(Update) March poor Retail Sales explained: not adjusted for Easter!

Earlier this week the Census Bureau reported a nasty decline in March retail sales. The report was jarring precisely because private sources of data -- same store sales, mall traffic, car sales -- all had indicated firmness.

And look at what this abrupt decline has wrought: Prof. James Hamilton at econbrowser calls March's poor retail sales "a particularly discouraging development". Prof. Krugman says that "things are still getting worse."

Now let's consider this little nugget: here is the March monthly and early April weekly update from Shoppertrak:

ShopperTrak ... reported retail sales slipped 5.0 percent for the [March] period.
The 2009 calendar shift which moved the Easter holiday to April as opposed to March last year was a key contributor to the month’s poor year-over-year performance....

On a weekly level, ShopperTrak’s NRSE reported sales for the week ending April 11 increased 12.5 percent compared to last year, while weekly sales increased 11.0 percent versus the previous seven-day period ending April 4.

Yesterday I wrote how I was at a loss as to why the private and public sources are so out of whack. Does the Census Bureau's "seasonal adjustment" not include Easter being a week later into another month? If so then their jarring March number is meaningless. Probably we will have to average the March and April numbers to take account of Easter.

UPDATE: So this morning, I found the answer, a/k/a when in doubt, read the instruction manual. According to the notes in the Census Bureau's release:

Estimated measures of sampling variability are based on estimates not adjusted for seasonal variation, or holiday, or trading-day differences. Medians are based on estimates for the most recent 12 months.

So the poor March retail sales number might be completely explained by the fact that Easter was April 12 this year as opposed to March 23 last year! According to Shoppertraik, retail spending increases in the weeks immediately before Easter.

So I wonder how many analysts and economists (who get paid to do this) are going to be as surprised by April's number as they were by March's?




that is strange

or is it possible shoppertrak wants to minimize a decline?

This drives me nuts actually with government. Of course they award contracts to their lobbyist pals who then cannot deliver on the software, infrastructure....but that aside,
why can't they get accurate solid real time data?

We have the banking industry, credit scoring companies, Google, even the grocery profiling our every move, in real time. Yet when it comes to economic stats they are stuck in the 1950's.

Consumer sentiment improved.


NDD, you are on to something.


I just posted an Instapopulist. It is only a modest improvement, but it is still the highest since the onset of Black September.