September retail sales growth drops from 0.8% to 0.2% without gas stations

The gap between the two figures is 0.6 percentage points ^[15], or 75% of the headline pace ^[16]. (Growth rates on different bases do not subtract cleanly, so treat the share as an order of magnitude.) The Census Bureau puts the margin of error on the monthly change at plus or minus 0.4% ^[9], so the headline sits between 0.4% and 1.2% ^[17]. Even the low end clears the 0.3% that economists expected, according to FactSet ^[8].

The two sources do not strip the same categories. Bloomberg, as relayed by qz.com, excludes motor vehicles and gasoline and gets 0.2% ^[3]. AP excludes gas stations alone and also gets 0.2% ^[4]. Within rounding, autos moved at about the pace of everything else, and the gap between headline and core sits at gas stations ^[18].

If higher pump prices were squeezing other budgets, department stores are where it would show. They rose 1.6%, furniture and home stores rose 1.3% and restaurants rose 1.1% ^[11]. Yet sales excluding gas stations rose only 0.2% ^[4], and with several categories above that rate, the rest of the basket must have averaged below it ^[19].

We think the energy-price reading is right in direction and unmeasured in size. AP notes that the figures are not adjusted for inflation, and that inflation accelerated in September as gasoline and many other goods and services became more expensive ^[7]. The reports do not include September gas-station sales, so fuel price cannot be separated from fuel volume.

There are ways we could be wrong. Drivers may simply have bought more gallons, and that would make the gap volume instead of price. The 0.2% core is also in current dollars ^[5], so with goods prices rising it could be flat or negative in real terms. And the estimate is not final: August was revised from 1.1% to 1.3% ^[6], so September can move too.

August is the closest comparison. Gasoline prices jumped and service-station sales rose 3.1% ^[12]. Year-ahead inflation expectations rose to 4.3% ^[14], and the University of Michigan's sentiment index fell 4.2 points ^[20], to 51.0 from 55.2 ^[13]. Spending rose as sentiment fell.

The view fails if the Census category tables show September gas-station sales up only modestly, or if the ex-gas figure is revised up toward the 0.8% headline.