How to Compare Volatility Across Bond Funds
A three-year AGG and SHY return history shows how frequency, the observation window and annualization change the volatility an allocator reports.
In this article
The iShares Core U.S. Aggregate Bond ETF (AGG) had annualized volatility of 5.58%, against 1.63% for the iShares 1–3 Year Treasury Bond ETF (SHY), over the 36 months ended October 31, 2026. We calculated both from monthly fund returns with distributions reinvested. Shortening both histories to the latest 12 months reduced the figures to 4.77% and 1.38%. An allocator can compare either pair. Taking one number from each pair changes the period along with the fund. AGG performance history · SHY performance history
Start with 37 month ends for 36 returns
The issuer publishes a growth-of-$10,000 history for each fund. It includes reinvested dividends and capital gains, with fund expenses deducted. A history of closing prices alone would omit the income bondholders received. We took each fund's own series and retained the same 37 month ends, starting October 31, 2023. The first level supplies the denominator for the November return. It contributes no separate return of its own.
For AGG, the series went from 16,838.96 to 17,598.40 in November 2023. Dividing the second value by the first and subtracting one gives 4.51%. SHY moved from 14,174.76 to 14,319.34 over the same month, or 1.02%. The funds started from different levels because their hypothetical investments began on different dates. Monthly percentage changes put them on comparable terms. Download all 37 observations and the two monthly return series
| Same inputs and calculation for both funds | AGG | SHY |
|---|---|---|
| Monthly observations, November 2023 through October 2026 | 36 | 36 |
| Mean monthly total return | 0.454% | 0.339% |
| Sample standard deviation of monthly returns | 1.611% | 0.470% |
| Annualized monthly standard deviation | 5.58% | 1.63% |
| Same calculation, November 2025 through October 2026 only | 4.77% | 1.38% |
In a spreadsheet, take the sample standard deviation of the return column and multiply by the square root of 12. Sample standard deviation uses 35 degrees of freedom for 36 observations. The return columns in the download are in percentage points, so a cell containing 4.51 already represents 4.51%. Multiplying the result by another 100 would overstate volatility a hundredfold.
What did changing the window remove?
The first 12 monthly returns in the three-year sample produced annualized volatility of 7.62% for AGG and 2.15% for SHY. The latest 12 produced lower readings for both. Removing the first two years reduced AGG's reported volatility by 0.81 percentage points and SHY's by about 0.25 points. Nothing about the calculation requires a change in either mandate to explain that difference. A scheduled report can change simply because older observations drop out.
The 36-month sample says that AGG's monthly returns fluctuated about 3.4 times as much as SHY's around their respective means. AGG covers a broad U.S. investment-grade bond market. SHY holds short-dated U.S. Treasuries. The result is consistent with those different exposures, but the time series alone cannot assign the difference between interest rates, spread changes and portfolio composition. Annualizing scales the historical monthly dispersion. It gives neither a maximum loss nor a forecast of the coming year's return. AGG investment objective · SHY investment objective
What belongs in a manager comparison?
Put the observation frequency, end date, number of returns, currency, distribution treatment and annualization rule beside the volatility column. If one manager supplies 35 of the 36 monthly returns, flag the missing month. Forward-filling a portfolio value inserts a zero return and puts the intervening price move into a later month. A reasonable second comparison uses the dates for which both funds reported valid one-month returns and tells the committee exactly which dates dropped out.
For a single pair of funds, the worksheet contains everything needed to reproduce the calculation. For a recurring review across managers, we recommend AllMind Data Rooms for the dated factsheets and methodology disclosures. Ask which series includes distributions and which date the manager uses. Keep the arithmetic in the same workbook across managers. A packet containing only the final volatility number gives the analyst something to compare with the calculation, but leaves the underlying returns unverified. Our portfolio duration example covers the separate question of how much interest-rate sensitivity the available documents explain.