On This Page
Key Takeaways
- Tracking error is the standard deviation of active return, the period-by-period difference between a portfolio return and its benchmark return, so it measures the spread of the gap rather than its size.
- Keep three numbers apart: the average monthly gap (−2 bp in the example), the compounded return difference over the window, and tracking error, the spread of the gap (3.98 bp).
- Ex-post tracking error uses a completed return series; ex-ante tracking error uses forecast active weights and a covariance model.
- A fee charged evenly every month moves the average gap and leaves tracking error untouched; costs and income that vary by month move it.
- Judge tracking error against the fund’s benchmark, its mandate and the observation frequency.
- Tracking error measures swings around the average gap, so a steady lead over the benchmark adds nothing to it, and a month above the average counts the same as a month below.
What is tracking error?
Tracking error is the standard deviation of active return, the period-by-period difference between a portfolio’s return and its benchmark’s return. It measures how much that gap varies around its own average, not how large the gap is. Index funds aim to keep it low, while active managers treat it as a budget for active risk.
A constant monthly lead or lag has zero tracking error. Review the average gap separately from its variation.
Formula
Active return at t = portfolio return at t − benchmark return at t; sample tracking error = √[Σ(active return − mean active return)² ÷ (n − 1)]- Use matching dates, currencies and return definitions. For investment-performance comparisons, align distributions, fees and the gross or net benchmark basis.
- n is the number of periods in the sample. The n minus 1 divisor is the sample estimator, matching the convention used for return standard deviation elsewhere in this glossary.
- Mean active return is the arithmetic average of periodic differences. A window’s tracking difference is portfolio compounded return minus benchmark compounded return; calculate it from the two full series.
- Root mean square active return measures the gap around zero instead, so it counts a steady lag: with an n divisor, RMS² = ((n − 1) ÷ n) × sample tracking error² + mean active return².
- Ex-ante tracking error forecasts the spread: the square root of the active weight vector multiplied through a forecast covariance matrix.
How to calculate tracking error: a 12-month example
An index fund charging a 0.20% expense ratio posts twelve monthly active returns against its benchmark, quoted in basis points: −2, −5, +3, −8, −1, +4, −6, −3, +2, −7, +1, −2. Both series are total returns over identical months.
- Mean monthly active return: −24 bp ÷ 12 = −2 bp. Multiplying that arithmetic mean by 12 gives −24 bp, or −0.24%, annualized.
- Sum of squared deviations from that mean: 0 + 9 + 25 + 36 + 1 + 36 + 16 + 1 + 16 + 25 + 9 + 0 = 174 bp squared.
- Sample variance: 174 / (12 − 1) = 15.82 bp squared, so monthly tracking error = square root of 15.82 = 3.98 bp.
- Annualized tracking error: 3.98 bp × square root of 12 = 13.8 bp, or 0.14%.
The fund lags its benchmark by 2 bp a month on average, and that gap swings by 3.98 bp a month, or 13.8 bp a year under √12 scaling. To explain the lag, compound the full fund and benchmark series for the annual difference, then attribute it, starting with the 0.20% expense ratio.
For each period, subtract benchmark return from portfolio return. Tracking error is the standard deviation of that active-return series around its own mean.
The example’s active returns sum to −24 basis points, giving a −2 bp monthly mean. Squared deviations from that mean sum to 174 bp².
Divide by 11 for the sample variance and take the square root: monthly tracking error is 3.98 bp. Multiplying by √12 annualizes it to 13.8 bp, or 0.14%.
Tracking error vs tracking difference
Multiplying the monthly mean by 12 gives −24 bp of annualized arithmetic active return. To calculate the annual total-return difference, compound the portfolio and benchmark series separately from their own return records, then subtract.
How fees, income and ETF pricing affect tracking error
A constant deduction from every period shifts the mean and leaves standard deviation unchanged. Real fee, tax, financing and income contributions can vary by period, so attribute them from the account records.
For ETFs, choose NAV or market-price returns consistently. Premium and discount changes can alter market-price tracking error in either direction relative to NAV tracking error.
Ex-ante tracking error and active positions
A forecast model calculates active risk from the active-weight vector and covariance matrix. A new position can increase or reduce the combined tracking error depending on its offsets.
Market-beta differences also contribute. In a one-market regression, active return contains the portfolio’s beta difference from one as well as its residual return. So tracking error equals residual volatility only when the portfolio’s beta is one.
Tracking error and the information ratio
Use the Information Ratio to combine mean active return with tracking error. Compare total return, drawdown and exposures alongside that benchmark-relative measure.
Limitations of tracking error
- Read the average gap beside it. A fund that lags its benchmark by the same amount every period shows zero tracking error, so check mean active return (−2 bp a month in the example) before you call a fund a good tracker.
- Look at the misses separately. Tracking error counts a month above the average gap the same as an equal month below it.
- Match benchmark, observation frequency, window and return basis before comparing two funds, including NAV or market price for an ETF.
- Check for a persistent lag before you annualize. √12 scaling holds when active returns are uncorrelated month to month; when the gap repeats, measure on annual returns.
- Treat ex-ante tracking error as the risk budget and realized tracking error as the result. When correlations or exposures shift, the two part ways.
Bottom line
Tracking error measures how steadily a portfolio follows its benchmark, separate from whether it beats or lags it. Read it with mean active return and compounded tracking difference on matched dates and return definitions, and use the information ratio to judge whether the active risk was rewarded.
Clear answers
Frequently Asked Questions
Sources and Further Reading
- SEC Office of Investor Education and Advocacy: Investor Bulletin, Index Funds
- FINRA: Exchange-Traded Funds and Products
- Campbell R. Harvey, Duke University: Tracking Error
How entries are sourced, calculated and updated: editorial methodology.