WEBVTT - Darrell Duffie On How to Fix the World’s Most Important Market 0:00:11.000 --> 0:00:14.160 Hello, and welcome to another episode of the Odd Lots Podcast. 0:00:14.240 --> 0:00:16.440 I'm Tracy Alloway and I'm Joe Wisenthal. 0:00:16.720 --> 0:00:18.920 Joe, have you been watching treasury yield s lately? 0:00:19.200 --> 0:00:19.320 Uh? 0:00:19.480 --> 0:00:22.080 They're up, They've I'm that's it. 0:00:22.120 --> 0:00:23.239 That's the show where. 0:00:23.079 --> 0:00:25.160 That the line has gone up into the right lance. 0:00:25.239 --> 0:00:28.960 Yes, but not just that, they've moved quite quickly up. 0:00:29.040 --> 0:00:32.200 And I think volatility in the treasury market has once 0:00:32.280 --> 0:00:35.479 again become a talking point. And I always get a 0:00:35.520 --> 0:00:38.519 little bit of a sense of deja vu because whenever 0:00:38.560 --> 0:00:42.800 things start happening in the market for US government bonds, 0:00:42.840 --> 0:00:46.800 extreme things, these extreme moves, it feels like everyone says, Oh, 0:00:46.800 --> 0:00:49.240 they shouldn't be happening in the world's most liquid market, 0:00:49.320 --> 0:00:52.320 the world's safest market. We shouldn't be seeing these types 0:00:52.479 --> 0:00:54.240 of dramatic shifts, right. 0:00:54.440 --> 0:00:58.000 The expectation is this is like a very extremely liquid, 0:00:58.240 --> 0:01:00.320 slow moving market. But it has been a very fast 0:01:00.360 --> 0:01:02.440 move and people don't really know why. There's a lot 0:01:02.480 --> 0:01:04.840 of debate. You have people talk about, well, look, the 0:01:04.880 --> 0:01:07.759 economy is proving to be more resilient than people might 0:01:07.800 --> 0:01:11.240 have guessed six months ago, even a month ago. And 0:01:11.319 --> 0:01:13.920 I think there's this expectation. There's this sense that things 0:01:13.959 --> 0:01:16.520 are moving faster. There's also a lot of talk about 0:01:16.640 --> 0:01:21.320 treasury supply, supplied demand imbalances and so forth. So things 0:01:21.360 --> 0:01:23.360 are on the move, to say the least, and you 0:01:23.400 --> 0:01:26.200 see it in mortgage spreads are very wide. You see 0:01:26.240 --> 0:01:30.000 it another risk assets not seeming to like this treasury volatility. 0:01:30.080 --> 0:01:31.560 So lots going on there. 0:01:31.600 --> 0:01:34.199 So you mentioned a couple of things there, which is, yes, 0:01:34.240 --> 0:01:37.880 the economy seems to be doing relatively well, and yes, 0:01:38.160 --> 0:01:41.800 the amount of treasury supply is exploding and has been 0:01:41.880 --> 0:01:45.039 going up for a very long time. There's another factor here, 0:01:45.160 --> 0:01:49.000 which is the actual inner workings of the treasury market, 0:01:49.080 --> 0:01:52.280 so how treasuries are actually traded and whether or not 0:01:52.400 --> 0:01:56.120 that is potentially contributing to some of the volatility that 0:01:56.160 --> 0:01:59.200 we've seen. And again, I feel like this topic keeps 0:01:59.360 --> 0:02:02.440 coming up every year there's some massive move in treasury 0:02:02.480 --> 0:02:06.120 markets and everyone starts talking about liquidity issues, and yet 0:02:06.240 --> 0:02:10.440 you don't really see any big solutions being proposed to it. 0:02:10.560 --> 0:02:13.160 You see talk about, you know, maybe loosening some of 0:02:13.200 --> 0:02:16.600 the regulation around supplementary leverage ratio or something like that, 0:02:16.680 --> 0:02:18.680 But it feels like we need to talk about this 0:02:18.800 --> 0:02:20.600 and maybe try to come up with a solution. 0:02:20.680 --> 0:02:22.440 Let's do it, and we're in a good place to 0:02:22.440 --> 0:02:23.000 talk about it. 0:02:23.080 --> 0:02:24.640 Yes, we're at Jackson Hole. 0:02:25.320 --> 0:02:26.639 I thought you were. I thought that was going to 0:02:26.680 --> 0:02:29.120 be the first is like Jackson I like how you 0:02:29.200 --> 0:02:30.720 like we eat it into it a little bit. 0:02:30.840 --> 0:02:33.280 Oh, I want this to be an evergreen episode for 0:02:33.320 --> 0:02:35.680 an evergreen problem. It feels like, but. 0:02:35.760 --> 0:02:37.280 The people here care about this topic. 0:02:37.440 --> 0:02:42.280 Yes, absolutely. The way US treasury debt is traded has 0:02:42.440 --> 0:02:46.440 a lot of implications, particularly for the cost of financing 0:02:47.080 --> 0:02:49.600 the US deficit. That's an obvious one, and so we 0:02:49.639 --> 0:02:50.360 should talk about it. 0:02:50.440 --> 0:02:51.160 Let's talk about it. 0:02:51.280 --> 0:02:54.080 I'm very pleased to say that we have the perfect guest. 0:02:54.280 --> 0:02:56.560 We are going to be speaking with Darryl Duffy. He is, 0:02:56.600 --> 0:03:00.000 of course, a professor of finance at Stanford's Graduate School 0:03:00.240 --> 0:03:03.800 of Business, and he is presenting a paper at Jackson 0:03:03.840 --> 0:03:07.399 Hole in front of the world's top central bankers all 0:03:07.440 --> 0:03:11.000 about this issue. It's called Resilience reducs in the US 0:03:11.040 --> 0:03:13.960 Treasury market. So, Darryl, thank you so much for coming on. 0:03:14.000 --> 0:03:17.600 All thoughts, thanks for having me, Tracy, Joe, It's terrific. 0:03:17.240 --> 0:03:19.359 To be here, you know, it's the perfect guest when 0:03:19.400 --> 0:03:21.280 it's the guest, and it's like not just it, like 0:03:21.320 --> 0:03:22.880 we want to have him on odd laws that the 0:03:22.960 --> 0:03:26.000 central the top central banker is around the world want 0:03:26.040 --> 0:03:29.080 to hear from a Professor Duffy. So yes on this issue. 0:03:29.160 --> 0:03:30.840 I'm very excited about this episode. 0:03:30.919 --> 0:03:34.760 When treasuries are moving, they call up Daryl Duffy. All right, well, professor, 0:03:35.240 --> 0:03:37.720 thank you again for coming on. Why don't we just 0:03:38.080 --> 0:03:40.800 start with the basics. You know, we sort of alluded 0:03:40.840 --> 0:03:43.200 to this in the intro, but it feels like this 0:03:43.240 --> 0:03:46.360 is a subject that just keeps coming up and never 0:03:46.440 --> 0:03:48.640 really seems to go away. What's going on here? 0:03:49.000 --> 0:03:53.120 Well, over the past century, there have been, as you suggested, 0:03:53.160 --> 0:03:57.000 many episodes of increased volatility and liquidity problems in the 0:03:57.040 --> 0:03:59.800 treasury market, but I do think these are happening more frequently. 0:04:00.160 --> 0:04:02.600 Just recently. I think you summarized pretty well some of 0:04:02.640 --> 0:04:06.320 the stresses in the treasury market coming from the fiscal side. 0:04:06.360 --> 0:04:10.240 The US is issuing more than people expected, there was 0:04:10.280 --> 0:04:14.320 a recent downgrade by Fitch, and the Fed is struggling 0:04:14.440 --> 0:04:18.039 with what to do about inflation. That additional monetary policy 0:04:18.080 --> 0:04:22.839 uncertainty also contributes to volatility. Let me back up a minute. 0:04:22.920 --> 0:04:25.080 I just spent most of the last year on a 0:04:25.080 --> 0:04:29.120 sabbatical at the New York FED and working with some 0:04:29.240 --> 0:04:36.039 terrific economists there, Michael Fleming, Frank Keene or Shekhar, Peter 0:04:36.160 --> 0:04:39.760 von tassl Claire Nelson. We decided we needed to look 0:04:39.800 --> 0:04:44.279 into the relationship between the volatility that you two discussed 0:04:44.600 --> 0:04:48.560 and the liquidity in the market. They're closely intertwined. So 0:04:48.839 --> 0:04:51.279 we dug deep and went into a lot of data, 0:04:51.279 --> 0:04:55.000 and yeah, volatility seems to be the main determinant of 0:04:55.040 --> 0:04:58.599 lil liquidity in the market. So when fiscal uncertainty or 0:04:58.720 --> 0:05:03.360 debt ceiling debate, or a COVID crisis or monetary policy 0:05:03.440 --> 0:05:09.320 uncertainty start to get a volatility higher and higher, market 0:05:09.360 --> 0:05:13.000 becomes less and less liquid. It's an extremely regular relationship. 0:05:13.000 --> 0:05:17.839 About eighty percent of illiquidity is explained simply by variation 0:05:18.000 --> 0:05:19.520 and yield volatility. 0:05:19.720 --> 0:05:24.600 So just to press on this relationship between rates volatility 0:05:24.640 --> 0:05:29.039 or yield volatility and illiquidity, like which way does the 0:05:29.080 --> 0:05:31.520 direction run? Or because we can come up with these 0:05:31.640 --> 0:05:34.680 fundamental stories right, like Okay, maybe the economy is better 0:05:34.720 --> 0:05:37.839 than expected. Maybe we're going to have more rate hikes 0:05:37.880 --> 0:05:40.599 than previously anticipated, in which case you say, okay, well 0:05:40.600 --> 0:05:43.880 that's sort of like that's a fundamental story of fundamental driver. 0:05:44.240 --> 0:05:46.880 And then you could look at certain nature of the structure, 0:05:46.960 --> 0:05:49.680 like the size of dealer balance sheets, et cetera. It's like, okay, well, 0:05:49.720 --> 0:05:53.680 this is something technical that contributes to also could be 0:05:53.680 --> 0:05:56.560 a contributor to volatility. How do you think about like 0:05:56.800 --> 0:05:59.800 the directions of causality when you look into a problem. 0:05:59.440 --> 0:06:04.240 Like this, Yeah, well, both the direct fundamentals, fiscal monetary 0:06:04.279 --> 0:06:08.640 fundamentals and global the global economy and geopolitics recently all 0:06:08.640 --> 0:06:11.800 play a direct fundamental role. And then, as you alluded, 0:06:12.240 --> 0:06:15.800 there's also kind of a feedback effect. When volatility rises 0:06:15.800 --> 0:06:20.200 for fundamental reasons, dealers are going to struggle with providing 0:06:20.200 --> 0:06:24.320 sufficient liquidity to the market, and to the extent that 0:06:24.400 --> 0:06:28.600 dealer balance sheets are not sufficiently flexible to accommodate the 0:06:28.640 --> 0:06:31.279 provision of liquidity to the market, that in and of 0:06:31.320 --> 0:06:36.440 itself increases illiquidity, increases volatility, and they kind of feedback 0:06:36.480 --> 0:06:39.080 on themselves, and you can get an episode like we 0:06:39.279 --> 0:06:43.839 had when COVID hit in March twenty twenty, where liquidity 0:06:43.880 --> 0:06:48.360 becomes even worse than would be suggested by volatility alone, 0:06:48.480 --> 0:06:51.000 much worse. And that's just a sign that the market 0:06:51.040 --> 0:06:56.720 is not capable of intermediating those extreme demands for liquidity. 0:06:56.960 --> 0:07:01.200 Right. So, this seems to be a distinct characteristic of 0:07:01.360 --> 0:07:05.280 treasuries in particular, which is when stuff starts to go 0:07:05.440 --> 0:07:08.599 really wrong like it did in March of twenty twenty, 0:07:09.000 --> 0:07:13.200 people often sell the safest stuff first, which means they 0:07:13.320 --> 0:07:16.280 sell treasury. So you get this big wave of selling 0:07:16.480 --> 0:07:19.240 at precisely the moment that a lot of dealers want 0:07:19.280 --> 0:07:23.120 to wind down or back away from risk. Is that 0:07:23.280 --> 0:07:26.520 just a fundamental tension in the market. Is that always 0:07:26.560 --> 0:07:27.360 going to be the case? 0:07:27.720 --> 0:07:30.960 As long as US treasuries are the world's most important 0:07:31.000 --> 0:07:35.080 safe haven, which is clearly the case by Miles, that's 0:07:35.120 --> 0:07:39.160 always going to be the result for basically two reasons. 0:07:39.240 --> 0:07:42.760 Number One, a whole lot of major investors like foreign 0:07:42.800 --> 0:07:48.080 exchange reserve managers firms that are storing a safe liquid 0:07:48.160 --> 0:07:51.320 asset just in case, Well, the just in case happened, 0:07:51.800 --> 0:07:54.760 and they are going to liquidate those positions. The other 0:07:55.120 --> 0:07:59.239 channel for this is as the volatility grows and uncertainty grows, 0:08:00.120 --> 0:08:02.640 of investors are you know, kind of finding it too 0:08:02.720 --> 0:08:05.240 hot to handle and they have to unload some risk, 0:08:05.320 --> 0:08:08.960 and treasuries are the easiest security to unload in the world. 0:08:08.960 --> 0:08:12.120 The market's got a good reputation for being the deepest 0:08:12.120 --> 0:08:14.400 and most liquid market in the world. 0:08:14.240 --> 0:08:16.560 So I know, we we've talked about this a little 0:08:16.560 --> 0:08:19.320 bit in the past with a few different guests, including 0:08:19.520 --> 0:08:21.920 Josh Younger. He made the point on one of the 0:08:21.920 --> 0:08:24.160 episodes that we did, which is like, if you're thinking about, 0:08:24.200 --> 0:08:26.520 like what do we want to do to have better 0:08:26.600 --> 0:08:30.800 treasury market structure, that it doesn't necessarily make sense to 0:08:30.920 --> 0:08:33.440 optimize for well, we never want to have a March 0:08:33.480 --> 0:08:36.680 twenty twenty again, because you don't necessarily want to have, 0:08:37.160 --> 0:08:39.440 you know, optimized for the one out of every one 0:08:39.559 --> 0:08:42.600 hundred year pandemic. But what would you say, like is 0:08:42.640 --> 0:08:45.040 the goal, Like if you're like, okay, there does seem 0:08:45.040 --> 0:08:49.040 to be this relationship between volatility and illiquidity. It does 0:08:49.080 --> 0:08:52.120 seem like some of these bouts of volatility and illiquidity 0:08:52.280 --> 0:08:55.439 become more frequent. If you think about designing sort of 0:08:55.520 --> 0:08:59.199 an optimal market structure for treasuries, what would you say 0:08:59.200 --> 0:08:59.880 we're trying. 0:08:59.640 --> 0:09:02.640 To each Josh I've known him since you worked at 0:09:02.720 --> 0:09:05.120 JP Morgan, and now that he's moved to the FED, 0:09:05.160 --> 0:09:07.280 we get to talk a lot more. This is a 0:09:07.400 --> 0:09:10.400 terrific insight that he has, the kind of do you 0:09:10.600 --> 0:09:12.679 really want to design a market for the worst day 0:09:12.679 --> 0:09:15.760 in a thousand? Isn't that very expensive? And maybe overdoing 0:09:15.760 --> 0:09:18.840 it because nine ninety nine days out of a thousand 0:09:18.880 --> 0:09:21.720 you didn't really need that kind of a market structure. 0:09:21.840 --> 0:09:24.760 I'm going to be a little provocative here. I think 0:09:24.800 --> 0:09:26.680 you do want to build a market for the worst 0:09:26.760 --> 0:09:29.920 day in a thousand for the following reason. If I'm 0:09:30.280 --> 0:09:33.800 let's say, managing the foreign exchange reserves of an emerging 0:09:33.840 --> 0:09:36.800 market central bank, when do I need to actually take 0:09:36.840 --> 0:09:39.760 advantage of the depth and liquidity of the US Treasury market. 0:09:40.080 --> 0:09:42.640 It's that one day in a thousand when all the 0:09:42.720 --> 0:09:46.400 other safe haven investors are trying to do the same thing. 0:09:47.400 --> 0:09:50.280 In the paper on that Tracy mentioned I'm giving here 0:09:50.280 --> 0:09:53.800 at Jackson Hole, I talk about this wrong way risk 0:09:54.000 --> 0:09:57.439 from the viewpoint of illiquidity. You don't want the market 0:09:57.559 --> 0:10:00.720 to be great except on that very same regular day 0:10:01.360 --> 0:10:05.800 on which everybody needs the liquidity. Why not? Well, because 0:10:06.080 --> 0:10:09.640 a this is the lynchpin of global financial market stability. 0:10:09.679 --> 0:10:12.840 You wanted to work day in, day out, and b 0:10:13.280 --> 0:10:17.640 If you discourage safe haven investors from believing that even 0:10:17.679 --> 0:10:20.720 though everybody else is liquidating that day, they could also 0:10:20.800 --> 0:10:24.000 liquidate at low cost with ease, then they won't use 0:10:24.040 --> 0:10:27.240 the US Treasury security as much as they're safe haven. 0:10:28.200 --> 0:10:31.080 They'll diversify. And that's what we've been seeing somewhat over 0:10:31.080 --> 0:10:34.320 the last couple of decades, a degree of diversification away 0:10:34.360 --> 0:10:38.839 from the US Treasury, still by far the dominant safe haven. 0:10:39.600 --> 0:10:42.360 Something like fifty nine percent of foreign exchange reserves are 0:10:42.360 --> 0:10:44.319 held in US treasuries. But from the viewpoint of the 0:10:44.440 --> 0:10:47.480 US taxpayer, you want everyone to believe that on the 0:10:47.520 --> 0:10:49.480 worst day NF thousand, that market is going to be 0:10:49.520 --> 0:10:50.520 there for them. 0:10:50.840 --> 0:10:53.880 Wait, I'm going to be provocative now or try to. Okay, 0:10:54.120 --> 0:10:57.480 So on this note, we did see in March twenty twenty, 0:10:57.559 --> 0:11:02.200 the FED unveiled all these different emergency programs aimed at 0:11:02.240 --> 0:11:04.199 supporting the US treasury market. So you know, we have 0:11:04.320 --> 0:11:07.680 the new standing repo facility, and I think you know 0:11:07.720 --> 0:11:10.520 they were buying US treasuries in exchange for reserves, and 0:11:10.520 --> 0:11:13.560 then they exempted all of those from the supplementary leverage 0:11:13.640 --> 0:11:19.280 ratios for banks. So it seems like the backstop is 0:11:20.000 --> 0:11:23.120 in place. If something bad were to happen again, I 0:11:23.160 --> 0:11:26.120 would presume that the FED would unveil either those exact 0:11:26.120 --> 0:11:30.559 measures again or something very similar. So is the issue solved? 0:11:31.080 --> 0:11:35.360 Yeah, you're tracy. That is provocative, So I would definitely. 0:11:35.120 --> 0:11:36.880 Say it's a polite way of saying you're so wrong. 0:11:37.480 --> 0:11:42.240 No, the FED. The FED came out guns blazing unlimited 0:11:42.400 --> 0:11:46.680 financing and the rebound market for anyone that had access 0:11:46.840 --> 0:11:49.880 to the FED a trillion dollars of purchases in the 0:11:49.880 --> 0:11:54.120 first three weeks, nearly a trillion of US treasuries, relieving 0:11:54.120 --> 0:11:57.800 the other dealer balance sheets of their overloading. Getting the 0:11:57.800 --> 0:12:02.000 supplementary leverage ratio back and it was causing problems took 0:12:02.040 --> 0:12:05.240 a little longer, and it took I think some backroom 0:12:05.360 --> 0:12:08.480 negotiations with the other bank regulators to come on board, 0:12:08.840 --> 0:12:11.040 so that got delayed, and that was a problem. But 0:12:11.080 --> 0:12:14.839 the FED did a terrific job at crisis management during 0:12:14.880 --> 0:12:19.320 those weeks, and I say weeks because they didn't solve 0:12:19.400 --> 0:12:23.120 the problem. They only made it less bad than it 0:12:23.200 --> 0:12:26.720 otherwise would have been. It took five, six, seven, eight 0:12:26.760 --> 0:12:31.040 weeks before market liquidity was restored. And again going back 0:12:31.040 --> 0:12:35.280 to my wrong way risk point, if I'm looking for 0:12:35.800 --> 0:12:38.280 a market that's going to work for me in a crisis, 0:12:38.800 --> 0:12:40.839 I don't want to have to wait weeks in order 0:12:40.960 --> 0:12:43.600 to get liquidity or to pay a low cost for liquidity. 0:12:43.640 --> 0:12:46.000 I wanted to be working all the time. Now, of 0:12:46.000 --> 0:12:49.320 course it's unrealistic that it should work every single day. 0:12:49.520 --> 0:12:52.040 But if we rely only on central banks and I 0:12:52.080 --> 0:12:56.319 speak more broadly, to bail out their government securities market 0:12:56.760 --> 0:12:58.920 when they get into trouble, it's not going to be 0:12:58.920 --> 0:13:01.920 one hundred percent effective and it raises moral hazard. It 0:13:02.000 --> 0:13:04.600 says to the rest of the world, we'll use the 0:13:04.640 --> 0:13:06.880 central bank balance sheet to bail you out. You don't 0:13:06.880 --> 0:13:11.160 need to focus on improving market structure reducing undue leverage. 0:13:11.840 --> 0:13:14.840 We have your backs that message. While it needs to be, 0:13:14.960 --> 0:13:18.600 there is not a substitute for improving the market structure. 0:13:18.840 --> 0:13:23.080 Yeah, you sort of anticipated my next question, but okay, 0:13:23.160 --> 0:13:26.600 it does seem like in an emergency the FED can 0:13:26.640 --> 0:13:29.200 say we're going to know by do QUI at a 0:13:29.240 --> 0:13:32.120 scale that we've never seen before, and as Tracy mentioned, 0:13:32.679 --> 0:13:35.880 unveil these unveil these new facilities kind of on the fly, 0:13:36.200 --> 0:13:38.679 and it seems like basically since two thousand and eight 0:13:38.679 --> 0:13:41.240 two thousand and nine, the FED has gotten really good 0:13:41.280 --> 0:13:44.200 at standing up new facilities very quickly. So that's like 0:13:44.240 --> 0:13:46.400 a skill that they've developed. But can you talk a 0:13:46.440 --> 0:13:48.640 little bit more about what you perceive is to be 0:13:48.800 --> 0:13:52.320 the cost of a sort of stability regime that sort 0:13:52.360 --> 0:13:56.520 of presumpt presumes that, yeah, we know there's some frailties, 0:13:56.800 --> 0:13:59.760 but our solution is that in that you know, seven 0:14:00.160 --> 0:14:03.280 more twelve sigma day that the FED is there, and like, well, 0:14:03.280 --> 0:14:05.200 talk about more about why that's not a good system. 0:14:05.320 --> 0:14:08.280 Okay, Well, just I want to re emphasize that the FED, I. 0:14:08.240 --> 0:14:10.839 Don't know how many sigmas it really is that billion, 0:14:11.080 --> 0:14:13.280 I don't know. I just put throughout a number. 0:14:13.480 --> 0:14:15.959 The FED does need to be there. It's not as 0:14:16.000 --> 0:14:20.120 though one should say, let's take the fed's balance sheet 0:14:20.120 --> 0:14:21.960 out of the equation and try to do without it. 0:14:21.960 --> 0:14:24.960 It needs to be there. It's a backstop. It's the 0:14:25.040 --> 0:14:27.840 last resort. The FED is the buyer of last resort. 0:14:27.880 --> 0:14:30.040 After it's become the lender of last resort. It can't 0:14:30.080 --> 0:14:33.560 do anything else. But bail at the market by buying securities. 0:14:33.880 --> 0:14:37.480 But relying on that has several problems. I already mentioned 0:14:37.720 --> 0:14:40.000 it's not one hundred percent of effective on the first day. 0:14:40.920 --> 0:14:43.280 And there's also the size of the Fed's balance sheet 0:14:43.480 --> 0:14:46.760 that's controversial. I mean, even if you think it's innocuous, 0:14:47.120 --> 0:14:50.600 it raises political concerns. There are those that say, well, 0:14:50.640 --> 0:14:53.040 maybe the Fed's balance sheet is too big, and we 0:14:53.080 --> 0:14:57.760 need to curtail the ability of central banks, including the FED, 0:14:57.840 --> 0:15:00.280 to expand their balance sheets to the extent that they 0:15:00.280 --> 0:15:02.840 have been and they've been using them very, very liberally 0:15:02.880 --> 0:15:06.120 over the past couple of decades. The other concern is 0:15:06.120 --> 0:15:08.920 once the balance sheet is large, it eventually is going 0:15:08.960 --> 0:15:11.560 to come back down and those treasuries are going to 0:15:11.600 --> 0:15:15.720 be adding to the stock of securities that other investors 0:15:16.040 --> 0:15:18.400 need to have, and that it means that the central banks, 0:15:18.440 --> 0:15:21.240 including the FED, need to do that very gingerly. There's 0:15:21.280 --> 0:15:23.160 a lot of volatility in the treasury market, and the 0:15:23.160 --> 0:15:26.320 FED is letting its balance sheet come down. Other investors 0:15:26.360 --> 0:15:29.160 are having to pick up the load. It's easier to 0:15:29.240 --> 0:15:31.360 expand the balance sheet than it is to bring it down. 0:15:31.560 --> 0:15:34.360 So using the Fed's balance sheet while it's necessary, is 0:15:34.520 --> 0:15:38.000 not a painless solution, and I would argue it's not 0:15:38.040 --> 0:15:41.080 the best solution anyway. We can do better by improving 0:15:41.120 --> 0:15:45.400 market structure. Pushing out into the extreme tails the number 0:15:45.440 --> 0:15:47.640 of events in which the FED needs to step in 0:15:47.680 --> 0:15:48.120 and linch. 0:15:51.040 --> 0:15:54.480 Why don't we talk about the market structure and maybe 0:15:54.520 --> 0:15:57.360 before we start talking about improvements that could be made, 0:15:57.440 --> 0:15:59.280 could you give us the sort of lay of the 0:15:59.400 --> 0:16:03.040 land when it comes to how treasuries are traded today. 0:16:03.120 --> 0:16:05.960 So there's primary dealers for the new issuance, and then 0:16:06.000 --> 0:16:08.600 there's your sort of run of the mill dealers for 0:16:09.000 --> 0:16:12.280 secondary market trades. But talk to us how it works 0:16:12.360 --> 0:16:12.840 right now? 0:16:12.920 --> 0:16:16.040 Terrific. Well, it's an extremely complex structure, but it can 0:16:16.040 --> 0:16:19.320 be summarized pretty simply. There's two segments of the market. 0:16:19.400 --> 0:16:23.520 There's the interdealer market, in which the dealers trade among themselves, 0:16:23.840 --> 0:16:26.120 and then there's the customer to dealer market, in which 0:16:26.520 --> 0:16:31.360 investors around the world trade with dealers. Notably, investors do 0:16:31.440 --> 0:16:34.720 not trade directly with other investors. There is no alto 0:16:34.760 --> 0:16:37.520 all trade in the US treasure market, no matter whether 0:16:37.520 --> 0:16:40.840 you're an insurance company, a hedge fund, a foreign exchange reserve, manager. 0:16:41.600 --> 0:16:44.000 You are going to be buying and selling with a 0:16:44.040 --> 0:16:46.960 dealer if you're a dealer. On the other hand, there 0:16:47.000 --> 0:16:50.320 is a very active interdealer market for the on the 0:16:50.400 --> 0:16:53.120 run securities. Those are the latest issues of the Treasury. 0:16:53.640 --> 0:16:55.960 There's an order book market, which is a high frequency 0:16:55.960 --> 0:16:59.440 trading market run by Brokertech, which is a subsidiary of 0:16:59.440 --> 0:17:02.960 the Chicago Mercantile Exchange, where you have the same kinds 0:17:03.000 --> 0:17:05.919 of high frequency trading that you see in the stock market. 0:17:06.080 --> 0:17:09.280 The only other participants on the broker Tech market are 0:17:09.359 --> 0:17:14.399 high frequency trading firms sometimes called principal trading firms like Jump, 0:17:14.680 --> 0:17:18.960 like DRW, like Citadel, firms that have a very special 0:17:19.000 --> 0:17:23.840 purpose of intermediating in the interdealer market, taking you know, 0:17:24.320 --> 0:17:27.560 little bit offer spreads from the dealers and from each other. 0:17:27.840 --> 0:17:30.840 That's the basic structure of the market. Again, the notable 0:17:30.880 --> 0:17:33.840 feature is if you're an investor, you can trade only 0:17:33.880 --> 0:17:36.480 with a dealer. If you're a dealer, you have the 0:17:36.560 --> 0:17:39.760 ability to lay off positions in the interdealer market. 0:17:39.920 --> 0:17:43.360 Again, before we get into sort of like optimal structure, 0:17:43.400 --> 0:17:45.399 I'm actually just curious. You know, you mentioned that you 0:17:45.440 --> 0:17:48.040 spent the last year at the New York FED looking 0:17:48.080 --> 0:17:50.439 into this and sort of getting this. What did you do? 0:17:50.640 --> 0:17:52.399 How did you let go about your research? Like, how 0:17:52.480 --> 0:17:54.720 much is it sort of like a sort of statistical 0:17:54.760 --> 0:17:58.760 base analysis versus how much was it talking to dealers 0:17:58.800 --> 0:18:01.680 and understanding how they are. I'd just be curious about. 0:18:01.680 --> 0:18:05.400 The research process for this particular project was a combination 0:18:05.560 --> 0:18:08.399 of meeting and discussing what needed to be done with 0:18:08.520 --> 0:18:11.600 the economists that I mentioned earlier, and those would be 0:18:11.640 --> 0:18:14.720 weekly meetings pretty in depth where we would go through 0:18:15.320 --> 0:18:17.480 what we've already learned and what we need to do next. 0:18:17.920 --> 0:18:22.120 And that happened for six months or so. And at 0:18:22.119 --> 0:18:26.000 the same time, in the background, we're collecting volumes of 0:18:26.119 --> 0:18:30.040 statistical data to FED, because it's a member of the 0:18:30.040 --> 0:18:33.080 official sector, has access not only to its own data, 0:18:33.480 --> 0:18:38.800 but to exceptionally fine grained data on at the transactions level. 0:18:38.880 --> 0:18:41.440 Let me give you one example. There is a data 0:18:41.480 --> 0:18:45.919 set called trace which records every single trade in the 0:18:45.960 --> 0:18:50.120 treasure market with a few minor exceptions. Those data are 0:18:50.160 --> 0:18:54.080 only available to the official sector. They're not available to 0:18:54.119 --> 0:18:56.520 the public. And by the way, I disagree with that policy, 0:18:56.960 --> 0:18:58.920 and we could talk about that. I think it actually 0:18:59.000 --> 0:19:02.600 is contributes to the problem of illiquidity. But in any case, 0:19:03.119 --> 0:19:05.720 the FED, as a member of that official sector group, 0:19:06.160 --> 0:19:11.080 can go to its sister agencies in the federal government 0:19:11.280 --> 0:19:14.320 and say, look, we have this project, here's its objectives. 0:19:14.320 --> 0:19:17.000 We want to use these trace data to analyze liquidity 0:19:17.000 --> 0:19:21.280 in the US treasury market. And then we get feedback saying, yeah, 0:19:21.320 --> 0:19:23.119 this looks good. The way that pre zata are being 0:19:23.200 --> 0:19:28.159 presented will not reveal proprietary information, so go ahead. And 0:19:28.200 --> 0:19:31.080 then we can do the same thing with dealer balance 0:19:31.080 --> 0:19:34.600 sheet data. We can get exposures of the dealers not 0:19:34.640 --> 0:19:38.399 only to treasury securities, but to agency mortgage backed securities, 0:19:38.400 --> 0:19:41.520 which turned out to be another big load on their 0:19:41.560 --> 0:19:45.720 balance sheet, particularly during March of twenty twenty. We can 0:19:45.800 --> 0:19:50.119 go to a wide range of data sets, and we 0:19:50.160 --> 0:19:53.680 wrote a paper it explains the extent to which we 0:19:54.640 --> 0:19:58.320 access all of these data bring them together. We developed 0:19:58.359 --> 0:20:02.720 eighteen different liquidity tricks and many different metrics on how 0:20:02.760 --> 0:20:05.680 dealer balance sheets are being loaded. And then we would 0:20:05.720 --> 0:20:13.600 analyze these using reasonably intricate econometric methods like quantile regressions 0:20:13.640 --> 0:20:17.800 and a number of other statistical approaches, and then we 0:20:17.840 --> 0:20:20.680 would start to see the patterns emerge very very clearly. 0:20:21.080 --> 0:20:24.680 That I describe two key patterns that came up over 0:20:24.760 --> 0:20:29.480 and over again in our discussion meetings. Where A volatility 0:20:29.480 --> 0:20:32.960 seems to explain most of the variation in liquidity, but 0:20:33.400 --> 0:20:37.280 b when it doesn't, it's dealer balance sheet loading that 0:20:37.440 --> 0:20:41.920 explains the remaining part of illiquidity. It's a highly nonlinear effect. 0:20:42.520 --> 0:20:46.720 When dealer balance sheets are normally loaded, they don't contribute 0:20:46.720 --> 0:20:50.160 to ill liquidity. But when they're reaching their extremes, where 0:20:50.200 --> 0:20:54.359 dealers are handling more treasury trades and more agency MBS 0:20:54.440 --> 0:20:57.400 trades and they've handled in the past, then you see 0:20:57.480 --> 0:21:01.760 illiquidity go up well beyond the level predicted by volatility. 0:21:02.320 --> 0:21:06.880 So after analyzing all these data and discussing what's driving these, 0:21:07.320 --> 0:21:10.440 then we turn to writing up our results and there's 0:21:10.480 --> 0:21:12.960 a lot of iterative work there which you can see 0:21:13.200 --> 0:21:14.200 in the paper that we wrote. 0:21:14.440 --> 0:21:16.960 So you can see the dealer balance sheets on a 0:21:17.040 --> 0:21:18.800 daily basis, not just at quarter. 0:21:18.720 --> 0:21:19.919 End, not quite. 0:21:20.040 --> 0:21:23.120 We can only see dealer balance sheets on a weekly 0:21:23.160 --> 0:21:26.480 basis because the FED has a data set called FR 0:21:26.520 --> 0:21:30.520 two thousand and four which collects those data only on 0:21:30.560 --> 0:21:33.720 a weekly basis and summaries of those data are available 0:21:33.720 --> 0:21:35.560 publicly on the New York Fed's website. 0:21:35.800 --> 0:21:39.480 So going back to this dealer balance sheet issue, I mean, 0:21:39.520 --> 0:21:41.960 this is something that has come up basically ever since 0:21:42.000 --> 0:21:45.240 the two thousand and eight financial crisis, and there have 0:21:45.320 --> 0:21:48.040 been a lot of complaints from the dealers about all 0:21:48.080 --> 0:21:51.560 this new regulation that limits their ability to take risk 0:21:51.720 --> 0:21:54.880 on their balance sheet. And the argument for doing that 0:21:54.920 --> 0:21:58.040 has always been one of financial stability. Well, we want 0:21:58.080 --> 0:22:00.520 the banks to be safer, and if they have to 0:22:00.560 --> 0:22:04.440