An Estimation of Taiwan Wage-Phillips Curve by Comparing Linear and Curve-Linear Regressive Model

本研究蒐集2009年9月至2013年2月之台灣總體經濟月資料，以線性迴歸模型與曲線線性化迴歸模型估計實質薪資型菲利浦曲線，並發現台灣實質薪資型菲利浦曲線的誤差修正項遠高於歐美國家，顯示台灣薪資成長率不僅受到通貨膨脹率與失業率影響，還明顯地受到勞動生產力影響。其次，台灣實質薪資型菲利浦曲線以曲線線性化迴歸模型估計更為精準且具有較高解釋力，並可觀察到台灣的實質薪資成長率並無成長跡象，此結果對決策者而言需注意政策擬定時慎重考量使用之數據。最後，台灣的實質薪資型菲利浦曲線不可以OECD國家之估計方程式標準進行估計，而需考量台灣的經濟結構特殊性，如高貿易依存度、小國特質或中國崛起因素所影響。 This study focuses on the estimation of Taiwan Wage-Phillips curve depending on Blanchard and Katz (1999) and uses the macroeconomic data from National Statistics database of R.O.C. and Taiwan Economic Journal database during September, 2009 to February, 2013. The study can compare the estimated values of Wage-Phillips curve of US and EU but compare the linear and curve-linear regressive models. The main findings are (1) Taiwan’s error correction term comes in a significant and more right-signed coefficient than US and EU; (2) curve-linear regressive model can accurately explain the wage-Phillips curve of Taiwan such as the decreasing real wage rate and increasing nominal wage rate of 3% by Taiwan government; and (3) curve-linear regressive model can forecast that the real wage rate has less increasing stretch. Finally, this study does not add the export-oriented variables in linear and curve-linear model, there might be that the raising China is an important factor to affect Taiwan’s wage rate.