NAICS Redux

NEWS ANALYSIS

It's been almost three years since government statisticians began publishing economic data using the North American Industry Classification System (NAICS), which replaced the old Standard Industrial Classification (SIC) method. At that time, C&EN found that many private economists, market researchers, and industrial planners were concerned about the new system.

Among the concerns, both for economists tracking the U.S. economy overall and for those following the chemical industry, was the loss of history. Under NAICS, data is seldom provided for years prior to the mid-1980s--a very short time for economists. SIC data were available going back decades.

The main purpose of NAICS was to base data on process rather than industry. Under SIC, firms classified as chemical companies that also made, for instance, chewing gum, would see all of their activities classified as "chemical." But using NAICS, chemical company employees who produced the chewing gum were removed from the chemical sector and placed under the chewing gum classification.

Now that most of the government data used by C&EN and by economists tracking the chemical industry have been converted to NAICS, it is time to look at how the exercise has worked. These data include shipments and inventories; production; employment, hours, and wages; and the producer price index for chemicals.

Economists seem to understand the reasoning behind NAICS; however, when they talk about the system, there is always a "but" that follows. Maurine A. Haver, president of Haver Analytics, a New York City consulting firm specializing in databases and software for economic analysis and business decision-making, says: "The obvious strength of NAICS is that newer high-technology industries and services are more clearly broken out. SIC really focused on manufacturing, but manufacturing has not been the dominant factor in the economy for a long time. So clearly, NAICS does a very good job of reflecting the economy we have."

Then comes the "but." Haver, who is a past president of the National Association for Business Economics and a member of the group's statistics committee, says: "The problem with NAICS is that those of us who use industry data, who care in the first place, also care about time series. The problem that continues to confront us is that there is no really nice mapping between NAICS and SIC. We lose history as a consequence.

"The Census Bureau came out with data that began in 1992, and they thought it was a great gift," she continues. "We were supposed to be thankful that we could work the data back to 1992. Most of us were used to working the data back to the '60s, '50s, and '40s." But, she says, when the Census Bureau published NAICS data three years ago, "we didn't even know we had a recession. We were looking at something that didn't have even one business cycle. That's really the weakness of NAICS."

But there are other, structural, problems with NAICS. Manufacturers, Shipments & Orders was the first of the monthly reports regularly used by C&EN to be converted to NAICS, with publication in mid-2001, and shortcomings in the data then have not been rectified. One niggling problem, pointed out at that time and still there, is that the Census Bureau uses the term "basic chemicals" when it really means "all chemicals" or "chemical manufacturing." Basic chemicals, according to NAICS nomenclature, is supposed to mean the subsector of chemical manufacturing composed of basic organic and inorganic chemicals. This leads to one of the biggest problems in the monthly report.

Basic chemicals--the real basic chemicals--is the largest chemicals subsector. Under SIC, it was called "industrial chemicals" and made up more than 40% of all chemical shipments. But shipments and inventories of basic chemicals under NAICS are not reported. Instead, the Census Bureau reports only the subsectors of pharmaceuticals and medicines; pesticides, fertilizers, and other agricultural chemicals; and paints, coatings, and adhesives. These subsectors are individually small potatoes compared with the basic chemicals subsector: Together, the three account for a little more than 30% of total chemical shipments.

When the changeover took place in 2001, T. Kevin Swift, senior director at the American Chemistry Council (ACC), said, "I would certainly have segmented it differently, especially to include basic chemicals." What is not reported, he pointed out, "includes everything from chlorine to ethylene oxide to printing ink."

While the Census Bureau has skimped on reporting of shipments and inventories, the Federal Reserve Board has done much better with its indexes of industrial production. For one thing, the Federal Reserve data go back to 1986; however, this is still not enough to cover an entire business cycle. More important, the production statistics get to a level of detail that makes tracking of an industry much more comprehensive. The monthly report not only provides the indexes for total chemical production, it also gives data for 22 chemical sectors and subsectors. Under SIC reporting, the Federal Reserve provided indexes for just 14 of the sectors and subsectors.

But if production data are good, employment data are another story. When reporting of industry employment is discussed with economists, the words "frustrating" and "confusing" are often heard.

The problem is essentially that NAICS, being process-based, primarily recognizes employees who are somehow involved in production--and few others.

As a result, there were large transfers of "service" employees out of the industry. Only two services sectors had a large enough percentage of chemical employees shifted to them to be mentioned--management of companies and enterprises, and professional and technical services. But these shifts were huge and extremely meaningful to anyone analyzing chemical industry employment.

The Bureau of Labor Statistics (BLS) used the first quarter of 2001 to construct an SIC-to-NAICS ratio that shows 7.4% of total chemical employees were shifted to management of companies and enterprises and 3.4% were put into professional and technical services. Thus, using the SIC employment data for the same quarter, more than 110,000 employees, or some 11% of total chemical employees at the time, were moved out of the industry.

THIS MEANS THAT, according to BLS, all sorts of managers, from company CEOs to branch office heads, no longer work in the chemical industry. And the industry no longer employs research chemists. They now work in the professional and technical services industry.

Martha G. Moore, formerly senior economist at ACC and now president of Moore Economics in Silver Spring, Md., says of BLS: "I understand why they've done what they've done. The nature of the economy is changing from a job function being in a certain company to its being outsourced. That makes some sense, but it's frustrating when you look at chemical industry employment."

She calls attention to what happened to employment data in Delaware, a state with several chemical company headquarters--one very significant--but a relative paucity of chemical plants. In 2000, the last year for which SIC data are available, Delaware had 21,318 chemical employees. The following year, the first under NAICS, the state had a mere 5,110 chemical employees.

In other words, nobody knows how many people work in the chemical industry in Delaware--or anyplace else.

If there weren't enough problems with NAICS, Haver warns that there are those in government statistical offices who believe there should be a NAICS-like international reporting system. She believes, though, that once other countries begin tinkering with NAICS, it will end up changing again. And she worries that, once again, the history will be lost.

"I do hope," she says, "having gone through all this, we're going to keep NAICS up to date so that, maybe not in my career, there will be history again."