In a post today, Matt Yglesias borrows Brad deLong's graph to show that "the back-and-forth swing of business investment has been the main motor of the recession...But the chart also shows us that contrary to a lot of myth-making, the recession is not identical to the downturn in housing." Nevertheless, Yglesias says, the lack of recovery in housing has kept the level of real output below what would be consistent with full employment.
This is a very important point, and I would like to develop Yglesias' argument a bit further. I find that economists, journalists, and other observers assign excessive blame to the housing bubble as the cause of the recession.
The "housing story" seems, at face value, to work. But the support in the data is weaker than the proponents of the housing story admit for two reasons. Both are captured in the graph I've produced below.(1) Private residential fixed investment peaked in 2006, whereas the rate of growth in real output did not begin to decline sharply until 2008. It is certainly possible to draw a causal connection between these two events, but the gap of two years should raise questions. Why did macroeconomic conditions seem able to withstand significant contraction in residential investment for two years, and then suddenly it could not? In other words, why were the structural problems not a problem until they were a problem? (Here is another post which reveals the apparent decoupling of residential fixed investment from the rest of the economy.) Structural problems "lurking" under the surface make for a good movie or novel -- but no so much for cogent economic reasoning. The "housing story" needs a more compelling account of how structural problems spent two years in transmission.
(2) The direct contribution of residential construction to the rate of real output growth never exceeded +/- 1 percent from 2000 to 2012. It is possible that the indirect effects of housing on real output dwarfed the direct impact, but the disproportionate difference should also raise questions. In essence, the housing story advocates lack an adequate explanation of their macro model, and in particular, how they explain this problem of proportions.
Beyond a reasonable doubt, the problems in housing were a contributory cause of the recession. But the case that they were a sufficient cause of the recession is far weaker than its advocates recognize.
To make housing their sufficient cause of the recession, they often end up incorporating in other causes through the back door, as it were. They invoke longer causal chains and transmission mechanisms -- the financial system, wealth effects, expectations for nominal income growth, unemployment, etc. -- and though they are not incorrect in doing so, they fail to appreciate that they've diluted their own argument. When problems in housing generate illiquidity problems in the financial system and reduce household wealth and income, the housing story is no longer just about housing. It's about macroeconomic conditions; it becomes a nominal story.
The evidence is consistent with housing as a contributory cause to a nominal story, in which monetary policy is passively tightened in 2008 as expectations of nominal income growth collapse, with housing exacerbating the depth and persistence of the decline in real output due to wealth effects and the contraction of credit.