How Reserve Currencies Rise and Fall: Arvind Krishnamurthy on the US Dollar and the History and Horizons of the World’s Money

Haoxiang Zhu: Arvind, today we'll talk about the dollar and its future as the global reserve currency. Maybe taking a step back—what does it mean to be a global reserve currency?

Arvind Krishnamurthy: Thanks, Haoxiang. So just to step back and think both about the financial data and the real side, you can see the footprint of the dollar being a reserve currency in a couple of places. 

I'll start from finance. One of the things that's interesting about the way the foreign exchange market works is if you want to trade, say, Aussie dollars for yen, you might think the way to do that is just trade Aussie dollar for yen. In fact, what actually happens is people trade Aussie dollars for US dollars—and then trade the US dollars for yen. The dollar sits there in the middle of these transactions. 

That tells you something about the superior liquidity of the dollar market. There are lots of buyers and sellers of dollar assets, dollar foreign exchange, dollar bonds. And so when you trade around the world, it's better to go through the liquid market than it is to go directly. That's a liquidity statement that you can see in the world. 

You see counterparts of this in asset prices. One of the measures I’ve used in some of my work is what I call the convenience yield on dollar safe assets. Imagine I were to compare two assets—for example, a US dollar repo and a foreign government bond, say, a Canadian government bond. Of course, you can’t just make that comparison directly because they have two different currencies. So you need to swap the Canadian dollar government bond into US dollars and then compare [its yield to that on] the US dollar repo.

When you do that comparison, you see that the US dollar repo has a much lower yield than the swapped Canadian one. There's a premium on US dollar safe assets: My work estimates that it averaged around 2% for the first 20 years of the 2000s. 

You can see [the reserve currency footprint] in other places, too. There's an enormous amount of dollar-denominated debt in the world. The dollar is the lingua franca of debt markets around the world. Sovereigns, corporations, multinationals all around the world borrow in dollars. And the flip side of that is lots and lots of investors around the world hold dollar safe assets. Think of foreign central banks or generally global banks holding dollars as reserves because it is the kind of liquidity, world money, in essence. All of that sits on the finance side. 

When you look at the real side, you also see patterns where the dollar looks very special. One particular pattern that trade economists have focused on is the use of the dollar as a denomination for invoicing contracts for trade. Even in trade with countries that aren't the US—say, Brazil trading with Japan—often, that contract will be specified to pay in dollars. 

There are other things you see on the real side. When the [Federal Reserve] changes interest rates, it doesn't just have impacts on the US economy; it has outsized impacts on the whole world. 

So there's a collection of characteristics that make the dollar a very special currency in the world, all of which I would put under the reserve currency framework.

HZ: That's a really useful overview. The dollar is a gateway for financial market transactions. It has a lower interest rate because of that status. It underlies the pricing in the global market. So the next question is: How does that come about? Why the dollar become the dominant currency? You have done research looking at hundreds of years of history. Maybe we can throw in the British pound and the Dutch florin—these were the dominant currencies in the past. Do you have a unified theory explaining why these currencies rose to dominance?

AK: Yeah, so first of all, this is perhaps the most intellectually fascinating question. What I told you first about the liquidity or the interest rates—we can see this very clearly in the data. Finance economists have ways to measure these things, and we can do a pretty good job of nailing these facts down. 

The broader question of the why is much harder. And the reason it's much harder is because there's very few transitions. It's a small sample, and if you want to understand why the dollar, you need to understand why not someone else. Why did the dollar end up rising? We have history as a guide. I'll give you my read on the history; this is work I've been doing with Chenzi Xu at Berkeley and Antonio Coppola at Chicago.

The central factor is the liquidity and safety and size of the safe asset market. Let’s just unpack that. The way we think about the world is, in a sense, that every economy uses safe assets within the country to facilitate and lubricate the financial system: broadly speaking, money. And I don't mean paper money. I mean, broadly speaking, the money markets within a country lubricate the financial system of the world.

HZ: Such as deposits or central bank money that is used for transactions.

AK: Absolutely. Deposits, central bank money, repo, this collection of short-term safe assets lubricate the financial system in just about every major currency that you look into.

What looks special about the reserve currency is that the world also adopts one as world money. And what the history has told me is that reserve currency status attaches itself to the world money that has the largest and most liquid money market, which is a position that the US dollar has held at least since the end of World War II.

You can see manifestations of this in many places. And it's important to recognize that it is size and safety and liquid use, all of these together. The US has a large supply of safe assets. Start from just US treasuries—there are probably tens of trillions of safe treasuries. Then, if you add to that the stock of treasuries and mortgages and agency bonds that back repo, you get even more. So when you look at the volume of safe assets, there's a large amount in the US.

Then if you adjoin that to, in a sense, world GDP, which is the underlying demand for using the dollars as world money, you can see why the world would want to concentrate on the dollars as the place in which you want to lubricate money. You could look at another currency—say, the Swiss franc, which is a very safe currency, as well. But there's too little of it to imagine the whole world running its money needs through the Swiss franc.

So the size and safety and liquidity of the money market seem important. That's the central theory. Just to provide some context, the other types of theories one should be thinking about here focus purely on economic size: The US has been the largest economy in the world for a while. The US has the largest military power or volume of trade flows. The US has been at the center of world trade, as well. So how do you tell apart different theories of the world? 

HZ: Exactly. 

AK: History is useful here. I'll start with the history that we have found most compelling, which is the historical comparison of the Dutch florin to the Spanish silver dollar. So let me take you back to the 1600s and the 1700s, when both Spain and Amsterdam were trading hubs of the world. Ships are going around the world carrying goods in different places. Spain is much larger than Amsterdam—just physically larger, economically larger, militarily larger. 

Yet when you look at the finance side, the center of finance is sitting in Amsterdam. And the Dutch florin, which is the ledger currency issued by a bank in Amsterdam called the Bank of Amsterdam, looks like it's performing a role similar to what the dollar is doing now. You have merchants who don't have anything to do with Amsterdam. Say a trader in Barcelona who wants to trade with someone in Russia settles the trade contract in florin at the Bank of Amsterdam. That looks like the reserve currency role of the dollar today. 

Amsterdam is much smaller in terms of economic size, military size, and trade volumes. Just physically smaller. But what Amsterdam has is a bank with a financial technology that creates a large supply of liquid and safe florin. So that's a clear example of a currency in a country without huge economic or military size acting as a reserve currency.

HZ: Maybe say more about how the Bank of Amsterdam created this new liquid asset.

AK: It was actually purely financial engineering, and we recognize it now. I'll tell you more about it. But you have to realize, in 1650, this looks very exotic. People just use this currency coin. At some point, coins become inconvenient. And as you know, economic transactions become more and more sophisticated. A better way to trade is to create a narrow bank where you use coin, you deposit the coin at a bank, and the bank gives you a ledger account. And when you want to make a payment, you debit your ledger account, so you do just everything on the ledger of a financial institution. 

That type of system had already been around in some form—I think the first version of this was in China 2,000 years ago, and it was common in Europe, as well, particularly in Italy at this time. But the ledger currencies were tied to the physical coin in their local jurisdiction. So, for example, in Venice, there was a local coin, the lira, issued by the local monarch. That was the coin that was being used as a ledger. You were limited in size by the amount of the local coin. 

What Amsterdam does very cleverly is create a repo facility where you can bring any coin from anywhere within Europe, deposit it, and get a credit of some amount of florin at the bank. That allows you to take many, many different types of coin and use them as backing to create a larger and larger florin supply. 

So I said it looks like financial engineering—it looks a lot like taking lots and lots of mortgages from all around the US, packaging them together, and creating some big security, which then ends up being collateral for repo. That's sort of the modern equivalent of money creation that happens in the world. And the simple version of it is you take heterogeneous coins and create a homogeneous florin, and that creates liquidity in the florin. 

At that point, if I am trader A trading with trader B, I know there are many, many traders who are trading that florin. I can borrow it, I can lend it, I can buy, I can sell. If you pay me in florin, I know I can pass that florin on to someone else. I am very comfortable that this thing is a very liquid asset. And that's the heart of their financial engineering.

HZ: That's really fascinating. I'd imagine that the Bank of Amsterdam also had a tool to control risk, credit risk, or default risk or some risk that maybe a coin from an issuer failed to redeem. Did they have tools like that back then?

AK: Absolutely, yeah. There's two versions of credit risk that pop up. The first version, which is very relevant when the bank starts, is that the traders who go and choose to deposit their coins at the bank are surely worried that the bank will steal the coin. So what the bank would do is, on Sunday mornings, it would open up its vaults for people to come and look and say, “Oh, there's your coin.” The bank tries to institutionally provide some commitment by allowing people to look at their books. That's important. I'll come back and talk about this later because, in a way, the bank failed because it lost that commitment in a war.

The other credit risk issue is a version of counterfeiting. You know, some trader brings in coins that are clipped, they're shaved. What's the metallic content? There's a lot of risk associated. So there's a whole industry of money changers whose business is verifying that a coin is worth something. 

Again, it's very similar to mortgage-backed securities. We have a whole creditors’ industry that verifies that this is the amount of payment that'll come back from this mortgage loan. The rating agencies and the valuation, all the securities traders of mortgages—all are in the business of collecting information around these underlying assets. So the bank was also involved in this. It had a kind of money-changing expertise in valuing the money. 

But the second thing is that to create florin against a coin, you don't need to know exactly what the coin is worth. You just need to know roughly what the coin is worth. You need to know at a minimum what the coin is worth. The trader would say, “Hey, you know, this coin is worth 100 florins.” The bank would look at the trader and say, “Well, maybe, maybe not. I'm fairly sure it's worth at least 80. So here's what I'm gonna do. I'm gonna give you 80 florins in the account, and I'll keep the coin. Any time you want, you can come and return 80 florins, and I'll give you your coin back. So if you think it's worth 100, fine, you can get your coin back. But I think it's worth at least 80.” And the 80 is what's important from a liquidity standpoint, because everyone is bringing coins in that are worth at least 80, and so the supply of florin and the usability of florin expand. 

In the modern world, this is a haircut in a repo, basically. The bank was active in setting haircuts on coins to ensure liquidity of the underlying florin market. These are the two sides of credit risk. 

But as I mentioned earlier, the Bank of Amsterdam, which was a fairly successful monetary institution, blew up around the late 1700s. The bank was a narrow bank. It just held the coin. But later in its history, the Dutch sovereign saw that there was coin at the bank. The sovereign was financing a war effort and so it took loans from the bank. But then the Dutch lost the war. The bank took losses. There was a run on the bank. And the whole thing collapsed by 1796, I think.

HZ: So that's the end of era for the first global safe currency. Maybe we can come to the British pound. For your theory about a very deep, liquid, and sizable financial market being the linchpin of the reserve currency—does that explain the rise and fall of the British pound?

AK: The British pound is also really interesting in a financial engineering sense. The Bank of England, which ended up being the kind of money creator of the world, certainly by 1850, 1860, was originally—like many other central banks, like the French central bank—was the book runner for issuing government bonds. But the Bank of England evolved in its role from 1800 to 1850 into also being a backstop for the merchant banks in England.

There's a charter for the Bank of England, I think it's from 1844, that formally names the Bank of England as the backstop for the domestic merchant banking sector. So it's interesting relative to the florin case because the Bank of Amsterdam was both a central bank and a bank, but the Bank of England looks much more like our modern central bank. It’s the backstop that sits above the banking system. So that's a layer of intermediation, financial technology that also plays a role in centering the world on sterling.

The answer to why it was sterling, why not [the French franc] or [the Spanish silver dollar], is less clear. The rise of England predated the empire—we normally think of empire as post-1850—whereas the rise of sterling was before that. But certainly with the rise of empire, the importance of the British in the world trading system was growing, with the world banking in terms of trade bills (the form of trade credit then). Something like 90% of world trade was settled in sterling, even when England was not involved. 

So sterling was absolutely a central reserve currency of the world from 1850. The shift that I think is interesting and identifies more was the shift from sterling to the US dollar. By roughly 1870, US GDP was larger than British GDP, if you include the British Empire and call that all British GDP. By around 1900, the US was larger. So the US, by the turn of the century, was the world's leading economic power. But sterling was still the liquidity and the central reserve currency.

That’s a case against [the theory that reserve currency status arises from] size per se. This is a time when corporate debt markets are taking off. They're all sitting in the London market. There's no Federal Reserve. There's no unified money market in the US. Financial markets are in their infancy. The US has had this history of banking crises and bank runs.

So this is a case in which you can see a superior financial system retaining reserve currency status even though a larger economy is sitting there using sterling.

Arvind Krishnamurthy | Fischer Black Visiting Professor in Finance

The way I think about this is: The US runs this gigantic parking lot where the gates work really well. You can go in and out, and that makes the whole thing work. You can have a giant parking lot, but if your gates are permanently closed, it's just not going to work.

HZ: So it's not size per se, and I'd imagine it's not military power per se because the US was also strong. But is it the other piece, which is the safety of the financial market? The Bank of England was established long before the Federal Reserve, the modern iteration of the central bank in the US. Maybe that's the contribution.

AK: I think that's right. I mean, there's some history dependence in this, which you would expect in any multiple-equilibria setting.

But the preconditions in 1900, financially, were not there in the US. They continued to hold in the UK. And in my mind, there are two events that are relevant: the Federal Reserve Act of 1913 and World War I.

The US has very little government debt in 1900. It’s paid off most of its Civil War debt. World War I has the US issuing more government debt to help finance the war effort in Europe and has Britain issuing lots of debt—in a way, too much debt—whereas the US now is in a position where it has enough debt to lubricate the system. The 1920s are the heyday of New York becoming a financial center. And then—it's hard to see over the next 20 years because the Depression and then the war are hard to read through—but certainly after the war, the world has shifted away from sterling and into the dollar. 

HZ: So it took another war. What’s really interesting here is the size of the debt. Usually we think about the size of the debt being a problem for any sovereign: You have too much debt, therefore the sovereign is less safe. But given the historical experience with the florin and the US dollar, you seem to be saying that the size of the debt market is actually important. You want to issue some debt regardless of whether you need it or not. In the florin experience, I think you're saying that [the Bank of Amsterdam] just created this debt, in an engineering sense, to support this very liquid florin market. But then, of course, if you issue too much, that can go too far. Can you elaborate on this point about the size of the debt market being crucial for creating a safe asset rather than damaging it?

AK: Yeah, absolutely. I think the two extremes you outlined are the clearest cases. If you don't have enough debt, effectively you have no liquidity. The liquid safe asset market is often the one that's created or backstopped by the government.

And even in the florin case, you know, I agree with you completely. The coin was what backed the debt, but the government was still back there because it's the government's commitment not to steal coin that ensures that you have enough of a safe asset market. The modern version of that has been government debt. If you don't have enough government debt, that's a problem. If you have too much government debt, then the marginal debt that you issue is risky. And people don't want to be using a risky asset as money. The characteristic of money is that it is effectively no-questions-asked: I want to be trading this because I'm doing something else. I don't want to be thinking about what I'm paying you with. And too much debt is a problem. 

So the two sides are very clear, and the world historically has settled somewhere in the middle. I don't have an answer to the biggest question, which is, “What numerically is the optimal size of debt?” This is, if you want, the elephant in the room in the world today.

HZ: It's literally a trillion-dollar question. 

AK: It's a trillion-dollar question. We have issued so much more debt in the last 10 years. Are we too far now? Some of the data suggest that the convenience [yield], particularly on long-term US treasury debt, has gone down substantially. That suggests the debt is too much. But I think this is the big question right now.

Insights on finance and policy, from MIT.

Subscribe to our newsletter

SIGN UP

HZ: How does one reserve currency lose its status and another gain it? Can you say a little more about this? I'd imagine looking forward, there is the euro’s project of trying to, in a way, become a strong competitor to the dollar. That has not happened yet. And then we see the Chinese yuan or maybe another emerging market currency.

Can you say a little more about what might be the tipping point and what happens in that process? For example, if [the eurozone] becomes a more important economic bloc, will that support the euro as a reserve currency? As China becomes bigger, will that support the Chinese yuan as a reserve currency? And how does the transition work?

AK: So there are two elements that need to be unpacked there. First of all, let's start with a country, let’s say Brazil, that really will not be [the reserve currency issuer] in the near future. But think about its usage of a world currency. It has been using the dollar, and the alternative is not really to use the renminbi. It's to use the dollar or use its own home currency. Every country around the world faces the trade-off of using the dominant currency, the dollar, or using its own home currency.

And so one aspect of transition is that if the dollar market becomes less liquid, there'll be a transition away from using the dollar toward using the home currency. And now, if you multiply that by world GDP, there is some fraction of world GDP that—with the dollar as a stable central currency—is using the dollar as its currency and, if you erode that fraction by 10%, everybody goes back to using their home currency by 10%. The dollar market shrinks.

Because the dollar market shrinks, its liquidity shrinks, in a sense, by more than the 10%. And so there's a multiplier. And then everyone goes back to their home currency a little bit more. In the near term, that’s the type of transition one should be thinking about: erosion at the margin of going back home, reducing the role of the dollar. 

The longer term question is: What is the competitor? In the cases we’ve talked about, the world chose the reserve currencies because it was economically efficient. There's an agglomeration benefit to everyone from using one. So this movement to home currencies is a movement in which you're giving up some of the efficiencies of having a world money. That creates a hole, a gap. What can fill the world money [gap]? For the last twenty-five years, the competitors have been the euro or the renminbi. 

The euro is a competitor in part because [the eurozone] is a large trading bloc. Economically, [the European Union] has a lot of the financial structure that the US has. So [the euro] is definitely a potential competitor, and some of the data suggest that it was catching up to the dollar until the European sovereign debt crisis. 

That’s also a really nice example of why safety is important, because you took a large stock of euro safe assets, all of [the eurozone’s] sovereign debt—and then a lot of it became less safe. Only a fraction of it post-2013 was safe. In the data, that looks like the euro’s catchup to the dollar turned and went down after 2012–2013. So the euro has been a competitor, but Europe has had sovereign debt problems. I mean, today, right now, France is in the news [for this reason].

HZ: Right, exactly. 

AK: So, that's an issue. Europe doesn't have a [eurozone-wide] safe asset. The US does. Treasury bills are a US-wide safe asset. So that's an issue to think about it. We should be thinking at the time scale of decades and asking: What will happen to Europe over the next 10, 20 years? 

The other alternative is China. China is economically very, very large. It has a debt market that's also sizable. But similar to Europe, [it’s subject to] questions about sustainability of debt and sovereign debt, local debt. There's a debt issue, credit risk issue floating around. But in my mind, the bigger issue for China is investability. Its capital market is basically closed. You can't [issue] a reserve currency if you're closed. 

The way I think about this is: The US runs this gigantic parking lot where the gates work really well. You can go in and out, and that makes the whole thing work. You can have a giant parking lot, but if your gates are permanently closed, it's just not going to work. China's experiments of trying to create an offshore yuan are little steps in that direction, but for China to be a serious competitor, it needs to [compete on] size and be very open, and it's not immediate. Maybe it'll be the case in 20 years.

HZ: Yeah, I think that's absolutely right. Maybe another commonality between China and the eurozone is that the strongest country in the eurozone, Germany, is a surplus country. It does not issue a lot of debt. From your historical perspective, it has to issue a lot more. But then of course, for historical reasons, it doesn’t want to issue more. So now I guess Germany is in a bind. And on China, I absolutely agree. For any country with capital controls, it's very hard for its currency to be the reserve currency. 

So to wrap up, Arvind, what do you think about the dollar’s status as the reserve currency? Is it going to be here for another 10, 20, 50, 100 years?

AK: I don't know about 100 years or 50 years. I can talk about the next 10 or 20 years. The data to me suggest erosion. What I described to you earlier—that feels like what’s happening right now. And it is to the detriment of both the US and the world that erosion is happening. I don't see a competitor on the horizon. 

HZ: Maybe that's the key point. You have to have a competitor to really switch. Otherwise, the relatively safest place will continue to be the reserve currency.

AK: We will continue. At a smaller scale than historically, but yes, we'll continue. That would be my forecast for the next 10 or 20 years.

About the Authors

Arvind Krishnamurthy

Email LinkedIn

Arvind Krishnamurthy is Fischer Black Visiting Professor in Finance at the MIT Sloan School of Management, the John S. Osterweis Professor of Finance at the Stanford Graduate School of Business and a research associate at the National Bureau of Economic Research (NBER). He formerly taught at the Kellogg School of Management (1998-2014).

Learn More

Haoxiang Zhu

Gordon Y Billard Professor of Finance, MIT Sloan School

Email LinkedIn

Haoxiang Zhu is the Gordon Y Billard Professor of Finance at the MIT Sloan School of Management, a Research Associate at the National Bureau of Economic Research, and a Co-Director of the MIT Golub Center for Finance and Policy. From 2021 to 2024, Zhu served as the Director of the Division of Trading and Markets at the U.S. Securities and Exchange Commission.

Learn More

• Authors’ Disclosures: The authors report no conflicts of interest. 
• The views expressed in this article are those of the authors and do not necessarily reflect the views of the MIT Golub Center for Finance and Policy, MIT Sloan, or the Massachusetts Institute of Technology.