Repository files navigation

The package features a framework for working with high-dimensional shrinkageoptimal portfolios. It allows constructing those in two ways: 1) by applyingshrinkage directly to the portfolio weights (functionMVShrinkPortfolio) and2) by obtaining shrinkage estimates of mean returns and covariance matrices(functionMeanVar_portfolio).

The latest stable release is always on CRAN:

install.packages('HDShOP')

The latest development version can be installed in the following way:

library("remotes")u<-"Otryakhin-Dmitry/"r<-"global-minimum-variance-portfolio"re<- paste(u,r,sep="")remotes::install_github(repo=re,subdir="")

In this example, returns of assets from S&P500 are loaded and an MV portfolio iscreated, for which methodssummary andplot are called.

library(HDShOP)# loading S&P daily asset returnsdata("SP_daily_asset_returns")assets<- t(SP_daily_asset_returns[2:301,2:201])gamma<-1p<- nrow(assets)b<-exp(-0.1*(1:p))# creating an MV shrinkage portfoliosh_mv_port<- MVShrinkPortfolio(x=assets,gamma=gamma,type='shrinkage',b=b,beta=0.05)# Making a summary and plotting the portfoliosummary(sh_mv_port)plot(sh_mv_port)

Taras Bodnar, Solomiia Dmytriv, Yarema Okhrin, Dmitry Otryakhin & NestorParolya (12 May 2025): High-Dimensional portfolio selection with HDShOP package, TheEuropean Journal of Finance,DOI: 10.1080/1351847X.2025.2501637