Mensch tracht, und Gott lacht

Wednesday, December 20, 2006

Deflating the PPI Inflation Report

You're all having way too much fun. Time for some truly boring economics.

On Larry Kudlow's radio show last Saturday, he reminded me that in August I wrote a couple of columns explaining how and why some economists and journalists were greatly exaggerating the reality or threat of inflation.

It should be reasonably clear by now that I was probably right about that. Yet here comes today's Associated Press report: "Wholesale prices surged in November by the largest amount in more than three decades, led by huge increases in the cost of gasoline and new cars and trucks." The comparison is made with "November 1974, back during a decade when repeated oil shocks sent inflation spiraling." That isn't history, but hysteria. In December 1974, the core PPI was up 17.4% from a year earlier, without any help from energy or food. The CPI was up 11.7% with energy excluded.

Before making such ridiculous historical comparisons, we have to at least leave out recent gyrations in energy prices, which caused the PPI to fall for 3 of the past 4 months. Aside from energy, the PPI in November was up 1.7% from the previous November which, in turn was up 1.5% from a year before. The November before that (2004), the non-energy PPI was up 2.3%.

None of these year-to-year increases is statistically much different from zero, because (1) the PPI cannot fully distinguish between changes in prices and improvements in quality and because (2) the PPI cannot fully distinguish between list prices and discounted prices. Dealer discounts are particularly important with cars and trucks, so any apparent spike in the dealer cost or consumer sticker price when new models come out is often ephemeral. It may also mean dealers are meeting demand for more luxurious vehicles with more options, which is not a true price increase.

In any case, the PPI has a terrible record for predicting consumer prices, partly because (1) the PPI is mostly goods while the CPI is half services, and because (2) global competition makes it difficult for producers or dealers to pass on higher domestic costs to consumers. A monthly change in the PPI obviously tells us nothing, or the press would have been fretting about "deflation" when the September and October PPIs came out.

Just as it was misleading to include energy in the PPI and CPI through June, when oil prices were rising, it can also be misleading to include energy prices when they fall. Aside from energy, U.S. inflation this year was and still is about as low as it has ever been in the postwar era. And surges in energy prices are typically followed by lower overall inflation in the following year, contrary to what Fed officials (who must not have looked at the data) have sometimes implied.

Don't say I didn't warn you that this would be dull. Reality often is.

1 comment:

Tom Van Dyke said...

Boring, yes, Alan, which is why the media have to invent something more interesting than truth.

On the other hand, you catching them red-handed at their mischief is damned interesting.

Quis custodiet ipsos custodes? Why, Alan Reynolds, of course.