Yingjie

回归方程带约束模型检验

薛英杰 / 2023-11-07


读入数据

pacman::p_load(data.table,plyr,stringr,dplyr,mongolite,future.apply,future,xlsx,
               zoo,readxl,foreign,plyr,stringr,pglm,doSNOW,mongolite,
               dplyr,tidyr,officer,flextable,formattable,semPlot,future.apply,
               stargazer,ggthemes,ggplot2,Rmisc,plot3D,panelvar,lubridate,
               readr,rlist,AER,psych,mediation,lavaan,imager,ldt,systemfit,car)


##读入数据
data<-read_excel("E:\\科研\\中航基金\\SUR\\模拟结果2.xlsx")
names(data)[10]<-"IRR_annR"

宽基

IIR检验

kuanji<-data|>
  filter(portfolio_name=="宽基")
kuanjifitIRR=lm(IRR~MKT+SMB+HML+CMA+RMW,kuanji)
summary(kuanjifitIRR)
## 
## Call:
## lm(formula = IRR ~ MKT + SMB + HML + CMA + RMW, data = kuanji)
## 
## Residuals:
##       Min        1Q    Median        3Q       Max 
## -0.117388 -0.012287  0.000365  0.013220  0.076020 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.041485   0.001387  29.906  < 2e-16 ***
## MKT          1.061054   0.008054 131.740  < 2e-16 ***
## SMB         -0.231025   0.012292 -18.794  < 2e-16 ***
## HML          0.115616   0.025115   4.603 4.69e-06 ***
## CMA         -0.160606   0.028045  -5.727 1.36e-08 ***
## RMW          0.044932   0.025831   1.739   0.0823 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.02158 on 994 degrees of freedom
## Multiple R-squared:  0.9472,	Adjusted R-squared:  0.9469 
## F-statistic:  3567 on 5 and 994 DF,  p-value: < 2.2e-16

IIR约束模型检验

linearHypothesis(kuanjifitIRR,c("MKT=0","SMB=0","HML=0","CMA=0","RMW=0"))
## Linear hypothesis test
## 
## Hypothesis:
## MKT = 0
## SMB = 0
## HML = 0
## CMA = 0
## RMW = 0
## 
## Model 1: restricted model
## Model 2: IRR ~ MKT + SMB + HML + CMA + RMW
## 
##   Res.Df    RSS Df Sum of Sq      F    Pr(>F)    
## 1    999 8.7725                                  
## 2    994 0.4631  5    8.3095 3567.2 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

annR检验

kuanjifitannR=lm(annR~MKT+SMB+HML+CMA+RMW,kuanji)

summary(kuanjifitannR)
## 
## Call:
## lm(formula = annR ~ MKT + SMB + HML + CMA + RMW, data = kuanji)
## 
## Residuals:
##        Min         1Q     Median         3Q        Max 
## -0.0256120 -0.0046797 -0.0002308  0.0046581  0.0279637 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.0252826  0.0004858   52.04  < 2e-16 ***
## MKT          1.0172863  0.0028208  360.63  < 2e-16 ***
## SMB         -0.2185653  0.0043052  -50.77  < 2e-16 ***
## HML          0.1269163  0.0087962   14.43  < 2e-16 ***
## CMA         -0.1793383  0.0098222  -18.26  < 2e-16 ***
## RMW          0.0591668  0.0090470    6.54 9.83e-11 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.00756 on 994 degrees of freedom
## Multiple R-squared:  0.9926,	Adjusted R-squared:  0.9926 
## F-statistic: 2.676e+04 on 5 and 994 DF,  p-value: < 2.2e-16

annR约束模型检验

linearHypothesis(kuanjifitannR,c("MKT=0","SMB=0","HML=0","CMA=0","RMW=0"))
## Linear hypothesis test
## 
## Hypothesis:
## MKT = 0
## SMB = 0
## HML = 0
## CMA = 0
## RMW = 0
## 
## Model 1: restricted model
## Model 2: annR ~ MKT + SMB + HML + CMA + RMW
## 
##   Res.Df    RSS Df Sum of Sq     F    Pr(>F)    
## 1    999 7.7039                                 
## 2    994 0.0568  5    7.6471 26762 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

IIR-annR检验

kuanjifitIIR_annR=lm(IRR_annR~MKT+SMB+HML+CMA+RMW,kuanji)
summary(kuanjifitIIR_annR)
## 
## Call:
## lm(formula = IRR_annR ~ MKT + SMB + HML + CMA + RMW, data = kuanji)
## 
## Residuals:
##      Min       1Q   Median       3Q      Max 
## -0.12122 -0.01140  0.00038  0.01207  0.07721 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.016202   0.001334  12.144  < 2e-16 ***
## MKT          0.043768   0.007747   5.650  2.1e-08 ***
## SMB         -0.012460   0.011823  -1.054    0.292    
## HML         -0.011300   0.024157  -0.468    0.640    
## CMA          0.018732   0.026974   0.694    0.488    
## RMW         -0.014235   0.024845  -0.573    0.567    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.02076 on 994 degrees of freedom
## Multiple R-squared:  0.03284,	Adjusted R-squared:  0.02797 
## F-statistic: 6.749 on 5 and 994 DF,  p-value: 3.378e-06

IIR-annR约束模型检验

linearHypothesis(kuanjifitIIR_annR,c("MKT=0","SMB=0","HML=0","CMA=0","RMW=0"))
## Linear hypothesis test
## 
## Hypothesis:
## MKT = 0
## SMB = 0
## HML = 0
## CMA = 0
## RMW = 0
## 
## Model 1: restricted model
## Model 2: IRR_annR ~ MKT + SMB + HML + CMA + RMW
## 
##   Res.Df     RSS Df Sum of Sq      F    Pr(>F)    
## 1    999 0.44296                                  
## 2    994 0.42842  5  0.014545 6.7493 3.378e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

主动基金

IIR检验

zhudong<-data|>
  filter(portfolio_name!="宽基")
zhudongfitIRR=lm(IRR~MKT+SMB+HML+CMA+RMW,zhudong)
summary(zhudongfitIRR)
## 
## Call:
## lm(formula = IRR ~ MKT + SMB + HML + CMA + RMW, data = zhudong)
## 
## Residuals:
##       Min        1Q    Median        3Q       Max 
## -0.155560 -0.037638 -0.001402  0.035994  0.189476 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.104876   0.003261  32.164  < 2e-16 ***
## MKT          0.920985   0.018932  48.648  < 2e-16 ***
## SMB         -0.116014   0.028893  -4.015 6.39e-05 ***
## HML         -0.699282   0.059034 -11.845  < 2e-16 ***
## CMA          0.124728   0.065920   1.892   0.0588 .  
## RMW          0.156617   0.060717   2.579   0.0100 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.05074 on 994 degrees of freedom
## Multiple R-squared:  0.7156,	Adjusted R-squared:  0.7141 
## F-statistic: 500.1 on 5 and 994 DF,  p-value: < 2.2e-16

IIR约束模型检验

linearHypothesis(zhudongfitIRR,c("MKT=0","SMB=0","HML=0","CMA=0","RMW=0"))
## Linear hypothesis test
## 
## Hypothesis:
## MKT = 0
## SMB = 0
## HML = 0
## CMA = 0
## RMW = 0
## 
## Model 1: restricted model
## Model 2: IRR ~ MKT + SMB + HML + CMA + RMW
## 
##   Res.Df    RSS Df Sum of Sq     F    Pr(>F)    
## 1    999 8.9949                                 
## 2    994 2.5586  5    6.4363 500.1 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

annR检验

zhudongfitannR=lm(annR~MKT+SMB+HML+CMA+RMW,zhudong)


summary(zhudongfitannR)
## 
## Call:
## lm(formula = annR ~ MKT + SMB + HML + CMA + RMW, data = zhudong)
## 
## Residuals:
##       Min        1Q    Median        3Q       Max 
## -0.033308 -0.006450  0.000064  0.006126  0.036040 
## 
## Coefficients:
##               Estimate Std. Error t value Pr(>|t|)    
## (Intercept)  0.1011244  0.0006257 161.611  < 2e-16 ***
## MKT          0.9367917  0.0036331 257.852  < 2e-16 ***
## SMB         -0.0933475  0.0055448 -16.835  < 2e-16 ***
## HML         -0.5396732  0.0113289 -47.637  < 2e-16 ***
## CMA          0.0510452  0.0126504   4.035 5.88e-05 ***
## RMW          0.1728307  0.0116519  14.833  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.009736 on 994 degrees of freedom
## Multiple R-squared:  0.9857,	Adjusted R-squared:  0.9857 
## F-statistic: 1.373e+04 on 5 and 994 DF,  p-value: < 2.2e-16

annR约束模型检验

linearHypothesis(zhudongfitannR,c("MKT=0","SMB=0","HML=0","CMA=0","RMW=0"))
## Linear hypothesis test
## 
## Hypothesis:
## MKT = 0
## SMB = 0
## HML = 0
## CMA = 0
## RMW = 0
## 
## Model 1: restricted model
## Model 2: annR ~ MKT + SMB + HML + CMA + RMW
## 
##   Res.Df    RSS Df Sum of Sq     F    Pr(>F)    
## 1    999 6.6024                                 
## 2    994 0.0942  5    6.5082 13731 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

IIR-annR检验

zhudongfitIIR_annR=lm(IRR_annR~MKT+SMB+HML+CMA+RMW,zhudong)
summary(zhudongfitIIR_annR)
## 
## Call:
## lm(formula = IRR_annR ~ MKT + SMB + HML + CMA + RMW, data = zhudong)
## 
## Residuals:
##       Min        1Q    Median        3Q       Max 
## -0.153221 -0.037540 -0.000542  0.032758  0.177668 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)   
## (Intercept)  0.003752   0.003208   1.170   0.2424   
## MKT         -0.015807   0.018624  -0.849   0.3962   
## SMB         -0.022667   0.028424  -0.797   0.4254   
## HML         -0.159609   0.058074  -2.748   0.0061 **
## CMA          0.073683   0.064848   1.136   0.2561   
## RMW         -0.016213   0.059730  -0.271   0.7861   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 0.04991 on 994 degrees of freedom
## Multiple R-squared:  0.01036,	Adjusted R-squared:  0.005378 
## F-statistic:  2.08 on 5 and 994 DF,  p-value: 0.06558

IIR-annR约束模型检验

linearHypothesis(zhudongfitIIR_annR,c("MKT=0","SMB=0","HML=0","CMA=0","RMW=0"))
## Linear hypothesis test
## 
## Hypothesis:
## MKT = 0
## SMB = 0
## HML = 0
## CMA = 0
## RMW = 0
## 
## Model 1: restricted model
## Model 2: IRR_annR ~ MKT + SMB + HML + CMA + RMW
## 
##   Res.Df    RSS Df Sum of Sq      F  Pr(>F)  
## 1    999 2.5020                              
## 2    994 2.4761  5   0.02591 2.0803 0.06558 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1