Hedge Fund Concentration Explained: Reading a Top-10 Book

What hedge fund concentration actually measures

Concentration is a property of the long book a manager reports on Form 13F: how much of total reported value sits in the largest positions. The headline metric is top-10 weight. Sum the value of the ten largest positions, divide by the total value of the filing. A fund at 85 percent top-10 is making a small number of bets that have to be right. A fund at 8 percent is running something closer to an index with tilts.

Top-10 weight is convenient but coarse, so it pays to read a small family of metrics together:

Effective positions deserves a note. Square each position's weight, sum the squares (that sum is the Herfindahl index), then take the reciprocal. A book with 100 names where one name is half the value has an effective count closer to 4 than to 100. It is the best corrective to the illusion that a long tail of tiny positions equals diversification.

One hygiene rule before comparing anything: aggregate by issuer first. A company can appear in a filing under several share classes and CUSIPs, and option legs are reported alongside common stock. Count raw rows and you will overstate diversification. Our guide to reading a 13F filing covers the field-level traps.

Conviction versus diversification: same number, different stories

A high top-10 weight is not automatically bullish and a low one is not timid. Concentration is mostly a strategy fingerprint, and the same number means different things in different hands.

Concentrated value and quality books. The canonical example is Berkshire Hathaway, where a handful of names has long dominated the reported equity portfolio. Managers in this mold argue that genuinely good ideas are scarce, that the eleventh-best idea dilutes the first, and that the correct response to deep research is size. For them, a very high top-10 weight is the design, not a risk failure.

Activists. Activist funds concentrate because the strategy demands it: you cannot influence a board from a 40-basis-point position. Stakes above 5 percent with control intent trigger a Schedule 13D within 5 business days, a much faster disclosure than the quarterly 13F. The interplay between the forms is covered in 13F vs 13D vs 13G vs Form 4.

Platforms and quants. Multi-manager platforms and quantitative funds deliberately run hundreds or thousands of positions under tight per-name risk limits. A 3 percent top-10 weight there is not a lack of conviction. The conviction lives in the process, the factor model, and the netting, none of which a holdings table shows.

Small and new filers. A manager just over the $100 million reporting threshold may hold five positions because the fund is young, not because it is fearless. Small books are mechanically concentrated.

So the first question to ask of any concentration number is: concentrated relative to what? Scan the dispersion across filers on the investors leaderboard and the strategy clusters separate quickly.

How to compute concentration from 13F holdings

The math is one line; the data hygiene is the work. Done by hand from raw EDGAR filings, you would need to parse the information-table XML, resolve CUSIPs to issuers, collapse share classes, and separate option legs from common stock before you can sum anything. Arkolith's Q1 2026 13F dataset (1,824 institutional filers, 1.87 million long positions, $53.7 trillion in reported value) ships with that normalization done, so the computation reduces to two API calls:

# 1. Resolve the fund to its CIK
curl -H "Authorization: Bearer YOUR_KEY" "https://arkolith.com/api/v1/search?q=berkshire+hathaway"

# 2. Pull its latest holdings, then sort by value and sum the top ten
curl -H "Authorization: Bearer YOUR_KEY" "https://arkolith.com/api/v1/funds/1067983/holdings"

From the response: exclude rows flagged as puts or calls so you are measuring the long equity book, aggregate by issuer, sort descending by reported value, and divide the top-10 sum by the filing total. Every holding carries provenance back to the SEC accession number of its source filing, so an agent computing this can cite the exact document on EDGAR full-text search instead of asserting a number from memory. If you are wiring this into an agent, the quickstart covers key minting and the MCP connection.

Two timing notes. 13Fs are due 45 days after quarter end, with 2026 deadlines of February 17, May 15, August 14, and November 16, so recompute after each window closes. And filers amend: a concentration figure computed mid-window can shift when an amendment supersedes the original, which is one of several reasons to recompute from the current dataset rather than caching a number forever.

What a concentrated book does and does not tell you

A 13F covers long positions in US-listed equities and certain related options. It does not show short positions, cash, bonds, most foreign-listed shares, or swap exposure. That asymmetry cuts both ways. A book that looks like five fearless bets may be hedged into something far tamer. The inverse failure is worse: a widely covered family-office collapse in 2021 involved an extremely concentrated, leveraged book built largely through total return swaps, which left almost no footprint in 13F data until it unwound. The SEC's own Form 13F FAQ is explicit about what the form does and does not capture.

There is also the lag. The filing is a snapshot as of quarter end, published up to 45 days later, so a concentrated position you see today may have been trimmed weeks ago.

On the performance question: academic work on managers' "best ideas" has generally found that the largest, highest-conviction positions in institutional portfolios have tended to perform better than the rest of those same portfolios, which is a key reason concentration screens stay popular. Treat that as a research motif rather than a law. Results vary with period, methodology, and how conviction is defined, and concentration amplifies error exactly as efficiently as it amplifies skill. For a fuller accounting of the trust boundaries in this data, see how accurate is 13F data.

Reading a concentrated book: a working checklist

When a fund prints 70 percent of its value in ten names, run through five questions before drawing conclusions.

Tenure. Are the big names long-held compounders or new this quarter? A position that compounded its way into the top slot tells you about patience. A brand-new top-3 position is a statement.

Crowding. Is the concentration idiosyncratic or shared? When the same mega-cap dominates hundreds of books at once, the number describes the market more than the manager. Compare any single fund against the most-owned stocks of Q1 2026 before crediting anyone with originality.

Faster paper. The quarterly 13F is the slowest disclosure in the chain. Corporate insiders file Form 4 within 2 business days of a trade, and activists file Schedule 13D within 5 business days of crossing the threshold. If a concentrated institutional holder is joined by insider buying at the same company (check a ticker's insider feed), independent signals are stacking. Arkolith tracks 51,000+ insider transactions alongside the 13F spine for exactly this cross-check.

Liquidity versus weight. A 20 percent weight in a mega-cap is exitable in days. The same weight in a small-cap is a commitment measured in months, which changes what the position means.

Options context. Check whether reported put legs sit next to the headline equity weight. A large equity position with a large put leg beside it is a different trade than the equity alone.

Concentration is a lens, not a verdict. It tells you how a manager expresses conviction. What it never tells you on its own is whether the conviction is right.

Frequently asked questions about hedge fund concentration

What is a good top-10 weight for a hedge fund?

There is no universal benchmark; the number is a strategy fingerprint. Concentrated value and activist books often hold most of their reported value in the top ten, while multi-manager platforms and quant funds sit in single digits by design. Compare a fund against peers running the same strategy and against its own history, not against an absolute bar.

Does high hedge fund concentration predict better returns?

Not reliably. Academic work on "best ideas" has generally found that managers' largest positions have historically outperformed the rest of their own books, but results vary with period and method. Concentration amplifies skill and error symmetrically, so concentrated books simply have wider outcomes in both directions.

How do I calculate a fund's concentration from its 13F?

Pull the holdings, exclude option legs, aggregate by issuer to collapse share classes and multiple CUSIPs, sort by reported value, and divide the top-10 sum by the filing total. With Arkolith that is one call to /api/v1/funds/<cik>/holdings on already-normalized data. The common mistake is counting raw rows, which overstates diversification.

Why might a 13F overstate or understate true concentration?

A 13F shows only long US-listed equities and certain options, so shorts, cash, foreign listings, and swaps are invisible. A five-name filing can sit inside a hedged, diversified total portfolio, and a swap-heavy book can be far more concentrated than its filing suggests. Filings are also up to 45 days stale, and small filers near the $100M threshold look mechanically concentrated.

This article explains public filings and data concepts. It is not investment advice.