Economic forecasting and policy-making require swift and accurate assessment of the current state of export activity. However, the existing export indicators, such as goods exports as a subcomponent of GDP (“goods exports” hereafter) and customs clearance exports, are prone to fluctuations resulting from short-term and uncommon factors, including facilities maintenance or strikes taking place at exporters and changes in the number of business days. The goods exports indicator also lacks timeliness as it is compiled only quarterly and published around one month after the quarter concerned. Accordingly, in this paper we construct the Export Coincident Indicator (ET-COIN), which can be used in the tracking of export trends and nowcasting of the current-quarter goods exports.
We estimated ET-COIN using the same method as existing indices such as BOK-COIN and Eurocoin, by projecting the target on the medium- and long-run components with waves of period longer than one year, extracted from export-related variables. Given the high volatility of export indicators, however, we adopt the elastic net regularization to identify only the variables closely related to exports.
A comprehensive assessment shows that ET-COIN is useful for analyzing the trend of exports. First, ET-COIN is found to capture the export trend well, with short-term and idiosyncratic factors removed. Unlike bandpass filters, it has high end-of-sample stability, as estimates do not fluctuate dramatically even after new time-series data are added. It also has excellent qualities as estimate for the current-quarter goods exports. Given the frequent changes in export conditions driven by uncertainties surrounding the external environment, we plan to expand the range of export-related variables included in ET-COIN, while continuing to monitor and improve the adequacy of the index for the assessment of export conditions.