The Effect of Missing Data on Estimates of Exponential Trend-Cycle and Seasonal Components in Time Series: Additive Case

Abstract

This study discusses the effect of missing data on Buys-Ballot estimates of trend parameters and seasonal indices. The method adopted in this study is based on the row, column and overall means of the time series arranged in a Buys-Ballot table with m rows and s columns.  The method assumes that (1) Only data missing at one point at a time in the Buys-Ballot table is considered. (2) the trending curve is either linear or exponential (3) the decomposition method is either additive or mixed. The article shows that, the estimation of the missing data as they occur consecutively with the errors being normally distributed. Result indicates that, under the stated assumptions, the differences between trend parameters in the presence and absence are insignificant, while that of seasonal indices are significant.