The story is that the Fed’s quantitative easing program injected large amounts of liquidity into financial markets, causing bond rates to fall and stock prices to accelerate. Consequently, the argument goes, the shrinking of the Fed’s balance sheet through maturity runoff will cause bond rates to increase and, presumably, stock prices to retreat. But what are the essential mechanics of Federal Reserve asset purchases and how might they affect liquidity in the market?
When the Federal Reserve began its quantitative easing program, it purchased Treasury obligations in the marketplace through the primary dealer facility and paid for those securities by writing up the reserve accounts of the sellers’ banks, simultaneously increasing the sellers’ bank deposits.[1] Effectively, the Fed created money in the purchase transactions, but as far as the public’s asset position is concerned, the purchases substituted demand liabilities for Treasury obligations. The sellers received deposits; their banks’ reserves increased by the same amount and the sellers’ Treasury holdings were reduced.
From the perspective of the consolidated government balance, Treasuries were removed from the public and on-demand Fed liabilities were substituted in their place, bearing a lower interest cost than the Treasuries they replaced.[2] One form of a very liquid asset (Treasury securities) was replaced by another (reserve deposits at the Fed, with corresponding deposits held by the public in its bank). The Fed’s purchasing Treasuries bid up bond prices and put downward pressure on interest rates.
One of the main effects of QE was to redistribute the ownership both of Treasuries and of bank reserves and their associated deposits. We can’t quantify or identify the sellers of securities, but we do know that a large portion of the excess reserves associated with those purchases ended up with US affiliates and subsidiaries of foreign banks. Presumably sellers were institutional investors, hedge funds, and money market mutual funds but could also include individuals.[3]
Foreign banks’ share of reserves peaked at about 50% in the fall of 2014 and is presently about 35%. Those excess reserves in US and foreign subsidiaries were potentially available to generate a large increase in bank loans and the money supply. A dollar of excess reserves would support an estimated $20 increase in credit and the money supply if it were converted to required reserves as part of the bank credit creation process.[4] But this obviously didn’t happen. Indeed, the ratio of bank loans and leases to bank reserves in Oct 2008 was 26.0, whereas that same ratio as of October 31, 2018, was only 5.6; so bank credit did not expand nearly to the same degree that bank reserves expanded. Interestingly, despite the series of QE experiments, in September 2014 a dollar of reserves was associated with only 2.9 dollars of bank loans and leases right before the Fed stopped adding to its portfolio in October 2014.[5]
In short, the degree of stimulus, as far as bank lending was concerned, was muted. In fairness, business investment demand for credit was not great. The NFIB (National Federation of Independent Businesses) reported in March 2014 that 53% of its respondents indicated no need for a loan. Only 2% reported that financing was a major problem; and only 30% reported borrowing on a regular basis, a near-record low. One of the reasons was pessimism about investment and expansion prospects. The report stated that “The small business sector remains in maintenance mode, no expansion beyond a few firm starts in response to regional population growth.”[6]
In December 2016 the Fed began a series of 25 bp increases in its target rate for federal funds, and in October of 2017 it began the process of shrinking its balance sheet by letting assets mature and run off naturally. As of December 19, 2018, the Fed’s balance sheet stood at $4.084 trillion, down from its peak of $4.5 trillion on October 14, 2015. Critics have complained that the balance sheet shrinkage process has contributed to a liquidity shortage; but they have not defined exactly what the nature of liquidity problem is, who is or is not constrained, and how that constraint is manifested.
The size of the Fed’s balance sheet is determined by the outstanding reserve balances (both required and excess reserves), the volume of currency in circulation (which is of course the most liquid of assets), the volume of funds in the Treasury’s account with the Fed, the volume of reverse repo transactions outstanding, deposits in foreign official accounts, and of course capital. The only way the Fed’s balance sheet can shrink in size is if outstanding currency declines, bank loans shrink, Treasury or other official balances decline, or assets are allowed to mature, in which case the Fed’s liability to the Treasury declines, offsetting the maturing assets.
Several factors have actually put upward pressure on the size of the Fed’s balance sheet since October 2014, including an increase of $330 billion in Treasury balances, an increase of $314 billion in added currency outstanding, and an increase of $82 million in foreign official and other deposits. This increase was offset by a decline of $1.07 trillion in bank reserves.
How can we explain the drop in reserves if other factors seem to be pointing to an increase in the balance sheet? The source of Treasury balances is tax revenues that are deposited in Treasury tax and loan accounts at commercial banks. When the Treasury transfers funds from those accounts, the reserve accounts at the affected commercial banks are drawn down. So, in fact, the increase in the Fed’s liability to the Treasury is offset by a decrease in the reserves of the tax and loan account banks. Similarly, when bank customers withdraw funds in the form of currency, currency demand increases. That currency is obtained from the Fed in exchange for a reduction in banks’ reserve accounts. So again, rather than actually increasing the size of the Fed’s balance sheet, the composition of its liabilities is changed – currency outstanding is increased, and bank reserves are decreased. Of the $1.07 trillion decline in bank reserves, there was a corresponding increase in Federal Reserve liabilities to the Treasury and the increase in currency outstanding together accounted for $653 billion of the decline. The remainder is largely associated with the runoff and shrinkage of the Fed’s asset holdings.[7]
It is important to note that when the Fed engages in what it calls reverse repo transactions, the securities sold remain on the Fed’s balance sheet. Bank reserves are temporarily reduced, but corresponding liabilities to banks under the account “reverse repurchase agreements” are increased. The composition of Fed liabilities changes but the volume does not. When the repo transaction is reversed, bank reserves go up and “reverse repurchase agreements” are reduced.
The bottom line is that the apparent decline in bank reserves, far in excess of the change in the decline of the Federal Reserve’s balance sheet, is offset by changes in the other factors absorbing reserve funds, which simply represent a reallocation of the ownership of Federal Reserve liabilities. In the case of the Treasury, it accumulates funds to spend on entitlements, purchases, salaries, etc., which when paid reduce Treasury balances but reappear as an offsetting increase in bank reserves when the funds are deposited with the banking system.
As for the notion that the Fed’s reducing its balance sheet holdings of Treasuries contributes to a so-called liquidity problem, again the mechanics are not clear, especially when we consider what has happened to Treasury debt issuance. To be sure, the Fed’s portfolio of Treasuries fell by $213 billion since the decision to let maturing issues run off, while MBS holdings declined by $132 billion.[8] Treasury debt held by the public increased by $956 billion through the end of the third quarter of 2018 as the Fed’s portfolio began to run off. But the actual net issuance of Treasury debt is even greater than that because the Fed’s portfolio is treated from an accounting perspective as part of the public’s ownership of the debt. Since the Fed’s ownership declined by $213 billion, the Treasury securities owned by the public, not including the Fed, increased by $1.169 trillion. This issuance dwarfs the rundown in the Fed’s portfolio and its potential impacts on securities markets. The decrease in the Fed’s marginal demand for Treasuries is far offset by the increase in supply. That supply, depending upon the maturity structure of the Treasury’s refunding, puts downward pressure on rates across the Treasury curve relative to the impact that the FOMC’s rate increases have had on short-term rates. This issuance pattern probably is the major explanation for the overall upward shift in the yield curve that we have experienced since the Fed began letting its portfolio run off.
In the meanwhile, more liquid assets are now in the marketplace as a result of the increase in currency outstanding and the increased supply of outstanding Treasuries, and banks still have a huge volume of liquid reserves. Note that, like cash and bank reserves, Treasuries satisfy the banking regulatory agencies’ liquidity requirements, so it isn’t clear what the nature of the claimed liquidity problem is or who is experiencing problems.[9]
Liquid assets are supposedly those that can be sold with little or no impact on their price. But we must be mindful that in order for an asset other than cash or deposits at the Fed to be liquid, there must be a buyer on the other side. If there is no buyer, then assets that were thought to be liquid suddenly are not. Indeed, the Fed in essence became the buyer-of-last-resort during the financial crisis. If Yogi Berra were asked to define liquidity, he might have said the following: “Liquidity is what you have when you don’t need it; but when you need it, you don’t have it.”
[1] The following discussion of asset purchases and sales omits much of the institutional detail and mechanics behind the transactions and focuses instead on the key results.
[2] We have noted before that the Treasury pays the Fed interest on its Treasury holdings, and the Fed pays interest on reserves out of those proceeds (as well as covering its other operating costs) and remits the remainder back to the Treasury. The effect is that the Treasury’s financing cost on the Fed’s Treasury portfolio is the cost of interest on reserves and not the interest payments on the Treasuries themselves.
[3] Foreigners and foreign institutions own about 50% of the outstanding debt held by the public.
[4] Author’s estimates
[5] Source: FRED FRB St Louis
[6] See http://www.nfib.com/Portals/0/PDF/sbet/sbet201403.pdf.
[7] It is important to note that when the Fed engages in what it calls its reverse repo transactions, the securities sold remain on the Fed’s balance sheet, and bank reserves are temporarily reduced, but liabilities to banks under reverse repos are increased. The composition of Fed liabilities changes, but the volume does not. When the repo transaction is reversed, bank reserves go up.
[8] Data from Oct 2017 through December 19, 2018
[9]If Secretary Mnuchin understood banking, the last institutions he would have called to inquire about liquidity problems would have been the nation’s largest banks, which hold the bulk of the excess reserves and have ample liquidity.