封閉型基金的掛牌買賣價格，與其投資標的淨資產價值（NAV, net asset value）間，總是存在一定程度的差距，且幾乎為10%~20%的折價現象，被視為財務學上的一種異常（anomaly）。早期研究學者們由經濟理性的角度解釋此種現象，認為是基金管理的績效表現、投資標的流動性或是稅制上的差異等等角度來解釋此一現象，但並沒有獲得一致的結論。 近期以來，有學者嘗試從行為財務學的角度，認為封閉型基金的折溢價現象及其幅度的變化，是由於投資人情緒（investor sentiment）的波動所造成。除了傳統的迴歸模式，更有不少學者以嚴謹的計量方式分析，例如財務時間序列的許多技巧。本研究即以時間序列之單根檢定（unit root test）、共整合現象檢定（co-integration test）及Granger因果關係檢定（Granger Causality test）等等方式，分析封閉型基金折溢價現象是否由投資人情緒所造成，及其兩者間是否有共整合現象。 雖然已有部分文獻探討類似議題，但跨國比較分析通常僅限於英、美兩國，且英、美兩國市場連動程度太高，可能影響分析結果的正確性；除此之外，樣本分析期間亦通常不超過十年。本研究以1991-2005年英國掛牌之封閉型基金為研究樣本，並包含投資標的為日本的封閉型基金，進行英國、日本之間的跨國性比較；不但有更足夠的樣本時間長度，亦能驗證是否不同跨國分析亦會有相同結論。 共整合現象檢定及Granger因果關係檢定大致支持行為財務學的角度。但英國、日本跨國比較的結果，似乎並不完全等同於前述英、美跨國比較的結果。建議後續可從掛牌國與投資標的區域連動程度較低的封閉型基金作為研究方向。 / The closed-end funds discounts have been an interesting phenomenon for a long period. Some theories based on economic rationale try to solve the puzzle but fail to get consistent conclusions. Recently some theories based on behavioral finance, such as the investor sentiment hypothesis, have been proposed to solve this puzzle. This study examines the investor sentiment hypothesis based on various time-series tests and finds some interesting results. Briefly, our conclusions are as follows: 1. The discounts can vary widely between funds and seem to be persistent in our sample period. 2. The local market indices are cointegrated with the domestic closed-end funds discounts and the information is flowed from the market to the closed-end funds, which support the investor sentiment hypothesis. 3. The causality relationship between the foreign closed-end funds and the local indices is not obvious. Based on this, the market segmentation hypothesis seems to hold in our sample, which indicates that investing in the foreign funds provide investors with the benefit of diversification. For future researches, we suggest that more the foreign funds should be included in the sample and the classification of the degree of investor sentiment and the categories of funds can also be improved.