Thread by @GeorgeSelgin

In my last thread on Gorton and Zhang's new working paper, I criticized their criteria, in Table 1 of their paper (p. 5) for selecting among options for regulating stablecoins. Here I turn to the paragraph immediately following that table, which itself calls for a thread.

Here they claim that the option of requiring that stablecoins be 100% backed by reserves or Treasurys "ties two forms of money together at a fixed ratio," thereby making "a shortage of one pf the forms of money" likely.

To illustrate, they point to the currency shortages that plagues the pre-Fed national currency system (1866-1014). They say these were caused by nat'l banks reluctance "to move all the Treasuries to back national bank notes," that this caused deposits to expand instead, and

that in consequence "demand deposits basically became the shadow banking system of their time, and there were runs on demand deposits."

I hardly know where to begin pointing out just how perverse this account is. It actually turns the received--and correct--account of what went wrong with the national currency on its head! The problem wasn't that deposits weren't backed by bonds: it was that notes had to be!

That note backing requirement, originally aimed at financing the Civil War (it was neither necessary nor sufficient to keep national banknotes at par), ended up artificially constraining note issuance because the requisite bonds became scarce and costly to acquire,

That happened because the government ran routine post-bellum surpluses, and used them to retire its debt. Treasurys soon commanded a premium. As the premium rose, it became harder and less profitable for national banks to maintain their note circulation.

Between 1880 and 1890, as the bond premium rose, the quantity of national bank notes fell by 50%! Yet the economy was growing. There were major "currency panics" in 1884 and 1893 and another, bigger panic in 1907.

All authorities understood then, and most understand today, that the bond-backing requirement was to blame, not just for cyclical currency shortages, but for seasonal ones: hence the complaint that the U.S. currency stock was "inelastic."

Had national banks been able to issue notes on the basis of their general assets, as banks elsewhere could, there would have been no shortages. Instead, the quantity of notes, and the note-to-deposit ratio, would have varied with demand.

That's why other countries did not experience U.S.-style crisis. Canada was the most famous example of this: same gold dollar, same seasons, highly depended on exports to the U.S., yet--no currency crises. alt-m.org/2015/07/29/the…

Yet there were several dozen Canadian banks of issue during the last decades of the 19th century, their notes were backed by general bank assets (mostly loans), and they weren't insured. Here's a picture showing how the currency supply behaved there and in the US:

(Bear in mind Canada was about 1/10th the size of the U.S. in wealth. Hence the difference absolute magnitudes.) Canada did experience occasional bank runs and failures. But they mostly involved small banks, and compared to the U.S. losses to bank liability holders were trivial.

GZ are, so far as I know, the only authors to try and suggest that pre-Fed U.S. banking crises were a result of too many bank deposits, rather than too few national banknotes!

I submit they offer this strange view because they need a story that squares with their claim that private money can't work unless its fully gov't backed. So: backed national bank notes = good private money; unbacked deposits = bad private money. Convenient. But wrong-Very wrong.