Following my post about the problems with the MMT and endogenous money banking theory, Scott Fullwiler, one of its proponents, briefly (and very nicely) commented on it, suggesting that I was not criticising the theory and that we were in fact in agreement. I beg to differ.

Another commenter, JH, also left a link to a much more comprehensive article describing the theory, written by Scott (you can find it here). I recommend this article to everyone. It is a very interesting piece about banking, and Scott clearly demonstrates in it that his knowledge of the banking system is superior to most of his fellow economists.

I had started to write a fairly long post criticising the (few, but in my view, important) errors in Scott’s paper, but decided against it eventually, and deleted most of it to get directly to the main point. I could probably write a whole academic article about the topic but I have no idea how to get it published so won’t do it! (any advice appreciated though!)

Overall, I agree that Scott’s description of the lending process is largely accurate. However, I believe that the conclusions that seem to ‘naturally’ follow fall into a fallacy of composition. The endogenous money theory correctly assumes that the lending decision regarding a single loan is independent of the reserve status of the bank. However, the theory incorrectly assumes that this description also applies to lending as a whole.

The endogenous money theory correctly describes one-off borrowings from banks that temporarily lack reserves for some technical reasons. But this lack of reserves is only temporary as they have the necessary liquid holdings to generate new primary reserves. If banks can lend independently of their reserve status, it is because they have beforehand secured enough liquidity (= claims on primary reserves) to face settlements.

The fallacy of composition involves applying this reasoning to a bank’s aggregate lending. What happens as a one-off event cannot happen continuously, as it would progressively deplete the bank’s secondary reserves until it ends up only relying on interbank or central bank funding for marginal lending, assuming funding costs were maintained at a stable level. But they are not. The more secondary reserves fall, the more the bank’s ratings are cut, its cost of funding increases and its share price falls. At some point, not only the marginal increase in lending is not profitable anymore, but also the bank might have endangered its very existence by becoming borderline liquid resulting in all market actors getting hesitant to provide it with any fund.

Therefore, Scott is right when he says that we agree that banks are pricing-constrained. Indeed. But we disagree on the backstop mechanism. The endogenous view considers central bank funding (and pricing) as the backstop mechanism, against which banks are going to benchmark their new lending to assess its profitability (the interest rate spread) if other sources of funds are unavailable. As Scott says:

Borrowings and reserve balances can always be had at some rate of interest; the question is whether or not this rate of interest is one at which the bank can make a profit that provides a sufficient return on equity.

Scott here mainly refers to the lender of last resort, the central bank. This implies that the supply of reserve by the central bank is perfectly elastic at a given interest rate, and therefore entirely driven by demand. But he does not take into account the impact for a bank of being able to raise funds only from a central bank. Sure, a bank that cannot seem to find any reserve anywhere else can still usually borrow from the central bank. Nonetheless, this bank, is, well… screwed. It just committed suicide.

By overexpanding, it depleted its liquidity position (through adverse clearing) to such an extent that no market actor was willing to lend to it anymore. By admitting its new reliance on central bank funding for survival, it admits its failure. As I have already said in my previous post, this is why banks are doing their best to avoid using central banks’ facilities. Nonetheless, this bank has the possibility to regain market confidence: it can reduce its lending in order to generate new liquidity. This shows the limit of the endogenous money theory: in the medium-term, lending is indirectly reserve-constrained.

We can see that the benchmark against which banks assess the profitability of new lending isn’t the central bank’s rate. It is the various market rates. Even without the central bank changing its target rate, an overexpanding bank will inevitably be forced to contract its lending by market forces, and consequently, the money supply.

Endogenous money theorists could respond that financial markets act irrationally: there is no point in punishing banks that could actually obtain funding from the central bank, making liquidity risk irrelevant (at least as long as they are only illiquid and not insolvent).  But investors have very good reasons: illiquid banks are usually either bailed-out or left to fail, with in either case serious consequences for shareholders, bondholders, and sometimes depositors. Think about Northern Rock and Bear Stearns. Those two banks were notoriously illiquid (rather than insolvent). You know what happened next.

Regarding the treatment of reserve requirements, I am not going to come back to the evidence from countries that actively use them as a policy tool. Nonetheless, the endogenous theory concludes that reserves are basically useless, except for interbank settlements and to comply with reserve requirements when needed (even cash withdrawals by customers are downplayed). I disagree; reserves are the most liquid asset banks can hold. When uncertainty increases in the markets and when even liquid securities suddenly become illiquid, banks increase their reserves holdings, which become precautionary reserves. Banks thus sacrifice some yield for liquidity. This does not mean that banks would not invest those reserves in loans or in securities should the economic conditions be normal.

I am going to stop here for today. There are a few other points I could have covered but as I said at the beginning of this post, this would require a 20-page article… Some of those include an overexpanding banking system as a whole (not just a single bank), the fact that, unlike what is implied in Scott’s article, banks are not maximising leverage, that leverage does seriously impact banks’ ratings, that deposits aren’t always the cheapest funding source, and even some details about banks’ regulatory capital calculations. Those are less important (or even very minor) points.

Update: See this post on the central bank funding stigma.