ETF overlap: how much of the same thing you actually own

ETF overlap is the company you own twice: once as a stock, and again inside a fund that holds it. A book with 19% in Apple and 19% in an S&P 500 fund does not own 19% of Apple. It owns more. You measure the gap by unwrapping every fund you own into its constituents and re-totalling by company instead of by ticker.

What ETF overlap is

Every ETF is a basket with a published list of what is in it and how much of each. Own the fund and you own a slice of every name in that basket, sized by that name's weight in the fund. Overlap is what happens when two lines on your statement point at the same company: two funds that share their largest positions, or a fund and the stock itself sitting side by side.

This is a risk question rather than a tidiness question. Diversification is about how many independent things your money is riding on. Overlap quietly turns what looks like eight positions into something that behaves like five, because a bad week for one company runs through the direct position and through every fund that carries it, at the same time and in the same direction. The count of tickers stays the same. The number of separate bets falls.

Nothing on an ordinary brokerage screen shows this. The screen shows tickers and dollar amounts, and every ticker takes one row and looks like one position. The arithmetic that collapses those rows back into companies happens somewhere else, which is why almost nobody runs it on their own book.

How to measure it

The measure

Look-through. You expand each fund into the companies inside it, then re-sum by company across the whole book. For one fund and one company, the arithmetic is a single multiplication:

contribution = (fund's % of your book) × (company's % of that fund) ÷ 100

Do that for every company in every fund you own, total the contributions company by company, then combine each total with whatever you own of that company directly. The result is your look-through exposure. The figure worth looking at is not the look-through number on its own, it is the distance between it and what you thought you owned. Clavix computes this under a versioned model, lookthrough-v1, and calls that distance hidden overlap.

A worked example

Take a $100,000 book. It carries $19,000 of Apple bought directly, which is 19%, and $19,000 in an S&P 500 fund, another 19%. The owner reads the statement and sees a 19% position in Apple.

Apple's weight inside a broad S&P 500 fund is roughly 7%. Clavix stores 7.1% in the fallback holdings table it keeps for when an issuer fetch fails, stamped 17 June 2026. That is an illustrative figure, not today's published number, and it moves with the fund every day. The arithmetic:

19% × 7.1% ÷ 100 = 1.35% of the book, or $1,349 of Apple inside the fund.
19% direct + 1.35% inside the fund = 20.35% look-through exposure to Apple.

So the real figure is 20.4%, not 19%, and the 1.35% gap clears the 0.25% floor Clavix uses before it will call anything hidden overlap. The whole distance between what the statement says and what you own is one input: how heavily the fund carries the name. A broad index fund spreads across hundreds of companies, so it adds a little. A fund built around one part of the market adds far more.

Sector funds really do run that concentrated. In the same stored table, Microsoft is 22.5% of the technology sector fund and Nvidia is 21.4%, so two companies account for 43.9% of it. A book with 19% in Microsoft directly and 19% in that fund carries 19% + 4.28% = 23.3% in a single company.

The other case is the one that catches people out, because there is no direct position at all. Someone with 20% of their book in a Nasdaq-100 fund and not one semiconductor stock anywhere owns about 1.8% of Nvidia: 20% × 9.2% ÷ 100 = 1.84%. Direct exposure 0%, look-through exposure 1.84%, all of it hidden. Those rows tend to be the surprising ones, because they name companies the owner never picked.

How much overlap is too much

Clavix publishes no percentage above which overlap is called excessive, and a fixed one would be wrong in both directions. Overlap is not itself a defect. It is an accounting correction: the same 3% of doubled-up exposure is trivial inside a forty-name book and material inside a six-name one, and 2% of extra exposure to a steady megacap is a different fact from 2% of extra exposure to one volatile small cap. What matters is the size of the gap between what you think you own and what you own, read against the rest of your book.

What Clavix does publish is the two floors that decide when a row is worth printing at all.

There is a second reason no single number would survive contact with real data, and it is about honesty rather than statistics. Clavix uses published constituent weights exactly as the issuer published them and never rescales them to 100. Some issuers publish every constituent. Where they do not, what Clavix has is a slice of the largest holdings, so a fund's stored rows might sum to 36% rather than 100%. Spreading the fund's whole weight across those rows would inflate every constituent by about 2.8 times and report a book that does not exist. Instead each fund contributes only the share it can genuinely account for, and the remainder is reported per fund as unattributed.

So an overlap figure computed this way is a floor. It is deliberately too low, and it says so. A number that is too low and admits it is recoverable, because you know which direction the truth lies in. One that is too high and stays quiet about it is not. The same principle covers funds nobody has a holdings table for: those are excluded from the arithmetic and named, never counted as one opaque position with zero overlap. Scoring an unopened fund as zero renders on screen as "no hidden overlap", which is a claim about the fund rather than an admission that nobody looked inside it.

Run your own version and the same discipline applies. Total what you can source, treat it as a minimum, and write down which funds you could not open. If you want a read on each ticker before you start, the free Clavix portfolio grader will grade every name in your book, funds included, on the same A+ to F scale the app uses.

Why overlap hides

The first reason is that funds with different names, different tickers and different issuers routinely carry the same top ten. A total market fund and a large cap index fund look like two decisions. In the fallback table Clavix stores, their lists lead with the same five companies in the same order: Apple, Microsoft, Nvidia, Amazon, Meta. The total market fund holds thousands of names the other one does not, and every one of them sits in the tail at fractions of a percent. Where the money actually is, the two funds are nearly the same fund.

The second reason is that overlap accumulates one reasonable decision at a time. A total market fund, then an S&P 500 fund for the core, then a technology fund because that is where the growth has been. Each purchase felt like a step toward more diversification, and each one picked up the same megacap block again at a different concentration. The technology fund is the one that does the damage, because it carries two of those same companies at 22.5% and 21.4%. Three funds, one bet, and nothing on the statement says so.

The third reason is the sharpest. You own a company directly because you thought about it and chose it, and you own a broad fund alongside it precisely to balance that choice out. If the company you picked is one of the largest constituents of the fund you picked to offset it, the fund is not balancing anything. It is topping the position up. The stronger your conviction in the direct name, the likelier it is a heavy constituent, because the funds retail investors reach for are weighted by market value.

None of this is a mistake anyone made. The reader bought diversification and got concentration with extra steps, because the only place the two views reconcile is an arithmetic that no statement performs and no fact sheet prints.

How Clavix reads this

Straight about scope: look-through is a real, versioned, deployed measurement model at Clavix, with the floors and the never-rescale rule described above. It is not a screen you can open on this site. So take the numbers on this page as Clavix's published measurement model, and the arithmetic as yours to run today.

What is live right now is the free web grader. It takes three to twenty holdings by ticker plus a share count or a dollar amount, and returns a single A+ to F grade for the whole book. Funds are graded on the same footing as companies: the fund-specific dimension relabelling the grader used to do is gone, and an all-fund book is now described in the same words a book of shares is. What a fund book does surface is the third layer behind the grade, independence, which compares how your holdings actually moved against each other over the past year. Enter a broad-market fund alongside a large-cap growth fund and that layer collapses, because the two are largely the same bet. That is the closest the grader gets to the subject of this page.

What it does not return is the subject of this page. The grader does not unwrap your funds into their constituents, does not re-total by company across them, and will not tell you that two of your funds share a top ten or that a fund is doubling a stock you own directly. It scores each ticker as a ticker. The independence layer will notice that two funds move together, but it cannot tell you why, and shared holdings are only one of the reasons two things move as one. That line is worth stating plainly, because internal concentration and cross-fund overlap are two different measurements and neither of them is on the screen today. For how the dimensions behind each holding's grade are built and what each grade band means, see the Clavix methodology.

Check your own portfolio

Look-through arithmetic starts with knowing what each ticker in your book is carrying on its own, and that part takes about a minute.

Frequently asked questions

How do I check if my ETFs overlap?

Pull the published holdings list for each fund from the issuer page. For every name on it, multiply the fund percent of your book by that name percent of the fund. Total those contributions across all your funds, then include whatever you own directly in the same names. Anything that turns up twice is overlap. Most published lists are a top ten, so your answer is a floor, not a total.

Do a large cap index fund and a total market fund overlap?

Heavily, at the top. In the holdings table Clavix stores as a fallback, the two lists lead with the same five companies in the same order: Apple, Microsoft, Nvidia, Amazon, Meta. The total market fund carries thousands of names the index fund does not, but those sit in the tail at fractions of a percent. Overlap concentrates where the money is.

How much of my portfolio is really in one stock?

Start with the percent you own directly. Then for each fund, multiply the fund percent of your book by that name percent of the fund, and total everything. A book with 19% in a stock directly plus 19% in a fund that carries the name at 7% owns about 20.4% of it. If the fund carries that name at 21% instead, the same book owns 23%.

What is ETF look-through?

Look-through means unwrapping every fund in a portfolio into the companies it holds, then re-totalling by company rather than by ticker. It converts a list of funds into a list of businesses. Clavix runs it under a versioned model called lookthrough-v1, which uses issuer published constituent weights exactly as published and never rescales them to 100.

Does Clavix show ETF overlap?

The free web grader does not unwrap funds. It scores every ticker you enter, funds included, and returns one grade for the whole book. It does compare your holdings against each other in one narrow sense: the independence layer measures how much they moved together over the past year, so two funds that are largely the same bet show up there. It does not unwrap funds into their constituents, does not re-total by company, and will not tell you which holdings two of your funds share.

Why does my ETF overlap number look too low?

Because most issuers publish only a top ten or a top twenty five, so the weights you can actually get sum to well under 100. Clavix uses them as published and reports the remainder per fund as unattributed, rather than spreading the fund weight across the rows it can see. Spreading a fund across rows summing to 36% would inflate every constituent about 2.8 times.

For educational purposes only. This page explains how a risk measure is calculated; it is not investment advice, not a recommendation about any security, and not a prediction of return. Clavix is not a broker-dealer or a registered investment adviser. Consider consulting a licensed professional before making investment decisions.