What a great question!

In today's webinar on GIPS 2010 one of the attendees (participants?) asked if for the new GIPS(R) 2010 (Global Investment Performance Standards) requirement to report 3-year annualized standard deviation, if an asset weighted version of standard deviation would work. Interesting, yes?

This is yet another example of the confusion which exists because standard deviation has been used for years to report the required measure of dispersion and an asset-weighted version can be used (although it isn't even shown in GIPS but was, at one time, recommended for the AIMR-PPS(R)).

One must separate the two ideas: risk vs. dispersion. Risk is across multiple time periods (e.g., the past 36 months) while dispersion is within a time period (e.g., for the year 2009). For risk there is no such thing as an asset-weighted version of standard deviation.