In my previous post, I presented evidence that the oil price increase over 2007:H2-2008:H1 made a significant contribution to the slowdown in consumption spending in general and decline in spending on domestic automobiles in particular. Here I discuss why this should be regarded as a key development that turned the slowdown in growth into a recession.
One way to put this in perspective is to compare what happened in 2007-2008 with what we saw in 1990-91. Prior to Iraq’s invasion of Kuwait in August 1990, the U.S. economy had been growing slowly due to a weak housing sector. The sudden loss in oil production from Iraq and Kuwait resulted in one of the biggest increases in oil prices on record, which unquestionably contributed to the downturn in consumption spending and the auto sector that were key factors in turning that slowdown into the recession of 1990-91. The graph above shows that although the oil price increase during 2007-08 was more gradual than what we saw in the fall of 1990, the cumulative magnitude was quite similar.
The next graph compares the path of employment in the auto sector during the two recessions. The red line shows the cumulative change in the number of workers employed in motor vehicles and parts manufacturing subsequent to July 1990. This sector had shed 94,000 jobs by the low point in March of 1991. For comparison, we’ve seen 150,000 jobs lost in this sector by October of this year. Although this is a small number relative to the total labor force, another hundred thousand jobs would have made the difference for calculations such as whether the year-over-year employment growth was positive or negative as of the end of the summer, which some might take as a criterion for declaring that a recession had begun in the first half of this year.
The decline in automobile production alone subtracted about $30 billion (in 2000 dollars) from U.S. real GDP in 1990:Q4-1991:Q1, which was enough to reduce the annual GDP growth rate by about 0.9% for each of these quarters. Autos subtracted more than $30 billion (in 2000 dollars) from U.S. real GDP in 2008:Q1-Q2, though due to a larger economy, this only subtracted about half a percent from the annual GDP growth rate for these quarters. But once again, I’d argue that autos by themselves would have been enough to make the difference between whether you characterize the first half of 2008 as slow growth or recession. The table at the right gives the actual annual growth rates for total GDP (left column) and gross domestic income (right column), which the NBER Business Cycle Dating Committee said it also looked at this time in declaring that the recession had actually started at the end of 2007. Without the slowdown in consumption in general or the strong hit to autos in particular, again I don’t think we would be characterizing these quarters as the beginning of an economic recession.
But why, some readers may ask, am I talking so much about the relatively modest contribution of autos, when housing was such an obviously bigger story? It is true that lost jobs and income from the housing sector were significantly bigger than what we saw in automobile manufacturing. But the problems in housing had been around for some time prior to 2008 without causing a recession. Between 2006:Q2 and 2007:Q3, the ongoing decline in residential fixed investment subtracted 1.04% from the average annual growth rate of U.S. real GDP. Over 2007:Q4 to 2008:Q3, housing only subtracted 0.91% on average. Something else made the difference in turning the latter episode into a recession, and it seems to me that consumption spending and automobile purchases were a key factor.
There is also an interaction effect between the spike in gasoline prices and the problems in housing, with rising gasoline prices delivering the fatal blow to many exurban communities. Joe Cortright (hat tip: MuniNet) notes for example that zip codes closest to the metropolitan center of Tampa actually saw increases in house prices, with the steepest declines in the outlying areas.
The same is true for the Los Angeles area.
And if I am correct that the oil shock contributed directly to slower GDP growth during 2007:H2 and 2008:H1, this in turn would have implications of its own for housing. For example, a recent study I did of the demand for housing (which appeared in the most recent issue of the Journal of Monetary Economics) estimated that a 1% drop in real GDP is typically associated with a 2.6% drop in new home sales. I see little basis for disputing that the oil price increase was one factor pushing home sales and house prices down.
Of course, we’ve now reached a point where defaults on home mortgages and the resulting financial chaos became a factor far bigger than both housing and autos and have put us in a hole much too deep for $1.50 gasoline to bring us back out.
How big would our current problems be if the blue line in the first graph had been flat instead of tilted so dramatically up? I don’t claim to know that. But here’s one thing of which I am very much persuaded: if gasoline prices had stayed at $2.50 a gallon through 2008, the NBER Business Cycle Dating Committee would not have declared that the current recession began in December 2007.
For those of you who will be attending the American Economic Association meeting in San Francisco, I’ll have more to say about these issues on Sunday in a session also featuring Olivier Blanchard and Lutz Kilian.