Texas Business reports: The Dallas Fed’s trimmed mean inflation rate for June was an annualized 1.4 percent, according to Jim Dolmas of the Federal Reserve Bank of Dallas. This follows back-to-back readings of 1.5 percent in April and May. The average rate for April, May and June—just under an annualized 1.5 percent—is the lowest for a three-month period since the three months from November 2010 to January 2011.
Looking over a bit longer period, the six-month trimmed mean rate was an annualized 1.7 percent in June, down from 1.9 percent in May. The index’s 12-month rate held steady at 1.9 percent for a third consecutive month.
The release of June PCE data always incorporates the results of the BEA’s annual revision to the national income and product accounts (NIPA). These annual revisions incorporate more complete or more reliable source data as well as occasional changes in the BEA’s methodology. The revisions—typically covering the previous three years—affect many of the underlying data series that are the grist for our calculation of the trimmed mean inflation rate.
This year’s annual revision—discussed in more detail below—affected data going back to 2009. As far as the trimmed mean is concerned, the main impact of the revision is to raise somewhat the inflation rate over mid- to late-2011. The six-month trimmed mean inflation rate recorded for August of 2011 had been an annualized 2 percent; it’s now 2.2 percent. That’s a small change, to be sure, but—in combination with the most recent reading of 1.7 percent—it does give the six-month rate a noticeable downward trajectory. Prior to the revision and the latest release, the six-month trimmed mean rate had looked fairly flat.
The headline PCE price index increased at a 1.3 percent annualized rate in June, following a decline of an annualized 2.2 percent rate in May. Prices of energy goods and services fell at a nearly 18 percent annualized rate in June, following an even larger decline in May. Gasoline, fuel oil and electricity all experienced substantial declines. Gasoline alone subtracted about a full annualized percentage point off June’s headline rate.
The six-month headline rate ticked up to an annualized 1.5 percent from 1.4 percent in May. The 12-month headline rate held steady at 1.5 percent.
Prices for Energy Goods and Services at 16-Month Low
The PCE price index for energy goods and services—which include gasoline and other motor fuel, fuel oil, electricity services and natural gas services—has declined nearly 8 percent (not annualized) over the three months of April, May and June, effectively erasing all the gains in that price index from March 2011 to March 2012.
Among energy goods and services, gasoline typically has the largest impact on the headline inflation rate for any given month, owing to its unique combination of high price volatility and large expenditure weight. (Fuel oil, for example, often experiences much larger price swings, but its weight in expenditure is very small compared with that of gasoline. Consumer spending on fuel oil is about 0.2 percent of PCE, versus a 3.5 percent share for gasoline.)
In June, the PCE price index for gasoline and other motor fuel fell 2 percent (or 22 percent at an annualized rate)—about exactly as we guessed it would in last month’s Inflation Update. As anyone who regularly buys gasoline knows, though, prices bottomed out sometime after the first week of July and have since been creeping back up. Nevertheless, the average price for the month of July—using weekly retail price data from the Department of Energy—remains about 2.7 percent below the average for June.
As it turns out, a 2.7 percent decline is very close to the typical seasonal decline in July (about 3 percent). One can thus expect—when PCE data for July are released—a seasonally adjusted change in the price index for gasoline of just about +0.3 percent. For gasoline, that’s a negligible change and would have a negligible impact on next month’s headline rate.
Food Price Inflation Set to Rebound; Impact on Headline Rate Should Be Minimal
Food prices jumped in June, increasing 3 percent at an annualized rate. Gains had averaged about 1 percent, annualized, over the prior six months.
June’s increase was driven entirely by the prices of less-processed food items, which increased at an 11 percent annualized rate. Prices for fresh vegetables had the biggest impact, rising at a roughly 45 percent annualized rate and contributing about 0.2 annualized percentage points to June’s headline inflation rate. In contrast, the index of prices for more-processed food items was essentially unchanged in June.
In spite of June’s increase in the overall and less-processed indexes, over the past six months all our price indexes—all food items, less-processed and more-processed—are up at a modest 1 percent annualized rate. One year ago, the six-month rate of increase in price for all food items had reached an annualized 6.3 percent.
Consumers may be in for another round of rapidly increasing food prices, though, as the spreading of drought conditions in the Midwest has led to expectations for sharply reduced yields of a number of crops, including feed corn and soybeans. Economists at the U.S. Department of Agriculture expect the drought’s impact on prices to be felt mostly in 2013, though for some items—poultry, for example—a larger impact in 2012 is forecast.
While painful for consumers, a pickup in the rate of food price inflation is unlikely to have a significant impact on headline PCE inflation. Just as an illustration of this point, over the past 12 months the PCE price index for all food items has increased 2.4 percent while the headline PCE price index has increased 1.5 percent. If food price inflation over the past 12 months had instead been 10 percent—keeping everything else the same—the 12-month headline inflation rate would have been higher by just 0.3 percentage points.
Increase in Core Goods Prices Includes Rare Gain for Computer Prices
The prices of core goods increased at a 1.4 percent annualized rate in June and are up 0.8 percent, annualized, for the six months ending in June. The biggest-impact items among core goods increasing in price were jewelry (up 26 percent at an annualized rate), prescription drugs (up 4.5 percent), computers and peripheral equipment (up 17 percent), and footwear (up 14 percent). Jewelry and prescription drugs together contributed about a quarter of an annualized percentage point to June’s headline inflation rate.
The truly odd item in that group, though, is computers and peripheral equipment. The 17 percent annualized increase follows an 11 percent annualized increase in May. Computer prices in the PCE are adjusted for improvements in quality and almost always decline—over the past 10 years they’ve averaged an annualized rate of decline of 12.5 percent. So increases are rather infrequent, back-to-back increases are rare and back-to-back double-digit increases are extremely rare.
At the opposite end of the spectrum—looking at price declines—used autos stand out for their outsized decline in price (13 percent annualized), though the component’s impact on the June headline rate was less than a tenth of an annualized percentage point. (Televisions and “computer accessories and software” were also in the bottom tail of the price distribution, but their behavior was nothing out of the ordinary. Like computers, their prices are quality adjusted and almost always experience large declines.)
Financial Components Lead Core Services Prices Higher
Among core services, most of the outsized, large-impact price movements were increases. Two components of financial services had the largest impact, among all PCE components, on the headline inflation rate. One is the imputed price of services provided without explicit fees by “other depository institutions and regulated investment companies,” which increased at a 23 percent annualized rate in June. The other is “financial service charges, fees and commissions,” which increased at a roughly 11 percent annualized rate. Together, they accounted for about half an annualized percentage point of June’s headline rate.
The price index for “other purchased meals”—basically, dining out—and the price of hotel and motel services also made noticeable positive contributions. Only the price of airline services—always volatile—showed an outsized decline and a noticeable (though still quite small) impact on the headline rate.
The two large shelter components within core services, rent and owners’ equivalent rent (OER), continue to post only modest increases. OER grew at a 1.2 percent annualized rate in June, after posting a 1.1 percent rate in May. For the first six months of 2012, OER has increased at an annualized rate of 1.8 percent, compared with a 2.3 percent rate over the last six months of 2011.
Rent recorded an annualized 1.5 percent increase in June, it slowest one-month rate since April 2011. The slowing in rent growth between the second half of 2011 and the first half of 2012 has been more dramatic than the slowing in the growth of OER. Over the last six months of 2011, rent averaged an annualized 3.2 percent rate of increase; over the first six months of 2012, rent has averaged a 2.2 percent rate.
Digging Deeper into the Annual Revision
What accounts for the effect of BEA’s annual revision on the trimmed mean? Revisions incorporate better data on both prices and quantities (the amounts of various goods and services households consume). In principle, either could impact the trimmed mean, since both price changes and expenditure patterns enter into the calculations.
In practice, though, the revisions to prices are what matter the most, at least for this year’s annual revision. How does one know this? Given both pre- and postrevision data for 178 underlying components, one can combine prerevision prices and postrevision quantities to calculate a trimmed mean inflation rate. That experiment—which corresponds to just revising the data on quantities—produces a trimmed mean that looks almost identical to the prerevision trimmed mean. This tells one that BEA’s revision to the quantities data had a negligible impact on the trimmed mean.
(One could also combine postrevision prices and prerevision quantities and calculate the trimmed mean using those data. As one might guess, that experiment—which corresponds to just revising prices—produces a trimmed mean that looks almost identical to the postrevision trimmed mean.)
Among all the price revisions, the ones that had the largest impact on the trimmed mean include rent and a number of medical care services. These were revised to show slightly more rapid price increases over the second half of 2011 and together go a long way toward accounting for the upward revision to the trimmed mean over that period.
Finally, one should note that while PCE is, of course, the primary focus, it’s only a small part of BEA’s responsibility. The annual revision updated not only PCE data, but also BEA’s estimates of gross domestic product, national income, and their various components. An overview of the annual revision can be found on the BEA website.
Next year’s annual revision, incidentally, is a “comprehensive” revision, which opens the books (potentially) on data going all the way back to 1929.