On a recent episode of my personal podcast, I laid out why I thought economist George Selgin—who is a leader in the “free banking” camp—was contradicting himself on the question of whether fractional reserve banking (FRB) caused the business cycle. Since then, George has responded to me, but even in light of his response I still think the contradiction is stark.
Inasmuch as stablecoins are the modern version of the classic banknotes—and government regulations insist upon “full reserves” in this context though not for standard checking accounts—this seemingly niche controversy is of relevance to the infineo audience. Consequently I’ll distill my arguments into print form to make it easier for George to continue the debate.
The Apparent Contradiction I Originally Flagged
What’s funny about this whole episode is that I thought I was going to use a series of simple thought experiments to lead George slowly but surely into my trap, when instead he just jumped right into it and shouted that he had done so with a megaphone. When I first saw his move, I was so surprised I even asked some of my colleagues, “If he believes this, then what we were arguing about in our public debate at the SoHo Forum?!”
So in this section of the post, let me summarize what happened on this count, and then in the next section I’ll explain how George reacted when I put out my earlier podcast episode making these very points.
Here was my list of simple questions to George posed on X:
In response, George readily admitted that scenario 1) constituted saving, but that the other scenarios didn’t constitute additional saving beyond the original $100. This didn’t surprise me; I had deliberately chosen my scenarios to be so simple that George’s answers were undeniable. Yet what floored me was that George went ahead and anticipated where I had planned on taking him:
So to be clear, here George was admitting that if commercial banks engage in FRB, then when a household deposits base money into a checking account, the banking system (other things equal) issues more loans which obviously don’t correspond to genuine saving on the part of the households. This is the essence of the “circulation credit theory of the trade cycle”—a.k.a. Austrian business cycle theory—that Ludwig von Mises developed. (I spelled out Mises’ views in this recent podcast episode.)
So to reiterate, I took the above Twitter exchange as the basis for my podcast victory lap, explaining that George (apparently) just conceded the very thing that I had been debating him about for several years at this point. In other words, all I have been trying to say is that when households save some of their income and then deposit those funds with a commercial bank, if that institution engages in FRB, then it provides more loans to the community than were originally saved. It is that mismatch—namely, between the savings lent to the banks and the savings lent out from the banks—that constitutes the Misesian theory of the business cycle. Again, I thought I was going to have to use my simple thought experiments to drag this conclusion out from George, kicking and screaming; but instead, he volunteered it himself!
Now it’s true, after my initial shock subsided, and in discussion with other experts on the topic, I had a hunch as to how George would argue that he had been consistent the whole time. Specifically, I re-read two of his blog posts (which he wrote after our SoHo forum debate, so I didn’t have access to them at the time we publicly clashed) that were devoted to this specific question of FRB and the Austrian theory of the business cycle. These posts (one and two) provide all one needs to explain George’s attempted reconciliation.
To be clear, I did anticipate George’s defense of himself in my podcast episode. But since that aired, he has written to me directly (and given permission to reprint). So rather than rehash what I anticipated George would say in reaction to my accusation above, let me go ahead and quote him verbatim in the next section, where I will then reiterate why I don’t think his answer rescues his position.
Selgin Attempts to Exit the Trap He Walked Into
Here is the full text of the email George sent after my episode aired, which (to reiterate) he gave me permission to reprint:
Dear Bob,
An increase of X ounces of gold into a banking system with an established reserve ratio of .5 will result in the very same _percentage_ increase in the total money stock and other nominal magnitudes as would result from the same absolute increase in a 100% reserve system. There is therefore no reason to suppose that it would have greater cyclical consequences. That the magnitude of nominal changes in the first case is double that of the second is irrelevant, because the initial nominal scale is also twice what it would be in the 100% case.
Sincerely,
George
Let me spell out what George is arguing, where I am confident in my elaboration because George’s email is consistent with his blog posts (one and two) in which he gave his thoughts on the connection—if any—between FRB and the Austrian business cycle.
First, suppose we had a 100% reserve banking system, and that the actually “base money” of the community consisted of one-ounce gold coins. Further suppose that we start out with a total money stock of one million gold ounces in the community. And just to keep the math as easy as possible, suppose that everybody keeps all of his or her money in the form of checking account balances at a commercial bank. (This isn’t realistic, but I don’t want readers getting mixed up between the bank’s reserve ratio versus the fraction of an individual’s money balances divided between gold coins in their pocket and gold coins deposited in their checking account.)
So initially, the various banks have all one million ounces of gold in their vaults, “backing up” 100% of their customers’ checking account balances, which sum to one million ounces.
In this framework, suppose that there is a sudden gold rush and miners quickly bring an additional one million ounces of gold into the community, which are stamped by the mints into one million new coins, and which quickly find their way into the various bank vaults where they serve to back up an additional million ounces in checking account balances. When the dust settles, the influx an additional million ounces of physical gold has led to an increase in one million ounces of money held by the community, in the form of checking account balances (which went from an initial 1 million to 2 million).
Second, George now wants us to instead imagine the same type of community, but instead of the banks practicing 100% reserves, they only maintain (on average) 50% reserves. So in the initial equilibrium, when there were 1 million gold ounces sitting in the vaults of the banks, there were now 2 million ounces of gold if you added up everybody’s checking account balances. (In other words, if there were a run on the bank and everybody tried to take his or her gold out, only half of the stated amounts could be redeemed immediately before the vaults ran empty.)
In this new scenario, what happens when the miners bring an additional million ounces of physical gold into the community? As before, we assume they are stamped into gold coins, which then quickly find their way into various bank vaults where they serve as 50% backing of outstanding customer checking account balances. When the dust settles, the influx of 1 million in new physical gold reserves increases the total amount of customer checking accounts from 2 million to 4 million.
And thus, we see George’s point: So long as the commercial banks’ reserve ratio is constant, its absolute level is (he thinks) irrelevant on this point. A given percentage increase in the base stock of money (gold in our example, but it could be Fed notes/reserves in our current real-world system) that finds its way into bank vaults serves to increase the total stock of money (measured as checking account balances) by the same percentage, regardless of the reserve ratio.
Therefore, George concludes, there can’t be any systematic mismatch between household saving and bank lending (as I was alleging) due to FRB per se, because an expansion in the base money has the same proportional effect on the total money stock regardless of the reserve ratio (so long as it doesn’t change in the meantime).
Now that I’ve spelled out George’s defense, let me explain why I think it fails.
Selgin’s Arithmetical Argument Doesn’t Eliminate the Economic Problem
There are two main problems I see with George’s attempted reconciliation of his position. But before I describe them, let me first explain to the reader an important nuance in the Misesian diagnosis of how FRB causes the business cycle.
Specifically, Mises argues that it is a flow issue, not a stock one. What I mean is that (according to Mises) the distortionary effects occur when newly created bank money is lent into existence and thereby temporarily pushes the market interest rate below its “natural” level. But once a new burst of bank-created money has worked its way through the economy, raising prices and wages along the way, the economy settles into a new equilibrium, and the market interest rate once again returns to its “natural” level. Only an additional flow of new bank money into the economy (via the loanable funds market) would once again push down the interest rate, and thereby contribute to an unsustainable boom.
Here is Mises in his own words explaining his view:
The term credit expansion has often been misinterpreted. It is important to realize that commodity credit cannot be expanded. The only vehicle of credit expansion is circulation credit. But the granting of circulation credit does not always mean credit expansion. If the amount of fiduciary media previously issued has consummated all its effects upon the market, if prices, wage rates, and interest rates have been adjusted to the total supply of money proper plus fiduciary media (supply of money in the broader sense), granting of circulation credit without a further increase in the quantity of fiduciary media is no longer credit expansion. Credit expansion is present only if credit is granted by the issue of an additional amount of fiduciary media, not if banks lend anew fiduciary media paid back to them by the old debtors. [Human Action, Scholar’s Edition, p. 431]
To relate Mises’ views to our numerical example above: If the banks maintain 50% reserves with 1 million gold ounces in their vaults, then the total money (broadly construed) held by the public will be 2 million gold ounces. As that excess 1 million ounces of “fiduciary media” are periodically returned to the banks as interest/principal payment on loans, the banks can continue to extend new loans and roll those credits over. So long as the amount of fiduciary media stays roughly fixed at 1 million ounces, however, there is no ongoing distortion of the credit markets, and no impetus to an unsustainable boom.
The First Problem With Selgin’s Defense (Even When We Assume the Base Money Is Government Fiat Currency)
We are now (finally!) able for me to quickly explain the fatal weaknesses in George’s defense of his position. In the first place, even if we grant that the underlying base money is a government fiat money, and so in a sense its injection (by the Federal Reserve in the modern US case) is “artificial” and distortionary, it is still true that the commercial banks compound the problem if they engage in FRB on top of the government/Fed’s fiat inflation.
George seems to be reasoning like this: Yes, if someone takes a $100 Federal Reserve Note and puts it into his checking account, then in a 50% reserve framework the bank creates an additional $100 that is not backed up by household saving. But—George seems to think—this proportionally is no worse than in the 100% reserve setting, because with 50% reserves prices will be double what they would have been with 100% reserves. And so, the economic significance of the Fed first injecting a new $100 bill, with the commercial banks then (collectively, in the aggregate) creating an additional $100 in credit, is just as high in “real” (inflation-adjusted) terms compared to the situation where the Fed just inject a new $100 bill, while the banks don’t create any new money. In other words, George seems to be arguing that with 100% reserves banking, the Fed injecting a new $100 bill is just as distortionary as the Fed injecting a $100 bill with 50% reserve banking, because prices are twice as high in the latter scenario (where a total of $200 is ultimately injected).
Yet even if we stipulate this reasoning, George has still left open the possibility that the public might take previously issued Federal Reserve Notes and deposit them into their checking accounts. In other words, suppose for some reason the community shifts its preferences, and wants to hold a higher proportion of its “total cash balances” in the form of checking account deposits rather than actual base money (in their wallets or purses). Then, according to George’s tweet quoted above, that necessarily means the banks contribute to an Austrian business cycle if they engage in FRB. In contrast, if the banks practice 100% reserves, then the public’s changing desire to hold different proportions of their cash in the form of base versus bank money, has no effect on the total money stock or interest rates. (To be clear, I’m not talking about the banks reducing their reserve ratios. I’m talking about the public deciding to shift its money from its wallets/purses into their checking accounts, with the banks holding a fixed reserve ratio.)
So to summarize, in a 100% reserve system, if for some reason that public decides to reduce the amount of $10, $20, and $100 bills in their wallets and purses and instead hold them in their bank checking accounts, that decision itself has no further inflationary consequences. When those $10, $20, and $100 bills initially entered the economy from the Fed/Treasury printing press, sure they caused prices to rise and may have distorted relative prices (including interest rates). But once things settled down, that was the end of the damage the Fed/Treasury had done. If the public then decides to deposit them in commercial banks, nothing further happens—if the banks maintain 100% reserves. Yet if the banks only keep (say) 50% reserves, then the decision of the public to change the composition of how they hold their money does have further inflationary/distortionary effects.
The Second Problem With Selgin’s Defense (If We Assume Base Money Is a Commodity Money or CryptoCurrency)
But there is a completely independent problem with George’s defense, beyond the one I just pointed out. Specifically, if we assume that the base money is not a form of central bank-produced fiat currency, then it’s crystal clear that George’s “proportionality” argument collapses.
Go back to our original gold example. Recall that we had tried to steelman George’s position in this way: If the community starts out with 1 million gold coins and then a new million ounces quickly flow in from the mines, and if we assume everybody prefers to hold all of his or her “cash balances” in the form of bank checking deposits (rather than physical coins clanking in pockets and purses), then we can see that the total money stock ends up doubling, whether the banks practice 100% or 50% reserves. In the former case, the total money stock doubles from 1 million to 2 million gold ounces. In the latter case, the total money stock doubles from 2 million to 4 million ounces. “So what’s the big deal?” asks George.
I’m not going to spell out the whole argument here—for that, interested readers should review my podcast episode for the Mises Institute—but suffice it to say, Mises didn’t think that the boom-bust cycle was caused by monetary inflation per se. Rather, it was specifically caused by monetary inflation entering the economy via the loan sector. When the “Cantillon effects” of new money come in through the banks making loans, interest rates are distorted early in the process. That’s what causes the boom-bust cycle. Or, to put it another way, if the Fed just did a “helicopter drop” and gave every household twice as much money, that wouldn’t cause an unsustainable boom; it would merely cause rampant price inflation.
No, the critical element of the Misesian story is that the new money enters the economy via the loan market. So if we have an economy where there is a central bank issuing government fiat currency, then George could be forgiven for assuming (when he answered me on Twitter) that the $100 in currency itself had come into the system via open market operations, i.e. by being lent into the economy by the Fed (and hence distorting interest rates).
But in general, if we are imagining an economy using gold, silver, or even Bitcoin as the underlying base money—which is just the type of “free banking” system that George and his frequent co-author Larry White endorse!—then this doesn’t hold at all. No, if the community starts out with 1 million ounces of gold in bank vaults, and then the miners dig up an additional 1 million ounces, that new money does not enter the economy via the loan market. No, it enters via whatever channels the miners initially spend their money on. Furthermore, even if the miners initially deposit the new gold into their bank accounts early in the process, under 100% reserve banking that in no way represents a mismatch between saving and lending. The miner depositing gold coins into his checking account isn’t “lending to himself.” As George had no problem admitting on Twitter, that man is simply changing the form of his savings.
And so we see that in general, and perhaps more significant, in the very framework of a market-produced commodity (or “synthetic commodity” as George describes Bitcoin) base money with free-banking to supplement it, the flow of new reserves into the banks doesn’t represent a mismatch between household saving and bank loans. So there’s nothing there to be “scaled up” by FRB. No, the mismatch between household and bank loans that George correctly admits on Twitter, is created entirely by the commercial banks practicing FRB in this type of framework.
Conclusion
As I hope I’ve demonstrated in this lengthy post, when George Selgin breezily answered my thought experiments on Twitter and then volunteered that they would trigger the Austrian business cycle, he unwittingly conceded the very thing we had been debating at the SoHo Forum years earlier.
At best, George can only rescue his position from his tweet by saying, “Bob, if we assume that the $100 in your scenario is newly issued from the Fed, and that the public is keeping a constant ratio of cash/bank balances in their portfolios, and that the commercial banks are keeping a constant reserve ratio, then we can avoid FRB triggering the boom-bust cycle that I admitted on Twitter would otherwise happen.” Even here, that hardly seems like a ringing defense.
Yet it’s worse, because even that walking-on-eggshells argument collapses if I change the thought experiment away from people using US dollars, and switch instead to George’s preferred system where they use gold as the underlying base money. In that case, I think it is literally impossible for George to avoid admitting that my side was the correct one at the SoHo Forum.
(In the next post, I will offer a completely independent problem with George’s Twitter responses. Stay tuned!)
Dr. Robert P. Murphy is the Chief Economist at infineo, bridging together the dependability of Whole Life insurance policies with the benefits of blockchain-based finance.
Twitter: @infineogroup, @BobMurphyEcon
Linkedin: infineo group, Robert Murphy
Youtube: infineo group
To learn more about infineo, please visit the infineo website.