Evaluating Investment Outcomes: Performance Measurement and Attribution

The Core Question: Skill vs. Luck

At its heart, performance evaluation is an exercise in separating a signal (manager skill, or alpha) from noise (market movements, or beta and random chance). As a statistician, you can think of this as a decomposition of variance problem. We want to attribute the observed return to its underlying drivers.

A concise guide from the Actuators Education Institute frames this perfectly.

This distinction is the foundation of performance attribution.

Performance Attribution: Deconstructing Returns

Performance attribution analysis aims to answer the "why" behind a portfolio's return relative to its benchmark. It provides a structured way to understand the impact of a manager's key decisions.

The most common framework for this is the Brinson-Fachler model, which decomposes the excess return into three primary effects.

The model breaks down the total excess return (the difference between the portfolio's return and the benchmark's return) into the following components:

1. Allocation Effect: Measures the impact of the manager's decision to overweight or underweight certain sectors or asset classes relative to the benchmark. A positive allocation effect means the manager successfully tilted the portfolio towards the better-performing parts of the market.
2. Selection (or Stock-Picking) Effect: Measures the manager's ability to pick individual securities within each sector that outperform the sector's benchmark return. A positive selection effect indicates skillful stock picking.
3. Interaction Effect: This is a cross-product term that captures the combined impact of allocation and selection decisions. For example, it measures the benefit of having good stock selection within the sectors that you chose to overweight.

To see how these are calculated and interpreted in practice, let's return to the video.

Interpreting an Attribution Report

In an exam or a real-world scenario, you won't be asked to calculate these effects from scratch. Instead, you'll be given an attribution table and asked to interpret it. The following video provides a perfect example.

From this example, we can see the practical power of attribution. The manager's total underperformance of -0.72% was primarily driven by the selection effect (-0.67%), not the allocation effect (-0.05%). This tells us the manager's big-picture bets on which countries to be in were mostly fine; their problem was picking the wrong stocks within those countries.

Risk-Adjusted Performance Measurement

Attribution tells us why we got a certain return, but it doesn't tell us if that return was worth the risk taken. A manager who generates a 20% return might seem brilliant, but if they took on a terrifying amount of risk to get it, they may not be as skilled as a manager who earned 15% with much lower volatility. This is where risk-adjusted metrics come in.

These metrics are ratios that relate the return earned to the risk taken. A higher ratio is always better, as it indicates a better return for each "unit" of risk.

Let's break down these essential ratios.

• Sharpe Ratio: This is the most common metric. It measures the excess return (above the risk-free rate) per unit of total risk (standard deviation).
  $\text{Sharpe Ratio}=\frac{R_p-R_f}{{\sigma }_p}$
  Where $R_p$ is the portfolio return, $R_f$ is the risk-free rate, and ${\sigma }_p$ is the portfolio's standard deviation.

• Treynor Ratio: This is similar to the Sharpe ratio, but it measures excess return per unit of systematic risk (beta). It is most useful for well-diversified portfolios where unsystematic (stock-specific) risk has been diversified away.
  $\text{Treynor Ratio}=\frac{R_p-R_f}{{\beta }_p}$
  Where ${\beta }_p$ is the portfolio's beta.

• Sortino Ratio: Your background in probability theory might make this ratio particularly interesting. Unlike the Sharpe ratio, which penalizes both upside and downside volatility equally, the Sortino ratio measures excess return per unit of downside risk only. It argues that volatility to the upside is a good thing and shouldn't be penalized.
  $\text{Sortino Ratio}=\frac{R_p-R_f}{{\sigma }_d}$
  Where ${\sigma }_d$ is the standard deviation of negative returns (downside deviation).

• Jensen's Alpha: This is not a ratio, but a direct measure of the return a manager generated above or below what would be expected from the Capital Asset Pricing Model (CAPM). A positive alpha indicates the manager added value.
  ${\alpha }_p=R_p-[R_f+{\beta }_p(R_m-R_f)]$
  Where $R_m$ is the market return.

The AnalystPrep video provides another excellent example of using these metrics to compare multiple managers.

The Cornerstone: A Proper Benchmark

All performance evaluation—both attribution and risk-adjusted measurement—is relative to a benchmark. If the benchmark is inappropriate, the entire analysis is meaningless. The qualities of a good benchmark are therefore a critical piece of knowledge.

A common mnemonic for these properties is SAMURAI:

• Specified in advance
• Appropriate
• Measurable
• Unambiguous
• Reflective of current investment opinions
• Accountable
• Investable

Choosing an appropriate benchmark can be especially challenging for alternative investments like hedge funds or private equity, which often lack transparency, liquidity, and investable indexes.

Conclusion

Today, we have equipped ourselves with the tools to critically evaluate an investment manager's performance. We've learned that this is not a simple case of looking at a single return figure. Instead, it's a two-part investigation.

First, we use performance attribution to decompose the excess return and understand its sources. This helps us answer why the manager achieved their results, separating the impact of their allocation decisions (market timing) from their selection decisions (stock picking).

Second, we use risk-adjusted metrics like the Sharpe, Treynor, and Sortino ratios to put the raw return into context. These measures help us answer how good the performance was, given the level of risk the manager assumed.

By combining these two approaches, an asset owner can move beyond simple outcomes and build a sophisticated, evidence-based view of a manager's true skill, helping them make better decisions about who to hire, retain, or fire.

In the next module, we will step back from the specifics of investment performance and look at the broader topic of Actuarial Modelling. The attribution and risk models we discussed today are just one application of the modelling principles that are fundamental to all areas of actuarial practice.

Key Takeaways & Cheatsheet

Here is a summary of the key performance evaluation frameworks for your CP1 exam revision.

Performance Attribution: The Brinson-Fachler Model

Risk-Adjusted Performance Metrics