This morning the Department of Commerce reported that advanced monthly U.S. retail sales for September were $413 billion. According to the release, retail sales rose +1.1% in September month-over-month and August was revised upwards to +1.2%. On a year-over-year basis, retail sales were up +5.4% for the month and +4.8% for the quarter. At face value, this is a strong report and the quarterly growth rate is slightly above the 20-year average of +4.6%.
In the year-to-date, there have been some interesting leaders and laggards in terms of retail sales. The top three on a year-over-year basis for the first nine months of the year were: autos up +8.7%, furniture and home furnishing up +8.8%, and shopping online up +11.5%. Meanwhile, the laggards for the first nine months of the year are department stores -0.4%, electronics and appliance stores up +0.1%, and pharmacies up +1.5%. A clear take-away from the line items is that autos and home furnishings continue to outperform based on easier comps and online sales continue to take share from old line department store retailers.
A broader question is whether this retail number, which admittedly is a lagging indicator, tells us much about the trajectory of GDP growth in the U.S. In the chart below we’ve compared real GDP versus retail sales going back more than twenty years. The obvious takeaway, not surprisingly, is that they follow the same general trajectory.
The caveat is that retail sales tends to be more volatile on the upside and downside. Even at the +5.4% level for the quarter, retail sales won’t necessarily translate into an above average GDP growth rate. As an example, retail sales were up 6.3% in Q1 2012 and real GDP was only up +2.0%. In Q2 2012, retail sales fell off a cliff to +3.5% year-over-year growth rate and GDP only decelerated to +1.3%.
Further, even as there are likely some worthwhile takeaways from the line items highlighted in bold above, we would caution from reading too much into the “better than expected” headline figure. Not dissimilar to the distortion we have noted in our work in non-farm payrolls, there also appears to be a distortion in the retail sales number when adjusted for seasonality.
On a non-seasonally adjusted basis, September was actually down -$31.6 billion sequentially from August. Now obviously there is seasonality, and we understand that. The larger issue is that in last year’s delta between August and September on the non-seasonally adjusted numbers, September 2011 was only $19.0 billion less than August 2011.
We’ve summarized the differences between seasonal and non-seasonal numbers in the table below. The key takeaway is that while the headline retail sales number was strong for September, it was the beneficiary of a very meaningful seasonal adjustment, so we would caution reading further into the number. This, sadly, is consistent with much of the government data that has been released as of late.
Daryl G. Jones
Director of Research