How to Choose Between Two Similar ETFs

Everything downstream depends on this. Compare the full holdings lists and weights, not just the top ten, and read the index each fund tracks. Two funds can share nine of their ten largest positions and then diverge across the remaining hundreds, because one index caps individual weights and the other does not, or one includes mid-caps and the other stops at large. If the holdings genuinely match, you have a cost-and-execution decision. If they do not, you have an exposure decision, and cost is a secondary consideration.

The practical shortcut is overlap: the share of each fund's assets sitting in the same securities. High overlap means the funds are substitutes; moderate overlap means the second one adds something, and the useful question becomes whether that something is what you wanted. How to find overlap in your ETFs covers the mechanics.

When two funds hold different things, the methodology document explains why, and whether the difference is structural or incidental. The lines that matter are the eligible universe (which companies can be selected at all), the selection rule (size ranking, a screen, or a factor score), the weighting scheme and any caps, and the reconstitution schedule. A gap caused by a capped weighting scheme will persist; a gap caused by a reconstitution that has not happened yet will close.

This is also where you find out whether a fund samples its index rather than replicating it fully. Sampling is normal in broad or illiquid indexes and it introduces a small, persistent source of tracking difference. None of this is visible from a price chart, and it is the reason two funds described identically can behave differently for years.

The expense ratio is charged every year against the full balance, so a small difference compounds over a long holding period and is nearly irrelevant over a short one. The way to make it concrete is to multiply the percentage-point difference by the amount you plan to hold, which gives a rough annual figure, then judge it against the holding period you expect. That converts an abstract comparison of decimals into a number you can weigh against the other fields.

Cost is the cleanest tie-breaker once sameness is established, and a poor first filter before it. A cheaper fund is sometimes cheaper because its index is narrower, its replication is sampled, or its asset base is smaller, all of which show up elsewhere in this list.

These two get used interchangeably and mean different things. Tracking difference is how far the fund's realized return sat from its index over a period, which is the number that actually reached your account. Tracking error is the volatility of that gap, meaning how consistent the tracking has been. A fund can post a low tracking error while lagging steadily by roughly its fee, and another can post a higher tracking error while landing closer on average.

For choosing between two funds on the same index, tracking difference is usually the more useful figure, because it silently includes everything the fund does behind the scenes: fees, sampling, rebalancing costs, cash drag, securities-lending revenue, and dividend-withholding treatment. It is the single best summary of execution quality, and it sits in the fact sheet and annual report.

Assets under management is a rough proxy for liquidity. Larger funds generally show tighter bid-ask spreads and deeper order books, so orders fill closer to fair value, which matters most for large or frequent trades and least for small automatic purchases. Very small funds also carry closure risk, and a closure forces a sale that may be taxable and is rarely timed to suit you.

Size is not a quality score, and a small fund from a large issuer is a different proposition from a small fund from a small one. The check is quick: your broker's quote shows the live bid and ask, and the gap between them is roughly the round-trip cost you pay on top of the fee.

In a taxable account, what you keep can differ from the headline return. Index ETFs are generally tax efficient because in-kind creation and redemption lets them avoid distributing many capital gains, so tax is mostly owed when you sell. That efficiency varies: higher turnover, active strategies, and options overlays all raise distributions. Dividend character matters too, since qualified and non-qualified distributions are taxed differently, and international funds add foreign withholding.

In a tax-advantaged account, most of this stops being a differentiator, which is why placement is part of the decision rather than separate from it. The mechanics are covered in how ETFs are taxed. This is general information, not tax advice.

Most index funds lend out a portion of their holdings to short sellers and earn a fee for it. Issuers return some or all of that revenue to the fund, which slightly offsets the expense ratio and shows up in the tracking difference. It also creates a small counterparty and collateral exposure, normally collateralized and disclosed in the annual report.

This is a minor tie-breaker, not a decisive one. But between two funds identical on index, holdings, cost, and spread, the lending policy and the share of revenue returned to shareholders is a genuine, checkable difference rather than a coin flip.

The last field is the one no screener knows: your own history. If you already hold one of the two funds in a taxable account at a gain, switching to the other realizes that gain, which is a certain cost weighed against an uncertain benefit. Very often that single fact settles a decision that looked evenly balanced on paper.

The reverse case is tax-loss harvesting, where a similar fund is the natural replacement for one sold at a loss. The wash-sale rule disallows the loss if you buy a substantially identical security within the surrounding 30-day windows, and two funds tracking the same index invite exactly that question. The rule is fact-specific and its application to near-identical index funds is not settled by bright-line guidance, so this is a point to take to a tax professional rather than a page on the internet.

Suppose you are deciding between Fund A and Fund B, both broad US equity index funds, both from large issuers, charts effectively on top of each other. Working the list in order:

Holdings and index. Fund A tracks a large-cap index of roughly 500 companies; Fund B tracks a total-market index of several thousand. Overlap at the top is nearly complete, but Fund B carries a slice of small and mid-caps that Fund A does not hold at all. So they are close substitutes, not identical, and the difference is a small-cap tail. That is now the primary question: do you want that tail, and does something else in your portfolio already cover it.

Cost and tracking. Suppose the fees are within a basis point or two. Multiply that difference by the balance you plan to hold and the annual gap is small enough that it should not outrank the holdings difference. Tracking difference for both sits close to the fee, as expected for large, liquid index funds, so neither shows an execution problem.

Size, spread, tax. Both are large with tight spreads, so execution is a wash. Both are structurally tax efficient index funds with long histories of minimal capital gains distributions, so the taxable-account penalty is similar.

What you already own. If you already hold a separate small-cap fund, Fund B's tail partially duplicates it and the choice tilts toward keeping the exposures separate and controllable. If you hold nothing else, Fund B covers more of the market in one line. And if you already own one of them at a gain in a taxable account, the realized tax on switching very likely outweighs everything above.

The honest conclusion is not that one fund wins. It is that seven of the eight fields came back a tie, the eighth (the small-cap tail, and whether you already cover it) is the actual decision, and that field is about your portfolio rather than about the funds. That is what the framework is for.

The order matters more than the list. Steps one and two decide whether this is an exposure question or a cost question, and everything after that only applies once you know which. The general per-metric detail lives in how to compare ETFs, and the category-specific version of the same problem is worked in how to choose a semiconductor ETF.

A long-term holder in a tax-advantaged account. Steps one and three dominate. Tax treatment is largely neutralized by the account, switching costs nothing, and the fee compounds over decades, so once sameness is established, cost is a reasonable decider.

A long-term holder in a taxable account. Step eight often overrides everything. An existing position with an embedded gain makes switching expensive, and distribution history matters more than a small fee difference.

An active trader. Step five moves to the front. Spread and depth are paid on every round trip, which can dwarf an annual fee difference within a handful of trades.

Someone still building a portfolio. Step one is the whole question. What each fund adds to the exposures you already hold matters far more than which of two near-identical funds you pick, and how to build a diversified portfolio is the wider frame for it.

Six of the eight fields are public and can be read off a fact sheet in a few minutes. The two that cannot are the two that usually decide it: how each fund overlaps with the portfolio you already hold, and what switching would actually cost you given your existing positions. Both need your real holdings, which is exactly what a public screener does not have.

Walnut is the assistant for that gap. With your brokerage linked, you can put two funds side by side through Claude, ChatGPT, or the built-in assistant and ask how much each overlaps with what you own, which exposures each would add or duplicate, and how each has tracked the S&P 500 over the window you choose. The connection is read-only and any order waits for your explicit approval. Walnut is not an investment adviser; it turns “which of these two is better” into “which one changes what I already hold”.

When two ETFs look the same, do not start with the fee. Start by proving whether they are the same, using holdings and index methodology. That single step reclassifies the decision: same bet means cost, spread, and tax settle it; different bet means the holdings gap settles it and cost is a footnote.

Then finish with your own position. What you already own decides whether a switch is free or expensive, and it decides whether the second fund diversifies or duplicates. Most stalled comparisons between two similar funds are resolved by that last field rather than by anything printed on either fact sheet.