Evan Soltas
Apr 3, 2012

Housing Bust ≠ Recession

The collapse of residential investment doesn't explain gross investment volatility

Yesterday, I thought about how investment causes recessions more often than consumption, how the huge swings in investment explain why the 2008 recession was so large, if not why it happened entirely, and therefore why monetary policy needs to place more weight on real stability in the future.

One more thing I want to add on that topic before we move on to Part II: investment matters. I mean that in the sense that, unlike consumption, there's a big opportunity cost to not investing, because of the effect of low investment on the capital stock, and thus productivity and future output and income.

Today's post works from a lot of the conclusions from yesterday, so I encourage you to give that a read before this. Yesterday, I looked at real gross domestic private investment in the aggregate. Macroeconomics loves aggregates, sure -- but perhaps a bit too much. I've found that some of the most effective research into explaining the behavior of aggregates actually goes one level down, into components.

So here we are. Most of us -- again I'm going to raise my hand here on the "misconception train" -- would endorse the view without further thought that the boom and bust in housing bears a lot of the responsibility for the recession, as well as its depth and length.

That's not really accurate, however.While we can see in this chart that residential investment -- the height of the gap between gross and gross nonresidential investment -- has fallen, the severe swing in 2008-9 that we identified yesterday was entirely in nonresidential components. In fact, residential investment had continued to shrink at the rate it had since mid-2006. We can't look at housing as an explanation for the volatility in investment, and thus the severity of the 2008 recession.

(Note: This last comment ignores the consideration of housing prices on real wealth, and through wealth effects onto consumption. I agree there are indirect pathways, but housing investment wasn't as important as most people think it was.)

What components of investment fell hard, then? First, nonresidential fixed investment -- "fixed" refers to the fact that the investment comes in a tangible form, like a factory. Second, other forms of nonresidential investment, such as equipment and software, fell sharply. Third, inventory change was also cyclical.Looking at index values from the peak of real output in Q4 2007, we see that residential investment is completely out of sync with the rest of the components of investment, as well as gross investment itself. Instead, gross investment looks much more like the nonresidential components.

This isn't an issue of scaling the components by their relative sizes, either. Residential investment is not a trivial fraction of gross investment -- it has varied between 36 and 17 percent -- but that leaves a great deal of other stuff going on, a great number of investment decisions in which monetary policy is directly relevant via interest rates, and indirectly relevant via real stability.

If we look at the raw numbers of the components of investment, we can see this is true: It seems evident that from 2006 to 2008, the Fed's policy was effectively allowing residential investment to shrink as a fraction of gross investment, without leading to headline losses. It kept real investment stable by maintaining nonresidential investment growth. And then everything changed in 2008 -- by which point, importantly, the reweighting process was almost entirely finished, that is, residential investment's scaling back was all but over. Suddenly, the Fed let real investment fall as monetary policy stopped accommodating and supporting nonresidential components. (Housing, true to the independence with which it had been moving, actually didn't change its trend amid the volatility in all other components of investment!)

The best way to look at this process by which residential investment was "decoupled" from the rest of investment is by looking at a correlation of the percent changes in the two variables.This is an amazing graph, if I can say that about my own work. The correlation weakened in 2006 through 2008, as residential investment declined amid rising gross investment. Then, true to our observations, gross investment behavior changes sharply, entirely due to the behavior of nonresidential investment. Since both have now begun declining, the correlation rises sharply in 2008, amid the collapse and crisis.