I’d like to think that I’m misunderstanding Arnold’s views on monetary econ.  But in his latest post, he states his position quite clearly.  He begins with the uncontroversial:

The Federal Reserve can manipulate some market interest rates by
printing money and using it to buy securities. If it buys massive
quantities of mortgage securities by printing dollar bills, it can
lower the interest rate on those securities. In theory, the Fed could
do this by exchanging mackerel for mortgage securities just as well as
by exchanging money for mortgage securities.

So far, so good.  My only quibble here would be to point out that mackerel have to be caught, but the Fed can produce money by adding zeros.  But then Arnold goes out of his way to challenge my quibble:

However, I am not sure how long they can keep it up. At some point,
market participants will think, “The Fed is not going to keep buying
these things forever. When they stop buying, the price will fall.” And
such thinking will dampen prices (raise interest rates) today. The
longer the Fed fights the market, and the more government keeps
spending on other stuff, the greater the likelihood that too much will
be financed by printing money, and we will have hyperinflation. As long
as we know that the Fed is afraid of hyperinflation, then it has a
limited supply of money, just as it would have a limited supply of
mackerel.

Arnold seems to grant that in principle, the Fed has the power to increase nominal GDP a million-fold.  After all, during hyperinflations, nominal GDP does skyrocket.  So if the Fed can increase NGDP a million times, why in the world can’t it raise it by a measly 5%?  His only answer, as far as I can tell, is that the market knows that the Fed won’t hyperinflate, so the Fed will shy away from measures that threaten hyperinflation.  “Therefore,” the Fed can’t raise nominal GDP by 5%. 

Say what?  You could just as easily argue that since I don’t want to get morbidly obese, I won’t be able to gain five pounds.

Is this a caricature?  I don’t think so.  Arnold spells it out for us:

Suppose that the Fed had decided to print a lot more money at some
point last year. The Scott Sumner thesis is that this would have raised
nominal GDP. In my view, the fall in nominal GDP was due to the fact
that real GDP had to fall. Real GDP had to fall, because the economy
was beginning a Great Recalculation… In terms of MV = PY, I see PY as
largely outside of the Fed’s control–the P part was determined by the
combination of habit and gradual adjustment in the Great Recalculation,
and the Y part was determined by the frictions involved in the Great
Recalculation. So if we could rewind the tape to some time in 2008,
hold everything else equal, and have more M, I think we would see
essentially a 100 % offset in V.

If Arnold’s position were that raising nominal GDP wouldn’t help real GDP, I could understand it.  But he’s actually questioning the ability of central banks to affect nominal variables.   Even stranger, he’s using hyperinflations – airtight proof of central banks’ near-infinite power over nominal variables – to prove that this power is illusory.

Arnold, please tell me I’m missing your point!

P.S. If nominal GDP and liquidity traps occupy your mind day and night, you are in luck this week.  The Money Illusion’s Scott Sumner is speaking at GMU not once but twice.