I have to admit -- Wednesday's retail sales number has been bugging me since it was released. Part of this is because it completely disrupts my personal thesis for the economy -- that the US economy is teetering on the edge of a recession largely due to housing. No one likes to have to actually rethink their position -- or God forbid admit they were wrong. (Please read that with a dose of extreme sarcasm).
So, let's go back to a couple of categories in the retail sames number to see what the internals say and compare those number to some other readily available retail sales numbers.
Let's start with the Bank of Tokyo Mitsubihi's Chain Store Sales Report. According to this survey, the seasonally adjusted month-to-month percent change in same store sales was -1%. October's number was slightly better at a -.6% drop. The year-over-year percent change in the number was 2.1%, which was the lowest reading of 2006 except for March when the number was 2%. We run into a problem with the Census data at this point because the Census states, "Data not adjusted for seasonal variations, holiday, trading-day differences, and price changes." So, read that as a disclaimer. However, most of the other retail data released is seasonally adjusted. So we have to work with what we have to figure this situation out.
Continuing our look at the BTM numbers for specific categories of retailers, specialty apparel, general merchandise and drug stores all advanced, but not at greatly accelerated rates. Specialty apparel increased 2.4%, general merchandise increased 1.6% and drug stores increased 8.4%. None of these numbers were radically higher than any previous months. In fact, general merchandise sales decreased sharply in October and November.
The Census data says that "Building material & garden eq. & supplies dealers" increased from (in millions) $29,289 to $29,817 The BTM survey says home supply sales sales decreased 15.7%. It's also important to remember the housing market is in a slump right now. Sales are down, inventory is up. Is this a real estate market where building materials and supplies sales will increase 1.8%?
According to the census data, car sales increased from (in millions) $77,197 to $77,906. This data is not seasonally adjusted. However, according to motor intelligence, the seasonally adjusted annual rate of US car sales decreased from 16.16 million in October to 16.04 million in November. This is the lowest seasonally adjusted annual rate for car sales this year.
In November WalMart reported their November sales decreased .1%, the first drop since April 1996. Other retailers showed increases. However, Wal-Mart is far bigger than their competitors.
So, while the Census data says retail sales increased at a high rate in November, all other surveys say sales of various retail sales components decreased. There is an important difference in methodology in the other reports because they are seasonally adjusted. However, using the old Sesame Street game "One of these things is not like the other one" we come to the conclusion the Census will probably lower their numbers for November.