The Information Ratio measures how much excess return you’re generating per unit of risk taken relative to a benchmark. It tells you how efficiently a strategy or manager beats their benchmark.
Beginner
What It Means
The Information Ratio shows how consistently you outperform your benchmark, adjusted for the extra risk you’re taking. High IR = consistent, efficient outperformance. Low IR = erratic or inefficient active management.
Portfolio Example
- Your portfolio returns 14% vs. S&P 500’s 10% (4% excess return)
- Your tracking error (active risk) is 5%
- Information Ratio = 4% / 5% = 0.80
This means for every 1% of additional risk taken relative to the S&P 500, you’re generating 0.80% of excess return.
Why It Matters
It separates skilled active management from luck. A high Information Ratio means consistent outperformance, not just occasional big wins. It’s the key metric for evaluating whether active management fees are worth paying.
Advanced
Mathematical Definition
Information Ratio (IR) = (Rp - Rb) / TE
Where:
- Rp = Portfolio return
- Rb = Benchmark return
- TE = Tracking Error = σ(Rp - Rb)
Tracking Error = Standard deviation of (Portfolio - Benchmark) returns
Interpretation Benchmarks
For active equity managers (annualized):
| Information Ratio | Interpretation |
|---|
| < 0 | Destroying value relative to benchmark |
| 0 to 0.3 | Weak active management |
| 0.3 to 0.5 | Moderate (median for active equity managers) |
| 0.5 to 0.75 | Good (top quartile) |
| 0.75 to 1.0 | Very good (top decile) |
| > 1.0 | Excellent (rare when sustained over 3+ years) |
Historical Context
The Information Ratio emerged from active portfolio management theory in the 1970s-80s, formalized by Richard Grinold and Ronald Kahn in “Active Portfolio Management” (1994). It became the standard for evaluating active management efficiency.
What Makes It Useful
- Active Management Focus: Specifically designed to evaluate active strategies against benchmarks
- Risk-Adjusted: Accounts for the additional risk taken to generate excess returns
- Consistency Measure: High IR requires consistent outperformance, not volatile alpha
- Portfolio Construction: Can use IR to optimally combine multiple active strategies
- Manager Evaluation: Standard metric for comparing active managers
Detailed Example
Strategy A:
- Returns 15% vs. benchmark 10% → Alpha = 5%
- Tracking Error = 10%
- Information Ratio = 5% / 10% = 0.50
Strategy B:
- Returns 13% vs. benchmark 10% → Alpha = 3%
- Tracking Error = 3%
- Information Ratio = 3% / 3% = 1.00
Strategy B is better: more efficient alpha generation despite lower absolute alpha
Data Requirements
| Requirement | Duration | Notes |
|---|
| Minimum | 36 months (3 years) | Initial IR estimate |
| Preferred | 60+ months (5 years) | Meaningful evaluation |
| Institutional standard | 3-5 years track record | Before considering IR credible |
- IR = 0.5 needs ~6 years to be statistically significant at 95% confidence
- IR = 1.0 needs ~4 years to be statistically significant
- IR = 1.5 needs ~3 years to be statistically significant
Annual IR estimates are nearly useless due to noise. IR is extremely volatile over short periods.
Relationship to Sharpe Ratio
If benchmark is the risk-free rate, Information Ratio equals Sharpe Ratio.
For most active strategies:
Simplified relationship (when active returns uncorrelated with benchmark):
SR_p² ≈ SR_b² + IR²
Important: This is primarily theoretical. In practice, long-only equity
strategies have correlated active returns, so the relationship is more complex.
The Fundamental Law of Active Management
E(IR) = IC × √BR (theoretical maximum)
Where:
- IC = Information Coefficient (manager skill, typically 0.03-0.10)
- BR = Breadth (number of independent investment decisions per year)
Critical Caveats:
- This represents theoretical maximum, not expected outcome
- Realistic IC: 0.03-0.10 (not 0.20+ often assumed)
- Effective BR is much lower than nominal positions due to correlations
- Real-world frictions reduce achievable IR to 20-50% of theoretical maximum
Limitations
- Benchmark Dependency: Only meaningful with appropriate benchmark
- Assumes Normal Distribution: Like Sharpe Ratio, assumes normally distributed returns
- Time Period Sensitive: Short periods give noisy estimates
- Doesn’t Capture Tail Risk: Can miss rare but severe underperformance
Alternatives
| Alternative | Description |
|---|
| Modified IR | Uses downside tracking error instead of total tracking error |
| Appraisal Ratio | Similar concept with different statistical properties |
| t-statistic of Alpha | Provides statistical significance of outperformance |
