What does “proof of reserves” prove?

Frances Coppola argues in a recent post that proof of reserves as practised by the crypto finance community proves nothing. I would be interested to read any rebuttals, but the arguments she advances in support of this claim looks pretty sound to me.

Frances starts with the observation that the concept of “reserves” is not well understood even in conventional banking.

In the banking world, we have now, after many years of confusion, broadly reached agreement that the term “reserves” specifically means the liquidity that banks need to settle deposit withdrawals and make payments. This liquidity is narrowly defined as central bank deposits and physical currency – what is usually known as “base money” or M0, and we could perhaps also (though, strictly speaking, incorrectly) deem “cash”.

“Proof of reserves is proof of nothing” Coppola Comment 16 Feb 2023

This certainly rings true to me. I often see “reserves” confused with capital when reserves are really a liquidity tool. If you are still reading, I suspect you are ready to jump ship fearing a pedantic discussion of obscure banking terminology. Bear with me.

If you have even a glancing interested in crypto you will probably have encountered the complaint that traditional banks engage in the dubious (if not outrightly nefarious) practice of fractional reserve banking. A full discussion of the pros and cons of fractional reserve banking is a topic for another day. The key point for this post is that the crypto community will frequently claim that their crypto alternative for a TradFi activity like deposit taking is fully reserved and hence safer.

The published “proof of reserves” is intended therefore to demonstrate that the activity being measured (e.g. a stablecoin) is in fact fully reserved and hence much safer than bank deposits which are only fractionally reserved. Some of the cryptographic processes (e.g. Merkle trees) employed to allow customers to verify that their account balance is included in the proof are interesting but Frances’ post lists a number of big picture concerns with the crypto claim:

1. The assets implicitly classified as reserves in the crypto proof do not meet the standards of risk and liquidity applied to reserves included in the banking measure; they are not really “reserves” at all as the concept is commonly understood in conventional banking
2. As a result the crypto entity may in fact be engaging in fractional reserve banking just like a conventional bank but with riskier less liquid assets and much less liquidity and capital
3. The crypto proof of “reserves” held against customer liabilities also says nothing about the extent to which the crypto entity has taken on other liabilities which may also have a claim on the assets that are claimed to be fully covering the customer deposits.

Crypto people complain that traditional banks don’t have 100% cash backing for their deposits, then claim stablecoins, exchanges and crypto lenders are “fully reserved” even if their assets consist largely of illiquid loans and securities. But this is actually what the asset base of traditional banks looks like. 

Let me know what I missing ….

Tony – From the Outside