Found: $840 billion in municipal bonds
The Federal Reserve has quietly admitted they had undercounted about $840 billion of municipal bonds. Bloomberg reports on this new pile of assets:
The U.S. municipal-bond market is 28 percent larger than reported in June, according to a quarterly Federal Reserve release, which used new data showing individuals own more state and local-government debt.
“The estimate of household holdings of municipal securities and loans is revised up by about $840 billion, on average, from 2004 forward,” according to the Fed’s Flow of Funds Accounts report for the third quarter.
The Federal Reserve may not have intended to bury the news, but I couldn’t find an official press release acknowledging the adjustment to their data. The massive adjustment only appeared as text in the quarterly flow of funds release.
Participants in the municipal bond market have been discussing the Fed’s bad numbers for months. Daniel Berger of Thomson Reuters wrote back in June about the possible reasons why the Fed’s municipal bond data were so understated. Here is how Berger explained the big discrepancy (emphasis mine):
The Federal Reserve calculates outstanding muni debt with a “perpetual inventory” approach. They use Thomson Reuters data for the new issue market and simply add it to their outstanding data, while adjusting for maturing bonds. The problem here is that if their outstanding number is understated to begin with, they will produce results that are consistently too low.
The most recent market data on new issues is what you are seeing in the top chart. Every quarter states and municipalities issue and redeem bonds. In some quarters, a lot of new debt is issued and only a small amount is redeemed, producing a net gain in new bonds outstanding. This new amount is added or subtracted to the “perpetual inventory” of municipal bonds outstanding. The Fed got this new issue data from Thomson and got that part right at least.
Daniel Berger also touches on another aspect the Fed got wrong:
Another issue is that the Federal Reserve calculates retail muni holders as a residual. This leads to inaccuracies and the problem is bigger in munis than any other sector because the household sector is such a big proportion of the total. Swings in the residual don’t mean as much where the residual -‐ the amount you can’t count directly -‐ is a much smaller proportion. In addition, foreign holders of munis (among the largest holders of municipal Build America Bonds) are difficult to calculate from reported sources as they are not subject to the same reporting requirements as domestic institutional investors.
Here is the new Fed data on retail, or household, holdings of municipal bonds, which was revised up by about $840 billion :
Determining the size and ownership of the municipal bond market is very important as Congress considers changing the municipal-bond tax exemption. Reuters reports:
[Thomson Reuters Daniel] Berger said an accurate count of who owns what bonds could help the market gain some respect from policymakers. ’(Someday) the municipal bond market might be under assault by a revenue hungry federal government,’ he said. ‘The higher number of outstanding bonds means that legislators must carefully consider the taxation and issuer policies impacting this market because they affect a much larger investor base than originally believed.’
Muniland has always been tightly controlled by a small group of Wall Street banks and much of the opacity comes from a lack of data. It would be really great for the Fed to due its part for market transparency and publish muniland data at the state and maturity level. If they are ready to create an advisory committee I’d be happy to serve and know other market professionals that are ready too.
Data source: Federal Reserve Flow of Funds Account of the United States - Flows and Outstandings Third Quarter 2011
Muniland: The Fed’s data snafus