Some people use the end of December as an opportunity for a retrospective on the year. But I decided to take a look back at the last three years, by way of updating some comparisons I made in April 2009 between the Great Recession and the average characteristics of other postwar recessions.
My inspiration at the time came from Bill McBride (as my better inspirations often do). Bill had commented on a typical pattern in recessions that I summarized using the figure below.
Average cumulative change in 100 times the natural log of real GDP or its respective component beginning from the business cycle peak for the 10 recessions between 1947 and 2001. Horizontal axis denotes quarters after the peak.
The horizontal axis indicates the number of quarters since the start of the recession. The vertical axis gives the average percentage change (measured logarithmically) in real GDP or its components from the value at the business cycle peak, with the average calculated across the 10 U.S. recessions between 1947 and 2001. GDP reached a low point on average in the third quarter of the recession, at a value 1.6% below the peak, and was back to its pre-recession value after 5 quarters. Consumption spending tended to be more stable, often showing no dip at all. Investment spending has always been one of the more volatile components of real GDP, with housing construction down 8% at the trough, but typically rebounding to make a positive contribution a year after the recession began. Business purchases of structures and equipment were on average still 7% below peak even after GDP had returned to pre-recession levels.
The corresponding magnitudes are reported for the most recent recession in the figure below. Note that a change in scale is necessary for both the horizontal and vertical axis in order to accommodate the fact that the last recession was significantly longer and deeper than normal. Real GDP was down more than 4% at the low point and had still not returned to its pre-recession value after 2-3/4 years, though presumably the 2010:Q4 numbers to be reported at the end of January will finally put us back to a new high. The substantial and prolonged decline in overall consumption spending was also unusual. But the real outlier was housing, which was still making a new low 11 quarters after the recession started at a value almost 50% (logarithmically) below its level when the recession started.
Cumulative change in 100 times the natural log of real GDP or its respective component from the most recent business cycle peak. Horizontal axis denotes quarters after 2007:Q4.
This is not to say that housing alone was responsible for the recession. Although the drop in percentage terms has been pretty spectacular, new home construction had been a relatively modest share of total spending to begin with. Home construction (quoted at a real annual rate) fell by $190 billion between 2007:Q4 and 2009:Q2, whereas real GDP fell by $554 billion.
With home sales bumping along the lowest levels on record, it’s hard to believe they could keep going lower. And any improvement, even to almost-but-not-quite record lows, would allow this sector to start making a positive contribution to GDP growth.
Source: Calculated Risk
Of course, that’s also what I thought last summer.
But fortunately, this time I see that Bill McBride agrees with me.
Quarterly percent change (at an annual rate) of real GDP, 1947:Q2 to 2010:Q3, and three of its components. Data source: BEA Table 1.1.3. Shaded areas denote NBER recession dates.
Posted by James Hamilton at December 29, 2010 05:58 AM