Paul Krugman writes:
How To Think About QE2 (Wonkish): Still on the run, so no long posts. But with all the talk about further quantitative easing by the Fed — QE2, for quantitative easing, the sequel — I think it’s worth sharing one way of thinking about what’s on the table — and why you shouldn’t be too optimistic about its effects. This isn’t original, although I don’t know who deserves the credit.
So, here it is: in effect, QE2 amounts to a decision by the US government to shorten the maturity of its outstanding debt, paying off long-term bonds while borrowing short-term. This should drive down long-term interest rates. But how much?
How do we get to this view? Think first of the Fed’s balance sheet. The Fed’s liabilities are the monetary base — currency in circulation, plus bank reserves. Those bank reserves are essentially short-term borrowing: the Fed pays a small interest rate on them, which is comparable to the interest rate on Treasury bills. More broadly, in a near-zero-rate world, cash — an official liability that pays no interest — is essentially equivalent to T-bills — another official liability that pays more or less no interest.
What happens when the Fed buys long-term government securities? If we consider the Fed and Treasury as a consolidated entity — which, for fiscal purposes, they are — then what happens is that some long-term federal debt is taken off the market, and paid for by issuing more short-term debt in the form of monetary base. It’s just as if Treasury sold 3-month T-bills and used the proceeds to buy back 10-year bonds.
So the question to ask is, how much do we think federal management of its maturity structure matters for the real economy? I think if you put it that way, most people wouldn’t be terribly optimistic.
Anyway, my jet-lagged thought for the day.
I would put it differently. The point--from one point of view, the neo-Wicksellian point of view--behind quantitative easing is to reduce the interest rate that matters for private business investment: the long-term, default-risky, systemic-risky, beta-risky, real interest rates at which private businesses finance their capital expenditures. You can reduce this flow-of-funds equilibrium interest rate and raise the level of economic activity in any neo-Wicksellian framework in two ways:
Reduce the "safe" real interest rate on short-term, safe government bonds.
Reduce the various premia--duration, default, systemic risk, and beta risk--between the rates the Treasury pays to borrow in T-bills and the rates businesses pay to borrow.
Conventional open-market operations that lower the nominal interest rate on T-bills accomplish the first. Once the nominal interest rate on T-bills has been pushed to zero, quantitative easing policies that create expectations of higher future inflation continue to lower the real interest rate on T-bills and thus help the situation.
Suppose, however, that the nominal interest rate on T-bills is zero and that you cannot alter inflation expectations--cannot commit to keeping your quantitative easing permanent, cannot commit to an exchange rate path, whatever, you cannot do it and inflation expectations are immovable. Then what?
Then, as Paul Krugman says, quantitative easing is working be altering the spread between the short-term safe T-bill rate and the long-term, systemic-risky, beta-risky, default-risky rate. How does it do that? Lloyd Metzler and James Tobin would say that it does so by altering relative asset supplies--by taking duration risk, systemic risk, beta risk, and default premia off of private savers' books and placing them on the government's books (and thus on the taxpayers, who are a very different group of people than are private savers). To the extent that quantitative easing thus involves assets whose risk characteristics are very similar--federal funds and two-year T-notes, say--we would not expect even a lot of quantitative easing to have much of an effect on anything.
Thus a quantitative easing program that is going to have bite should involve Federal Reserve purchases of long-term risky private assets rather than merely long-term U.S. Treasuries. Hiring PIMCO as an agent to manage a long bond index portfolio naturally comes to mind--if one could avoid its front-running.
And, of course, the most effective quantitative easing program of all would involve the Federal Reserve issuing reserve deposits and using that purchasing power to buy the assets that are the furthest away in their risk characteristics from short-term government bonds: bridges, dams, the human capital of American citizens, police protection, research and development. The best quantitative easing program of all is a money-financed fiscal stimulus, as Jacob Viner said back in 1933:
It is often said that the federal government and the Federal Reserve system have practiced inflation during this depression no and that no beneficial effects resulted from it. What in fact happened was that they made mild motions in the direction of inflation, which did not succeed in achieving it.... [If] a deliberate policy of inflation should be adopted, the simplest and least objectionable procedure would be for the federal government to increase its expenditures or to decrease its taxes, and to finance the resultant excess of expenditures over tax revenues either by the issue of legal tender greenbacks or by borrowing from the banks...