Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=

# Dollar Erosion: Understanding the Loss of Reserve Currency Status

Authors:   
Discussant: None  
Video: https://www.youtube.com/watch?v=iWc7Gz_UmKc&t=0s


## Talk (00:00:00 – 01:11:35)

[00:00:00] Everybody, welcome to the first seminar of winter quarter. We had conflicts with distinguished visitors the first two weeks, that's why we're starting on the third week. We are very happy to have  
[00:00:13] Arvind Krishnamurthy talk today and and let me just remind people so we're going to go for an hour and 15 minutes till 1:45.  
[00:00:20] And Arvind said it's fine to interrupt with questions. So we're doing it kind of like econ finance seminar rules.  
[00:00:26] Although if people get out of hand, feel free to uh you know, >> [laughter] >> recapture the floor.  
[00:00:33] Yes. So I think that two of your co-authors are online, Zing Yang Cheng and Robert Richman and Hanolustig who's at another seminar might be able to join us. And so the title of the talk is dollar erosion, understanding the loss  
[00:00:48] of reserve currency status, very very topical. Take it Okay.  
[00:00:52] >> from here. All right, thank you Valerie. Thank you for having me here. Thank you all for coming to listen. All right, so here's what I'm going to uh do.  
[00:01:01] I'm going to walk you through some data. Uh particularly around the tariff shock last year.  
[00:01:08] Um and what I use you've probably seen some of this data, but I would just want you to put put a bunch of data together and show you that the behavior in the dollar and treasuries was was weird.  
[00:01:21] Right? It was uh a real change in correlations relative to what we have seen in the past and I'll show you some data on that. I think just looking quickly at what has happened in the last two days with the Greenland shock, it sort of looks similar again. So there's  
[00:01:34] been something going on around that and I'm going to interpret as I talk you through the data, I'm going to interpret that the shifts in correlation are due to shifts in the perception of  
[00:01:47] dollar safe assets as the world's reserve asset. That's what I'm going to argue to you.  
[00:01:52] Um I'm not going to argue to you that the world has switched completely to another safe asset as the world reserve asset, but if you imagine that what has happened over time is we were in a world where we were sort of 100% sure that the  
[00:02:06] dollar asset was the safe asset and the events [clears throat] over the last year have shifted some of those probabilities.  
[00:02:15] That's a way of understanding some of the changes in asset prices.  
[00:02:19] So I'll walk you through some data around that.  
[00:02:21] Then I'm going to shift over uh a model calibration exercise.  
[00:02:27] And basically I want I want to try to put some numbers down on how to think about end points for where this could go. And as I said, I don't think this is where the world is, but I'm going to put an  
[00:02:41] end point of suppose the reserve asset demand for dollar safe assets just disappears completely.  
[00:02:49] Um where would things end up? And in particular, I want to walk you through I'll I'll take you through a computation of where uh equilibrium interest rates in the US will end up.  
[00:03:01] Uh an exercise on to where the real exchange rate in the US will end up. And just to tell you where I'm going, I'm basically going to be comparing across two steady states.  
[00:03:11] Uh one in which there's world demand for dollar assets as safe assets and one in which I turn that off to zero. And I'll compare and see how asset prices compare across these two.  
[00:03:23] Um the comparison I think is useful for a couple of reasons. One, you know, there's been what I called uh we called the Miran hypothesis, which is that the dollar safe at the fact that the US has  
[00:03:38] been the reserve currency has led to an appreciated dollar and that has had, you know, various knock-on effects.  
[00:03:44] So one of the things you can take away from the computation I'm going to do is I'm going to effectively tell you how much higher the dollar was has been because of uh particularly reserve asset demand. Okay?  
[00:03:58] I'll I'll give you a number on that. And then the other side of looking at that is to say, suppose we've been moving this probability down from one of we're always in this safe asset world.  
[00:04:10] >> [clears throat] >> Um where where's the end point? Right? If we just let this all play out, where will exchange rates re-equilibrate?  
[00:04:20] Um so I'll take you through exchange rate and interest rate computations. And the interest rate computation in particular is very closely related to another computation that I'm going to take you through, which is uh what is the wealth loss to the US?  
[00:04:31] I'm going to try to put some numbers on what it'll mean in in terms of a measured wealth. And you'll see what I mean by measured wealth if we end up in this state in which reserve asset demand completely disappears. Okay? Does this  
[00:04:45] give you any leverage? I'm guessing not cuz they're steady states about Uh-huh.  
[00:04:49] how likely this is to happen and what what it means and what it would what require of some other currency to take the place of the US.  
[00:04:55] >> Yeah, so I'll I'll I'll I'll one could imagine I I I think there's an exercise that we could do and we've thought about, which is as I said, you can interpret the data as shifting probabilities.  
[00:05:07] If I impose a model on that and tell you this is where things would end up if the probability goes to one, then I can I can put some I can actually probably extract a time series of these probabilities. We haven't done that. Um  
[00:05:21] and in part you'll the reason you'll see when I do the exercise part, we're missing many things.  
[00:05:26] Like I'm going to do an exercise where I'm going to take reserve asset demand for dollars away.  
[00:05:32] Um and I'm just going to show you what happens. You should really think about it almost as like a it's a partial derivative of the world with respect to one thing.  
[00:05:41] Now surely if that one thing happens, 17 other things will happen, too. So I don't I don't know if I want to go too far down that road because I'm not sure I have the model that has all the things that would happen in that state of the  
[00:05:54] world. Right? So I I view this more as an exercise to give me an idea of magnitudes as to what this means for the world. Can you talk about the time period this occurs over? Yeah, I mean I'll I'll take you through the data and you'll see the  
[00:06:08] data I'm talking about. As opposed as I think you might be asking, over to what time period might reserve asset demand disappear?  
[00:06:16] Uh and there's nothing in the data that's going to tell me that. In fact, you know, I think those of us who have looked at quantity data, it's hard to discern big shifts in dollar holdings.  
[00:06:27] Uh there's some uncertainty about the current reserve positions because we don't measure that very well. Uh but at present it doesn't look like quantities have shifted dramatically.  
[00:06:37] But I mean the the reason to look at asset prices, which is what I'm going to do, is asset prices are forward-looking. So if I tell you in two years it'll disappear, today's price will already pick that up.  
[00:06:46] And so I'm going to look at everything through the lens of asset prices.  
[00:06:51] Okay, with that let me walk you through some data. Um I mean this is the thing that I think is really striking if you paid attention to movements in dollars  
[00:07:02] around times of global financial stress. Um I hope you can see this. The top there, that's the the 2008 crisis and I'm plotting the VIX in orange against the value of the dollar.  
[00:07:17] And you know, the typical pattern during periods of financial panic is the VIX spikes and the dollar appreciates.  
[00:07:25] Right? There's sort of this flight to dollar, flight to treasuries thing that you see in the world. There are 5 to 10% in the Yeah. global financial crisis go Absolutely fair. You know, fairly significant movements in the in dollar exchange rates around these type of  
[00:07:39] periods historically. This is the COVID period. The correlation is not quite as strong, but it's still there.  
[00:07:46] It's interesting it's not quite as strong. Maybe something had already shifted and I'll I'll come back and say this.  
[00:07:51] This is the March to May 2025 after the tariff war.  
[00:07:55] And you can see it just goes in the totally the opposite direction. Right?  
[00:07:58] This is uh a a core correlation that looked like was playing out in asset markets, which is when the world blows up, people look to park their safe safety in dollar safe assets, looked like it  
[00:08:12] wasn't happening here. Right? Here's another version of this computation.  
[00:08:19] Uh I'm graphing here uh the exchange rate, euro-dollar exchange rate, that's the orange curve.  
[00:08:26] And then in blue here is the yield differential between um 10-year European government bonds and and uh treasury bonds.  
[00:08:37] And again, you see this pattern that the dollar depreciates, but the yield spread widens.  
[00:08:44] Right? In particular, it looks like treasury yields are rising relative to foreign yields.  
[00:08:51] And uh if you just to give some magnitudes here, The same thing be true of JGBs?  
[00:08:57] I haven't looked at JGBs. I'm So I can't I'm not going to uh tell you the number. I have some Japanese uh yields uh later in the talk, but I haven't looked specifically at this one.  
[00:09:09] Um this is true across a whole bunch of other currencies I looked at. I looked at it with respect to the UK, that was true there.  
[00:09:16] Um but I I don't have a full answer for you.  
[00:09:19] Just to give you an idea of magnitudes here, you know, if you if you look at these 10-year yield differences, they widen by around 50 basis points. If you use just long-run uncovered interest parity as a benchmark and you will move yields  
[00:09:31] by 50 basis points, present value that, the exchange rate should appreciate by 5%. Higher higher returns in the US.  
[00:09:40] Instead, the dollar depreciates by 6 and 1/2%. So again, a very clear sign that something is sort of odd around this period.  
[00:09:47] Here's another metric. So I'm guessing most of you have seen those two spreads.  
[00:09:52] I'm going to show you some other numbers that you probably haven't seen before.  
[00:09:56] Uh and talk you through them. This is a spread that, uh, we have, uh, constructed and looked at extensively.  
[00:10:04] It is take the yield on a one-year US government bond.  
[00:10:11] So, a Treasury bill. Compare that to the yield on, say, a Canadian government bond.  
[00:10:18] Now, if I just take that yield differences, I'm not doing apples to apples because one is in dollars and one is in Canadian dollars.  
[00:10:24] Take the Canadian dollar bond. Tack on a foreign exchange swap that converts the Canadian dollar cash flows into US dollars.  
[00:10:35] Now, the package then of the Canadian government bond with the FX swap is now in equivalently a dollar bond, right?  
[00:10:42] Uh, sort of a Canadian package to dollar bond, but it's a dollar bond. Compare that yield to the US Treasury yield.  
[00:10:50] Uh, and I'm going to do that as an average across the G10 countries. So, Japan is included in this one.  
[00:10:56] Why is that an interesting measure? Because it tells you something about world demand in particular for dollar assets, right? If if if what has happened and, uh, is historically, which is the the quantity data certainly  
[00:11:09] suggest, the world has heavily held safe assets in US dollars and particularly in Treasuries, then this will be an indication of something like excess demand for Treasuries relative to other  
[00:11:22] world safe assets. Uh, Mhm. The cost of swaps change.  
[00:11:27] You know, at at a at high frequencies, uh, um, there are financial intermediation frictions that can influence the, uh, FX hedging costs. Uh, that certainly happens at times.  
[00:11:41] This is I'm looking at a sample period from '88 to 2017. And just let me talk you through this. I don't think that's the main thing that that's going on here, right?  
[00:11:50] Uh, here's that spread. Um, so, I'll first notice that spread is basically typically positive. And the way to think of a positive means that foreign above, uh, Treasuries. So, there's been kind of a premium on  
[00:12:03] Treasuries historically. The premium blows up at certain times.  
[00:12:08] These look These are typically times where, you know, the US is in some type of or the world is in some type of recessionary conditions.  
[00:12:15] What's the first blowup? That's the, uh, a a recession, uh, in the US in '93. That's, um, you know, obviously the tech bubble burst thing, that's 2008.  
[00:12:26] Um, this thing averages about 22 basis points.  
[00:12:30] Not a huge number, um, but it blows up at certain times. I'll come back and want to talk to you about the 22 basis points and not what not a huge number means because I'm really constructing two assets that, you know, from a finance perspective, these things should  
[00:12:44] be zero, right? Um, and the fact that there's even 22 is is kind of remarkable.  
[00:12:50] >> [clears throat] >> Uh, and but the time series is also interesting cuz the time series goes up uh, during these periods of financial stress, which is very much consistent with  
[00:13:02] a way of thinking about flows into these dollar safe assets, in particular during times of financial stress, right? So, in 2008, there's a flight, you know, I showed you the VIX and the dollar, there's sort of a flight to Treasuries, which would exactly pick up in this type  
[00:13:16] of spread. So, this is one of the reasons we have used this spread. It's a sort of an attractive spread to to an a very fine-tuned level measure the specialness of dollar safe assets. So, Arvin, one of the >> Yeah. One of the other things that  
[00:13:29] happened in, uh, in April, Yeah. uh, was that, uh, the credit default swap Mhm.  
[00:13:36] on, uh, on dollars, >> Yeah. uh, you know, went way up. I think it would be interesting to include that in also because I'm it's a reflection of the fact that, uh, people no longer view Mhm. uh, US government bonds as safe.  
[00:13:50] >> just I'm going to show you some some numbers on that in just the next few slides, okay? So, um, this is a kind of a historical fact about 22 basis points.  
[00:14:01] Um, we've gone back as far as '88 because that's about as far as we can go back with FX swap data to construct this consistently.  
[00:14:08] I have a feeling if you went back further, we would see this back into the early '80s. We've done some comparisons of the UK and the US back into the '70s.  
[00:14:16] There the correlations look like they flip sign much more. So, if you think of the '70s as a period where the dollar wasn't established as sort of the world's reserve asset, that's very consistent with some of these patterns.  
[00:14:27] All right, let me flip bring you to the to around the terror shock.  
[00:14:32] Um, here's the same basis, uh, right around April. And what I want you to notice is that as the exchange rate, uh, is depreciating,  
[00:14:45] this thing goes negative, not positive, which is what it has historically been. So, again, it's very consistent with what looks like sort of a flight away from Treasuries.  
[00:14:57] This is one-year Treasuries, Bob, not long-term stuff. So, you know, I'm Shouldn't be default to It should This is not the thing that you would be worried about default. It's kind of remarkable that it goes negative.  
[00:15:10] Uh, when we first looked at this, this this is what sort of jumped out at us that that the the this looked >> you to interrupt for a second? So, the one-year Treasury, uh, credit default swap in April went from 16 basis points to 52.  
[00:15:24] Okay. 52 basis points is a lot. It's a big number.  
[00:15:27] >> Okay, bigger than 22. Absolutely. Let me show you some some other numbers as I go through them. Mhm. What makes you emphasize the short-term change as opposed to where it comes back as if you're thinking about this the Sorry. I I I agree. So, it I I When we  
[00:15:42] first looked at that, I was really struck by the opposite correlation. I think that's really what I want to emphasize. The thing moved in the wrong direction relative to historical correlations.  
[00:15:52] Then, here's the same construction. Um, and now I'm going backwards in time.  
[00:15:59] That around the blue line and the the the spread going down is March, April.  
[00:16:07] If you go backwards in time, uh, first of all, notice it turned negative sometime last summer, which, uh, is striking, you know, consistent with, uh, Bob, what you're saying.  
[00:16:21] Um, but it's really this spread has been going down since 2022.  
[00:16:27] Right? Like, you know, historical numbers, it bounces around as I showed you from uh, the previous graph, you know, averaging 22, but going up as high as one, one and a half percent in the financial crisis. So, if I just looked  
[00:16:40] at this pattern, I sort of stopped this at the beginning of 2023, this I would say this is in the range of normal.  
[00:16:46] But from, uh, 2022 down onwards, it's basically been falling.  
[00:16:51] And if you look at it in that light, what had happened in March, April was a event that marked a change in a correlation that already taken place.  
[00:17:04] Rather, the world had already shifted away to some extent in asset pricing land away from dollar safe assets. So, is there any reason to believe that the spreads should be be invariant to levels?  
[00:17:18] The level of interest rates? Yeah, I mean, we had a big surge in inflation.  
[00:17:21] Inflation started coming down, for example. Yeah. The Europe to Eurozone has likewise had, but they differ a bit from country to country. Mhm. And they had so similar things.  
[00:17:31] >> So, I mean, I I can answer >> short-term. So, I'm just saying Yeah.  
[00:17:34] Um, I can answer the question >> a theorem that I don't know about?  
[00:17:37] >> No, no, there's no theorem. I I Look, if if you want me to put a model on this, which we have a model of this, the the model would be something about demand for dollar safe assets. And then if you kind of unpack that and says, "Where is the demand for dollar safe assets coming from?" There are surely macro and  
[00:17:52] monetary factors underlying that, of [clears throat] which, you know, world growth, what is happening to interest rates around the world, uh, definitely play a role.  
[00:18:03] So, I'm sure that that's there. If I go back, I showed you the longer time series for which we have good data on, we can just take that time series and project that on interest rates. And if you wanted to, we could do that and sort of strip that out, not much, uh, happens  
[00:18:16] to this computation. But I think the other thing though, which is the the the number that really jumps out at me is the fact that it went basically down to zero and then eventually crossed zero, which, you know, if you told me we were moving  
[00:18:30] from 40 to 20 because of something about two relative monetary policies, yeah, but to go all the way the other side, that feels like some there's a fundamental change in it in, uh, a relationship.  
[00:18:42] >> We looked at Mhm. based on what you're saying.  
[00:18:45] The secular changes of interest in the background is growing problem of public debt. Yes. What the secular issue? Yes.  
[00:18:50] The great story of our time is Yes. the assault on taxation in Congress.  
[00:18:54] Absolutely. And then there's the timing with the tariffs, isn't it? Not just the sort of crazy, but they really are this crazy. So, it's really just an updating of priors interacting with the sort of deteriorating fiscal situation or thinking about the role of the secular  
[00:19:07] versus the sort of Yeah, so, th- this is sort of this is the sense in which I mean, I think to me, the way I interpret this is there's events that are moving slowly.  
[00:19:16] There many things that are going on in the background. The fiscal situation is surely one of them.  
[00:19:22] Uh, I'm sure there especially after tariff and there are questions about the Fed's independence that are probably playing out here. There's a bunch of more macro stuff that's playing out here. And I'm going to interpret this all through the lens of just put that all together and  
[00:19:35] put a probability on it. Uh, and that's about all I can say. I I can't unpack how much of each of these things is happening.  
[00:19:45] Um, in the paper, we do an exercise, we do a kind of a calibration exercise of what happens if you just double US debt?  
[00:19:54] Uh and make the fiscal situation worse. What what what type of effects are but I I'm not I'm uncomfortable ascribing this specifically to one factor or the other.  
[00:20:04] So there has to be some binding um constraints on arbitraging this situation, right? Uh so >> [clears throat] >> This is this is so >> read I was reading yesterday about uh the the Treasury swap spread. So it's  
[00:20:19] the difference between the fixed the fixed rate on the swap and the and the corresponding Treasury and and there are four different margins there that uh you know, that it are involved with with the  
[00:20:31] repo market and you know, and is it a is it a safe repo? Is it is it uh in um you know, in real terms? Um you know, and it would be interesting, I think, to  
[00:20:45] look at what you know, what what margins are active here because it you know, your your story about the um you know, the safe asset, you know,  
[00:21:00] doesn't address, you know, why you know, if there's you know, the big demand for safe asset, you say, "Okay, then the dollar you know, the dollar should be strong. The the the rate should be positive. The differential should be positive." But but there has to be some  
[00:21:13] reason why it's not arbitraged. Yeah, I mean, I think one one way of thinking about the arbitrage margin is the arbitrageurs, the banks for taking the other side to some extent, they surely are are effectively creating safe assets by shorting.  
[00:21:27] So if you think imagine the way the way to think about the world is there's a demand for dollar safe assets. It's provided by in different forms by the US government, by different parts of the private sector.  
[00:21:39] Suppose at the the quantities, there's still a spread. Then the the the financial system comes in and says, "We'll short, that is, create repo and buy some of the foreign assets to convert them into dollar safe assets."  
[00:21:53] If in if after all of that, there's still a spread, it tells you that there's there's some cost that's left that they're not fully dealing with. I think that relates to what Mike was asking as well. Um so that's how I kind of interpret uh these spreads. I'm going  
[00:22:07] to show you one more number, Bob, and this is related to your Treasury point.  
[00:22:11] We did another computation in which take a look at the the blue line here.  
[00:22:17] This is European safe rates relative to US repo on a swap basis.  
[00:22:23] Treasuries are off the table here. The these are uh repo rates um against safe collateral. This is primarily Treasury collateral, but these are repo rates where the underlying repo index is overnight.  
[00:22:38] I think it's a reasonable to think that we'd be worried about US government default over the next 30 years. This is overnight repo. That's the underlying index here.  
[00:22:48] I'm Yeah. Basically, don't believe that the CDS number you mentioned has any bearing on this construction directly. And you can see here that that blue line looks about the same as what I showed you for  
[00:23:01] one-year Treasuries. It is kind of noteworthy that it doesn't go through zero.  
[00:23:05] Right? It remains it it follows the same pattern, but it remains slightly above zero during this period. And that tells you something that surely within the US market the safe asset market within the US, there's been some discrimination away  
[00:23:19] from Treasuries and towards uh stuff like overnight repo, which I think is consistent with worries about the fiscal position of the government.  
[00:23:27] But if there the zooming out, it's still the case that overall, we've had a deterioration in the spread relative to um in blue line euro asset.  
[00:23:40] >> up while that's going down? Pardon? Cash, paper money, $100 bills. Is there any relationship between like more $100 being issued as rates go down? I haven't looked. I have no idea. You do you don't know. I have  
[00:23:53] Okay, cuz some people think that's worth focusing on, namely Scott Sumner writes about that stuff all the time. Now, I don't know whether you like it or not, believe it or not, but he's been making a very strong argument on that point.  
[00:24:04] I guess I mean, the lens I've always taken through this to all of this stuff is this is about global portfolios.  
[00:24:11] Um large reserve asset managers, private sector agents around the world looking for And is China going to show up somewhere down the road on your data?  
[00:24:21] >> No. No, I this is the last bit of data I'm going to show you. I'm going to look at a model in a second. Okay. No China?  
[00:24:26] I mean, China is an outcome of what I mean, this is in the background, China must be driving some of these prices.  
[00:24:32] >> Okay. Um I don't but I'm I don't have any quantity on China to tell me that this is being driven by China or the other.  
[00:24:40] That's the So when I looked at this, I mean, I think this is what I I it I'm comfortable saying it's not Treasury CDS. It's a broad statement about safe assets in the US that's coming off of the blue line.  
[00:24:53] Um The other comparison I'm doing here is the same spread relative to yen safe assets. So take a yen safe asset rate, swap it into dollars, compare it to the US repo rates.  
[00:25:06] Right? And what I want you to notice is basically, there's a decline as well.  
[00:25:11] Um it's a decline that has left uh a more positive spread at the end.  
[00:25:17] So, you know, you could interpret that as some of the thing there's been shifts in Maybe what's happened is some set of investors have shifted or are expected to shift into euro safe assets. There's less of that that's happened with Japan.  
[00:25:31] Um The red line is Danish kroner. I thought Danish kroner was an interesting spread to look at because it's basically pegged back to the euro, but it doesn't have the size of the euro market. Uh you can see the same pattern.  
[00:25:45] Uh you know, fairly big drops in in in uh spreads, but levels that are higher.  
[00:25:50] So, you know, I if I look at these three numbers one second, Marcus, my interpretation is there's they look the the financial data look like it's been deterioration of the dollar with respect to everything.  
[00:26:03] Um I'm not sure that the financial data tell me that there's been a rotation particularly to euros or to anything else. It's just sort of diminishing of the dollar safe asset premium with respect to all assets.  
[00:26:14] Yeah. So, Arvind, uh everything you said makes perfect sense in a world where the relative supplies of the different assets in the world are the same.  
[00:26:24] Mhm. I want to ask about too much of a good thing. If you keep increasing a really terrific liquid safe haven like the US Treasury, Yeah. Yeah, eventually people are going to say, "Look, I you know, I don't need that many safe asset. Not enough. If you  
[00:26:39] want me to hold more, I'm I'm only going to do it if you pay me an extra yield." >> Mhm.  
[00:26:44] Yeah. That's not really I mean, I don't know. It it gets interpretation, but I'm not turning away from the asset. I'm just saying I yes, it's terrific. I love it, but I can only take so much of my wealth put it into uh lock it into uh a safe asset like the Treasury. I got to  
[00:26:58] do something with the rest of the money. Yeah, fair enough. Okay, so I mean, there's sort of two things going on here. There's supply and demand. There's got to be a there's got to be an upper limit to the amount people want to >> Absolutely.  
[00:27:09] Perfect. There I mean, there's clearly supply and demand here that's going on.  
[00:27:11] >> has to turn inelastic at some point, and people have been arguing this for 30 years. We're about to get there. Yeah, yeah. We haven't. I mean, I I have to say when you know, I we look at the data before 2020, there's a spread.  
[00:27:25] We have way more debt um post 2020. Remarkably, there was still a spread. So, supply went up, but it sure looked like um world demand kept up with the  
[00:27:38] increase in supply until recently. So, um I mean, in response to Daryl's question, I don't think that this action could be supply. Supply hasn't changed that much over this period. I don't have any instruments here, and I'm not going to run any regressions.  
[00:27:52] >> come back in every quantity. Yeah, yeah. So, we we it could be that the right way to think about this is as supply has changed, the world is shifting to thinking about an equilibrium where it would be better  
[00:28:05] if demand spread itself across currencies.  
[00:28:09] That's entirely possible. That That is uh we were tracing a world in which demand was stable, supply was expanding. At some point, supply got to a point where the world recalibrated demand and said, "Let's move to a new equilibrium in which we're  
[00:28:23] not 100% dollars, and now let's spread ourselves." That would be consistent with this. That's the sense in which I mean, I was telling you put a probability one to a lower probability.  
[00:28:32] I think that would be entirely consistent. You can't tell a pure fixed demand uh expanding supply story to hit this type of thing.  
[00:28:39] >> the debt is 80% this year? Absolutely. And that that fits exactly with you know, if I told you that what's happened is we were 100% in dollar safe assets, and we're not so sure, and we start to diversify into gold.  
[00:28:52] That fits very consistently with uh this uh with what I'm seeing here.  
[00:28:57] >> Does the rise of crypto play into this as well?  
[00:29:00] No. >> [laughter] >> Gold, I feel comfortable that there's something in that, but crypto, I have really no idea. Have we seen because of the sanctions, probably that made a bit of a major role. Yeah, I'm sure. Yeah.  
[00:29:13] Mhm. The official [clears throat] sector is moving out, and it's replaced by less stable funding, isn't it? Yeah, absolutely. So, I mean, if you say if you again, if if the world demand was a mixture of a bunch of these things, some of it was demand that previously never  
[00:29:27] thought about sanctions, suddenly thinks about sanctions, and that's the shift that's happening. That's entirely possible. This is I I there's a sense in which I feel like I can't here there's a I cannot unpack how much is equilibrium shift, how much is some of the stuff.  
[00:29:42] What I can tell you is it looks like that's uh thrown away clearly. Move away from the macro stuff.  
[00:29:47] >> Yeah. There's sort of two kind of background things going on on reserve currency status. One is the efficiency of making two trades in dollars versus pairwise trades and other things. The markets are too too thin.  
[00:30:01] The bid ask spreads are too large. And then there's what you've estimated as convenience yields with a series of papers. And that's basically that it's convenient to hold this as collateral in a variety of ways. It's kind of a stock of I don't know if you want to call it  
[00:30:16] quasi quasi buffer or something of that sort.  
[00:30:19] Are you able to sort of get at any of that in here? Again, the state of the economy tell you. I mean your model >> I do think they're connected. I mean if if you ask me to I would tell you the following which is that the payment system relies on moving  
[00:30:34] collateral around. It relies on a stock of safe assets in the background to facilitate payment flows. So those two things are definitely connected. The fact that the dollar market has been the the liquid asset  
[00:30:48] safe asset of choice in part because it's large, liquid, and safe supports a payment system that allows for easy transactions and why the world typically goes through the dollar to do trades.  
[00:31:00] Um At the risk of hitting something, the blockchain is changing that system.  
[00:31:06] Absolutely. And therefore could that be part of what's going on? Yeah.  
[00:31:11] Um but it did for the blockchain to change that system we have to ask the question does the blockchain change it in a way that uses dollar safe assets as the backing asset?  
[00:31:23] Like kind of today what we're thinking about stable coins which in which case we're sort of staying saying the same world or are we moving to a world in which there's some diversified portfolio that's the backing asset, right? And those are kind of important questions that need to be sorted out.  
[00:31:38] This is my last data slide. I know I have I haven't answered all the questions that have come up.  
[00:31:44] What I was going to do next is you know I think I've I've I explained this. My my take on what has happened in the world is this probability has been shifting around. It was stuck at one for a long time.  
[00:31:56] And it's just looks like it's been bouncing around, right?  
[00:32:00] And it's really kind of striking to see these correlations flip signs.  
[00:32:04] So what I'm going to do next is I'm going to walk you through a basically a calibrated model exercise.  
[00:32:11] Um and the model I'm going to I'm just going to give you sort of a uh schematic version of the model. There's more equations and details in the paper that was circulated.  
[00:32:21] But the model which I'm going to write down is one in which I'm going to argue that the US has been exporting liquidity services to the rest of the world. That is the rest of the world has been holding  
[00:32:34] uh assets in dollars and particularly in dollar safe assets. And as I've emphasized at various times in assets that are particularly low yielding.  
[00:32:44] That has meant that the US collectively has been financing itself cheaply.  
[00:32:49] That's export of a liquidity service, a liquidity good.  
[00:32:52] So I'm going to ask what happens if we just cut that off. Right? That's the the the the the model computation I'm going to do is around a very particular model.  
[00:33:00] It's a model in which think of the world as being two blocks in which the rest of the world has held a whole bunch of dollar safe assets in which it's earned low yields. Suppose we turn that off.  
[00:33:12] And the exercise that I'm going to do is the following.  
[00:33:18] I think the way to think about it is that's the hopefully familiar usual exchange rate um expression that comes from iterating on the UIP condition. Think of the exchange rate today as reflecting you know future paths of interest rates  
[00:33:33] between two currencies. Uh this is our special thing something about convenience yields on currencies, something about risk premium on currencies, and then some terminal value in 10 years of say the exchange rate.  
[00:33:44] Okay? So today's exchange rate at any one point in time and even over the last year is probably reflective of all of these components.  
[00:33:52] I don't know how much each of these components have changed. I'm going to do a computation that asks how much has this component changed?  
[00:33:58] And uh could it what would happen to an exchange rate far out in the future? That's why I'm sort of thinking 10 years.  
[00:34:08] Uh if we were to move to a world in which the new steady state ends up being in one in which reserve asset demand disappears.  
[00:34:15] I'm not going to take a stand on where the reserve asset demand goes to. Does it go to euros or anything else? I'm as I said I'm just doing kind of a partial exercise to benchmark how big these numbers are as to what would happen if we drop this.  
[00:34:29] Um and so these are the I'll I'll take you through an exercise and I'll give you answers to these three questions.  
[00:34:35] And so when you have the sterling to dollar with the sterling to dollar switch, maybe the other things were happening in the war and so on may make it I was wondering whether you see something similar to how that could be.  
[00:34:45] Mhm. >> [clears throat] >> Also just wondering to what extent was the dollar helped by the fact that there weren't good substitutes.  
[00:34:50] And so if there was a good substitute out there maybe things would be worse in terms of this uh Yeah.  
[00:34:56] Yeah, I mean so we've looked at at um correlations between these convenience yield measures and the dollar sterling exchange rate starting um around that period the early '70s.  
[00:35:08] They covary very strongly. Um and during that period the signs flip. Like the convenience yield is in some periods on the dollar asset, some periods on the sterling assets.  
[00:35:18] More recently. Um and the the signs flip and but it tracks exchange rates fairly well during that period.  
[00:35:26] In the '70s. In the '70s. '80s and '90s. Yeah.  
[00:35:30] >> Sterling's disappearing. >> Sterling's gone by the by the '80s '90s.  
[00:35:32] I mean I think by the time our other the picture I showed you shows up like in the late '80s the convenience yield is pretty much sitting in the dollar asset.  
[00:35:42] And your treasuries rather than assets. Yeah, I mean we were looking at one year treasuries but one year treasuries is where it was.  
[00:35:49] Okay. So as I said this is kind of the model I'm going to put on the table, one in which we have a I want to think about the US as you know exporting goods, importing goods from the rest of the world. And there's the second thing that the US does which is exports liquidity  
[00:36:04] services. Okay? I'm going to try to measure how much liquidity services are being exported.  
[00:36:09] Um the way I want you to think about these liquidity services is the rest of the world has been holding assets in dollars and earning low rates of return.  
[00:36:20] So you can think of almost like holding money in a bank and getting a very low interest rate on your deposits.  
[00:36:26] You're paying some liquidity services over to the US in order to hold that. So in order to put some numbers on this I need to measure the size of these liquidity services and I'll tell you how we're going to do that.  
[00:36:38] Okay. So take those three blocks and just think about [clears throat] sort of the simplest steady state type of analysis.  
[00:36:48] Uh in a steady state you need imports minus exports to equal the liquidity services.  
[00:36:55] Effectively if the US is exporting liquidity services to the rest of the world that is a that's a net it's a form of a good that the US is exporting which allows it to  
[00:37:07] uh facilitate a larger trade balance. Um plus some constant that maybe has to do with net foreign asset position or stuff that has to do with the capital account.  
[00:37:16] Okay? Um write the import minus export relation as increasing in the real exchange rate in the US. And effectively the exercise I'm going to do is going to measure where liquidity  
[00:37:31] services were. I'm going to do it particularly in 2016.  
[00:37:36] And that hits some equilibrium with some value of the exchange rate.  
[00:37:41] I'm going to imagine shutting off liquidity services which means we got to move to the left on this curve.  
[00:37:49] And that means the dollar has to depreciate.  
[00:37:52] Okay? So we're in a world in which the trade balance was in part buffered by the liquidity services.  
[00:37:59] I turn off liquidity services, the trade balance has to readjust by the amount of lost liquidity services. The way that's going to happen is the exchange rate has to depreciate.  
[00:38:09] How much the exchange rate will depreciate actually just depends upon the slope.  
[00:38:13] Uh what is that slope? That's basically the elasticity of trade with respect to exchange rates which the trade literature has given us some numbers on.  
[00:38:20] So I'm going to pin down a slope from the trade literature.  
[00:38:25] I'm going to measure liquidity services, turn that off.  
[00:38:28] Given a slope I can tell you how much exchange rates have to move. Okay? And again as I said this is just it's really comparing across two steady states in which I'm holding a whole bunch of other things constant.  
[00:38:39] The second exercise which we'll do is the rest [clears throat] of the world has been holding things like US treasuries.  
[00:38:47] And they dump all the US treasuries because not such a great asset anymore. US interest rates have to rise.  
[00:38:56] How much do they have to rise? They basically have to rise depending upon the slope of the domestic bond demand curve. So if I have an estimate of the slope and I know how much foreign holdings are going to be dumped I can figure out how much interest rates, right? So my we do this in a more  
[00:39:10] less pictorial way in the model but basically this is what I'm going to do.  
[00:39:14] I'm going to take we're going to take you through an exercise in which we put some numbers on these slopes, figure out how much liquidity services are involved, and figure out where exchange rates and interest rates will have to readjust.  
[00:39:24] Foreign hold foreign demand for US bonds can decline because of a perceived deterioration in the quality of US bonds or because something else comes along that's better. Yes. Those two things have seem to have very different implications for domestic  
[00:39:39] demand for US bonds. I see. So but you're but you're holding the the the curves uh stable there.  
[00:39:47] So, I don't think it I don't think it'll affect this guy. I think you're saying it's going to affect this guy.  
[00:39:52] >> Yeah. Yeah, exactly. I'm asking a question about that. So, So, there's an implicit assumption there about nature It's a weird world in which suddenly the rest of the world doesn't want to hold US. And the US wants to continue to hold it. Yeah, at the same  
[00:40:05] on the same terms as before. Okay. It's a fair point. I I agree with you. So, I'm going to give you a number and I'm going to pick the slope based upon historical data and I think you're telling me the slope is probably changed as well.  
[00:40:17] Um so, I it's totally a fair point. Maybe I I mean my Just if I just think that through logically, probably that slope is increased.  
[00:40:28] I'm less willing to absorb and I'm going to give you a number on interest rates and increase [clears throat] it. You just uh just really a slightly off.  
[00:40:40] Based on you're having this one variable model rather than 20, Yeah.  
[00:40:45] do you feel comfortable enough to make investment decisions on currencies and do you?  
[00:40:50] >> [laughter] >> I I I I I I kind of I told you this at the start, which is all I'm doing is putting a size on an effect.  
[00:40:56] >> I understand. I mean, obviously you're informing yourself and others about >> Yeah.  
[00:41:02] things that are have been happening and trend lines and how you're attributing a lot in the trend lines to a variable Yeah. even though others would affect it. So, just seems to me that the certain form  
[00:41:16] investment decision trading currencies Mhm. out to the furthest limit, maybe you can so you don't avoid the day-to-day fluctuations and so, do you? To answer that question, I have You don't have to. Never mind. I would answer the question I have you have to  
[00:41:30] tell I have to tell you what my I keep asking everybody what the answer is. the research too. So, maybe I should might be priced in by everyone.  
[00:41:38] >> [laughter] >> Uh okay. So, here's here's kind of the the underlying calibration data.  
[00:41:45] Um we we Oops. Um we go and measure the total quantity of safe bonds in the US. So, these are treasuries, uh commercial paper,  
[00:41:57] uh repoable mortgage-backed securities, bank deposits. So, we kind of construct of our an aggregate of what we think of as safe assets that contain convenience yields within the US.  
[00:42:09] That number in 2016 is about 150% of GDP.  
[00:42:13] And then from the flow of funds, we can figure out what the share is of that stuff is held by foreign investors.  
[00:42:19] Right? So, that's another number. It's 30% of the total holding. So, it's about 45% of US GDP.  
[00:42:25] Um I need and I need to So, that's effectively what foreign is holding.  
[00:42:31] The liquidity services that are being exported on that is take that quantity and multiply it by the convenience yield.  
[00:42:38] Right? How much lower are US interest rates? How much are is the rest of the world paying on a flow basis in order to keep their money in dollar safe assets?  
[00:42:47] I'm going to put a number on that. I'm going to share give you a number 2%.  
[00:42:50] I'll tell you explain to you where I get that number from.  
[00:42:52] Uh we need a slope for for this trade elasticity.  
[00:42:56] Uh it's basically coming from the trade literature. I'm going to use 0.3.  
[00:43:00] Um Steven, you can tell me if that's uh a reasonable number.  
[00:43:05] And then I need a slope here and this is related to uh Steve's question. I'm going to use a bond demand curve elasticity. I'll start with that. It's basically from uh my JP paper with Annette, which is admittedly historical data, annual data.  
[00:43:20] It gets us a nice demand curve. I'm going to use that as my curve. Okay? And if you tell me it's wrong and it's become steeper, I'm totally with you.  
[00:43:28] >> [laughter] >> The data in your paper with your JP paper uh You had a convenience yield of 50 basis points. 75. Mhm.  
[00:43:37] Okay, 75. Mhm. Now, you've got something 2% three times. Yeah. Yeah, from if this is estimated one on more recent data.  
[00:43:44] No, it's I think um let me talk talk you through the 2% and I'll I'll explain. I I've I've come to understand the following, which is we were measuring that in that paper the spread between AAA corporates and treasuries  
[00:43:58] and calling that a convenience yield on treasuries. That is making a a zero assumption.  
[00:44:05] The zero assumption being that AAA corporates had no convenience yield.  
[00:44:08] I've come to realize that's not true. That in fact AAA AAA treasury bonds are also a very kind of special asset and that um when we when I take that 75, it's  
[00:44:22] actually a lower bound. And I really should be if I if I really wanted a full convenience yield, I should be using something even higher.  
[00:44:31] Like my my colleagues at the GSB uh >> [clears throat] >> Ben Ebers, Sebastian Talla and co-authors have tried to use equity returns as the benchmark relative to treasuries to measure a convenience yield.  
[00:44:44] Right? Which I think conceptually is the correct thing to do, but of course equity returns are super noisy.  
[00:44:49] So, when you do that construction, you're going to add lots of measurement error. The nice thing about looking at AAA's versus treasuries, which is what Annette and I did, is once I both sides are relatively low free of risk, so you you feel like you can pick up something much more cleanly.  
[00:45:04] And that has been a general issue in these convenience yield measurements.  
[00:45:07] The more accurate the measure, the smaller the number you're going to get.  
[00:45:12] And so, you need to make some guestimates around where I'll tell you where where I'm going to get the 2% from.  
[00:45:19] >> transfer you're going to use a 2%. I'm going to use the 2%. If you want to, you can double you can have my numbers.  
[00:45:24] Um if you like 75 better, then adjust Everything is linear here. So, it's all you This is just to put some numbers in the table. Okay? So, this one is clear and I think I agree with Steve's caveat.  
[00:45:37] Here's where the 2% is coming from. So, I showed you this this spread before and I told you this thing average around 22 basis points. What is this? This is a CIP wedge between two government bonds.  
[00:45:51] Okay? It's an excess premium on holding treasuries relative to uh rest of the world's foreign government bonds.  
[00:45:58] Now, the right number for the liquidity service computation is an uncovered interest parity computation. It's on an uncovered interest parity uh basis, how much do I want to hold dollar  
[00:46:11] safe assets relative to say your your European safe assets? That's really the number I should be looking at. That's not what this number is.  
[00:46:19] And it's it's related my it's this is taking the Canadian bond and converting it in dollars. So, I've sort of I'm putting taking two assets, one of which is effectively like a better dollar than the other. Right?  
[00:46:30] It's the Canadian bond FX swapped into dollars. Now, it's a dollar asset and I'm comparing to treasuries and that's a that's a that's a number.  
[00:46:37] It's it's going to be a lower bound on what I want to do.  
[00:46:40] I really what I'm really interested in is if you just offered me and asked me asked the world investor, you could hold your money in Canadian or you could hold your money in dollars, how much more are you willing to pay on an uncovered basis? Okay.  
[00:46:55] That's a different number. Pardon? Depends on the inflation differential.  
[00:46:59] >> Okay. So, the way what we have done in our papers is we've said we've noticed the following. There's a very strong relation Oops. Uh between the CIP wedge movements and the movements of the exchange rate.  
[00:47:12] So, when CIP wedges grow, uh increase dollar exchange rate appreciates.  
[00:47:20] The dollar exchange rate capitalizes the full UIP deviation.  
[00:47:24] Right? If you just think about the exchange rate expression, it reflects UIP differences.  
[00:47:29] So, I the based on the co-movement of this wedge and the dollar exchange rate, I can extract how much CIP implies for UIP.  
[00:47:39] Uh that ratio from our estimates is about 10 to 1.  
[00:47:42] So, we get about a 20 basis point average number here. Multiply that by about 10, that gets us the 2%. So, what we our our underlying computation in this exercise is the the what I really like to know is how much has the world  
[00:47:56] been on paying to hold dollar assets relative to other world assets. That's really a UIP thing. I need to I need to find some way to get at that and this is the way This is where the 2% comes from. Okay? The  
[00:48:09] covering is uh There's an There's some assumptions here that gets you to a bigger number. If you if you want 75, I'm happy with that. I'm you can just have my numbers. Here's another uh type of number you can look at to try  
[00:48:23] to benchmark this. And now, this is within the US. Take investment grade corporates, CDS hedge them and subtract that from a safe rate. This is similar to the type of things that Annette and I have used at in the past.  
[00:48:36] This is a premium. The red is the premium relative to repo, which is, you know, over the last five years is numbers ranging from, you know, something like 75 to 90 basis points.  
[00:48:48] Um so, I'm you know, if you told me, "Look, your 2% number is too high. I'm only going to give you 1%." Have my numbers. I'm fine with that, too. Okay?  
[00:48:58] Uh it's interesting and this is related Bob to your point. This is the same spread. Take investment grade corporates relative to CDS, subtract that to treasuries.  
[00:49:06] That looks like, you know, treasuries don't look like as good collateral. Repo still looks like a safe asset. That's I think very one thing that's uh apart from what I'm saying sort of an interesting fact in the US. Okay? So, those are the two numbers I'm going to  
[00:49:20] put in. And then I'm going to state state those numbers and just do the computation. So, effectively what I'm doing is in the background there's a model in which services are being exported and I just compute the model under two steady states, one in which I turn off liquidity demand.  
[00:49:34] The uh baseline we're doing is think about the world in 2016 and then I'm going to compare it to a world in which we shut the whole thing off.  
[00:49:43] This is the difference in the exchange rate across these two uh steady states.  
[00:49:47] So, seven that's 7.6% depreciated. Not a huge number. I just want you to notice that that's the number.  
[00:49:56] Interest rates in the US and admittedly uh Steve, to your point, with the historical slope end up rising by about 92 basis points.  
[00:50:05] Okay, so let me just kind of interpret some of those numbers for you.  
[00:50:10] This is really what uh we're doing. We are comparing across steady states. So, as I said, we're holding everything else fixed.  
[00:50:17] It's a real exchange rate. You know, when we first did this number, 7.6% doesn't look like a very big number. In fact, I I think one of the takeaways I had from doing this computation is the exchange rate impact of shifting out  
[00:50:31] is actually small. 92 basis points is a big number, right?  
[00:50:36] I I think the my takeaway from these computations was exchange rate impact small, interest rate impact large and perhaps even larger.  
[00:50:44] Why is the exchange rate impact small? And I've put big numbers in here. I mean, I've put in a 2% convenience yield. I've given this thing a fair amount of room to work.  
[00:50:54] Um and the answer I mean, it's it's we're using a long-run trade elasticity and basically that's what's going on. So, it's reasonable to think, you know, we have more inelastic uh measures at the short  
[00:51:08] run and maybe you can use those and get bigger numbers. In the paper, we also use asymmetric elasticities. Like, I think there's reason to think that the US export elasticity is different than the US import elasticity. You can try numbers like that, but that's really  
[00:51:22] what's driving it. It's the point three coupled with this which is giving us about 7.6%. As I said, it's a long-run real exchange rate. You know, inflation, paths of monetary policy, all will matter, of course, about the current exchange rate, but this is sort of just  
[00:51:35] a partial exercise to see to size this effect.  
[00:51:38] >> Yeah. So, you've said why, you know, you say much about dynamics, which I get, but is there potentially another plausible, more plausible maybe, counterfactual which is not that demand disappears for the US as a safe asset, but  
[00:51:51] becomes more like cuz other other other Yeah. currencies are used, right? Other assets are used. It becomes more like other countries. Yeah. And I I don't have no idea what those numbers look like.  
[00:52:01] And that's still a big reduction in Easy to do, right? Like, I we could we could another exercise which we could easily do is make the US move reserve asset to the world portfolio.  
[00:52:11] All right, we set it to zero. So, we could move it to the world portfolio and I I could tell you what those numbers are. I don't have them off the top of my head, but it'd be easy to do, right? Or another exercise we could do is shift it from the US to the rest of the world.  
[00:52:25] Right? Kind of do the reverse and then these numbers would be double what they are. Follow up my previous comment.  
[00:52:30] Yeah. There's a useful way to think about it. Or potentially useful, you can tell me. Mhm.  
[00:52:35] >> So, if you think that the reduction in the foreign demand for US assets is a repugnance effect or a fear of expropriation or subject to sanctions if your assets are held in dollars, >> Yeah. you could see why that might not  
[00:52:49] alter the uh slope for domestic investors. Yes. On the other hand, if you think it's about a general deterioration in the creditworthiness Mhm. of the US government, then you would expect to get bigger >> Slope change. Yeah, slope change  
[00:53:02] implying bigger numbers. So, so to some extent this which which interpretation you prefer, the one you've done or the slope or combined with the slope uh steepening Yeah. depends on what you think is really going on underlying, you  
[00:53:17] know, behind this uh Yeah. this change in the demand for dollars. Yeah. Tell me the fiscal problem with Trump. Yeah.  
[00:53:24] That what you Yeah, exactly. But But the the thing about sanctions, the sanctions one it seems like uh I don't know >> That This This is I I think what you're telling me is this is probably most con- The calibration I've done is most consistent with the sanctions story.  
[00:53:37] It's not consistent with uh increased Yeah, that's that's that's how I would think about it. I think that's that's a fair point. Yeah. I mean, I I think on an increasing US risk, I wouldn't know I off the top of my head what to do. [clears throat]  
[00:53:51] Like, I I know that we should increase the slope. How much? I don't have a benchmark. I'll think about it. But it's a it's a totally valid point. Especially since what it looks to us is the action is in interest rate land.  
[00:54:03] >> Yeah, so that is a big effect. I don't know if I said it could be bigger if it It could be >> even bigger than this. So, the Mhm. Are you holding the uh euro interest rate constant? I'm hold As I said, this is I'm holding every So, I mean, presumably they're moving into something. Yeah, yeah, yeah.  
[00:54:17] All All the caveats I've told you before apply. Yes. I'm not This is why I'm not trying to explain the high frequency movements in the dollar. Not a general equilibrium model of the world asset Yeah. It's I'm It's a general equilibrium model in which I'm holding a  
[00:54:30] bunch of reaction functions the same. >> [laughter] >> Um I mean, you decide. I I still find the 7.6% number It's at least it gave me an idea as to how big this effect is. On the exchange rate, it seems pretty  
[00:54:44] plausible given what we saw in COVID and the great recession. Sharp right. And it's also, as I said, it's an answer to 7.6% is not a big number. If the Miran hypothesis was that was the cause of manufacturing jobs,  
[00:54:58] I think one would have to look elsewhere.  
[00:55:00] It just doesn't look like that could be >> an important That's That's a could just couldn't possibly be that big according to this, right?  
[00:55:07] Here's another way of thinking about this and this is really building off of the the 90 basis point number, the point nine percent number.  
[00:55:15] Um you can you know, you a different way of interpreting the equilibrium we've been in is that the US has an asset uh in the background that pays a dividend.  
[00:55:25] And that dividend, which is the liquidity services the rest of the world buys, is what allows us to uh run uh uh partially offset the trade deficit.  
[00:55:37] So, another question I could do is what is the value of that asset?  
[00:55:40] That gets very directly to this 90 basis point number that I've shown you. So, I'm going to do a computation where I'm just going to take my Now, I'm going to use 45% of GDP of dollar bonds. That's what the rest of the world holds. I'm going to use 2%. Uh  
[00:55:55] Mike, if you want to have it, please go ahead and have it and as I do this computation.  
[00:55:59] And I I'm going to ask what is the value of that asset, right? That's a That's a a flow of goods that the US has had for a while.  
[00:56:08] Um suppose you were to lose that point nine percent as a flow of GDP, it should something should change, right? That asset valuation should be capitalized in a bunch of places. Like, if I think about it, the most direct places it should be capitalized is in  
[00:56:23] the franchise values of safe asset issuers like banks. You know, there's been kind of Yeah. And you look at the Chinese yuan. Is that that as as an alternative.  
[00:56:34] Uh I'm in what sense? As a as a currency. Oh, you're saying in the world? Could that be the case?  
[00:56:42] Absolutely. I Again, I'm not taking a stand at this point on where the money goes, right? I'm just taking a stand on it goes away from the dollar. Right?  
[00:56:50] Obviously, if it depending upon where it's going to go to, the constellation of currencies will um and interest rates will move differently. All right.  
[00:56:58] I think it So, asset values should be somewhere um in in equities, in collateral assets like housing, uh in the financing cost of the US government.  
[00:57:10] So, something about the PV of the future tax burden. I don't know where exactly it's sitting. I'm just going to take a very macro exercise and just PV this thing, okay?  
[00:57:18] So, here is another computation. Take point nine percent of GDP. Um it's a loss on a slice of GDP, right? Which think of GDP [clears throat] as a risky growing stream and now this is just simple  
[00:57:31] finance at this point. We used AQR's um valuation model to put some numbers down.  
[00:57:39] Uh AQR's valuation model is a 1.7% risk-free plus equity risk premium 1.6% and we need a growth estimate for GDP, which we're going to use 1.8%.  
[00:57:50] Uh we need a beta on GDP. So, to figure out what the discount rate should be, we're use you're going to use a beta of 2/3. This is from some of my co-author's other work. If you just take those two those numbers, take the senior edge, do uh  
[00:58:04] just a PV formula R minus G, it's 93% of GDP or roughly 29 trillion.  
[00:58:10] Okay? Uh if you want to drop my convenience yield from 2% to 1%, just have that number.  
[00:58:15] That's a big number. >> Agree on that. Yeah, it's a lot of Yeah.  
[00:58:18] So, I I think as I said, I I the the thing I took away from this exercise is that the big effects are not in exchange rate, they're in in the asset valuation market uh for the US. And you know, this is these are just rough estimates. I  
[00:58:32] don't want I'm not going to tie my hands to 29 trillion, but it's in that's the type of number that if you want exorbitant privilege has been worth uh to the US.  
[00:58:42] This is assuming about the term structure of the debt? No, no, I I'm not not taking any The dollar holding say they're holding one year treasuries. It doesn't matter. So, I all I'm I'm present valuing the the convenience yield flow.  
[00:58:56] Whether it's coming from the rest of the world It assumes that that's going to be rolled over continually, right? Point nine percent of GDP forever. That's That's what That's Mhm.  
[00:59:06] So, it the the underlying assumption was the world was one in which the rest of the world wanted to hold dollar safe assets, was paying up this 2% convenience yield on an annual flow basis.  
[00:59:18] They were paying it to to the Treasury, they were paying it to banks, they were paying it to to the housing market in the form of lower cheaper mortgages. They were paying it in all sorts of places.  
[00:59:28] All right? Um I'm not I I I'm not going to distinguish between all those different places. That's that's my only point that you're just assuming it goes on forever and Yes. Okay. You could consider other scenarios where this unfolds in a way where people change their term  
[00:59:42] structure. They have They hedge that in various ways.  
[00:59:46] This gets adjusted. So um >> Axel No question out there. Mhm. Axel Yes, hi. Um I come as a practitioner and uh so apologies I don't I don't perfectly fit into the model.  
[01:00:00] When I think about this liquidity reserve, I think about it in the context and you you referenced April 2nd last year.  
[01:00:09] I think about it in terms [clears throat] of barriers to trade interrupting not just the flow of goods, but also the context of a trade deficit, the flow of currency. And so to me uh  
[01:00:24] the liquidity services are not external, but an integral part. And then you can You can also see that that when the tariffs were imposed that the bond yields were rising because the financing happened more domestically. I was just  
[01:00:37] wondering how that drives with the sort of model that you have presented um whether it's part or not. And then the other part I just wanted to to to point out a lot of the times when there's a discussion about the the equivalent of  
[01:00:51] liquidity services, the US acting as a bank or a hedge fund or the exorbitant privilege, however one wants to phrase that.  
[01:00:58] People say, "Oh, the the US won't lose that because there's no alternative." And then of course in currencies we always think about the currency pair.  
[01:01:07] If I'm not mistaken, there does not need to be an alternative. That the That the alternative that we're heading towards is a global disintegration where there's less liquidity in the world. Um and I I I just wanted to have your thoughts on those. Um that would be great.  
[01:01:22] I mean on on your first point I think the the uh economics that you outlined are in the model.  
[01:01:28] That is it's it it there's a coupling in the in the model between the what's happening in the goods market and what is happening in the financial market of holding liquidity in dollars.  
[01:01:38] Those two things are coupled. So I think it captures the economics that you're after. The The broader point you're making about what an a new equilibrium would look like. Again, I There's nothing I've shown you that sheds any light on that. I'm happy to talk about  
[01:01:52] it because I thought about it. And you know, my read on the world and in history is that that times the world has grappled towards other reserve currencies and those periods are  
[01:02:05] volatile times which are not so great. Uh the '70s weren't great. The '30s weren't great.  
[01:02:13] So That gets to the point that if one envisions the dollar no longer being the global reserve currency >> Yeah. that's something much worse than the $29 trillion will probably be going on. Yeah. This bunch I mean so as I said worse things like that. Many other  
[01:02:27] things could be happening along that probably something really ugly. I agree.  
[01:02:32] I I That's all I have to show I show you. So that's my conclusion. No, that's very helpful. Thanks. So it's just one of these G should be actually global growth rate. G should be global. Because the demand comes from based on the global bond market. It should be global growth. You're You're right. It should  
[01:02:47] be global growth. Use your mic. >> [laughter] >> So I Mark has just pointed out I'm using US GDP growth to grow this dividend, but if you think that it's world demand and it's certainly the case that world  
[01:03:00] GDP and world demand for safe assets is in dollar assets has grown faster than US GDP. That's probably the right number to use. So it could be even larger than this.  
[01:03:10] Yeah. What's the present value of GDP to compare this to?  
[01:03:13] Do the multiplying. >> [laughter] >> $29 trillion kind of shrieks out as a huge number.  
[01:03:18] >> Yeah. Yeah. Relative to the present value of GDP, it's kind of modest. Absolutely.  
[01:03:24] 100%. Can I just go back? Um so I think your your exchange rate point Mhm. that the even this you know, just shutting off the demand for foreign US foreign demand for US  
[01:03:36] Treasuries entirely as this 7.6% fact that that is a that is a striking result in the modesty of the size. I want And And so [clears throat] I And it speaks to these issues that Moran and others have raised and that many take  
[01:03:50] quite seriously. So are there Can you help me think about whether there are reasons beyond your framework Mhm.  
[01:03:57] that could let lead that effect to be much bigger than you've estimated it to be?  
[01:04:03] Or should I really think of this as a a solid number that I can say at least what's an upper bound?  
[01:04:10] I feel like I've I've I mean I I think this has come up. I feel like I've given it within the context of within the context of your framework room as possible. You have, but I I'm I want to I mean uh Yeah. I'm probing to help me  
[01:04:24] think about your own framework and its potential deficiencies in this respect.  
[01:04:29] So is there something Is there some other reasonable framework I might bring to bear that would make that number bigger?  
[01:04:37] Exchange rate effect? Say it again. How about the exchange rate effect? Yeah, the exchange rate effect cuz cuz that is There's many interesting things that come out of your analysis, but that's one >> That's the That's thing that jumps I find quite striking that I have not seen  
[01:04:50] before. >> [clears throat] >> Um I mean again I So it's a real exchange rate.  
[01:04:57] Yeah, well that that seems like the right one. It seems like the right I mean when when people look at the world, they're used to thinking about paths of monetary policy which are moving at a high frequency. I'm sure that the movements in the dollar last year have been driven by a bunch of others. This  
[01:05:12] is just a little stub of that valuation equation.  
[01:05:16] But this is the right stub for the long run question, right? That's why What is missing?  
[01:05:23] I um I don't have an answer. I guess an an offsetting factor would be that if if this were to occur, then presumably the US would no longer be the world's policeman and we would benefit  
[01:05:37] from uh reduced defense spending, although there would might be in more wars.  
[01:05:43] >> [laughter] >> Yeah, well, we What would happen to sea lanes?  
[01:05:47] So there may be a sharp decline in global trade. Yeah. I I Steve, there's one thing that I think >> Along with that, the decrease in defense spending. Um which we've we've run through it. So the If you If you take the historical trade elasticity  
[01:06:01] and you slap on tariffs, then effectively the size of trade is shrinking. Right. Which means that in order to readjust to a loss of liquidity services, which is the same loss,  
[01:06:15] with a smaller trade >> smaller adjustment you it's a bigger percentage adjustment, which means the exchange rate has to move more.  
[01:06:22] So you know, one way of going to to make this number bigger is to effect you have to change slow change the slope.  
[01:06:30] Right? And there's different ways of changing that slope. One way of One exercise which we've done in the paper is do a combination of tariffs and lose um and lose the dollar >> lose the dollar  
[01:06:45] >> status. Yeah. Oh, okay. So that maybe that's one answer.  
[01:06:50] I'm I'm all ears though because this is If you want as I said, when we did this computation, this was the thing that jumped out at me. That It jumped out at me, too. I don't know if it's in the paper. I apologize for not reading the paper in advance. Um but something that would be useful would  
[01:07:03] be just lay out a table with Mhm. different parameter values and seeing what what happens. Yeah, well so I I'm I'm showing you kind of our baseline.  
[01:07:12] The paper has a bunch of different You know, we we have different slopes that we use from different papers to try to get some benchmarks.  
[01:07:22] So in in the model there um Mhm. The foreign country has uh demand for safe assets for US safe assets, but there's there's no reverse part.  
[01:07:32] >> I'm turning I'm turning it off. Yeah. Yeah. Uh shouldn't that be there?  
[01:07:37] Yeah, okay. So actually maybe that's a that's a better answer to Steve's question as well. Yeah, that's a good Um if we flipped it made the the the new steady state one in which the US buys the foreign asset,  
[01:07:52] then I mean symmetrically in a sense it would double the numbers.  
[01:07:56] Maybe that's a Right? And that's I think that I I agree with the bottom. That's a better answer to Steve's question. So you I your response to the the other question I guess was the same question. You said You said, you know, your framework's more like the the sanctions effect than  
[01:08:10] the fiscal position long run fiscal position effect.  
[01:08:14] But in your early work, you know, the the empirical work at the [clears throat] beginning, maybe I mean maybe there aren't clean events that raise the you know, that sort of raise the relative demand of a country to try  
[01:08:27] to replace the US. But think about Yeah. war with Ukraine, sanctions, alternative currencies to trade in. I mean do any of those show up in the data? I mean there there Over the Over the other the other paper that I showed you a little bit of,  
[01:08:40] we ended our sample in 2018. There's nothing that looks like Everything looks like dollars, more dollars, please more dollars.  
[01:08:50] So I have no hint of an another correlation. The only hint is as Steve asked about this is the the '70s period with the UK and the US.  
[01:08:59] There's There's some hints of other things going on there.  
[01:09:04] Sorry, [snorts] I Steve, just to come back to you. I think I I could I could put the other side and I could get a bigger number.  
[01:09:10] Um and that would be interesting. Uh it would not be directly about the Moran hypothesis though, which is really about the 7.6. Right.  
[01:09:18] >> Right? Because that's about how much has world demand for US assets led to an appreciated dollar.  
[01:09:25] Is there There's an asset pricing question about where the Moran view is.  
[01:09:28] Yeah, the benchmark ought to be something like um the US is just part of the global portfolio share of safe assets, maybe equal to its share of world GDP.  
[01:09:40] >> Perfect. And then then then it would be even smaller than 7.6. And you said community yields maybe too high. Yeah, and that would be an even smaller number than 7.6 if that's the case.  
[01:09:49] But if you were to sell some other goods, also it changed the exchange rate.  
[01:09:53] I'm staying stating, right? So Just what particular good are you selling?  
[01:09:59] You're saying if the But again, like this is all comes back to the partial exercise, right? You're saying suppose tariffs came or US was exporting or importing totally different goods, then that's a different exercise.  
[01:10:13] What what I'm trying to grope towards and is it it is important if you can put out into the public discourse in some way that people can understand that the particular concern that Moran raised  
[01:10:27] cannot be a very big source of US deindustrialization. That that is an important I mean, I'm not particularly sympathetic to that view anyway, but some people are. And if one could show that no, that really can't be what's  
[01:10:41] going on, that would be a very useful public service.  
[01:10:44] >> So Steve, what more would you like to see?  
[01:10:46] >> No, I like what you've done already. I'm >> Yeah, I mean I I I hadn't seen it before and so it's resonating with me.  
[01:10:53] What I would like to see what I would like to see what I would like to see I'll talk to you about this later is I'd like to see you write this up in five or 10 pages in a way that the Morans of the world might understand. Okay. That's a challenge. I'm not sure I'm capable of >> Big demand.  
[01:11:06] >> [laughter] >> Ideally.  
[01:11:08] >> So um if we take a bigger step back the forecast accuracy in the exchange market and bond markets has has missed almost every big trend badly.  
[01:11:23] Yeah, they missed the inflation of the '70s disinflation of the '80s when the Fed lowered the discount rate and uh Fed funds rate to zero in December of 2009, the bond market was expecting it to stay there for 9 months, it stayed there for 7 years. I could go on and on.  

## Q&A (01:11:35 – 01:13:00)

[01:11:35] So if something as big as you're talking about happens like we're shutting off um I guess one question is I I'd be interested in a wide range of understanding of how financial markets  
[01:11:48] would react. They've got Brexit, right?  
[01:11:53] The pound depreciated greatly and British growth prospects in the aftermath have been pretty So that was like one one one on the other side of the ledger of your example.  
[01:12:04] I'm not saying they're always wrong. I'm just saying for for many of the big events in US economic history since Bretton Woods I mean that you know that the historical shifts of reserve currencies happens  
[01:12:17] over decades to decades. Yeah. This is slow moving stuff. It's not Brexit. So I'm not sure one should ask for a quick response. Yeah, you know, it's also um  
[01:12:31] I think Rogoff makes this point quite frequently that there's never been a global reserve currency as dominant in the in the exchange system as the dollar is.  
[01:12:43] The pound was never that, the guilder and all this before the time Yeah, in the financialized world it matters even more. Your answer will have to wait.  
[01:12:51] >> [laughter] >> He knows what Marie Christine says.  
[01:12:54] Well, lots of food for thought. Thank you so much Arthur.