Benchmarking of Korean Economic Time Series (in Korean)

For a target economic variable, two economic time series with different periodicities may be available. There are discrepancies between low frequency series such as annual series and high frequency series such as quarterly series In this case the high frequency series are adjusted to make them consistent with low frequency series in a process termed benchmarking. Benchmarking procedures one of two approaches : a numerical approach such as Denton (1971)'s method or a statistical modelling approach such as the methods of Chow and Lin (1971), Fernández (1981) and Litterman (1983). Though most monthly and quarterly Korean economic time series have been benchmarked to annual series, the method is frequently not explicitly disclose, making it difficult to engage in a discussion of benchmarking procedures and carry out related studies. In this paper, we set out various definitions of benchmarking and the case for its necessity, before proceeding to introduce several benchmarking methods. Moreover, the quarterly benchmarked series of manufacturing GDP and facilities investment based on several methods are estimated using related series and are compared in terms with RMS%E and correlation. The comparative study shows that Litterman's method and Chow & Lin's method perform well. In addition, we estimate monthly manufacturing GDP and benchmarked seasonally-adjusted manufacturing GDP.