中文题名: | 多因子模型的因子选择与策略构建 |
姓名: | |
学科名称: | 金融工程 |
学生类型: | 学士 |
学位名称: | 经济学学士 |
学校: | 中国人民大学 |
院系: | |
专业: | |
第一导师姓名: | |
完成日期: | 2016-05-26 |
提交日期: | 2016-05-26 |
外文题名: | The Factor Selection and Strategy Construction in Multi Factor Alpha Model |
中文关键词: | |
外文关键词: | Arbitrage Pricing Theory ; Multi Factor Model ; Factor Selection |
中文摘要: |
随着套利定价理论的问世,多因子模型受到了大量理论与实证研究的关注。本文旨在 发现影响股票收益的基本面及市场信息因子,在沪深A股的股票池中构建有较强预测能力 的多因子模型;并基于此模型,通过有效因子选择股票,建立市场中性策略来评估因子的 预测、解释股价能力。本文参照BARRA(1998)的因子分类方式,本文将40个因子共分为 八组进行组内、组间筛选:传统估值类、相对价值类、历史成长类、预期成长类、盈利趋 势类、销售增长类、价格动量类、价格反转类、市场规模类。 根据主成分分析、IC等指标筛选出的最显著的共同因子为:TV组PC5,HG组PC3,PT 组PC2,PM组PC1,PR组PC1及PC3,RV组PC4,SS组的PC1。根据打分分类、及策略回 测,都可以看出共同因子能产生持续的超额收益。 关键词:套利定价理论;多因子模型 因子筛选;回归预测;市场中性策略 |
外文摘要: |
With the advent of Arbitrage Pricing Theory, the focus of theoretical and empirical research has been attracted to the Multi-Factor Model. This study also aims to develop a multi-factor model especially for China Shanghai and Shenzhen A-share stocks. By analyzing the most efficient fundamental and market information factors, the study picks up 40 common factors for model building and constructs a market neutral strategy to test factors’ forecastability. Referring to the BARRA’s handbook (1998), the study grouped the factor universe into eight categories: Traditional Value, Relative Value, Historical Growth, Expected Growth, Profit Trend, Accelerating Sales, Price Momentum, Price Reversal and Small Size. After applying the PCA and IC criteria to factor selection, the study yields ten most efficient common factors: TV’s PC5, HG’s PC3, PT’s PC2, PM’s PC1, PR’s PC1 and PC3, RV’s PC4, SS’s PC1. Those common factors proved to generate consistent excessive returns during the quintiles examination and the back tests of market neutral strategies stemmed from them. Key Words: Arbitrage Pricing Theory; Multi Factor Model; Factor Selection ;Regression and Prediction; Market Neutral Strategy |
总页码: | 37 |
参考文献: | ﹀ |
开放日期: | 2016-05-27 |