馬克維茲的投資組合理論是現代投資組合理論的基礎,平均數-變異數效率投資組合在其中扮演著重要的角色,在運用平均數-變異數效率投資組合所建立的投組中往往包含極端的空倉及多倉,然實務上許多機構投資人是無法持有空倉部位的,因此在其建立效率投資組合時,常給予投資組合不得放空及投資上限等投資限制。 本文以台灣股市為研究對象,資料來源為TEJ資料庫。自1975年至2004年,每年隨機選取55支上市股票作為投資組合的投資標的,並以夏普指數來衡量是否可以進行放空交易及在有最大投資上限的投資組合其樣本外的績效表現。 我們的實證結果發現,不得放空的投資組合夏普值大於可以放空投組的夏普值,在不得放空的情況下對投資組合訂定投資上限,並不會明顯改善投資組合的績效。 The Markowitz portfolio theory is the foundation of modern portfolio theory. The mean-variance efficient portfolios play an important role in the theory. However, the efficient portfolios constructed by the mean-variance method often involve taking on extreme long and short positions, but institutional investors are generally not allowed to hold short positions and are required to meet certain diversification requirements. Hence, the institutional investors often impose non negativity and some kind of upper bounds on portfolio weights when constructing portfolio. In this paper, we investigate the merit of short selling constraints on the Taiwan stock exchange (TSE). From 1975 to 2004, we randomly select 55 stocks listed on TSE to construct equity portfolios. We use the Sharpe Ratio to compare the out-of-sample performances of portfolios with and without short-selling and maximum weights constraints. Our empirical results show that the Sharpe Ratios of no-short selling portfolios outperform the unconstrained portfolios. The imposed upper bounds on portfolio weights do not lead to a significant improvement in performance when no-short selling restrictions are in place.