| Abstract: | 本文採用1998年第一季至2018年第四季之樣本期間，重新檢視歐肯法則(Okun's law)(失業率(Unemployment rate)及經濟成長率(Economic growth rate)在台灣之抵換關係(Trade-off))。考量季頻變數，包括經濟成長率、失業率及固定資本形成毛額(Gross fixed capital formation)；月頻變數，包括通貨膨脹率(Inflation rate)、廣義貨幣供給(Broad money supply)、匯率(Exchange rate)、重貼現率(Rediscount rate)、股票成交值及景氣對策信號(Monitoring indicator)等。利用混頻迴歸模型(Mixed-Data Sampling，簡稱MIDAS)，進行參數推估及樣本內預測(模擬)。實證發現：(1)台灣歐肯法則失業率與經濟成長率之抵換關係為；失業率上升1%，經濟成長率下降4至6%；高於美國失業率上升1%，經濟成長率下降3%。(2)除失業率外，固定資本形成毛額、通貨膨脹率、景氣對策信號、重貼現率及股票成交值等變數，皆會顯著影響經濟成長率。(3)由於混頻迴歸模型引入月頻資訊，故其預測績效優於普通最小平方法。
Based on the 1998Q1~ 2018Q4 data, reexamining Okun's law in Taiwan, the results of this study show that the trade-off between unemployment rate and economic growth rate in Taiwan is about 4 to 6% decline for every 1% of unemployment. This decline is higher than 3% of that of the Americans. Different from the econometric model in the past, the research considers seasonal frequency variables, like economic growth rate, unemployment rate and gross fixed capital formation, also monthly frequency variables, such as inflation rate, broad money supply, exchange rate, rediscount rate, turnover of the stocks, and monitoring indicator, etc.at the same time. By using the Mixed-Data Sampling (MIDAS) regression model to estimate the regression parameters, and imitate the sample. The results of this study show that in addition to the unemployment rate, the gross fixed capital formation, inflation rate, monitoring indicator, rediscount rate, turnover of the stocks, etc. will effect the economic growth rate notably. Besides, the forecasting performance of mixed data sampling regression model is better than the ordinary least squares method due to including monthly frequency information. |
