Wednesday, July 27, 2011

Autos in the US Economy

Yesterday, I did a simple rough calculation about General Motors to try to get at how much manufacturing job loss one could attribute to foreign competition versus automation and productivity increases. Various commenters objected that my assumptions were too simplistic, particularly in neglecting the domestic content in foreign autos and the importance of shifts in the supply chain over time.

To try to address these concerns, I'm going to repeat the exercise using national statistics for the components of GDP and employment by industry.  The GDP statistics are supposed to keep track of the value-added via supply chains and correctly account for foreign versus domestic components of production.  For time reasons I'm going to have to spread this out over a couple of posts. Today, I look at autos as a fraction of GDP from 1967 to 2010. In particular from BEA Table 1.5.5 I took the total US GDP and also US consumption of "Motor Vehicles and Parts". Then from Table 4.2.5 I took imports and exports of "Automotive vehicles, engines, and parts" which I can also divide by US GDP. Then, by adding the exports to the consumption and subtracting the imports, I can get an estimate of domestic manufacture.

All of these are shown in this graph:

You can see the rise in imports from much less than 1% in the 1960s to about 2% of US GDP in the mid 2000s (until the great recession hit). The other interesting factor is that consumption of autos (as a fraction of the economy) started a serious decline in the early 2000s. That's going to complicate the analysis which I will take up again in the next post.

But it seems an interesting observation in its own right - the triple combination of the early 2000s recession, the oil shock of the mid 2000s, and then the great recession, have combined to lower the share of auto consumption in the US economy by about a third over the last decade.


Nick G said...

I believe car sales plateaued in 1973. It's interesting to see that the car industry managed to increase the "value" per car sufficiently from 1973-2003 to mostly maintain it's share of GDP.

OTOH, a mature market can only keep that up for so long, by adding features, etc. Car reliability/longevity has increased dramatically since 1973, and the overall inventory of light vehicles has been increasing, with older ones staying around but getting much less use. I'd say that the car market was seriously oversupplied with new and used cars by 2003. The decline that started at that point was overdue.

porsena said...

The odd-looking line in this graph is the one for domestic consumption of autos and parts as a fraction of GDP; the line starts to drop after 2002, only reversing in 2010. It's odd to me because a graph of US vehicle sales is essentially flat from 2001 to 2007 at between 16 and 17 million vehicles per year. Reading off Stuart's graph above, the gross consumption line starts at about 3.7% of GDP in 2001 and drops to around 2.8% in 2007, so roughly a 25% drop. Looking at BEA Table 1.1.5. US GDP rose 36% during that period.

My interpretation is that the GDP-normalized graph of sales of autos and parts has more information about US GDP from 2001-2007 than it has about the auto industry. Do I remember from somewhere that the statistics for US GDP include the financial sector, which was waving its magician's wand vigorously during this period?

Anonymous said...


Complicating your exercise even more is the fact that people tend to keep their cars longer nowadays than they used to. This is in large part because quality has gone up, reducing the frequency needed to buy a new car. I suspect this has far more to do with the declining share of GDP dedicated to auto purchases than any of the other factors you mentioned.