US Retail Sales ‘Control Group’ Unexpectedly Tumbled In October After Huge ‘Seasonal Adjustment’

If the omniscient chaps at BofA are right, the soft landing narrative is about to crash as they forecast a well below consensus tumble in retail sales this morning…

Source: BofA

…but notably this is mainly due to seasonals, apparently. Seasonal adjustments are likely to impact the October retail sales report. The Census Bureau’s projected seasonal factor (SF) for October 2024 is considerably less favorable than the October 2023 SF. Meanwhile, the October SF in the BAC card data is little changed from last year.

Source: BofA

So, given all that, what happened!?

Well, a lot!

The headline retail sales print beatexpectations rising 0.4% (+0.3% exp).

Core retail sales (ex-autos and gas) disappointed, rising just 0.1% (+0.3% exp).

But the Control Group – which is used for GDP calculations – tumbled0.1% MoM (+0.3% exp).

BUT – the disappointments are likely driven by major upward revisions to the prior month

Headline September revised up from +0.4% MoM to +0.8% MoM.

Core revised up from +0.7% to +1.2% MoM.

Control Group revised up from +0.7% MoM to +1.2% MoM.

Source: Bloomberg

The driver of the headline beat was all cars and food…

Core retail sales growth slowed on a YoY basis…

Source: Bloomberg

On a non-seasonally-adjusted basis, retail sales jumped 6.8% MoM…

Source: Bloomberg

A big outlier of an adjustment for an October… it’s almost as if they wanted to make the numbers look bad on purpose…

This is odd: Census Bureau used a surprisingly weak – and a big negative outlier – seasonal adjustment factor for retail sales. If comparable adj factor had been used as prior Octobers, retail sales would have been a 1% increase pic.twitter.com/TbtKyyNQop

— zerohedge (@zerohedge) November 15, 2024

The Control Group data is ‘noisy’ – September was revised up to its strongest MoM jump since Jan 2023 (but this was the second monthly decline in nominal retail sales in three months)…

Source: Bloomberg

A real Goldilocks of a data set there – hot headline, cool core, ugly control group – take your pick, dove or hawk!