siat 2.13.43__py3-none-any.whl → 2.14.2__py3-none-any.whl

siat/capm_beta2.py

@@ -0,0 +1,615 @@
+# -*- coding: utf-8 -*-
+"""
+本模块功能：CAPM beta
+所属工具包：证券投资分析工具SIAT 
+SIAT：Security Investment Analysis Tool
+创建日期：2024年3月22日
+最新修订日期：2024年3月22日
+作者：王德宏 (WANG Dehong, Peter)
+作者单位：北京外国语大学国际商学院
+作者邮件：wdehong2000@163.com
+版权所有：王德宏
+用途限制：仅限研究与教学使用，不可商用！
+特别声明：作者不对使用本工具进行证券投资导致的任何损益负责！
+"""
+
+#==============================================================================
+#关闭所有警告
+import warnings; warnings.filterwarnings('ignore')
+#==============================================================================
+from siat.common import *
+from siat.translate import *
+from siat.security_prices import *
+from siat.grafix import *
+
+import pandas as pd
+import numpy as np
+#==============================================================================
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    ticker="600519.SS"
+
+def get_market_index_code(ticker):
+    """
+    功能：基于股票ticker确定其所在市场的大盘指数代码
+    """
+    
+    _,_,suffix=split_prefix_suffix(ticker.upper())
+    
+    if suffix in ['SS']:
+        mktidx='000001.SS' #上证综合指数
+    elif suffix in ['SZ']:
+        mktidx='399001.SZ'
+    elif suffix in ['BJ']:
+        mktidx='899050.BJ' #北证50指数   
+    elif suffix in ['CN']:
+        mktidx='000300.SS' #沪深300指数          
+    elif suffix in ['HK']:
+        mktidx='^HSI'      #恒生指数
+    elif suffix in ['TW']:
+        mktidx='^TWII'     #台湾加权指数    
+    elif suffix in ['SI']:
+        mktidx='^STI'      #新加坡海峡时报指数
+    elif suffix in ['T']:
+        mktidx='^N225'     #日经225指数
+    elif suffix in ['KS']:
+        mktidx='^KS11'     #韩国综合指数
+    elif suffix in ['NS','BO']:
+        mktidx='^SNX'      #孟买敏感指数
+    elif suffix =='':
+        mktidx='^SPX'      #标普500指数
+    elif suffix in ['L','UK']:
+        mktidx='^FTSE'     #英国富时100指数
+    elif suffix in ['DE']:
+        mktidx='^DAX'      #德国DAX30指数
+    elif suffix in ['F']:
+        mktidx='^CAC'      #法国CAC40指数
+    else:
+        mktidx='^SPX'      #标普500指数
+        
+    return mktidx
+
+#==============================================================================
+if __name__=='__main__':
+    ticker="600519.SS"
+    ticker={'Market':('US','^SPX','中概教培组合'),'EDU':0.7,'TAL':0.3}
+    
+    start2='2022-10-31'
+    end='2024-3-23'
+    ret_type='Annual Ret%'
+    RF=0.01759
+    regtrddays=252
+    
+    mktidx='auto'; source='auto'
+    
+    reg_result,dretdf3=regression_capm(ticker,start2,end,ret_type,RF,regtrddays)
+
+def regression_capm(ticker,start2,end,ret_type='Annual Ret%',RF=0,regtrddays=252, \
+                    mktidx='auto',source='auto'):
+    """
+    功能：进行CAPM回归，R-Rf=beta*(Rm-Rf)，无截距项回归
+    x为(Rm-Rf)，y为R-Rf，均为日收益率，默认回归样本长度一年（365日历日或252交易日）
+    返回：beta系数
+    注意：回归基于传统的日收益率，而非滚动收益率
+    """
+    DEBUG=False
+
+    #抓取股价，计算股票日收益率
+    if DEBUG:
+        print("*** DEBUG:",ticker,start2,end)
+    #pricedf=get_price(ticker,start2,end,source=source)
+    pricedf=get_price_security(ticker,start2,end,source=source)
+    if pricedf is None:
+        print("  #Error(regression_capm): info of security",ticker_name(ticker),"not found or inaccessible")
+        return None,None
+    
+    #计算股票滚动收益率
+    pricedf1=calc_daily_return(pricedf)
+    
+    #抓取大盘指数，计算指数日收益率
+    """
+    if 'auto' in mktidx.lower():
+        mktidx=get_market_index_code(ticker)
+    """
+    if isinstance(ticker,dict):
+        _,mktidx,pftickerlist,_=decompose_portfolio(ticker)
+        if 'auto' in mktidx.lower():
+            mktidx=get_market_index_code(pftickerlist[0])
+    else:
+        if 'auto' in mktidx.lower():
+            mktidx=get_market_index_code(ticker)
+        
+    marketdf=get_price(mktidx,start2,end,source=source)
+    if marketdf is None:
+        print("  #Error(regression_capm): info of market index",mktidx,"not found or inaccessible")
+        return None,None
+    
+    marketdf1=calc_daily_return(marketdf)
+    
+    #合并股票和大盘指数日收益率
+    dretdf1=pd.merge(marketdf1,pricedf1,how='inner',left_index=True,right_index=True)
+
+    #准备CAPM回归文件
+    ret_type_lower=ret_type.lower()
+    dretname='Daily Ret'
+    #计算日无风险利率
+    RF_daily=RF / 365 
+    if '%' in ret_type_lower:
+        dretname='Daily Ret%'
+        RF_daily=RF_daily * 100
+        
+    dretx=dretname+'_x' #指数日收益率
+    drety=dretname+'_y' #股票日收益率
+    dretdf2=dretdf1[[dretx,drety]]
+    dretdf2.dropna(inplace=True)
+
+    #计算股票和指数收益率的风险溢价R-RF
+    dretdfcols=list(dretdf2)    
+    for c in dretdfcols:
+        dretdf2[c]=dretdf2[c].apply(lambda x: x-RF_daily)
+    dretdf2=dretdf2.reset_index()
+    #dretdf2.rename(columns={'index':'Date'},inplace=True)
+    if 'Date' not in list(dretdf2):
+        dretdf2['Date']=dretdf2['index']
+    
+    #CAPM回归，计算贝塔系数
+    dretnum=len(dretdf2)
+    if regtrddays >= dretnum:
+        regtrddays=dretnum - 31 *2
+        
+    import statsmodels.api as sm
+    reg_result=pd.DataFrame(columns=('Date','beta'))
+    for i in range(dretnum):
+        i2=dretnum-i
+        i1=i2-regtrddays
+        if i1 < 0: break
+    
+        regdf=dretdf2[i1:i2]
+        lastdate=regdf.tail(1)['Date'].values[0]
+        
+        X=regdf[dretx] #无截距项回归
+        Y=regdf[drety]
+        model = sm.OLS(Y,X)	#定义回归模型R-Rf=beta(Rm-Rf)，X可为多元矩阵
+        results = model.fit()	#进行OLS回归
+        beta=results.params[0]	#提取回归系数
+        
+        row=pd.Series({'Date':lastdate,'beta':beta})
+        try:
+            reg_result=reg_result.append(row,ignore_index=True)        
+        except:
+            reg_result=reg_result._append(row,ignore_index=True)
+        
+    reg_result.set_index(['Date'],inplace=True) 
+    reg_result.sort_index(inplace=True) #按日期升序排列 
+    
+    dretdf3=dretdf2.set_index(['Date']) 
+    if 'index' in list(dretdf3):
+        del dretdf3['index']
+    
+    return reg_result,dretdf3
+
+#==============================================================================
+if __name__=='__main__':
+    ticker="600519.SS"
+    ticker={'Market':('US','^SPX','中概教培组合'),'EDU':0.5,'TAL':0.3,'TEDU':0.2}
+    
+    start="2024-1-1"
+    end="2024-3-23"
+    RF=0.01759
+    regtrddays=252
+    mktidx='auto'; source='auto'
+    
+    beta1=get_capm_beta(ticker,start,end,RF,regtrddays)
+    beta1.plot()
+
+def get_capm_beta(ticker,start,end,RF=0,regtrddays=252,mktidx='auto',source='auto'):
+    """
+    功能：套壳函数regression_capm，仅返回滚动的贝塔系数，基于日收益率
+    滚动窗口长度为regtrddays，默认为一年的交易日
+    注意函数regression_capm没有向前调整日期，需要本函数内进行调整。
+    """
+    start2=date_adjust(start,adjust=-365/252 * regtrddays -31*2)
+    
+    reg_result,_=regression_capm(ticker=ticker,start2=start2,end=end, \
+                                 ret_type="Annual Ret%",RF=RF, \
+                                 regtrddays=regtrddays,mktidx=mktidx,source=source)
+    
+    startpd=pd.to_datetime(date_adjust(start,adjust=-7))
+    endpd=pd.to_datetime(end)
+    reg_result2=reg_result[(reg_result.index >= startpd) & (reg_result.index <= endpd)]
+    
+    return reg_result2
+    
+#==============================================================================
+#==============================================================================
+if __name__=='__main__':
+    ticker=["600519.SS","000858.SZ"]
+    ticker={'Market':('US','^SPX','中概教培组合'),'EDU':0.5,'TAL':0.3,'TEDU':0.2}
+    
+    start="2024-1-1"
+    end="2024-3-23"
+    RF=0.01759
+    regression_period=365
+    
+    graph=True; axhline_value=1; axhline_label=''
+    printout=False
+    annotate=False
+    mktidx='auto'
+    source='auto'
+    
+    betas=compare_mticker_1beta(ticker,start,end)
+
+def compare_mticker_1beta(ticker,start,end, \
+                         RF=0,regression_period=365, \
+                         graph=True,axhline_value=1,axhline_label='', \
+                         annotate=False,mktidx='auto',source='auto'):
+    """
+    功能：多只股票，对比其贝塔系数
+    """
+    
+    import os,sys
+    class HiddenPrints:
+        def __enter__(self):
+            self._original_stdout = sys.stdout
+            sys.stdout = open(os.devnull, 'w')
+
+        def __exit__(self, exc_type, exc_val, exc_tb):
+            sys.stdout.close()
+            sys.stdout = self._original_stdout
+    
+    #转换字符串和列表，避免下面的循环出错
+    if isinstance(ticker,str) or isinstance(ticker,dict):
+        ticker=[ticker]
+    if isinstance(RF,list):
+        RF=RF[0]
+    if isinstance(regression_period,list):
+        regression_period=regression_period[0]
+    print("  Starting to retrive and calculate capm beta, please wait ......")
+        
+    #计算日历日regression_period对应的交易日数
+    regtrddays=int(252 / 365 * regression_period)
+    
+    df=pd.DataFrame() 
+    for t in ticker:
+        #关闭print输出
+        with HiddenPrints():
+            df_tmp=get_capm_beta(t,start,end,RF,regtrddays,mktidx,source)
+        
+        if df_tmp is None:
+            break
+        else:
+            dft=df_tmp[['beta']]
+            dft.rename(columns={'beta':ticker_name(t)},inplace=True)
+            
+        if len(df)==0: #第一个
+            df=dft 
+        else:
+            df=pd.merge(df,dft,how='outer',left_index=True,right_index=True)
+
+    if len(df)==0:
+        print("  #Error(compare_mticker_1beta): beta data not available for",t,"between",start,end)        
+        return None
+    
+    #仅用于绘图和制表
+    df1=df.copy()
+    for c in list(df1):
+        #是否绘制水平线
+        if df1[c].max() > axhline_value and df1[c].min() < axhline_value:
+            axhline_label='零线'
+        #df1.rename(columns={c:codetranslate(c)},inplace=True)
+    
+    #共同脚注    
+    footnote1="注："
+    if RF !=0:
+        footnote2="年化无风险利率为"+str(round(RF*100,4))+'%。'
+    else:
+        footnote2="假设年化无风险利率为零。"
+        
+    footnote3="贝塔系数基于日收益率，回归期间跨度为"+str(regression_period)+"个自然日。"
+    
+    import datetime; todaydt = datetime.date.today()
+    footnote4="数据来源: 综合新浪/stooq/Yahoo，"+str(todaydt)+"统计"
+    if footnote3 !='':
+        footnotex=footnote1+footnote2+footnote3+'\n'+footnote4
+    else:
+        footnotex=footnote1+footnote2+footnote3+footnote4
+
+    #绘图
+    if graph:
+        title_txt="CAPM贝塔系数"
+        y_label="贝塔系数"
+
+        draw_lines(df1,y_label,x_label=footnotex, \
+                   axhline_value=axhline_value,axhline_label=axhline_label, \
+                   title_txt=title_txt,data_label=False,annotate=annotate)
+ 
+    return df
+
+#==============================================================================
+if __name__=='__main__':
+    ticker="600519.SS"
+    ticker="000858.SZ"
+    ticker={'Market':('China','000300.SS','白酒组合'),'600519.SS':0.2,'000858.SZ':0.3,'600809.SS':0.5}
+    
+    start="2024-3-11"
+    end="2024-3-23"
+    RF=[0,0.01759,0.05]
+    regression_period=365
+    
+    graph=True; axhline_value=1; axhline_label=''
+    annotate=False
+    mktidx='auto'
+    source='auto'
+    
+    betas=compare_1ticker_mRF(ticker,start,end,RF)
+
+def compare_1ticker_mRF(ticker,start,end, \
+                         RF=[0,0.02,0.05], \
+                         regression_period=365, \
+                         graph=True,axhline_value=1,axhline_label='', \
+                         annotate=False,mktidx='auto',source='auto'):
+    """
+    功能：一只股票，不同的无风险收益率
+    """
+    
+    import os,sys
+    class HiddenPrints:
+        def __enter__(self):
+            self._original_stdout = sys.stdout
+            sys.stdout = open(os.devnull, 'w')
+
+        def __exit__(self, exc_type, exc_val, exc_tb):
+            sys.stdout.close()
+            sys.stdout = self._original_stdout
+    
+    #转换字符串和列表，避免下面的循环出错
+    if isinstance(ticker,list):
+        ticker=ticker[0]
+    if isinstance(RF,float):
+        RF=[RF]
+    if isinstance(regression_period,list):
+        regression_period=regression_period[0]
+    print("  Starting to retrive and calculate capm beta on different RF, please wait ......")
+        
+    #计算日历日regression_period对应的交易日数
+    regtrddays=int(252 / 365 * regression_period)
+        
+    df=pd.DataFrame() 
+    for t in RF:
+        #关闭print输出
+        with HiddenPrints():
+            df_tmp=get_capm_beta(ticker,start,end,t,regtrddays,mktidx,source)
+        
+        if df_tmp is None:
+            break
+        else:
+            dft=df_tmp[['beta']]
+            dft.rename(columns={'beta':"基于无风险利率"+str(round(t*100,4))+'%'},inplace=True)
+            
+        if len(df)==0: #第一个
+            df=dft 
+        else:
+            df=pd.merge(df,dft,how='outer',left_index=True,right_index=True)
+
+    if len(df)==0:
+        print("  #Error(compare_1ticker_mRF): data not available for",ticker,"between",start,end)        
+        return None
+    
+    #仅用于绘图和制表
+    df1=df.copy()
+    for c in list(df1):
+        #是否绘制水平线
+        if df1[c].max() > axhline_value and df1[c].min() < axhline_value:
+            axhline_label='零线'
+        #df1.rename(columns={c:"基于无风险利率"+c},inplace=True)
+    
+    #去掉提前的数据
+    start1=pd.to_datetime(date_adjust(start,adjust=-2))
+    df1=df1[df1.index >= start1]
+    
+    #共同脚注    
+    footnote1="注："
+    footnote2=""
+        
+    footnote3="贝塔系数基于日收益率，回归期间跨度为"+str(regression_period)+"个自然日。"
+    
+    import datetime; todaydt = datetime.date.today()
+    footnote4="数据来源: 综合新浪/stooq/Yahoo，"+str(todaydt)+"统计"
+    if footnote3 !='':
+        footnotex=footnote1+footnote3+'\n'+footnote4
+    else:
+        footnotex=footnote4
+
+    #绘图
+    if graph:
+        title_txt="CAPM贝塔系数："+ticker_name(ticker)
+        y_label="贝塔系数"
+
+        draw_lines(df1,y_label,x_label=footnotex, \
+                   axhline_value=axhline_value,axhline_label=axhline_label, \
+                   title_txt=title_txt,data_label=False,annotate=annotate)
+            
+    return df
+
+#==============================================================================
+if __name__=='__main__':
+    ticker="600519.SS"
+    ticker={'Market':('China','000300.SS','白酒组合'),'600519.SS':0.2,'000858.SZ':0.3,'600809.SS':0.5}
+    
+    start="2024-1-1"
+    end="2024-3-15"
+    RF=0.01759
+    regression_period=[365,183,730]
+    
+    graph=True; axhline_value=1; axhline_label=''
+    annotate=False
+    mktidx='auto'
+    source='auto'
+    
+    betas=compare_1ticker_mregression_period(ticker,start,end,RF,regression_period)
+
+def compare_1ticker_mregression_period(ticker,start,end, \
+                         RF=0, \
+                         regression_period=[183,365,730], \
+                         graph=True,axhline_value=1,axhline_label='', \
+                         annotate=False,mktidx='auto',source='auto'):
+    """
+    功能：一只股票或一个投资组合，不同的回归期间
+    """
+    
+    import os,sys
+    class HiddenPrints:
+        def __enter__(self):
+            self._original_stdout = sys.stdout
+            sys.stdout = open(os.devnull, 'w')
+
+        def __exit__(self, exc_type, exc_val, exc_tb):
+            sys.stdout.close()
+            sys.stdout = self._original_stdout
+    
+    #转换字符串和列表，避免下面的循环出错
+    if isinstance(ticker,list):
+        ticker=ticker[0]
+    if isinstance(RF,list):
+        RF=RF[0]
+    if isinstance(regression_period,int):
+        regression_period=[regression_period]
+    print("  Starting to retrive and calculate capm beta on different regression period, please wait ......")
+        
+    df=pd.DataFrame() 
+    for t in regression_period:
+        #计算日历日regression_period对应的交易日数
+        regtrddays=int(252 / 365 * t)
+        
+        #关闭print输出
+        with HiddenPrints():
+            df_tmp=get_capm_beta(ticker,start,end,RF,regtrddays,mktidx,source)
+        
+        if df_tmp is None:
+            break
+        else:
+            dft=df_tmp[['beta']]
+            dft.rename(columns={'beta':"基于"+str(t)+"自然日回归"},inplace=True)
+            
+        if len(df)==0: #第一个
+            df=dft 
+        else:
+            df=pd.merge(df,dft,how='outer',left_index=True,right_index=True)
+
+    if len(df)==0:
+        print("  #Error(compare_1ticker_mregression_period): data not available for",ticker,"between",start,end)        
+        return None
+    
+    #仅用于绘图和制表
+    df1=df.copy()
+    for c in list(df1):
+        #是否绘制水平线
+        if df1[c].max() > axhline_value and df1[c].min() < axhline_value:
+            axhline_label='零线'
+        #df1.rename(columns={c:"基于"+str(c)+"自然日回归"},inplace=True)
+        
+    #共同脚注    
+    footnote1="注："
+    footnote2=""
+        
+    footnote3="贝塔系数基于日收益率，无风险利率为"+str(round(RF*100,4))+'%'
+    
+    import datetime; todaydt = datetime.date.today()
+    footnote4="数据来源: 综合新浪/stooq/Yahoo，"+str(todaydt)+"统计"
+    if footnote3 !='':
+        footnotex=footnote1+footnote3+'\n'+footnote4
+    else:
+        footnotex=footnote4
+
+    #绘图
+    if graph:
+        title_txt="CAPM贝塔系数："+ticker_name(ticker)
+        y_label="贝塔系数"
+
+        draw_lines(df1,y_label,x_label=footnotex, \
+                   axhline_value=axhline_value,axhline_label=axhline_label, \
+                   title_txt=title_txt,data_label=False,annotate=annotate)
+            
+    return df
+
+
+#==============================================================================
+# 合成函数
+#==============================================================================
+if __name__=='__main__':
+    ticker="600519.SS"
+    ticker=["600519.SS","000858.SZ"]
+    ticker={'Market':('US','^SPX','中概教培组合'),'EDU':0.5,'TAL':0.3,'TEDU':0.2}
+    
+    start="2024-1-1"; end="2024-3-20"
+    
+    RF=0.01759
+    RF=[0.005,0.01759,0.05]
+    
+    regression_period=365
+    
+    graph=True
+    annotate=False
+    source='auto'
+    
+    betas=compare_beta_security(ticker,start,end,RF)
+
+def compare_beta_security(ticker,start,end,RF=0,regression_period=365, \
+                          graph=True,annotate=False,mktidx='auto',source='auto'):
+    """
+    功能：组合情况，可能多只股票、多个投资组合或投资组合与股票的混合，多个无风险收益率
+    
+    """
+    
+    #情形1：多个证券
+    if isinstance(ticker,list):
+        if len(ticker) > 1:
+            if isinstance(RF,list):
+                RF=RF[0]  
+                
+            df=compare_mticker_1beta(ticker,start,end,RF,regression_period, \
+                                     graph, \
+                                     annotate=annotate,mktidx=mktidx,source=source)
+            return df
+        else:
+            #实际上是单个证券
+            ticker=ticker[0]                
+    
+    #情形2：1只证券，多个RF。时间区间要尽可能短，不然难以看出差异！
+    if isinstance(RF,list):
+        if len(RF) > 1:
+            df=compare_1ticker_mRF(ticker,start,end,RF,regression_period, \
+                                     graph, \
+                                     annotate=annotate,mktidx=mktidx,source=source)
+            return df
+        else:
+            #实际上是单个RF
+            RF=RF[0]       
+    
+    #情形3：1只证券，多个回归天数
+    if isinstance(regression_period,list):
+        if len(regression_period) > 1:
+            df=compare_1ticker_mregression_period(ticker,start,end, \
+                                     RF,regression_period, \
+                                     graph, \
+                                     annotate=annotate,mktidx=mktidx,source=source)
+            return df
+        else:
+            #实际上是单个regression_period
+            regression_period=regression_period[0]  
+            
+    #情形4：1只证券，1个RF，1个回归天数
+    df=compare_mticker_1beta(ticker,start,end,RF,regression_period, \
+                             graph,annotate=annotate,mktidx=mktidx,source=source)
+    return df
+
+
+#==============================================================================
+#==============================================================================
+
+#==============================================================================
+#==============================================================================
+#==============================================================================
+#==============================================================================
+#==============================================================================