The Synchronization of Output Fluctuations Between Taiwan and the United States

Abstract

本文在於探討於1961年至l987年這一段期間臺灣與美國總體產出波動的特性。希望能夠檢定(1)確定趨勢過程(deterministic trend process)或隨機趨勢過程(stochastic trend process)何者較適於代表臺灣與美國總體產出變數演化(evolution)的過程，(2)在臺灣與美國總體產出的波動中，趨勢成分(trend component)或循環成分(cycle component)何者扮演比較重要的角色。比較以OLS模型及ARIMA模型所導出的產出循環(output cycles)，吾人發現以OLS模型導出的產出循環較以ARIMA模型導出的產出循環波動幅度(magnitudes)較大但頻率(frequencies)較低。依據理論的背景，應以ARIMA模型而非OLS模型所導出的產出循環來代表臺灣與美國之產出波動的一致性。This paper investigates the nature of output fluctuations in both Taiwan and the U.S. for the period from 1961 to 1987. It attempts to test (1) whether the evolution of the output series is better represented by the deterministic trend process or by the stochastic trend process, and (2) which component, trend or cycle, plays a more important role in the output fluctuations.Using the Dickey-Fuller test and the Phillips-Perron test to examine three different unit root models, the results show that we cannot reject the hypothesis that the evolutions of Taiwan`s and U.S. real output follow a stochastic trend process. Both estimates of the structural model and the ARIMA model indicate that the trend component plays a more important role than the cycle component in both countries` output fluctuations. In addition, innovations in output have persistent effect on the output fluctuations of Taiwan and the U.S.At the end, the output cycles derived by the OLS and ARIMA models are compared. We find that the output cycles derived by the OLS model have a larger magnitudes but lower frequencies than the output cycles derived by the ARIMA model. Referring to theoretical background, the output cycles derived by the ARIMA model rather than by the OLS model should be used to represent the synchronization of output fluctuations between Taiwan and the U.S.