Separating Hyperplanes

The Housing Bubble

Matthew Martin 7/24/2012 11:21:00 AM

I've been thinking about what people have been saying about the role monetary policy played in the housing bubble, contrasted with this graph:

The red line is the overall price level and the blue line is house prices, both indexed so that the trough after the 1990 recession is 100, just to give us a common reference point that was, presumably, before the bubble. The green line is the federal funds rate, which is targeted by the federal reserve. When people say that the Fed held interest rates too low in the early 2000s, they mean specifically that they think this federal funds rate was too low.

I'm rather puzzled by the argument that the interest rate was too low. First of all, there is actually a positive correlation between the interest rate and the growth in home prices, not the negative relationship that is typically assumed. Just look at any of the interest rate hikes in the graph: an interest rate hike has never reduced home prices relative to CPI anywhere in the available data, and in all but one case (in the mid 1990s) was actually associated with higher home prices. I'm not trying to argue causality here--I just don't think that an interest rate hike would have been effective at all in reducing or slowing the growth in home prices relative to inflation.

Now, the fact is that we have had very low inflation since 1997 when the current housing bubble appears to have begun. Inflation was even lower than it appears when you consider that housing also represents a significant chunk of the CPI index. My view is that if we had aimed for a higher overall inflation rate, we would have been able to devalue home prices relative to the overall price level, and thus have avoided the bubble. That is to say, I think that higher inflation would have increased wages faster than it increased home prices, so that the ratio of the two would never have grown to the proportions it did prior to the current recession.

As for the very low interest rates, it would be a mistake to cite these as a reason to believe that monetary policy was too loose. This fallacy comes from failing to fully understand how monetary policy influences interest rates: there are two effects, the liquidity effect and the inflation effect. The former is what people who associate low rates with loose money are thinking of--in the short run under tight enough monetary conditions, the primary effect of expanding the money supply is to lower interest rates due to the fact that lenders are more liquid, and therefore more willing to lend. But in the long run, and if monetary conditions are loose enough, the inflation effect dominates--the effect of expanding the money supply is simply an increase in the inflation rate, which in turn increases interest rates. Thus, "easy money" conditions should be characterized by a small liquidity effect and a relatively larger inflation effect--further expansion of the money supply increases interest rates, while "hard money" conditions should be characterized by a relatively long period of depressed interest rates following an increase in the money supply (note that in any conditions, the liquidity effect is temporary, in the long run printing money always raises interest rates).

So we can tell if monetary policy was tight or loose by looking at what happens to interest rates following an increase in monetary base: if there is a long period of low interest rates, then the liquidity effect dominates and money is tight; if there is an increase in interest rates not too long after the increase in base (maybe following a brief decline) then monetary policy is relatively loose.

Unfortunately the data is not quite that kind. It looks to me as though it was possible that monetary policy was too loose in the 1998-9 years, but once the 2001 recession started, monetary conditions were tight and remained tight ever since. We might conclude, then, that the Fed's response to the 2001 recession was inadequate. They managed to keep the CPI growing at their implicit target rate of 2% per year, but a disproportionate amount of that inflation was just housing, while the rest of the economy suffered from lack of demand. My hypothesis is that an increase in the CPI target would have affected non-housing prices more substantially than house prices themselves, leading to a gradual devaluation of the housing bubble relative to the rest of the economy.

I certainly do not think that making monetary conditions even tighter would have solved anything, and might actually have made the recession worse by forcing banks into even more precarious leverage ratios.