Connect the Retail Sales Dots

The usually astute WSJ blows a front page story on yesterday’s retail sales data. Actually, you can asign 60% of the blame to whoever came up with the headline, leaving the writer only 40% blameworthy — he gets many (but by no means all) of the details correct.

"Consumers went on a postholiday shopping spree in January, a strong sign of economic vigor that increases the likelihood the Federal Reserve will keep raising short-term interest rates.

The Commerce Department said yesterday that retail sales surged a seasonally adjusted 2.3% in January from December, largely because of gift-card redemptions and abnormally mild winter weather. The January jump followed a tepid 0.4% rise in December. January sales were up 8.8% from a year earlier." (emphasis added)

That is dead on. Recall my prior perspectives on the holiday retail season as so very dependent on gas prices. (Although December 2005 was revised downwards) Fuel costs have moderated, and we had a fair (not great) holiday season, right within my 3-4% range. 

However, the writer goes astray here:

"Coming after an earlier report that employers added nearly 200,000 jobs in January and pushed the unemployment rate down to 4.7%, the retail-sales report was cheered as evidence that the economy has roared back from a fourth-quarter lull."

My divergement from the standard reporting is based on their fundamental misunderstanding of how BLS reports data, and what it actually means: Citing the the January sales data as a sign of hiring is simply a fundamental misunderstanding of what seasonal adjustments are. In actual numbers, the economy lost 2.625m jobs in January. When the BLS runs the data through their meat grinder, they seasonally adjust them them to a +193k. I have no problem with that, as it makes sense for BLS to present the seasonal adjustments this way. (You can see more details on the NFP and Unemployment data here). 

However, financial reporters should be savvier, and realize that 193k jobs were no more created in January then were Enrons’ profits created in Q1 2000.  Its an accounting adjustment (Only Enron’s were fraudulent, while the BLS data is real — just seasonally adjusted). Further, the unemployment rate drop has been fully explained as a drop in the Labor Participation Rate, and note an actual decrease in the number of people looking for work.

But note how putting those two adjusted data points together creates a very false read of the strength of the economy. Its an awesome sucker play, and helps set up the eventual top.

Back to the Journal:

"Excluding the 2.9% increase in sales of motor vehicles and parts and a 5.5% rise in gasoline-station sales driven by higher pump prices, the remainder of January retail sales — everything from department stores to bars — were up 1.8% from December, when such sales rose just 0.3%."

Gas sales are hardly signs of economic strength, nor is Detroit giving away cars at cost. Month over month, the gain was less than 2% — easily explained by a big gift-card season, and the abnormally mild winter weather.

Further, the WSJ’s Justin Lahart counsels us to consider the seasonal adjustments to gift cards. That’s right, not unlike NFP for January, the seasonal adjustments made in January exaggerate the actual data. (Note that this adjustment predates the rise of gift cards as a Janaury phenomenon:

"The gift-card effect may be exaggerated by "seasonal adjustments" made by the Commerce Department. These adjustments, meant to smooth out monthly sales data, assume January’s sales will be paltry next to December’s. As a result, a small bump to January’s sales level could be magnified in the seasonally adjusted numbers that get reported."

Surprisingly, the person counseling the most restraint about these numbers is Rosalind Wells, chief economist at the National Retail Federation:

"Ms. Wells of the retail federation and other economists cautioned that two primary drivers of January’s sales growth — balmy weather and gift-card purchases — will likely evaporate this month, particularly after a blizzard last weekend snowed in consumers throughout the Northeast. Discount chain Target Corp. yesterday said colder weather will hurt February sales and that it now expects sales of stores open more than a year to climb 2.5% to 3.5% in February, down from its previous prediction of 2.5% to 4.5% growth.

The January bounce was "a one-month shot," Ms. Wells said. "We’re likely to see weak February sales because of the blizzards in the Northeast, and then generally I think the economy is going to slow as the year goes on."

Take this Retail Sales data, tie it into the recent Account Deficit numbers, ands its not to hard to figure out why the Savings Rate is null.

Obviously, I am talking my own beliefs — but I am also trying to counter the widespread misread of the data.


UPDATE: February 15 2006 9:47am

 The NYT is no better:

But few analysts now fear that consumer spending will fall off a cliff. Job
gains and wage growth, typically the leading factor in consumer behavior, appear
to be on an upward trajectory after their post-Katrina sluggishness of last
fall. The economy added 193,000 jobs in January; wages rose 0.4 percent; and the
unemployment rate dropped to 4.7 percent, its lowest level in four and a half

"The deep fundamentals, the personal income numbers, are healthy," said
Gregory Miller, chief economist at SunTrust Bank in Atlanta.

With growth picking up again, economists now expect the Federal Reserve to
increase its benchmark short-term interest rate, currently at 4.5 percent, at
least once but possibly twice more over the next few months before it brings its
tightening cycle to an close. They will be paying particularly close attention
today and tomorrow to the Fed’s new chairman, Ben
S. Bernanke
, who is scheduled to testify to Congress for the first time
since replacing Alan


Retail Sales Surge 2.3%, Underlining Economy’s Health
Data Help Dow Industrials Jump 1.3% to Top 11000; Another Rate Increase Likely
WSJ, February 15, 2006; Page A1

Buying Power
Justin Lahart
WSJ, February 14, 2006; Page C1

U.S. Trade Deficit Ballooned To a Record in 2005
February 11, 2006; Page A1

 What Fuel Bills? U.S. Consumers Still Spending
NYT, February 15, 2006

Print Friendly, PDF & Email

What's been said:

Discussions found on the web:
  1. B commented on Feb 15

    Doggone it! Why didn’t I get a gift card to spend frivolously? Must be body odor.

    I still believe a correction is a resetting of expectations and a purging of the excesses rather than the consumer going broke. May come to pass but not yet. Btw, I don’t think I have seen a single article talking about the fact that our savings rate could go negative and stay negative for a long time at some point. Will that be when enough people are in retirement and are not working any more or are only working to supplement their savings and, thus, not saving anything or little. One of those long wave economic variables that changes forever? I’m not an economist but I suspect if the models were changed to take this into account, it could turn out to be true. Take away the youth, take away the massive retiree population and you have 25%, or whatever the number is, of the population capable of significantly contributing to savings.

    I’m beginning to wonder if the correction scenario that may arise is more along the lines of the economy has been goosed so much that the data may convince the Fed to raise well beyond blah blah blah. But 4Q GDP just plain stunk so may not come to pass.

    But one thing that has come to pass is the dividend paying stocks have significantly underperformed in the last month. That is a very new phenomenon. The banks have had bouts of underperformance for some time but now it is all dividend yielding stocks. That, to me, signals their yield is possibly too low to compete with upcoming long rates. In it’s psychotic gyrations, the market apparently thought that might be a good thing yesterday. The final hurrah? The glory of large cap leadership? Yea, it steepens the yield curve for banks but at the expense of …………the consumer and alot of other interest rate sensitive industries.

    The S&P dividend yield is still at a 100 year low RANGE that started with the mid 90s blow off and still exists today. If long rates are starting to rise off of their unsustainable low, that situation may need to rectify itself. That ain’t going to happen by increasing the dividend yield.

    PE contraction. Just one scenario which may help lead a market correction that does not necessarily involve the consumer puking. Who knows. Just another possible scenario. Where are all of those economists that are so busy updating their charts? Do they look up long enough to contemplate such possibilities? I doubt most do until the new trend smacks them and it has already happened.

  2. miami commented on Feb 15

    Jan06 sales were up 8.8% over Jan05 sales, thus accounting for ‘seasonal adj’ and use of gift cards.

    If you think the consumer is going broke this year, you will likely be very, very, very wrong.


  3. Patrick (G) commented on Feb 15

    I know it doesn’t work this way for retailers, but from a consumer POV, gift cards in January are money that was spent in December.

    Also, I don’t know about you folks, but Not-Spending in January and February is how I rebuild the savings I depleted paying for Christmas.

    If you don’t have savings to dip into, you either have to Not Spend, or extend the months/years/debt to Master Card.

    So considering that avg income has been flat and savings declining, a Great December should be balanced by a lousy January February, or vice versa, or both can be mediocre.

  4. David Silb commented on Feb 15

    Hey good analysis Barry glad you picked up on some of the obvious and glaring mistakes made.

    “However, financial reporters should be savvier, and realize that 193k jobs were no more created in January then were Enrons’ profits created in Q1 2000. Its an accounting adjustment (Only Enron’s were fraudulent, while the BLS data is real — just seasonally adjusted). Further, the unemployment rate drop has been fully explained as a drop in the Labor Participation Rate, and note an actual decrease in the number of people looking for work.”

    Very good point. I can not believe they are still flogging the lower unemployment rate without mentioning the lower participation rate.

    Good read Barry.

  5. Emmanuel commented on Feb 15

    As always, a very good explanation Mr. Ritholtz. Unlike my sometimes reflexive skepticism, you explain very well how the numbers behind the numbers show they’re not all they’re cracked up to be.

    The problem now is that I hardly believe any economic indicators anymore. Call it the political economy of statistics, but it’s very interesting how fudging figures serves ideological purposes.

    That said, on to a -1.0% savings rate!

  6. miami commented on Feb 15

    ‘Further, the unemployment rate drop has been fully explained as a drop in the Labor Participation Rate, and note an actual decrease in the number of people looking for work.”

    This is traditionally the case in Jan as students return to school, no?

  7. pete Preissle commented on Feb 15

    Barry, LOL with the “…financial writers should be savvier…”

    You and I both know these are the same jokers who do this:

    NEW YORK (AP) — Declining oil prices and weakness in other commodities sent stocks lower Tuesday,


    NEW YORK (AP) — A drop in oil prices and strong earnings at PepsiCo Inc. and Cisco Systems Inc. propelled stocks higher Wednesday,


    The stories ran last week.

Posted Under