Two weeks ago we reported that July auto sales were a disaster: recall sales for bloated with inventory GM were down 15% YoY, Ford off 7% and Chrysler down 11% - despite record incentive spending - as overall auto sales declined and disappointed for yet another month. And yet, according to this morning's retail sales report from the Census Bureau, sales for "motor vehicle & parts stores" rose much more robustly than anyone had anticipated, rising 1.2%, the fastest pace since December.
This number was so bizarre, and so out of context with recent sales data, that SouthBay Research threw up all over it in its morning note today. Here's why:
- Retail Sales m/m: 0.6%
- Retail Sales ex Autos m/m: 0.45%
- Retail Sales ex Autos & Amazon m/m: 0.3%
Consumer Retail Spending was Actually Mild, As Expected
- Auto Sales growth unbelievable
- Amazon Prime Day juiced the results
Don't believe the auto sales data. Per the BEA, unit sales were flat m/m (+90K). Meanwhile, per JD Power, July average retail prices were $950 lower than June's as auto dealers struggled to make sales and incentives averaged $3.9K, the highest on record and $100 higher than June.
- Hmmm, no rise in auto sales per the real world and the BEA. Coupled with a fall in net prices. But in fantasy land, the Census Bureau announces a $1.2B m/m jump in sales and a 7%+ y/y rise.
Amazon Prime Day Was Huge...and will Cut August Sales
- Nonstore Retail Sales jumped $700M m/m. That's the Amazon Prime Day effect. I modeled it lower and that's the source of my miss this month
Reasons for Caution: Government Data is Overstating Reality
- The Retail strength does reinforce my view that macro data favors the US in 2H and that the dollar is oversold. But the Retail headline figure is wrong and analysts were correct: consumer spending as captured by Retail is sluggish. The fact that reality is badly captured by the Retail figures is concerning insofar as it affects the Fed's decision making.
The opportunity is to recognize that consumer spending in the real world will pull back and it will also be missed by the official data. With Consensus unprepared for the pull back, it will deliver a greater shock.
Meanwhile, here's a quick look at SouthBay's proprietary "Vice Index."
For those who are unfamiliar, the vice index tracks US consumer spending on alcohol, marijuana, prostitution and gambling, Vices are a special form of discretionary spending that is highly sensitive to near-term
economic conditions: i) Cash based: depends on free cash flow; ii) Luxury spending: wants not needs; iii) Significant dollar amount: not pricey but not cheap. Vice spending is broadly representative of the US consumer: i) Broad-based: Every socioeconomic and demographic group participates; ii) High-volume transactions: Over 100M discrete events per year.
The reason why this index is of particular interest, is because vices predict retail spending with a 4-month lead. Luxury spending is the 1st thing to be affected by changes in household finances.
This is what the index shows: