Reality Is Too Often Elusive in Those Statistics We Revere

Oh, how we revere official statistics, those precise, seemingly authoritative numbers that purportedly reflect realities beyond personal knowledge: exports, productivity, price changes, educational levels. The latest data in these and other areas are widely received as established truths, as reliable as official temperature recordings or batting averages. In response, government and industry set policy and plans and the stock market gyrates.

Not often examined is whether the numbers do indeed approximate reality. When that’s done, they are often found to be so plainly inaccurate or statistically bizarre that their acceptance suggests a gullibility epidemic or a national hunger for assurance that we really know what’s going on out there.

The reality is that we seem to know less and less as industry goes international and new high-tech products and services upset traditional statistical score-keeping.

A new study by the Office of Technology Assessment, Congress’ own think tank, notes the haphazard nature of crucial import-export data. A review of U.S.-Canadian trade figures for 1986, it says, resulted in a 42% downward revision of the U.S. trade deficit, from $23 billion to $13 billion. The explanation, quoted from a study by another congressional agency, the General Accounting Office, is that “there is a strong possibility that U.S. exports are not fully counted; as a result, the U.S. merchandise trade deficit possibly has been overstated for the past several years.”

A basic problem, the technology office says, is that the record-keeping of international trade has failed to keep up with the new patterns of international production: “In 1985, nearly one-third of all U.S. exports were exports from U.S. companies to overseas affiliates. Over one-fifth of all imports to the United States came from these affiliates.” However, the report says data on what’s produced where is skimpy, with the result that “an uncomfortable number of assumptions must be made to see which industries are affected directly and indirectly by changes in the U.S. trade position.”

The low level of increase in American productivity is a persistent issue in economic policy-making, but here, too, reliable information turns out to be scarce. The technology office cites a study that concluded that “as much as 30% of the productivity decline after 1973 is the result of errors in measuring both outputs and inputs.” It adds that productivity figures compiled by the Bureau of Labor Statistics are especially poor on the service industries, which now account for a majority of employment. “The services that are included tend to be industries whose output can be quantified, such as ton miles in the railroad transportation industry.” This means that “important and quickly growing service industries whose output is difficult to measure, such as business services, health care and private education, do not have published productivity estimates.”

The unreliability of many basic official statistics even extends to the number of scientists and engineers in the United States, a seemingly simple matter that’s arousing concern because of indications of serious shortages in the next decade. The National Science Foundation, the federal agency that tracks these professions, put the number at 4,626,500 in 1986. But the Bureau of Labor Statistics, using census information, has its own figure: 3,287,000. Within its total, the foundation says, the number of “biological and life scientists” is 411,800. The figure from the bureau is 65,000.

Asked to account for these widely varying counts, an expert panel of the prestigious National Academy of Sciences said that “they are not readily explained,” and added that the academy was “greatly troubled by these differences.”

Many federal statisdical programs were gutted in the early years of the Reagan Administration and began to make a comeback only after industry felt the loss of the missing data. The recovery still has a way to go. But even if it eventually soars beyond previous levels in scope and sophistication, both caution and humility are in order.

The complexities of the real world are not easily captured by statistics.