Census Bureau data released today show retail sales rising by 3.0% in January, with essentially zero revision to the December sales estimate. The more widely watched “control” sales measure (described in the accompanying chart), rose by 1.7% with only a slight -0.3% revision to December. These gains were above consensus forecasts.
The retail report comes on the heels of payroll jobs and Consumer Price Index (CPI) reports that also rose faster than was expected by the markets. Call us stubborn, but we are not buying into the widely held assertion that the economy suddenly revived in January. Rather, we’ll repeat our warning of February 3 (with respect to the jobs data) that the economy always sees wild swings around the turn of the year, and government seasonal adjustment routines often fail to accurately track these.
Concerning the jobs data, we pointed out that before seasonal adjustment, private-sector jobs declined by 2.166 million, only for that to morph into a 0.446 million gain after adjustments. For retail sales, headline sales declined 16.2% prior to seasonal adjustment, but rose 3.0% after adjustment. Control sales declined 23.2% prior to adjustment, with that decline morphing into the 1.7% gain after adjustment. Once again, the signal-to-noise ratio for these reports is very low.
Last month, when sales came in very weak, we were also cautious about that swing, pointing out that seasonal shopping patterns could be evolving, with more Christmas shopping occurring in October and November, so that unadjusted sales don’t rise as much from November to December as they did previously. If that indeed was the case, then it is only logical to think that unadjusted sales wouldn’t decline as much from December to January as they did previously. In that case, seasonal adjustments—based on past data—would convert those swings into reported declines in December and increases in January. This could be all that is going on with the reported data.
Indeed, looking at the January data in Exhibit 1, the reported gains look something less than earthshaking both in raw dollar terms and after adjusting for price changes. They merely offset preceding months’ declines. We saw even larger month-to-month “chop” this time of year one and two years ago.
As large as the January sales gains appear on the surface, when one looks at individual store types, there is little sign of noteworthy gains. Vehicle dealer sales rose especially sharply, but that gain merely offsets declines seen over preceding months, leaving sales unchanged over the past one to two years. Restaurants did show a very large gain that propelled them to a new high, but sales there had flagged over the previous two months, and we would not be surprised to see much of the January gain in restaurant sales reversed—or revised away—in a month. Other store types saw only circumspect gains that mostly just reversed preceding months’ declines.
Yet one more time, we’ll bring up the adage, “one month does not make a trend,” and this is especially true during volatile seasonal months such as December and January. Government statisticians do as good a job as they can, but in an evolving world, seasonal adjustment is an imprecise art, and one would do well to be just as skeptical of today’s reported sales gains as we were of the reported sales declines a month ago.
Spontaneous combustion rarely occurs in the natural world or in real-world economies. We see no fundamental reason why the economy would suddenly revive in January, and we are skeptical of both the jobs and retail sales data pointing in that direction.