台商在中國大陸投資對出口的影響

Abstract

本文研究的對象是針對1991-2006年, 台灣上市公司投資中國大陸對母公司營運的影響。理論模型中分為兩種論點, 其一為“水平式對外投資”(horizontal foreign direct investment), 其二為“垂直整合式的對外投資”(vertical foreign direct investment); 前者強調對外投資與出口兩者為替代的關係, 後者強調投資與出口為互補關係。本文使用經濟部投資審議會「對外投資核准金額」與財政部關稅總局針對海關進出口貿易的統計資料, 合併取自「台灣經濟新報」的上市公司財報資料, 以固定效果的panel data模型進行分析。發現對台灣對中國大陸的投資, 顯著得減少廠商本身的出口金額, 而由估計結果得到投資中國大陸對出口的替代彈性約為3% 。由於出口資料具有截切(censored) 的問題, 因此, 後來使用random effect的Tobit model 解決這個問題, 並確認結果與原先固定效果的迴歸模型一致。

In the literature, two types of theoretical models describe the relationship between foreign direct investment (FDI) and exports. While the theory of 'horizontal FDI' predicts that FDI and exports are substitutes, the theory of 'vertical FDI' suggests that FDI and exports are complements. This paper examines how the amount of FDI in China affects exports of the parent firms using panel data of Taiwanese manufacturing firms. The panel data set combines data on firms listed on Taiwan Stock Exchange with the amount of FDI in China (from the Investment Commission, Ministry of Economic Affair) and customs data on exports and imports. We start with the empirical model of panel regressions with fixed effects to estimate the effect of FDI on exports. We find that investment in China significantly decreases exports. Since firms have to pay a sunk cost for entering the export market, if a firm cannot reach the threshold of exporting, its amount of exports is zero. To deal with the problem of censored dependent variable, we use random-effects Tobit model for a robustness check. The results are consistent with the findings using the model of panel regressions with fixed effects. Thus, we conclude that FDI in China significantly substitutes the exports of the parent firms in Taiwan.