景氣對策信號建構程序之檢視

Evaluating the Procedure for Constructing the Monitoring Indicators of Business Cycles

近年來，國內、外經濟情勢快速變化，各界對於景氣對策信號即時反映經濟體質的期盼日增，因而對現行景氣對策信號採用年增率測度，容易受到基期因素影響的現象，提出質疑與檢討建議。有鑑於此，本文旨在透過合適的計量方法，藉由實證分析與蒙地卡羅模擬之應用，嘗試對景氣對策信號各種編製程序的相對優劣進行評價，據以探究景氣對策信號之編製程序是否尚存改善空間，期能藉此對我國景氣燈號之改善，提供若干助益。較具體地說，透過景氣對策信號構成項目之實證剖析與各種資料性質的模擬試驗可發現，在本研究檢視數種年增率的替代做法中，確實具有同時提升即時性與精確性的可能性，尚存柏拉圖式的改善空間，值得研究者持續投入心力，進行更為審慎、周延的評估。

Recent years have seen swift changes in the economic situation at home and abroad. This has generated rising demand from all quarters for a set of monitoring indicators that enable a timely response to shifts in the state of the economy. However, many doubts have been voiced about the monitoring indicators currently used in Taiwan, which measure year-on-year rate of change and hence are easily affected by base effects. In this paper, we conduct some econometric investigations to evaluate several alternative procedures for constructing monitoring indicators. More specifically, based on empirical analysis of the components of the current monitoring indicators and a small-scale Monte Carlo experiment with various data properties, the feasibility of simultaneously increasing timeliness and accuracy has been verified. That is, we demonstrate the possibility of a Pareto improvement. Therefore, we conclude that it would be worthwhile to conduct more advanced evaluations, employing more sophisticated techniques and covering more comprehensive considerations.