Monthly unemployment rates are considered an important gauge of how the economy is doing -- and are widely anticipated and used by economic forecasters and policy makers.

The states use the survey on which the national figure is based to determine their own jobless rates. But there is a hitch. Both national and state unemployment rates are calculated based on samples. While the national sampling is sufficiently large to arrive at a fairly reliable figure, the samples taken by states can be small enough to produce unreliable data.

Ohio's recent experience is a case in point.

• When Ohio issued its latest unemployment rate last month, it reported that the figure for April was 3.9 percent -- and then announced that the figure had to be wrong, since the number of people filing for unemployment compensation was up 68.8 percent from a year ago, when the jobless rate was 4.1 percent.

• State officials became suspicious when the February rate dropped sharply to 3.7 percent from January's 4.2 percent -- but they chalked it up to aberrations which routinely occur once or twice a year.

After checking and rechecking their figures and finding no obvious errors, they were left with essentially two explanations.

• One is the difficulty in measuring a dynamic and bipolar economy in which certain businesses are expanding and others contracting, and indicators such as inventories and orders plummet one month and inch forward the next.

• The other is the imperfections in the process of calculating the unemployment rate.

• While the national unemployment rate is calculated based on interviews with members of 60,000 households across the country, Ohio's rate is based on just 2,000 household interviews -- only 0.04 percent of the state's total households.

• Moreover, the bulk of people sampled in Ohio live in major metropolitan areas -- which have higher employment in services than manufacturing.

Unreliable figures can exasperate not only the business community -- which needs sound data to estimate future production and sales - but also state and local officials who use the information for a variety of policy and planning needs.

Source: Clare Ansberry, "States Discover It Is Hard Work to Figure Their Jobless Rate," Wall Street Journal, June 4, 2001.

For text (WSJ subscribers) http://online.wsj.com/articles
/SB991601114569535571.htm