Bryan Caplan  

Are We in a Liquidity Trap After All?

Central Bankers and History... Corrupt Bargains...
A while back, I argued that we're not in a liquidity trap.  But then Greg Mankiw pointed out that the two series on his webpage had different endpoints, leaving the matter again in doubt.  Now Jeff Hummel says that excess reserves have indeed risen almost 1:1 with the rise in the monetary base:

[Y]ou can tease out recent estimates by going to the Fed's weekly H.3, H.4.1, and H.6 releases and by checking against how much of the base increase has ended up as currency in the hands of the general public. Using these means, I put total reserves for the entire banking system (not adjusted for changes in reserve requirements and not seasonally adjusted but counting all vault cash and clearing balances) at $72 billion in August. Currently, as of October 22, total reserves are somewhere between $343 and $358 billion. Notice how close this comes to matching the corresponding increase in the base, from $847 billion to $1,149 billion.
But Jeff adds that it's premature to cry "liquidity trap":

But this hardly indicates a liquidity trap, for at least four reasons: (1) Base money can only be held as reserves or currency, and the allocation of a massive base increase between the two tells you absolutely nothing about the overall demand for base money. (2) At the same time that the Fed stomped on the monetary accelerator, it began paying interest on bank deposits at the Fed, obviously increasing the demand for reserves. (3) A sudden SHIFT outward in the demand for base money does not in and of itself demonstrate a liquidity trap, as the history of bank panics teaches us. (4) You must allow for lags to see whether this incredibly sudden base increase works its way into the broader monetary aggregates.

Comments and Sharing

COMMENTS (1 to date)
dWj writes:

If I construct a time series of the S&P500 + days since January 1, 1900, I will find that it has a correlation of practically 1 with time; a regression with time will give a very good fit, and the data will rise about 1:1 with time. This is not to indicate that there is no interesting content in the series other than what is captured by the time-covariance; a large source of variance is simply swamping any other signal.

The 1:1 variation in recent monetary base change to recent excess reserves over a period of time in which both have changed a lot more than any historical norm, demand for liquidity seems to have gone up, and the Fed has started paying interest on excess reserves can be explained as readily by fed accommodation of a velocity shock as by a view that the fed has somehow been fecklessly pouring money into the system, only to see it pool up. It doesn't really suggest that monetary easing is having no effect, so much as that something has caused a large run-up in both reserves and the money supply that makes these rather poor indicators to watch. I wouldn't worry about liquidity traps until interest rates get to 50bp. So, you know, may December or something.

Comments for this entry have been closed
Return to top