Performance measurement without attribution is meaningless. Did you outperform the benchmark? By how much? Why? Understanding the sources of return distinguishes skill from luck and reveals where improvements can be made.
Benchmark Selection
Benchmarks should represent the opportunity set available to investors. An actively managed portfolio should be compared to an appropriate active benchmark, not a passive index. A global portfolio should compare to global benchmarks. Mismatched benchmarks create false conclusions.
Benchmark construction matters. Market-cap weighting weights index toward the largest companies. Equal-weighting treats all companies equally. Different weighting methodologies produce different benchmarks. Selecting appropriate benchmarks requires domain knowledge.
Return Attribution
Attribution analysis decomposes returns into: benchmark return + active return. Active return = specific bets + timing decisions. Further decomposition identifies which decisions created excess return and which destroyed value.
Allocation effect = (portfolio weight - benchmark weight) × (benchmark return - total benchmark return). Selection effect = benchmark weight × (security return - benchmark return). These effects separate asset class decisions from security selection decisions.
Statistical Significance
Outperformance by 0.1% annually might be luck. Outperformance by 3% annually over 10 years shows skill. Statistical significance testing determines if results are unusual enough to assume skill rather than luck. Most active managers show results consistent with luck alone.
Tracking error quantifies deviations from benchmark. A low tracking error strategy tightly follows the benchmark. A high tracking error strategy makes large active bets. Information ratio (excess return / tracking error) measures active management quality.
Performance Persistence
Does past outperformance predict future outperformance? In most cases, no. Outperformance tends to be mean-reverting. Lucky managers don't consistently beat their benchmarks. Identifying persistent alpha (not luck) is the challenge of performance analysis.
Fees and Net Returns
Gross returns mean nothing; net returns (after fees) are what investors earn. A strategy with 1% annual fees must generate 1% alpha just to break even. Most active managers underperform their benchmarks on a net basis after fees.
Educational content only. Not investment advice.