綠能產業越趨重要的年代，太陽能產業瞬息萬變，自從中國廠商崛起，擁有低勞動成本、原物料成本及政府補助的優勢下，大量供給破壞全球太陽能產品供需。美國政府因此對中國廠商提起雙反訴訟，台灣從中獲得貿易分散效果及轉單效應，但美國旋即展開新雙反訴訟，台灣便被包含於起訴對象中，於是貿易分散效果轉變為貿易損害效果。因此本研究透過量化分析2010年1月至2016年12月的出口資料以及廠商營收，探討台灣太陽能廠商受到兩次雙反案訴訟的實質影響。 本研究利用最小平方法（Ordinary Least Squares）及固定效果模型進行回歸分析（Fixed Effect Model）。再以F檢定（F Test），檢測兩種方法何者更為適合本研究。研究結果發現，從對五個貿易對象出口量及廠商營收縱橫資料（Panel Data）當中，並無法看出貿易分散效果，但在針對單一對象—美國的實證結果中，可以看出貿易分散效果，因此可以確定台灣廠商的貿易分散及轉單效應集中於單一對象。而新雙反案所產生的貿易損害效果，雖可從實證結果中發現，但因為反傾銷、反補貼具有延遲效果，再加上新雙反案調查與終審時間點過於接近，都可能是使得貿易損害效果並不顯著的原因。 ;In the years when the green energy industry is becoming increasingly important, the solar energy industry is changing rapidly. Since the rise of Chinese manufacturers, with the advantages of low labor costs, material costs and government subsidies, large supplies have disrupted the supply and demand of global solar energy products. The U.S. government therefore filed a double-anti lawsuit against Chinese manufacturers, from which Taiwan obtained trade diversion effects and the order transfer effects. However, the United States immediately launched a new double-anti lawsuit. Taiwan was then included in country which was issued by The U.S. government, and the trade diversion effect turned into the trade destruction effect. Therefore, this study used the export data and the revenue of manufactures from January 2010 to December 2016, and found the impact of double-anti lawsuits on the solar manufacturers in Taiwan. This study uses the Ordinary Least Squares and the Fixed Effect Model. In addition, I used the F test to test which of the methods is more suitable for this study. The study found that the trade diversion effect cannot be seen from the export volume of five trade partners and the revenue of manufactures. However, the trade diversion effect could be seen in the U.S.A empirical results. Therefore, it could be determined that the trade diversion and the order transfer effect of manufacturers in Taiwan were concentrated in a single object. The effect of trade destruction caused by the new double-anti could be found in the empirical results, but it was not significant. It might be due to the delaying effects of double-anti, and the fact that the time of investigation and the final judgment were too close to each other. This might make the trade destruction effect insignificant.
