Date: Sat, 5 Mar 2005 11:44:09 CST
From: "Elle" 
Newsgroups: misc.invest.financial-plan
Subject: Portfolio Allocation "Accuracy" [was Re: Expense ratios]
	iQBVAwUAQinv6/l/I4+O31e5AQEXowH/TqIrJrqlUPc4/+HRcPk1u4GV1IQ0oQrd
	jeY4wztnnhq6KnXlwkrEDrlUGECpwXQW7/R9mdsTFU7deURfyRARLQ==
	=ZVdg

Just curious: Are you aware of the claimed accuracy of the results of
algorithms that use the principles you indicate below?

I know software is available or can be easily designed that
follows the tenets of the latest Markowitz or Harvey "theorem." It should
spew out suggested portfolio allocations, factoring in the investor's years
to retirement and maybe some other particulars. The results might be
something like 40% S&P 500, 20% long-term high grade bonds, etc.

Given the historical variation in returns of each asset class, surely there
is a margin of error of some kind associated with each suggested
percentage. On about what order would this be? E.g. is the S&P 500
allocation more accurately given as 40% +/- about 10%?

One of the reasons I haven't spent time reading the minutiae of Markowitz,
Modern Portfolio Theorists, etc. is because the underlying assumptions seem
to me to be so enormous. The only practical value their work may have is in
providing very general guidelines.

 wrote
re the work of Markowitz and Modern Portfolio Theory
> A good book on his
> work is "Mean-Variance Analysis in Portfolio Choice and Capital
> Markets" (1990), by Markowitz, Todd, and Sharpe. Although the book is
> heavily mathematical, it is accompanied by VBA code implementing the
> algorithms that can be used in an Excel spreadsheet.
>
> One recent paper on portfolio optimization of assets with non-normal
> returns is
> "Portfolio Selection With Higher Moments" by Harvey, Liechty, Liechty
> and  Mueller , available from
> http://www.fuquaworld.duke.edu/www/fsc/frp2.jsp?pid=123 .