One hedge fund buying a stock is an opinion. Five funds buying the same stock in the same quarter is a pattern. 13F consensus is the practice of reading institutional filings across funds instead of one at a time, to find the positions where multiple respected managers agree. It is the most useful thing you can do with 13F data, and also the easiest to misuse. This article explains how consensus is measured, what it signals, and why crowded positions deserve both attention and caution.
The raw material: quarterly 13F filings
The data source is SEC Form 13F. Institutional investment managers with at least $100 million in US-listed 13(f) securities must report their long equity holdings within 45 days after each calendar quarter ends, as described in the SEC's Form 13F FAQ. Every filing is public and free on SEC EDGAR.
A single filing answers one question: what did this fund hold at quarter end. Consensus analysis asks a different question across a chosen set of funds: which stocks do several of them hold, and which of those were they actively buying. The unit of analysis shifts from the fund to the stock.
How consensus is measured
A basic consensus build has four steps.
Pick the funds deliberately. Consensus is only as meaningful as the panel behind it. A useful panel mixes styles, so agreement between funds is informative rather than automatic. A concentrated value investor, a macro fund, an activist and a distressed specialist agreeing on one stock says more than four clones of the same strategy agreeing. Panels are usually small, six to twenty funds, because beyond that the aggregate drifts toward looking like an index.
Extract and normalize holdings. Pull each fund's latest 13F and the prior quarter's, and match positions by CUSIP rather than by company name, since names are abbreviated inconsistently across filings. Respect the put and call flags, because an option row is not a plain long position.
Compute the deltas. For each fund, classify every position as a new buy, an add, a trim, an exit or unchanged, comparing share counts rather than dollar values so that price moves are not mistaken for trading.
Aggregate by stock. For every stock, count how many funds in the panel hold it, and how many were buyers this quarter. A simple conviction score combines the two: breadth of ownership plus recent buying activity. A stock held by five of six funds where three added or initiated ranks above a stock held by five funds that all sat still, which in turn ranks above a name held by two.
The ranking that falls out of this is the consensus table. The top of the table is where the panel's independent research processes converged in the same quarter.
Why crowded positions matter: the bull case
The case for paying attention to consensus rests on what a 13F position actually represents. Each line is real capital committed after a research process, by a team with resources most individual investors do not have. When several such teams with different mandates reach the same name in the same window, three useful things follow.
First, the idea has survived multiple independent filters. Different funds have different analysts, different valuation frameworks and different risk committees. Convergence means the thesis is robust to more than one way of looking at it.
Second, buying activity dates the signal. Broad ownership alone can be stale, a legacy of positions built years ago. Fresh adds and new buys across funds indicate that the thesis is attractive at recent prices, which partially offsets the 45 day reporting lag baked into all 13F data.
Third, consensus compresses your research funnel. Nobody can study every stock. A ranked consensus table across respected managers is a defensible way to pick the ten names worth your next month of reading, which is a far more realistic use of 13F data than mechanical copying. Studies of naive 13F-copying strategies report mixed and often negative results, largely because of the reporting lag, so the honest framing is that consensus tells you where to look, not what to buy.
Why crowded positions matter: the risk case
Crowding is not only a signal. It is also a risk factor, and any serious use of consensus data holds both ideas at once.
A stock heavily owned by similar institutions embeds a shared assumption. If that assumption breaks, many large holders may want out through the same narrow door at the same time, and the exit itself moves the price far more than the news alone would justify. Episodes of rapid, correlated unwinding in widely held hedge fund positions are a recurring feature of markets, and funds themselves track crowding metrics precisely to manage this exposure.
There is also a valuation effect. By the time a stock reaches the top of every consensus screen, a lot of informed buying has already happened. The expected return left on the table for a latecomer is smaller than what the early funds captured, and the downside if the crowd turns is larger.
The practical takeaway is not to avoid crowded names, and not to chase them, but to read the crowding level as context. High consensus with fresh buying flags a strong, current, shared thesis, and also flags that the position will be volatile if the thesis cracks. High ownership with net trimming across the panel is a very different message from the same ownership count, which is exactly why buying activity belongs in the score.
Building consensus yourself, or reading it ready-made
Everything above can be done by hand with free tools. EDGAR full-text search and company search give you every filing. Dataroma shows overlap across its curated superinvestors for free, including which stocks appear in many tracked portfolios. WhaleWisdom offers broader screening across the full filer universe, with some features paid.
The manual route is a spreadsheet exercise repeated every filing season: two filings per fund, CUSIP matching, delta classification, then aggregation. It is genuinely doable and instructive to build at least once. It is also the kind of work that quietly stops happening by the third quarter you have to redo it.
The ready-made route is a maintained report. The Smart-Money 13F Consensus Report tracks six widely followed managers: Berkshire Hathaway, Citadel, Bridgewater, Pershing Square, Appaloosa and Scion. It ranks 50 stocks by a conviction score based on how many of the funds hold each name plus their buying activity, lists each fund's new buys, adds, exits and trims, and refreshes each filing season. Every figure traces back to the underlying EDGAR filing, so any line can be verified at the source in a minute.
However you build it, the discipline is the same. Consensus is the start of the research process, not the end. The table tells you where smart money agrees. The work of deciding whether they are right is still yours.
The Smart-Money 13F Consensus Report distills the latest filings of six top managers into one ranked consensus table. Get the free preview.
DataSignals Lab publishes data and research. This is not investment advice.
Prefer the tool to the write-up? Browse all data feeds or connect the free MCP server.