Yesterday, Citigroup released a report claiming that Federal Reserve is far underestimating the amount of municipal debt outstanding. The Fed reported (page 91) that there was $2.925 trillion in outstanding muni debt at the end of 2010. Citi believes that figure to be closer to $3.7 trillion. Not a small difference!
How does Citi explain the error? The full report isn’t public, but according to the Bond Buyer, they say it’s a problem with the Fed’s data on household holdings of municipal debt. Citi also says this is not a new error, and that the last several years’ reported changes in debt have been more or less correct, with the base figure far too low—which is to say, if there’s a hidden pile of muni debt, it’s not a hidden recent explosion in muni debt.
Color me skeptical about the claim that the Fed data are so far off. While (as I understand it) the Fed gets its data by surveying bondholders, the Census Bureau also reports a total state and local debt figure that it calculates by collecting information directly from municipal bond issuers. The Census data are only available far in arrears—the most recent estimates are as of July 2008 (last page)—but as of that time the Census estimate of state and municipal debt outstanding was actually about three percentage points lower than the Fed’s contemporaneous estimates.
So, if the Fed is way off on muni debt, so is Census. And it’s hard to see why the Census figures, calculated with the benefit of a couple of years’ hindsight and actual reporting from issuing governments, would be so inaccurate.