Use of Time Series Analysis as an Alternative Methodology for Estimating Value of Missing Data in the MISSI and PPS
Missing data is evident in the Monthly Integrated Survey of Selected Industries (MISS) and Producer Price Survey (PPS) due to late responses among sample establishments. A complete data set of sample establishments is key in yielding indices. Currently, the carry forward technique or cold deck imputation is the practice for handling missing data. This method uses historical imputation without trend adjustment, that is, the latest available data from the establishments is used to estimate for missing data. It produces rather crude estimates for other variables such as the value of production used in the MISSI. This study determined an alternative methodology for estimating value of missing data for both the MISSI and the PPS that will generate more reliable estimates. Using Eviews, time series analysis was executed to establish an operational methodology to forecast missing data resulting from non-responses in the MISSI and PPS. Comparisons on how the forecast values differ from the imputed values and actual data, and how the resulting indices using forecast values differ from using the imputed values and actual data were both obtained. This study recommends the implementation of the most appropriate forecasting methodology, however, the forecasting techniques have limitations. Yet the upper and lower limits of the forecasts values may be used as an option to determine the impact of unusual events, shocks or interventions.