Thread by @GeorgeSelgin

Reading Gorton and Zhang's new WP on stablecoins, I have the sinking feeling I'd have to write a book to expose all of its errors. The first pages alone contain another paper's worth. papers.ssrn.com/sol3/papers.cf…

Consider this closing claim of their paper's opening paragraph: "'Stablecoins'...aspire to be used as a form of private money and so are allegedly backed one-for-one with government fiat currency." "And so"?

Do CZ mean to say that anything that "aspires" to be private money must pretend to be 100% fiat backed? What, then, of transferable bank deposits? Are they not "private money" in fact? Do their issuers pretend that they are 100% fiat backed? I don't think so.

Likewise, throughout history, private (commercial) banknotes that have served as money have hardly ever been so backed, and (despite what some tin-foil hat Austrians and crappy textbooks say), their issuers have seldom claimed otherwise. cambridge.org/core/journals/…

Finally, most stablecoins themselves don't claim 100% fiat backing. Tether suggested it was so backed--and got called to the carpet for it. Some others have been less than fully transparent. But not all are equally irresponsible. coindesk.com/tethers-report…

If that GZ sentence is bad, the next is far worse. Here they assert that "private money is a subpar medium of exchange and is subject to runs." The first part of that claim is clearly untrue of bank deposits, which once again are also "private money."

If GZ mean, "private currency," they should say so. But if they did say so, their sweeping generalization would still be false. Based as it presumably is on antebellum U.S. experience, it's symptomatic of the parochialism of which all too many U.S. trained economists are guilty.

In fact the world offers numerous examples of competitively-issued commercially issued banknotes that traded throughout national markets at par w/ official money, despite being backed mainly by bank loans.

In fact, outside of the antebellum U.S. case, I'm unaware of any exception. Sometimes (as in Scotland during the Bank restriction) notes were discounted relative to former metallic standard, but they stayed at par relative to a temporary standard--here the BofE paper pound.

If GZ are aware of what happened in dozens of other countries that allowed numerous private banks to issue redeemable banknotes, they show no sign of it. Instead, to support their sweeping generalization, they resort to a priori reasoning. And what reasoning!

Their "proof" rests on appeal to what they call money's "most obvious," essential property, namely. that it "must satisfy the no-questions-asked ('NQA') principle, which requires the money be accepted in a transaction without due diligence on its value."

What's that? You say you've never hear of this supposed principle? Well, join the club! But let's consider it. Obviously it can't "prove" that commercial banknotes backed by diverse assets can't circulate at par, anymore than physics can prove that bumblebees can't fly.

Yet it's true that asking people to do due diligence on any banknote proffered them could be asking a lot. So how do we reconcile these facts? Once again, we do so by resisting the temptation to generalize from the United States unique antebellum experience.

That experience was unique because, thanks to legal restrictions, almost all U.S. banks back then were "unit" banks, that is, banks with one office and no branches. The flip side of this lack of branch banking was a very large number of banks--some 1400 or so by 1861.

In contrast, in most historical note issue systems, there were at most several dozen note brands only. For that reason alone, the "due diligence" burden was much lighter. Nor is that all.

In practice most people only had to practice due diligence in choosing one particular bank to do business with. They could then rely on that bank's determination of which other banks' notes were "current" ones they'd accept at par, to decide which notes to accept themselves.

A bank's membership in a clearinghouse association could also serve as a "seal of approval" of its notes' reliability. Despite unit banking, New England achieved a uniform banknote currency as early as 1824...

...thanks to the diligence practiced by the Suffolk Bank in its capacity as an early clearinghouse (encyclopedia-of-money.blogspot.com/2013/02/suffol…). What, I wonder, do GZ have to say about the Suffolk System? What do they have to say about numerous later clearinghouses that also policed their members?

CZ's claim that privately-produced monies are vulnerable to runs also reflects their narrow outlook. Bank runs may have been common enough in some plural note issue systems. But they weren't common in others.

Here again, unit banking played a part. It weakened banks confined to it by limiting their opportunities for asset and liability diversification. Branch banking systems, in contrast, could rely on sound branches to help others suffering from exposure to local shocks.

Of course, size matters here. Consider Canada's pre-1935 experience. Bank runs and failures there were almost entirely confined to small (sometimes tiny) banks. The larger Canadian banks were largely unaffected. (It isn't clear, BTW, that the optimal # of riskier banks is zero.)

GZ also fail to consider that a private note issuing bank's vulnerability of runs also depends on the nature of its banknote and deposit contracts. In principle, those contracts can be designed to rule-out panic based runs.

In fact, there's a very good paper on this by a fellow named...Gary Gorton! Wonder whatever happened to him: …d1a-wp-offload-media.s3.amazonaws.com/faculty/wp-con…

For my part, I've shown that, despite a claim made by Diamond and Dybvig in their famous paper on runs, suspension contracts can be perfectly consistent with consumers' welfare: link.springer.com/article/10.100…

In practice, in the U.S. and elsewhere, governments often prevented banks from including suspension clauses in their contracts, so they could not limit runs that way. Even so, runs were generally far less common than they were in the pre-FDIC U.S. cato.org/sites/cato.org…

And let's not forget that runs are not necessarily bad in themselves: they can ultimately spare an irresponsibly managed bank's creditors bigger losses. They are bad when they have contagion or other spillover effects on responsibly-managed banks. fraser.stlouisfed.org/files/docs/his…

(The failure of a stablecoin issuer or two might likewise prove salutary!)

So much for GZ's first page. I hope it serves as a warning to all that Gorton and Zhang's arguments are not much more worthy of implicit trust than a ca. 1838 Michigan bank note.