Note 2-50

(50) If the effects of a merger are presumed to take several years to emerge, data may be limited for mergers in and after fiscal 1999. In order to prevent a fall in the number of samples as much as possible, the gap between the average ROA for three years from the year of the merger implementation (for up to two years after the merger) and that for the year before the merger was adopted as the first-year effect, and the gap between the average ROA for three years after the merger year (for the first to third years after the merger year) and the first-year effect as the second-year effect.