基于财务指标量化选股的alpha策略可行性研究

发布时间：2018-04-02 22:00

本文选题：财务指标　切入点：量化选股　出处：《浙江工商大学》2017年硕士论文

【摘要】：Alpha策略以其风险敞口小、不受市场所处经济周期的影响等优势在量化投资领域占有不朽的一席之地。寻找有效的alpha因子成为能否获取正的alpha收益的关键,而资本市场中公开披露的财务信息与alpha收益是否存在紧密相关性,能否利用财务信息构建alpha因子,都成为量化策略投资特别关注的问题。本文从alpha收益与财务指标的相关性分析出发,探讨基于财务指标构建alpha因子的可行性。首先,本文根据alpha收益构建的基本原理,通过剔除Fama-French的三因子和五因子模型中财务指标,提出了两因子模型,即只考虑了市场系统风险和市值风险因子的资本资产定价模型,从而使得实证过程中alpha收益包含了不确定的公司财务指标的影响,以此研究alpha收益与财务指标的相关性。在样本时间区间内依据财务指标大小对股票进行排序,得到财务指标排名前20%的股票的平均α收益α1时间序列和财务指标排名后20%的股票的平均α收益α2时间序列,对两组时间序列做t检验来判断两者之间是否有显著差异,并由此判断财务指标与股票alpha收益之间的相关性。其次,通过滚动回归计算所需alpha,并依据财务报表披露日期对财务数据做滞后处理。由于在一段时间内财务指标数值是不变的,alpha是时变的,这就使得两者之间的spearman相关系数是时变的。最后,基于样本内分析所得有效财务因子构建alpha策略,并通过样本外数据检验是否可获取正alpha收益。本文共做了三次样本外投资组合的构建,对所有可能涉及到的持仓区间均进行了实证检验。通过对样本内32个财务指标与股票alpha收益之间的相关性分析得到如下结论:若持仓期为半年,则在10%显著性水平下,不存在与股票alpha收益相关性较为显著的财务指标。若持仓期为四个月,则在5%显著性水平下,每股收益同比增长与alpha收益正相关,资本固定化比率与股票alpha收益负相关。在10%显著性水平下,流动比率、扣除非经常性损益后归属于母公司净利润、流动资产总资产占比、流动负债总负债占比与股票alpha收益正相关。若持仓期为两个月,则在5%显著性水平下,流动比率、应收账款周转天数与股票alpha收益正相关,资本固定化比率与股票alpha收益负相关。在10%显著性水平下,现金比率、现金循环周期、流动负债总负债占比与股票alpha收益正相关,资产负债率与股票alpha收益负相关。反映了相同财务指标在不同的时间区间与股票alpha收益之间的相关性是有所不同的。三组样本外的实证结果表明,即便是在熊市,alpha策略仍然可以使投资者获取正的收益。且与沪深300指数和投资组合每日的涨跌幅相比,投资组合alpha收益的波动率更为平稳,尤其是在市场震动较为剧烈的时候,此优势更为突出。
[Abstract]:The Alpha strategy has an immortal place in the quantitative investment field because of its small exposure and its advantages such as being free from the impact of the business cycle in which the market is located.Finding effective alpha factor becomes the key to obtain positive alpha income. However, whether there is close correlation between the publicly disclosed financial information and the return of alpha in the capital market, and whether to use financial information to construct the alpha factor, can be used to construct the alpha factor.All become the question that quantification strategy investment pays special attention to.Based on the analysis of the correlation between alpha income and financial index, this paper discusses the feasibility of constructing alpha factor based on financial index.Firstly, according to the basic principle of alpha income construction, this paper proposes a capital asset pricing model which only takes market system risk and market value risk factor into account by excluding the financial indexes of Fama-French 's three-factor and five-factor models.Thus, the empirical process of alpha earnings includes the impact of uncertain corporate financial indicators, so as to study the correlation between alpha returns and financial indicators.In the sample time interval, the stock is sorted according to the size of financial index, and the average 伪 return 伪 1 time series of the top 20% stock and the average 伪 return 伪 2 time series of 20% stock after the financial index rank are obtained.T test is made on the two groups of time series to judge whether there is a significant difference between the two groups, and the correlation between the financial indexes and the stock alpha returns is judged.Secondly, the alphaa is calculated by rolling regression, and the financial data is delayed according to the date of financial statement disclosure.The spearman correlation coefficient between the two is time-varying because the value of the financial index is invariant for a period of time.Finally, the alpha strategy is constructed based on the effective financial factors obtained from the in-sample analysis, and the positive alpha income can be obtained by the out-of-sample data.In this paper, we construct the portfolio outside the sample three times, and make an empirical test on all possible positions.By analyzing the correlation between 32 financial indexes and stock alpha returns in the sample, the following conclusions are drawn: if the holding period is half a year, there is no significant correlation between the financial indicators and the stock alpha returns at the level of 10% significance.If the holding period is four months, the growth of earnings per share is positively correlated with alpha earnings at a significant level of 5%, and the capital fixation ratio is negatively correlated with stock alpha earnings.At the significant level of 10%, the current ratio, after deducting the non-recurrent profit and loss, belongs to the net profit of the parent company, the proportion of the total assets of the current assets and the proportion of the total current liabilities are positively related to the alpha return of the stock.If the holding period is two months, at the level of 5% significance, the ratio of current and the days of turnover of accounts receivable are positively correlated with the return of stock alpha, while the ratio of capital immobilization is negatively correlated with the return of stock alpha.At the significant level of 10%, the cash ratio, the cash cycle, the ratio of current liabilities to total liabilities are positively correlated with the stock alpha returns, while the asset-liability ratio is negatively correlated with the stock alpha returns.It reflects that the correlation between the same financial index and the stock alpha return is different in different time range.Empirical results out of three groups of samples show that even in the bear market alpha strategy can still enable investors to achieve positive returns.Compared with the daily rise and fall of Shanghai and Shenzhen 300 index and portfolio, the volatility of portfolio alpha returns is more stable, especially when the market trembles more intensely, this advantage is more prominent.
【学位授予单位】：浙江工商大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：F275;F832.51

【参考文献】