An abnormal return is the difference between the return a security or portfolio actually earns and the return expected under a selected benchmark or model. It measures performance relative to that reference, rather than judging the realized return alone. The result can be positive, negative, or zero.
How It Is Calculated
The formula is AR = realized return − expected return. Both returns must cover the same period and use a consistent basis, including the treatment of dividends, fees, and currency. For example, comparing a monthly realized return with an annual expected return would not give a meaningful result. When percentage returns are subtracted, express the difference in percentage points, not as a percentage change.
Expected return is an estimate from the selected benchmark or model, not a guaranteed outcome or necessarily a consensus forecast. For example, the capital asset pricing model (CAPM) estimates it as the risk-free rate plus beta multiplied by the expected market risk premium. That premium is the expected market return minus the risk-free rate. Beta measures sensitivity to market returns, not total volatility.
Interpreting the Result
A positive abnormal return indicates outperformance relative to the benchmark; a negative value indicates underperformance. Zero means the realized and expected returns match. These labels describe relative performance, not whether the investment gained or lost value.
In a hypothetical period, a security returns −3% against an expected −5%. Its AR is −3% − (−5%) = +2 percentage points, so it outperforms while still losing money. Conversely, a realized gain of 5% against an expected 8% produces −3 percentage points: an absolute gain but relative underperformance.
Looking for liquidity, exploring on-ramp/off-ramp services, or seeking expert guidance?
Get startedPerformance Review and Event Analysis
Investment-performance reviews use abnormal return to assess whether a security or portfolio performed above or below its chosen reference. A positive or negative result can prompt a review of holdings, risk exposures, or developments during the period. The figure is most useful alongside the realized return, since relative success does not show whether the investment made money.
Cumulative abnormal return (CAR) is the sum of period-by-period abnormal returns across a specified window. Event studies use CAR to examine deviations around announcements such as earnings releases or mergers. The window may include periods before and after the announcement to examine the timing of the reaction. CAR is an arithmetic sum, not a compounded investment return, and it is not inherently a long-term measure.
Benchmark Choice and Limitations
Interpretation depends on the benchmark and its risk assumptions. The same realized return can produce different results under different expected-return estimates, so risk adjustment does not automatically make comparisons fair. An abnormal return can reflect chance, model error, or omitted risks as well as unusual performance. One observation does not establish investment skill or company problems. Even persistent outperformance is not sufficient evidence of mispricing or riskless arbitrage.
In event analysis, overlapping news and the choice of event window can affect the measured result. A deviation around an announcement therefore does not, by itself, establish that the announcement caused it. Statistical testing can help assess whether a result is distinguishable from ordinary variation, but it cannot eliminate a flawed benchmark or isolate overlapping events automatically. Market efficiency does not require every realized return to match an estimate; the issue is whether available information permits predictable, consistently exploitable outperformance after accounting for risk and costs.