Optionally, you can decompose the Total Spread plus Residual (or only Residual) into the Brinson-Fachler equity style effects. And you can calculate multiple key rate duration effects as well as calculating the parallel shift and curve reshape effects. Other factors beyond those already described are considered by further decomposing the total spread (or residual) using equity attribution techniques. Performance attribution uses a standard Brinson-Fachler single-factor equity style attribution approach to evaluate the breakdown of the total spread into segment allocation, selection, and interaction on a benchmark relative basis. The following table displays the formulas for local equity style attribution effects.
Effect | Formula | Description |
|---|---|---|
Allocation | (Fund Segment Base Weight − Benchmark Segment Base Weight) * (Benchmark Segment Local Return − Benchmark Total Local Return) | Measures the effect of overweighting or underweighting the sectors. |
Selection | Benchmark Segment Base Weight * (Fund Segment Local Return − Benchmark Segment Local Return) | Measures the impact of choosing securities within an asset segment that provide different returns from the benchmark. It evaluates the manager's skill in choosing better performing securities than those in the benchmark. |
Interaction | (Fund Segment Base Weight − Benchmark Segment Base Weight) * (Fund Segment Local Return − Benchmark Segment Base Return) | Measures the interaction in the other effects. |
Allocation, Selection, and Interaction explain the residual return. There are options to specify which types of spread should be included in the Residual return decomposition. The default Residual return selection includes the Total Spread. There are options that exclude the Price Spread or exclude both Price Spread and Yield Spread. If trading return is mapped in the Options and Field Map field, then the trading contribution is subtracted from the residual. See the Residual options settings in Configure the Global Attribution Group for Fixed Income.
0 Comments