QM013 Hands-on Lab

Benchmark-Relative Portfolio Evaluation

Objective

Compute active return, tracking error, information ratio, relative wealth, and relative drawdown from the same aligned portfolio and benchmark return series. The exercise emphasizes that these quantities answer different questions.

Run

From the QM013 folder:

python labs/python/qm013_benchmark_relative.py

What to inspect

active = portfolio - benchmark is an arithmetic period-by-period difference.
Tracking error is the volatility of that active-return series after annualization.
Information ratio combines average active return with tracking error.
Relative wealth is a ratio of compounded wealth paths, so its drawdown is not the same object as ordinary portfolio drawdown.

Experiment

Change one portfolio return while leaving the benchmark unchanged, then change both by the same amount in the same period. Trace which metrics move and which relative quantities remain similar. Also try a benchmark series with nearly zero active-return volatility and note why the information ratio becomes unstable or undefined.

Interpretation boundary

A strong benchmark-relative statistic is conditional on the benchmark choice and sample. It is not a general statement about absolute portfolio quality.