The usually reliable Arnold Kling misses one, I think.

[Y]esterday I was trying to teach the consequences of paying interest on reserves to my bright high school students. They got, correctly, that paying interest on reserves lowers the money multiplier (it makes banks want to hold more reserves, which lowers the amount of lending they do for a given supply of reserves from the Fed) and is therefore contractionary.

But one student asked, "Doesn't paying interest on reserves increase the deficit, and isn't that expansionary?" My response was to say that this is correct...

But the Federal Reserve does not draw money from the Treasury to pay interest. It pays it out of its own earnings on its portfolio. Look at its (audited!) financial statements here, and see page 5. You'll see a line item for "depository institution deposits" under "interest expenses." The Fed remits the excess of its income less expenses to the Treasury (see the line "payments to U.S. Treasury as interest on Federal Reserve notes") which was in 2008 lower. But net interest income didn't change, so you would be hard pressed to say that paying interest increased the U.S. federal deficit.

Likewise would have been my response to increasing reserve requirements. However, I agree with Kling that the signal that would have created would have been potentially more damaging than paying interest on excess reserves. Doing the latter does, I think, increase the marginal cost of lending to the bank, and thus reduce loans. (Paying interest on required reserves, on the other hand, is just a pure transfer of income from the Fed to commercial banks.)

My thanks to my colleague Eric Hampton for thinking this through with me.

Labels: economics, Federal Reserve