| dc.contributor.advisor | 羅秉政 | zh_TW |
| dc.contributor.advisor | Kendro Vincent | en_US |
| dc.contributor.author (Authors) | 陳碩川 | zh_TW |
| dc.contributor.author (Authors) | Chen, Shou-Chuan | en_US |
| dc.creator (作者) | 陳碩川 | zh_TW |
| dc.creator (作者) | Chen, Shou-Chuan | en_US |
| dc.date (日期) | 2025 | en_US |
| dc.date.accessioned | 1-Jul-2025 15:18:00 (UTC+8) | - |
| dc.date.available | 1-Jul-2025 15:18:00 (UTC+8) | - |
| dc.date.issued (上傳時間) | 1-Jul-2025 15:18:00 (UTC+8) | - |
| dc.identifier (Other Identifiers) | G0112352024 | en_US |
| dc.identifier.uri (URI) | https://nccur.lib.nccu.edu.tw/handle/140.119/157838 | - |
| dc.description (描述) | 碩士 | zh_TW |
| dc.description (描述) | 國立政治大學 | zh_TW |
| dc.description (描述) | 金融學系 | zh_TW |
| dc.description (描述) | 112352024 | zh_TW |
| dc.description.abstract (摘要) | 本論文應用主成分投資組合（Principal Portfolio）方法，分析台灣股市中基於異常報酬的投資策略。研究中以 64 個來自財務報表的異常因子構成預測信號矩陣，透過對預測報酬矩陣進行特徵值分解，萃取出三類正交主成分投資組合：主成分投組（PP）、對稱主成分投組（PEP）與反對稱主成分投組（PAP）。
實證結果顯示，反對稱組合（PAP1– PAP3）在風險調整後的報酬表現上優於其他類型，特別是在一個月落後報酬作為信號的設定下（S_lag1）。進一步與 Fama-French 五因子模型的迴歸分析亦顯示，PAP 組合的報酬較少受到傳統風險因子的解釋，凸顯其潛在的交叉預測能力。
此外，本研究亦比較不同信號定義（如落後報酬與累積報酬）對組合表現的影響。結果指出，雖然累積報酬可提升對稱組合的穩定性，但同時可能削弱反對稱組合的預測能力。整體而言，本研究驗證主成分投資組合分解法能有效提升信號解釋性、捕捉異常報酬中的交叉預測結構，並具備應用於新興市場資產配置的潛力。 | zh_TW |
| dc.description.abstract (摘要) | This thesis applies the Principal Portfolio framework to anomaly-based strategies in Taiwan’s stock market. Using 64 firm-level accounting signals, we construct a predictive matrix and derive three types of orthogonal portfolios: Principal Portfolios (PP), Symmetric Principal Portfolios (PEP), and Antisymmetric Principal Portfolios (PAP), via eigenvalue decomposition.
Empirical results show PAP1– 3 achieves the highest risk-adjusted returns under the one-month lag signal (S_lag1), outperforming other strategies. Regression analysis confirms these returns are less explained by Fama-French factors, suggesting strong cross-predictive signals.
We also compare lagged and cumulative signals. While cumulative returns benefit symmetric portfolios, they weaken antisymmetric performance. Overall, the PAP structure best captures anomaly interactions, offering superior alpha and robustness in portfolio construction. | en_US |
| dc.description.tableofcontents | 1 Introduction 1
2 Literature Review 3
2.1 Cross Predictability 3
2.2 Fundamental Signals 4
2.3 Taiwan Anomaly Portfolios 4
3 Methodology 6
3.1 Data 6
3.1.1 Data Source 6
3.1.2 Data Preprocessing 7
3.2 Constructing Anomaly Portfolios 7
3.2.1 Framework and Notation 7
3.2.2 Linear Trading Strategies 8
3.3 The Prediction Matrix 10
3.3.1 Positive Own-Predictability 10
3.3.2 General Linear Strategies 11
3.4 Construction of PP, PEP, and PAP 11
3.4.1 Principal Portfolios (PP) 11
3.4.2 Decomposing Alpha and Beta Strategies 13
3.4.3 Symmetric Strategies: PEPs 14
3.4.4 Antisymmetric Strategies: Principal Alpha Portfolios (PAPs) 16
3.5 Features Introduction 18
3.5.1 Profitability Variables 18
3.5.2 Solvency Variables 22
3.5.3 Other Variables 25
3.6 In-Sample and Out-of-Sample Settings for PP, PEP, and PAP 30
3.6.1 Rolling Window Specification 30
3.6.2 Training-Period Optimization and Decomposition 30
3.7 Portfolio Evaluation Criteria 32
3.7.1 Return and Risk Measures 32
3.7.2 Factor-Based Regression Analysis 33
4 Empirical Results 35
4.1 Descriptive Statistics of Anomaly Portfolios 35
4.2 Performance of Principal Portfolio 45
4.2.1 In-sample Average Returns of Principal Portfolios 45
4.2.2 Time Variation of Returns of Portfolios During In-Sample and Out-of-Sample Periods 48
4.2.3 Anomaly Portfolio Weights in the Principal Portfolios 56
4.2.4 Performance Evaluation of Principal Portfolios 64
4.3 Cross Predictability of Anomaly Portfolios 65
4.4 Analysis of Different Signals 67
5 Conclusions 72
Reference 74
Appendices 75
A Mathematical Foundations of Principal Portfolio Analysis (PPA) 75
A.1 Strategies Based on Π 75
A.2 Optimization Framework for Principal Portfolios 75
A.2.1 Link to Predictive Regression 76
A.2.2 Connection to Mean-Variance Portfolio Optimization 76
A.2.3 Dimension-Reduced Objective in PPA 77
A.2.4 Principal Portfolio Objective 77
A.2.5 Matrix Norm Constraint 78
A.2.6 Robustness to Risk Scaling 78
A.2.7 Summary and Implications 78
B Mathematical Derivation for Principal Portfolios 79
B.1 Optimal Linear Strategies of Principal Portfolios 79
B.2 Construction and Expected Return of Symmetric Strategies: PEPs 80
B.3 Antisymmetric Strategies: Principal Alpha Portfolios (PAPs) 81
B.3.1 Rank-2 Antisymmetric Strategies 81
B.3.2 Decomposition of Antisymmetric Matrices 81
B.3.3 Beta-Neutral Strategies 82
C Static and Dynamic Bets 83
C.1 Decomposition of the Prediction Matrix 83
C.2 Static Bets 84
C.3 Dynamic Bets 84
C.4 Implications for Portfolio Design 85
D Two-Month Lag Returns as Signals 86
E Three-Month Cumulative Returns as Signals 102
F Four-Month Cumulative Returns as Signals 118
G Six-Month Cumulative Returns as Signals 134 | zh_TW |
| dc.format.extent | 45394988 bytes | - |
| dc.format.mimetype | application/pdf | - |
| dc.source.uri (資料來源) | http://thesis.lib.nccu.edu.tw/record/#G0112352024 | en_US |
| dc.subject (關鍵詞) | 主成分投資組合 | zh_TW |
| dc.subject (關鍵詞) | 異常策略 | zh_TW |
| dc.subject (關鍵詞) | 信號矩陣 | zh_TW |
| dc.subject (關鍵詞) | 對稱與反對稱分解 | zh_TW |
| dc.subject (關鍵詞) | 風險調整報酬 | zh_TW |
| dc.subject (關鍵詞) | Fama-French 五因子模型 | zh_TW |
| dc.subject (關鍵詞) | 交叉預測能力 | zh_TW |
| dc.subject (關鍵詞) | Principal Portfolio | en_US |
| dc.subject (關鍵詞) | Anomaly Strategy | en_US |
| dc.subject (關鍵詞) | Signal Matrix | en_US |
| dc.subject (關鍵詞) | Symmetric and Antisymmetric Decomposition | en_US |
| dc.subject (關鍵詞) | Risk-adjusted Return | en_US |
| dc.subject (關鍵詞) | Fama-French Five-Factor Model | en_US |
| dc.subject (關鍵詞) | Cross-Predictability | en_US |
| dc.title (題名) | 使用主投資組合探討台灣股票市場的基本面指標信號 | zh_TW |
| dc.title (題名) | Exploring Fundamental Indicator Signals in the Taiwan Stock Market by the Principal Portfolio Analysis | en_US |
| dc.type (資料類型) | thesis | en_US |
| dc.relation.reference (參考文獻) | Abarbanell, J. S. and Bushee, B. J. (1998). Abnormal returns to a fundamental analysis strategy. The Accounting Review, 73(1):19–45.
Cakici, N., Fieberg, C., Metko, D., and Zaremba, A. (2024). Do anomalies really predict market returns? new data and new evidence. Review of Finance, 28(1):1-44.
Dong, X., Li, Y., Rapach, D. E., and Zhou, G. (2021). Anomalies and the expected market return. The Journal of Finance, 77(1):639–681.
Greig, A. C. (1992). Fundamental analysis and subsequent stock returns. Journal of Accounting and Economics, 15(2):413–442.
Harry Markowitz, John, G. G. X. B. B. (2021). Financial anomalies in portfolio construction and management. The Journal of Portfolio Management, 47(6):51–64.
Kelly, B., Malamud, S., and Pedersen, L. H. (2023). Principal portfolios. The Journal of Finance, 78(1):347–387.
Müller, S. (2017). Economic links and cross-predictability of stock returns: Evidence from characteristic-based “styles”. Review of Finance, 23(2):363–395.
Olson, D. (2001). Cross-correlations and predictability of stock returns. Journal of Forecasting, 20(2):145–160.
Wang, M.-C. and Ding, Y.-J. (2020). Does the quarterly accrual anomaly exist in taiwan’s stock market? evidence from manager’s earnings management. Managerial and Decision Economics, 42(3):688–701.
Yan, X. S. and Zheng, L. (2017). Fundamental analysis and the cross-section of stock returns: A data-mining approach. The Review of Financial Studies, 30(4):1382–1423. | zh_TW |
