[R-SIG-Finance] portfolioFrontier nonsense

[Diethelm Wuertz]

ssmith88 at umd.edu wrote:
> I'm attempting a very basic optimization with fPortfolio using the default constraints.  Yet, the portfolio frontier that is generated is nonsense.  As the return increases, the risk decreases!  

Let me start with a comparison: When you fill up your Ferrari with water 
instead
of gas, then the "return in speed decreases" and the "risk of damage 
increases".
This is a problem arising from the kind of fuel you use and not of the 
Ferrari.

I think it is a data problem - when you send me your data, I will try to 
fnd out
what was possibly going wrong.

Kind regards
Diethelm



> Any help would be greatly appreciated.  Below is my code:
>
>
>
> library(fPortfolio)
> #Gets date,price, and return data from excel spreadsheet
> #------------------------------------------------------------
> sp500Reuters<-read.csv("C:\\data\\sp500Reuters.csv")
> spgsciReuters<-read.csv("C:\\data\\spgsciReuters.csv")
> msciWorldReuters<-read.csv("C:\\data\\msciWorldReuters.csv")
> cshfReuters<-read.csv("C:\\data\\cshfReuters.csv")
>
> # Creates return Matrix, timeDate object, and timeSeries object
> #------------------------------------------------------------
> returnMatrix<-cbind(sp500Reuters[(2:185),3],msciWorldReuters[(2:185),3],spgsciReuters[(2:185),3],cshfReuters[(2:185),3])
> colNames<-c("US Equity","World Equity","Commodities","Hedge Fund")
> colnames(returnMatrix)<-colNames
> returnSeries<-as.timeSeries(returnMatrix)
> timeObject<-timeSequence(from="1994-1-31",to="2009-4-30",by="month",format="%Y-%m-%d",FinCenter="GMT")
> dataSet<-timeSeries(returnSeries,timeObject,units=NULL,format=NULL,zone="GMT",FinCenter="GMT")
> ------------------------------------------------------------
> # Creates portfolioData object and generates frontier
> portfolioData<-portfolioData(dataSet,spec=portfolioSpec())
> newFrontier<-portfolioFrontier(portfolioData)
>
> #Note that I've also tried 
> newFrontier<-portfolioFrontier(dataSet)
>
> In both cases I get an output such as the following:
>
> Target Return and Risks:
>      mean     mu    Cov  Sigma   CVaR    VaR
> 1  0.0380 0.0380 0.5285 0.5285 1.3375 0.9721
> 13 0.0500 0.0500 0.4391 0.4391 1.0920 0.7691
> 25 0.0620 0.0620 0.3609 0.3609 0.8590 0.5488
> 37 0.0740 0.0740 0.3014 0.3014 0.6797 0.3874
> 50 0.0869 0.0869 0.2732 0.2732 0.6036 0.4280
>
>
> Thanks for any help in advance.
>
> _______________________________________________
> R-SIG-Finance at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-finance
> -- Subscriber-posting only.
> -- If you want to post, subscribe first.
>
>