Every quarter the same headlines appear. A famous fund "bought" a stock. Another "dumped" one. The number attached is almost always a dollar figure, and the dollar figure is almost always the wrong thing to look at.
A 13F is a list of long US equity positions held by an institutional investment manager with at least 100 million dollars in qualifying assets. The SEC describes the form and who must file it on its Form 13F page. The filing tells you what a manager held at the end of a quarter. It does not tell you why. If you want to extract anything close to intent from that list, you have to look at how a position is sized relative to the rest of the portfolio, not at the raw dollar amount.
This article is about that shift in reading. It is not investment advice. It is a guide to the mechanics of a filing and the handful of ratios that separate a real bet from noise.
Why dollar value misleads
A 200 million dollar position sounds big. For a manager running 500 million dollars it is enormous. For a manager running 150 billion dollars it is a rounding error, probably held for index tracking or as a hedge leg.
Dollar value scales with the size of the fund. Conviction does not. A manager who puts 15 percent of their book into one name has made a decision that will visibly affect their returns. A manager who puts 0.1 percent into the same name may not even remember owning it.
The headline writers lean on dollar value because it produces the biggest numbers and because the largest funds generate the largest numbers. That bias rewards exactly the managers whose individual positions carry the least information. A trillion dollar asset manager that holds every stock in the S&P 500 will show up as a "top buyer" of almost anything that rose in price last quarter, simply because the market value of its holding went up.
So the first rule is simple. Convert every position to a percent of the manager's total reported 13F value before you read anything into it.
Percent of portfolio is the base unit
Each 13F lists, per security, the number of shares and the market value at quarter end. Sum the market values and you have the total reportable portfolio. Divide each position by that sum and you have the weight.
Once you have weights, several things become visible that dollar values hide:
- A new position that enters directly at 5 percent of the book is a different event from one that enters at 0.3 percent. The first is a decision. The second is a probe, a hedge, or an arbitrage leg.
- A position that grows from 2 percent to 8 percent through added shares is a conviction signal. A position that grows from 2 percent to 8 percent because the stock tripled while the share count stayed flat is not a decision at all. It is drift.
- A manager with 15 positions and a top holding at 25 percent is telling you something very different from a manager with 1,200 positions and a top holding at 1.5 percent.
The share count column matters here. Market value changes for two reasons, price movement and trading. Only trading reflects a choice. Always compare share counts quarter over quarter, then look at the weight. If shares rose and weight rose, the manager added into the position. If shares were flat and weight rose, the market did the work.
Concentration tells you how the manager thinks
Concentration is a portfolio level property. It describes how much of the book sits in the top few names. A manager running a concentrated book has a reason for every position, because every position is large enough to hurt. A manager running a diversified book may hold a name because a model told them to, because an index includes it, or because a client mandate requires it.
Two measures are useful and both are easy to compute from the raw filing:
- Top ten weight. Add the weights of the ten largest positions. A value above roughly half the book indicates a concentrated manager. A value below a fifth indicates a broadly diversified or quantitative one.
- Number of positions. Fewer than 30 names usually signals a fundamental, high conviction approach. Several hundred or more usually signals quantitative, index, or multi strategy management.
Neither number tells you who is right. It tells you whose individual positions are worth reading. A concentrated manager's new position is informative because they chose to displace something else to make room for it. A diversified manager's new position is one of hundreds of adjustments made that quarter.
This is also why the same stock appearing across many managers only becomes interesting when you check what kind of managers they are. Ten concentrated funds each putting 4 percent of their book into one name is a stronger signal than two hundred index-like funds each holding it at the weight it carries in a benchmark.
Signals worth isolating
Once positions are in percent terms and managers are sorted by concentration, a few patterns carry most of the information in a filing.
New positions above a threshold. A manager opening a name at 3 percent or more of the book, in a single quarter, has placed a real bet. Filter for this rather than for "new positions" in general. Most new positions in most filings are tiny.
Adds into existing large positions. A manager who already holds 6 percent of the book in a stock and raises the share count by 40 percent is doubling down. That is a stronger statement than the initial purchase, because they have had time to watch the thesis play out and chose to increase exposure.
Full exits of meaningful positions. A position that was 5 percent last quarter and is gone this quarter is a decision. A position that was 0.2 percent and is gone may be a mandate change or a reshuffle.
Trims that do not match price moves. If a stock fell 30 percent and the manager reduced shares by 10 percent, the weight fell for two reasons and you can separate them. If a stock rose 30 percent and the manager cut shares by half, they took profit deliberately.
Position rank changes. Watch which names move into or out of a manager's top five. Rank is a weight ranking, so it already normalizes for fund size and ignores dollar noise.
What a 13F cannot tell you
Reading weights carefully does not remove the limits of the form itself.
The data is stale. Filings arrive up to 45 days after quarter end, so a position may have changed or closed before you see it. The mechanics of that delay are covered in our piece on 13F deadlines and the 45-day lag.
The form covers long positions in listed US equities and certain options and convertibles. It does not include short positions, most derivatives, bonds, cash, or non US listed stocks. A manager who appears to hold 10 percent of the book in one name may be hedged with a short that never shows up. The SEC's official list of 13F securities defines exactly which instruments are reportable each quarter, and anything outside it is invisible.
Put and call options are reported with a flag in the filing, and they are reported at the notional value of the underlying shares. A large "position" flagged as a put is a bearish or protective stance, not a long bet. Always check the put/call column before counting a position as conviction.
Managers can also request confidential treatment for positions they are still building, which delays disclosure. When those positions are eventually revealed, they can appear as large new holdings that were actually accumulated over several quarters.
Doing this yourself with free sources
Everything described here can be done with public data. The SEC publishes every 13F filing through EDGAR full text search and through the structured 13F data sets on its Division of Economic and Risk Analysis page. Those data sets are quarterly bulk downloads of every filing's information table in a machine readable format.
With that data you can compute total portfolio value per manager, weights per position, quarter over quarter share count changes, and concentration measures with a spreadsheet or a short script. Several free websites also present individual manager filings with weights already calculated. The computation is straightforward. The work is in cleaning the data, matching filers across quarters, handling amended filings, and doing it for several thousand managers every quarter rather than for one.
How we use this
The Smart-Money 13F Consensus report applies the approach above across a curated set of managers who run concentrated, fundamental books. Every position is expressed as a percent of the manager's reported portfolio. Share count changes are separated from price drift. New positions, meaningful adds, and full exits are flagged only when they cross a weight threshold, and the consensus score counts how many independent concentrated managers hold or added the same name in the same quarter. The result is a short list of stocks where conviction, not dollar volume, is doing the talking. You can see the latest edition and the methodology at Smart-Money 13F Consensus.
Want the signal instead of the raw filings? Get the free 13F Consensus preview. Prefer the tool to the write-up? Browse all data feeds or connect the free MCP server.