Net Net: Promoting innovation and managing change
Net Net: Promoting innovation and managing change

John Cochrane on why the Fed's QE program has 'no effect'

The Federal Reserve Building in Washington
Raymond Boyd | Michael Ochs Archives | Getty Images

The Richmond Fed has a tremendously engrossing interview with the economist John Cochrane, the AQR Capital Management Distinguished Service Professor of Finance at the University of Chicago Booth School of Business.

Cochrane is fascinating because he is seriously engaged in discovering the way fiscal and monetary policy mix and interact. Here he is on why quantitative easing doesn't really do very much:

In my opinion, QE has essentially no effect. Interest rates are zero, so short-term bonds are a perfect substitute for reserves. QE creates a minor change to the maturity structure of government debt — and doubly minor because the Fed's effort to shorten maturity is essentially matched by the Treasury's new sales of long-term bonds. We've had much larger changes in the quantity and maturity structure of debt in the past with no big effect on the level of interest rates. You have to buy some new theory of very long-lasting flow effects, but I think coming up with new theories to justify policies ex post is a particularly dangerous kind of economics.

So I don't think the theory suggests QE can have a big effect. What about the evidence? Most of it comes from announcement effects. Even there, it's pretty weak: a 15-or-so basis point change in interest rates in return for a pledge to buy trillions in Treasuries. But interpreting announcements is tricky, and tells you a lot less about QE's effectiveness than you might think.

Markets tell you what they think will happen — mixed with what risks they're willing to take — but not why. If the Fed announces more QE or delayed tapering of QE and bond prices rise on that announcement, is that because QE itself is moving the markets? Or is it because bond investors think, "Wow, the Fed is scared, so it will keep interest rates low for a lot longer than we expected"? Without a solid economic reason to believe QE on its own has much of an effect, the latter interpretation seems more likely.

Also, the market's reaction to an announcement doesn't tell you for how long QE could have an effect. QE advocates take these reaction estimates, assume they are causal, and assume they are permanent. There are more than $17 trillion in U.S. Treasury bonds outstanding, and another $1 trillion are being issued every year. Why would the Fed buying even $1 trillion of them — in exchange for reserves, which are really just floating-rate overnight debt — have a permanent effect? Microstructure studies might see price pressure in Treasury markets but for a day, not for years. Also, if market reactions prove anything, they prove that markets think QE has an effect. But this is a policy we've never seen before, so we don't have much rational expectations-based reason for believing markets are right about it. Markets are great at correlations and unconditional forecasting, and less so at structural cause and effect for things they have never seen before.

So neither the theory nor the evidence make me think QE is effective. But the good news is that we therefore can't worry too much about its reversal. It's neither going to cause hyperinflation, nor need it cause much trouble when the Fed "tapers."

That really sounds like what Warren Mosler and I were saying back in September:

To understand quantitative easing you first need to understand that the Fed is a bank with two kinds of accounts that the Fed calls reserve accounts and securities accounts, which normal banks would call checking and savings accounts (or CD's). A "reserve account" is just a fancy name for a checking account. And a U.S. Treasury bill, note, or bond is just a fancy name for the dollars on deposit in those securities accounts at the Fed.

So when the Treasury spends, the Fed just adds the dollars to some bank's 'checking account' at the Fed. And when Treasury securities are sold, the Fed shifts those dollars to a 'savings account' at the Fed. Both are just dollar deposits at the same Fed, but the dollars in checking accounts are not counted in the national debt and the dollars in savings accounts are the national debt.

So why are reserve accounts not the debt, while securities accounts are the debt? It's all just a holdover from way back before most of you were born, when only dollars in checking accounts were legally convertible into gold at a fixed price. It's all entirely inapplicable today.

Apparently Cochrane's been saying this sort of thing for quite a while, although I hadn't noticed until now. It's definitely good to learn that this view of money and debt has a perch at the University of Chicago.

(Read more: For new Fed chair, headaches may be only starting)

—By CNBC's John Carney. Follow him on Twitter @Carney