Economists and financial market analysts are wrestling with the tough job of getting a handle on the repurchase market without comprehensive data, and some are calling on the Federal Reserve to reinstate its publication of the M3 measure of the money supply.
The M3 measurement included the broadest selection of repurchase agreement data collected. The Federal Reserve discontinued publication of M3 in 2006, saying it was too erratic to be useful – just when it would have been a big help in spotting the housing bubble.
The housing bubble was fed by securitization, which was largely financed by the repurchase market. Comprehensive data on repo volume could have revealed the explosion of leverage in both the traditional and shadow banking systems 2000-2008.
Without M3, the best data on repo volume probably comes from the primary dealers who represent about half the market, the tri-party system which represents about one-fourth of the market, the Depository Trust & Clearing Corp. which tracks repos collateralized by U.S. government securities, and the European Repo Council which publishes an annual survey of the European repo market. (Update: In June 2012 New York Fed analysts said primary dealers represented about 90 percent of the U.S. repo market.)
The Federal Reserve began publishing weekly and monthly money supply reports in 1971:
-M1 includes currency, traveler’s checks, deposits at commercial banks, credit unions and thrifts, negotiable order of withdrawal and automatic transfer service accounts at banks,
-M2 includes M1 plus savings deposits, certificates of deposit smaller than $100,000, Individual Retirement Account and Keogh balances at banks, and retail money market mutual funds.
-M3 included M2 plus repurchase agreements longer than one day at commercial banks, large time deposits, and institutional money market accounts.