What would an IS shock look like?
By Scott Sumner
One thing that makes it hard to discuss macroeconomics with people is the widely held assumption that changes in interest rates reflect changes in monetary policy. That misconception comes from the fact that people misinterpret the implications of two true facts:
1. In the IS/LM model, an easy money policy lowers interest rates.
2. The Fed targets short term interest rates.
People think that if the Fed is controlling rates, and if an easy money policy causes rates to fall in the IS/LM model, then a fall in rates must be an easy money policy. But of course this doesn’t at all follow from the previous assumptions. Recall that in the IS/LM model, low interest rates can be caused by either an increase in the money supply or a leftward shift in the IS curve (perhaps due to bearish expectations). Indeed it’s even possible that the IS curve might shift left because the public expected a tight money policy that would produce a depression.
So falling interest rates do not necessarily indicate easy money, they might reflect a decline (shift left) in the IS curve. How could we tell the difference? If the fall in interest rates were caused by a leftward shift in IS, then you’d see GDP growth decline at the same time. If it was caused by easy money, then GDP growth should accelerate. In that case, why not cut out the middleman and simply look at what’s happening to NGDP growth?
There’s nothing controversial about what I’m saying here. It’s simple IS/LM. But for some reason lots of people have a lot of trouble clearly seeing the implications of this model. Thus when rates fell in late 2007 and early 2008, it was clearly due to a decline in the IS curve (or its growth rate to be more precise.) After all, NGDP growth slowed at the same time. The monetary base was stable. And yet I think you could find many people, even professional economists, arguing that the Fed eased policy in late 2007 and early 2008, when they continually cut rates from 5.25% to 2.0%. And they made this claim despite the fact that the LM curve did not shift, rather the IS curve shifted left. It’s as if they thought the IS/LM model was simply a fixed IS curve, with movements along the curve. It would be like people who think the supply and demand model is actually a “demand model”, and who are puzzled when people drive more “despite” high oil prices.
Is the onset of the 2007-09 recession typical? This graph suggests the answer is yes:
The correlation is far from perfect, but there is a definite tendency for interest rates to fall during recessions (grey bands on the graph). Yet many economists talk about those interest rate declines as if they are “easy money” policies, whereas the IS/LM model implies they are actually caused by a leftward shift in the IS curve. This is a very basic error, made all the time.
The question is why? I think it’s due to the fact that the Fed targets interest rates. Since they control short-term interest rates, it seems plausible that any change in interest rates is caused by a shift in monetary policy. Of course this is not true, most changes are simply the Fed adjusting its target to reflect changes in market rates induced by shifts in the IS curve. But that’s not how it looks to people.
Is this mistake costly? Previously I’ve argued that the misdiagnosis of the stance of monetary policy was one of the most important causes of the Great Recession. If nominal short-term rates had risen to 8% during 2008, most people would have viewed monetary policy as getting much tighter, and the Fed might well have been blamed for the recession. That blame would have caused the Fed to shift course. So the fact that people wrongly thought monetary policy was being eased ended up distracting attention from the Fed, and gave them a sort of free pass—even from liberals like President Obama. Of course conservatives were even worse, many of them predicting high inflation.
Supply and demand is a very simple model, but that doesn’t stop Nobel Prize winners from making the basic error of reasoning from a price change. And IS/LM is also a very simple model, but once again elite economists often draw the wrong conclusions about the stance of monetary policy. It’s an unfortunate mistake, which cost millions of workers their jobs in 2008-09.
PS. Fortunately, the mistake is beginning to be rectified.