This paper compares different methods for handling missing values in the "Short-term Economic Survey of ALL Enterprises in Japan (Tankan)". Empirical research on missing value issues in official statistics has been limited. Our study shows that, contrary to the fact that weighting method is widely used in sample surveys in Japan including the Tankan survey, the "last value carried forward" procedure is the most accurate for the Tankan survey. The result reflects characteristics of the Tankan survey: namely that the variance observed in the data of a given enterprise over time is smaller than the variance observed within an individual stratum of the survey at a given point in time.
