George Selgin, senior fellow at the Cato Institute and director emeritus of its Center for Monetary and Financial Alternatives, is retiring in July 2025. Selgin is a renowned monetary theorist and historian whose work on free–market–centered banking and monetary systems spans volumes. In this post, Cato is republishing an interview he gave in 2008, but with a twist: It adds a new question—and answer—regarding the implications newer technologies have for the private production and issuance of money. (Selgin is interviewed by economist and Cato Senior Fellow Stephen Slivinski.)
The post covers core myths surrounding free markets and money and explains why scholars carefully distinguish between a free-banking regime and a monetary-base regime. It pushes back against the standard view that banking systems are inherently fragile and discusses both reasons for optimism and pessimism regarding cryptocurrency’s potential role in a freer and more sound monetary system.
Stephen Slivinski: Describe free banking. How does it differ from the sort of system we have in the United States today?
George Selgin: I use the term to mean laissez-faire banking—banking without any special government regulations or restrictions. Like free trade, it’s an ideal concept. It doesn’t refer to any specific or actual banking system, although some, like Scotland’s in the early 19th century, came close. My own ideal version of free banking would have no special requirements for note issuance. Private banks would be able to issue their own notes on the same basis as they create demand deposits. They would also be free to open branches and invest in all kinds of securities. Finally, there wouldn’t be any sort of implicit or explicit government guarantees, like deposit insurance.
SS: Is a commodity standard necessary in your hypothesized free banking environment? Or, to put it another way, is “fiat money” incompatible with free banking?
GS: I think a distinction needs to be made between the banking regime on the one hand and the monetary base regime on the other. The way I envision free banking, it does not rely on a particular base regime. It’s true, as a matter of history, that if you had free banking from the get-go, you wouldn’t have central banks, and you would almost certainly have a commodity money standard, probably gold. But one can conceive of free banking in a modern fiat money setting. What would make it free is that the central bank would not have a monopoly on issuing paper currency, the way central banks do today, almost everywhere.
A modern proposal for free banking that doesn’t radically alter the monetary base regime is one that freezes the monetary base, lets banks issue any sort of liabilities—including currency—and gets rid of deposit insurance. The central bank would still maintain the monetary base, but, in principle, it would just be a question of making sure it mopped up old central bank notes and otherwise maintained a fixed stock of reserve credit for banks to settle with. In that case, you’d have free banking with a fiat money standard.
The fiat money we currently have is purely the product of central banks. I think it’s pretty clear that if we never had central banks, we wouldn’t have fiat money. Instead, we’d still have commodity money. I don’t think there were any evolutionary forces at work that would have weaned monetary systems off established commodity standards, particularly gold and silver. What would have happened instead, and what was tending to happen while we still had those standards, was that the actual need for gold and silver as money would have fallen, thanks to financial innovations, to very trivial amounts.
In the Scottish free banking system, for example, actual gold coin reserve ratios had already fallen to as low as 1 percent to 2 percent of the banks’ outstanding demand liabilities by the 1820s. Most of the liabilities were banknotes back then—deposits weren’t so important. At any rate, the Scottish banks didn’t need a lot of gold, and the system was always finding new ways to economize on it. But the ultimate standard was still gold, and I think it would have remained so in the absence of government interference.
SS: Do you consider fractional reserve banking inherently problematic? Does free banking require a commodity standard so that private banks don’t issue too much currency?
GS: The advantages of free banking are distinct from those of the gold standard or any commodity standard. That doesn’t mean that I think there is no advantage to a gold standard. As a matter of history, I think it’s a shame that the gold standard was dismantled. That dismantling really began in earnest during World War I, and the gold standard that was restored afterward was a jury-rigged and, ultimately, very unstable standard. But one can have a better banking system under free banking, whether there is a gold standard or not. Fiat money would also work better with free banking than without it.
As for fractional reserve banking, I think it’s a wonderful institution and that it’s crazy to argue that we need to get rid of it to have a stable monetary regime. Those self-styled Austrian economists, mostly followers of Murray Rothbard, who insist on its fraudulent nature or inherent instability, are, frankly, making poor arguments. I don’t think the evidence supports their view, and think that they overlook overwhelming proof of the benefits fractional reserve banking has brought in the way of economic development by fostering investment.
The main thing to keep in mind is that a competitive bank of issue is one that can issue circulating currency but has no monopoly on doing so. So it isn’t in a position to print up its own reserves or to print anything that other banks can be counted on to treat as reserves. Free banks compete, as it were, on an even playing field in issuing paper IOUs, which are basically what banknotes are. They have to redeem those IOUs on a regular basis: The competition among different issuers means that their notes will be treated the same way that checks are treated by banks today. They will be accumulated for a day or so and then sent through the clearing system for collection. It’s this competition among issuers that assures that none of them has the power to lead the system into a general overexpansion.
That’s quite unlike the situation you have when you have a monopoly bank of issue. Even in the presence of a gold standard, when the privileged banks’ IOUs are themselves claims to gold, a monopoly bank of issue can expect other banks to treat its paper notes and its deposit credits, which are close substitutes, as reserve assets—that is, to treat them as if they were gold themselves. As a result of that tendency, which exists only because the recipient banks are deprived of the right to issue their own paper currency, the less privileged banks become dependent on the monopoly currency provider and therefore treat its notes as reserve money. Now that monopoly bank has the power to generate more reserves for the whole system and it, in turn, is free of the discipline of the clearing mechanism. That’s where central banks’ power comes from. This is what allows central banks to promote a general overexpansion of credit and inflation.
What I just described is exactly the sort of thing that triggered many of the financial crises of the 19th century. The irony is that people now see these periodic crises, especially in England, as proving the need for a central bank and a lender of last resort. Walter Bagehot, on the other hand, recognized that the boom-and-bust cycles were a product of a monopoly in currency issuance.
Today, poor Bagehot must be spinning in his grave, because your average central bank apologist likes to cite him as having argued that every country should have its own central bank. That is a calumny. Bagehot, in fact, wrote very explicitly that he thought it would have been best had there never been a Bank of England, and if England instead had a competitive banking system like Scotland’s. In that case, there would have been no need for any lender of last resort. In recommending that the Bank of England serve as such a lender, Bagehot wasn’t recommending a solution to problems inherent in unregulated banking. He was just trying to get an inherently flawed Bank of England to behave itself.
SS: You’ve already mentioned the Scottish banking system of the early 19th century as the best historical example of a functional free banking system. How did the Scottish system emerge?
GS: The Scottish system was unique, and that’s because of politics. After the 1707 Treaty of Union, English authorities did not want Scotland to end up with an institution with the same power and prestige as the Bank of England. They more or less insisted that Scotland allow open entry into the note-issuing business. So the Bank of Scotland, chartered in 1695, was followed by the Royal Bank of Scotland, and then by other note-issuing banks, until Scotland had a couple of dozen banks of issue—some big, some small—all competing. In this way, the English quite unintentionally gave Scotland the world’s most stable, most envied banking system, and one far superior to its own. For one thing, Scotland was relatively free of crises while England was buffeted by one crisis after another.
By the way, the same comparison can be made between the US banking system and the Canadian banking system in the last half of the 19th century. Neither was a free banking system, but the Canadian system was freer in crucial respects, like allowing banks to issue notes without special collateral requirements and allowing nationwide branch banking. This greater freedom made the Canadian banking system the envy of US commentators at the time.
SS: Supporters of central banking claim it is superior to free banking because the central bank can serve as a lender of last resort in a crisis or a contagion. Are there characteristics of a free banking environment that would obviate the need for a lender of last resort?
GS: The standard view is that banking systems are inherently fragile and that they’ll be subject to frequent bank runs, which, with fractional reserve banking, will have very serious consequences. But there’s no good evidence for this view.
Two things need to be said. First, truly irrational and random runs on banks, out of pure ill-informed panic, are the exception. In most cases, the runs turn out to be based on relatively accurate information about which banks are insolvent and which ones aren’t. In other words, so-called bank contagions tend to be very limited.
Secondly, the tendency for banking systems to suffer failures, especially big clusters of failures, depends on the regulatory environment. Had we had nationwide branch banking all along in the United States, that alone would have allowed us to avoid many of the bank failures and problems we’ve experienced.
So, the question that has to be asked is not whether heavily regulated and structurally weak banking systems in the past could have benefited from a lender of last resort. Perhaps they could have. It’s whether the first-best solution is to get rid of the regulations that rendered these systems so artificially fragile in the first place. I don’t think that laissez-faire or free banking systems, or the closest approximations we have been able to study, have demonstrated the sort of fragility that suggests the need for a lender of last resort at all. In my opinion, a lender of last resort is a second-best solution to problems caused by misguided regulation of banking systems.
Freedom to issue notes is important, too. When banks can’t issue their own notes, well, they need a lender of last resort to supply them with notes. If we told companies that manufacture shoes that henceforth they could only make shoes for left feet, lo and behold, there would be a need for an “emergency” source of shoes for right feet, which could be created by establishing a new government agency for the purpose. Eventually, people would say, “Thank goodness for the Government Shoe Agency. How would anyone be able to walk otherwise?”
New Question:
SS: In 2009, the world was introduced to Bitcoin, blockchains, and cryptocurrency. What implications does this technology have for the private production and issuance of circulating money? Can they be part of a system that still uses the US dollar as its base currency?
GS: It’s useful here to distinguish between two sorts of private money. One consists of private claims to some “outside” base money that are themselves convenient exchange media. Modern bank deposits are an example, as were the commercial banknotes of yore. The other consists of privately produced base monies, like the gold coins struck by private mints during the California gold rush or the private fiat-like currencies Hayek envisions in his pamphlet Denationalisation of Money.
Bitcoin comes as close as anything ever has to realizing Hayek’s vision. It’s an irredeemable financial asset that’s not a commodity, like gold, yet it shares gold’s inherent scarcity: The Bitcoin stock will never quite reach 21 million bitcoins. This actually gives Bitcoin an advantage over “Hayek” money, which relied on private fiat-type money issuers to deliberately keep their money scarce for their reputations’ sake. As Lawrence White and others have shown, those issuers might instead profit more by hyperinflating, causing a Hayek-type system to fall apart. In an essay I wrote a few years after Bitcoin was introduced, in recognition of its inherent scarcity, I classified it as a potential “synthetic commodity money,” while noting that it or something similar might have advantages not possessed by either fiat money or a genuine commodity money.
But Bitcoin hasn’t yet achieved that potential. Instead of being widely used as an exchange medium, it has become a very popular long-term investment, and the more inclined people are to hold (or “HODL”) Bitcoin, the less chance it has of ever becoming a popular exchange medium. That’s true even setting aside the huge network-effect advantage enjoyed by incumbent currencies, the US dollar above all.
Also, were Bitcoin somehow to end up serving not only as an economy’s main medium of exchange but as its medium of account, its absolutely fixed long-term stock could pose macroeconomic problems, particularly should there be any substantial increase in the real demand for it, especially relative to people’s earnings. In that case, painful deflation could occur unless the stock of actual Bitcoins is supplemented by a more elastic quantity of second-layer Bitcoin substitutes. Here, Bitcoin-based banks could help, particularly if they are “free” banks that can issue circulating Bitcoin substitutes—“Bitstablecoins,” if you like—without being constrained by minimum reserve requirements. Hal Finney, who is widely considered to have been one of Bitcoin’s developers, was looking forward to a Bitcoin-based free banking system as long ago as December 2010.
Were a true Bitcoin standard to arise anywhere, it would, of course, mean a reduced role for, if not complete elimination of, whatever national currency it displaced. As I’ve said, that remains a distant prospect, except perhaps in some very small economies: Even in El Salvador, whose government has gone to great lengths to promote Bitcoin’s adoption as money, the US dollar still dominates. Non-Bitcoin stablecoins denominated and often redeemable in established national currencies might themselves take the place of existing official paper currencies, limiting official monetary authorities’ direct money-supply role to that of supplying bank reserve credits. But in that case, those authorities would still retain overall control of their nations’ money supplies. Consequently, this sort of monetary privatization would have far less radical consequences than a switch to a Bitcoin-based monetary regime or to one based on any other privately supplied base money.
Interview excerpt reprinted with permission of the Federal Reserve Bank of Richmond. The views expressed in Region Focus are those of the contributors and not necessarily those of the Federal Reserve Bank of Richmond or the Federal Reserve System.