Almost exactly four years ago, economist Paul Krugman wrote the following explanation of the difficulty in defining the money supply for purpose of monetarist analysis. I think it is worth the resurrection. Surely we don’t mean to identify money with pieces of green paper bearing portraits of dead presidents. Even Milton Friedman rejected that, more than half a century ago. For one thing, a lot of those pieces of green paper are pretty much inert — sitting outside the United States, in the hoards of drug dealers and such. For another, checking accounts are clearly a close substitute for cash in hand. Friedman and Schwartz dealt with this by proposing broader aggregates –M1, which adds checking accounts, and M2, which adds a broader range of deposits. And circa 1960 you could argue that those aggregates were good enough. But now we have a large shadow banking system, in which things like repo serve much the same function as deposits; M3 used to capture some of that, but...