There’s been plenty of debate in the blogosphere recently about the IS-LM model. In particular, while Paul Krugman and Brad Delong have been defending it one of its critics has been Matt Rognlie. To my understanding Matt’s critique comes down to the assumption of an exogenously fixed money supply, central banks don’t fix the money supply anymore (if they ever did) so you can’t get reasonable conclusions out of the assumption that they do.
While I tend to side with Krugman and DeLong on this I actually think the model is best saved by re-interpreting the LM curve as the set of (Y,i) pairs that equate the supply and demand for liquidity, not the set of points that equate the supply and demand for money. One thing that needs to be stressed here is that the standard IS-LM model only has two assets, money, which pays no interest but serves a liquidity purpose and bonds, which pay a yield but are relatively less liquid. In the model money demand is liquidity demand and money supply is liquidity supply. That’s not true of the real world. So, I tend to think the model always really meant the LM curve to be the set of (Y,i) points that equate the supply and demand for liquidity.
The Aggregate Supply of Liquidity
Holmstrom and Tirole have an entire book (Inside and Outside Money) based on the premise that the aggregate supply of liquidity is limited because there is a limit to how much output can be pledged externally, basically this comes from various things like agency problems and information asymmetries that induce some sort of limited commitment on the part of the insiders to any productive enterprise.
The important point though is that this is basically a technological constraint on the economy. Furthermore, since their are similar limits to the amount of output that the government can extract to back its claims we see that their is effectively an exogenous limit to the aggregate supply of liquidity.
If we combine this sort of market incompleteness with the “liquidity supply = liquidity demand” interpretation of the LM curve then I think we can get back to thinking about the economy in pretty much the original IS-LM framework with an exogenously given liquidity supply.
This framework can then be applied to several issues, in particular to understanding the difference between what the Fed was doing in QE1 versus QE2 and twist.
QE1: Liquidity Provision
The first round of asset purchases, in which the Fed almost entirely bought agency debt was about provision of liquidity. Essentially you had a large stock of claims that were liquid and had ceased to be so, the Fed was attempting to replace some of them with claims that would still serve the liquidity purpose. To see that this is so one only needs to notice that, as Stephen Williamson had pointed out, the Fed was actually a net seller of Treasuries during that time.
In the IS-LM framework this was classic monetary policy, there had been a reduction in aggregate liquidity that shifted the LM curve to the left. The fed responded by increasing the supply of liquidity to shift the LM curve back out to the right.
In so far as the fall in output was quickly arrested and deflation averted I’d say this worked in exactly the way expected. Of course, we haven’t returned to full employment, or even much in the way of employment growth. That’s what QE2 and twist are about.
The Liquidity Trap
Of course the fall in output didn’t go with a rise in the risk-free real interest rate so to be consistent with observation the model must be consistent with the IS curve shifting as well. I think this is a reasonable interpretation of events.
Along with the liquidity shock we had a large increase in risk premia, that means that a given risk-free real interest rate will now imply less aggregate investment than previously and that would show up in the IS-LM model as a shift in the IS curve. In this case the IS curve has shifted so far left that the full employment level of Y corresponds to a negative i, the liquidity trap has sprung.
QE2 and Twist: Stimulus
In the context of IS-LM the second round of QE and twist, where the fed buys up yielding assets with low yielding reserves, are about lowering risk premia and thus shifting the IS curve back out. This is stimulus to investment in the same sense that a lowering of the interest rate is in normal times, this is an attempt at “normal” policy stimulus. Of course getting the amount right is much harder since we don’t have so much experience with this, but it’s not really any different from normal Fed policy.
PS: I think it’s very important to notice that in neither case did QE have anything at all to do with the supply of base money per se. Matt is exactly right when he says “The messy regulatory and technical issues that determine banks’ demand for reserves on the fed funds market have virtually nothing to do with the effects of monetary policy on the economy. Obsessing over money demand is a waste of time.” With that I couldn’t agree more.
Senior Quantitative Portfolio Manager, Co-Fund Manager
Oct 16, 2011 at 7:04am