The Bank of Japan (BoJ) has reorganized the key TANKAN survey, its widely followed quarterly measure of business sentiment. In the initial report, issued yesterday, April 1, the new Business Conditions reading for March is -5, compared with -11 in December. On the old survey basis, December was -15. Due to changes in the selection and grouping criteria for companies in the sample, the new data are not comparable with the old. To the extent possible, December data were reaggregated according to the structure of the new survey, so there is one historical period for comparison with actual results and with a forecast for the current period.

The BoJ made three major changes in the sample composition. Like many other countries, it reorganized the industry divisions. Similar to the US shift from "SIC" to "NAICS", Japan is moving to the "Japan Standard Industrial Classification". This captures more of the new kinds of industries that have emerged during the recent technology boom; for example, "information services" and "communication", once subsumed in a broad "services" sector, are now shown separately. Secondly, the BoJ applied the results of a business census taken in 2001 to select firms for sampling. Third, it switched the method of designating businesses by size. Formerly this was judged on the basis of number of employees, but now capitalization is the criterion. This recognizes that some companies might utilize a small number of employees who actually operate a large business, such as a holding company. These changes together are thought by the BoJ to capture better the newer, more vibrant segments of the Japanese economy, which may have experienced better performance in recent years than the former industries and companies in the survey. They believe that the former survey results may have been biased downward due to its concentration on older, now less relevant industries and companies. Additionally, some concepts in the survey were redefined.

In terms of the survey results themselves, the following table indicates that business fortunes appear quite segregated by business size. Small firms are generally struggling, medium firms believe they abide in a mixed environment, and large firms are experiencing net positive performance. At the same time all major groups of companies performed better in the first quarter than they expected to. That is, the March "actual" is more positive or less negative than the December "forecast" value. Since there is no more history, it is not possible to judge whether this underestimation is a random circumstance or a habitual conservatism. Whichever would apply, large firms expanded on balance in the first quarter in both manufacturing and non-manufacturing sectors, and medium-sized manufacturing firms shifted into positive territory. Even small manufacturing companies had minimal erosion in the latest period, and they expect that to continue during the spring.