ETF overlap checker

How the overlap is computed

1. Weighting basis: the latest published holdings snapshotEach fund contributes the top holdings of its most recent successful snapshot, with the weight the issuer prints for each line. The snapshot date is the fund's dataAsOf; it is shown on the fund page and in the worked example below, so you always know which portfolio date the comparison is based on.
2. Security matching by normalised holding nameFund houses spell the same company differently. Before comparing, every holding name is upper-cased, punctuation is dropped and corporate suffixes and share-class markers (Inc, Corp, PLC, Ltd, AG, SA, Class A, ADR, ORD…) are removed, giving one key per issuer. “Apple Inc” and “APPLE INC.” therefore match; two lines of the same issuer inside one fund are summed. ISINs of the underlying securities are not used, so two share classes of the same issuer count as one company.
3. Holdings without weights count as emptySome issuers publish their largest positions as names only. Such a fund is matched by name, so the shared names are still listed, but its overlap is reported as null rather than as 0%: the figure cannot be measured, which is not the same as no overlap. The fund is listed under namesOnly in the API response.
4. Substitute baskets are not comparedA synthetic, swap-based ETF holds a basket of collateral that backs the swap, not the index it tracks. Its holdings say nothing about its exposure, so they are treated as an empty portfolio, the fund appears under excluded and every pair involving it carries comparable: false. Its sector and country tables, which describe the index, are unaffected and are what the look-through tool uses.
5. Overlap = sum of min(weight in A, weight in B)For each shared key the smaller of the two weights is taken, and these minima are added. Identical portfolios give 100%; disjoint ones give 0%. This is the common-holdings measure most overlap tools use, and it is symmetric: A against B equals B against A.
6. The result is a lower boundOnly disclosed positions can be matched. An ETF usually publishes its whole portfolio, an active fund often only its top ten, so the disclosed share of each fund tells you how much of the portfolio was visible. Positions that are not published cannot add to the overlap, so the true figure can only be higher than the one shown.
7. At most 25 shared holdings are listedThe shared lines are sorted by their counted weight and the list is capped at 25 per pair. The overlap percentage and the shared count are computed over every matched holding, so the cap only shortens the table, not the figure.

Worked example

How it works