ICSC and Redbook release their data along with commentary each Tuesday morning. The reports, which as seen above compare pretty closely, are based on separate samples of roughly 75 general merchandise and apparel chains. The graph below compares ICSC's final monthly reading with the Commerce Department's GAFO category (general merchandise, apparel, furniture & other and representing about 25 percent of total retail sales & food services).

Note this graph has two axes, reflecting ICSC's same-store tally on the right with its smaller percentage changes and GAFO's total sales on the left with larger percentage changes. The usefulness of the same-store measure is under debate but the idea is sound -- to gauge changes in demand against a fixed base. "Same" stores or "comparable" stores are those open at least one year, to the exclusion of new stores, acquired stores, and sometimes remodeled stores.

ICSC appears to do a respectable job predicting directional changes in unadjusted general merchandise trade data. ICSC correctly trended lower through the first half of 2004, hitting a convincing bottom, along with the GAFO data, in August of that year. ICSC did indicate phantom strength in early 2005, but then settled in for nearly a year of good calls. Note this year's Easter shift, from March to April, sharply skewed the data (which are unadjusted). As seen below, Redbook's results are very similar. Note these data are monthly; the weekly data of course are more volatile.

ICSC and Redbook offer a long history of data and their weekly commentary is often picked up in market chatter. But the chain-store niche is changing. Enter Wal-Mart. The massive general merchandise chain has nearly 4,000 domestic stores where annual sales are approaching $300 billion. And the chain is unusually transparent, issuing a same-store range each month and keeping the public closely apprised, often on a weekly basis, of success or failure.

Though May '04 stands out as a bad head fake as perhaps does August '05, Wal-Mart by itself does a pretty good job of signaling directional shifts in GAFO sales. Let's narrow the comparison further to the smaller category of general merchandise stores, where Wal-Mart's effect is greater. The general merchandise category is made up of department stores, discount stores, and club warehouses, and the data, unlike GAFO data, are included with the initial retail trade report.

But does having a handle on general merchandise offer an inside track at overall retail sales? That is, retail sales excluding cars & light trucks and gas stations. Though these are major categories, especially vehicles, they are subject to unusual month-to-month volatility, with vehicle sales skewed by the presence or absence of incentive programs and gas station sales skewed by swings in prices. The graph below compares non-auto & non-gas sales with general merchandise sales. While there are similarities, there were several times when the data offered a wrong signal, especially in 2004 and again in 2005 - so caution is needed when evaluating these data.