[Uwgarp-commits] r38 - pkg/GARPFRM/R

[noreply at r-forge.r-project.org]

Author: tfillebeen
Date: 2014-01-03 06:20:10 +0100 (Fri, 03 Jan 2014)
New Revision: 38

Added:
   pkg/GARPFRM/R/CAPM_CRSP.R
Log:
Quick CAPM function improvement: tstat/pvalue

Added: pkg/GARPFRM/R/CAPM_CRSP.R
===================================================================
--- pkg/GARPFRM/R/CAPM_CRSP.R	                        (rev 0)
+++ pkg/GARPFRM/R/CAPM_CRSP.R	2014-01-03 05:20:10 UTC (rev 38)
@@ -0,0 +1,143 @@
+# CAPM Testing and Fitting
+
+# 'Load the GARPFRM package and the CAPM dataset.
+suppressMessages(library(GARPFRM))
+options(digits=3)
+data(crsp.short)
+data(cons)
+stock.df <-cbind(largecap.ts,cons[,"CONS"])
+colnames(stock.df)= c(colnames(largecap.ts),"CONS")
+colnames(stock.df)
+
+# Summarize Start, End, and Number of Rows
+#stock.z = returns
+start(stock.df)
+end(stock.df)
+nrow(stock.df)
+
+# Estimate excess returns: subtracting off risk-free rate
+# To strip off the dates and just return a plain vector/matrix coredata() can be used.
+# as.data.frame to check if an object is a data frame, or coerce it if possible.
+returns.mat = coredata(stock.df)
+exReturns.mat = returns.mat - returns.mat[,"t90"]
+exReturns.df = as.data.frame(exReturns.mat)
+
+# Run CAPM regression for Microsoft (MSFT) using first 5 years
+# 60 months divided by 12 months in a years = 5 years
+# capm_data use AAPL and MARKET (uppercase)
+capm.fit = lm(MSFT~market,data=exReturns.df,subset=1:60)
+summary(capm.fit)
+
+# Plot data with regression line
+plot(exReturns.df$market,exReturns.df$MSFT, main="CAPM for MSFT",
+     ylab="Excess Return: MSFT",
+     xlab="Excess Return: MARKET")
+
+# Plot CAPM regression estimate
+abline(capm.fit)    
+# Create Axis 
+abline(h=0,v=0,lty=3)
+# Placing beta & tstat values on the plot for APPL
+alpha = coef(summary(capm.fit))[1,1]
+a_tstat = coef(summary(capm.fit))[1,3]
+beta = coef(summary(capm.fit))[2,1]
+b_tstat = coef(summary(capm.fit))[2,3]
+
+legend("topleft", legend=c(paste("alpha =", round(alpha,dig=2),"(", round(a_tstat,dig=2),")"),
+                           paste("beta =", round(beta,dig=2),"(", round(b_tstat,dig=2),")")), cex=.8, bty="n")
+
+# Use a capm.tstats function:
+# Estimating CAPM with alpha=0 for asset using first 5 years of data
+capm.tstats = function(r,mkrt,type = FALSE) {
+  # Fiting CAPM and retrieve alpha specific tstats or pvalues
+  capm.fit = lm(r~mkrt)    
+  # Extract summary info
+  capm.summary = summary(capm.fit) 
+  if(is.null(type) | type=="pvalue"){
+    # Retrieve p-value if specified
+    p.value = coef(capm.summary)[1,4]  
+    p.value
+  }else{
+    # Otherwise retrieve t-stat if specified or on default
+    t.stat = coef(capm.summary)[1,3]  
+    t.stat
+  }
+}
+
+# Retrieve tstats from function for assets
+# Filter out rf and market before running
+# For capm_data use -c(1,6,7)
+colnames(exReturns.mat[,-c(21,22,23)])
+tstats = apply(exReturns.mat[1:60,-c(21,22,23)],2, capm.tstats,
+               exReturns.mat[1:60,"market"])
+tstats
+
+# Test Hypothesis for 5% CI: H0: alpha=0
+abs(tstats) > 2
+any(abs(tstats) > 2)
+
+# Plot expected return versus beta
+# Estimate expected returns over first 5 years
+mu.hat = colMeans(exReturns.mat[1:60,-c(21,22,23)])
+mu.hat
+
+# Compute beta over first 5 years
+capm.betas = function(r,market) {
+  capm.fit = lm(r~market)    			
+  # Fit capm regression
+  capm.beta = coef(capm.fit)[2]				
+  # Extract coefficients
+  capm.beta
+}
+
+betas = apply(exReturns.mat[1:60,-c(21,22,23)],2,
+              
+              FUN=capm.betas,
+              market=exReturns.mat[1:60,"market"])
+betas
+
+# Plot expected returns versus betas
+plot(betas,mu.hat,main="Expected Return vs. Beta")
+# Estimate regression of Expected Return vs. Beta
+sml.fit = lm(mu.hat~betas)
+sml.fit
+summary(sml.fit)
+# Ideally intercept is zero and equals the excess market return
+mean(exReturns.mat[1:60,"market"])
+
+# Plot Fitted SML
+plot(betas,mu.hat,main="Estimated SML")
+abline(sml.fit)
+legend("topleft",1, "Estimated SML",1)
+
+
+# The Consumption-Oriented CAPM is analogous to the simple form of the CAPM. Except that 
+# the growth rate of per capita consumption has replaced the rate of return on the market 
+# porfolio as the influence effecting returns.
+
+# Run C-CAPM regression for CONS (Consumption) using first 5 years
+# 60 months divided by 12 months in a years = 5 years
+end = nrow(stock.df)
+capm.fit = lm(CONS~market,data=exReturns.df,subset=(end-60):end)
+summary(capm.fit)
+
+# Plot data with regression line
+plot(exReturns.df$market,exReturns.df$CONS, main="CAPM for CONS",
+     ylab="Excess Return: CONS",
+     xlab="Excess Return: market")
+# Plot C-CAPM regression estimate
+abline(capm.fit)    
+# Create Axis 
+abline(h=0,v=0,lty=3)
+# Placing beta & tstat values on the plot for CONS
+beta = coef(summary(capm.fit))[2,1]
+b_stat = coef(summary(capm.fit))[2,3]
+alpha = coef(summary(capm.fit))[1,1]
+a_stat = coef(summary(capm.fit))[1,3]
+legend("topleft", legend=
+         c(paste("alpha =",round(alpha,dig=2),"(",round(a_tstat, dig=2),")"), 
+           paste("beta =",round(beta,dig=2),"(",round(b_tstat,dig=2),")")), cex=.8, bty="n")
+# NOTE: CCAPM it has two puzzles: the equity premium puzzle (EPP) and the
+# risk-free rate puzzle (RFRP). EPP implies that investors are extremely
+# risk averse to explain the existence of a market risk premium. While RFRP
+# stipulates that investors save in TBills despite the low rate of return.