美國、日本為臺灣最密切的兩個傳統貿易國,但先前研究探討臺灣∕美國與臺灣∕日本購買力平價 (purchasing power parity, PPP)是否成立及其影響因素尚無明確答案。本研究使用最長可利用的新臺幣匯率資料以增進檢定力、考慮Balassa- Samuelson (BS) 效果(由Balassa (1964) 與Samuelson (1964) 提出)及結構性變動兩個影響因素、及使用晚近發展的傅立葉形式單根及定態檢定,探討臺灣∕美國與臺灣∕日本長期PPP 及不同匯率體制時期PPP 的成立情況與影響因素。實證結果發現,臺灣∕美國與臺灣∕日本實質匯率皆存在顯著的BS 效果及結構性變動,長期及各匯率體制時期PPP 都無法成立。影響臺灣∕美國與臺灣∕日本PPP 成立的因素不同,無論是長期或各匯率體制時期,結構性變動都是影響臺灣∕美國PPP 成立的主因;BS 效果則是影響臺灣∕日本長期PPP 成立的主因,只有在浮動匯率時期結構性變動及BS 效果都會影響臺灣∕日本PPP 的成立。本研究的結果顯示,於PPP 的相關研究應考慮BS 效果及結構性變動;再者,即使都是貿易密切的國家影響PPP 成立的因素也可能不同,值得應用PPP 假說時留意。
The U.S. and Japan are traditionally the two closest trading partners of Taiwan. Previous studies on Taiwan/U.S. and Taiwan/Japan purchasing power parities (PPPs) and their influence factors, however, provide unclear answers. This study explores the evidence of and the influence factors on the long-run Taiwan/U.S. and Taiwan/Japan PPPs as well as PPPs under different exchange rate regimes. We employ the longest available exchange rate data series of the New Taiwan Dollar, consider two influence factors of the Balassa-Samuelson (BS) effect and structural change, and apply the newly developed Fourier form unit-root and stationarity tests. The empirical results show no evidence of long-run Taiwan/U.S. and Taiwan/Japan PPPs as well as PPPs for different regimes. Infact, the influence factors on Taiwan/U.S. and Taiwan/Japan PPPs are different. For the long-run Taiwan/U.S. PPP and PPPs of different regimes, structural change is the main factor. In contrast, the BS effect is the main factor for the long-run Taiwan/Japan PPP, whereby only under a floating regime can both factors influence the support for PPP. The empirical findings reveal that both structural change and the BS effect should be considered in PPP-related studies. Moreover, the influence factors on PPP might be different even if all the countries explored are close trading partners, which is noteworthy for the application of the PPP hypothesis.