What is Money?

1/14/2013 03:22:00 PM
Steve Randy Waldman has an interesting post pointing out that because the Fed will pay interest on reserves from here on out (a policy started in 2008), reserves and treasury bonds will always be perfect substitutes. Krugman has a rebuttal that is, in my opinion, no good--he only points out that currency drain exists, which is quite uninteresting and ignores the deeper point about the relationship between money, inflation, and interest rates. Indeed, I've always thought that Krugman's argument about the substitutability of bonds and money in a liquidity trap was a moot point.

I wrote a short tract on a related topic some time ago, in response to "endogenous money" arguments that were flying around at the time. Admittedly, the paper was not quite as rigorous as it should have been, but it includes a mathematical proof of the fact that monetary base equals the sum of currency and reserves. One point that I want to add here is this: my paper makes an assumption about the definition of currency. If you want to play a logic game, go see if you can find it? Otherwise, spoiler alert: I defined currency as any asset backed by the US treasury that banks would accept as a deposit into a checking account. Banks will only accept deposits if they are backed by someone they trust: either the federal reserve (reserves), other banks (loans), or the treasury (currency). On Waldman's post, I claimed that this would always pose an institutional constraint on the substitutability of treasury bonds and actual currency, because you can't (to my knowledge) go to a bank, deposit a treasury bond, and then write checks against it. But suppose you can--this would mean that treasury bonds are a form of cash, and that the Federal Reserve is no longer in control of how much currency is in circulation.

The caveat, however, is that the Federal Reserve can decide for itself what banks will and will not accept as a legitimate form of cash. To see this, just imagine that a bank tries to deposit a treasury bond in its federal reserve account. The Fed can choose to accept it and credit the account with an amount equal to the value of the bond--meaning that the treasury bond has just increased the money supply. Or the Fed can return the bond to the bank with a note saying that it is not a legitimate form of payment. In that case no new money is created. If banks can't deposit the treasury bonds at the Fed, then they won't let households deposit them in their checking account (keep in mind that when people write checks, that money comes out of the banks' accounts at the Federal Reserve, so the deposit does have to be credible at the Fed). So I maintain that the Fed has total control over whether treasury bonds can be used as currency.

This is not a moot point, either. The Federal Reserve recently announced, essentially, that it would not recognize deposits of large-denomination platinum coins. That means that the Treasury can mint the $1 trillion coin, but it won't be currency.

This is, of course, all tangential to the point that Waldman and Krugman are discussing. But I thought it worth pondering the nature of currency for a moment.