The Global Monetary System Is Not As a Backwards Looking Fed Sees It
The origins of the Federal Open Market Committee make for quite a (relevant) story. By accident, the twelve branches of the early Federal Reserve System discovered that what they did for their own account directly influenced bank behavior. As such it became an important part, and then the only part, of monetary policy.
The Federal Reserve Act required, among other things, that the new central bank not burden the American taxpayers. The expenses for all its operations must be covered by internal means, no subsidy from the Treasury allowed (to the Treasury is another story, of course). Because of this, and because nothing in this world is free, even actual money printing, the early Fed was made to hold some quantity of assets to gain interest income on its own account.
The branches already did so as a consequence of several of its basic duties, including what was thought the main policy action in rediscounting. What changed was the securities market as a consequence of World War I. The reserve banks had been buying up independently more and more Treasury debt to finance their operations, if largely because government debt increasingly became all that was available. In the eight months leading up to and including May 1922, collectively they bought $400 million in so-called earning assets, largely UST’s, upon only the consideration of obtaining income.
The effect on the money markets was obvious. In response, in May 1922 an ad hoc committee of officers from the Eastern branch reserve banks was formed to try and coordinate purchases to minimize market disturbances.
That wasn’t their only goal, however, established by market reaction to this coordination in 1923. The more formal Open Market Investment Committee for the Federal Reserve System replaced the prior format, and throughout that year it ran what was described in the Fed’s Tenth Annual Report a test on the effectiveness of these operations as a policy matter.
These early central bankers discovered what became known as the scissors effect. Essentially, when the Fed purchases securities (almost all either UST’s or bankers’ acceptances) from the open market, the member banks from whom the purchases are made tended to rediscount in lower volumes as a result; that is, they would borrow less from the Fed on collateral (rediscounted as a haircut of sorts). When the reserve branches sold securities into the open market, the member banks tended to rediscount more to borrow back the cash (reserves) used to pay for them.
The clear “scissors” was shown in these differences in 1923. The level of total earning assets, and thus the total volume of reserve credit, remained constant throughout. The Fed began the year combined with equal proportions of rediscounted bills and open market assets (of those, about two-thirds UST’s, one-third acceptances).
Through especially the middle part of 1923, the twelve branches in more coordinated fashion drew down their open market holdings (sold UST’s). As a result, member banks added to their volumes of rediscounted paper (borrowing from the Fed). Despite large changes in open market operations, the total level of earning assets (reserve credit) remained almost perfectly constant as lower open market operations were nearly perfectly offset by borrowing from the reserve banks.
What had happened was that the member banks shifted from reserve balances held and then drawn down by the Fed’s sale of open market assets to reserve balances borrowed at the Discount Window. It is the familiar textbook account of Open Market operations in the modern era; the Fed buys assets and increases the volume of reserves affecting overall money conditions in predictable fashion; or it sells assets into the market and removes them.
But that isn’t quite what the Tenth Annual report describes. There is another part to the equation:
“The difference between discount operations and open-market operations is that the initiative in rediscounting lies with the member banks, while in the purchase and sale of securities the initiative may be taken by the reserve banks.”
That makes perfect sense; discount operations are the effects of market agents acting upon market forces and using the Fed as one money alternative to meet those market demands. Open market operations, by contrast, are the Fed intruding itself into the market and forcing the market to act in the way it wants (believing that intrusion to produce predictable effects).
“The extent to which member banks borrow in order to replace the funds withdrawn by the reserve banks through the sale of securities is a measure of the demand for reserve bank credit. The sale of securities by a reserve bank may thus serve as a test of the degree of adjustment between the demand for reserve bank credit and the outstanding volume of such credit.”
In 1923, obviously, this was true. The Fed sold bonds from its twelve portfolios, the result of which was the higher demand for borrowedreserve credit in order to maintain a constant volume of reserve credit throughout. Combined with changes in the Discount Rate as a matter of further policy, the Fed could then influence all sides to the money market.
Thus, if the Fed wanted to really tighten reserve credit (base money), and therefore total bank credit, it could sell down its holdings of assets while at the same time raising the rate at which banks would be expected to borrow more at the Discount Window. This squeeze or scissor should sound familiar given the FOMC statement issued earlier this month making plain the Fed’s intentions to do just that.
But not quite that. There are some important differences to consider. After all, it is 2017 not 1923.
For one, nobody uses the Discount Window anymore. Even during the height of the 2008 crisis, volume in Primary Credit (the Discount Window was totally rearranged in 2002) was never anything more than immaterial. Banks were well aware of the stigma associated with accessing it; they knew well that going to the Fed and borrowing reserves meant painting a bright red target on their liquidity needs.
Strange as it may seem, the very fact that there could have arisen any dishonor from using the Discount Window did first come about as a consequence of the Federal Reserve operations in the twenties. Borrowed reserves were a huge part of the monetary expansion of that decade as a regular and direct funding mechanism. Without acknowledging explicitly its role in that expansion, the Fed during the Great Depression of the thirties began to discourage their use – insofar as the Discount Window was concerned.
If the private money market would no longer regularly borrow reserves from the Fed, it would borrow them instead from itself. And if it could borrow them from itself, it could also dictate the terms and really the form in which it would accept liabilities to satisfy these liquidity obligations. The federal funds market, for one, arose in the fifties as a matter of monetary expansion in the same way as the twenties, only substituting simple interbank borrowed funds for Discount Window borrowed funds.
As a matter of monetary policy, as modest as it was this shift was still enormous and as usual misunderstood and unappreciated by monetary officials (it took them six years to conduct a study on the federal funds market and then put together enough information for what became the federal funds effective rate). Not only that, the way in which banks borrowed reserves changed, too. There were federal funds in sufficient volume, and then throughout the sixties and seventies there was also the burgeoning market for repurchase agreements. Besides those, there was a fair amount of currency swaps, often offshore, and often with foreign central bank counterparties effectively cheating on Bretton Woods.
The entire monetary system transformed, and yet monetary policy remained wedded to an outdated model. That is one reason why throughout the Great Inflation the Fed changed its mind several times about how to go about carrying out monetary policy – and failing each time until the last after a decade and a half of great social as well as economic costs.
As Milton Friedman later said in 1999 speaking to Forbes Magazine:
“We’re in a period like the 1960s, when no one paid any attention to the money supply. Then we got inflation. By the 1980s everyone was obsessed with the money supply. Now they’re forgetting again. And it will turn around and surprise them.”
That assessment has proved to be totally correct; just not at all in the way Friedman envisioned. The father of monetarism saw that the world was finally listening to him particularly about Japan. He was clear on the malady there, where low interest rates in Japan indicated tight money, not loose or stimulative conditions. This interest rate fallacy is dispositive, as witnessed during both the Great Inflation (where rates were high, not low) and the Great Depression (where rates were low, not high).
Friedman’s recommendation to the Bank of Japan was to stop focusing on interest rates (then achieving ZIRP for the first time) and start focusing on money supply. Or, as he told Forbes that time, “The mechanism of recovery is monetary expansion plus reform of their banking system. In the short run, nothing other than monetary expansion [will work].”
He believed that by increasing the level of bank reserves that would have the effect of increasing active money in the real economy. QE was the result of this idea, and the Bank of Japan did get to it a few years after the Forbes interview. If it had been successful, Friedman said in 1999, it could mean higher global interest rates as a result of Japan’s recovery. In other words, he was expecting more inflation in the 21st century rather than a continuation of the later-named Great “Moderation.”
His error was more basic. Friedman contended that Economists and central bankers had stopped paying attention to the money supply when in fact they really stopped paying attention to money. This is a profound difference, and it goes back to what the Tenth Annual Report was really getting at in 1923.
You can’t have scissors with one blade missing. After the Great Inflation, the Federal Reserve carried out monetary policy in almost circular fashion, focusing only on the one end. In other words, they did very little in terms of Open Market operations, instead preferring an interest rate (federal funds) target; a target that came about because of the threat of Open Market operations which the Fed never did.
In general terms, the federal funds target was meant to accomplish both parts. It was widely believed that the quantity of bank reserves, especially those borrowed not from the Discount Window but from the private dollar markets in federal funds and so much more, could be fully influenced by the rate; if not, as some came to believe, completely directed by it.
Was this ever really true? That’s the question. It seems on common sense grounds that meeting private demand for bank reserves only through their borrowing and lending and only within the private market makes a big difference in system operation; capacity as well as control. Furthermore, the methods for meeting reserve demand could at times be paramount, more important even than the quantity of those borrowed reserves (however borrowed).
We have witnessed the rise of repo as one of those devices, placing great importance on collateral given that one aspect of unchecked repo growth is its fluidity (rehypothecation primary among collateral dimensions).
But that isn’t the only “new” way in which banks have satisfied the function of reserves. In fact, the very idea of reserves is what has changed. It is a word that no longer truly applies, at least not as it once did. It has a technical definition, of course, but that definition actually highlights all the ways in which the Fed has failed to keep up.
A more apt description would be something like liquidity, a more comprehensive term that incorporates bank reserves as one element within it. A bank may borrow in federal funds, repo, or FX from another bank. Each requires several components that in most cases have less to do with bank reserves the more esoteric each form. And, as noted last week in the BIS’s belated “discovery” of $13 or $14 trillion in offshore, footnote debt used primarily in these sorts of interbank channels, there is more often now almost no direct relationship with bank reserves whatsoever.
What the Fed actually did was (much) smaller than any official ever thought possible; they lowered the rate on borrowed federal funds and increased the level of bank reserves as a byproduct of QE thinking they were following closely the old textbook. In the end, it wasn’t the same at all as raising global dollar liquidity. We know this unambiguously by the results; a panic in 2008; a secondary shock in 2011; and then one more in 2014-16.
Go back to what the Federal Reserve wrote in 1923, changing “sales” of securities for “purchases” in QE. It did indeed serve as “a test of the degree of adjustment between the demand for reserve bank credit and the outstanding volume of such credit.” The test showed unequivocally that the demand for reserve credit in its more comprehensive global liquidity formats was no longer related to the outstanding volume of strictly reserves.
It’s almost a tautology in its obviousness; if the private market depends onvarious forms of borrowed “reserves” to carry out financial and economic business, it would therefore require the full range of these various borrowed “reserves” to fix any problems within it. The Fed offered only a single form, and one that was more than outdated.
The rest is really quite simple. If there weren’t any positive effects (other than psychology largely based on 1923) due to central bank balance sheet expansion and cutting the federal funds rate, why should we expect the opposite processes to have any?
The Fed may indeed raise the federal funds rate and even selloff rather than just runoff its balance sheet. Many people are expecting that this will make a big difference. It won’t.
There aren’t any good analogies to break this down in simple terms, but I’ll attempt one anyway. Imagine the Fed as the only filling station in town in 1923, and that it discovered it could affect how much you drive by changing the quantity of gas it sells as well as its price (because it makes its own gasoline at negligible cost, the quantity and price are not related except by the sales policy).
If it restricted how much it would sell while also raising the price, if you really wanted to buy it because you really needed to drive somewhere you would go through the trouble of waiting for it to be available and then pay the higher rate anyway (validating the test for demand).
What happens, however, when a private business opens a similar station in the same town offering much the same product at often comparable enough terms? The Fed’s ability to affect your driving habits becomes more remote; they might have some, but it isn’t clear how much or even how because you can get largely the same gasoline from elsewhere. Thus, the terms of those private transactions would be dictated only partially by what the Fed did in restricting its own supply and at what price.
Now imagine that several other filling stations open and offer much better quality gasoline, the real high octane stuff at radically different formulations and then sold at cheap prices. Most people in town will inevitably switch to that over time because it fits better the vehicles they really want to drive.
Then, out of nowhere, those other stations suddenly can’t or won’t sell their fuel anymore. The Fed in seeing this genuinely and even desperately offers to help by manufacturing huge quantities of the same gasoline it has always sold and selling it for as cheaply as possible. A few people may be able to take advantage of the offer, but most of the rest of the town really can’t because their modern vehicles are made to run on the other stuff. They can get away with the occasional fillup at the Fed, but they can’t do it all the time nor is it cost-effective or realistic to go back to driving the older vehicles that could.
Given this situation, would it matter if the Fed’s gas station after being so unused for so long decided to cut back on the quantity of its brand available for sale, raising the price slightly while doing so? No, it wouldn’t. The town’s drivers in the aggregate would be stuck largely in the same driving-less position regardless.
The issue is borrowed reserves; it has always been borrowed reserves. The Fed believes still today that they haven’t changed much since 1923, and therefore its policies didn’t need to be changed much, either. QE sounds so very 21st century, but it isn’t; it is so very 1920’s.
The way in which borrowed reserves function as global liquidity is in many ways nothing like the way in which borrowed reserves functioned back then. To repeat what I wrote above, if borrowed reserves were left as strictly a private matter then the private market could be left to decide the terms and really the form in which it would accept liabilities to satisfy these liquidity obligations. That is what happened, slowly, over the decades.
The analogy above isn’t a perfect one, but perhaps one anyway that is easy enough to understand for all the complex monetary evolution that makes it apt. The global monetary system hasn’t run on simple monetary gasoline in half a century and more. The quantity and price at which it sells today simply doesn’t matter all that much, if at all.