摘要 本文採用2001年至2014年共25國的追蹤資料(Panel Data)檢測外人直接投資與避稅天堂之間的關係,其他控制變數包括租稅負擔、國家發展程度、GDP、教育支出及人口密度等。本文除使用傳統最小平方法進行分析外,也使用Koenker and Bassett (1978)所提出的分量迴歸法(Quantile regression)進行探討,分量回歸法可以透過不同分量給予不同權數進行估計,用以說明解釋變數對被解釋變數在特定分位數下的邊際效果。 根據研究結果發現,首先,使用分量迴歸的結果顯示出,避稅天堂只對高分量的外人直接投資有顯著影響。 第二,租稅負擔對高分量的外人直接投資具有較大的影響。 第三,本文發現避稅天堂對開發中國家的外人直接投資影響比已開發國家還大。
ABSTRACT This paper investigates the impact of tax havens on foreign direct investment (FDI) for 25 selected countries among 2001 to 2014, which employed quantile regression approach, based on Koenker and Bassett (1978). Quantile regression approach can analyze the effects across quantiles in the conditional distribution. Our analysis yields several interesting findings. First, by means of quantile regression model, the results showed that tax haven increase FDI at the top of conditional FDI. Second, tax burden has a relatively greater impact for high quantiles of the conditional FDI. Third, we find the positive effect of tax haven on FDI among developing countries. Finally, it is also noted that country characteristic factors, in part, responsible for the FDI.