In September 2008, the Federal Reserve initiated a series of quantitative easing (QE) programs that dramatically transformed the Fed's balance sheet--in size, liability mix, and asset mix. The "exit strategy" questions now facing the Fed, and the dollar-using public who are its captive customers, are when and how to reverse those transformations.
On the liability side of the Fed's balance sheet, QE swelled the stock of base money (the subset of the Fed's liabilities consisting of currency held by the nonbank public plus depository institutions' reserves) more than four-fold. Contrasting October 2015 to August 2008, the base rose to $4.06 trillion from $0.85 trillion. The mix of Fed liabilities shifted as approximately $2.6 trillion of the $3.2 trillion in new base money was added to the reserve balances of depository institutions (the other $0.6 trillion was added to currency held by the public). Total bank reserves have grown more than 50-fold, to $2.7 trillion from a mere $0.05 trillion. Only a tiny share of the added reserve holdings (about $0.1 trillion) are accounted for by the growth in required reserves accompanying growth in commercial bank deposits held by the public; the bulk are voluntarily held as excess reserves (balances over and above legally required reserves against deposits). Excess reserves have risen to $2.5 trillion and 62 percent of the monetary base, from only $0.002 trillion and close to zero percent (about two-tenths of 1 percent) pre-QE. (1)
[FIGURE 1 OMITTED]
While the QE programs accelerated the monetary base (hereafter MO) at an unprecedented rate, Figure 1 shows that they did not accelerate the quantity of money held by the public as measured by the standard broad-money aggregate M2 (currency in circulation plus all bank deposits). During the pre-QE decade of September 1998-September 2008, the Fed expanded M0 at a compound rate of 5.99 percent per annum. The expansion rate jumped to 23.69 percent per annum during September 2008-September 2015. The growth rate of M2 has fluctuated a bit but hardly changed over the longer term: 6.3 percent per annum in the pre-QE decade and 6.6 percent since the beginning of QE. The fact that M2 has hardly budged from its established long-term path indicates that quantitative easing was not a change in monetary policy, in the sense that it was not used to alter the path of the standard broad monetary aggregate in a sustained way. (2)
The growth rate of the M1 component of M2 (currency plus only checking deposits) did rise faster, to 11.5 percent per annum from 3.0 percent. But because M2 as a whole did not grow faster, this only indicates that households have reduced the share of their total bank deposits in savings (non-Ml) accounts and increased the share in checking accounts. This shift can be explained primarily by households responding to a collapse in the spread between savings and checking account interest rates, both rates hilling to near zero. The national average rate on three-month CDs, for example, tumbled to 16 basis points in September 2015 from 359 in September 2008, while the rate on interest checking declined far less, to 4 basis points from 20. (3)
Why didn't M2 grow faster? As money-and-banking textbooks tell us, the growth rate of M2 mirrors the growth of M0 when the commercial banking system sheds excess reserves by banks making loans and securities purchases such that system deposit liabilities grow in proportion to system reserves. After September 2008, however, banks began sitting on the additional reserves the Fed was creating. They did so largely because the Fed almost simultaneously--and not coincidentally--began paying interest on reserves in early October 2008. With a higher reward for holding reserves, banks began holding greater reserves in excess of legal requirements, which meant that the system began creating fewer deposit dollars per reserve dollar. The ratio of excess reserves to deposits rose from a fraction of 1 percent in September 2008, before QE began, to 24 percent today. This enabled M2 to continue along its pre-QE path despite the huge increase in M0.
The initiation of interest on excess reserves (hereafter IOER) and QE at the same time was no accident. The Fed chose to start paying IOER in order to neutralize the flood of excess reserves that QE1 and other Fed lending programs were creating. Fed spokesmen have at times described the rationale for initiating IOER as a move to counteract downward pressure on the federal funds rate (the overnight interest rate at which banks lend reserves to one another) from excess reserves (Dr. Econ 2013). Instead of trying to get rid of excess reserves by lending them and in the process driving the fed funds rate too low, banks would now be happy to hold the reserves.
This is a curious account given that the fed funds rate fell to near-zero anyway. A better explanation begins by noting that IOER, by getting banks to hold more reserves, has allowed the Fed to greatly expand its assets and consequently MO, while keeping M2 from ballooning. The combination of QE with IOER enables the Fed to finance a hugely expanded portfolio of assets without inflationary consequences, essentially by borrowing from the banking system. Without IOER, purchasing assets...