TLT vs. VGLT vs. EDV: Treasury ETF Holdings and Overlap

The word “long” can refer to an index’s maturity boundary, a fund’s interest-rate sensitivity, or an investor’s holding period. Those meanings are related, but they are not interchangeable. The benchmark rules explain why these three tickers can respond differently even though all three are backed by U.S. Treasury obligations.

TLT seeks to follow the ICE U.S. Treasury 20+ Year Bond Index. ICE’s index universe is fixed-rate U.S. Treasury debt with at least 20 years remaining, weighted by market value and rebalanced monthly. That design makes TLT a focused basket of long-maturity coupon bonds. The iShares TLT profile names the benchmark; the ICE index description explains its maturity and weighting rules.

VGLT follows the Bloomberg U.S. Long Treasury Bond Index, whose eligible nominal Treasury bonds have at least 10 years remaining. Vanguard uses sampling, so the fund seeks to represent index risk and return characteristics without necessarily holding every constituent in the same weight. Its 10-year boundary admits bonds in the 10-to-20-year range that do not qualify for TLT’s index. The Vanguard VGLT profile and its summary prospectus describe the fund and its approach.

EDV takes a different route. Its benchmark is the Bloomberg U.S. Treasury STRIPS 20–30 Year Equal Par Bond Index. STRIPS separate a Treasury bond’s principal and individual interest payments into distinct zero-coupon securities. Each piece has one payment at maturity and its own CUSIP. EDV therefore holds stripped cash flows rather than a portfolio of ordinary coupon-paying bonds. See the Vanguard EDV profile, EDV summary prospectus, and TreasuryDirect’s STRIPS explanation.

The figures below use sponsor files available by September 26, 2026. TLT’s file is dated September 24; Vanguard’s latest VGLT and EDV holdings feeds are dated August 31. They are different snapshots, not a synchronized portfolio comparison.

Counts exclude cash and reserve-fund lines. They describe each file’s direct Treasury securities, not the number of issuers: all three portfolios ultimately reference obligations of the same sovereign borrower. The top-ten percentages use the ten largest reported security rows in each relevant bond or STRIPS basket. A lower top-ten figure indicates weights are spread over more lines in that snapshot; it does not remove the funds’ shared exposure to U.S. interest rates or long-maturity Treasury pricing.

The largest positions are different kinds of Treasury cash flows

The top positions put the basket distinction in concrete terms. TLT’s three largest September 24 lines were a 4.75% coupon bond due May 15, 2055 at 4.54%, a 4.625% bond due May 15, 2054 at 4.31%, and a 4.125% bond due August 15, 2053 at 4.21%. VGLT’s August 31 leaders were the 5.00% May 15, 2056 bond at 2.18%, the 4.75% August 15, 2055 bond at 2.13%, and the 4.75% May 15, 2055 bond at 2.12%.

EDV’s top three August 31 lines were not whole coupon bonds: they were principal STRIPS due November 15, 2055 at 1.90%, February 15, 2056 at 1.85%, and May 15, 2056 at 1.83%. The weights are rounded to two decimals. A principal STRIP and a coupon bond can be connected to the same original Treasury issue while remaining different securities with distinct CUSIPs and cash-flow timing.

Vanguard’s product-page tables can lag the holdings API: the public profile still showed July 31 data for several portfolio fields when the August 31 holdings feeds were available. This comparison uses the newer sponsor API files for VGLT and EDV holdings counts, positions, and overlap. The duration and SEC-yield dates are stated separately below because those metrics were not published on the same date as the August holdings snapshot.

Matching direct Treasury CUSIPs produces a specific result: all 47 TLT bond identifiers appear in the August 31 VGLT file. Those matched bonds account for 99.99% of TLT’s reported bond weight and 58.87% of VGLT’s. VGLT’s remaining bond exposure includes issues outside TLT’s 20+ year maturity rule, as well as weight differences created by their distinct index and portfolio rules.

The “sum of lower weight” is calculated by taking the smaller reported portfolio weight for each shared CUSIP and adding those minima. It is a symmetric way to express matched position weight; it is not a return correlation, an overlap statistic published by the fund companies, or a prediction. Eight of TLT’s ten largest issues also appear among VGLT’s top ten. Those eight account for about 32.54% of TLT and 16.67% of VGLT in the respective snapshots.

The table intentionally counts direct security identifiers, not parent Treasury issues. TreasuryDirect notes that stripping a bond creates separate principal and interest securities, each with a unique CUSIP. That is why EDV can show zero direct-CUSIP overlap with coupon-bond funds while still holding cash flows created from the same U.S. Treasury market. To compare EDV’s STRIPS with the original bonds, one would need to map each stripped component back to its parent issue and payment date; the simple CUSIP match does not do that.

The sponsors reported TLT effective duration of 14.88 years as of September 24. Vanguard’s VGLT and EDV product pages reported average durations of 13.4 and 23.9 years, respectively, as of July 31. These dates and provider labels differ, so the figures are a rough comparison of published interest-rate sensitivity, not a same-day ranking built on one identical methodology.

For a small, approximately parallel yield move, a duration-based first-order estimate is: percentage price change is roughly negative duration multiplied by the yield change. If yields rose by one percentage point, a mechanical calculation using those reported figures would imply about −14.9% for TLT, −13.4% for VGLT, and −23.9% for EDV before convexity and other effects. A one-point decline gives the opposite first-order signs. These are sensitivity illustrations, not forecasts or expected returns; actual price changes depend on curve shape, changing duration, convexity, income, and market pricing.

EDV’s zero-coupon pieces concentrate each security’s cash flow at maturity instead of paying coupons along the way. That produces greater price sensitivity per dollar of market value for long-dated cash flows, consistent with EDV’s longer reported duration. Duration and rate changes, convexity, and key-rate duration explain why a single duration number cannot capture every yield-curve move.

Duration is not a maturity date. TLT and VGLT are ordinary bond ETFs whose holdings change as bonds age, indexes rebalance, and funds sample their benchmarks. Buying ETF shares does not create a date on which the fund promises to return the investor’s purchase price. The guide to whether bond ETFs mature covers that distinction.

The latest displayed 30-day SEC yields were 5.41% for TLT as of September 24, 5.26% for VGLT as of September 8, and 5.56% for EDV as of September 10. These are not identical-date observations and should not be treated as a synchronized yield contest. SEC yield is a standardized annualized measure based on recent fund income and expenses; it is not a guaranteed distribution rate, total return, or forecast of what a shareholder will earn. See the guide to SEC yield versus distribution yield.

TLT and VGLT list monthly distributions; EDV lists quarterly distributions. Distribution frequency does not make EDV’s underlying STRIPS coupon-paying bonds: each STRIP itself has one payment at maturity. Fund distributions can vary and should be read alongside the fund’s current documents and tax information.

The dollar column is simple arithmetic using the stated expense ratios, not a bill charged separately to an account. Actual fund expenses accrue through net assets and a balance changes with market value. It excludes bid–ask spreads, brokerage charges, taxes, trading costs inside a portfolio, and premiums or discounts to NAV. For another $10,000, the arithmetic would be $15 for TLT, $3 for VGLT, and $5 for EDV before those other costs. Lower fees do not make two funds interchangeable if they hold different maturity ranges or cash flows.

The direct overlap result connects TLT and VGLT unusually closely at the security level: almost all of TLT’s listed bond weight appeared in VGLT’s later snapshot. That does not mean holding both doubles diversification. It can instead increase exposure to many of the same long-maturity coupon bonds, while VGLT adds some 10-to-20-year holdings and applies different weights.

EDV belongs in the same comparison because it owns U.S. Treasury-linked cash flows, but its direct securities and duration profile differ. The Treasury STRIPS guide explains the principal-and-coupon pieces; the long-term Treasury yield guide discusses why long yields can move for reasons beyond the latest policy-rate decision. For market-price deviations from portfolio value, see ETF NAV versus market price.

These relationships do not establish how closely the funds’ share prices moved over a particular period. Return correlation would require choosing a measurement window, total-return series, and frequency. Security overlap also omits duration, coupon, curve exposure, inflation sensitivity, and changing weights. Two ETFs can share no direct CUSIPs and still react to many of the same Treasury-market forces.

The portfolio distinction is clearer as a set of exposure descriptions than as a winner list. TLT concentrates on nominal Treasury coupon bonds with 20 or more years remaining. VGLT includes nominal Treasury coupon bonds from 10 years onward and overlaps heavily with TLT’s dated holdings. EDV holds 20–30 year Treasury STRIPS, placing more of each security’s cash flow at a distant maturity and producing a longer duration in the sponsor’s reported figures.

Before comparing these funds, check the latest benchmark, holdings date, duration date, SEC-yield date, distribution policy, and expense ratio. Then decide whether the intended exposure is a narrower 20+ year coupon-bond basket, a broader 10+ year coupon-bond basket, or a long-dated STRIPS basket. That is a description of fund mechanics, not an individualized recommendation. Long-maturity Treasuries still carry substantial mark-to-market, inflation, and yield-curve risk even when their issuer is the U.S. government.