本研究主要探討不動產投資信託是否可以增加投資人分散投資之效益，以Markowitz (1952)
所提出的平均數-變異數模型為基礎建構多條效率前緣，並使用Kan and Zhou (2008) 提出的平均數-變異數擴張檢定方法進行效率前緣是否擴張之檢定。本研究建立三個假設情況並選擇台灣與幾個國家而進行檢驗，探討REITs 對於該國投資人之分散投資效果，以及投資期間，還有投資
人之基礎投資組合對於REITs 分散效果之影響。實證結果顯示美國、日本、新加坡與台灣地區
之REITs 可替該國投資人獲得多元投資之效益，對於台灣地區之股票投資者而言，納入多國的REITs 投資比僅單一納入台灣REITs 更具分散投資之效益。其次在投資期間檢驗方面，發現澳洲REITs 對於該國股票投資人的分散投資效益會隨著時間而有顯著的改變，因此必須適時調整投資組合內資產的投資權重以獲得最適之資產配置，但美國、日本、新加坡與台灣地區之REITs
的分散投資效果並未隨著投資期間的改變而有顯著之改變。最後實證研究也發現若投資人除了納入台灣股票投資與台灣REITs 外，再新增債券投資，則可更進一步降低資產組合之風險。

The research examines whether the addition of real estate investment trust can significantly expand the mean-variance efficient frontier beyond stocks to provide substantial diversification benefits to investors. This study estimate efficient frontiers based on the mean-variance model by Markowitz and also uses the mean-variance spanning tests to examine whether adding a REIT portfolio can significantly improve return or reduce risk. We proposed three hypothesized portfolio to
analyze several selected markets. The empirical results indicate that in America, Japan, Singapore and Taiwan, REITs are able to provide diversification benefit. For the Taiwan stock market, the portfolio with adding REITs of various countries jointly can outperform the one with adding only T-REITs. Also, the empirical result shows that the diversification benefit by Australian REITs is unstable and investor should rebalance their portfolio weight in LPTs. Finally, our result indicate that in addition to adding T-REITs, investors can further increase the investment performance substantially by adding the investment in government bonds.