WEBVTT - The Regulatory Blunder That Gave Us the Silicon Valley Bank Disaster 0:00:10.200 --> 0:00:14.200 Hello, and welcome to another episode of the Odd Lots podcast. 0:00:14.280 --> 0:00:19.040 I'm Joe Wisenthal and I'm Tracy Allow. So Tracy, we obviously, 0:00:19.280 --> 0:00:23.000 I think have a sense of why Silicon Valley Bank failed. 0:00:23.040 --> 0:00:25.680 We just published a really good episode with Dan Davis, 0:00:25.840 --> 0:00:28.520 like sort of like talking about where things went wrong 0:00:28.840 --> 0:00:32.200 on the sort of deposit side and failing to balance 0:00:32.280 --> 0:00:35.800 assets and liabilities and the issues of strengths and weaknesses 0:00:35.800 --> 0:00:38.519 of the business model. But then the other question, there 0:00:38.520 --> 0:00:42.320 are many, many more questions beyond just like why they failed. Yes, 0:00:42.440 --> 0:00:44.920 I mean some of the big ones that are emerging 0:00:45.120 --> 0:00:49.440 are where were the regulators right it? You know, people 0:00:49.520 --> 0:00:54.040 were already analyzing spb's balance sheet, you know, certainly the 0:00:54.080 --> 0:00:57.279 week before it collapsed and for some time before that, 0:00:57.320 --> 0:01:00.600 and you could see these vulnerabilities when it comes to 0:01:00.680 --> 0:01:03.640 duration exposure. And again that's something we talked about with 0:01:03.680 --> 0:01:06.960 Dan Davies. And then the other thing I guess just 0:01:07.400 --> 0:01:12.200 in general is it's not like bank failures are that 0:01:12.720 --> 0:01:15.960 unusual over the course of history. So this is the 0:01:16.000 --> 0:01:18.840 first big one, I guess since the two thousand and 0:01:18.840 --> 0:01:22.119 eight financial crisis. And so it's obviously garnering a lot 0:01:22.120 --> 0:01:26.840 of attention, But we do have bank failures throughout history 0:01:27.000 --> 0:01:29.319 from time to time, and it kind of begs the 0:01:29.400 --> 0:01:32.640 question of, well, if we're going to keep having them 0:01:33.120 --> 0:01:36.319 in different ways, and if the government or the Federal 0:01:36.360 --> 0:01:39.320 Reserve are going to keep coming in and rescuing them 0:01:39.400 --> 0:01:43.120 in various ways, should we maybe do something to the 0:01:43.240 --> 0:01:47.040 system to make it different. What were the failures and 0:01:47.120 --> 0:01:49.240 can we at least you know, we're always going to 0:01:49.280 --> 0:01:52.680 be fighting the last War cliche obviously, but what does 0:01:52.720 --> 0:01:55.400 it tell us about weaknesses in the system? And there 0:01:55.440 --> 0:01:56.920 may be things that we could do, And of course 0:01:56.960 --> 0:01:59.240 there's things that people are talking about on the sort 0:01:59.280 --> 0:02:03.680 of written law side, which is like, Okay, maybe we 0:02:03.720 --> 0:02:08.560 need more banks to have greater liquidity, ability to meet withdrawals, etc. 0:02:08.840 --> 0:02:12.040 I think some of the smaller, more regional banks don't 0:02:12.080 --> 0:02:15.000 have a stringent requirements on that front as the really 0:02:15.080 --> 0:02:18.120 large banks. And then people are talking about supervision, and 0:02:18.600 --> 0:02:22.120 I don't think supervision gets as much attention as the 0:02:22.160 --> 0:02:25.080 sort of written laws, but it's essentially, well, why did 0:02:25.200 --> 0:02:29.240 the supervisors the bank regulators allow the bank to create 0:02:29.280 --> 0:02:31.960 this confluence of risks, this big like sort of very 0:02:32.000 --> 0:02:36.280 specific mismatch between the nature of its assets, the nature 0:02:36.320 --> 0:02:40.880 of its deposit base that allowed it to unravel really quickly. Absolutely, 0:02:40.919 --> 0:02:44.359 and then of course, with the FED announcing this new facility, 0:02:44.720 --> 0:02:47.560 which is quite a dramatic one, you have a question of, Okay, 0:02:47.639 --> 0:02:51.040 if we're just going to guarantee all the US bank 0:02:51.120 --> 0:02:55.040 deposits out there, then should we maybe make a more 0:02:55.160 --> 0:02:59.880 fundamental change to the banking industry itself? Right? And actually, 0:03:00.320 --> 0:03:03.560 Matt Klein over at the Overshoot, I think he tweeted this, 0:03:03.760 --> 0:03:06.480 but he had that great first line in his most 0:03:06.480 --> 0:03:11.919 recent newsletter about how basically banks are these private investment 0:03:12.000 --> 0:03:16.560 funds that are grafted on top of critical infrastructure, and 0:03:16.600 --> 0:03:20.639 that structure is designed to extract subsidies from the rest 0:03:20.680 --> 0:03:25.799 of society by basically threatening people with banking crises whenever 0:03:25.840 --> 0:03:27.960 one of them is allowed to fail. And we saw 0:03:28.000 --> 0:03:31.520 that last week, right, We saw especially a bunch of 0:03:31.600 --> 0:03:34.440 vcs coming out and saying, if you don't rescue all 0:03:34.440 --> 0:03:38.160 the SVB depositors right now, this is going to happen 0:03:38.160 --> 0:03:40.360 to all the banks. And so you kick off that, 0:03:40.480 --> 0:03:45.720 you know, privatization of profits versus publicization of loss's argument 0:03:45.880 --> 0:03:48.520 over and over and over again. Well, that's been like 0:03:48.720 --> 0:03:52.840 really clear in this particular episode, like there's something about 0:03:52.840 --> 0:03:56.080 this story really raises some like uncomfortable days because it's 0:03:56.080 --> 0:04:00.400 not coming in like a wholesale financial collapse that's related 0:04:00.440 --> 0:04:02.760 to the collapse of the economy like in two thousand 0:04:02.760 --> 0:04:04.480 and eight or two thousand nine. It's like it's this 0:04:04.640 --> 0:04:08.320 very specific industry that sort of got it in trouble. Anyway, 0:04:08.680 --> 0:04:11.520 we could go on and on, but I'm excited. We 0:04:11.600 --> 0:04:14.160 do have the perfect guest for us to talk about 0:04:14.440 --> 0:04:18.039 the role of regulator, the role of regulatory failure, the 0:04:18.120 --> 0:04:20.159 role of the FED in all of this in the 0:04:20.240 --> 0:04:22.200 history of banking, and how we got here. We're gonna 0:04:22.200 --> 0:04:25.200 be speaking to a Levmanon Here's a professor at Columbia 0:04:25.279 --> 0:04:27.880 Law School written a lot about the FED and regulation. 0:04:28.000 --> 0:04:30.360 So Lev, thank you so much for joining us, Thank 0:04:30.400 --> 0:04:32.359 you so much for having me. So just you know, 0:04:32.880 --> 0:04:36.680 very top line view. You know, what would you say 0:04:36.800 --> 0:04:47.160 was the main regulatory failure with SIV with SVB, So yeah, yeah, 0:04:47.200 --> 0:04:50.640 we can distinguish between maybe regulation bright line rules yeah, 0:04:50.680 --> 0:04:53.840 put down in advance, and sort of supervision, yes, which 0:04:53.920 --> 0:04:58.400 is discretionary safety and soundness oversight by by examiners and 0:04:59.480 --> 0:05:03.720 federal federal regulators, federal officials, and both the sort of 0:05:03.720 --> 0:05:06.680 bright line rules and the sort of supervision failed here. Yeah, 0:05:06.720 --> 0:05:08.880 there was an overreliance on the bright line rules and 0:05:08.920 --> 0:05:11.600 a failure to do the discretionary oversight, the safety and 0:05:11.640 --> 0:05:14.760 soundness oversight effectively. So on the bright line rules side, 0:05:15.839 --> 0:05:19.480 SVB figured out a way to take additional risk without 0:05:19.520 --> 0:05:24.839 holding additional capital. Because of what's called the risk capital rules, 0:05:25.040 --> 0:05:28.000 treasury securities are risk weighted zero. That means that a 0:05:28.040 --> 0:05:32.599 bank has to hold zero equity against their treasury positions. 0:05:32.800 --> 0:05:36.240 And so SEP was able to go and buy a 0:05:36.279 --> 0:05:38.320 lot of long dated treasuries and actually build up quite 0:05:38.320 --> 0:05:40.440 a bit of interest rate risk without that being reflected 0:05:40.560 --> 0:05:43.440 in the capital that was required of them under the 0:05:43.760 --> 0:05:46.880 under the bright line rules, under the under the regulatory framework. 0:05:46.960 --> 0:05:50.760 Now we have a whole supervisory framework that's designed to 0:05:50.880 --> 0:05:54.520 deal with these sorts of holes in the rules. Everybody 0:05:54.560 --> 0:05:58.080 knows that the rules are insufficiently precise, and in fact, 0:05:58.320 --> 0:06:00.720 we didn't even have the rules until the nineteen eighties 0:06:00.880 --> 0:06:04.200 in meaningful sense. We used to just do discretionary safety 0:06:04.200 --> 0:06:06.240 and soundness oversight. And so the real question here in 0:06:06.279 --> 0:06:09.000 some sense is how come the supervisors didn't pick up 0:06:09.000 --> 0:06:11.880 on the fact that SVB had gamed the rules to 0:06:12.440 --> 0:06:15.120 take on a lot of interest rate risk without holding 0:06:15.120 --> 0:06:18.040 an adequate amount of capital against it. It's a pretty 0:06:18.080 --> 0:06:22.200 obvious maneuver. It's not nearly as complex as some of 0:06:22.240 --> 0:06:27.039 the maneuvers the gaming and it's not novel. It's not novel. Yeah, 0:06:27.080 --> 0:06:29.520 this is an old move. It's not like they camp 0:06:29.520 --> 0:06:35.279 with something new. You would think any seasoned supervisor looking 0:06:35.320 --> 0:06:38.480 at the balance sheet could pick up on this pretty quickly. 0:06:38.839 --> 0:06:41.960 So so what happened, I think is exactly the right 0:06:42.040 --> 0:06:46.280 question to be asking, and I think the answer requires 0:06:46.960 --> 0:06:50.720 maybe And this might just be my approach to these 0:06:50.760 --> 0:06:55.200 issues twenty thirty years worth of history to understand, because basically, 0:06:55.720 --> 0:07:00.840 I think contemporary supervision is broken in some sense and 0:07:00.960 --> 0:07:04.680 this is a manifestation of that. Okay, um, well, I'm 0:07:04.839 --> 0:07:07.839 I'm just gonna go ahead and buy and say, please, 0:07:07.960 --> 0:07:10.440 please give us the you know, thirty or forty years 0:07:10.480 --> 0:07:13.200 of banking history building up to those. Yeah, So let me. 0:07:13.360 --> 0:07:15.120 I'll say why I think it's broken, and then I'll 0:07:15.160 --> 0:07:17.000 tell you how we got sure, how it got so broken. 0:07:17.520 --> 0:07:22.800 So it's broken because outside of the stress testing framework, 0:07:22.840 --> 0:07:27.040 which I think we should definitely talk more about. Supervisors 0:07:27.680 --> 0:07:34.000 primarily now focus on process and procedures. Our insight into 0:07:34.000 --> 0:07:37.400 what actually goes on in supervision is very limited, and 0:07:37.640 --> 0:07:44.080 that's because supervisory materials are all confidential and can't be 0:07:44.120 --> 0:07:47.840 disclosed by the bank and aren't disclosed by by the regulators, 0:07:48.120 --> 0:07:51.360 and often never made public. So there's a lot of 0:07:51.400 --> 0:07:54.720 opacity into into what supervisors are actually doing. But it's 0:07:54.720 --> 0:07:57.440 it's fairly obvious, and I'll go through a few examples 0:07:57.640 --> 0:08:01.240 that what supervisors today tend to focus on is the process, 0:08:01.320 --> 0:08:04.080 and so they will look to see does the bank 0:08:04.200 --> 0:08:08.160 have a good risk management process, does it have the 0:08:08.280 --> 0:08:11.920 right board committees, does it have the right management committees 0:08:12.200 --> 0:08:16.760 looking at its risk decisions? Are there three lines of defense? 0:08:17.120 --> 0:08:22.120 And if the supervisors see the requisite process, they are 0:08:22.200 --> 0:08:26.120 very reluctant to make judgments about the actual decisions that 0:08:26.160 --> 0:08:29.000 are coming out of that process, and so they don't 0:08:29.040 --> 0:08:33.160 want to impose. They are reluctant to impose their own view. Oh, 0:08:33.280 --> 0:08:37.240 that is excessive risk, that is too much interest rate 0:08:37.360 --> 0:08:41.120 risk as opposed to we like the procedures that you're 0:08:41.160 --> 0:08:43.120 set up to manage interest rate risk. Just to be clear, 0:08:43.200 --> 0:08:47.840 though they certainly could under current law, they're allowed to 0:08:48.040 --> 0:08:51.280 say they're supposed to. Arguably, that's what current law is about, 0:08:51.320 --> 0:08:55.920 and the process approach being grafted onto this is an innovation. 0:08:56.240 --> 0:08:59.559 The purpose of safety and soundness law is really very 0:08:59.640 --> 0:09:04.559 much to address risk, and the process angle is born 0:09:04.640 --> 0:09:07.199 of the view that the best way to do that, 0:09:07.559 --> 0:09:10.079 the most efficient way to do that across a massive 0:09:10.120 --> 0:09:12.760 banking system, is just to make sure that the procedures 0:09:12.760 --> 0:09:15.199 are good. If the procedures are good, the problem will 0:09:15.240 --> 0:09:17.760 sort of take care of itself. So as long as 0:09:17.760 --> 0:09:22.959 you see the bank kind of debating its risk exposure internally, 0:09:23.240 --> 0:09:26.000 which you know seems to have been the case at SVB, 0:09:26.160 --> 0:09:28.319 and I know I brought this up in the previous episode, 0:09:28.400 --> 0:09:31.000 but you know, there were some internal documents which I've 0:09:31.000 --> 0:09:34.640 seen where they're talking about interest rate exposure and they're 0:09:34.679 --> 0:09:37.760 debating it, you know, with their asset liability committee and 0:09:37.840 --> 0:09:42.000 presumably with their risk specialists. But if they come to 0:09:42.040 --> 0:09:46.040 the conclusion that actually we're okay, the regulators are just 0:09:46.080 --> 0:09:47.800 going to look at that and take out on face 0:09:47.880 --> 0:09:50.640 value because the process is there, and they assume that, 0:09:50.760 --> 0:09:52.800 you know, the bank is kind of doing what it 0:09:52.840 --> 0:09:55.080 should be doing. Yeah, exactly, if you're not one of 0:09:55.120 --> 0:09:58.640 the big gesips, the global systemically important banks, and you're 0:09:58.679 --> 0:10:02.160 not in the stress testing route game, I think that 0:10:02.360 --> 0:10:04.679 is what tends to happen. It doesn't have to happen, 0:10:04.720 --> 0:10:07.240 As Joe said, they're certainly going to be examples where 0:10:07.240 --> 0:10:11.880 supervisors exercise them independent judgment. But there is a tendency 0:10:12.080 --> 0:10:16.000 still in the supervisory process to look at compliance with 0:10:16.040 --> 0:10:19.520 the rules, to check for processes, and if you see 0:10:19.520 --> 0:10:22.280 compliance with the rules and you see processes in place, 0:10:22.720 --> 0:10:26.719 to give the bank a clean bill of health, as 0:10:26.720 --> 0:10:28.880 it were. And so the question is how did we 0:10:28.920 --> 0:10:31.720 get to this place, because actually this was an innovation. 0:10:32.200 --> 0:10:35.640 At one point, we used to do safety and soundness 0:10:35.679 --> 0:10:41.079 substantive supervision without much bright line rules at all. Capital 0:10:41.160 --> 0:10:45.280 rules date only to the mid eighties, and this focus 0:10:45.400 --> 0:10:49.440 on process is really an innovation from the nineteen nineties. 0:10:49.760 --> 0:10:52.439 And so it's part of what I think is the 0:10:52.520 --> 0:10:56.079 toxic brew of regulatory supervisory policies that brought us the 0:10:56.120 --> 0:10:59.160 two thousand and eight crisis. And we still sort of 0:10:59.200 --> 0:11:05.160 have supervision guided by this nineties approach, and I think 0:11:05.160 --> 0:11:09.680 the SVB failure is one of several really significant examples 0:11:09.800 --> 0:11:14.160 of post two thousand and eight supervisory failure where the 0:11:14.240 --> 0:11:20.600 supervisors are still focused on process and too unwilling to 0:11:20.640 --> 0:11:24.600 make substantive judgments, reflecting the sort of approaches that were 0:11:24.600 --> 0:11:27.880 developed by primarily the greenspan fed but also the Ludwig 0:11:27.880 --> 0:11:31.520 occ in the nineties. Can you explain what it was 0:11:31.720 --> 0:11:34.440 or what happened in the nineties, was a directive that 0:11:34.640 --> 0:11:38.120 came down what caused this philosophical shift or just maybe 0:11:38.200 --> 0:11:40.880 mechanical shift in the approach to supervisory Let me start 0:11:40.880 --> 0:11:43.240 with the eighties, actually just with the birth of the 0:11:43.320 --> 0:11:47.240 capital rules. So what's the baseline. So supervision, safety and 0:11:47.280 --> 0:11:50.280 signed supervision dates all the way back to the nineteenth century, 0:11:50.520 --> 0:11:53.480 and so the way that the government's managed the incentive 0:11:53.520 --> 0:11:58.040 misalignment between bank shareholders and managers and bank depositors and 0:11:58.080 --> 0:12:03.920 the public has been through discretionary supervisory oversight safety and 0:12:03.960 --> 0:12:07.800 soundness oversight. That's been the bywords of federal law since 0:12:07.840 --> 0:12:11.760 the nineteen thirties. And the way that supervisors would do 0:12:11.800 --> 0:12:14.280 their job is they would make judgments about the riskiness 0:12:14.280 --> 0:12:18.520 of banks assets and the riskiness of banks leverage and 0:12:18.640 --> 0:12:22.680 the amount of capital and they would write letters, and 0:12:23.760 --> 0:12:27.240 they would job bone and they would take enforcement actions. 0:12:27.520 --> 0:12:29.640 They would issue cease and desist orders if they thought 0:12:29.640 --> 0:12:31.520 banks were under capitalized, or like in the case of 0:12:31.559 --> 0:12:35.120 Silicon Valley Bank, they would tell Silicon Valley Bank that 0:12:35.160 --> 0:12:37.920 it had to shorten its duration risk. That's what supervisors 0:12:37.920 --> 0:12:41.920 would do. In the eighties. You have a moment that's 0:12:41.960 --> 0:12:45.520 quite similar to today in that the banking system business 0:12:45.520 --> 0:12:48.720 model comes under a lot of pressure for macroeconomic reasons. 0:12:49.000 --> 0:12:51.640 Inflation goes up and then interest rates go way up 0:12:51.640 --> 0:12:54.680 because of the Vulker shock. This causes the yield curve 0:12:55.040 --> 0:12:57.839 to change in a way that's very ugly for a bank. 0:12:57.880 --> 0:13:00.640 Because a bank's business is a positive net interest margin. 0:13:00.679 --> 0:13:02.320 You earn more on your assets than you pay on 0:13:02.360 --> 0:13:07.520 your liabilities, and your liabilities are short duration, and so 0:13:07.800 --> 0:13:10.640 if interest rates go way up quickly, you can end 0:13:10.720 --> 0:13:13.200 up in a position where you are paying more on 0:13:13.200 --> 0:13:15.360 your liabilities than you are in your assets, which is 0:13:15.720 --> 0:13:17.559 which is going to run right through your capital. That's 0:13:17.600 --> 0:13:20.480 not a profitable business model. This happened in the eighties 0:13:20.520 --> 0:13:25.160 and a lot of banks became undercapitalized and supervisors were 0:13:25.160 --> 0:13:29.400 swamped with cease and desist orders and supervisory directives to 0:13:29.480 --> 0:13:32.000 banks all over the place to raise more capital. Some 0:13:32.080 --> 0:13:35.400 banks sued. One bank was able to prevail in the 0:13:35.440 --> 0:13:39.480 fifth Circuit, and Congress intervened and pass a new law 0:13:39.760 --> 0:13:45.200 authorizing forward looking, ex anti brightline capital rules for the 0:13:45.280 --> 0:13:49.920 first time and saying that capital judgments of supervisors can't 0:13:49.960 --> 0:13:52.760 be second guests by courts. And so, in a sort 0:13:52.760 --> 0:13:55.599 of accidental way, you have the birth of capital regulation. 0:13:56.000 --> 0:14:00.400 The supervisors are overwhelmed and there are laws, suits, and 0:14:00.440 --> 0:14:03.199 you have Congress saying, actually, just write a rule that 0:14:03.800 --> 0:14:06.120 the banks all have to comply with so that you 0:14:06.160 --> 0:14:08.600 don't have to get into litigation over whether this bank 0:14:08.679 --> 0:14:12.160 or that bank is undercapitalized. Fast forward a few years. 0:14:12.480 --> 0:14:15.600 You get these rules, and you get you get Basel one, 0:14:16.000 --> 0:14:19.480 you get an effort to align these rules internationally, and 0:14:19.600 --> 0:14:23.480 you get Alan Greenspan as fed share. Going into the nineties, 0:14:23.520 --> 0:14:26.440 the banking system starts to transform. You have the emergence 0:14:26.480 --> 0:14:32.600 of large, complex banking institutions, and you have a lot 0:14:32.680 --> 0:14:36.440 of soul searching in Washington and the Federal Reserve at 0:14:36.440 --> 0:14:40.160 the occ about whether supervisors are really up to the 0:14:40.280 --> 0:14:46.200 task of assessing the risk taking of these new large 0:14:46.280 --> 0:14:50.200 complex financial institutions, these large complex banking organizations that we 0:14:50.320 --> 0:14:53.920 never had in this country before through the traditional means, 0:14:54.240 --> 0:14:57.120 or whether we should actually embrace these new rules that 0:14:57.160 --> 0:15:01.240 have developed up and rely primarily on the rules and 0:15:01.440 --> 0:15:04.520 shift supervisors to a task that they are that they're 0:15:04.560 --> 0:15:07.920 more capable of performing. This is this is very self 0:15:07.960 --> 0:15:10.160 conscious for the policymakers at the time. You can go 0:15:10.200 --> 0:15:12.560 back and read some of Alan Greenspan's speeches about the 0:15:12.640 --> 0:15:15.640 changes that he's making, and he basically thinks that, especially 0:15:15.640 --> 0:15:18.200 for large banks, supervisors are just not going to be 0:15:18.240 --> 0:15:20.000 able to do it the way they used to. What 0:15:20.120 --> 0:15:24.400 we need are capital rules that require shareholders to have 0:15:24.520 --> 0:15:27.280 enough skin in the game, and then the shareholders will 0:15:27.280 --> 0:15:30.720 do it. The shareholders were supervised banks. And so Alan 0:15:30.760 --> 0:15:34.320 Greenspan says, it's not about needing net less regulation. It's 0:15:34.360 --> 0:15:37.480 about whether it should be public sector regulation or private 0:15:37.520 --> 0:15:40.520 sector regulation, and we need to reorient the banking system 0:15:40.560 --> 0:15:43.960 so that we have more private regulation. So that obviously 0:15:44.000 --> 0:15:48.280 goes terribly wrong in two thousand and eight. It's not funny, 0:15:48.440 --> 0:15:51.640 but this happens over and over and over it does, 0:15:51.680 --> 0:15:54.120 and so by two thousand and eight you have supervisors 0:15:54.160 --> 0:15:58.280 have more or less unilaterally disarmed. They have shifted to 0:15:58.800 --> 0:16:03.480 enforcing the capital rules. The banking agencies are relying almost 0:16:03.640 --> 0:16:07.360 entirely on the capital rules for making judgments about whether 0:16:07.400 --> 0:16:10.320 a bank has adequate capital for the risks that it's taking, 0:16:10.880 --> 0:16:14.880 and instead they are looking at processes. And there's a 0:16:14.920 --> 0:16:18.760 real theory behind this. The theory is, in order for 0:16:19.400 --> 0:16:22.480 market regulation to work, private regulation to work, there has 0:16:22.520 --> 0:16:24.720 to be disclosure, and so a bank has to be 0:16:24.760 --> 0:16:27.720 transparent about the risks that it's taken. A bank can't 0:16:27.720 --> 0:16:30.040 be transparent about the risks that it's taking if it 0:16:30.040 --> 0:16:33.200 doesn't have processes to monitor and disclose those risks. And 0:16:33.240 --> 0:16:36.080 so the job for supervisors is to make sure banks 0:16:36.080 --> 0:16:39.520 are monitoring their risks and disclosing them to the shareholders 0:16:39.560 --> 0:16:42.280 so that the shareholders can discipline the banks. And by 0:16:42.280 --> 0:16:46.120 two thousand and eight, supervisors have stopped bringing in cease 0:16:46.160 --> 0:16:48.760 into sists. There's no safety and soundness enforcement actions against 0:16:48.800 --> 0:16:50.720 any of the major banks, any of the major banks 0:16:50.720 --> 0:16:52.800 that take TART for years running up to two thousand 0:16:52.840 --> 0:16:54.920 and eight, because if they're in compliance with the rules 0:16:54.920 --> 0:16:58.280 and they're disclosing to the market, the judgment is that 0:16:58.440 --> 0:17:02.240 system is going to work. What goes wrong is that 0:17:03.000 --> 0:17:05.600 bank shareholders have an incentive to take much more risk 0:17:05.640 --> 0:17:07.600 than is in the interests of the government or depositors. 0:17:07.760 --> 0:17:10.800 It's sort of basic economic stuff. The shareholders have an 0:17:10.840 --> 0:17:13.959 incentive to extract wealth from the depositors and from the public. 0:17:14.200 --> 0:17:16.480 And so the shareholders is gonna be much more comfortable 0:17:16.840 --> 0:17:19.240 with a lot more risk than the public should be. 0:17:19.520 --> 0:17:21.199 And so if you're going to rely on them to 0:17:21.320 --> 0:17:22.919 monitor risk, cake and you're going to get a much 0:17:23.000 --> 0:17:26.520 riskier bank. And this is Silicon Valley Bank story. I mean, 0:17:26.560 --> 0:17:29.040 it's very much Silicon Valley Bank story. So you get 0:17:29.040 --> 0:17:33.200 two thousand and eight, and you get a modification after 0:17:33.320 --> 0:17:36.119 two thousand and eight, which is stress tests, but a 0:17:36.119 --> 0:17:40.439 lot of ordinary day to day supervision continues to be 0:17:40.720 --> 0:17:44.520 I think procedurally oriented. And you see this with the 0:17:44.600 --> 0:17:48.159 London whale, and you see this with the fake account 0:17:48.240 --> 0:17:50.880 scandal at Wells Fargo, both of which I think are 0:17:51.040 --> 0:17:55.399 really important data points for outside observers to think about 0:17:55.600 --> 0:17:59.080 how much did we fix supervision after two thousand and eight, 0:17:59.080 --> 0:18:01.679 how much do we move away from the green span 0:18:01.800 --> 0:18:06.440 nineties cocktail of bright line capital rules and procedural oversight 0:18:06.520 --> 0:18:10.560 oriented to shareholder discipline. And I think the answer is 0:18:10.600 --> 0:18:13.360 for the small banks that are not subject to stress tests, 0:18:13.520 --> 0:18:15.919 and even for the big banks that are subject to 0:18:15.960 --> 0:18:20.160 stress tests, outside of the stress testing regime, we still 0:18:20.200 --> 0:18:24.840 have a lot of procedural oversight. So this is actually 0:18:24.840 --> 0:18:28.359 a very quick mechanical question. But for a bank like 0:18:28.480 --> 0:18:33.000 Silicon Valley Bank, is it as supervisor? Is it a 0:18:33.040 --> 0:18:34.919 panel like hot Like what do you know? Do you 0:18:34.920 --> 0:18:37.119 have a sense of like how many people? I mean, 0:18:37.160 --> 0:18:40.760 because it's someone at the SFED presumably, Now assume there's 0:18:40.760 --> 0:18:43.800 thousands of banks in California probably or at least hundreds, 0:18:43.840 --> 0:18:46.400 Like what kind of just like human resources could even 0:18:46.480 --> 0:18:50.200 go currently to paying attention to Silicon Valley a bank 0:18:50.240 --> 0:18:54.119 like Silicon Valley Bank. So there are thousands of supervisors 0:18:54.119 --> 0:18:58.000 across the federal system. Okay, they're split across three agencies, 0:18:58.040 --> 0:19:01.680 the Federal Reserve, the Federal Deposit Insurance Corporation, and the 0:19:01.760 --> 0:19:07.280 Controller of the Currency. They split up responsibility for supervising bank. 0:19:07.359 --> 0:19:11.120 Silicon Valley Bank is a state chartered bank and it's 0:19:11.160 --> 0:19:13.719 a member of the Federal Reserve system. So as a result, 0:19:13.760 --> 0:19:16.480 as you say, it's the FED that would have responsibility 0:19:16.560 --> 0:19:20.520 at the federal level for primary responsibility for supervision if 0:19:20.560 --> 0:19:23.560 there's always overlap, So the FDIC has some ability to 0:19:24.160 --> 0:19:26.880 come in because it's insuring the deposits obviously, and then 0:19:27.000 --> 0:19:30.040 it's very involved now. But the day to day job 0:19:30.119 --> 0:19:35.440 here is FED personnel in San Francisco who are really 0:19:35.480 --> 0:19:39.760 actually exercising delegated authority of the Board of Governors, which 0:19:39.960 --> 0:19:45.800 is the federal agency with the power to supervise member banks. 0:19:46.320 --> 0:19:48.560 And the Reserve Bank of San Francisco is actually a 0:19:48.560 --> 0:19:53.000 federal bank of Federal corporation, and so it's it's facilitated, 0:19:53.040 --> 0:19:56.520 it's helping the board, and the board has supervisory staff 0:19:56.600 --> 0:19:59.880 that are supposed to sort of oversee what is going 0:20:00.200 --> 0:20:02.520 at the reserve banks. So is San Francisco doing a 0:20:02.560 --> 0:20:04.840 good job, And obviously at the top of that is 0:20:04.880 --> 0:20:25.120 Michael Barr, the Vice champer supervision. I'm going to have 0:20:25.160 --> 0:20:28.520 some more questions on the San Francisco FED and the 0:20:28.720 --> 0:20:32.920 fhlbs as well. But just going back to the evolution 0:20:33.480 --> 0:20:37.720 of bank capital rules. So one of the big things 0:20:38.440 --> 0:20:41.040 that happened, and you sort of outlined it in the 0:20:41.160 --> 0:20:43.000 lead up to the two thousand and eight crisis, but 0:20:43.080 --> 0:20:46.639 like it definitely hardened after two thousand and eight. Is 0:20:46.640 --> 0:20:52.080 this idea that banks should be holding more bonds in general, 0:20:52.160 --> 0:20:55.439 the safest bonds, so, you know, US treasuries in the 0:20:55.440 --> 0:20:59.679 case of US banks, maybe agency mortgage backed securities that 0:20:59.760 --> 0:21:03.400 are implicitly guaranteed by the US government, things like that 0:21:03.480 --> 0:21:08.719 for their regulatory capital and liquidity buffers. And it seems 0:21:08.760 --> 0:21:11.800 to me like that probably made a lot of sense 0:21:11.960 --> 0:21:15.440 in the low inflation environment of two thousand and eight, 0:21:15.960 --> 0:21:18.879 But now that you have the FED raising rates, you 0:21:18.920 --> 0:21:22.840 have a lot of volatility. It seems like these bonds 0:21:23.160 --> 0:21:26.919 might not be I don't think safe is the right word, 0:21:27.080 --> 0:21:31.720 but not as unproblematic as maybe we imagine them to 0:21:31.920 --> 0:21:35.520 once be. Could you talk a little bit more about 0:21:35.560 --> 0:21:39.200 basically how we built the modern banking system on top 0:21:39.240 --> 0:21:41.760 of a bedrock of bonds that are presumed to be 0:21:41.920 --> 0:21:45.760 somewhat stable in price. So I think that you're right 0:21:45.840 --> 0:21:52.879 to sort of highlight the appeal of treasuries and agencies 0:21:53.040 --> 0:21:57.240 to both bankers and supervisors in the wake of two 0:21:57.640 --> 0:22:01.680 and eight. Bank that loads up on treasury, that's, like, 0:22:01.760 --> 0:22:05.240 you know, very wholesome. It seems very wholesome for a 0:22:05.280 --> 0:22:08.120 bank to do right. The government likes that. The government 0:22:08.200 --> 0:22:10.480 is like banks that buy treasuries since the Civil War, 0:22:10.920 --> 0:22:16.560 and so it's gonna cut against even the most ambitious 0:22:16.600 --> 0:22:20.000 and confident, you know, safety and soundness overseas. It's going 0:22:20.040 --> 0:22:23.320 to cut against their their impulses to to sort of 0:22:23.359 --> 0:22:26.000 fault a bank for loading up on treasuries. I mean, 0:22:26.119 --> 0:22:29.200 that's that seems that seems that's a good thing, right, 0:22:29.400 --> 0:22:33.840 And so it's it that helps to explain part of 0:22:33.840 --> 0:22:38.919 what's happening here. And it's also true that banks do 0:22:39.040 --> 0:22:44.960 have the ability to weather usually a fair amount of 0:22:45.440 --> 0:22:50.240 interest rate losses on their assets. So many banks think 0:22:50.280 --> 0:22:56.920 of themselves as structurally hedged against interest rate increases because 0:22:57.680 --> 0:23:02.040 while their assets, if they have law assets like long 0:23:02.119 --> 0:23:04.960 dated treasuries, that's those are going to lose value when 0:23:04.960 --> 0:23:08.600 the interest interest rates rise, their liabilities or deposits, and 0:23:08.640 --> 0:23:12.600 their deposits are sticky, they don't pass through, and so 0:23:12.720 --> 0:23:16.160 actually their deposit funding becomes much more valuable when interest 0:23:16.280 --> 0:23:18.960 rates rise. So if in a zero interest rate environment, 0:23:19.400 --> 0:23:23.200 interest rates or zero deposit rates or zero deposits aren't 0:23:23.200 --> 0:23:25.719 that useful. You're getting a little benefit that you have 0:23:25.760 --> 0:23:28.920 deposit funding. But if interest rates go way up, you're 0:23:28.960 --> 0:23:30.800 not gonna if you're the bank, you're not actually going 0:23:30.880 --> 0:23:34.000 to be forced to raise your extracting rents right from 0:23:34.000 --> 0:23:35.960 the deposit public. But you're not going to be forced 0:23:35.960 --> 0:23:39.200 to raise your deposit rates. And this is actually strengthening 0:23:39.240 --> 0:23:42.479 your business. And Silicon Valley Bank is going to think, yeah, okay, 0:23:42.520 --> 0:23:44.960 so we take some we take some hits on our 0:23:45.040 --> 0:23:48.600 long assets, but actually our net interest margin is going 0:23:48.640 --> 0:23:51.000 to position, is going to main strong, our deposits are 0:23:51.000 --> 0:23:53.560 going to be much more valuable, and we're just gonna 0:23:53.640 --> 0:23:56.320 we're gonna work through, you know, year to eighteen month 0:23:56.400 --> 0:23:59.200 period and be totally fine, right, which it seems like 0:23:59.240 --> 0:24:01.640 they lose was sort of the assumption that they had, 0:24:01.680 --> 0:24:04.680 like they knew they it wasn't great and maybe even 0:24:04.960 --> 0:24:07.320 technical and solvent, but I mean, I think it was 0:24:07.359 --> 0:24:09.440 in this leek one of matt Leview's news and letters, 0:24:09.480 --> 0:24:12.240 he's like, this was actually like a very profitable time 0:24:12.320 --> 0:24:13.800 for them, like it did. It would have been fine 0:24:13.840 --> 0:24:17.200 if everyone's did. And presumably their expectation was, well, we're 0:24:17.200 --> 0:24:19.359 just making a lot of income right now. So the 0:24:19.359 --> 0:24:20.760 fact that we took it a hit on the asset 0:24:20.800 --> 0:24:23.520 side is not really well. They had they had a 0:24:23.560 --> 0:24:27.000 specific estimate in one of those internal documents where they said, 0:24:27.160 --> 0:24:30.840 we could shorten duration, but that would mean an eighteen 0:24:30.880 --> 0:24:33.560 million dollar hit to our net interest margin in one 0:24:33.640 --> 0:24:36.720 year alone, going up to like thirty six million over 0:24:36.760 --> 0:24:39.760 the next three years. So they knew that if they 0:24:39.840 --> 0:24:44.400 reduced duration, they would be sacrificing earnings to some extent. Yeah, 0:24:44.400 --> 0:24:46.119 I mean, I think it's fair to say that in 0:24:46.160 --> 0:24:48.440 twenty twenty one they were making huge profits because this 0:24:48.520 --> 0:24:52.679 strategy was really working. Interest rates went way up, and 0:24:53.000 --> 0:24:57.440 I think it would have impaired their profitability, but they 0:24:57.440 --> 0:25:00.240 were They weren't wrong to think that they were what 0:25:00.400 --> 0:25:02.760 structurally hedge, even though they had no interest rate hedges 0:25:02.840 --> 0:25:06.040 or anything, by virtue of the fact that they would 0:25:06.080 --> 0:25:08.920 slowly be able to replace their treasuries with much higher 0:25:08.960 --> 0:25:12.359 yielding treasuries while being able to pay depositors very little. 0:25:12.400 --> 0:25:15.400 And we also, you know, we talked about this deposit 0:25:15.520 --> 0:25:18.760 betas on a recent episode with Joe Batte and why 0:25:18.800 --> 0:25:20.920 they're off at low and what if His points was like, well, 0:25:20.960 --> 0:25:23.840 you know, you have a bank, you have an individual 0:25:23.960 --> 0:25:26.880 has an account at Chase or something like that they're 0:25:26.960 --> 0:25:29.679 providing a lot of services along with that. People are 0:25:29.760 --> 0:25:33.040 not that inclined to move their checker account just because 0:25:33.040 --> 0:25:35.520 the interest rate doesn't bump up a little bit. And 0:25:35.560 --> 0:25:40.040 I imagine for Silicon Valley depositors these companies, the whole 0:25:40.080 --> 0:25:43.560 story about Silicon Valley Bank was all of the products, 0:25:44.080 --> 0:25:47.240 the startup specific products that they are that they offered, 0:25:47.400 --> 0:25:51.000 which presumably to their mind, insulated them to some extent 0:25:51.040 --> 0:25:55.120 against losing deposits. Absolutely, and I mean in the case 0:25:55.160 --> 0:25:58.000 of SVB, there was also what an antitrust law we 0:25:58.160 --> 0:26:03.880 call tying, where accompany ties one product to another product. Yeah, 0:26:03.960 --> 0:26:07.479 so the bank would require that if you wanted to 0:26:07.520 --> 0:26:09.919 borrow from the bank, that you would have Is that 0:26:10.000 --> 0:26:12.360 unusual because people some people obviously some people are like, oh, 0:26:12.440 --> 0:26:14.000 that's weird, and some people are like no, of course, 0:26:14.080 --> 0:26:17.040 like any commercial loan. But is that unusual in your view? 0:26:17.240 --> 0:26:20.520 I don't think it is unusual. There are strict rules 0:26:20.520 --> 0:26:24.359 about bank tying in other areas. But my understanding is 0:26:24.359 --> 0:26:28.920 that banks are explicitly permitted to tie deposit account services 0:26:28.960 --> 0:26:32.199 to lending services. And historically it was core to the 0:26:32.200 --> 0:26:36.240 banking business that you were the depository institution for the 0:26:36.800 --> 0:26:40.879 for your borrowers that that went together, and you know, 0:26:40.920 --> 0:26:44.320 we've moved away from that technology lots of things have 0:26:44.440 --> 0:26:47.520 allowed people to borrow from banks that aren't the banks 0:26:47.560 --> 0:26:51.600 where they bank at. But a long time, the idea 0:26:51.720 --> 0:26:53.719 was that there's a lot of synergies between that no 0:26:53.760 --> 0:26:55.880 one's going to be in a better position to determine 0:26:55.880 --> 0:26:59.159 how much to lend to you than your own banker. 0:27:00.800 --> 0:27:03.000 I imagine like there was also a bit of a 0:27:03.160 --> 0:27:06.879 prestige element to banking at SVOB as well, given that, 0:27:07.160 --> 0:27:10.120 you know, it was so popular among a particular type 0:27:10.119 --> 0:27:13.720 of tech slash VC person. But you know, Joe touched 0:27:13.720 --> 0:27:18.040 on the episode we did on deposit betas with Joe Abode. 0:27:18.320 --> 0:27:21.360 But I'd love to hear from you, like why didn't 0:27:21.480 --> 0:27:26.000 people pull more deposits from a bank that was essentially 0:27:26.480 --> 0:27:29.439 paying them nothing? Because to some extent, this is the 0:27:29.480 --> 0:27:35.080 big question, like why did SVB have so many deposits? 0:27:35.080 --> 0:27:39.119 Well into twenty twenty two, at which point we started 0:27:39.160 --> 0:27:41.760 to see some of the I guess most interest rate 0:27:41.800 --> 0:27:45.800 sensitive parts of the economy, ie the tech industry lose 0:27:45.840 --> 0:27:47.720 a bunch of money and have to pull funds. But 0:27:48.000 --> 0:27:50.840 why why were people accepting of that for so long. 0:27:51.760 --> 0:27:55.359 So we mentioned a couple of the sort of rational 0:27:55.480 --> 0:27:59.840 explanations that you know, there's a strong brand, you want 0:27:59.840 --> 0:28:02.800 to they're they're lending to you, they've required you to 0:28:02.880 --> 0:28:05.960 keep deposit there. But you can't discount the fact here 0:28:06.040 --> 0:28:08.919 that a big piece of this was a lack of 0:28:08.920 --> 0:28:13.560 sophisticated financial management on the part of startup companies that 0:28:14.520 --> 0:28:17.800 maybe didn't have CFOs, didn't have anybody on their teams 0:28:17.920 --> 0:28:22.600 with any experience in managing cash. They're focused on their 0:28:22.640 --> 0:28:26.960 their business and it's very hard to justify five hundred 0:28:26.960 --> 0:28:29.960 million dollar bank account balance, and I think we have 0:28:30.040 --> 0:28:33.000 one example of that that just there's there's no reason 0:28:33.080 --> 0:28:37.600 for that's that's very bad management. No well run, mature 0:28:37.680 --> 0:28:42.680 company would operate in that way. Among other things, you 0:28:42.680 --> 0:28:45.640 you you have a huge uninsured deposit risk that we 0:28:45.840 --> 0:28:48.360 that we saw, but you're also just giving up lots 0:28:48.360 --> 0:28:52.400 of return. You could have that money invested in laddered 0:28:52.520 --> 0:28:58.640 treasury bills or something. And significantly more so, there's a 0:28:58.680 --> 0:29:00.640 huge amount of money that's just been left on the 0:29:00.640 --> 0:29:03.120 floor here there, and it's you know, it doesn't it 0:29:03.120 --> 0:29:06.040 doesn't really make sense. So we have to understand that 0:29:06.080 --> 0:29:09.520 these that these customers, despite having lots of money, are 0:29:09.560 --> 0:29:12.480 not actually very sophisticated. So I want to go back 0:29:12.480 --> 0:29:15.400 to the supervisory question and ask about it, kind of 0:29:15.400 --> 0:29:18.880 come at it from a different angle. You know. Obviously 0:29:18.920 --> 0:29:21.880 the two thousand and eight two thousand and nine crisis 0:29:22.880 --> 0:29:26.000 was very focused on the asset side of the business 0:29:26.080 --> 0:29:28.520 and were these really high quality assets? And then part 0:29:28.600 --> 0:29:31.000 of the reasonable bunch of banks failed is because the 0:29:31.080 --> 0:29:34.080 assets like weren't very good that they held. And as 0:29:34.120 --> 0:29:38.000 people have discussed with Silicon Value Bank, a lot of 0:29:38.000 --> 0:29:40.640 the issues. Yes, maybe they made a wrong bet on 0:29:40.760 --> 0:29:43.800 treasuries or they put too much, but is the flightiness 0:29:43.800 --> 0:29:46.480 of the deposits? Can you talk a little bit more. 0:29:46.520 --> 0:29:51.840 I know that regulators do bucket deposits from the most 0:29:51.840 --> 0:29:53.960 sticky to the least sticky, but could you talk a 0:29:54.000 --> 0:29:58.120 little bit about like how good supervisory in a sort 0:29:58.120 --> 0:30:02.520 of like active pre nineties way might have approached the 0:30:02.720 --> 0:30:06.480 uniformity of SVB deposits and the risk of them all 0:30:06.560 --> 0:30:08.360 leaving at once. Yeah, I mean I think that if 0:30:08.360 --> 0:30:11.560 you showed svb's balance sheet to supervisor brought up during 0:30:11.560 --> 0:30:14.480 the new deal system. So let's say it's nineteen seventy five, 0:30:15.240 --> 0:30:20.560 they would be horrified at the enormous concentration of uninsured 0:30:20.600 --> 0:30:27.520 deposits controlled by a group of businesses with very similar 0:30:27.680 --> 0:30:30.880 risks to their business and so all of your depositors, 0:30:30.880 --> 0:30:33.000 and this is something that and this is something that 0:30:33.040 --> 0:30:36.600 a new deal era supervisor would be familiar with from 0:30:36.600 --> 0:30:39.400 press experience. Well, I think I think it would have been. 0:30:39.720 --> 0:30:42.040 They would have been unfamiliar with it in the sense 0:30:42.080 --> 0:30:44.440 that it would have been so unusual back then everyone 0:30:44.440 --> 0:30:47.440 would have looked at it and said, WHOA, this bank 0:30:47.680 --> 0:30:50.840 has a very unstable deposit base. It would not have 0:30:50.920 --> 0:30:54.160 been novel to view this with concern. It would be 0:30:54.400 --> 0:30:57.560 it would be even more concerning because of how risk 0:30:57.600 --> 0:30:59.760 it would have been to operate a bank in this 0:30:59.800 --> 0:31:02.040 way at that point in time, when when, of course, 0:31:02.040 --> 0:31:05.200 people still remembered the bank runs of the thirties much 0:31:05.240 --> 0:31:08.400 more than they do than they do today. And you know, 0:31:08.440 --> 0:31:10.640 part of what went wrong at SVP is it's not 0:31:10.680 --> 0:31:13.760 just that they had ninety seven percent or something in 0:31:13.920 --> 0:31:18.040 uninsured deposits, but that all of their depositors were going 0:31:18.040 --> 0:31:20.520 to withdraw at the same time. And so there's sort 0:31:20.520 --> 0:31:24.200 of a classic issue in the banking business always is 0:31:25.400 --> 0:31:27.360 in what circumstances am I going to be subject to 0:31:27.400 --> 0:31:30.240 a deposit train? You know, you get to model your 0:31:30.240 --> 0:31:35.960 deposits as sticky if you're a bank, because over time, 0:31:36.000 --> 0:31:38.280 for the banking system, the deposit base is always sort 0:31:38.280 --> 0:31:40.160 of growing. I mean, with the exception of over the 0:31:40.240 --> 0:31:42.520 last year where monetary policy is trying to shrink the 0:31:42.520 --> 0:31:45.160 money supply, but you know, over time it's a constant 0:31:45.160 --> 0:31:47.760 growing base, and so deposits are really, in some simes 0:31:47.760 --> 0:31:50.920 a very long duration asset, except if you're the one 0:31:50.960 --> 0:31:52.880 bank that experiences a drain to the rest of the 0:31:52.880 --> 0:31:55.400 system where everyone withdraws from you, and so if your 0:31:55.480 --> 0:31:59.160 customers are all going to face hardship your depositories or 0:31:59.720 --> 0:32:02.960 depot it was hard. The same time, you really can't 0:32:02.960 --> 0:32:06.360 treat yourself as structurally hedged. You're the opposite of structurally 0:32:06.400 --> 0:32:09.360 head And that's what that's what Silicon Valley Bank found out, 0:32:09.680 --> 0:32:12.160 is it thought that, oh, you know, when when my 0:32:12.320 --> 0:32:15.040 assets lose value, my deposits will become more valuable. But 0:32:15.080 --> 0:32:17.880 actually all their depositors started to draw down their accounts 0:32:17.960 --> 0:32:21.200 and the opposite happened, and so they were just very 0:32:21.480 --> 0:32:25.240 very long low interest rates, silicon valuing. The whole business 0:32:25.320 --> 0:32:27.520 model was tied to low interest rates, I think to 0:32:27.600 --> 0:32:30.000 an extent that they did not appreciate, and to an 0:32:30.040 --> 0:32:34.720 extent supervisors clearly didn't depreciate, but maybe weren't even thinking 0:32:34.800 --> 0:32:37.320 as hard about as they might have in an earlier 0:32:37.480 --> 0:32:40.640 period where they were more empowered to make those sorts 0:32:40.680 --> 0:32:43.640 of judgments. Yeah, this is exactly what I said on 0:32:43.720 --> 0:32:46.600 our episode with Dan Davies. It was interest rate exposure 0:32:46.840 --> 0:32:49.880 kind of squared, but just on the deposit side. Because 0:32:49.960 --> 0:32:53.320 to me, this is kind of the most novel or 0:32:53.520 --> 0:32:56.480 interesting thing about all of this, because we know that 0:32:56.560 --> 0:33:00.080 a lot of banks have unrealized losses on bonds, and 0:33:00.320 --> 0:33:04.800 you know too. Seems like broadly they've been managing their 0:33:04.880 --> 0:33:08.920 interest rate risk so far. But with SVB, the big 0:33:09.000 --> 0:33:15.560 difference was that group of highly concentrated, extremely unreliable depositors 0:33:15.800 --> 0:33:19.120 who themselves had significant interest rate exposure. And we're pulling 0:33:19.240 --> 0:33:24.040 money over the past years. So what could regulation do 0:33:24.800 --> 0:33:28.080 on that front? So I guess instead of the asset side, 0:33:28.120 --> 0:33:33.080 looking more at the liability side. Yeah, So what you 0:33:33.160 --> 0:33:38.880 want to see is a coherent asset liability management strategy 0:33:38.960 --> 0:33:42.920 for a bank, and so a bank that anticipates deposit 0:33:43.040 --> 0:33:47.560 trains for a bank that has flighty deposits, and there 0:33:47.600 --> 0:33:49.280 are many banks that can fall into this category. This 0:33:49.400 --> 0:33:51.800 is something that regulator supervisors do think a lot about. 0:33:52.160 --> 0:33:54.840 If you're in that category, then you need to hold 0:33:55.440 --> 0:33:59.479 liquid assets that you can sell and that at their 0:33:59.520 --> 0:34:04.080 fair market value to cover the withdrawals. And so part 0:34:04.080 --> 0:34:07.440 of the problem here is that Silicon Valley Bank did 0:34:07.560 --> 0:34:12.840 not actually have available for sale securities at fair market 0:34:12.960 --> 0:34:17.560 values sufficient to cover the withdrawals, and so the fix 0:34:17.760 --> 0:34:22.000 for this would have been to have much less duration 0:34:22.160 --> 0:34:25.160 in the in the in the asset portfolio, or many 0:34:25.400 --> 0:34:28.600 many more reserves. This is the same problem, by the way, 0:34:28.800 --> 0:34:31.959 that took down silver Gate Bank and to some extent, 0:34:32.080 --> 0:34:38.360 Signature Bank. They had deposit bases that were flighty, that 0:34:39.120 --> 0:34:43.840 their depositors suffered and their deposit the banks experienced deposit 0:34:43.960 --> 0:34:46.680 drains because they were concentrated in a group of people 0:34:46.760 --> 0:35:07.040 that were exposed to interest rate hiking. I just realized 0:35:07.080 --> 0:35:11.600 I promised to ask about discount lending and the fhlb's 0:35:11.640 --> 0:35:14.600 the Federal Home Loan Bank. So you know, in theory, 0:35:14.840 --> 0:35:17.480 when you have this type of banking crisis or you know, 0:35:17.640 --> 0:35:21.000 some sort of liquidity issue with the financial institution. You 0:35:21.080 --> 0:35:23.840 would expect them to either go to the FED, to 0:35:23.960 --> 0:35:27.440 the discount window and for all blots listeners, we recorded 0:35:27.600 --> 0:35:30.239 an episode on this a month or two ago, or 0:35:30.480 --> 0:35:33.719 they can borrow from the FHLBS. And some of the 0:35:33.840 --> 0:35:41.120 talk out there is that SVB got cut off by FHLB. 0:35:42.080 --> 0:35:45.560 Why would that have happened and why wouldn't those two 0:35:45.719 --> 0:35:49.480 lenders of last resort do everything they can in order 0:35:49.560 --> 0:35:51.920 to step in and support the bank. Or is it 0:35:52.000 --> 0:35:54.439 the case that at some point, you know, maybe they're 0:35:54.480 --> 0:35:56.520 talking to the FDIC and they just say this is 0:35:56.680 --> 0:35:59.480 untenable and no matter how much money we provide, like 0:35:59.600 --> 0:36:01.200 the bank is not going to be able to get 0:36:01.280 --> 0:36:04.520 up and running again. So I'm speculating a bid here, 0:36:04.760 --> 0:36:08.880 and you might want to talk to the FHLB expert 0:36:08.960 --> 0:36:12.120 in the legal academy, K Judge, who's my colleague. But 0:36:12.880 --> 0:36:16.640 the FHLBS are not a lender of last resort in 0:36:16.680 --> 0:36:21.719 the way that the FRBs are. Right, the FHLBS they 0:36:21.840 --> 0:36:25.400 do provide sort of lender of second to last resort 0:36:25.520 --> 0:36:30.239 services to their members, but they are much more operated 0:36:30.800 --> 0:36:34.280 by their members, and they pay dividends to their members 0:36:34.719 --> 0:36:37.920 than the FRBs. So the FRBs were set up in 0:36:37.960 --> 0:36:42.719 a similar model, but today basically function as public banks, 0:36:42.760 --> 0:36:45.400 so they have no interest in profits or anything like that, 0:36:45.840 --> 0:36:48.480 and they're willing to sort of take one for the 0:36:48.600 --> 0:36:51.880 team in a way that the fhlbs are not. So 0:36:51.960 --> 0:36:54.319 I think it's a mistake to look at the FHLBS 0:36:54.320 --> 0:36:56.680 and say, oh, well, you really ought to have lent 0:36:56.880 --> 0:37:00.200 into an insolvent institution and took on that potential. Rik k. 0:37:00.520 --> 0:37:03.160 The FRBs are are are are wary of that for 0:37:03.880 --> 0:37:06.879 various reasons we could get into. But the flhol bes 0:37:06.920 --> 0:37:09.399 have even more reason to sort of to pull back. 0:37:09.600 --> 0:37:12.840 We definitely have to do an FHLB episode at some 0:37:12.960 --> 0:37:15.160 point with k Judge, because I don't know much about 0:37:15.200 --> 0:37:17.600 them at all, and it definitely it's been too long 0:37:17.719 --> 0:37:22.520 or it's far too long without having yeout having her on. 0:37:22.920 --> 0:37:26.120 I want to ask another dimension. You know, people pointing 0:37:26.239 --> 0:37:31.320 to the twenty eighteen law change to Dodd Frank that 0:37:31.840 --> 0:37:34.759 seemed to exempt banks like Silicon Valley Bank from some 0:37:34.960 --> 0:37:37.279 of these liquidity requirements that you were talking about, like 0:37:37.440 --> 0:37:39.960 do you have enough liquid assets? Hey? Could you sort 0:37:39.960 --> 0:37:42.600 of characterize the change that was made there in b 0:37:43.239 --> 0:37:46.600 had that not been in place, like is that is 0:37:46.680 --> 0:37:48.839 that change that was made in twenty eighteen? Does does 0:37:48.920 --> 0:37:51.480 that tell the story of the demise? Had the old 0:37:51.680 --> 0:37:54.279 Dodd Frank laws remained in place for a bank the 0:37:54.360 --> 0:37:57.160 size of Silicon Valley Bank, would they have been able 0:37:57.200 --> 0:37:59.160 to weather the storm? You can never know for sure, 0:37:59.239 --> 0:38:03.960 but I would suggests yes, that is decisive. And so 0:38:04.560 --> 0:38:07.120 this brings us back to how the government responded to 0:38:07.200 --> 0:38:09.960 the two thousand and eight failure of supervision regulation. And 0:38:10.120 --> 0:38:13.840 they responded with a set of tiered new requirements. And 0:38:14.680 --> 0:38:17.680 for the very biggest banks you had the CCR stress 0:38:17.719 --> 0:38:21.799 testing regime, and for all of the banks with more 0:38:21.840 --> 0:38:25.120 than fifty billion dollars of assets, you had stress testing 0:38:25.280 --> 0:38:30.520 as well as collection of other enhanced prudential standards. And 0:38:31.400 --> 0:38:35.800 this new cocktail of regulation and supervision was geared towards 0:38:35.920 --> 0:38:38.520 preventing a repeat of something on the scale of two 0:38:38.560 --> 0:38:41.680 thousand and eight. And so we weren't going to impose 0:38:41.760 --> 0:38:43.719 this on the whole banking system, on all of the 0:38:43.760 --> 0:38:46.640 sort of smaller banks. But the thinking was, if we 0:38:46.719 --> 0:38:50.800 could just really get back to serious government oversight of 0:38:51.280 --> 0:38:54.279 banks over fifty billion dollars that would really go a 0:38:54.360 --> 0:38:58.800 long way towards preventing another another calamity like two thousand 0:38:58.840 --> 0:39:03.480 and eight. And what happened was immediately there was litigation 0:39:03.719 --> 0:39:08.719 over the threshold for this new regulatory supervisory cocktail. And 0:39:08.800 --> 0:39:11.279 so this fifty billion dollars threshold came under a lot 0:39:11.360 --> 0:39:15.000 of political pressure from the banking agencies and from various 0:39:15.040 --> 0:39:19.520 people in Washington, and the bank lobby fought a battle 0:39:19.640 --> 0:39:22.840 over many years to raise the threshold. One of the 0:39:23.920 --> 0:39:28.480 important figures lobbying for raising the threshold was the CEO 0:39:28.560 --> 0:39:31.600 of Silicon Valley Bank. He was growing his bank and 0:39:31.680 --> 0:39:34.960 he did not want to grow his bank into additional 0:39:35.080 --> 0:39:41.000 regulatory and supervisory requirements. And in twenty eighteen, after winning 0:39:41.120 --> 0:39:44.680 over people in both parties to this cause of raising 0:39:44.719 --> 0:39:49.239 the threshold, Congress changed the law and the threshold moved up. 0:39:49.600 --> 0:39:52.120 A bunch of thresholds moved around, but the relevant threshold, 0:39:52.280 --> 0:39:54.920 I think, moved up to two hundred and fifty billion dollars. 0:39:55.000 --> 0:39:58.400 And the result was Silicon Valley Bank and its peers 0:39:58.480 --> 0:40:03.200 had successfully exempted themselves from the enhanced prudential standards that 0:40:03.280 --> 0:40:06.600 Congress had created. After two thousand and eight, and the 0:40:06.760 --> 0:40:10.160 result was you didn't have the stress testing, which is 0:40:10.200 --> 0:40:14.360 the primary means by which supervisors now exercise substantive judgment 0:40:14.400 --> 0:40:18.759 about risks, and so you had a much more I 0:40:18.880 --> 0:40:24.319 think light touch process focused oversight that allowed rule compliant 0:40:24.440 --> 0:40:29.759 balance sheet configurations like svbs to go relatively unchallenged. But 0:40:29.880 --> 0:40:32.680 if you had been in the in the Enhanced Prudential 0:40:32.719 --> 0:40:35.440 Standards bucket, I think that they would have been challenged. 0:40:35.480 --> 0:40:38.200 They would have been challenged through all of these additional 0:40:38.320 --> 0:40:43.279 rules and also supervisory programs. And it's unlikely. You can 0:40:43.320 --> 0:40:46.960 never know, but it's unlikely that they could have taken 0:40:47.080 --> 0:40:52.279 so much duration risk and not raised capital earlier, been 0:40:52.440 --> 0:40:55.399 permitted to go for so long in a position where 0:40:55.440 --> 0:40:59.040 their liquidation value was possibly negative. Yeah, since we're on 0:40:59.160 --> 0:41:01.880 the topic of the blame game, I mean, one of 0:41:01.920 --> 0:41:05.520 the things that you see people saying now is that, well, 0:41:06.239 --> 0:41:09.120 it's the Fed's fault. The FED kept interest rate slow 0:41:09.280 --> 0:41:13.240 for too long, and it basically forced people to assume 0:41:13.360 --> 0:41:17.560 additional duration risk in order to seek out yield. And 0:41:17.680 --> 0:41:21.040 I think, you know, financial repression is a real thing. 0:41:21.200 --> 0:41:23.680 But on the other hand, you cannot ignore the individual 0:41:23.719 --> 0:41:27.480 actions of one or a few specific banks, their managers, 0:41:27.520 --> 0:41:31.920 their shareholders, and their depositors. But how would you describe 0:41:32.480 --> 0:41:38.520 the overall monetary policies role in the current predicament. So 0:41:39.040 --> 0:41:43.080 overall monetary policy is of course central to the current predicament. 0:41:43.200 --> 0:41:47.120 But there's a sort of false dichotomy underline, the view 0:41:47.400 --> 0:41:51.480 that somehow it's like monetary policy happening up here at 0:41:51.520 --> 0:41:54.680 the FED that's then causing problems down here at the 0:41:54.760 --> 0:41:57.080 banking system. The whole thing is monetary policy. The whole 0:41:57.080 --> 0:41:59.279 reason we have banks is monetary policy. Banks are creating 0:41:59.320 --> 0:42:01.960 the money supply, and the question is how much money 0:42:02.000 --> 0:42:04.200 do we want to be created. So it would be 0:42:04.280 --> 0:42:07.800 the tail wagging the dog if we had to change 0:42:07.920 --> 0:42:10.000 our judgment about how much money should be created because 0:42:10.040 --> 0:42:12.640 like somehow the system couldn't create that amount of money 0:42:12.960 --> 0:42:16.320 safely and stably. We have a broken system if it 0:42:16.440 --> 0:42:18.719 can't create the amount of money that the FMC says 0:42:18.800 --> 0:42:22.800 is appropriate for macroeconomic conditions. And so I would not 0:42:23.000 --> 0:42:26.600 blame the FOMC for thinking that we need to adjust 0:42:26.800 --> 0:42:29.360 the amount of money that the banking system and the 0:42:29.480 --> 0:42:31.920 financial system are creating. I would blame the banking and 0:42:32.040 --> 0:42:35.000 financial system and the banking and financial laws. If they're 0:42:35.080 --> 0:42:39.000 incapable producing the amount of money and changing the amount 0:42:39.040 --> 0:42:42.760 of money they're producing overtime consistent with the fomc's directives 0:42:42.760 --> 0:42:45.400 and the needs of the economy, that's a really big problem. 0:42:45.480 --> 0:42:48.200 And it does look like we're facing that problem now, 0:42:48.280 --> 0:42:50.640 where in the coming months we could be in quite 0:42:50.640 --> 0:42:54.320 a predicament where the FOMC may make the judgment and 0:42:54.440 --> 0:42:56.160 we can brack it whether it would be the correct 0:42:56.239 --> 0:42:59.919 judgment that the economy needs less money and the banking 0:43:00.080 --> 0:43:05.759 system may be incapable of functioning properly under that directive. 0:43:06.000 --> 0:43:08.520 And that's the flip side of, you know, the suggestion 0:43:08.560 --> 0:43:10.760 that the banking system should also be able to function 0:43:10.880 --> 0:43:13.640 under the judgment that interest rate should be zero and 0:43:13.760 --> 0:43:16.080 function in a way that is sustainable over time. And 0:43:16.160 --> 0:43:18.440 so to the extent that is what's going on, I 0:43:18.560 --> 0:43:21.520 think it's a real indictment not of monetary policy, like 0:43:21.600 --> 0:43:23.680 the high level decision about how much money we need, 0:43:23.920 --> 0:43:28.160 but the structure being unable to follow through to execute 0:43:28.200 --> 0:43:30.839 on those decisions. Leveman on that was an amazing answer, 0:43:31.040 --> 0:43:34.120 That was an amazing conversation that was so helpful that 0:43:34.320 --> 0:43:37.160 was so helpful and so clear. I've said it every 0:43:37.239 --> 0:43:40.080 time we talked to I'm like, oh, I've finally had 0:43:40.360 --> 0:43:42.480 There's been a few more, but I so that was 0:43:42.520 --> 0:43:45.319 a that was very good. I'll just leave it there. 0:43:45.560 --> 0:43:48.360 I really I really appreciate that. So thank you so 0:43:48.440 --> 0:43:50.600 much for coming back on up. Thank you so much 0:43:50.640 --> 0:44:06.560 for having me that, Tracy. I love that whole conversation, 0:44:06.800 --> 0:44:08.960 starting from the very end, which I think is a 0:44:09.080 --> 0:44:15.440 really excellent way to sort of reconceptualize the monetary policy problem, 0:44:15.560 --> 0:44:17.480 which is that if the FED is going to be 0:44:17.640 --> 0:44:22.000 tasked with like the sort of like big sweep macro management, right, 0:44:22.320 --> 0:44:27.200 getting employment inflation at its targets and so forth, in theory, 0:44:27.680 --> 0:44:31.200 we want to have a financial system that can operate 0:44:31.360 --> 0:44:35.239 under any you know, operate relatively safely under whatever rates 0:44:35.280 --> 0:44:37.960 the FED deems to be appropriate. Absolutely, I mean, I 0:44:38.080 --> 0:44:42.520 do think there is a fundamental tension between monetary policy, 0:44:42.880 --> 0:44:46.279 which you know, like the big thing monetary policy does 0:44:46.440 --> 0:44:48.879 is basically impact the price of bonds and then having 0:44:48.960 --> 0:44:52.759 the financial system and banks specifically have to hold a 0:44:52.880 --> 0:44:56.160 bunch of bonds as part of their capital and liquidity mandates. 0:44:56.239 --> 0:45:00.200 Like that tension is there, but there are way is 0:45:00.280 --> 0:45:03.520 to manage it such that we can avoid failures and 0:45:03.760 --> 0:45:08.480 also provide for the effective implementation of monetary policy as 0:45:08.520 --> 0:45:11.200 a whole. The other thing that really struck me is 0:45:11.560 --> 0:45:15.480 this is kind of an incentives episode, right, And I 0:45:15.600 --> 0:45:19.799 thought Lev's point about basically outsourcing a lot of bank 0:45:19.920 --> 0:45:25.880 supervision to private shareholders and expecting them to you maybe 0:45:26.200 --> 0:45:29.359 press the brakes on risk when things start to get 0:45:29.440 --> 0:45:33.080 out of hand, that was a really interesting one. And SVB, 0:45:33.360 --> 0:45:35.280 I think, is going to end up as a classic 0:45:35.360 --> 0:45:38.960 case where you know, there was an acknowledgement that there 0:45:39.200 --> 0:45:43.400 was an issue here. There was the asset liability duration mismatch, 0:45:43.480 --> 0:45:46.120 and there was too much exposure to long bonds and 0:45:46.520 --> 0:45:50.480 too much of an assumption that deposits would be around 0:45:50.600 --> 0:45:54.640 forever or that they might even return or start growing again, 0:45:55.280 --> 0:45:59.520 and it was a conscious decision to pick up net 0:45:59.640 --> 0:46:02.440 interest margin, or at least it looks like that. I'm 0:46:02.480 --> 0:46:04.719 sure more will come out over the course of all 0:46:04.760 --> 0:46:06.880 of this, but for now, it certainly seems like it 0:46:07.040 --> 0:46:09.640 was a decision to do that. That's really interesting, like 0:46:09.760 --> 0:46:12.680 thinking about like, it's pretty fascinating that a lot of 0:46:12.760 --> 0:46:16.680 these like capital requirements and ratios and regulations that we 0:46:16.760 --> 0:46:19.240 think of as courtA how we manage the banking system 0:46:19.440 --> 0:46:21.799 are all pretty young, and that for the most part, 0:46:21.920 --> 0:46:24.320 for a long time in history it was like active 0:46:24.360 --> 0:46:29.239 super advisory people making judgments based on the operations of 0:46:29.280 --> 0:46:33.200 the bank, whether their decisions on loans and deposits were healthy. 0:46:33.640 --> 0:46:36.080 But it does make total sense that it is sort 0:46:36.120 --> 0:46:39.080 of like you know, I hate usually where it's like neoliberal, 0:46:39.640 --> 0:46:43.960 but that like that the role of supervisors would essentially 0:46:44.120 --> 0:46:49.759 transform to making sure that shareholders were getting adequate information. 0:46:49.920 --> 0:46:51.960 That you start with the assumption that the market is 0:46:52.000 --> 0:46:54.400 the best regulator, and then what is the role of 0:46:54.440 --> 0:46:56.360 the government well, and the role of the government, and 0:46:56.480 --> 0:46:59.319 that point is to make sure that market regulators get 0:46:59.360 --> 0:47:02.600 good information. Like that seems like a very like big 0:47:02.760 --> 0:47:05.279 theme that like you could like characterize across a lot 0:47:05.360 --> 0:47:07.960 of different industries, a lot of different government and then 0:47:08.040 --> 0:47:10.360 the failure of just like well, the problem is like 0:47:10.520 --> 0:47:13.480 shareholders can lose everything and not be accountable for the 0:47:13.600 --> 0:47:16.719 spill over when a bank fails. And so yeah, the 0:47:16.880 --> 0:47:19.400 need perhaps to get back to a type of supervisory 0:47:19.719 --> 0:47:22.560 that actually takes decisions into own hands and rather than 0:47:22.640 --> 0:47:25.839 just outsourcing it. You know, I realized we didn't even 0:47:25.920 --> 0:47:28.880 get into FED checking accounts, which is one of them. 0:47:28.960 --> 0:47:30.400 Now that's the next step. So then I think the 0:47:32.160 --> 0:47:37.520 series is if all deposits post SVB are presumed to 0:47:37.560 --> 0:47:40.640 be ensured, which almost seems like implicitly the case, now 0:47:41.239 --> 0:47:44.640 why do we have private deposit taking institution? So I 0:47:44.719 --> 0:47:46.600 kind of think that's the next one in this series 0:47:47.040 --> 0:47:49.319 to look at, Well, what does this tell us now 0:47:49.360 --> 0:47:52.320 about even the point of private deposit taking institutions? And 0:47:52.360 --> 0:47:54.840 should there be a point? I mean, that is what 0:47:54.960 --> 0:47:56.840 I was kind of hinting at in the intro. But 0:47:56.920 --> 0:47:59.320 we'll just have to leave that. You know, it was 0:47:59.360 --> 0:48:01.879 a trailer not for this episode but for the next one, 0:48:02.160 --> 0:48:04.440 so plenty more to come, but shall we leave it there? 0:48:04.520 --> 0:48:07.920 Let's leave it there. This has been another episode of 0:48:08.040 --> 0:48:10.880 the Odd Thoughts podcast. I'm Tracy Alloway. You can follow 0:48:10.960 --> 0:48:14.319 me on Twitter at Tracy Alloway and I'm Joe wisn't All. 0:48:14.360 --> 0:48:17.320 You can follow me on Twitter at the Stalwart, follow 0:48:17.400 --> 0:48:21.239 our guest Levmnond on Twitter at levmanand follow our producers 0:48:21.280 --> 0:48:25.280 Carmen Rodriguez at Carmen Arman and Dash Bennett at dashbot 0:48:25.719 --> 0:48:28.960 Check out Bloomberg's podcast son to the handle at podcasts, 0:48:29.120 --> 0:48:31.480 and for more Odd Lots content, go to bloomberg dot 0:48:31.520 --> 0:48:34.440 com slash odd Lots, where we post all the transcripts. 0:48:34.600 --> 0:48:37.080 Tracy and I have a blog and weekly newsletter that 0:48:37.120 --> 0:48:40.000 comes out every Friday. Go there and sign up. Thanks 0:48:40.040 --> 0:48:40.440 for listening.