January 2004 marks the 10th anniversary of retail deposit sweep programs at U.S. banks. Some regulators initially were skeptical about the effects of retail sweep programs. This anniversary seems an appropriate time to re-examine their effects.
Retail deposit sweep programs increase bank earnings by reducing the amount of noninterest bearing deposits that banks hold at Federal Reserve banks. A bank's transaction deposits beyond approximately the first $50 million are subject to a 10 percent reserve requirement ratio, which is satisfied by holding vault cash or noninterest-bearing deposits at Federal Reserve banks. In contrast, savings deposits are subject to a zero percent ratio.
Retail deposit sweep programs take advantage of this difference by "sweeping" transaction deposits into savings deposits—that is, relabeling transaction deposits as savings deposits for reserve-requirement purposes. Currently, banks are limited by law to no more than six transfers per month between a customer's transaction deposit account and this type of savings account. (Withdrawals, including checks, are paid from the transaction deposit. When the transaction deposit is depleted, the computer must move funds from the savings deposit back to the original transaction account.)
Sweep programs allow banks to increase their earnings by redeploying excess Federal Reserve deposits into earning assets. In many cases, the bank is able to reduce its reservable transaction deposits by 70 percent or more. Some analysts have regarded the increase of retail deposit sweep programs as the end of binding statutory reserve requirements, because many banks find they are able to reduce their required reserves below the amount of vault cash necessary for day-to-day business.
During the early years of retail deposit sweep programs, some analysts were concerned that federal funds rate volatility would increase as banks held fewer deposits at the Federal Reserve. Some recalled the winter of 1991, when higher volatility followed the December 1990 cut in reserve requirements. The spread of retail sweep programs among banks, however, has not increased the volatility of the federal funds rates. This experience demonstrates that—at least for the United States—banks and the clearing system can operate effectively with low levels of deposits at Federal Reserve banks.
This "win-win" experience with retail deposit sweep programs—higher bank earnings without increased federal funds rate volatility—has led some members of Congress to propose relaxing regulatory constraints on retail deposit sweeping. Proposed legislation would increase that limit to 24 transfers per month, more than one for each business day. Such a change would be economically equivalent to reducing the reserve-requirement ratio to zero for banks with sweep programs—effectively, the end of binding statutory reserve requirements in the United States.
For financial market analysts, retail deposit sweep programs have distorted available data. Larger banks report to the Federal Reserve their daily close-of-business amounts of transaction and saving deposits; however, they do not report the amounts of deposits involved in retail deposit sweeps. Hence, the amount of transaction deposits swept into savings is excluded from published figures on M1.
Today, this amount is approximately half of all transaction deposits held by households and firms, compared with 10 years ago. For example, the Federal Reserve's Flow of Funds accounts showed households held currency and transaction deposits of $615 billion in 1994; in 2003, those deposits were down to $322 billion. The reason for the drop is clear: The household sector is calculated as a residual from the deposit figures reported by depository institutions to the Board for reserve-requirement purposes, and half of all transaction deposits are being swept into savings deposits.
The widespread availability of retail deposit sweep software now makes binding statutory reserve requirements a voluntary constraint for most banks. Fears of an increase in federal funds rate volatility have proven unfounded. Perhaps the most serious remaining issue is distortion to published M1 and other aggregate measures of financial intermediation, a small issue at best.