This is an unedited draft for my projected banking manuscript. It might be an idea to embed some of this content into earlier articles that discussed the self-funding nature of the banking system. Although I planned to do a different article first, I decided to add this discussion in response to some reader feedback.
When discussing the self-funding nature of the banking system, the risk is that my arguments may suggest to some readers that a bank can make whatever loans it wishes without ever worrying about funding. I was conscious of this comprehension risk when setting out my examples — I tried to emphasise the practical limits of what the bank can do. However, readers might skip over the numbers, or might just see out-of-context quotations of what I write. Rather than bury everything I previously wrote under a layer of waffle, I want to break out the concerns herein into a stand-alone discussion.
Why Self-Funding Matters
My discussion of self-funding is aimed at dubious attempts to resurrect loanable funds theory — the idea that the pre-existing pool of “savings” is fixed, and all that banks do is redirect the existing savings from “savers” to “borrowers.” We could delete banks from the economic model (or even the real world economy) and nothing would allegedly change — other than the greater effort expended by “savers” to find “borrowers.” The key point is that the amount of savings is fixed, and so we can use “supply and demand curves” to make statements about the effects on interest rates in order for “the savings market to clear.” To the extent that people try to pretend that loanable funds is a reasonable theory, they need to ignore the “pool of savings is fixed” part of loanable funds, and instead play word games to obscure the argument.
We need to keep track of transactions — if a bank loses deposits due to outflows, that is an injection of settlement balances into the rest of the financial system. Those settlement balances have to go somewhere — and the implication is that the bank can theoretically draw on those balances in the inter-bank or wholesale funding markets.
There is no guarantee that a bank will get those funds. However, in practice, large banks have treasury teams that go out and get the requisite funding (or place excess funding into other financial assets) every working day, year in, year out. Although I can imagine some small banks have failed due to incompetence, I am unaware of any major banks in the modern era suddenly failing solely because their treasury teams miscalculated. (Central banks being the lender-of-last-resort of course helps.)
Fragmentation
It is possible to question my assumption of looking at banking from the perspective of modern, large, money-centre banks that have full access to almost all domestic funding markets (and can even draw on foreign funding via instruments like currency swaps). These are the banks that dominate activity in most developed financial systems. This is less true of the United States — the country of most of my readers (as based on book sales data). And if one wants to discuss historical experiences from before the deregulation of banking and finance (as late as 1990, depending upon the jurisdiction), traditional banks were less integrated with other funding markets.
If a bank is not fully integrated with all funding markets, the observation that the financial system is self-funding may not be entirely helpful. Such a bank cannot blindly extend loans without worrying about funding — since the settlement balances could “leak” into markets that it cannot draw from.
An Artisanal Bank
Let us imagine we looking at an artisanal bank that has funding that relies solely upon demand deposits, term deposits, equity (and in the worst case, discount window borrowing). It is cut off from funding markets — so the self-funding nature of the system is perhaps not helpful. It holds high quality money market assets, and possibly the liabilities of a “senior” bank.
Such a bank cannot hope to survive without having a large liquidity buffer — implying that is going to have much more deposits than loans outstanding. The excess of deposits ends up in the high quality money markets. Inflows and outflows from the bank are going to be matched with outflows/inflows from that market (or the senior bank). How the self-funding property of the system shows up is that if the bank needs to liquidate some Treasury bills, it has injected the funds into the system that matches the liquidity needs of the entity that buys the Treasury bills. This is not too surprising — nobody would expect that a tiny bank selling some of its Treasury bill portfolio is going to cause the bill market to seize up. (In any event, the central bank acts as the de facto back stop of the bill market.)
The need for a large liquidity buffer is not contradictory to my previous examples. The example bank has a target liquidity ratio as well as a regulatory minimum — and both of those were my arbitrary choices because they made the numbers easier to follow. (Since banks cannot completely control their inflows and outflows, they really need to have a target liquidity ratio range — which will be a band around some target level. I just refer to a “target ratio” rather than “target band” for simplicity.) The principle that matters is that real banks have to set a target liquidity ratio — and that ratio is going to depend upon the circumstances of the bank. If the bank has no alternative short-term funding sources, it has no alternatives but wait for new deposits to expand its liquidity buffer. Large banks set up connections to wholesale funding markets for a reason — they can cut back on their liquidity buffers (which have a lower expected returns than other assets).
A traditional bank cannot expect to survive if it does not make loans. Which means that it has to expect to face outflows. What matters is that they need to rebuild their liquidity — either via waiting for deposit inflows, or via issuing funding instruments. If a bank decides to cut itself off from all funding instruments, it has no choice but to wait for new deposits to “walk in the front door” before ramping up lending again. However, banks that decide to follow that strategy are not a significant portion of the financial system in modern economies.
In addition to the passage of time, even the least sophisticated banks can attempt to convert their demand deposits into term deposits to lock in term funding. (Savings accounts may also help.) A big bank can issue bonds, but there is a minimum size for a viable bond issue — term deposits are just a retail-sized bond (with a bit more flexibility).
Another possibility in the modern era is mortgage securitisation. In the United States, the main agencies work with banks of all sizes. As per Freddie Mac “Those divisions work with lenders of all sizes – national, regional, and community lenders and credit unions – to buy conventional, conforming mortgage loans for one- to four-unit homes – including condominiums and manufactured homes – up to a certain dollar amount set by our regulator.” (URL: https://www.freddiemac.com/about/business, quote taken on 2024-05-01.) So long as the bank is big enough to qualify as a “community lender,” the resources exist to allow the bank to get mortgages off its balance sheet. And for a community bank, that plus term deposits might be enough to deal with “strategic” management of the liquidity ratio — the liquidity portfolio would absorb short-term net flows.
History
One of the key problems with economist writing about banking is that they spend far too much time looking at historical banking systems, particularly in the United States. There is partly for ideological reasons — hard money bugs want to paint the Gold Standard era as being “normal.” For those of us without fixations on shiny rocks, this is not a concern. (As a disclaimer, I own units in a Canadian gold mine fund at the time of writing, but I am not entirely sure why.)
Historical behaviour of banks only matters if you have access to a time machine. Current operating procedures at large banks will bear no resemblance to banks of earlier eras.
There are two key differences. Firstly, old school management would be viewed as horribly amateurish by modern standards. This certainly reflects the technological gap — the increase in capabilities of digital computers allows much more effective monitoring of a bank’s situation, as well as risk analysis. Secondly, many of the funding markets that large banks use did not exist.
The situation in the United States was even more complex because of the fragmentation of banking. Small banks in rural areas had highly seasonal cash flows — courtesy of farmers’ needs to finance the purchase of seed and fertiliser, then lumpy cash flows as crops are harvested and sold. This created large seasonal net flows between various segments of the banking system. (In countries like Canada with a more centralised banking system, those seasonal flows would be moving between branches of the same bank.) Although some readers might find a discussion of this era interesting, one needs to be very careful — it will offer almost no useful information for understanding the dynamics of a modern banking system.
It is completely unsurprising that banks with treasury teams with less analytical support and with very few tools for managing liquidity would have to be quite cautious with cash management. However, this is not something that helps understand the core of modern financial systems.
Concluding Remarks
Banking is a cyclic business. Banks extend loans that leads to deposit drawdowns, which is then compensated by tapping into funding markets (or the passage of time). Treasury teams do not worry about funding existing, rather, they are worried about the price of funding, and whether it will be extended to them in particular.
The discussion of the self-funding nature of bank lending is aimed at loanable funds theories, which suggests that the funding might not exist anywhere if the savers do not arrive first.
No comments:
Post a Comment
Note: Posts are manually moderated, with a varying delay. Some disappear.
The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.
Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.