The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.
A number of utilities for dealing with the Markowitz portfolio.
– Steven E. Pav, shabbychef@gmail.com
This package may be installed from CRAN; the latest version may be found on github via devtools:
if (require(devtools)) {
# latest greatest
install_github(repo = "MarkowitzR", username = "shabbychef",
ref = "master")
}
The (negative) Markowitz portfolio appears in the inverse of the uncentered second moment matrix of the ‘augmented’ vector of returns. Via the Central Limit Theorem and the delta method the asymptotic distribution of the Markowitz portfolio can be found. From this, Wald statistics on the individual portfolio weights can be computed.
First for unconditional returns:
set.seed(1001)
<- matrix(rnorm(1000 * 3), ncol = 3)
X <- mp_vcov(X, fit.intercept = TRUE)
ism <- ism$W/sqrt(diag(ism$What))
walds print(t(walds))
## X1 X2 X3
## Intercept 0.83 -0.15 -1.8
Now for conditional expectation:
# generate data with given W, Sigma
<- function(W, Sigma, Feat) {
Xgen <- Sigma %*% W
Btrue <- Feat %*% t(Btrue)
Xmean <- chol(Sigma)
Shalf <- Xmean + matrix(rnorm(prod(dim(Xmean))), ncol = dim(Xmean)[2]) %*%
X
Shalf
}
<- 3
n.feat <- 5
n.ret <- 2000
n.obs set.seed(101)
<- matrix(rnorm(n.obs * n.feat), ncol = n.feat)
Feat <- 5 * matrix(rnorm(n.feat * n.ret), ncol = n.feat)
Wtrue <- cov(matrix(rnorm(100 * n.ret), ncol = n.ret))
Sigma <- Sigma + diag(seq(from = 1, to = 3, length.out = n.ret))
Sigma <- Xgen(Wtrue, Sigma, Feat)
X <- mp_vcov(X, feat = Feat, fit.intercept = TRUE)
ism
<- ism$W/sqrt(diag(ism$What))
walds print(t(walds))
## X1 X2 X3 X4 X5
## Intercept -0.61 0.4 0.11 0.61 0.2
## Feat1 0.30 -20.7 14.56 -10.70 -24.6
## Feat2 -14.82 -4.0 12.01 29.75 -1.8
## Feat3 16.62 -5.3 -30.41 0.83 3.5
# results are not much changed when using robust
# s.e.
library(sandwich)
<- mp_vcov(X, feat = Feat, vcov.func = sandwich::vcovHAC,
ism.rse fit.intercept = TRUE)
<- ism.rse$W/sqrt(diag(ism.rse$What))
walds.rse print(t(walds.rse))
## X1 X2 X3 X4 X5
## Intercept -0.61 0.4 0.11 0.60 0.19
## Feat1 0.29 -20.4 14.43 -10.66 -24.45
## Feat2 -14.49 -3.9 11.88 29.60 -1.80
## Feat3 16.37 -5.3 -29.85 0.83 3.48
# errors should be asymptotically normal with the
# given covariance.
<- 5
n.feat <- 15
n.ret <- 3000
n.obs set.seed(101)
<- matrix(rnorm(n.obs * n.feat), ncol = n.feat)
Feat <- 5 * matrix(rnorm(n.feat * n.ret), ncol = n.feat)
Wtrue <- cov(matrix(rnorm(100 * n.ret), ncol = n.ret))
Sigma <- Sigma + diag(seq(from = 1, to = 3, length.out = n.ret))
Sigma <- Xgen(Wtrue, Sigma, Feat)
X <- mp_vcov(X, feat = Feat, fit.intercept = TRUE)
ism
<- cbind(0, Wtrue)
Wcomp <- ism$W - Wcomp
errs dim(errs) <- c(length(errs), 1)
<- solve(t(chol(ism$What)), errs)
Zerr print(summary(Zerr))
## V1
## Min. :-2.05
## 1st Qu.:-0.55
## Median : 0.14
## Mean : 0.20
## 3rd Qu.: 1.00
## Max. : 2.49
library(ggplot2)
<- ggplot(data.frame(Ze = Zerr), aes(sample = Ze)) +
ph stat_qq() + geom_abline(slope = 1, intercept = 0,
colour = "red")
print(ph)
# qqnorm(Zerr) qqline(Zerr,col=2)
Now load the Fama French 3 factor portfolios.
if (!require(aqfb.data, quietly = TRUE) && require(devtools)) {
# get the 10 industry data
::install_github("shabbychef/aqfb_data")
devtools
}library(aqfb.data)
# fama
data(mff4)
# will not matter, but convert pcts:
<- 0.01 * mff4
ff.data
# risk free rate:
<- ff.data[, "RF"]
rfr
# subtract risk free from Mkt, HML and SMB:
<- ff.data[, c("Mkt", "HML", "SMB")] - rep(rfr,
ff.ret 2)
Now analyze the Markowitz portfolio on them.
<- mp_vcov(ff.ret, fit.intercept = TRUE)
ism <- ism$W/sqrt(diag(ism$What))
walds print(t(walds))
## Mkt HML SMB
## Intercept 4 0.32 -2
# now consider the hedging constraint: no
# covariance with the market:
<- matrix(c(1, 0, 0), nrow = 1)
Gmat <- mp_vcov(ff.ret, fit.intercept = TRUE, Gmat = Gmat)
ism <- ism$W/sqrt(diag(ism$What))
walds print(t(walds))
## Mkt HML SMB
## Intercept 1.5 0.32 -2
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.