On the willingness of economists to believe data that doesn’t make sense

Suppose I told you that the economy was about to go through its greatest financial convulsion in 75 years, causing a vicious disruption in the credit system and the global trading system. Consumers and small businesses would suddenly be cut off from access to credit, and large businesses would go into a fetal crouch and stop investing. A priori , would you expect that such a scenario would lead to

a) An acceleration in productivity growth,  or

b) A deceleration in productivity growth or even a drop in productivity.

My guess is that a majority–perhaps even a great majority–of economists would have answered (b) if I had thought to put the question to them.  After all, disruptions to credit, trade, and investment are terrible for the normal operations of business.

Why, then, were economists willing to accept the productivity surge of 2007-09 as gospel? It made no sense that businesses should suddenly function better  in the face of a financial crisis that nearly dragged down the whole economy.

I’ve noticed this willingness of many economists (and journalists, for sure) to accept the official data as gospel before.  It’s not a good thing.

P.S. You may guess that I’m revising my textbook.

 

 

 

Productivity “Surge” of 2007-09 melts away in new data

Until this morning, the official data showed that the U.S. productivity growth accelerated during the financial crisis. Nonfarm business productivity growth supposedly went from a 1.2% annual rate in 2005-2007, to a 2.3% annual rate in 2007-2009.  Many commentators suggested that this productivity gain, in the face of great disruptions, showed the flexibility of the U.S. economy.

Uh, oh. The latest revision of the national income accounts, released this morning, makes the whole productivity acceleration vanish. Nonfarm business productivity growth in the 2007-09 period has now been cut almost in half, down to only  1.4% per year.

This revision has political and policy consequences. Back in March, I analyzed the apparent productivity surge and  argued that it was statistically suspect.  I pointed out that:

First, the measured rapid productivity growth allowed the Obama Administration to treat the jobs crisis as purely one of a demand shortfall rather than worrying about structural problems in the economy.  Moreover, the relatively small size of the reported real GDP drop probably convinced the Obama economists that their stimulus package had been effective, and that it was only a matter of time before the economy recovered.

A more accurate reading on the economy would have–perhaps–cause the Obama Administration to spend more time and political capital on the jobs crisis, rather than on health care. In some sense, the results of the election of 2010 may reflect this mismatch between the optimistic Obama rhetoric and the facts on the ground.

Now the productivity surge of 2007-09 has vanished, and so is the pretense that the U.S. economy was able to sail   through the financial crisis with barely any problems.  It’s time to set a new economic course.

Does Larry Summers Support Countercyclical Regulatory Policy?

Here’s an interesting excerpt from Ezra Klein’s interview with Larry Summers:

EK: What about ideas to simply speed up the regulatory process during downturns?

LS: This could be very important. Harvard built a whole football stadium in the 1890s with less than two years from the idea’s conception to the stadium opening. That kind of thing does not happen today. My attitude is we probably slow things down unnecessarily with our various approval processes. So I’m sympathetic speeding the regulatory process during recessions. But I have to be honest and say my enthusiasm for reducing regulatory processes in recessions is in significant part a reflection of a more general concern about regulatory delays.

 

 

 

(Hat tip to William Hoffman)

Misinterpreting Data: How the WSJ Got the Wireless Jobs Story Wrong

On July 17 the online edition of the WSJ published a widely-cited story entitled Wireless Jobs Evaporate Even As Industry Expands. The main point of the story  (my emphasis):

In May, on the heels of a record year for industry revenue, employment at U.S. wireless carriers hit a 12-year low of 166,600, according to U.S. Labor Department figures released earlier this month. That’s about 20,000 fewer jobs than when the recession ended in June 2009 and 2,000 fewer than a year ago. While the industry’s revenue has grown 28% since 2006, when wireless employment peaked at 207,000 workers, its mostly nonunion work force has shrunk about 20%.”

In addition, the Journal digs further into the official data and claims that:

The number of customer-service workers at wireless carriers dropped to 33,580 last year from 55,930 in 2007, according to the Labor Department

Seems like a pretty straightforward story, doesn’t it?  The Journal is quoting directly from authoritative BLS data to demonstrate that the  wireless industry has been losing jobs, despite the mobile boom. The big picture message: Innovation does not equal job growth.

Unfortunately, the reporters and editors at the WSJ  fell into the same trap that has ensnared many other journalists, policymakers, and even economists. They looked at the label on a piece of official economic data, and assumed that they understood it.  But as we saw during the financial crisis and subsequently,  government economic data can all too easily be misinterpreted.

In this case,  the article was based on the Journal’s analysis of jobs in the “wireless telecommunications carrier industry,”  as defined by the BLS. However,  despite the name of the data series, it turns out that:

  • The BLS definition of the “wireless” industry does not include company-owned  retail stores or stand-alone company-owned call-centers.
  • The customer service numbers cited do not include  company-owned retail stores or stand-alone company-owned call centers.
  • The 2007 occupational data in telecom cited in the story cannot be compared with later years, because the telecom industry classifications in the occupational data were substantially redone in 2008.

As a result:

  • The data cited in the WSJ article completely misses the growth of jobs at company-owned retail stores (see Metro PCS chart below)
  • The data cited in the WSJ article potentially misses call center job growth such as the expansion of Verizon’s Nashville call center (see example below)
  • The phrase “employment at U.S. wireless carriers hit a 12-year low”  simply cannot be supported by the available data.  Data from the industry trade association (CTIA), which shows wireless employment up 36% since 2000, is much more plausible (see chart below).
In my view,  the WSJ article is a classic case of misinterpreting official statistics.

Before getting into the details, why am I taking the time and trouble to disassemble this particular article?Historically innovation and job creation have been closely linked, as I have argued in multiple papers and articles. With Washington now fighting tooth and nail over the budget, it’s very important for policymakers to understand that successful innovation creates jobs, not the opposite.

Second,   journalists, policymakers, and economists need to understand how easily government statistics can be misinterpreted.  For example, the statisticians at the BLS have reported huge U.S. productivity gains over the past decade, including the years following the financial crisis–a fact that has been duly repeated by journalists and applauded by economists.  However, in a recent paper, Sue Houseman of the Upjohn Institute and I argued that these reported U.S. productivity gains could be interpreted, in part,  as  an increase in the efficiency of global supply chains.  It matters enormously for jobs and wages whether productivity increases are coming from more efficient domestic operations, or more efficient offshoring.

Or consider  consumer spending.  Journalists regularly report that  “consumer spending accounts for 70 percent of economic activity.”  (see, for example, this recent Associated Press story that ran on the New York Times website). However this number,  calculated by dividing consumer spending into GDP, is pernicious nonsense. Nonsense,  because consumer spending includes a big chunk of  imports, which does not correspond to economic activity in the U.S.  Pernicious,  because it perpetuates the fallacy that the U.S. cannot recover without gains in consumer spending (see my blog post on the subject here).

Details

Now let me turn to the details of the WSJ’s mistake, or if you’d like, misintepretation.  The WSJ analyzed BLS jobs data for the “wireless telecommunications carrier industry”, (with the NAICS  ID 5172). That data looks pretty bleak (if you want to download the data for yourself, instructions are at the end of this post).

However, the WSJ apparentlydid not realize  that the BLS collects industry employment by establishment, not by company. The BLS defines an establishment in this wa:y

An establishment is an economic unit, such as a farm, mine, factory, or store, that produces goods or provides services. It is typically at a single physical location and engaged in one, or predominantly one, type of economic activity for which a single industrial classification may be applied.

Whenever possible,  the BLS assigns each establishment to an industry, and counts all the employment at that establishment at part of that industry.

Viewed from this perspective, a single wireless carrier, such as Verizon Wireless or Metro PCS,  will typically include several different types of establishments, each of which will be assigned to a different industry.

  • Wireless operations are in NAICS 5172 (“Wireless telecommunications carriers”)
  • Company-owned call centers are in NAICS 56142 (“telephone call centers”)
  • Company-owned retail stores are in retail trade, probably NAICS 443112 (“Radio, TV and electronics stores”)
  • Mobile tower and base construction could be in NAICS 23713 (“Power and Communication Line and Related Structures Construction”)
There might even be more different types of establishments in the wireless industry…it’s hard to tell.

This has several implications. First, retail expansion by wireless providers is counted in the retail trade industry, not  the BLS “Wireless Industry” numbers that the WSJ used.  This is true even if the store is carrier-operated.

For example, the tremendous expansions of retail stores by Metro PCS  in recent years, with added jobs,  did not show up in the WSJ data (for a related example,  employees at Apple stores are counted in the retail industry, not the computer industry).

Second, to the degree that wireless carriers are expanding stand-alone call centers, those additional jobs are not being picked up by the WSJ data.  We don’t know exactly how many there are, but we do know that overall national employment at telephone call centers have been rising, surprisingly enough.  It’s likely that the expansion of the wireless industry is a factor in that rise in call center employment.

We also know that at least some wireless telecom companies have been hiring at their call centers. For example, it took the work of five minutes to find this example of Verizon hiring workers for a call center outside of Nashville. Here’s an excerpt from the July 1, 2011 story in the Nashville Post:

Verizon Wireless has announced via Facebook and Twitter that it will expand its Sanctuary Park Center of Excellence Loyalty Retention Center by opening an office in Franklin. The company plans to add some 300 jobs in the Franklin area over the next 18 months.

“We’re excited about this expansion for several reasons. It allows us to continue to provide customers with the high quality of service they expect from Verizon Wireless,” said James Nelson, associate director of customer service. “It’s also great to be a source for new job opportunities – especially in this economy.”

Further, the company said, “Many of the thousands of calls handled by LRC representatives each month are from customers requesting to discontinue service. It is their responsibility to convert as many of those disconnect requests into satisfied customers.”

The company ran its first training sessions in May and plans to begin taking calls at the center starting July 5.

I didn’t research this example any further. But it looks like these new call center jobs are not counted in the BLS data that the WSJ was using.

Finally, those customer service figures that the Journal made such a big deal about. Let me repeat the quote from the Journal story.

The number of customer-service workers at wireless carriers dropped to 33,580 last year from 55,930 in 2007, according to the Labor Department

Actually, that sentence is not correct as it stands.  The WSJ is citing occupational data pertaining to the “wireless” industry as defined by the BLS (NAICS 5172).   By definition, the WSJ’s figure for customer-service workers excludes  company-owned stand-alone call centers (like the previous example for Verizon Wireless). As a result, the figures cited by the Journal are absolutely useless for determining whether  wireless carriers are hiring or firing customer-service workers.

Just to add insult to injury, there’s a subtle twist that no reporter could be expected to know.  Buried deep in the documentation, the BLS explains that:

In 2008, the OES survey switched to the 2007 NAICS classification system from the 2002 NAICS. The most significant revisions were in the Information Sector, particularly within the Telecommunications area.

The implication is that telecom occupational data from 2007 simply cannot be compared to later years (I believe that the BLS would agree with that, if asked).

What’s the bottom line here? Let me show you again the chart of the jobs in the  BLS “wireless industry” (the data the WSJ used), and compare it to the survey of wireless industry employment done by CTIA, the wireless industry association.

The industry association figures-rose by 46% from 2000 until 2008, before dipping by 7% from 2008 to 2010.  By contrast, the BLS “wireless” data, which does not include call centers, retail stores, and tower construction, rose by only 8% from 2000 to 2008. Now, honestly, in the middle of a wireless boom of historic proportions, which figure do you think is more likely to reflect “employment at wireless carriers”, the phrase used in the WSJ story?

Now, that brings me to my final ethical question: Does the Journal have an obligation to run a retraction or a corrective story?  The article did not slander or libel anyone, and the reporter used the government statistics in good faith.  However, because the statistics did not mean what the Journal thought they meant, the story is filled with statements which leave readers with the wrong impression.  The typical reader  would read the story and naturally conclude that the phrase ” employment at U.S. wireless carriers hit a 12-year low”  referred to the number of workers who receive paychecks from Verizon Wireless, Metro PCS, the wireless part of AT&T, and the like.  But as we have seen, that phrase is based on government figures that only reflect a portion of wireless carrier employment.

More importantly, the story’s big picture conclusion–that innovation does not equal job growth–is not supported by the statistics. In this era of distrust of the press, should publications make an effort to clarify the record if their original story is faulty?

 

 

Coda: How to Get the Government Data that the WSJ used

 

Go to  http://www.bls.gov/data/#employment

Click on “Employment, Hours, and Earnings – National, Multiscreen data search”

Check ‘Not seasonally adjusted’, and click on ‘next form’

Scroll to ‘information’, and  click on ‘next form’

Click on ‘all employees, and  click on ‘next form’

Scroll to “wireless telecommunications carriers (except satellite)”, and  click on ‘next form’

Click on ‘retrieve data’

The data for customer service representatives in the wireless industry in 2007 can be found at

http://www.bls.gov/oes/2007/may/naics4_517200.htm#b41-0000

More on Cedar Balls

Remember that about two months ago I put up a post entitled Cedar Balls–”Grown in USA, Made in China”?. It created a fair bit of discussion (see, for example, “Wood to China: Cedar balls and rum” ).

I recently had an extended phone conversation with  Howard Goldman, who owned and ran Cedar Fresh from 2003 until recently, when he sold it (yes, he’s looking for a new business opportunity). Goldman confirmed that the company does ship cedar logs to china in containers where they are made into cedar balls, cedar hangers, and the like. “We used to do all our production in the U.S.”, says Goldman, “but we couldn’t be competitive.” It’s much less expensive to manufacture in China, he says, even with shipping both ways.

I asked Goldman to estimate the difference between producing in China and the U.S. He replied that it varied significantly between items, but cedar hangers tended to be 20-30% less expensive when made in China, including shipping and customs duty.

Goldman pointed out that Cedar Fresh is one of the few (“if not the only”) cedar storage accessories companies that still does manufacturing here in the U.S., with his primary remaining U.S. facility being an adult handicapped day care center in rural Arkansas.

Finally I asked if he could reduce the cost of producing in the U.S. by using machinery to do the most labor intensive parts of the production process (such as packaging). Goldman noted that a lot of the machines he would buy are actually made in China, and a lot cheaper there. So China may be cheaper than the U.S. both for labor and for capital investment.

Archives

Follow

Get every new post delivered to your Inbox.

Join 65 other followers