WEBVTT - Eric Balchunas on How Important Fees Are to Fund Performance 0:00:04.760 --> 0:00:08.080 Welcome to the Bloomberg P and L Podcast. I'm pim Fox. 0:00:08.119 --> 0:00:11.200 Along with my co host Lisa Abramowitz. Each day we 0:00:11.280 --> 0:00:14.480 bring you the most important, noteworthy, and useful interviews for 0:00:14.520 --> 0:00:16.880 you and your money, whether at the grocery store or 0:00:16.920 --> 0:00:20.680 the trading floor. Find the Bloomberg p L Podcast on iTunes, 0:00:20.840 --> 0:00:29.319 SoundCloud and at Bloomberg dot com pim Fox. It takes 0:00:29.360 --> 0:00:31.840 a lot to survive in the jungle, in particular the 0:00:31.880 --> 0:00:34.080 E t F jungle. I want to bring in my 0:00:34.159 --> 0:00:37.000 colleague Eric Valtunes. He has an E t F analyst 0:00:37.040 --> 0:00:41.320 for Bloomberg Intelligence, and he has great ideas. I love 0:00:41.400 --> 0:00:44.680 talking with Eric. I'm very glad you could join us. Uh. 0:00:45.040 --> 0:00:48.320 You know, we were talking yesterday about a couple of 0:00:48.440 --> 0:00:51.440 moves that Fidelity has has made, uh and black Rock 0:00:51.520 --> 0:00:57.840 has made to cut expenses as passive management takes a hold, 0:00:58.080 --> 0:01:02.840 how much our investors prioritizing the cheapness of funds over 0:01:03.240 --> 0:01:07.679 what they actually offer right now, it's it's everything, uh 0:01:07.319 --> 0:01:09.600 that you know. The phrase I use his expense ratio 0:01:09.680 --> 0:01:12.520 is the new past performance chart. I think this really 0:01:12.560 --> 0:01:16.319 gets down to trust. I think investors just trust a 0:01:16.480 --> 0:01:18.880 fee more than a pretty chart. Right now. And I 0:01:18.880 --> 0:01:21.400 don't know if it's from decades of seeing the chart 0:01:21.600 --> 0:01:23.720 buying into it. It's not working out because a lot 0:01:23.760 --> 0:01:25.880 of times investors will buy in it after the chart 0:01:25.920 --> 0:01:28.320 goes up and you know, it's hard to maintain our performance. 0:01:29.080 --> 0:01:31.720 Or it's just this spread of information that's been going 0:01:31.760 --> 0:01:34.640 on for about ten years that the Internet really helped 0:01:34.640 --> 0:01:36.720 to get out there, which is how important fees are 0:01:37.240 --> 0:01:40.600 in your in your final performance. Fees are a big 0:01:40.640 --> 0:01:43.679 contributor and the one thing you can control. So give 0:01:43.760 --> 0:01:45.520 us a sense of just how much more money has 0:01:45.520 --> 0:01:47.840 gone to the cheapest fees versus the most expensive. I'll 0:01:47.840 --> 0:01:50.760 throw out a few stats here. So last year, six 0:01:50.760 --> 0:01:53.440 billion dollars went into ETFs. Over half of that money 0:01:53.480 --> 0:01:56.800 went to products charging nine basis points or less UM. 0:01:56.880 --> 0:01:59.960 So that's once that so basically free, basically free. Yeah, 0:02:00.040 --> 0:02:02.640 I mean if you look at UM I've did this 0:02:02.880 --> 0:02:05.400 study where I you know, put everything into different buckets. 0:02:05.680 --> 0:02:08.359 It's going in active mutual funds. Funds that charge less 0:02:08.360 --> 0:02:11.440 than forty basis points that are active saw inflows d 0:02:11.560 --> 0:02:14.600 F A Vanguard or two big issuers in that area. 0:02:15.080 --> 0:02:17.880 Index funds that charge less than twenty saw more inflows 0:02:17.880 --> 0:02:20.840 than once charged forty So even within index funds, which 0:02:20.840 --> 0:02:23.480 are already passive, there's the cost migration. So I think 0:02:23.520 --> 0:02:25.960 when you look at the we put the flow every 0:02:26.080 --> 0:02:29.280 mutual fund, share class and e T F on a 0:02:29.280 --> 0:02:32.920 big spreadsheet, divided into buckets by expense ratios, and we 0:02:33.000 --> 0:02:35.800 found that by the vast majority of money goes to 0:02:35.919 --> 0:02:38.880 zero to ten basis points, then the next is ten 0:02:38.919 --> 0:02:42.040 to twenty, and it cascades down to about forty. Then 0:02:42.080 --> 0:02:45.839 at forty six it starts going down until plus ninety 0:02:45.919 --> 0:02:49.799 is the biggest outflow. So there's definitely a big correlation 0:02:49.919 --> 0:02:52.200 between the how much you charge and the flows you're seeing. 0:02:53.040 --> 0:02:55.640 Are the e t F issuers as well as the 0:02:55.680 --> 0:02:58.600 brokerage firms. Are they victims of their own success? And 0:02:58.680 --> 0:03:01.120 how many people does it to run a business that 0:03:01.240 --> 0:03:04.240 is going in this direction? Right? Good question. I have 0:03:04.280 --> 0:03:06.359 a piece out today that looks at the fact that 0:03:06.720 --> 0:03:09.080 you need human beings to do this. You also need 0:03:09.160 --> 0:03:12.200 human beings to analyze ETFs. It's not all robots, but 0:03:12.639 --> 0:03:16.320 I do think it's gonna definitely put some pressure on 0:03:16.800 --> 0:03:20.160 the operations and the human element of running an asset manager. 0:03:20.639 --> 0:03:22.320 We've already seen that. I mean, you wrote a story 0:03:22.360 --> 0:03:25.480 just yesterday calling, uh, you know, with the situation of 0:03:25.560 --> 0:03:27.680 pressure cooker for asset managers, and I think that's a 0:03:27.680 --> 0:03:31.560 pretty good term. So they're probably gonna have to do 0:03:31.639 --> 0:03:34.440 some things you saw Janice. I think this is the biggest, 0:03:34.440 --> 0:03:38.360 biggest example, Janice got together with Henderson. So I think 0:03:38.360 --> 0:03:40.320 you're gonna see more and more companies hooking up with 0:03:40.360 --> 0:03:43.080 other companies to get big enough so they can lower 0:03:43.120 --> 0:03:45.200 their fees and get scale going. And that will be 0:03:45.280 --> 0:03:47.320 the new game for the next decade. All right, So 0:03:47.600 --> 0:03:51.400 you see this migration toward passive, it doesn't this sort 0:03:51.440 --> 0:03:55.040 of raise alarms that people are going at the wrong time. 0:03:55.080 --> 0:03:57.640 I mean, the more people go into passive, the more 0:03:57.640 --> 0:04:01.960 opportunities there ought to be for active investors to outperform. Now, 0:04:02.760 --> 0:04:05.200 yes and no. So the fact that everybody's going into 0:04:05.200 --> 0:04:09.560 passive has helped lift assets in the same indexes. So 0:04:09.680 --> 0:04:11.800 right now, it's actually helping that there's this sort of 0:04:12.000 --> 0:04:13.560 call we'll call it a mini bubble. Right a lot 0:04:13.560 --> 0:04:15.800 of people are going into passive. It's lifting up stocks 0:04:15.840 --> 0:04:17.799 in the SMP. In other words, it's a self fulfilling 0:04:17.800 --> 0:04:20.279 prophecy because the more people go in, the more stocks rise, 0:04:20.320 --> 0:04:22.480 and then they see good performance and they keep going in. 0:04:23.320 --> 0:04:26.839 You talk to really smart managers that know their factors 0:04:26.880 --> 0:04:29.320 back and forth. They claim that I'm just gonna sit 0:04:29.360 --> 0:04:32.400 in value and wait till this little mini passive bubble 0:04:32.400 --> 0:04:34.920 pops and I'll then I'll be the hero. The question 0:04:35.000 --> 0:04:37.640 is when, because remember this isn't just the move to passive, 0:04:37.640 --> 0:04:40.080 it's a cost migration. Just so happens that the cheapest 0:04:40.080 --> 0:04:43.800 stuff is you know, index based, so passive is drawing 0:04:43.839 --> 0:04:46.039 those assets. The quint is when will this all play out? 0:04:46.120 --> 0:04:49.960 Right now, passive investments only own about twelve percent of 0:04:50.000 --> 0:04:53.279 the stock market, so the majority is still in other things, 0:04:53.360 --> 0:04:56.440 mostly active that so this pendulum could swing a lot further. 0:04:56.640 --> 0:04:58.960 So if you're waiting, how long do you wait until 0:04:58.960 --> 0:05:01.440 this happens? The second thing a smart Beta. You just 0:05:01.440 --> 0:05:03.400 made the sales pitch for smart Beta rob or not? 0:05:03.560 --> 0:05:05.839 He says, why would you want to buy an index 0:05:05.839 --> 0:05:08.200 where they reward price and they give more waiting to 0:05:08.279 --> 0:05:11.080 stuff that's bigger and pricier. You should come with me. 0:05:11.240 --> 0:05:14.560 I actually put higher weightings and value things that are cheaper. 0:05:14.960 --> 0:05:17.080 So smart Beta has funneled some money for people who 0:05:17.080 --> 0:05:19.680 are investing based on what you just said, but still 0:05:19.760 --> 0:05:23.520 using passive products to try to reward stocks that are 0:05:23.560 --> 0:05:26.039 trading cheaper or or in a momentum spurt. That's the 0:05:26.040 --> 0:05:28.200 whole smart beta sales pitch. You know, as you talk, 0:05:28.320 --> 0:05:31.359 I just keep thinking, how much does passive have to 0:05:31.400 --> 0:05:34.520 own before they've become a dominant force in a market? 0:05:34.560 --> 0:05:36.280 And I think this is one big question a lot 0:05:36.320 --> 0:05:39.280 of people are analyzing. There's not really a good answer 0:05:39.360 --> 0:05:42.320 to it. You might say that there is, but well, 0:05:42.920 --> 0:05:45.000 well we had an event. We had John Vogel he 0:05:45.040 --> 0:05:47.920 said it could be nine before it's a problem. Wow. 0:05:48.520 --> 0:05:51.880 Thanks very much. Eric Balcunis, he is an expert when 0:05:51.920 --> 0:05:54.040 it comes to E t F. He's our senior et 0:05:54.480 --> 0:05:58.240 F analyst for Bloomberg Intelligence and you can follow him 0:05:58.320 --> 0:06:13.680 on Twitter at Eric Valtunist. Well, when we want to 0:06:13.720 --> 0:06:16.599 know more about municipal bonds, we call on one person, 0:06:16.680 --> 0:06:19.560 Joe Maisak. He is editor of Bloomberg Briefs for the 0:06:19.640 --> 0:06:22.520 municipal market and he joins us. Now, Joe, thanks very 0:06:22.600 --> 0:06:25.920 much for being here. Let's talk about municipal bonds and 0:06:26.080 --> 0:06:29.200 what may or may not happen to the municipal bond market. 0:06:29.520 --> 0:06:32.720 If the Fedow Reserve A deigns to increase interest rates 0:06:32.720 --> 0:06:36.960 twenty five basis points in a couple of weeks, tons 0:06:37.120 --> 0:06:41.120 of uh of actually good things follow on the hills 0:06:41.160 --> 0:06:45.800 of of higher yields PIM. If we see the triple 0:06:45.880 --> 0:06:49.680 a ten year ago to maybe four percent, I see 0:06:50.240 --> 0:06:54.719 things like insurance rising. Uh, there will be the muni 0:06:54.800 --> 0:06:58.800 junk bottle of revive swaps will come back. The auction 0:06:58.880 --> 0:07:03.480 rate securities market, which has been sort of dead for um, 0:07:04.440 --> 0:07:09.560 you know, since two thousand seven, that will probably revive. Um, 0:07:09.720 --> 0:07:13.640 We're probably going to see a little help for public pensions. 0:07:13.760 --> 0:07:16.120 And why don't I toss in a little bit uh 0:07:16.600 --> 0:07:20.000 more stadium construction, uh to sort of go hand in 0:07:20.040 --> 0:07:23.480 hand with muni junk. And right now the returns are 0:07:24.000 --> 0:07:27.680 pretty good, right, muni bond returns in February, we're the 0:07:27.840 --> 0:07:33.120 highest since two tho correct, Yes, and uh, you know 0:07:34.080 --> 0:07:36.880 it's still you know, these are the yields have been 0:07:36.960 --> 0:07:42.800 solo for so long. Uh, it's it's almost incredible how 0:07:42.880 --> 0:07:45.280 flat the market has been about two and a half 0:07:45.320 --> 0:07:47.760 percent and ten years. Okay, So one place that has 0:07:47.800 --> 0:07:51.240 not been flat is with Puerto Rico municipal bonds. And 0:07:51.280 --> 0:07:53.280 we did get some news this week with the new 0:07:53.320 --> 0:07:58.680 governor ra Cio. We're proposing a fiscal plan that included, 0:07:59.400 --> 0:08:02.440 uh it was it, coverage of about two thirds of 0:08:03.040 --> 0:08:06.760 bond payments over the next I don't know, nine years 0:08:06.800 --> 0:08:09.640 or so, a little little less than that. Uh, he 0:08:10.240 --> 0:08:14.000 said about one point two billion. I think the control Board, 0:08:14.040 --> 0:08:18.600 the oversight, the Federal Oversight Committee, Uh, you know, they 0:08:18.640 --> 0:08:23.360 see him paying maybe eight million. The kind of surprising 0:08:23.440 --> 0:08:26.960 thing is he also asked for a complete stay on 0:08:27.000 --> 0:08:29.760 any litigation to be extended to the end of the year. 0:08:30.200 --> 0:08:34.040 So bond holders are sort of up in arms about that. Yeah, 0:08:34.040 --> 0:08:37.720 but they don't seem to care. I think they didn't 0:08:37.760 --> 0:08:41.520 really respond much to this, did they. But the bond prices, Yeah, 0:08:41.559 --> 0:08:43.280 the fact that, like, you know, all of a sudden, 0:08:43.320 --> 0:08:47.120 the amount that's covered under this fiscal plan is completely 0:08:47.120 --> 0:08:51.120 insufficient for what for to cover their losses, and people 0:08:51.160 --> 0:08:53.200 are just sort of like, yeah, whatever, you know, it's 0:08:53.480 --> 0:08:57.000 right now. They're going to enter the negotiation period with 0:08:57.040 --> 0:08:59.600 the governor. And it's the surprising thing about the governors. 0:08:59.600 --> 0:09:03.280 He can panned as the uh as the man who's 0:09:03.320 --> 0:09:07.040 going to repay the debt, and now that he's in office, 0:09:07.840 --> 0:09:12.400 repaying the debt is a lot less appetizing. Shall we say, 0:09:12.480 --> 0:09:14.200 can we talk about some numbers, because I was just 0:09:14.200 --> 0:09:18.440 looking at a comparison. California, which actually went re entered 0:09:18.520 --> 0:09:21.960 the green muni market, uh this week with an issuance 0:09:22.280 --> 0:09:26.560 California for a ten year averages two point six seven percent. 0:09:27.000 --> 0:09:30.920 That's obviously triple tax free. New York State two point 0:09:31.000 --> 0:09:34.280 three one percent, and then you've got these outliers Illinois 0:09:34.360 --> 0:09:38.760 four point five seven percent. Tell us a little bit 0:09:38.760 --> 0:09:43.160 about the volume of issuance, because supply also affects the price. 0:09:44.160 --> 0:09:47.599 You know, this year we started off first year, I 0:09:47.640 --> 0:09:51.200 think January was around thirty billion. We're probably up to 0:09:51.960 --> 0:09:56.280 over the first two months maybe fifty five billion or so. 0:09:56.720 --> 0:09:59.960 And this is off a little bit. Of course, last 0:10:00.040 --> 0:10:02.920 you was a record pace. We had forward six billion, 0:10:03.240 --> 0:10:05.440 So this year we're off a little bit. It's a 0:10:05.440 --> 0:10:09.960 little slack, however, you know, you bring this up and 0:10:10.320 --> 0:10:12.959 next week, all of a sudden, we see a whole 0:10:13.080 --> 0:10:16.720 batch of state general obligation bonds, including two point four 0:10:16.760 --> 0:10:21.120 billion from California. So the states seemed to be looking 0:10:21.120 --> 0:10:25.440 to take advantage of maybe the last shot at rates 0:10:25.480 --> 0:10:30.520 this low. One reason why mini bonds initially sold off 0:10:30.640 --> 0:10:33.880 after President Trump was elected was because people were ratcheting 0:10:34.120 --> 0:10:37.280 uh back their expectations for how high their taxes might 0:10:37.320 --> 0:10:41.199 go and expect expecting some tax relief. This basically gives 0:10:41.360 --> 0:10:44.400 less of a benefit to municipal bonds, which are tax free. 0:10:44.840 --> 0:10:48.000 Now that we really have not seen a tax plan, 0:10:48.559 --> 0:10:51.280 uh and we don't know how soon it will be implemented. 0:10:52.000 --> 0:10:53.640 Is that part of what's bringing people back to the 0:10:53.640 --> 0:10:56.960 mini market too. Well, they have to put their money 0:10:56.960 --> 0:11:02.320 to work, and you have, uh, you know, tax concerns, um. 0:11:02.360 --> 0:11:06.240 But you know, the whole tax picture I think has 0:11:06.280 --> 0:11:09.520 sort of been put on the back burner for a while. 0:11:09.600 --> 0:11:12.839 Like you say, it's was something that instead of coming 0:11:12.880 --> 0:11:15.560 here in June, July or August, which is what the 0:11:15.559 --> 0:11:18.720 administration has talked about, most people really think it's going 0:11:18.800 --> 0:11:20.960 to be very late in the year, probably more like 0:11:21.200 --> 0:11:25.040 next year. The administration has a lot uh to worry 0:11:25.040 --> 0:11:28.640 about between now and then, especially the Healthcare Act. Thanks 0:11:28.720 --> 0:11:31.120 very much for joining us. Joe Maisak. He is the 0:11:31.240 --> 0:11:35.880 editor of Bloomberg Briefs on Municipal Markets. Always enjoyed listening 0:11:35.920 --> 0:11:37.800 to you, and of course meanings We're gonna have to 0:11:37.800 --> 0:11:40.160 wait and see what the foto reserved does. And maybe 0:11:40.200 --> 0:11:43.240 this will increase the issuance of municipal bands well. And 0:11:43.280 --> 0:11:46.000 this might actually help with the whole idea of infrastructure 0:11:46.040 --> 0:11:50.000 spending uh, as President Trump has proposed, perhaps by a 0:11:50.080 --> 0:12:04.800 sort of a back channel with private money pim fox. 0:12:04.880 --> 0:12:08.640 There is so much data that is available today. There 0:12:08.640 --> 0:12:11.000 are some good statistics about just how much the n 0:12:11.080 --> 0:12:15.079 s A has, for example, compared to prior administrations during 0:12:15.120 --> 0:12:17.280 the communist era, when people thought, you know, oh my gosh, 0:12:17.280 --> 0:12:19.320 the government has such a handle on what everyone is doing. 0:12:19.760 --> 0:12:21.640 But now we have so much data that we need 0:12:21.679 --> 0:12:26.280 to rely on algorithms to sort through it and find patterns. Uh. 0:12:26.360 --> 0:12:30.960 But are these algorithms really better at sifting through the data, 0:12:31.000 --> 0:12:34.240 albeit the massive tropes that we have, and coming up 0:12:34.280 --> 0:12:37.160 with conclusions and humans? I want to bring in Kathy 0:12:37.240 --> 0:12:42.080 O'Neill Bloomberg View columnists h and also a mathematician who 0:12:42.080 --> 0:12:45.760 has been a professor, hedge fund analyst and data scientist, 0:12:46.120 --> 0:12:48.560 and she wrote a fabulous column about how in the 0:12:48.600 --> 0:12:53.080 world of big data, more isn't always better. Kathy uh, 0:12:53.480 --> 0:12:56.720 first of all. Just to put this into perspective, just 0:12:56.920 --> 0:13:02.880 how prevalent are algorithms in decision making processes throughout the 0:13:02.920 --> 0:13:06.640 economy at this point. How dependent are we upon them? 0:13:06.640 --> 0:13:09.400 That's a great question, um. And people don't really realize 0:13:09.400 --> 0:13:11.520 this because a lot of the algorithms are actually happening 0:13:11.520 --> 0:13:13.680 behind the scenes and we can't even see them. But 0:13:13.720 --> 0:13:17.120 it turns out algorithms are being used on basically every 0:13:17.160 --> 0:13:19.880 important decision in somebody's life, every time they have an option, 0:13:19.920 --> 0:13:22.640 and they're sort of competing with other people. So that 0:13:22.720 --> 0:13:25.600 means you know, college admissions, getting a job, even while 0:13:25.640 --> 0:13:28.640 you're on the job, how you're being evaluated, how much 0:13:28.679 --> 0:13:31.400 you pay for insurance, you know, how much you you 0:13:31.559 --> 0:13:33.959 what kind of ap are you get for credit? Um. 0:13:34.000 --> 0:13:36.640 Even things like policing and how long you're going to 0:13:36.760 --> 0:13:40.480 jail if you get if you if you're guilty. Um, 0:13:40.520 --> 0:13:43.280 those are all determined by algorithms. Never mind all the 0:13:43.320 --> 0:13:45.920 things that are happening online of course, and the political 0:13:46.000 --> 0:13:48.920 micro targeting, all the ads you see on politics, which 0:13:49.160 --> 0:13:53.760 or are all algorithmically defined. So yeah, it's it's absolutely everywhere. Um. 0:13:53.800 --> 0:13:57.880 And what's especially concerning is the moments like when you're 0:13:57.880 --> 0:13:59.640 trying to get a job and you send your resume 0:13:59.760 --> 0:14:03.720 in and there are algorithms that filter your resume based 0:14:03.760 --> 0:14:06.120 on you know, used to be just keyword searches, like 0:14:06.160 --> 0:14:08.480 what kind of words did you say on your resume? 0:14:08.559 --> 0:14:11.000 But now it's all sorts of other kinds of correlative 0:14:11.200 --> 0:14:14.120 information about your resume and you will never know that. 0:14:14.240 --> 0:14:17.040 Just to be clear, like, you'll either get a callback 0:14:17.160 --> 0:14:18.840 or you won't. But if you don't get a callback, 0:14:18.880 --> 0:14:21.600 you won't know why, and it might be because of 0:14:21.600 --> 0:14:23.880 an algorithm. I wonder if you can give us an 0:14:23.880 --> 0:14:28.520 example of an algorithm that's gone wrong. Yeah, well, I 0:14:28.520 --> 0:14:31.680 actually wrote a book about this called Weapons of Mass Destruction. UM. 0:14:32.840 --> 0:14:35.120 Very good book. I did read it. Oh, oh, thank 0:14:35.120 --> 0:14:38.280 you so much, FIM. UM. So I would say, a 0:14:38.320 --> 0:14:42.520 lot of algorithms go wrong, um for the people they 0:14:42.520 --> 0:14:45.320 are targeted, but not necessarily for the people that build 0:14:45.320 --> 0:14:47.760 the algorithms. So to be clear, UM, one of the 0:14:47.920 --> 0:14:51.120 main points of the book is that, UM, what you 0:14:51.160 --> 0:14:53.520 know It depends on this on your perspective on whether 0:14:53.560 --> 0:14:56.160 something's going wrong. But one of the one of the 0:14:56.160 --> 0:14:58.760 examples of my book, UM, it comes from the world 0:14:58.800 --> 0:15:03.480 of education. UM. In their teacher assessment algorithms, and teachers 0:15:03.480 --> 0:15:06.400 are basically being scored between zero and a hundred UM. 0:15:06.520 --> 0:15:09.640 And I think that is probably the best example of 0:15:09.640 --> 0:15:12.920 a terrible algorithm, because it's very inconsistent. I found a 0:15:12.920 --> 0:15:15.400 teacher who got a ninety six one year and a 0:15:15.520 --> 0:15:19.040 six another year UM. And another teacher got fired for 0:15:19.080 --> 0:15:21.480 a bad score, even though she thinks that her score 0:15:21.560 --> 0:15:24.960 was artificially low because other teachers cheating. UM. And there's 0:15:25.120 --> 0:15:27.560 I mean, I think the critical point about these algorithms 0:15:27.640 --> 0:15:30.119 isn't that they're bad, because of course there's bad algorithms 0:15:30.120 --> 0:15:32.560 out there. The critical point is that they're being used 0:15:32.600 --> 0:15:36.760 as if they're scientifically they have scientific authority. UM. So 0:15:36.800 --> 0:15:40.720 people trust them in a kind of under an overexaggerated 0:15:40.760 --> 0:15:43.600 way because of their mathematical nature. And so that's one 0:15:43.640 --> 0:15:45.440 of the things I was trying to get go after 0:15:45.560 --> 0:15:48.600 in my book. As a mathematician, I don't want people 0:15:48.600 --> 0:15:51.880 to blindly trust mathematics. But the point of mathematics is 0:15:51.880 --> 0:15:54.720 that it's actually supposed to clarify things, not obfuscate them. 0:15:54.840 --> 0:16:00.320 Well to that point, how can companies and universe these 0:16:00.440 --> 0:16:04.320 make sure that their algorithms are doing the right thing? 0:16:04.360 --> 0:16:06.320 I mean, what's the check here on how to make 0:16:06.320 --> 0:16:08.720 the algorithms better. The answer isn't necessarily for a human 0:16:08.760 --> 0:16:11.080 being to be trying to sift through all the data themselves, 0:16:11.320 --> 0:16:15.280 because at this point it's that's an unsustainable solution given 0:16:15.280 --> 0:16:17.760 the amount of data that a lot of these algorithms 0:16:17.800 --> 0:16:22.320 are tasked with the processing. Absolutely true, and I'm not 0:16:22.720 --> 0:16:25.560 I am not anti algorithm whatsoever. What I'm trying to 0:16:26.080 --> 0:16:29.880 um suggest um is that we create standards of evidence 0:16:30.640 --> 0:16:34.600 that the algorithms are meaningful, that they have statistical meaning, 0:16:35.000 --> 0:16:38.680 but they're also that they're fair, and they're they're legal. Um. 0:16:38.720 --> 0:16:41.280 A lot of the algorithms I examined in my book 0:16:41.360 --> 0:16:45.160 actually I think are probably illegal. But because regulators don't 0:16:45.200 --> 0:16:48.800 know how to examine algorithms, um, they're you know, companies 0:16:48.840 --> 0:16:52.160 are getting away with stuff. So in particular with the 0:16:52.440 --> 0:16:57.359 algorithms for hiring, they're kind of replacing their HR divisions 0:16:57.440 --> 0:17:00.840 with algorithms without making sure that all those all the 0:17:00.960 --> 0:17:05.440 algorithms are actually you know, they reflect fair hiring practices, 0:17:05.480 --> 0:17:08.080 which there's plenty of laws around that. So my point 0:17:08.200 --> 0:17:11.840 is like we need to create evidence and to demand 0:17:11.880 --> 0:17:14.560 evidence from the people that build these algorithms and use 0:17:14.600 --> 0:17:17.520 all these algorithms that what they're doing is actually legal 0:17:17.560 --> 0:17:21.399 and fair. A point to you, is it possible that 0:17:21.480 --> 0:17:25.240 what's going on is algorithms are being used to cut 0:17:25.320 --> 0:17:29.560 costs in a corporate setting, and that the reason it's 0:17:29.600 --> 0:17:33.400 being done is because then there is no personal accountability 0:17:33.440 --> 0:17:36.440 for the result, that is, you nailed it on the head, 0:17:36.480 --> 0:17:41.160 Like I observe just many, many examples of these algorithms 0:17:41.200 --> 0:17:42.920 that are what I call weapons of math destruction, and 0:17:42.920 --> 0:17:46.200 they're very powerful, their secret and their destructive. And one 0:17:46.240 --> 0:17:48.800 of the things I noticed sort of after writing the 0:17:48.840 --> 0:17:51.520 book is that there's a kind of certain characteristics of 0:17:51.520 --> 0:17:54.040 a situation that are ripe for these kinds of algorithms. 0:17:54.080 --> 0:17:56.359 And it's exactly what you just said. It's when people 0:17:56.400 --> 0:17:59.840 don't want to take personal responsibility for tricky decisions. Tricky 0:17:59.920 --> 0:18:01.840 is asians like is this a good teacher or not? 0:18:02.000 --> 0:18:04.040 Is this a good applicant or not? Those are tricky 0:18:04.359 --> 0:18:06.359 and they know that it can go wrong, and they 0:18:06.359 --> 0:18:09.159 would like to say it's not me, it's the algorithm. 0:18:09.160 --> 0:18:11.679 And that's when the situation is right for for pretty 0:18:11.720 --> 0:18:14.919 nasty things happen. Well, it certainly reminds me of the 0:18:14.920 --> 0:18:20.080 Stanley Kubrick movie, Uh, Doctor Strange Love, because right, I 0:18:20.119 --> 0:18:22.160 mean the doomsday machine. It just takes it takes over, 0:18:22.160 --> 0:18:24.320 and there's no way to really get a human being 0:18:24.320 --> 0:18:27.320 in between that and and the dire consequences. Thank you 0:18:27.440 --> 0:18:30.720 very much. Cathy O'Neil is a Bloomberg View columnists and 0:18:31.160 --> 0:18:35.920 you can follow her on Twitter at math Babe dot org. 0:18:36.240 --> 0:18:38.520 Yes indeed, and she is also the author of the 0:18:38.560 --> 0:18:55.359 book Weapons of Math Destruction. A lot has been said 0:18:55.440 --> 0:19:00.080 about President Trump's immigration policies, particularly as they pertain to 0:19:00.240 --> 0:19:03.840 Mexico and our other southern neighbors. I want to bring 0:19:03.880 --> 0:19:06.720 in someone who could talk more about that. Hector Barretto, 0:19:06.800 --> 0:19:10.440 who is chairman of the Latino Coalition and former US 0:19:10.480 --> 0:19:15.320 Small Business Administrator under George W. Bush. Hector, I'm glad 0:19:15.359 --> 0:19:17.119 you could join us. First, I want to start with 0:19:17.840 --> 0:19:22.960 a question, how different are President Trump's proposed immigration policies 0:19:23.000 --> 0:19:27.680 to the ones that were implemented under former President Barack Obama. 0:19:28.280 --> 0:19:31.840 I think we're saying under Bush or who I worked for. UM, 0:19:31.920 --> 0:19:34.960 you know, the policies are very different. Obviously, every administration 0:19:35.000 --> 0:19:36.800 has a different take on it. We don't know all 0:19:36.800 --> 0:19:39.960 the specifics about what the immigration policy is going to be, 0:19:40.320 --> 0:19:44.399 especially going forward with regards to countries like Mexico. You know, 0:19:44.520 --> 0:19:47.240 I worked for President Bush, and the first thing he 0:19:47.280 --> 0:19:50.480 did when he uh got into office was he put 0:19:50.520 --> 0:19:53.960 forth in an initiative with Mexico called the Partnership for Prosperity. 0:19:54.000 --> 0:19:56.520 We worked on that for eight years while I was 0:19:56.960 --> 0:20:01.240 in the administration. Uh. Less so in the last administration 0:20:01.280 --> 0:20:04.399 in terms of things like the Partnership for Prosperity. Obviously, 0:20:04.480 --> 0:20:08.400 there was never comprehensive immigration reform under the Obama administration, 0:20:08.640 --> 0:20:12.800 and we don't know what the immigration reform proposals are 0:20:12.800 --> 0:20:15.399 going to be uh in in this administration. Do you 0:20:15.400 --> 0:20:18.080 think that immigration policies need to be reformed in the 0:20:18.200 --> 0:20:22.080 last Absolutely? Well, I mean, you know, I think everybody 0:20:22.160 --> 0:20:28.159 stipulates that our immigrations UH system are policies towards immigration 0:20:28.440 --> 0:20:31.159 don't really work for anybody. They don't work for the 0:20:31.200 --> 0:20:33.440 immigrants that are here, they don't work for the economy, 0:20:33.640 --> 0:20:36.160 they don't really work for anybody. So, you know, everybody 0:20:36.200 --> 0:20:38.480 basically says it's a broken system and we need to 0:20:38.520 --> 0:20:41.720 fix it, but nobody can coalesce around what that looks 0:20:41.760 --> 0:20:44.200 like and so that's why we keep kicking this can 0:20:44.240 --> 0:20:47.359 down the road. Remember, we haven't had immigration reform in 0:20:47.400 --> 0:20:51.720 the United States since nineteen six that was President Reagan. 0:20:51.960 --> 0:20:54.600 So we've been thirty years dealing with a lot of 0:20:54.600 --> 0:20:56.919 these issues. And I think a lot of people on 0:20:56.960 --> 0:20:59.960 both sides, they all are very frustrated and and hope 0:21:00.040 --> 0:21:03.480 bowl that we can get something done in this administration. Hector, 0:21:03.520 --> 0:21:05.920 I want to just bring a little bit more detailed 0:21:05.960 --> 0:21:09.960 to the conversation. You previously before your role in a 0:21:10.720 --> 0:21:14.560 business on a large scale, you worked in your family's restaurant, uh, 0:21:14.760 --> 0:21:20.720 import export business, construction accompany, and so on. In US, 0:21:20.800 --> 0:21:25.280 Hispanic buying power was larger than the gross domestic product 0:21:25.720 --> 0:21:30.639 of Mexico and it is growing substantially. Uh, we're talking 0:21:30.640 --> 0:21:36.240 about a buying power of about thirteen point nine trillion dollars, right, 0:21:36.280 --> 0:21:42.800 that's uh, that's the total. Is there a way for 0:21:42.880 --> 0:21:46.480 this immigration plus a I don't know if you want 0:21:46.520 --> 0:21:51.400 to call it immigration, but a nafter revamp, trade policy 0:21:51.440 --> 0:21:56.159 revamp that could actually make that an even more a 0:21:56.200 --> 0:21:59.840 stronger figure. How can it help the business about it? Look, 0:22:00.040 --> 0:22:02.480 you know, I serve in a lot of different capacities. 0:22:02.480 --> 0:22:04.760 I'm also on the board of the US Chamber of Commerce. 0:22:04.800 --> 0:22:08.200 This has been a major, major focus and an initiative 0:22:08.240 --> 0:22:12.160 for us for many, many years. And these things are connected. Uh, 0:22:12.240 --> 0:22:14.520 you know you were talking about the family businesses. You know, 0:22:14.560 --> 0:22:17.560 I'm the son of an immigrant. My father, Hector Bretto Sr. 0:22:17.880 --> 0:22:20.639 Was an immigrant from Mexico in the fifties and he 0:22:20.720 --> 0:22:23.200 was the founder of the U. S Hispanic Chamber of Commerce. 0:22:23.440 --> 0:22:27.520 We've learned about these issues all my life in our family. 0:22:27.680 --> 0:22:31.000 We know how important that they are, and they are interconnected. Look, 0:22:31.040 --> 0:22:33.399 purchasing power in the United States. Just Hispanics in the 0:22:33.440 --> 0:22:36.600 United States is one point five trillion dollars. My father 0:22:36.680 --> 0:22:38.920 used to always say, you know, the Hispanics in the 0:22:39.000 --> 0:22:41.800 United States have more purchasing power than all the the 0:22:41.800 --> 0:22:44.280 the Hispanics in Mexico and a lot of countries in 0:22:44.359 --> 0:22:49.120 Latin America. And and that is going to grow Hispanic business, 0:22:49.119 --> 0:22:52.359 which I'm very passionate. About four million companies that are 0:22:52.359 --> 0:22:55.000 generating close to seven hundred billion in revenues and they 0:22:55.000 --> 0:22:57.960 could double every five years. And many of those have 0:22:58.119 --> 0:23:01.399