Does Copying Superinvestors Actually Work?

Every quarter, the world's best-known fund managers are forced to show their hands. Institutional investment managers with over 100 million dollars in US equities must file Form 13F with the SEC, disclosing their long positions within 45 days of quarter end. That creates an obvious temptation: why do your own research when Warren Buffett, Seth Klarman, and Bill Ackman are doing it for you, and publishing the results for free?

This idea goes by several names: coattail investing, copycat investing, 13F cloning. The pitch is simple. A great manager might spend millions of dollars and thousands of analyst hours deciding to buy a stock. When that decision shows up in a public filing, you can piggyback on all of that work for the cost of reading a document. The question is whether the strategy actually holds up once you account for the delay, the blind spots, and human nature. The answer, perhaps surprisingly, is a qualified yes, with some large caveats worth understanding before you copy a single trade.

The theory: free-riding on professional research

Markets reward information and analysis. A concentrated, research-driven fund manager holds a stock because a team concluded, after serious work, that it is worth more than its price. If that judgment has any skill behind it, the stock should outperform on average over the manager's holding period. A cloner who buys the same stock shortly after disclosure captures most of that same holding period, minus the first 45 to 135 days.

For a manager who holds positions for three to five years, missing the first few months costs relatively little. For a manager who trades in and out within a quarter, the filing may describe a position that no longer exists. That single observation explains most of what works and what does not in 13F cloning.

What the evidence actually says

Academic research on copycat investing is more supportive than skeptics tend to assume. Several studies have found that portfolios constructed from the disclosed holdings of skilled managers can earn meaningful excess returns, even when the copier waits until the filing is public. The information in a good manager's book does not fully decay in 45 days, largely because the market is slow to absorb the theses behind long-horizon positions.

The most cited piece of evidence concerns concentration. In their working paper "Best Ideas," Randolph Cohen, Christopher Polk, and Bernhard Silli examined the positions that managers weighted most heavily relative to a benchmark: their highest-conviction bets. They found that these best ideas significantly outperformed both the market and the rest of the same managers' portfolios. The implication is uncomfortable for the fund industry and useful for cloners: managers do appear to have stock-picking skill, but it is concentrated in a handful of names, and it gets diluted as portfolios expand for business reasons rather than investment ones. If you are going to copy anyone, the research suggests copying their largest active positions, not their fortieth-largest holding.

The practitioner case is best represented by Mohnish Pabrai, who has spent years publicly advocating what he calls shameless cloning. Pabrai has said repeatedly that most of his best investments were ideas lifted from other investors' 13F filings and then verified with his own work. He treats cloning not as an embarrassment but as a discipline: why generate original ideas when you can select from ideas already vetted by people with better resources?

Both strands of evidence point the same direction. Copying works best when applied to low-turnover, high-conviction managers, and when focused on their biggest positions. It works worst when applied indiscriminately across a manager's whole book, or to managers whose edge is speed.

Why it fails in practice

If cloning were easy money, everyone would do it and the returns would vanish. In the real world, several things go wrong.

The 45-day lag kills fast signals

A 13F filed in mid-August describes holdings as of June 30. Positions opened in early April are already four and a half months old by the time you see them. For a quantitative fund or an event-driven trader, that information is stale beyond usefulness. Even for fundamental managers, a stock can run 30 percent between purchase and disclosure, changing the risk-reward calculation entirely. The lag does not destroy the strategy, but it filters which managers are worth following.

You see longs, not the whole trade

13F filings cover long positions in US-listed equities and certain options. They exclude short positions, most bonds, foreign-listed shares, currency hedges, and swaps. A position that looks like a bullish bet may be one leg of an arbitrage, a hedge against something invisible, or a merger play with a defined exit. Copying one leg of a two-legged trade means taking risk the original manager deliberately neutralized.

Sizing and timing differ

Even when you copy the right stock, you rarely copy the position. The manager bought at a different price, sized it against a portfolio you cannot see in full, and may add on weakness in ways you will only learn about a quarter later. Many cloners buy after a filing-driven pop, then discover in the next filing that the manager was already trimming.

Taxes and discipline

A fund's after-fee, pre-tax return is not what you get. Cloning quarterly means realizing gains on the manager's schedule rather than your own, which can be tax-inefficient in a taxable account. The bigger killer is behavioral. The moment a copied position drops 30 percent, you have no thesis of your own to lean on. Managers hold through drawdowns because they understand why they own the stock. Cloners who never did the work tend to sell at the bottom, capturing the volatility of the strategy without its returns.

Survivorship in who you follow

Most people decide whom to clone by looking at recent performance, which builds survivorship bias directly into the strategy. The manager who looks brilliant today may simply be the lucky tail of a large distribution, and hot streaks attract cloners at exactly the wrong time. Past performance does not guarantee future results, and that caution applies doubly to a strategy whose entire premise is extrapolating someone else's past performance.

How to clone intelligently

The evidence and the failure modes suggest a fairly clear playbook.

• Pick slow, concentrated managers. Favor investors with multi-year average holding periods and portfolios of 10 to 30 names over funds holding hundreds of positions. Browse the manager list and look at turnover and concentration before you look at returns. A book like Berkshire Hathaway's portfolio, where top positions are held for years or decades, loses very little signal to the filing lag.
• Weight their best ideas. Pay attention to the top five positions and to new positions that immediately appear at meaningful size. A 6 percent starter position says far more about conviction than a 0.4 percent tag-along.
• Watch for cluster buys. When several unrelated managers you respect initiate the same name in the same quarter, that is a stronger signal than any single filing. Pages like consensus buys and most bought stocks surface exactly this pattern.
• Size positions yourself. Your portfolio, liquidity needs, and risk tolerance are not the manager's. Copy the idea, not the weight, and never let a cloned name become a position you could not defend on its own merits.
• Treat 13Fs as a research pipeline, not a buy list. This is Pabrai's actual method, often lost in the retelling. He clones the idea generation, then does the valuation work himself. A filing tells you a smart investor found something interesting. It does not tell you the price at which it stops being interesting.

Testing it before you trade it

The honest way to evaluate cloning is to measure it, not to argue about it. You can use the 13F clone simulator on this site to see how a portfolio that mechanically copied a given manager's filings would have performed, lag included, and compare that against simply holding an index fund. Running that comparison across a few managers is instructive: some books clone well, others lose most of their edge in the reporting delay, and the difference usually comes down to turnover.

Copying superinvestors can work. The research supports it, and one of the better-known value investors of the past two decades has built a career on it. But it works as a disciplined process applied to carefully chosen managers, with your own analysis layered on top, not as a shortcut around thinking. Read the filings, follow the money, and then do the part the filing cannot do for you. None of this is investment advice; it is a description of how the evidence and the practical constraints fit together, and every investor's situation differs.