Loading...
Loading...
What actually works in smart-money tracking — every strategy tested against real price data with honest, forward-looking entry timing and full methodology. No hype, no cherry-picked windows.
We run a site built on smart-money data — 13F filings, congressional trades, insider filings — so we had a choice: sell the fantasy that copying billion-dollar funds beats the market, or actually test it and tell you what we found. We tested it. Across a dozen backtests — cloning the legends, the consensus buys, the concentrated bets, every congressional purchase and sale — almost nothing beat the index in a way you could repeat. Here's the whole scorecard, the two whispers that survived, and what this data is honestly good for.
Every quarter brings the same omen: Buffett is sitting on a record pile of cash, so a crash must be coming. We pulled Berkshire's balance sheet from its SEC filings for every year back to 2010 and measured what actually matters — how much of the company is deployed in stocks. From 2010 to 2018 the equity portfolio averaged about 21% of assets; it peaked at 36.6% in 2021 (mostly a ballooning Apple stake), then reverted to 23% by 2024. The 'record cash' is a dollar record because Berkshire tripled in size. As a share of the company, today's Buffett is simply the Buffett of most of the last sixteen years.
'Follow the smart money' assumes the pros have a durable edge you can borrow. So we borrowed it — every year back to 2014, we cloned every 13F fund we track, held the year, and compared it to the S&P. Cloning beat the market in exactly six of twelve years. The average looks like a small edge until you notice one year, 2020, is carrying all of it: strip out the COVID rebound and cloning trails the index. And the two most recent bear-ish years, 2021 and 2022, were double-digit disasters.
Two groups get accused of knowing something you don't: members of Congress and the executives buying their own companies' stock. We priced every disclosed buy from both. The best stock-picker in Congress beat the market by 58 points a year across 20 trades; the best insider beat it by far more — on a bet almost no one would have copied. Here's who, and what it actually means.
The headlines pick a winner — a senator up triple digits — and call it a signal. We did the opposite: for every year back to 2014, we copied every disclosed congressional purchase, held twelve months, and measured it against the S&P. The average is a modest +2 points a year of alpha. But it's negative in four of ten years, fewer than half of individual trades beat the market, and the whole edge rides on 2016 and 2020. 'Follow Congress' is a coin flip dressed up by a couple of great years.
Everyone says to follow the high-conviction picks. So we correlated portfolio concentration against 3-year clone alpha across 58 funds. The relationship is negative: the most concentrated funds' clones averaged −3.8 points a year and beat the S&P only 32% of the time, while the most diversified averaged +6.1 and beat it 76%. The reason is not what the headline implies.
The buy side gets the headlines, but the sharper accusation is on the sell side: if members of Congress trade on what they know, the stocks they quietly sell should fall afterward. We took every individual-stock sale we could price back to 2014 and measured what happened over the next twelve months. The typical sold stock does lag the market a little — but they dump future winners just as often, and whether any of it looks smart depends entirely on which way the market went. It's a whisper of a signal, not a secret.
The whole case against congressional stock trading is the committee edge: a senator on Armed Services buying defense, an Energy member buying oil, trading on what they see in briefings. We tagged 3,552 disclosed purchases back to 2017 by whether the stock's sector falls under the member's own committees, then measured each trade's next-12-months return against the S&P — across the 2018 correction, 2020 crash and 2022 bear, not just one rally. Overseen-sector trades beat the market by 3.6 points; everything else by 2.9. The committee edge is essentially zero.
Our manager-cloning study said you can't pick winners ahead of time. We tested that directly: split 60 funds' clone records into two back-to-back periods. Past alpha does predict future alpha — 70% of top-half funds stayed top-half, correlation +0.32. But the magnitude collapses: the +91-point star reverted to +20. Persistence is real; extrapolation is a trap.
The unspoken promise of following billion-dollar funds is capital protection: when the market breaks, the pros are supposed to be positioned for it. We tested that directly — cloned every fund we track and measured how far each fell against the S&P inside the three worst drawdowns of the last decade. In the fast 2020 crash the smart money cushioned the fall a little. In the slow 2022 repricing it fell half again as hard as the index, and seven funds in ten did worse than just holding the S&P. 'Smart money' isn't a hedge; it's leveraged beta to the exact thing that breaks.
A federal contract award is real revenue, and 'follow the government money' sounds like an edge. We priced every verified award we track against the stock. The three companies pulling in the most contract dollars outside of pure defense — Accenture, Booz Allen, and Palantir — are down 47%, 25%, and 23% this year. Contract flow tells you what a company sells to Washington, not where the stock is going.
Buffett, Ackman, Icahn, Druckenmiller, Tepper, Loeb, Burry, Peltz. Equal-weight, cloned from their 13F filings with no lookahead over ten years, the group returned 15.2% a year versus the S&P's 15.4%. Six of the eight individually lost to the index. The 'follow the legends' pitch, tested honestly.
We built a daily clone portfolio for every tracked fund — top-20 holdings, rebalanced at filing dates, no lookahead — and measured 3-year alpha vs SPY. The median fund's clone added exactly nothing. The winners are real, concentrated, and impossible to pick in advance.
We replayed every disclosed trade of 54 members of Congress with honest execution — you only trade when the disclosure becomes public. The spread between the best and worst member is enormous, and that spread is the whole story.
Holding the 8 stocks that the most concentrated hedge funds overweight beats SPY in 73% of rolling one-year starts — a real, repeatable edge. Then you compare it to QQQ, look at the drawdown, and the story gets honest.
When 3+ tracked hedge funds open the same new position, is that a buy signal? We replayed every such trade since 2021 with honest entry timing. Win rate: a coin flip. Excess return vs SPY: zero.
Across 34,256 trades ≥$1M in the latest 13F cycle, active managers pulled $10.7B out of energy, poured $17.1B into tech — and staged the quarter's best argument in Microsoft, where 23 funds sold $22.7B while Bill Ackman built a $2.1B stake.
One weekly email with notable insider buys, sells and congressional disclosures. Unsubscribe anytime.