[00:56.030 --> 01:01.310] This is Paul Graham. He's going to be speaking on inequality and risk. [01:04.570 --> 01:06.050] That was quick. [01:07.030 --> 01:08.290] Hey, guys. [01:11.600 --> 01:14.300] Hello! Well... [01:14.800 --> 01:18.680] The last speaker in here trained the audience to chant, [01:18.680 --> 01:22.980] fuck you, in unison, whenever authority commanded them to do anything. [01:23.100 --> 01:24.940] Then at the end of the talk, everybody said, [01:24.940 --> 01:26.940] okay, talk's over, and the guys in the red shirts said, [01:26.940 --> 01:30.460] everybody out that into the room, and they just filed out as meekly as sheep. [01:33.460 --> 01:36.260] All right, so... [01:40.670 --> 01:42.530] Are we ready to start? [01:43.210 --> 01:45.610] I think we're ready to start. All right, let's go. [01:47.870 --> 01:50.830] Suppose you wanted to get rid of economic inequality. [01:50.830 --> 01:52.570] By the way, I have no slides. [01:55.370 --> 01:57.710] Asking is what I do for talks. [02:00.230 --> 02:03.070] Suppose you wanted to get rid of economic inequality. [02:03.570 --> 02:05.870] There are two ways to do it, basically. [02:05.950 --> 02:09.550] You can give money to the poor, or take it away from the rich. [02:09.970 --> 02:12.250] But they both amount to the same thing. [02:12.370 --> 02:15.930] Because if you want to give money to the poor, you've got to get it from somewhere. [02:16.750 --> 02:19.710] And if you take it from the poor, they end up just where they started. [02:19.810 --> 02:22.130] So you basically have to take it from the rich. [02:22.730 --> 02:25.750] There is, of course, a way to make the poor richer, [02:25.750 --> 02:28.950] without merely shifting money from the rich to them. [02:29.210 --> 02:32.450] You could make the poor more productive, for example, [02:32.450 --> 02:36.210] by giving everybody equal access to education. [02:36.790 --> 02:40.810] That is, instead of taking money from engineers and giving it to checkout clerks, [02:40.810 --> 02:45.150] you could enable the people who would have become checkout clerks to become engineers. [02:45.250 --> 02:47.390] Well, that works. [02:47.750 --> 02:51.030] That's an excellent strategy for making the poor richer. [02:51.330 --> 02:53.770] But the evidence of the last 200 years [02:54.270 --> 02:59.270] shows that it doesn't reduce economic inequality, [02:59.270 --> 03:01.810] because it tends to make the rich richer, too. [03:02.490 --> 03:08.350] If there are more engineers, then there are more opportunities to hire them, to sell them things. [03:08.630 --> 03:11.490] Henry Ford couldn't have made a fortune making cars [03:11.490 --> 03:16.010] in a society in which people were all still subsistence farmers, [03:16.010 --> 03:19.510] because there would have been no one to buy his products and no one to build them. [03:19.510 --> 03:23.310] If you want to reduce economic inequality, [03:23.310 --> 03:26.570] instead of just improving the overall standard of living, [03:26.570 --> 03:29.350] it's not enough just to raise up the poor. [03:29.650 --> 03:34.190] What if one of your newly minted engineers goes on to become the next Bill Gates? [03:34.350 --> 03:37.090] Economic inequality will be as bad as ever. [03:38.430 --> 03:39.990] Whoa, dudes! [03:40.970 --> 03:42.470] Yes, go ahead. [03:44.630 --> 03:46.470] Really? Can people not hear? [03:46.470 --> 03:47.850] Oh, God. [03:51.950 --> 03:54.690] This talk, Inequality and Risk. [03:54.690 --> 03:56.190] Are you guys in the wrong place? [03:57.310 --> 04:00.910] Am I? I'm where they brought me. [04:11.130 --> 04:14.250] There's been a rescheduling. That's at 2 o'clock. [04:14.250 --> 04:15.970] This is Paul Graham. [04:18.810 --> 04:20.990] It's all on the website. [04:25.660 --> 04:27.220] Okay, guys. [04:27.220 --> 04:32.180] I was supposed to talk today, [04:32.180 --> 04:35.180] and then they switched me to yesterday, but they didn't tell me. [04:35.180 --> 04:39.060] So, I look on the website yesterday to find out where the hotel is, [04:39.060 --> 04:44.740] and I'm astonished to discover I'm supposed to speak four hours before my plane arrives. [04:46.240 --> 04:48.940] So, they put me in this lot today. [04:50.280 --> 04:52.700] I'm also not even in the printed thing. [04:53.940 --> 04:56.480] I'm on the website. There's a description of the talk. [04:56.480 --> 04:58.800] Let me tell you what the talk's going to be about. [05:00.480 --> 05:01.600] Sorry, what? [05:02.840 --> 05:04.000] What's the talk about? [05:04.100 --> 05:07.520] I'll tell you what the talk's about, and you can tell me. [05:07.540 --> 05:10.400] Well, actually, you can decide if you want to hear it. [05:11.380 --> 05:13.380] What did the abstract say? [05:14.180 --> 05:16.840] The abstract was about, [05:16.840 --> 05:21.960] where's the right place to hack the connection between power and wealth? [05:24.460 --> 05:26.460] Alright, so if that sounds interesting, [05:26.460 --> 05:30.240] there is no hardware component to this talk. [05:36.120 --> 05:37.740] Out you go. [05:39.980 --> 05:42.160] No, no, no, I totally understand. [05:42.160 --> 05:45.660] Believe me, I feel the same way you guys do. [05:53.090 --> 05:55.590] After these guys all leave, [05:55.590 --> 05:58.870] I'll tell you all the secret behind-the-scenes stories [05:58.870 --> 06:03.330] of the creation and launching of the internet worm of 1988. [06:05.390 --> 06:07.010] Just kidding. [06:10.150 --> 06:12.050] Robert would kill me. [06:16.500 --> 06:17.560] Alright. [06:18.840 --> 06:20.560] Let's start again. [06:20.960 --> 06:24.040] I feel like what doesn't kill you makes you stronger. [06:24.940 --> 06:26.320] And in the speaker department, [06:26.320 --> 06:28.600] this is definitely one of those experiences. [06:31.380 --> 06:33.660] So, inequality and risk. [06:33.660 --> 06:35.620] Does everybody remember where we are? [06:35.960 --> 06:38.980] I'll start somewhat over. [06:40.220 --> 06:41.120] Alright. [06:42.660 --> 06:44.260] Bill Gates. [06:44.520 --> 06:45.760] That's right. [06:49.900 --> 06:52.920] If you want to reduce economic inequality, [06:52.920 --> 06:56.300] instead of just improving the overall standard of living, [06:56.300 --> 06:58.620] it's not enough just to raise up the poor, [06:58.620 --> 07:00.820] because what if one of your newly minted engineers [07:00.820 --> 07:03.180] goes on to become the next Bill Gates? [07:03.380 --> 07:05.400] Or maybe in this case, Larry Ellison. [07:05.720 --> 07:08.320] Economic inequality will be as bad as ever. [07:09.000 --> 07:12.980] If you actually want to compress the gap between rich and poor, [07:12.980 --> 07:15.800] it's not enough to just push up on the bottom, [07:15.800 --> 07:18.240] you also have to push down on the top. [07:19.020 --> 07:21.920] So, how do you push down on the top? [07:22.180 --> 07:24.560] Well, you could try to decrease the productivity [07:24.560 --> 07:26.800] of the people who make the most money. [07:27.300 --> 07:31.480] You could make all the best surgeons operate with their left hands, [07:31.480 --> 07:34.140] force the most popular actors to overeat, [07:34.140 --> 07:35.260] and so on. [07:35.600 --> 07:39.700] But, this approach has implementation difficulties. [07:40.580 --> 07:42.960] It seems like the only practical approach [07:42.960 --> 07:46.140] is to let people do the best work they can, [07:46.140 --> 07:50.160] and then, either by taxation or by limiting what they can charge, [07:50.160 --> 07:52.980] to confiscate whatever you deem to be surplus. [07:53.920 --> 07:58.020] So, let's be clear what reducing economic inequality means. [07:58.020 --> 08:00.860] It's identical with taking money from the rich. [08:01.380 --> 08:05.340] When you transform a mathematical expression into a new form, [08:05.340 --> 08:07.720] you often notice new things about it. [08:07.720 --> 08:09.600] So it is in this case. [08:10.300 --> 08:13.080] Taking money from the rich turns out to have consequences [08:13.080 --> 08:16.800] you might not foresee when you phrase the same idea [08:16.800 --> 08:19.660] in terms of reducing economic inequality. [08:20.560 --> 08:25.920] The problem is, risk and reward have to be proportionate. [08:26.160 --> 08:28.740] A bet with only a 10% chance of winning [08:28.740 --> 08:30.480] has to pay a lot more, [08:30.480 --> 08:33.060] in fact, somewhat more than 5 times more, [08:33.060 --> 08:35.620] than a bet with a 50% chance of winning. [08:35.940 --> 08:39.320] So if you lop off the top of the possible rewards, [08:39.320 --> 08:42.360] you decrease people's willingness to take risks. [08:43.480 --> 08:47.780] Transposing into our original expression we get, [08:47.780 --> 08:49.640] decreasing economic inequality [08:49.640 --> 08:54.740] means decreasing the risk people are willing to take. [08:55.400 --> 08:57.580] There are whole classes of risks [08:57.580 --> 08:59.120] that are no longer worth taking [08:59.460 --> 09:02.220] if the maximum return is decreased. [09:02.420 --> 09:05.400] One reason high taxes are so disastrous, [09:05.400 --> 09:08.060] as the English found in the 1970s, [09:08.900 --> 09:12.840] is that this class of risks includes starting new companies. [09:13.700 --> 09:16.720] Startups are intrinsically risky. [09:16.720 --> 09:19.420] That's just what the deal is. [09:19.420 --> 09:23.060] A startup is like a small boat in the open sea, [09:23.060 --> 09:25.240] one big wave and you're sunk. [09:25.440 --> 09:28.680] A competing product, a downturn in the economy, [09:28.780 --> 09:31.700] a delay in getting funding or regulatory approval, [09:31.700 --> 09:34.180] the loss of a big account, a patent suit, [09:34.180 --> 09:35.940] changing technical standards, [09:35.940 --> 09:38.280] any one of these can sink you overnight. [09:39.020 --> 09:41.200] It seems like the odds of a startup succeeding [09:41.200 --> 09:44.700] are a little worse than 1 in 10, which is not very good. [09:44.700 --> 09:49.200] Our startup paid its first round of investors 36x, [09:49.200 --> 09:51.780] which meant, with current US tax rates, [09:51.780 --> 09:53.740] it made sense to invest in us [09:53.740 --> 09:58.100] if it seemed like we had a better than 1 in 18 chance of succeeding. [09:58.440 --> 10:00.080] That's probably roughly how we looked [10:00.080 --> 10:03.400] when we were a pair of nerds operating out of an apartment. [10:03.960 --> 10:06.060] If that kind of risk doesn't pay, [10:06.060 --> 10:08.660] venture investing as we know it doesn't happen. [10:09.120 --> 10:10.720] Now, that might be okay [10:10.720 --> 10:14.160] if there were other sources of capital for startups. [10:14.440 --> 10:16.900] So, for example, why not just have the government [10:16.900 --> 10:20.320] or some almost government organization like Fannie Mae [10:20.320 --> 10:23.960] be in charge of supplying venture capital for new companies? [10:24.400 --> 10:26.560] Well, I'll tell you why that wouldn't work. [10:26.740 --> 10:28.620] Because then you're asking the government [10:28.620 --> 10:32.000] or almost government employees to do the one thing [10:32.000 --> 10:34.960] they're least able to do, take risks. [10:35.700 --> 10:38.620] As anyone who's worked for the government knows, [10:38.620 --> 10:40.100] the important thing in the government [10:40.100 --> 10:42.960] is not to make the right choices. [10:43.300 --> 10:47.980] It's to make choices that can be justified later if they fail. [10:49.220 --> 10:50.820] If there's a safe option, [10:50.820 --> 10:53.120] that's the one a bureaucrat will choose. [10:54.000 --> 10:57.660] But that is exactly the wrong way to do venture investing. [10:57.800 --> 11:00.360] You want to take the most risky choices you can find [11:00.360 --> 11:02.440] if the upside looks good enough. [11:03.580 --> 11:05.760] VCs currently get paid in a way [11:05.760 --> 11:07.820] that makes them focus on the upside. [11:07.820 --> 11:10.260] They get a percentage of the fund's gains. [11:10.260 --> 11:12.560] And this is what enables them to overcome [11:12.560 --> 11:15.960] their understandable reluctance in investing in a company [11:15.960 --> 11:19.840] that seems to be run by a pair of college students. [11:20.700 --> 11:23.220] If VCs weren't allowed to get rich, [11:23.220 --> 11:25.260] they'd behave like bureaucrats. [11:25.260 --> 11:29.220] Without hope of gain, they'd have only fear of loss. [11:29.760 --> 11:32.760] And so they would make the wrong choices. [11:32.760 --> 11:34.880] They'd turn down the nerds [11:34.880 --> 11:38.260] in favor of the smooth-talking MBA in a business suit [11:38.260 --> 11:42.580] because that choice would be easier to justify later if it failed. [11:43.880 --> 11:46.500] Even if you could somehow redesign venture funding [11:46.500 --> 11:49.540] to work without letting VCs get rich, [11:49.540 --> 11:53.300] there's another kind of investor you really can't do without. [11:53.380 --> 11:56.260] The startup's founders and early employees. [11:56.800 --> 11:59.340] What they invest is their time and ideas. [11:59.340 --> 12:01.300] But these are equivalent to money, [12:01.300 --> 12:05.080] as you can tell from the fact that the VCs, if forced, [12:05.080 --> 12:07.680] will grant the same status to sweat equity [12:07.680 --> 12:11.100] that they do to the equity they've paid for with cold cash. [12:11.800 --> 12:14.420] The fact that you're investing time and ideas [12:14.420 --> 12:17.720] doesn't change the ratio between risk and reward. [12:18.200 --> 12:20.560] If you're going to invest your time in something, [12:20.560 --> 12:24.080] you'll only do it if the payoff is proportionate to the risk. [12:25.800 --> 12:28.280] If large payoffs aren't allowed, [12:28.280 --> 12:30.340] you may as well play it safe. [12:30.340 --> 12:33.100] So, for example, most people who start startups [12:33.100 --> 12:37.140] do it with the hope that as a result, if it pays off, [12:37.140 --> 12:39.720] they'll end up much richer than they were before. [12:39.720 --> 12:42.200] So if your society has some kind of policy [12:42.200 --> 12:46.240] designed to prevent anyone from being much richer than anyone else, [12:46.240 --> 12:48.240] then unless your startup founders were before [12:48.240 --> 12:50.460] the only poor people in the country, [12:50.460 --> 12:53.060] your economic policy is going to be tending [12:53.060 --> 12:55.360] to prevent startups from even happening. [12:55.980 --> 12:58.760] Like many startup founders, I did it to get rich. [12:58.760 --> 13:02.860] Not because I wanted expensive things, but because I wanted security. [13:03.340 --> 13:05.420] If I'd been forbidden to make enough [13:05.420 --> 13:07.420] from a startup to do this, [13:07.420 --> 13:10.400] I would have sought security by some other means. [13:10.460 --> 13:13.860] For example, by going to work for a big, stable organization [13:14.360 --> 13:16.840] from which it would have been hard to get fired. [13:16.840 --> 13:19.420] Instead of busting my ass in a startup, [13:19.420 --> 13:22.280] I would have tried to get a nice, low-stress job [13:22.740 --> 13:25.980] in a big research lab or tenure at a university. [13:26.240 --> 13:30.000] That's what everyone does in societies where risk isn't rewarded. [13:30.020 --> 13:32.560] If you can't ensure your own security, [13:32.560 --> 13:35.240] the next best thing is to make a nest for yourself [13:35.240 --> 13:37.220] in some large organization [13:37.640 --> 13:40.880] where your status depends mostly on seniority. [13:41.240 --> 13:42.800] So it's interesting, actually. [13:42.800 --> 13:46.140] You'll find in societies you either get economic inequality [13:46.140 --> 13:49.820] or a world where seniority counts for a lot. [13:49.820 --> 13:51.940] You'll always get one or the other. [13:55.250 --> 13:58.090] Even if we could somehow replace VCs, [13:58.090 --> 14:00.550] I don't see how we could replace founders. [14:00.550 --> 14:02.350] VCs mainly contribute money, [14:02.350 --> 14:04.910] which in principle is the same no matter where you get it. [14:04.910 --> 14:08.530] Founders contribute ideas. Those are not interchangeable. [14:08.970 --> 14:11.490] So let's rehearse the chain of arguments, [14:11.490 --> 14:14.850] because I'm heading to a conclusion many people will have to be dragged, [14:14.850 --> 14:16.790] kicking, and screaming towards. [14:17.230 --> 14:19.950] I've tried to make every link unbreakable. [14:20.090 --> 14:24.230] Decreasing economic inequality means taking money from the rich. [14:24.230 --> 14:27.150] Since risk and reward are interchangeable, [14:27.150 --> 14:31.090] decreasing potential rewards decreases people's appetite for risk. [14:31.090 --> 14:33.670] Startups are intrinsically risky. [14:33.930 --> 14:37.330] Without the prospect of rewards proportionate to the risk, [14:37.330 --> 14:39.870] founders won't start them. [14:39.870 --> 14:41.990] Founders are irreplaceable. [14:42.030 --> 14:46.950] So eliminating economic inequality means eliminating startups. [14:47.610 --> 14:50.610] This argument applies proportionately. [14:50.610 --> 14:53.730] It's not just that if you eliminate economic inequality, [14:53.730 --> 14:55.290] you get no startups. [14:55.290 --> 14:58.630] To the extent that you reduce economic inequality, [14:58.630 --> 15:02.310] you decrease people's appetite for risk. [15:02.470 --> 15:05.210] And that seems bad for a lot of people. [15:05.210 --> 15:09.530] New technology and new jobs both come disproportionately from new companies. [15:09.530 --> 15:12.110] In fact, if you don't have startups, [15:12.110 --> 15:14.710] you'll find, just as if you didn't have any more children, [15:14.710 --> 15:19.010] you wouldn't have any established companies either after a while. [15:19.290 --> 15:21.190] It sounds benevolent to say [15:21.190 --> 15:23.770] we ought to reduce economic inequality. [15:24.450 --> 15:27.650] But when you phrase it like that, who can argue with you? [15:27.730 --> 15:30.130] Inequality has to be bad, right? [15:31.330 --> 15:33.530] It sounds a good deal less benevolent [15:33.530 --> 15:36.030] to say that we ought to reduce the rate [15:36.030 --> 15:37.970] at which new companies are founded, [15:37.970 --> 15:41.110] and yet the one implies the other. [15:42.390 --> 15:46.110] Indeed, it may be that reducing investors' appetite for risk [15:46.110 --> 15:48.910] doesn't merely kill off larval startups, [15:48.910 --> 15:52.410] but kills off the most promising ones disproportionately. [15:52.990 --> 15:54.770] Startups yield faster growth [15:54.770 --> 15:57.150] at greater risk than big companies. [15:57.550 --> 16:00.930] Does this trend also apply within startups? [16:00.930 --> 16:04.290] Are the riskiest startups also the ones that pay off the most? [16:04.290 --> 16:06.730] I mean for everyone, not just for the investors. [16:07.110 --> 16:09.170] I suspect the answer is yes. [16:09.170 --> 16:11.150] And if so, that is a chilling thought, [16:11.150 --> 16:14.610] because it means if you cut investors' appetite for risk, [16:14.610 --> 16:18.190] the most beneficial startups are the first to go. [16:20.870 --> 16:22.490] Well, not all rich people [16:22.490 --> 16:24.630] get that way from startups, of course. [16:26.630 --> 16:29.790] What if we let people get rich by starting startups, [16:29.790 --> 16:32.710] but taxed away all other surplus wealth? [16:32.710 --> 16:35.070] Wouldn't that at least decrease inequality? [16:36.370 --> 16:38.030] Less than you might think. [16:38.150 --> 16:41.570] If you made it so that people could only get rich by starting startups, [16:41.570 --> 16:44.310] people who wanted to get rich would all start startups. [16:44.570 --> 16:46.570] And that might be a great thing, [16:46.570 --> 16:49.870] but I don't think it would change the distribution of wealth a lot. [16:50.230 --> 16:53.130] People who want to get rich will do whatever they have to do. [16:53.330 --> 16:55.070] If startups are the only way to do it, [16:55.070 --> 16:57.470] you'll just get a lot more people starting startups. [16:57.470 --> 17:00.590] That is, if you write the laws very carefully, [17:00.590 --> 17:03.070] more likely what you'll get is a lot of people doing stuff [17:03.070 --> 17:05.570] that can be described on paper as startups. [17:06.410 --> 17:09.430] If we're determined to eliminate economic inequality, [17:09.430 --> 17:11.070] there's still one way out. [17:11.330 --> 17:14.470] We could say that we're willing to go ahead and do without startups. [17:14.610 --> 17:16.290] What would happen if we did? [17:16.690 --> 17:20.410] At a minimum, we get lower rates of technological growth. [17:20.710 --> 17:22.990] If you believe that large companies [17:23.610 --> 17:27.830] could somehow be made to develop technology as fast as startups, [17:27.830 --> 17:30.510] the ball's in your court to explain how. [17:31.030 --> 17:32.310] The good news is, [17:32.310 --> 17:36.130] if you can come up with even a remotely plausible explanation, [17:36.130 --> 17:40.050] you can make a fortune writing books and consulting for large companies. [17:41.090 --> 17:43.170] Okay, so we get slower growth. [17:43.170 --> 17:44.630] Is that so bad? [17:44.910 --> 17:46.770] Well, one reason it's bad in practice [17:46.770 --> 17:51.550] is that other countries might not agree to slow down and jog alongside us. [17:51.550 --> 17:54.150] If you're content to develop new technologies [17:54.150 --> 17:56.410] at a slower rate than everyone else, [17:56.410 --> 18:00.270] what actually ends up happening is you don't invent anything at all. [18:00.770 --> 18:04.250] Anything you might discover has already been discovered elsewhere. [18:04.390 --> 18:06.670] And the only thing you can offer in return [18:07.390 --> 18:10.730] is cheap labor and raw materials. [18:11.130 --> 18:12.370] Once you sink that low, [18:12.370 --> 18:14.790] other countries can do whatever they want with you. [18:14.990 --> 18:17.390] Install puppet governments, [18:17.390 --> 18:19.990] come and pacify you, [18:19.990 --> 18:21.630] siphon off your best workers, [18:21.630 --> 18:23.110] use your women as prostitutes, [18:23.110 --> 18:25.330] dump their toxic waste on your territory, [18:25.330 --> 18:28.470] all the things we do to poor countries now. [18:29.270 --> 18:32.190] The only defense is to isolate yourself, [18:32.190 --> 18:35.170] as communist countries did in the 20th century. [18:35.370 --> 18:36.850] But the problem then is, [18:36.850 --> 18:40.230] you have to become a police state to enforce the isolation. [18:41.790 --> 18:43.390] Eliminating economic inequality [18:43.770 --> 18:45.390] means eliminating startups, [18:45.390 --> 18:48.150] and to the extent you do that, you're hosed. [18:48.350 --> 18:51.190] I realize startups are not the main target [18:51.190 --> 18:53.730] of those who want to eliminate inequality. [18:53.830 --> 18:56.650] What they really dislike is the sort of wealth [18:56.650 --> 19:00.990] that becomes self-perpetuating through an alliance with political power. [19:00.990 --> 19:04.670] For example, the construction firm that funds a politician's campaign [19:04.670 --> 19:07.610] in return for government contracts. [19:07.610 --> 19:10.610] But if you try to attack this type of wealth [19:10.610 --> 19:12.550] through economic policy, [19:13.470 --> 19:17.830] it's hard to hit without destroying startups as collateral damage. [19:18.390 --> 19:20.490] The problem here is not wealth, [19:20.490 --> 19:21.690] but corruption. [19:21.770 --> 19:24.510] So why not go after corruption directly? [19:25.030 --> 19:27.970] We don't need to prevent people from becoming rich [19:27.970 --> 19:31.810] if we can prevent wealth from translating into power. [19:32.090 --> 19:36.510] And there has been progress on that front. [19:36.510 --> 19:40.250] Before he died of drink in 1925, [19:40.250 --> 19:43.210] Commodore Vanderbilt's wasteful grandson Reggie [19:43.210 --> 19:47.010] ran down pedestrians on five separate occasions, [19:47.010 --> 19:48.550] killing two of them. [19:48.690 --> 19:51.570] By 1969, when Ted Kennedy [19:51.570 --> 19:53.830] drove off the bridge at Chappaquiddick, [19:53.830 --> 19:56.570] the bag limit seemed to be down to one. [19:56.890 --> 19:59.570] Today, it may well be zero. [19:59.730 --> 20:02.770] But what's changed is not variation in wealth. [20:02.770 --> 20:05.930] That's as bad, or as extreme, as ever. [20:06.070 --> 20:09.770] What's changed is the ability to translate wealth into power. [20:09.790 --> 20:12.650] Rich people can't get away with as much, he assured. [20:12.950 --> 20:15.030] How do you break the connection [20:15.030 --> 20:16.850] between wealth and power? [20:17.270 --> 20:19.210] Demand transparency. [20:19.710 --> 20:23.110] Watch closely how power is exercised and demand an account [20:23.110 --> 20:25.390] of how decisions are made. [20:25.570 --> 20:29.230] Why aren't all police interrogations videotaped? [20:29.230 --> 20:33.430] Why did 36% of Princeton's class of 2007 [20:33.430 --> 20:34.670] come from prep schools [20:34.990 --> 20:38.930] when only 1.7% of American kids attend them? [20:39.050 --> 20:40.850] Why did George Bush make it harder [20:40.850 --> 20:43.350] to see previous presidents' papers? [20:43.750 --> 20:46.830] Why does the U.S. so often put Israel's interests [20:46.830 --> 20:48.210] before its own? [20:48.590 --> 20:52.230] Why don't government officials disclose more of their finances? [20:52.230 --> 20:54.490] And why only during their term of office? [20:54.490 --> 20:56.390] Why not for 10 years after? [20:57.110 --> 21:00.210] A friend of mine who knows a lot about computer security [21:01.250 --> 21:03.670] says the single most important step [21:03.670 --> 21:05.510] is to log everything. [21:05.810 --> 21:08.650] Back when he was a kid trying to break into computers, [21:08.650 --> 21:11.950] what worried him most was the idea of leaving a trail. [21:12.410 --> 21:15.490] He was more inconvenienced by the need to avoid that [21:15.490 --> 21:18.810] than by any obstacle deliberately put in his path. [21:19.270 --> 21:21.670] Like all illicit connections, [21:21.670 --> 21:24.970] the connection between wealth and power flourishes in secret. [21:25.350 --> 21:26.790] Expose all transactions [21:27.710 --> 21:30.190] and you will greatly reduce it. [21:30.190 --> 21:33.630] Log everything. That's a strategy that already seems to be working [21:33.630 --> 21:35.870] and it doesn't have the unintended side effect [21:35.870 --> 21:38.570] of making your entire country poor. [21:38.870 --> 21:41.890] I don't think many people realize there's a connection [21:41.890 --> 21:44.450] between economic inequality and risk. [21:44.450 --> 21:47.210] I didn't fully grasp it until recently. [21:47.210 --> 21:49.370] I'd known for years, of course, [21:49.370 --> 21:51.630] that if one didn't score in a startup, [21:51.630 --> 21:55.630] the next best alternative was to get a cozy research position [21:55.630 --> 21:57.810] somewhere where they couldn't fire you. [21:57.810 --> 22:01.330] But I didn't understand the equation governing my behavior. [22:01.790 --> 22:04.330] Likewise, it's obvious empirically that a country [22:04.330 --> 22:07.410] that doesn't let people get rich is headed for disaster, [22:07.410 --> 22:10.990] whether it's Diocletian's Rome or Harold Wilson's Britain. [22:11.430 --> 22:15.310] But I didn't until recently understand the role that risk played. [22:15.870 --> 22:17.850] If you try to attack wealth, [22:17.850 --> 22:21.650] you end up nailing risk as well, and with it, growth. [22:21.990 --> 22:24.110] If you want a fairer world, [22:24.110 --> 22:26.990] I think you're better off attacking one step downstream [22:26.990 --> 22:29.510] where wealth turns into power. [22:30.710 --> 22:33.110] Thank you for listening to this talk. [22:33.110 --> 22:36.330] Sorry about the circumstances. Thank you. [22:45.360 --> 22:46.820] That was hair-raising. [22:46.820 --> 22:48.860] Do we have any time left? Are we done? [22:48.860 --> 22:51.860] Are we done? We are done? [22:52.820 --> 22:54.820] We have two minutes left. [22:54.820 --> 22:57.180] Should anybody have any questions about this talk [22:57.180 --> 23:00.280] that was obviously so enthralling, by all means, let me know. [23:13.630 --> 23:15.010] The question is, [23:15.010 --> 23:18.150] why do you want to reduce economic inequality in the first place? [23:18.310 --> 23:21.090] Well, actually, [23:21.090 --> 23:22.790] I'm saying you don't have to. [23:23.030 --> 23:26.050] But a lot of people seem to think that it's a really bad idea. [23:26.050 --> 23:29.950] I mean, it's one of those things you see in newspaper stories, [23:29.950 --> 23:31.770] not as a statement, [23:31.770 --> 23:33.830] but as an axiom underlying the story. [23:33.830 --> 23:36.050] They say inequality has increased, [23:36.050 --> 23:39.370] as if this were ipso facto bad. [23:39.730 --> 23:41.650] I think it's not necessarily. [23:51.140 --> 23:53.640] If government control can't result in growth, [23:53.640 --> 23:55.420] how do I explain what's happening in China? [23:55.420 --> 23:57.900] What's happening in China is the government has a lot less control [23:57.900 --> 23:59.460] than it did in the time of Mao. [23:59.460 --> 24:01.500] That's why China is growing faster. [24:02.040 --> 24:02.700] Yes? [24:19.370 --> 24:22.030] Oh, okay, that's an interesting question. [24:25.070 --> 24:26.610] I think what you're saying is, [24:26.610 --> 24:28.690] should you have this sort of wild west economy [24:28.690 --> 24:30.570] where everybody can do what they want? [24:31.010 --> 24:33.690] Or would it be better to have some kinds of protections? [24:49.280 --> 24:52.440] How much is an acceptable level of damage to startups [24:52.440 --> 24:54.820] in order to reduce economic inequality? [24:54.820 --> 24:56.440] I don't know. I don't know. [24:56.440 --> 25:03.720] I just discovered this equation relating these two things. [25:03.920 --> 25:06.180] You yank on the bed sheet here, [25:06.180 --> 25:09.380] and someone on the other end of the bed gets their covers pulled off. [25:09.700 --> 25:13.260] People yanking on the economic inequality end of the bed sheet [25:13.260 --> 25:16.300] don't often realize that it's attached to all these other things. [25:16.400 --> 25:18.440] But what's the right relationship? [25:18.440 --> 25:19.820] God only knows. [25:19.820 --> 25:24.280] It's enough for me to point out that this relationship exists. [25:27.280 --> 25:29.100] All right. Yes? [25:47.610 --> 25:49.070] Well, for example, [25:49.070 --> 25:51.850] I think bloggers [25:51.850 --> 25:55.730] have put a little bit of a fire under government [25:55.730 --> 25:57.670] that didn't used to be there before. [25:58.190 --> 26:02.250] If everybody demanded more, [26:02.250 --> 26:04.670] asked more hard questions of politicians, [26:04.670 --> 26:05.950] that would be a good thing. [26:05.950 --> 26:08.770] Is everybody going to do it? No. [26:08.770 --> 26:11.690] But, you know, these things might change slowly [26:11.690 --> 26:13.410] in the right direction. [26:21.940 --> 26:23.800] Would the politicians just ignore you? [26:23.800 --> 26:25.060] Well, they'll try. [26:25.920 --> 26:27.860] But look what happened in Northern Ireland. [26:27.860 --> 26:32.300] I just read recently that an analyst looking at what was happening in Northern Ireland [26:32.300 --> 26:34.840] said basically the people got tired of the violence [26:34.840 --> 26:37.520] and the politicians were forced to go along with them. [26:37.920 --> 26:41.240] That bumper sticker, if the people leave, the politicians will follow. [26:41.240 --> 26:42.900] I definitely agree with that one. [26:44.160 --> 26:46.600] All right. We are done. Thank you very much. [26:46.600 --> 26:48.580] Sorry about the strange circumstances.