Forensic semantics: the meaning of liquidity trap
I promise not to do too many more posts about a) macro or b) Paul Krugman. I don’t just love macro, these are not my most popular posts, and Krugman is too shrill to read on a regular basis. Nevertheless, I think I can sort through some of the recent disagreement about liquidity traps.
The term “liquidity trap” comes from Keynes. He described it as a theoretical possibility under which monetary policy would be ineffective; it would be unable to stimulate an economy in recession. Consequently, fiscal policy would be needed. To my knowledge, Keynes was not claiming that the economy ever has been in a liquidity trap; it’s simply a possibility that occurs under very specific conditions. Those conditions define “liquidity trap,” but there is disagreement over when they hold.
What is at stake here is the status of fiscal policy. If the economy sometimes experiences liquidity traps, then that is perhaps a good reason to keep fiscal policy in our toolbelt. If the economy never experiences a liquidity trap, then monetary policy strictly dominates fiscal policy: it is faster, less wasteful, and does not increase sovereign debt. It’s sometimes hard to say whether advocates and detractors of fiscal policy take those sides because of their position on liquidity traps or vice versa.
Krugman is among those who think that the world does experience liquidity traps, that we are in one now, and that we need more fiscal policy. Why does he think this? Arnold Kling does some of the forensic work and uncovers an old and a recent statement from Krugman on liquidity traps. Arnold says they are inconsistent, but I think they are perfectly consistent. The old statement:
My view … is that the liquidity trap is real: no matter how much the Fed increases the monetary base, it has no effect, because it just substitutes one zero-interest asset for another.
The economy is in a liquidity trap when even a zero nominal interest rate isn’t enough to restore full employment. That’s it.
What Paul is saying is that the economy is in a liquidity trap when the nominal interest rate on short-term Treasuries is zero. When the Fed tries to expand the money supply by buying up short-term Treasuries, it is swapping cash for Treasuries. Normally, cash and Treasuries have different properties: cash has a nominal interest rate of zero and Treasuries bear some positive nominal interest rate. However, when Treasuries bear an interest rate of zero, they are basically the same as cash. They are backed by the US Government and they don’t carry interest. Why should swapping one asset for an identical asset make any real difference in the world? On this narrow point, Krugman is clearly correct: it wouldn’t make a difference at all.
Krugman is wrong, however, that this constitutes a liquidity trap, either in the sense that Keynes meant it or in the looser sense that monetary policy is ineffective, because swapping cash for short-term Treasuries is not the only (or even necessarily the best) way to conduct monetary policy. First of all, it is important to recognize that there is not just one nominal interest rate. There is an infinity of nominal interest rates. If the interest rate on short-term Treasuries is zero, the Fed can swap cash for longer-term Treasuries. It could in theory buy private bonds, or stock, or mortgage-backed securities, or even non-financial assets. In any of these cases, the Fed is increasing the amount of money in circulation, and it is removing less liquid assets. This is expansionary except in extraordinary circumstances I’ll discuss below.
Incidentally, the Fed can also conduct monetary policy by other means. It can simply print money and distribute it, the infamous “helicopter drop.” It can buy foreign currency. It can lower the interest rate or raise the penalty on excess reserves that banks hold at the Fed. It can promise to inflate more in the future. All of these actions are expansionary, again except perhaps in extraordinary circumstances.
What are the extraordinary circumstances in which all monetary policy is ineffective? Keynes got it right. Monetary policy is ineffective when people want to hoard whatever cash they can get their hands on. In technical terms, the demand for money is infinitely elastic. The point is that increasing liquidity in the system (buying illiquid assets with liquid assets, say) does not translate into more spending because people soak up whatever liquidity there is.
When is demand for money infinitely elastic? Basically never. This is what Tyler is saying in his most recent post on liquidity traps. In Tyler’s terminology, there are multiple margins on which people express preferences for liquidity. There is the money-bonds margin, and in fact, there are multiple money-bonds margins. When the nominal interest rate on short-term Treasuries is zero, that is one margin on which people are expressing a preference for liquidity. But as I argued above, there are other bonds, and people are generally willing to sacrifice liquidity for a non-zero rate of return. There is also the money-goods margin. People are generally willing to sacrifice liquidity for stuff. That is, if you give them money, they spend some of it. But since Keynes is all about aggregate spending, you can see how it would be the case that if people infinitely preferred liquidity to goods (they were unwilling to spend even if you gave them more money), then it would be desirable to have the government to engage in direct spending (fiscal policy) to boost aggregate demand.
So why does Krugman fixate on only one interest rate, on only one particular money-bonds margin? I think that it’s just a lack of imagination about what monetary policy consists of. Traditionally, monetary policy in the US has consisted primarily of open market operations on short-term US Treasuries. But there is nothing special about this particular kind of monetary expansion. If Krugman wants to call it a liquidity trap when the nominal interest rate on short-term Treasuries is zero, he needs to abandon the conclusion that fiscal policy is called for in a liquidity trap. I prefer to retain Keynes’s original meaning and conclusions by defining a liquidity trap as an infinitely elastic demand for money.