[[alternative]]The investigation on the volatility of stock returns in Taiwan stock market

碩士[[abstract]]本研究以GARCH(1,1)估計股票報酬波動，並將”個別風險佔市場風險比率”加入Farm & Franch三因子模型作為第四因子。探討Beta、公司規模、淨值市價比及新加入的個別風險佔市場風險比率因子與股票報酬之關係。並且，本研究依獨特性風險(Idiosyncratic Volatility)將樣本資料由小到大分為5個投資組合，觀察是否在不同獨特性風險下，各因子對報酬解釋能力有不同的改變。 本研究以panel data模型進行分析研究，實證結果得到： 1. 由三因子模型觀察，Beta與淨值市價比對股票報酬為顯著負相關，公司規模為顯著正相關。 2. 由四因子模型觀察，新加入第四因子-個別風險佔市場風險比 率，與報酬呈顯著正相關，Beta與淨值市價比對股票報酬為顯著負相關，公司規模為顯著正相關。 3. 依獨特性風險分類後觀察四因子模型，Beta與淨值市價比在所有投資組合下，對股票報酬皆為顯著負相關；公司規模僅於最小及最大獨特性風險(portfolio 1、portfolio 5)時顯著正相關，其餘則不顯著；個別風險佔市場風險比率因子在最小獨特性風險下時，對報酬關係顯著負相關，在portfolio 2時對報酬沒有顯著解釋能力，而在portfolio 3 ~ portfolio 5則為顯著正相關。[[abstract]]In this study, we estimate the volatility of return by GARCH(1,1). We use the ratio of the volatility of each stock return to TAIEX proxy for aggregate volatility risk, and we add Fama & French three-factor model as the fourth risk factor-ratio. Here we are going to discuss the relationship among these four risk factors (Beta, ln(ME), BE/ME, ratio) and the stock return. And, we sort stocks based on idiosyncratic volatility into 5 portfolios. Portfolio 1(5) is the portfolio of stocks with the lowest (highest) idiosyncratic volatility. We want to know if these four factors are still significant in explaining stock returns in different portfolios. We analyze our empirical data with Panel data model. The empirical results show that: 1. From the three factors model: We find there is negative relation between Beta（or BE/ME) and stock return and positive relation between market value and stock return. 2. From the four factors model: We find there is negative relation between Beta（or BE/ME）and stock return and positive relation between market value（or ratio）and stock return. 3. After sort stocks based on idiosyncratic volatility into 5 portfolios. We find there is also negative relation between Beta（or BE/ME）in all portfolios but market value is only significant in the portfolio 1 and portfolio 5. The fourth risk factor-ratio is negative relation to return in the portfolio 1 and positive in portfolio 3,4 and 5. But the factor-ratio has no significant explanation power of return in portfolio 2.[[tableofcontents]]目 錄 第一章 緒論...............................................1 第一節 研究背景與動機.....................................1 第二節 研究目的......................................2 第三節 研究流程......................................3 第四節 研究架構......................................4 第二章 文獻回顧...........................................5 第一節 風險因子相關文獻探討..........................5 第二節 股票報酬波動相關文獻探討.....................14 第三章 研究方法..........................................20 第一節 Panel data模型...............................20 第二節 GARCH模型....................................30 第三節 獨特性風險之估計..................................34 第四章 實證結果..........................................36 第一節 資料來源與變數定義...........................36 第二節 Panel data模型分析...........................39 第五章 結論與建議........................................51 第一節 研究結論.....................................51 第二節 研究限制.....................................53 第三節 建議.........................................54 參考文獻.................................................55 附錄: portfolio 1~ portfolio 5之成分股...................58 表次 【表一】三因子－Hausman Test.............................41 【表二】三因子－ F Test..................................41 【表三】三因子－Panel data模型...........................41 【表四】因子間相關係數...................................42 【表五】四因子－Hausman Test.............................43 【表六】四因子－ F Test..................................43 【表七】四因子－Panel data模型...........................43 【表八】各投資組合之Hausman Test及F Test.................47 【表九】各投資組合之Panel data分析.......................49 【表十】各投資組合之敘述統計.............................50[[note]]學號: 694490441, 學年度: 9