本研究運用相對評價法中之PBV(Price to Book Ratio)乘數，進行橫斷面 迴歸分析，判斷股價低估(undervalue)或高估(overvalue)，並利用判斷結果進行投資組合部位建立。選取樣本利用台灣50 指數成分股排除公司規模之影響，而研究結果顯示，利用PBV 低估的股票進行多部位建立，在研究期間 全期中，可以明顯超越市場報酬，但在空部位建立則無法超越市場報酬。 研究中並發現，極端值對績效造成影響，其因素為金控及銀行類股與景氣 循環股產業特性所致。 / This research applied the multiple of the price to book value in the relative valuation to begin cross-section regression analysis and to evaluate if the stock price was overpriced or underpriced. The result can be used to develop the investment portfolios. This research used the constituents in Taiwan 50 Index as samples in order to avoid the size effect. During the whole period, we found that we could gain excess profit obviously and defeat the market return by utilizing those stocks underpriced in PBV to develop the long position. However, we could not defeat the market return by developing the short position. In the research, we also found that the extreme value could affect the investment return, which result mainly from both stocks of financial holding companies and banks and stocks of business cycle industries.

投資組合保險的概念興起於1980年代，其對於無法承受巨大損失風險或是對於股市未來走勢不清楚的投資人而言，是一種很好的投資策略，既可以保障原本所投資的本金，又可以參與上方的獲利。近年來隨著市場發展與金融創新，亦有許多新金融商品依據其概念設計商品，為投資人提供許多較存放定期存款更好的投資選擇。 本研究以台灣上市公司（金融保險類除外）為研究對象，探討選股策略搭配投資組合保險操作進行實證研究。在選股策略的採用以相對評價法中股價/盈餘比、股價/淨值比與股價/銷售比三項選股指標策略所形成之高低兩類投資組合，並搭配固定比率投資組合保險策略、時間不變性投資組合保險策略、固定比例策略及買入持有策略，尋找面對多頭、空頭及盤整不同狀態市場下之最佳投資策略與探討搭配投資組合保險策略後對投資組合績效表現與避險效果的影響。 實證發現，低股價/盈餘比、低股價/淨值比及低股價/銷售比這三類組合在市場下跌及盤整的狀況報酬率均勝過大盤，表現相對抗跌。搭配投資組合保險策略後皆能達到設定的保本比率要求。從機會成本的角度分析，低股價/盈餘比及低股價/銷售比二類組合搭配投資組合保險策略皆損失了機會成本，至於高股價/盈餘比、高股價/淨值比及高股價/銷售比此三類組合則在大盤上漲的年度內有機會成本的損失；從長期相對平均成本的觀點，買入持有策略仍較投資組合保險策略為佳。 在市場上漲年度，高股價/盈餘比、高股價/淨值比及高股價/銷售比此三類組合搭配CPPI策略有最佳的報酬率；大盤盤整年度，六類組合均以搭配CM策略表現最佳。至於六類投資組合在大盤下跌時分別該搭配何種策略較佳，在實證期間內則沒有一致的結論。

In this article, the author is trying (1) to know more about the Internet industry, theoretically and practically, (2) to know the strengths and weaknesses of Internet companies with different business models, and the criteria that can be used to tell a winner from a would-be loser, and (3) to arrive fundamental approaches that could be used or that would be suitable to the valuation of a given Internet stock. They are mainly because the law of increasing return, first mover advantage, brand name advantage and model consolidation that make the competitiveness of leading Internet companies increase continually, and make it harder and harder to survive for new entrants. The author compares different valuation models including dividend discount model, discount free cash flow method, adjusted discount free cash flow method, EVA method, P/S, P/B, P/E multiples, and other non-financial multiples such as page view multiples and unique visitor multiples. Among them, the adjusted discount free cash flow method is the most suitable model to the valuation of Yahoo! and other leading portals that already report positive earnings or can be sure to make money in the near future. Other valuation models cannot be used to the valuation of portals.