Behind the Numbers: PCE Inflation Update, September 2011
This update, prepared by Dallas Fed Senior Economist Jim Dolmas, provides an in-depth analysis of the latest personal consumption expenditures (PCE) inflation data. Updates will be posted monthly, following the release of the official PCE data by the Bureau of Economic Analysis. NOTE: Terms in bold are defined in the Inflation Update Glossary.
The headline PCE price index increased 2.0 percent at an annualized rate in September, a downshift from the more rapid rates of increase we saw in July and August (4.6 percent and 3.2 percent, respectively). Typically, when we see a sharp deceleration in the headline rate like this, we expect that it owes to decelerations or declines in the index’s volatile energy or food components. Not so this time, as both food and energy (especially gasoline) experienced robust price gains. In fact, of the 2.0 annualized percentage points making up the September headline rate, about 1.4 of them came from gasoline alone—very close to our prediction of 1.2 percentage points in the last Inflation Update. About 0.3 percentage points came from food.
So, whence the tamer headline number? Naturally enough, the answer lies with the index for items apart from food and energy, the core PCE price index. The core PCE index was essentially unchanged in September, declining at an annualized rate of less than 0.03 percent, which rounds to 0.0. This follows five months of readings at or above annualized 2 percent (2.4 percent, on average). A sharp decline in core goods prices, together with a very modest rate of increase for core services prices, combined to produce September’s negligible change.
The trimmed mean also points to a softer underlying rate of inflation in September, though the difference between September and the recent past is less pronounced in the much smoother trimmed mean. September’s trimmed mean rate came in at an annualized 1.6 percent, compared with an average annualized rate of 2.1 percent over the prior five months.
The 12-month trimmed mean rate inched up to 1.8 percent from 1.7 percent last month, while the six-month trimmed mean rate held steady at 2.1 for the third month in a row. Note that revisions by the BEA to the PCE data for July and August raised the six-month trimmed mean rates for those months from 2.0 percent to their current 2.1 percent.
The 12-month core rate ticked down to 1.6 percent from 1.7 percent a month earlier, while the six-month core rate dropped to 2.0 percent from 2.3 percent. The 12-month headline rate held steady at 2.9 percent, while its six-month rate dropped to 2.4 percent from 2.9 percent in August.
September also saw a jump in the fraction of items in the PCE basket experiencing price declines, from the neighborhood of 30 percent, which we’ve observed since the end of 2010, to roughly 43 percent. We never want to make too much of one month’s data, though this number will bear watching over the next few months.
In sum, September was definitely a softer inflation month, compared with the last several. But, with the trimmed mean’s six- and 12-month rates at 2.1 and 1.8 percent—and fairly stable—we see no reason (yet) to depart from our conclusion of the past several months: that the underlying trend in PCE inflation remains around 2.0 percent.
Gasoline Prices on Track to Subtract from Next Month’s Headline Rate
We noted above the large contribution of gasoline prices to September’s headline rate. The PCE index for gasoline increased 3 percent (not annualized) in September, about double the jump we saw in August. Over the three months of July, August and September, the gasoline index rose nearly 10 percent.
Looking ahead to October’s release, we should finally see some reversal of that pattern. Weekly retail price data collected by the Department of Energy show October gasoline prices on track to be down about 4.4 percent from September. Taking account of the normal seasonal variation in gasoline prices—October usually brings a 1.2 percent decline—we can expect that seasonally adjusted prices (which are the ones relevant for the PCE index) will show a 3.2 percent decline in October.
Given gasoline’s weight in the PCE, that 3.2 percent decline would swing gasoline’s contribution to the headline rate from the +1.4 annualized percentage points we had in September to –1.4 annualized percentage points. Barring outsized positive contributions from other components, that swing points to an October headline rate of close to zero.
Food Prices: Another Helping of the Same
Regarding food, we’re tempted to just say, “See last month’s Inflation Update,” and leave it at that—the numbers for September are roughly similar to the numbers for August.
The price index for food as a whole increased at a 6.5 percent annualized rate in September, compared with a 7.5 percent rate in August. Prices for less-processed items increased at a 9 percent annualized rate, while prices for more-processed items increased at a 5.6 percent rate. The latter, as we’ve argued before, should be more indicative of the underlying trend in food price inflation. That trend clearly remains robust.
Over the last six months, our more-processed index increased at an average annualized rate of 5.4 percent. Over the same period, the index for all food items is up at an average 4.9 percent rate, and the index for less-processed items is up at a 3.6 percent rate.
What Core Goods Giveth, Core Goods Taketh Away
We’ve often noted that large movements in core goods prices tend not to be durable—big increases or decreases generally get reversed in short order. The string of large price increases for apparel (four months) and, earlier, new motor vehicles (five months) lasted longer than we expected them to. Vehicle prices, if not reversing course, at least were essentially unchanged in August and now again in September. Apparel prices, which increased 6.1 percent from April to August (not annualized), tumbled in September, falling 1.4 percent. Women’s and girls’ clothing alone subtracted about 0.5 annualized percentage points from September’s core PCE rate.
Flat vehicle prices and sharply falling apparel prices, together with outsized declines in a few other goods categories—jewelry and the “games, toys and hobbies” component—combined to produce a hefty 3.1 percent annualized drop in the price index for core goods, its largest decline since November 2006. (April 2010 and December 2008 also saw declines almost as large, each at –3.0 percent.)
Core Service Inflation Slower in September
The decline in core goods prices would not have yielded a zero core rate in September if core services prices had increased at a rate more in line with their average pace over the last several months—annualized 2.1 percent at a six-month horizon.
But, September saw only a modest increase in the price index for core services, which increased at an annualized rate of 1.1 percent. As was the case in August, the price index for commercial bank services—in particular, those services provided without explicit fees—made a significant negative contribution, subtracting about 0.3 annualized percentage points from September’s core rate. The price index for final consumption expenditures of nonprofit institutions serving households shaved off another 0.5 percentage points.
Other important factors in the deceleration of core services prices were rent and owners’ equivalent rent (OER), both of which slowed somewhat in September, as compared with August. Rent posted a 2.8 percent annualized rate of increase in September, compared with a 4.6 percent rate in August. OER slowed to a 1.4 percent rate from a 2.6 percent rate in August.
A Jump in the Number of Falling-Price Components
Just when we thought it was almost safe to stop referencing the fraction of components in our PCE basket experiencing price declines, that fraction jumped, as noted above, to 43 percent in September.
Regular readers will recall that values around 40 percent—which is unusually high—were recorded over much of 2009 and 2010, a period when our measures of underlying PCE inflation drifted down to near all-time lows. Beginning around the start of this year, the fraction of falling-price components fell back to a much more normal range of around 30 percent. In fact, we had seen a long enough string of normal values that we were seriously considering discontinuing references to that number in the Inflation Update.
As always, though, one month’s data do not make (or break) a trend. It will, however, postpone for the time being any retirement of the falling-price components part of the update.
October 28, 2011