Paul Krugman on why the liquidity trap really matters

Read his latest post, which outlines many key but usually unstated assumptions behind monetary theory and policy.  It is one of the most instructive econ posts to appear in some time. 

That said, on the policy issue I think one of Krugman’s earlier posts (I can’t find it) is closer to the mark.  With or without a liquidity trap, monetary policy can’t fix negative real shocks and — here is now the earlier Krugman — monetary policy can’t make insolvent (or potentially insolvent) banks whole.  That’s my take on why the Fed is relatively powerless, not because of a liquidity trap.  If you believe, as a Keynesian would, that insufficient aggregate demand is the problem in the first place, you will be relatively worried about liquidity traps.  If you believe, as a neo-Austrian would, that malinvestments and coordination problems are the key issues, you will look toward other factors which limit the power of central banks to restore order. 

In my view sometimes the Keynesian perspective is relevant, but not so much today.  As the contraction of credit spreads through the Fed-regulated banking sector, however, and the broader money supply aggregates come under stronger negative pressure, the Keynesian perspective is likely to become more relevant.  That is in fact my major medium-term worry and we probably should be pessimistic in this regard.

There is a separate and very important liquidity issue about restoring the markets and valuations for bank loans, but this is not a liquidity issue in the sense of Keynes’s portfolio theory or the traditional liquidity trap.

Addendum: Brad DeLong adds comment.  Another way of putting my point is this: in the situations where a liquidity trap might be binding, there is usually some even worse constraint which is more binding, thereby making the potential liquidity trap not so much a problem at the relevant margin.


Has anyone noticed that oil has become the new gold?

With or without a liquidity trap, monetary policy can't fix negative real shocks

Hi Tyler,

I can think of 3 possible definitions of real shocks:

1. Shocks that change the level or growth rate of potential GDP

2. Cost (oil) or "supply" shocks

3. "Real" Demand shocks

I see that monetary policy has no influence (can't offset #1, there are tradeoffs in dealing with #2 and can certainly be useful (ignoring the liquidity trap) #3

Just curious : what do you have in mind when you discussed "real" shocks?

Brent: Has anyone noticed that oil has become the new gold?

Yes and yes.

travis: Can't the Fed just keep buying treasuries or any other debt by printing cash (or just crediting accounts)?

The result would be further devaluation of the dollar and an increasing chance of a "run" on the federal government itself (i.e. a mass dumping of Treasuries, potentially doubling or more interest costs on the national debt). And if you think gas prices are bad today, try for example tripling oil prices to $350/barrel (the probable result if long-run inflation expectations increase by 2.5%/year, e.g. from 4.0% to 6.5%).

Mmmm... a substantial post on core economic issues. We need more of those.

Historical and political posts, as well as posts describing economic institutions are good too -- but hard-core economics posts are the best.

travis: the $800 billion worth of Fed assets is not directly relevant...Krugman should be moaning that the problem is too big to solve by creating fiat money because it will create inflation.

I agree with the latter half of your statement, but the assets certainly are quite relevant, even if there are some alternatives to relying on them. The Fed's assets become relevant because relatively free trade in global oil, and the freedom of commodity markets via the dreaded speculators to immediately reflect changing inflation expectations, creates political constraints on Fed behavior. If it weren't for the rapid reaction of the dollar and gas prices to monetary policy, the pain of inflationary policy would be too diffuse to make it an important political constraint on the actions of the Fed.

It's also the case that Treasuries aren't created out of thin air. They reflect the budget deficit passed by Congress and resulting future U.S. taxpayer obligations. And as Krugman pointed out, a Treasury with a nominal return of near 0% is practically the same thing as a dollar. In the current flight-to-safety environment with short-term Treasuries paying near 0%, if the Fed wants to inflate the dollar faster than the inflation implied by the federal budget deficit, they have lower the rate they lend to banks or they have to buy up something besides Treasuries, for example mortgage paper. That's why gold and oil went way up when the Fed was buying mortgage paper in 2007 and earlier this year, went back down when the Fed realized they were damaging the dollar by playing John Law and these purchases subsided, and went back up spectacularly when Paulson announced the bailout late last week, which involves the federal government itself buying mortgage paper and increasing the budget deficit (and thus issuing more Treasuries for the Fed to buy) to do so.

(Oil has also recently gone down a bit relative to gold because of the looming worldwide recession -- the two commodities don't track 100%, as oil is obviously also influenced by fundamental industrial expectations and surprises -- but over entire business cycles, allowing for the slightly faster geological depletion of oil than gold, they track each other pretty closely, and there is also significant short-term correlation caused by monetary policy).

Comments for this post are closed