Alpha (Jensen's alpha) in mutual funds

Jensen’s alpha (commonly called simply “alpha” in mutual fund analysis) is a risk-adjusted performance measure that quantifies how much excess return a fund manager has delivered above the return that would be expected given the fund’s systematic risk exposure, as predicted by the Capital Asset Pricing Model (CAPM). A positive alpha indicates outperformance attributable to the manager; a negative alpha indicates underperformance after accounting for market risk.

The concept was introduced by Michael C. Jensen in his 1968 paper “The Performance of Mutual Funds in the Period 1945–1964” published in the Journal of Finance, making it one of the oldest formal measures of active management skill.

Formula

\[ \alpha = R_p - \left[ R_f + \beta_p \times (R_m - R_f) \right] \]

Where:

Alternatively, alpha is the intercept term in the CAPM regression:

\[ R_p - R_f = \alpha + \beta_p \times (R_m - R_f) + \epsilon \]

Where \(\epsilon\) is the error term. The regression is typically run on monthly excess returns over a rolling 36-month window.

Worked example

Consider a mid-cap equity fund with the following 3-year data:

• Fund return (annualised): 18.5 per cent
• Benchmark (Nifty Midcap 150 TRI) return: 16.0 per cent
• Risk-free rate (91-day T-bill): 6.5 per cent
• Fund beta: 0.92

\[ \text{Expected return} = 6.5 + 0.92 \times (16.0 - 6.5) = 6.5 + 8.74 = 15.24% \]

\[ \alpha = 18.5 - 15.24 = +3.26% \]

The fund generated 3.26 percentage points of alpha, excess return attributable to manager skill rather than market exposure.

Interpretation

Alpha vs simple excess return: Many publications report “excess return” as \(R_p - R_m\), which is not the same as Jensen’s alpha. Simple excess return does not adjust for the level of systematic risk taken. A fund that beats the index by taking on 1.4x market risk has not demonstrated skill, it has demonstrated leverage. Alpha corrects for this by using beta.

Statistical significance

Alpha should be evaluated alongside its t-statistic. An alpha of +2 per cent with a t-statistic below 1.5 is not statistically distinguishable from zero at conventional confidence levels. Over a 3-year monthly return series (36 observations), the standard error of alpha is typically 1.5–3 percentage points, meaning only alphas above 3–4 per cent are clearly significant.

This is why most academic studies (including Jensen’s original 1968 paper) find that the average actively managed fund produces zero or negative alpha net of costs, positive alpha is rare, inconsistent over time, and difficult to distinguish from luck.

Typical alpha ranges in Indian mutual funds

Based on rolling 3-year data for actively managed equity funds in India (2015–2024):

Large-cap funds struggle to generate positive alpha because Indian large-cap indices are relatively efficient and SEBI mandates that at least 80 per cent of assets be in the top 100 stocks by market cap, constraining the opportunity set.

Mid and small-cap categories show higher frequency of positive alpha, consistent with less efficient price discovery in those segments. However, survivorship bias inflates these figures, funds that consistently underperform often close or merge.

Alpha and the total expense ratio

Every percentage point of TER directly reduces net alpha by the same amount. A fund generating 2 per cent gross alpha but charging 1.50 per cent TER delivers only 0.50 per cent net alpha to the investor. This is why the regular plan version of an actively managed fund (with 0.80–1.10 per cent higher TER than the direct plan) must generate that much additional gross alpha just to break even with the direct plan, a bar few regular plan investors examine.

Alpha vs Sharpe ratio vs Treynor ratio

Alpha and the Treynor ratio are both beta-based, but they answer different questions: alpha is an absolute excess return; Treynor is a ratio of excess return to beta.

Limitations

• Beta instability: Beta is estimated from historical data and can change, especially in market stress periods. A fund with β = 0.85 in a bull market may behave like β = 1.10 in a sharp drawdown.
• Benchmark sensitivity: Changing the benchmark changes the beta estimate and therefore the alpha. Choosing a too-easy benchmark (e.g., a mid-cap fund benchmarked to Nifty 50 instead of Nifty Midcap 150) artificially inflates alpha.
• Single-factor model limitation: CAPM explains only market risk. Multi-factor models (Fama-French 3-factor and 5-factor models) account for size, value, profitability, and investment style; what appears as alpha in CAPM may be factor exposure in a multi-factor model.
• SEBI’s total return index requirement: Since February 2018, SEBI mandates that fund performance be reported against the total return index (TRI) of the benchmark, which includes dividends reinvested. Alpha calculated against price-return indices (which exclude dividends) is overstated because dividends represent a real return delivered by the index but excluded from the price index.

Alpha in AMFI factsheets

AMFI does not publish alpha centrally. AMCs may report it on their factsheets, and third-party portals (Value Research, Morningstar India, MF Central) compute and display it. Investors should check whether the reported alpha uses TRI or price-return benchmark before drawing conclusions.

See also

• Sharpe ratio in mutual funds
• Sortino ratio in mutual funds
• Beta in mutual funds
• Treynor ratio
• Information ratio
• Total expense ratio
• Total return index benchmarking
• Mutual fund

References

1. Jensen, M. C. (1968). “The Performance of Mutual Funds in the Period 1945–1964.” Journal of Finance, 23(2), 389–416.
2. SEBI circular SEBI/HO/IMD/DF2/CIR/P/2018/007 dated 4 January 2018, TRI benchmarking mandate.
3. Fama, E. F. and French, K. R. (1993). “Common risk factors in the returns on stocks and bonds.” Journal of Financial Economics, 33(1), 3–56.
4. AMFI, Mutual fund industry data and factsheets at amfiindia.com.
5. Value Research, Fund alpha database at valueresearchonline.com.