Every few months a chart goes viral showing some member of Congress up 200% and the caption writes itself: they know something. We've made that mistake ourselves — it's easy to find a winner after the fact. So this time we did the un-viral thing. We went back to the earliest disclosures we have, and for every single year, we copied every congressional purchase, held each for twelve months, and compared it to the S&P 500 over the same window. No cherry-picking a member, no cherry-picking a period.
Year by year, the "edge" flickers on and off
| Year bought | Purchases | 12-mo return | Alpha vs S&P | Beat S&P |
|---|---|---|---|---|
| 2014 | 298 | +13.3% | +4.1 pt | 57% |
| 2015 | 258 | +2.3% | −1.4 pt | 50% |
| 2016 | 241 | +28.6% | +9.2 pt | 52% |
| 2017 | 318 | +17.9% | +2.4 pt | 44% |
| 2018 | 328 | +13.0% | +2.2 pt | 53% |
| 2019 | 184 | +8.3% | −4.2 pt | 41% |
| 2020 | 428 | +59.4% | +14.8 pt | 55% |
| 2023 | 227 | +26.2% | −5.8 pt | 32% |
| 2024 | 932 | +14.1% | −2.3 pt | 40% |
| 2025 | 1,479 | +25.9% | +2.4 pt | 41% |
Averaged across the years, copying Congress beats the S&P by about +2 points a year. Stop there and it sounds like a real edge. Keep reading the column and it falls apart:
- It's negative in four of the ten years — 2015, 2019, 2023, and 2024. You don't get to skip the bad ones in advance.
- The average is two years in a trench coat. Strip out 2016 (+9) and the COVID-rebound class of 2020 (+15) and the remaining eight years average roughly zero alpha. The signal isn't broad; it's two lucky vintages.
- Most individual trades lose to the market. Across the whole sample, only
45% of congressional buys beat the S&P over the next year — worse than a coin flip. The positive years are carried by a handful of big winners, not by members being right more often than they're wrong.
- The recent years are the worst. 2023 (−5.8) and 2024 (−2.3) are both negative. Whatever edge existed in the 2010s is not showing up now.
Why the flattering average is a trap
This is the same trap as "follow the smart money" and "clone the legends": a metric that looks like skill until you demand it be repeatable. A strategy that beats the market in six years, trails in four, and depends on two outliers isn't an edge — it's market exposure with extra steps and a 45-day disclosure lag stapled on. And as we showed separately, the committee-oversight story doesn't hold either: the returns aren't coming from inside information about the industries these members regulate.
The honest version of "follow Congress"
Congress, in aggregate, is a slightly above-average buyer of large-cap US stocks — no more, and wildly inconsistent year to year. If a specific member has a genuinely great multi-year record (a few do), that's about that person's picking, not a congressional superpower you can index. Watch the disclosures on /congress as a curiosity and an accountability tool — not as a strategy that reliably beats buying the index and doing nothing.
Methodology: every disclosed purchase from 2014 to mid-2025 with a resolvable ticker, measured over the 12 months after the transaction date against the S&P 500 over the same window, equal-weighted, grouped by the year of purchase. Years shown have 20+ priced trades; 2021–2022 are omitted for sparse coverage. A rolling forward window means each year is judged against its own market, so a rising tape doesn't flatter the result. Not investment advice.