How to Copy Congress Trades: What Works & What Doesn't

Can You Actually Copy Congress Trades?

Yes. Under the STOCK Act of 2012, all 538 members of Congress are required to publicly disclose every stock, bond, and commodity trade over $1,000. These filings are public record. You can read them, analyze them, and use them to inform your own investment decisions. There is nothing illegal about acting on publicly disclosed congressional trading data.

The appeal is obvious. Members of Congress sit on committees that regulate entire industries. They receive classified briefings about national security, economic policy, and regulatory actions. They vote on legislation that directly moves stock prices. 343 of 538 members (63.8%) actively trade stocks while in office. If their trades reflect non-public knowledge, copying those trades should be profitable.

And academic research suggests it is — for the politicians. Studies have shown that congressional portfolios outperform the market by a meaningful margin. But there is a critical difference between their timing and yours. They know what they're buying when they buy it. You find out an average of 49 days later.

That delay is not a minor inconvenience. It is the fundamental problem with every copy-trading strategy that relies on disclosed filings.

The 49-Day Disclosure Delay Problem

The STOCK Act gives members of Congress 45 calendar days to file a disclosure after executing a trade. Based on GovGreed's analysis of 189,595 congressional trades, the actual numbers are worse than the law requires:

The median delay of 28 days means most filings do arrive within the legal window. But the distribution has an extremely long right tail. Serial late filers like Rep. Michael McCaul (R-TX) — who has 6,670 late filings out of 32,302 total trades — create a world where you simply cannot trust that you are seeing trades in a timely manner.

And even the "compliant" filings arrive nearly a month after the trade was executed. In a market where stocks can move 10-20% in a month, that is an enormous information disadvantage.

The Math: Why Delayed Copies Underperform

The alpha from congressional trades — the excess return above the market — is concentrated in the period immediately following the trade. This makes intuitive sense: if a senator buys a stock because she knows a favorable regulation is coming, the stock moves when that regulation is announced, not when her filing is published 49 days later.

The Timeline of Alpha Decay

GovGreed's backtest data from the trade_returns table — covering 51,600 trades with measurable outcomes — shows a clear pattern:

• Days 1-7 — The sharpest price movement occurs immediately after the trade. This is when the informational advantage is highest. The politician knows something; the market does not. Nobody outside Congress sees this trade.
• Days 7-30 — Alpha continues to accrue as the catalyst (bill markup, regulatory announcement, earnings surprise) plays out. Most of the excess return has been captured by this point.
• Days 30-45 — The legal disclosure window closes. Most compliant filings are now public. By this point, the fundamental reason for the trade has often already been priced in by the market.
• Day 49 (average filing) — You see the trade. The stock has already moved. You are buying after the alpha has been captured.

This does not mean every delayed copy trade loses money. If a politician is buying for long-term fundamental reasons — a sector with sustained growth, a company with improving earnings — the stock may continue to appreciate after you enter. But the edge from congressional knowledge is largely gone by the time you act on it.

Copy-Trade Services: What They Actually Do

Several services have emerged to automate the process of copy-trading Congress. Platforms like Autopilot, Quiver Quantitative, and various fintech startups let you follow congressional trades and, in some cases, automatically execute matching trades in your brokerage account.

These services do solve one real problem: speed of reaction. When a filing hits the disclosure database, automated services can notify you (or execute on your behalf) within minutes or hours, rather than waiting for you to manually check the filings. That shaves hours or days off the reaction time — but it does not solve the fundamental 49-day disclosure delay.

What Copy-Trade Services Typically Offer

• Filing alerts — Notifications when new STOCK Act disclosures are published
• Auto-execution — Automatically place trades in your brokerage when selected politicians file
• Politician filtering — Choose which members to follow based on past performance or committee membership
• Portfolio mirroring — Attempt to replicate a politician's overall portfolio allocation

The Honest Assessment

These services are well-built tools that serve a real demand. The problem is not the software — it is the underlying data constraint. No amount of engineering can fix the fact that you are reacting to a trade that happened 28-49 days ago. Even if the filing alert reaches you in 5 minutes, you are still acting on information that is, at minimum, nearly a month old.

There is a meaningful difference between "copy the trade as soon as the filing appears" and "predict the trade before the filing exists." The first is reactive. The second is predictive. Both are legitimate approaches, but they produce very different results. For a walkthrough of all available data sources and platforms, see our guide on how to track congressional stock trades.

A Better Approach: Predictive Signals

Instead of waiting for a filing and reacting to it, what if you could predict which politicians are likely to trade which stocks — before the trade is disclosed, or even before it happens?

This is the approach GovGreed takes. Rather than copy-trading from filings, GovGreed's intelligence engine analyzes 7 layers of data to identify the conditions that precede congressional trades:

1. Politician quality (20% weight) — Historical win rates, trading styles, committee alignment. 247 politicians profiled with quality scores from S-tier to F-tier.
2. Herd signals (20% weight) — When 3+ politicians buy the same stock within a short window, it is a stronger signal than any individual trade. 31 active herd signals tracked.
3. Bill correlation (16% weight) — 256,112 historical trade-bill correlations mapped. When a bill advances, politicians with correlated trading histories are predicted to trade again.
4. Technical context (12% weight) — RSI, SMA, volume surge analysis for 32,000 ticker-date combinations. Are politicians buying at technically favorable levels?
5. Sector momentum (12% weight) — Congressional buying/selling patterns aggregated at the sector level. Front-running sector-wide moves.
6. Contribution patterns (10% weight) — 565 patterns linking campaign donations to subsequent trades in the donor's sector.
7. Lobbying alignment (10% weight) — 2,101 patterns linking lobbying activity to politician trades in the lobbied sector.

Each layer provides a partial signal. Combined through a weighted scoring model with convergence multipliers (3 signals = 1.3x, 4 signals = 1.5x, 5+ signals = 2.0x), the result is a master score from 0-100 that grades the conviction behind any given trade or prediction.

Signal Tier Performance

Not all congressional trades are equal. A junior representative buying $1,001-$15,000 of an S&P 500 stock through a financial advisor is very different from a senior committee chair buying $500,000 of a stock in the sector she regulates, on the same day her committee marks up a related bill, while two other committee members buy the same stock.

GovGreed's signal scoring engine assigns every qualifying trade a tier based on its master score. These tiers have been backtested against actual 30-day returns using the signal_backtest_stats view:

The threshold for signal scoring is deliberate: only trades with an estimated value of $50,000 or more are scored. This excludes the thousands of small, routine trades ($1,001-$15,000) that are likely advisor-managed or portfolio rebalancing rather than conviction-driven insider purchases. Currently 61 politicians meet this threshold and receive signal scores.

Practical Strategies That Actually Work

If naive copy-trading is undermined by the disclosure delay, what approaches actually produce results? Based on GovGreed's analysis of 189,595 trades, 256,112 bill-trade correlations, and 2,790 signal scores, here are the strategies with the strongest empirical backing:

1. Watch for Herd Signals

When 3 or more politicians buy the same stock within a short time window, this "herd" behavior is a meaningfully stronger signal than any individual trade. It suggests that multiple members — potentially from different committees or parties — have independently concluded that a stock will appreciate.

GovGreed currently tracks 31 active herd signals. Historically, herd convergence has been one of the strongest individual components in the 7-layer scoring model.

2. Track Committee-Aligned Trades

Politicians trade disproportionately in sectors their committees regulate. A member of the Senate Armed Services Committee buying defense stocks before a procurement bill markup is a qualitatively different signal than a random representative buying the same stock. GovGreed's committee alignment score measures the correlation between a politician's committee assignments and their trading sectors.

When an upcoming committee markup involves a bill with high investability (scored against the bill_impacts table's 908,000 bill-ticker mappings), and committee members have a history of trading in that sector, a prediction is generated before the trade happens.

3. Detect Recurring Patterns

Some politicians trade with predictable regularity — DCA-like purchasing intervals in the same stocks or sectors. GovGreed's pattern detection engine identifies politicians with at least 3 purchases of the same stock within a consistent time interval (up to 120 days between buys) and predicts the next purchase within a 21-day window.

This currently covers 11 politicians with sufficiently consistent patterns. It is a narrow but high-confidence signal: if a politician has bought MSFT every 45 days for the past year, the next purchase is predictable.

4. Use Bill-Trade Correlations as Leading Indicators

GovGreed has mapped 256,112 correlations between specific bills and specific trades. When a bill from the 119th Congress advances through committee, politicians who traded on similar bills in the past are predicted to trade again. This correlation-based approach covers 76 politicians and generates the highest volume of predictions.

• ✗ Don't: Wait for filings and copy blindly. You are 49 days behind. The alpha is gone.
• ✗ Don't: Follow high-volume traders indiscriminately. Ro Khanna has 48,257 trades across 1,372 tickers — most are noise, not signal.
• ✓ Do: Focus on high-tier signals (A+ and above) where multiple intelligence layers converge.
• ✓ Do: Monitor herd behavior — 3+ politicians buying the same stock is statistically significant.
• ✓ Do: Track committee markups and cross-reference with member trading histories for forward-looking predictions.

GovGreed's 819 Active Predictions

GovGreed runs 4 prediction engines daily through the refresh_all_predictions() pipeline. Together they produce 819 active predictions covering 76 politicians — more than 10x the coverage of the 61 politicians who receive signal scores.

All predictions are exposed through the predictions_latest view, which deduplicates daily accumulation and keeps only the most recent prediction per (politician, sector, source). The pipeline runs daily at 11:45 PM UTC on weekdays, with stale predictions automatically expired after 30 days.