Separating Overweight and Underweight Effects in Brinson Attribution

Summary

This document raises a portfolio-attribution question: how to extend geometric Brinson-style analysis so that overweight and underweight positions in individual securities and groups can be examined separately. It distinguishes the allocation component, where group-level active weights are comparatively straightforward to identify, from the selection component, where the author is unsure how to isolate the contribution of security-level benchmark-relative weights.

The text provides no proposed formula, cited paper, example, or empirical evidence, so it does not establish a particular attribution method. It is useful as a statement of an analytical problem for researchers designing portfolio performance decompositions. Any implementation would need to define the benchmark and grouping scheme, specify how geometric linking is handled, and clarify how interaction effects are assigned. Those choices affect whether the resulting overweight and underweight contributions add consistently to total active return, and the document leaves them open.

Key ideas

• The question concerns separating overweight and underweight contributions in geometric Brinson attribution.
• Group-level active weights are presented as easier to handle in allocation analysis.
• The unresolved issue is isolating security-level benchmark-relative weights within selection effect.
• The document proposes no formula, references, or empirical validation.
• A usable decomposition would need explicit rules for linking and interaction effects.

Tags

# Brinson Attribution - Overweight Underweight


# Brinson Attribution - Overweight Underweight












I am looking to perform a geometric Brinson type attribution that is able to separate out the effects of underweights and overweight's of individual securities and groups individually.

From an allocation standpoint, this is relatively easy as an overweight/underweight allocations to groups can be determined relatively easily.

I am stuck on how to think about this from a selection standpoint. I would like to be able to break down selection effect from benchmark relative overweight's and underweights.

Are there any reasonable approaches or papers for this?

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.