In a recent post – The End of Retail – Scott Sumner writes:
During the early stages of the recession there were all sorts of “structural” theories of unemployment. One popular theory claimed that housing had been overbuilt, and hence the job losses in home-building were due to “reallocation” of labor, not falling AD. That idea never made much sense as an overall theory of unemployment, as aggregate employment held up well during the great housing crash of January 2006 to April 2008.
The idea that there was a house building “frenzy” in the 1990s and 2000s is illustrated in Chart 1. The somewhat cyclical pattern of house starts is broken in 1990, after which starts increase almost continuously for the next fifteen years to 2005.
As Chart 2 indicates, that is consistent with the rise in the rate of population growth, which rose form 1% in the 1965 to 1989 period to 1.2% for the next 10 years. A substantial portion of this rise was due to immigration, with the number of immigrants doubling from an average of 500 hundred thousand prior to 1990 to 1 million thereafter.
Chart 3 shows that the house stock population ratio reached a ‘steady state’ in the late 1980s. It appears that the rising trend in house starts was just sufficient to maintain the house stock-population ratio constant.
Chart 4 shows that after house construction peaked in early 2006, the fall in residential construction employment was partially compensated by the continuing rise in non-residential construction. Nevertheless, the after-recession pattern of overall construction employment is very different from the pattern of private employment or even retail employment shown in Charts 5 and 6, respectively.
Chart 7 illustrates the behavior of the ‘forcing variable’ – NGDP. “Structural excuses” are just a cop-out for Fed incapacity to keep nominal spending at an adequate level! By not doing anything to offset at least part of the previous drop in nominal spending, we get the so called ‘new normal’ But as shown, that´s the responsibility of the Fed!
Chart 8 explains the dearth of residential construction employment. Residential investment has only recently begun to slowly pick up. Chart 9 helps explain why non-residential employment also languishes. Non-residential investment is mostly being determined by equipment & software investment, not by nonresidential construction.
Historinhas 
Pingback: About that housing boom…. | worldofinterest
I don’t understand how this negates the “structurak problems” hypothesis. As I understand it, the structural problems are too many resources (human, capital, etc.) were invested in housing but once the housing boom was over (for whatever reason), the economy needed time to reallocate this resources in other investments (maybe retail? which is why recovery is slow. How is this inconsistent with your charts? I must be missing some insight or is there another “structural problems” hypothesis/explanation I don’t know about?
Eric, the economy is dynamic and there is always some reallocation going on. And in the case of housing, after the peak, this was also true. Employment in residential construction was falling but was rising in other construction activity and overall. This adjustment process which is natural and occurs all the time was ‘blocked’ by bad monetary policy which reduced nominal income and nominal income expectations.
Marcus Nunes,
That third chart simply blew my mind. I almost saw it as “Housing stock/Households”. Chart 3 is that counterintuitive!
The first time I presented it people were dumbfounded! It is really counterintuitive and shows how the “bubble” argument had week foundations.
Pingback: The “there was no bubble” reading list | The Market Monetarist