A Contained Downturn?

The headlines tell us that the U.S. is going through a rough patch which could last much longer than economists had previously thought.

But official government data suggests that the weakness in the economy has been mostly felt only in residential real estate investment and automobiles. The first chart below shows the average contribution of real estate and autos to economic growth for inflation-adjusted G.D.P. for all recessions since 1945 compared with the current cycle. The second chart shows the same thing, except for all other components of G.D.P. The typical recession lasts for four quarters, as indicated by the gray region in each chart.

It looks like the current cycle for real estate and autos is roughly inline with past recessions, except that real estate has been subtracting from G.D.P. for a much longer period than average. Meanwhile, all other components of G.D.P. have been adding to growth, which is inconsistent with past recessions. (This is assuming a current recession started in the fourth quarter of 2007.)

So what accounts for the discrepancy? It's likely to be in the way G.D.P. is measured. The hot debate on econ blogs since last week's Q2 G.D.P. release is that the government's measure of inflation used to calculate growth -- the G.D.P. deflator -- is understating real inflation, and overstating growth (which came in at 1.9 percent).

The problem with calculating the G.D.P. deflator, says Wrightson-ICAP economist Lou Crandall, is that when things are moving rapidly, as energy prices recently have, getting the timing of these effects right is a huge challenge for government statisticians.

Take oil imports as an example. The government tries to estimate the price of crude oil imports at many different points in the process: when they first cross the border, when they show up in inventories, when they are turned into refined energy products, and when those refined products sold to businesses as an intermediate product and to consumers as a final product. In many cases, BEA must try to apply price estimates collected at one point in the month to product flows spread over the month as a whole, or to inventory snapshots collected at the end of the month. The inventory data themselves must be massaged to correct for the various accounting methods used by different companies.

The upshot is that, over time, these discrepancies even out. What that means for current conditions is that the downturn is probably not as contained as the official numbers suggest -- and even though recent lower energy prices will probably help the economy -- the biases in how the G.D.P. deflator is measured will make growth appear lower in the next couple of quarters.