Skip to main content
ETF Classification is the label layer FocusAlpha computes for every US-listed ETF. Each fund carries two independent label systems side by side: the wrapper tier, which says what kind of instrument the fund is, and the four-level taxonomy, which says what its holdings actually are. They are never merged, because a leveraged fund and a plain index fund can wrap the same exposure — and separating the two axes is what makes both usable.
Plan: Fund · Credits: 1 per call; the taxonomy endpoint and screen facets are free

​
The Two ETF Label Systems

Wrapper tier

Nine mutually exclusive values — plain_beta, leveraged_inverse, structured_outcome, option_income, physical_commodity, futures_commodity, crypto, currency, alternative — plus the exposure_hidden flag.

Four-level taxonomy

asset_class → category → segment → specializations, over 900 dotted-path codes computed from holdings, each with a stated semantic.

​
What You Can Use ETF Classification For

Build a peer group that is genuinely comparable, using the level-3 segment code as the peer key. Separate structural risk from exposure: find every leveraged fund regardless of what it tracks, or every health-care fund regardless of how it is wrapped. Aggregate flows or assets by segment. Detect a fund whose portfolio drifted away from the category its name implies — the taxonomy is computed from holdings, so the drift shows up as a dated label change.

​
What the Fund Itself Declares

Separately from both computed label systems, the record carries the issuer’s own claims: the fund name, the filed index flag, the fund-of-fund flag, and the filed leveraged/inverse flag. These are kept apart from FocusAlpha’s computed labels so the two can be compared rather than confused — a fund whose filed flags and computed labels disagree is a finding, not an error.

​
ETF Classification Date Convention

Labels are monthly; numbers are daily. The classification is recomputed from holdings once a month, while NAV, flows and the other measured columns refresh daily. labels_as_of.holdings_as_of is the portfolio date the current labels were computed from; labels_as_of.label_since is when the current label was assigned; the history array dates every previous label change, so a drift can be read directly rather than inferred. Two “unknown” values mean different things: unknown.no_holdings (no holdings feed yet) and unknown.ambiguous (holdings arrived but did not decide a label).

​
ETF Classification Sources

The taxonomy is calculated by FocusAlpha from issuer-published holdings; the wrapper tier is derived from the fund’s filed structure and holdings shape; the declared flags are read from SEC filings. The vocabulary itself is a published artifact served by the free taxonomy endpoint.

​
Query ETF Classification

A fund’s labels are returned in the labels block of the ETF profile. The full vocabulary comes from the free taxonomy endpoint. To select funds by label, use the ETF screen. See also ETF Wrapper Tiers and ETF Taxonomy for each system in full, ETF Holdings for the data the labels are computed from, and ETF Screening for filtering by them.