Methodology
Turning congressional disclosures into a systematic signal
Members of the U.S. Congress disclose their securities transactions under federal disclosure requirements. Thousands of transactions become public every year — but an isolated purchase by a single lawmaker tells us very little.
Congress Consensus takes a different approach.
Instead of following individual politicians, we look for consensus: situations where multiple lawmakers independently purchase the same publicly traded company within a defined period.
Our model then ranks those consensus signals using several dimensions, including the number of distinct buyers, the size of disclosed purchases, the recency of the transactions, the presence of potentially contradictory selling activity and the historical quality of the lawmakers behind the signal
No private information. No prediction of individual lawmakers. Just a systematic analysis of public disclosures.
How the strategy works
Congress Consensus transforms raw congressional trading disclosures into a weekly, rules-based portfolio through four successive steps.
Collect publicly disclosed stock transactions from members of the U.S. Congress.
Identify companies purchased by multiple lawmakers within the observation window.
Rank qualifying companies using the Congress Consensus scoring model.
Build a diversified portfolio from the highest-ranked signals and reassess it weekly.
No look-ahead bias
A transaction only enters the Congress Consensus model once it has been publicly disclosed.
Congressional transactions can be reported days or weeks after they actually occur. Using the transaction date alone in a historical backtest would therefore give the model information that investors did not yet have.
To avoid this, Congress Consensus uses the public filing date to determine when a transaction becomes available to the model. A trade executed earlier but disclosed later cannot influence the portfolio before its disclosure.
This point-in-time approach is applied throughout our historical analysis.
If the information was not public at the time, the model cannot use it.
Portfolio construction
Congress Consensus is designed as a concentrated but diversified equity portfolio built from the strongest signals identified by the model.
At each weekly review, eligible companies are ranked using the Congress Consensus scoring framework. The portfolio holds a maximum of eight positions, subject to diversification and data-availability rules.
Individual stock exposure is capped at 15% of the portfolio. The strategy uses no leverage and no short selling.
When a qualifying security cannot be reliably traded or priced, the corresponding allocation may remain in cash rather than being artificially reassigned to another position.
Signals are frozen at the end of each week using only information publicly available at that time. Portfolio changes are implemented at the next eligible market session.
The Congress Consensus score
Every qualifying company receives a systematic score based on five independent dimensions:
Consensus — How broadly the buying signal is shared across distinct lawmakers.
Purchase size — The relative magnitude of the disclosed purchases.
Recency — More recent public disclosures receive greater relevance.
Contradictory activity — Selling activity is incorporated to avoid treating conflicting signals as pure consensus.
Lawmaker quality — The model evaluates the historical outcomes of previously disclosed purchases associated with each lawmaker.
A point-in-time quality model
Lawmaker quality is calculated point-in-time. At any historical portfolio date, the model can only evaluate outcomes that would already have been observable on that date.
Recent transactions whose evaluation period has not yet elapsed cannot contribute to a lawmaker’s historical quality score.
The model also applies conservative statistical shrinkage when a lawmaker has a limited track record. Rather than assigning extreme scores based on only a handful of trades, insufficient evidence pulls the assessment toward a neutral baseline.
This reduces the influence of small samples and makes the quality factor deliberately harder to earn.
Backtesting methodology
The Congress Consensus portfolio was backtested from 2020 onward, using earlier historical disclosures where required to establish point-in-time lawmaker quality.
Historical simulations reconstruct the information set available at each portfolio date. Securities cannot enter the model before the corresponding congressional transaction has become public, and lawmaker-quality information is subject to the same point-in-time constraint.
Portfolio signals are generated weekly and implemented at the next eligible market session. The simulation includes 10 basis points of transaction costs on portfolio changes.
Performance is evaluated not only over the full historical period, but also across calendar years and rolling 12-month and 24-month periods. The strategy is compared against broad U.S. equities and congressional-trading benchmarks rather than judged solely on its cumulative return.
Limitations
Congressional disclosure data has important limitations.
Transactions are generally disclosed after they occur, sometimes with a substantial delay. Reported transaction values are typically expressed as ranges rather than exact amounts. Disclosures may contain amended filings, inconsistent security identifiers or securities for which reliable historical market data is unavailable.
Congress Consensus is designed to account for these limitations where possible, but it cannot eliminate them.
Historical simulations also involve assumptions about execution, transaction costs and data availability. Backtested results are hypothetical and do not represent the performance of an actual account.
Some historical securities could not be included where reliable market data was unavailable. These observations were excluded rather than estimated or reconstructed.
Past performance, including backtested performance, is not indicative of future results. Congress Consensus provides research and informational content and does not provide individualized investment advice.