When the Federal Reserve began paying interest on bank reserves on October 9, the justification in its press release was to permit the Fed to hit its Federal funds target interest rate more reliably. But the full explanation is more complicated than that.

The Fed also serves as a clearinghouse for banks, and that function is in tension with monetary policy. When the Fed was first created in 1914, it provided clearing services to all member banks for free, driving out of business the various private clearinghouses that had arisen and were solving some of the liquidity problems associated with the destabilizing National Banking System. Then in 1980, the Depository Institutions Deregulation and Monetary Control Act required the Fed to offer its clearing services to all depository institutions--whether banks or not, and whether members of the Fed or not--but at a fee that allowed the reemergence of private alternatives.

There are two ways to run a clearinghouse. The first, net settlement, involves waiting until the end of the settlement period (traditionally a day) and then transferring only net amounts. The second, a real-time gross settlement (RTGS) system, settles each payment as it occurs. The private Clearing House Interbank Payments System (or CHIPS, created in 1970), which now handles more than one-quarter of bank clearings in the U.S., netted all settlements at the end of the business day until 2001, when it switched to intraday payments. The Fed's current system, Fedwire, in contrast, has always conducted real-time settlements. Because banks may therefore lose all their reserves to other banks before any offsetting receipts come in, the Fed provides banks with intraday overdrafts to assist with their clearings.

Before paying interest on reserves, the value of these overdrafts was climbing to an amount exceeding bank reserves. Thus since 1987, U.S. banks reserves (counting vault cash) have hovered around $65 billion, but the average daylight overdrafts outstanding at any minute during the day rose from around $15 billion to nearly $50 billion. The peak value of daylight overdrafts at any moment could rise even higher, to over $100 billion. Alex Tabarrok over at Marginal Revolution provides a colorfully apt description of this process:"in essence, the banks used to inhale credit during the day--puffing up like a bullfrog--only to exhale at night. (But note that our stats on the monetary base only measured the bullfrog at night.)"

A clearing system using net settlement leaves the potential losses from a bank's failure to pay on the bank(s) owed money. But Fedwire's RTGS system transfers that risk to the Fed itself. The Fed has consequently tried to limit bank use of daylight overdrafts, first imposing net debit caps in 1985 and then imposing minute-by-minute interest charges in 1994. Once the Fed began paying interest on reserves, the banks had an incentive to substitute excess reserves for now more costly Fed credit. A technical article that models this change, Ennis and Weinberg, reveals that the Fed expected and hoped for this result in order to reduce the Fedwire risk from daylight overdrafts.

Being able to earn on interest on reserves would have caused the banks to hold more of them anyway, but the fee on daylight overdrafts only augments that effect. Much of the Fed's recent, more than ten-fold increase in bank reserves, to a current total of about $675 billion, has so far caused banks primarily to increase their reserve ratios rather than expand their loans. Some, like Paul Krugman, have interpreted this as a huge, deflationary flight to liquidity. Others, like myself, contend that it is only a matter of time before this leads to inflation. But as Alex suggests, the payment of interest on reserves could render both predictions wrong. The increase in reserve ratios could simply be a one-shot response of banks, hiking ratios to a new level. No one knows for sure, including I would add, the Fed itself.

The European Central Bank (ECB), like the Fed, has been pumping up its monetary base with wild abandon. And as in the U.S., private European banks seem to be increasing their reserve ratios, at least in the short term. But the ECB has paid interest on reserves since introduction of the Euro in 1999. In fact, because interest on reserves reduces central bank seigniorage, confining it to currency alone, this feature may have eased the negotiations over the inevitable conflicts in creating a European monetary union. The ECB also oversees a real-time clearing system with intraday credit, known as TARGET (Trans-European Automated Real-Time Gross Settlement Express Transfer).

So again, the accumulation of excess reserves may reflect the perverse impact of central banks paying interest on them. The Fed started doing this, confident from the experience of the ECB, and other central banks, such as those of Canada, New Zealand, and Australia, that have been doing so for some time. But the policy had never been tested in a period of falling interest rates, rising risk premiums, and rising preferences for shorter maturities.

I predict that future economic historians will look back on this change as a major blunder during the current credit tightening, making traditional monetary policy less effective. True, the broader monetary aggregates are already beginning to respond to the Fed's base explosion, with M1 annual growth up from 0 to 20 percent over the last three months, and M2 annual growth up from 5 to 10 percent over the same period. Yet irrespective of whether the long run brings deflation, inflation, or neither, paying interest on reserves has certainly applied deflationary pressure in the short run. It may eventually rank with the Fed's doubling of reserve requirements in the 1930s and bringing on the recession of 1937 within the midst of the Great Depression.

Moreover, the paying of interest on reserves was motivated by the misguided focus on interest rates, rather than money supply measures, as an indicator and target of monetary policy, a focus that has dominated central bank operations worldwide for the last two decades. It is also a focus that seems sadly to have taken in many libertarian and free-market economists. Although done in the name of controlling inflation, this focus actually reflects a move toward centralized economic planning on the part of central banks, given that the interest rate is a relative price, with a significant real as well as a nominal component, compared with such purely nominal targets as the money supply or the price level.

The irony is that the Fed is now less able to hit its interest rate target than ever before. It first adopted the corridor or channel system of the ECB, setting the interest rate on reserves below its Federal funds target, as a lower bound, with the discount rate above the target as an upper bound. But as the effective Federal funds rate fell not only below target but below the interest rate on reserves, the Fed on November 5 moved to the New Zealand system, where the interest rate on both required and excess reserves is set right at the target Federal funds rate. So far, this hasn't worked either. (For accessible descriptions of these two systems, see an article by Keister, Martin, and McAndrews.)

Why does the effective Federal funds rate remain below the Fed's target rate of 1 percent, despite the fact that the Fed is now paying interest on both required reserves and excess reserves at that target rate? The best explanation for the anomaly has been offered by Jim Hamilton. Fannie, Freddie, the Federal Home Loan Banks, and other GSEs, plus some international institutions have deposits at the Fed, and these do not earn interest. These institutions are also players in the Fed funds market. So their Fed funds loans would not be affected by the interest paid on excess reserves.