by Patrick Burns.
Abstract: Realized tracking errors are examined for a series of optimized portfolios using various estimates for the variance matrix. It is clear that the benchmark should be added mathematically to the variance matrix using the constituent weights — this dramatically outperforms the case where the benchmark is a separate asset in the return matrix or where relative returns are used. The common belief that factor models are to be preferred to sample variance estimates is confirmed, but only on condition that the benchmark is added mathematically to the variance matrix.
This version: 2003 May 06 (pdf)
See also the blog post ”How to add a benchmark to a variance matrix”.
jun 25, 26
Music and snow. Poke my eyes out Perhaps your immediate response is: ”I’d rather poke my eyes out with a burning stick than do data analysis.” There’s a completely different [...]
jun 25, 26
Move your data analysis to a computing environment specifically designed for it. Why R and not spreadsheets? Here are three reasons: complexity graphics money Spreadsheets are easily overwhelmed. Very complex [...]
jun 25, 26
The steps taken to fix an R problem. Task To prepare for the Portfolio Probe blog post called ”Implied alpha and minimum variance”, I tried to update a matrix of [...]