33:44.536 borrowing and you can explain how they lived in a house, 1:33:44.543 --> 1:33:47.833 in a sense, rent-free for a significant period of time; 1:33:47.829 --> 1:33:55.759 but, I think it would be mistake to discount the trauma 1:33:55.763 --> 1:33:59.293 that has been imposed. 1:33:59.289 --> 1:34:03.629 But, there is a sense in which more regulated entities have 1:34:03.632 --> 1:34:07.532 performed worse and suffered more than less regulated 1:34:07.526 --> 1:34:10.066 entities. Last question? 1:34:10.069 --> 1:35:03.329 Student: How does our current crisis compare with 1:35:03.331 --> 1:35:28.511 Japan's since the 1990s? 1:35:28.510 --> 1:35:33.240 Professor Larry Summers: Look, I think there are 1:35:33.239 --> 1:35:39.019 structural similarities between our problem and Japan's problem. 1:35:39.020 --> 1:35:41.450 I think there are a variety of reasons to think that the bubble 1:35:41.449 --> 1:35:42.899 was substantially greater in Japan. 1:35:42.899 --> 1:35:45.519 I think, if one compares U.S. 1:35:45.517 --> 1:35:50.747 policy in the last nine months with Japanese policy between 1:35:50.753 --> 1:35:53.643 1989 and 1992, we are acting with 1:35:53.641 --> 1:35:57.971 substantially more aggressiveness to get in front 1:35:57.974 --> 1:36:00.374 of the problem. U.S. 1:36:00.365 --> 1:36:06.245 regulators, fortunately from my perspective, have been very 1:36:06.245 --> 1:36:11.005 resistant to what was a major Japanese idea, 1:36:11.010 --> 1:36:14.790 which was mark assets in ways that will--that it is thought 1:36:14.790 --> 1:36:18.830 will--contribute to confidence rather than in ways that reflect 1:36:18.831 --> 1:36:20.201 market realities. 1:36:20.199 --> 1:36:24.799 I think that the rather arbitrary accounting that was 1:36:24.804 --> 1:36:29.504 pursued as a major tool of policy in Japan inhibited a 1:36:29.496 --> 1:36:32.946 great deal of necessary adjustment. 1:36:32.949 --> 1:36:38.099 That said, while I think the probability is low, 1:36:38.099 --> 1:36:44.339 one can't rule out the kind of situation that developed in 1:36:44.344 --> 1:36:48.464 Japan. I think we do--I think these 1:36:48.460 --> 1:36:54.930 issues of levels of capital in troubled financial institution 1:36:54.931 --> 1:36:58.061 are very important issues. 1:36:58.060 --> 1:37:06.010 One way of responding to an externality is with a subsidy. 1:37:06.010 --> 1:37:09.760 The idea of subsidizing institutions to raise capital 1:37:09.756 --> 1:37:13.786 seems quite untenable in the context of the institution's 1:37:13.791 --> 1:37:16.891 shareholders holding onto their shares. 1:37:16.890 --> 1:37:20.350 So in a way, we operate a system where the 1:37:20.347 --> 1:37:25.407 government can't infuse capital into an institution until the 1:37:25.406 --> 1:37:29.846 situation is very far gone, but once the situation is very 1:37:29.849 --> 1:37:31.739 far gone, it is very far gone. 1:37:31.739 --> 1:37:35.259 I don't think it's inconceivable that the 1:37:35.260 --> 1:37:40.540 government will have to infuse capital into institutions on a 1:37:40.541 --> 1:37:45.031 significant scale in the course of this crisis; 1:37:45.029 --> 1:37:50.289 but, I don't think we're ready yet to start planning for Nordic 1:37:50.289 --> 1:37:55.949 style or Mexican style, or Japanese style socialization 1:37:55.951 --> 1:38:02.241 of the financial system, though I wouldn't--I think that 1:38:02.235 --> 1:38:05.315 probability is very low. 1:38:05.319 --> 1:38:10.289 I wouldn't preclude that as a possibility and I think, 1:38:10.291 --> 1:38:14.421 in retrospect, it came significantly later in 1:38:14.418 --> 1:38:17.328 Japan than would, in retrospect, 1:38:17.326 --> 1:38:20.996 have been ideal. Thank you very much.