Guide
Information ratio explained
The Sharpe ratio asks how much return you earned per unit of total volatility. Active managers face a different question: how much excess return did I deliver relative to my benchmark, and was that edge large enough to justify the tracking error I introduced? The information ratio (IR) divides average active return by the volatility of that active return series. Pension consultants, fund selectors, and quantitative allocators use it to separate skilled stock pickers from closet indexers and benchmark huggers. This guide defines active return and tracking error, walks through the IR formula and annualization, explains interpretation thresholds (including the informal 0.5 rule), connects IR to Grinold-Kahn fundamental law intuition, works a Harbor Capital active equity sleeve example against the S&P 500, compares IR to Sharpe, alpha, and Sortino, lists benchmark-gaming pitfalls, and provides an allocator checklist alongside diversification discipline.
Active return: beating (or trailing) the benchmark
An active equity fund charges higher fees because it promises to outperform an index like the S&P 500 or MSCI World. Active return (also called excess return or relative return) is the portfolio return minus the benchmark return for the same period:
active return_t = R_portfolio,t − R_benchmark,t
If the fund returned +12% and the benchmark returned +10%, active return is +2% for that year. Over many months or years you compute the average active return — the headline number marketing decks cite as “alpha,” though true Jensen's alpha adjusts for beta and is a related but distinct concept. Positive average active return is necessary but not sufficient: a manager who beats the index by 3% per year while wandering 8% away from it may be a worse bet than one who beats by 1.5% with tight tracking.
Why benchmarks must match the mandate
IR is only meaningful when the benchmark reflects what the manager is actually hired to beat. A small-cap value fund compared to the S&P 500 will show high tracking error from style exposure alone — that is factor risk, not stock-selection skill. Pair IR with factor attribution or a style-appropriate index (Russell 2000 Value, not the S&P 500) before concluding a manager adds value.
Tracking error: how far you wander from the index
Tracking error is the standard deviation of the active return series. It measures how consistently the portfolio deviates from its benchmark — not whether deviation is good or bad, only how large the swings are:
tracking error = std( active return_t )
Low tracking error (1–3% annualized) suggests a closet indexer: the fund looks active but largely mirrors the benchmark, making its fee hard to justify. High tracking error (6–10%+) means the manager takes real active bets — sector tilts, concentrated positions, cash timing. High TE with negative average active return is the worst outcome: you pay active fees for guaranteed underperformance with extra volatility.
Annualization
From monthly active returns, annualize average active return by multiplying by 12 and tracking error by √12. Daily data uses √252. As with Sharpe and Sortino, inconsistent annualization across databases is a common source of conflicting IR figures on the same fund. Always confirm frequency and whether returns are gross or net of fees.
Tracking error vs beta
Beta measures sensitivity to market moves; tracking error measures deviation from a specific index path. A fund can have beta ≈ 1.0 and high tracking error if it holds different names with similar sector weights. Conversely, low tracking error with beta far from 1.0 is rare but possible with offsetting factor bets.
The information ratio formula
With average active return AR and tracking error TE:
Information Ratio = AR / TE
Interpretation: IR tells you how many units of benchmark-relative return you earned per unit of active risk. An IR of 0.50 means the manager delivered 0.5% of annualized active return for every 1% of tracking error — or equivalently, 2% active return at 4% tracking error. Higher is better. Negative IR means consistent underperformance relative to the benchmark on a risk-adjusted basis.
The informal 0.5 threshold
Practitioners often cite IR ≈ 0.5 as a rough dividing line between mediocre and genuinely skilled active management over long horizons. This is not a physical law — it is a heuristic from decades of pension consultant experience. Top quartile equity managers over 10+ years sometimes achieve IR of 0.6–1.0; most active funds cluster below 0.3 after fees. Treat 0.5 as a screening bar, not a guarantee of future results.
Grinold-Kahn fundamental law (intuition)
Richard Grinold and Ronald Kahn framed active management as:
IR ≈ IC × √breadth
where IC (information coefficient) is the correlation between forecasted and realized returns on individual bets, and breadth is the number of independent bets per year. Skill (IC) matters, but so does opportunity set: a manager who makes many small, weakly correlated bets can achieve high IR with modest per-stock skill. A concentrated manager needs very high IC to match that IR. This explains why quant diversified portfolios and fundamental concentrated portfolios can both show strong IR through different paths.
Worked example: Harbor Capital active equity sleeve
Harbor Capital runs a $120M institutional portfolio with a mandate to beat the S&P 500 by 1.5% per year net of fees, with tracking error capped at 5%. Over the trailing 36 months (monthly data, net returns):
- Average monthly portfolio return: 0.92%
- Average monthly S&P 500 return: 0.78%
- Average monthly active return: 0.14%
- Standard deviation of monthly active return: 1.05%
Annualized active return: 0.14% × 12 ≈ 1.68%. Annualized tracking error: 1.05% × √12 ≈ 3.64%. Information ratio: 1.68 / 3.64 ≈ 0.46.
The sleeve cleared its 1.5% active return target slightly and stayed inside the 5% TE budget — but IR just under 0.5 flags “good, not exceptional.” Attribution showed sector overweight in technology (+80 bps) offset by stock selection in healthcare (−20 bps). The investment committee kept the mandate but tightened position limits on single-name bets above 4% weight to reduce idiosyncratic TE spikes. They also compared IR on a rolling 12-month basis: one strong quarter had pushed the 36-month IR up; the rolling window showed IR falling to 0.38 — a reminder that window selection changes the story.
IR vs Sharpe vs alpha vs Sortino
| Metric | Numerator | Denominator | Best when |
|---|---|---|---|
| Information ratio | Average active return vs benchmark | Tracking error of active return | Evaluating active managers against a stated benchmark |
| Sharpe ratio | Excess return vs risk-free rate | Total return volatility | Comparing absolute risk-adjusted performance across strategies |
| Jensen's alpha | Return above CAPM expected return | Implicit (single regression coefficient) | Isolating skill after market beta adjustment |
| Sortino | Return above MAR hurdle | Downside deviation | Retiree spending hurdles, asymmetric return profiles |
| Calmar | Annualized return | Maximum drawdown | Crisis-sensitive allocators, trend followers |
A manager can show high Sharpe (smooth absolute returns) but low IR if they hug the benchmark and barely beat it. Conversely, high IR with mediocre Sharpe is common for concentrated active funds that take benchmark-relative risk but not extreme absolute volatility. Report both when deciding whether to pay active fees or switch to index funds.
Decision table: when IR is the right metric
| Your question | Start here | Also check |
|---|---|---|
| Should we renew this active equity mandate? | 3-year IR net of fees vs 0.5 hurdle | Rolling IR, max drawdown, fee drag |
| Is the fund a closet indexer? | Tracking error below 2% with active fee | Active share, holdings overlap with index |
| Quant strategy vs long-only manager | IR on same benchmark and window | Capacity, turnover, transaction costs |
| Multi-manager lineup overlap | Pairwise correlation of active returns | Combined portfolio IR vs sum of parts |
| Passive vs active allocation decision | Expected IR × active weight vs fee savings | Tax, factor exposure, governance |
Who uses the information ratio
Pension and endowment consultants
Institutional due diligence often ranks manager candidates by IR over 5- and 10-year windows, with minimum track-record length requirements. Consultants also stress-test IR after removing the best single year to detect luck-driven scores.
Fund-of-funds and multi-manager platforms
Combining managers requires knowing whether their active returns are correlated. Two managers each with IR 0.4 can produce a combined sleeve with IR above 0.4 if their active bets diversify — or below 0.4 if they crowd the same sectors.
Smart beta and factor overlays
Factor tilts (value, momentum, quality) produce active return relative to cap-weighted benchmarks. IR helps separate factor premia harvesting from genuine security-selection skill — though factor exposures should be regressed out before crediting alpha.
Common pitfalls
- Wrong benchmark — comparing a global fund to a domestic index inflates TE and distorts IR.
- Gross vs net returns — IR on gross returns overstates skill; allocators eat fees.
- Short windows — three years is a minimum; one great year can manufacture IR above 1.0 that mean-reverts.
- Survivorship bias — databases drop dead funds; published IR percentiles look better than live experience.
- Benchmark gaming — managers choose easy benchmarks or change them after underperformance.
- Ignoring active share — low active share with low TE and low IR is an expensive index clone.
- Style drift — a manager who drifts from value to growth may show temporary IR from factor luck.
- Capacity constraints — high historical IR on $200M AUM may collapse at $5B as edge dilutes.
Allocator checklist
- Define the benchmark in the investment policy statement before hiring.
- Compute IR on net returns over at least one full market cycle when possible.
- Report rolling 12- and 36-month IR alongside point-in-time figures.
- Compare IR to an explicit hurdle (e.g. 0.5) and to peer universe median.
- Run factor attribution; separate style exposure from stock-selection IR.
- Check active share and holdings overlap to detect closet indexing.
- Stress-test by removing best and worst years from the active return series.
- Pair IR with max drawdown, Sharpe, and fee drag for a complete picture.
- Document data frequency, annualization method, and benchmark version for audits.
Key takeaways
- The information ratio measures active return per unit of tracking error — the core metric for benchmark-relative skill.
- High average active return with high tracking error can score lower IR than modest outperformance with tight tracking.
- IR ≈ 0.5 over long horizons is a common heuristic for genuine skill; most active funds fall short after fees.
- Grinold-Kahn links IR to per-bet skill (IC) and the number of independent bets (breadth).
- Always use the right benchmark, net returns, and long windows — and pair IR with factor attribution and drawdown analysis.
Related reading
- Sharpe ratio explained — absolute risk-adjusted return vs risk-free rate
- Index funds explained — when passive exposure beats active IR after fees
- Factor investing explained — separating style premia from manager skill
- Portfolio rebalancing explained — keeping active sleeves inside risk budgets