運用Fama and French模型與財務指標建構股票投資策略之研究---以台灣五十指數成分股公司

An Empirical Application of Fama-French Model and Financial Ratio to Structure Stock Investment Stategies—an Evidence from Taiwan Top 50 index company

陳文岷(WEN-MIN CHEN)

指導教授 ： 古永嘉

國立臺北大學/商學院/國際財務金融碩士在職專班/碩士(2014年)

摘要

本研究運用Fama and French模型與財務指標，試圖建構有效的選股投資策略。以台灣證券交易所上市之台灣五十指數成分股為研究對象，擷取TEJ資料庫具完整資料的28家公司，自2008年10月至2013年9月連續五年之60筆資料，共計1,680個樣本。藉由益本比、股價淨值比、股利殖利率、營收成長率、營業毛利成長率、營業利益成長率、總資產成長率、淨值成長率等8項財務比率對股價之迴歸模型，並利用移動窗格法，預測該股理論股價與實際股價之差異，進行買進或賣出之投資決策。以持有期間1～6個月進行模擬交易，並計算其年化獲勝率、獲利金額及報酬率，再以GLM/DUNCAN檢定其是否具顯著差異。實證研究結果顯示：
一、股價淨值比、營收成長率、總資產成長率及淨值成長率對股價有正向顯著影響;股利殖利率對股價有負向顯著影響。
二、以移動窗格法之模擬交易結果：
（一）整體28家公司若不分個股及持有期間，則年化平均每股交易獲利金額及報酬率，均呈現虧損情形，獲勝率為43.82%不足五成。
（二）就個股分析，寶成、華碩、大立光、中鋼及仁寶等五家公司呈現獲利情形。其中寶成及華碩之獲勝率超過64%，且年化報酬率達14%以上。

關鍵字

移動窗格法 ； 逐步迴歸 ； 迴歸分析 ； 交易策略

並列摘要

The purpose of this study is to search an effective stock investment strategy by using Fama-French model and financial indicators. Based on Taiwan top 50 Index companies listed in Taiwan Stock Exchange and data availability from the TEJ database, 28 company’s monthly data from Oct. 2008 to Sept. 2013, a total 1,680 observations, were collected. Eight explanatory variables, including earning-price ratio, price to book ratio, dividend yield, revenue growth, operating margin growth, operating income growth, total assets growth, and equity growth, were selected for regression analysis. Using moving window method and comparing actual price to predicted theoretical price, buy-sell investment decision was made for each stock. Holding periods are from 1 to 6 months. The annualized winning rates, the amount of profit/loss and investment returns were calculated. The GLM/DUNCAN statistical tests were applied to check differential effects on holding periods. Empirical results showed that：
1. The price to book ratio, revenue growth, total assets growth, and equity growth have significant positive impacts on stock price; while dividend yield has significant negative impact on stock price.
2. Using moving window method for simulating trading strategies, the results showed: (1) For the overall 28 firms without differentiating holding periods, the average annualized net profits and returns resulted in negative values, and the winning rate is a poor 43.82%. (2) For individual stock simulation, 5 company stocks, including ASUS, POU CHEN, LARGAN, CSC and COMPAL, showed positive profits. In particular, POU CHEN and ASUS had winning rate exceeding 64% and annualized trading returns exceeding 14%

並列關鍵字

moving window method ； stepwise regression analysis ； regression analysis ； trading strategy

參考文獻

6. 林昭芃(2007)，股市之價值溢酬及多因子模型之探討-以台灣股票市場為例，中央大學產業經濟研究所未出版之碩士論文。

11. 黃雅新(2009)，投資組合最適權重動能策略之實證研究-以塑化類股及半導體類股為例，台北大學企業碩士在職專班碩士論文。

7. 許哲瑋(2010)，Fama and French 三因子模型與財務指標對股票報酬之影響--以台灣上市電子公司為例，台北大學企業管理學系碩士論文。

1. Banz,R.W.(1981)“The Relationship between Return and Market Value of Common Stocks, ”Journal of Economics, 6, pp.103-126。

2. Barber,B.M. and Lyon,J. D.(1997) , “Firm Size,Book-to-Market Ratio, and Security Returns︰A Holdout Sample of Financial Firm,”Journal of Finance, 52, pp. 875-883。