Appraisal Ratio: Alpha per Unit of Residual Risk

The appraisal ratio measures estimated alpha relative to the residual risk taken to generate it.

A common form is:

Appraisal Ratio = Alpha / Residual Risk

Here, alpha is usually the intercept from a specified asset-pricing or factor model, while residual risk is the standard deviation of the return component left unexplained by that model.

A simple interpretation

Suppose a manager has an estimated annual alpha of 2% and annualized residual risk of 4%. The appraisal ratio is 0.50.

A second manager with the same 2% alpha but 8% residual risk has an appraisal ratio of 0.25. Under the same model and estimation conventions, the first manager generated more estimated alpha per unit of residual risk.

That does not automatically mean the first manager is superior. The ratio inherits the assumptions and estimation error of the underlying model.

Appraisal ratio versus information ratio

The appraisal ratio is related to, but not interchangeable with, the Information Ratio.

The information ratio generally compares benchmark-relative return with Tracking Error. The appraisal ratio instead compares model-based Alpha with residual risk after accounting for the model's systematic risk factors.

If the benchmark itself captures the relevant risk exposures perfectly, the concepts can move closer together. In real analysis, benchmark choice and factor-model choice can produce materially different conclusions.

Why the risk model matters

A manager can appear to have positive alpha under a simple model and little or no alpha under a richer model that recognizes additional compensated exposures.

For example, returns that look like stock-selection skill under a market-only model may partly reflect persistent value, size, momentum, quality, duration, credit, or other factor tilts when a broader model is used.

The denominator is also model-dependent because residual risk is what remains after the selected factors explain return variation.

Gross versus net performance

Performance appraisal can be performed using gross returns or returns net of fees and expenses. Those answer different questions.

Gross performance is more useful for isolating the investment process before fees. Net performance is closer to the investor's realized economic experience. An appraisal ratio should therefore state which return convention is being used.

Statistical uncertainty

An estimated appraisal ratio is not a direct probability that a manager is skilled. Alpha estimates can be noisy, samples can be short, residuals can be non-normal, and manager processes can change over time.

Backfilled data, survivorship bias, multiple testing, benchmark changes, and regime dependence can further distort apparent performance.

What the appraisal ratio cannot establish

A high historical appraisal ratio does not prove persistent manager skill, and a low ratio does not prove that future value added is impossible.

It is one performance-appraisal lens among several. It should be interpreted alongside the manager's process, benchmark, exposures, costs, drawdowns, sample length, and evidence that the underlying source of returns is economically plausible and repeatable.

Sources

• CFA Institute, Portfolio Performance Evaluation, 2026
• CFA Institute, Analysis of Active Portfolio Management, 2026

Explore more topics in the Financial Research Encyclopedia.