Part 1 is here.
Having used monetary policy and proved its ability to manage big things and even force onto the monetary system and the economy big changes, the Federal Reserve grew in esteem and prestige so that what followed was the hardening assumption that monetary policy could be used similarly for all circumstances. It was not just the burgeoning class of activist and interventionist policymakers who had bought into this assumption, as by and large the public did, too.
That was, in fact, the notable distinction for the whole decade of the 1920’s. Bank reserves overall swelled enormously at that time, not via actually money in bank vaults but instead through this derivative claim of money on the central bank. Bank reserves throughout that decade were increasingly of the “borrowed” format. A lot of that was due to technical innovations, including the first federal funds transactions in 1920. Banks that held “excess” reserve balances in the FRB deposit format could lend them out overnight to banks that were deficient of their regulatory requirement by the barter of checks; the borrowing bank issuing a draft on its house account payable at the close of the next business day in exchange for a check issued by the lending bank drawn on its FRB account payable immediately. The transaction was easy, discreet, and above all self-eliminating.
Since these federal funds transactions were unsecured, they were carried out by banks that were quite well acquainted, likely located in the same areas if not the same cities and towns, sharing common contacts and openness. If we assume Bank A is the member bank and Bank B the non-member across the street, then the federal funds relationship transfers excess reserves from A to B. From a systemic perspective, there is no change as the total quantity of reserves remains exactly the same. The only difference is the microscale in which Bank A is reduced of its excess reserves for one day, transferring them to Bank B so that it, too, could be statutorily satisfied. But what of each banks’ depositors?
The idea of borrowed reserves in the legal sense is straightforward, but it isn’t so in terms of the ultimate moneyholder. Already in the correspondent setup, both banks A and B have reserves one step removed from money, as correspondent balances are claims on some other bank’s claims. In the Federal Reserve System, that is still the case where reserve balances at Reserve Banks stand for what were correspondent balances. From the depositor’s perspective, the member bank substitutes the Reserve System as a pool of potential claims with the Federal Reserve Banks (the 12 district banks that made up the business end of what we now call the central bank Fed); it is still remote, as what has changed is who ultimately might be responsible for delivering funds upon request.
In the federal funds transaction, that still appears to be the case as non-member Bank B has intertwined its federal funds borrowing from Bank A within its whole vault cash and money distribution format as if there is no distinction – despite the fact that the federal funds transaction with Bank A is “borrowed reserves.”
Federal funds did not remain one-off measures to be occasionally dusted off here and there. Instead, the country’s plethora of Bank B’s came to be more and more dependent on Bank A’s for ongoing, even permanent funding needs and over greater lengths of time. Federal funds, though overnight, came to exhibit the now-familiar rollover characteristics. If you went to Bank B expecting to claim money as is your right as a depositor, you were likely unaware that Bank B was allowed to meet yours as well as every other depositor claim to that money by promises from other banks, including the central bank, to get it to you. Your money isn’t in the vault, it is in “the market” for money.
The true money multiplication in this format is therefore not strictly vertical; it is horizontal, as well, which can be an inordinately more difficult entanglement. That was the Great Crash and Great Depression.
Throughout the 1920’s, the level of available vault cash dwindled though outwardly it appeared banks were relatively well-supplied via “borrowed reserves.” The implicit assumption is that if a single bank is short of liquidity, it can more widely appeal to the “market” which is sure to have it. In conceptual terms, it appears as if the national system, with a powerful central bank at its center, had broadened its monetary base so as to functionally provide currency elasticity. No one need doubt a bank’s ability to supply dollars upon convertibility request, for if it didn’t have them the “market” surely did especially as backed by the central bank as a matter of regular policy as well as a last resort.
In response to growing faith in this arrangement, the public chose to hold far less currency on its own, trading them instead for deposit accounts. At the start of 1915, before the grand WWI non-Q QE experiment, but after the Fed had been operational, Federal Reserve member banks had multiplied their monetary holdings into about $8.7 billion in deposits. While there were deposits to consider among the proportion non-member banks still operating at that time, those estimated deposits were supported by $1.62 billion in bank reserves, of which less than $300 million were FRB balances.
In the fifteen years thereafter to the end of 1929, while total bank reserves almost doubled the level of deposit balances more than quadrupled. The fraction of the money multiplier had enormously expanded, and did so to a far, far greater extreme in terms of just money and currency. Thus, the rough estimate of vault cash alone supporting deposits (again, of just member banks) was about $6.40 in deposits for every $1.00 of cash and money. At the end of 1929, every $1 in vault cash was supporting $43.47 in deposits.
The level of vault cash continued to dwindle throughout the 1920’s because all facets of the monetary and banking system had become comfortable with the notion that FRB “reserves” were perfect or nearly perfect substitutes – that the “market” held sufficient dollars by which to absorb any wave of convertibility that might ever develop, as well as the assumption that if in the worst case the broad, national market also needed aid it could depend on the central bank.
What happened after the stock market crash, which, again, was inescapably intertwined with the monetary system via the vast levels of call money and street loans made, was that these assumptions proved to be catastrophically false. The money multiplier when pushed into reverse was hugely disproportional, as while the system built up this higher fractional lending over more than a decade, it was stressed and thus fell apart almost all at once. Milton Friedman and Anna Schwartz put it best in their book:
During the five months when highpowered money rose by $330 million, currency held by the public increased by $720 million. The extra $390 million had to come from bank reserves. Since banks were unwilling and unable to draw down reserves relative to their deposits, the $390 million, amounting to 12 per cent of their total reserves in August 1931, could be freed for currency use only by a multiple contraction of deposits. The multiple worked out to roughly 14, so deposits fell by $5,727 million or by 15 per cent of their level in August 1931. It was the necessity of reducing deposits by $14 in order to make $1 available for the public to hold as currency that made the loss of confidence in banks so disastrous. Here was the famous multiple expansion process of the banking system in vicious reverse. That phenomenon, too, explains how seemingly minor measures had such major effects. The provision of $400 million of additional high-powered money to meet the currency drain without a decline in bank reserves could have prevented a decline of nearly $6 billion in deposits.
The last sentence is the verdict and assumption that has prevailed for the half century since it was written. Had the Fed injected just $400 million in additional FRB balances of bank reserves, the worst of the Great Depression, the utterly absurd monetary contraction, would have been prevented. Yet, what the authors wrote at the start of that passage might have been much more significant, “banks were unwilling and unable to draw down reserves relative to their deposits.” In other words, bank reserves, especially in the format of FRB balances weren’t actually perfect substitutes for cash; they were, as the massive crash proved, very poor substitutes.
Much of the reason for that can be appreciated by unpacking even further our conceptual model of money multiplication from still bound if in a lower degree of aggregation. Above I have drawn FRB balances as itself a monolithic block standing in for all “borrowed reserves.” They are still treated today in the same way, even the language by which we use to describe the process gives off that impression: “pools” of funds. In reality, however, they are no such things, more aptly described as chains of transactions that are intertwined so tightly so as to suggest a solid block upon cursory inspection.
If we think of Bank B trying access funds from Bank A, it’s not at all likely that Bank A actually has funds available in the form of spare vault cash that it could immediately send to Bank B to meet Bank B’s depositor convertibility demand. Yet, that is exactly how the accounting treats these “borrowed reserves”, and even the way in which it is semantically describes sees them as if that were the case. Friedman and Schwartz’s quoted passage above shows very well the wording of this inappropriate reduction (“draw down reserves”).
In operational reality, it was very likely that all the Bank A’s that were funding the Bank B’s were doing so in the legal distinction of “excess reserves”, which they were themselves derived from some other source, either the “market” or the central bank. If the federal funds rate were above the discount rate, for example, as it so often was in the 1920’s, any bank with sufficient collateral could have accessed the Fed’s Discount Window with which to gain a further surplus of bank reserves to then relend into the federal funds market for those Bank B’s who were at a legal deficit to their own requirements. This situation, however, in terms of the perspective of depositor is not at all like how it is conceived of and described; actual money and currency is even further remote from the deposit-holder.
The longer these chains of interbank transactions and funding relationships, the higher the probability of one link in them breaking down; one of the various Bank A’s who could deny an attempt to “draw down” on money market lending, thus causing other Bank A’s to adjust their own supplies, too, setting off a chain reaction to where bank reserves which were supposed to be able to meet depositor claims were instead nothing more than troublesome and unreliable numbers on individual bank account ledgers. If these “borrowed reserves” aren’t actually reserves in a manner from which they could be actually mobilized, then they aren’t actually reserves in the first place from the perspective of either the bank or the depositor.
From this more complicated understanding, we can begin to appreciate why all the Bank B’s and a good many Bank A’s started to instead call back and cancel deposit accounts in order to gain liquidity. That was one of the many complaints expressed at the time of the Crash and even afterward contributing to the regulatory reform that somehow handed the Fed expanded powers. On the branch level, at least, the Fed during the crisis better appreciated these distinctions, with many banks reporting to FRB officials that though there were often large balances of reserves in the aggregate at some reserve and central reserve city banks, they were hoarding them and not lending them out.
There were some $2.3 billion in deposits at FRB banks at the end of 1929, but as these interbank chains broke down that had the effect of concentrating them rather than having them dispersed to where most needed, as was assumed would happen. Further, it amplified the stresses because each bank on the wrong end of these individual breakdowns would react more violently to them by increasing its own margin of liquidity, including calling in loans, cancelling deposits, and redeeming notes. The quantity of reserves didn’t contract, at least not right away, but functional money did.
Orhtodox economics largely as a product of Milton Friedman’s approach decided that the Fed’s actions during the Great Crash were to blame; that if it had supplied just $400 million more, history would have been far different. From a more nuanced view of “borrowed reserves”, we have very good reason to suspect that $400 million wouldn’t have done much more than sit idle further concentrated on the ledgers of only a few banks.
That is, after all, what should be most striking about this review; how familiar it is and sounds to us in light of the events of 2008. The behavior and even the specific conceptual breakdowns were the same in theory; what was different about them was only the means and methods by which they occurred.
Part 3 is here.