Bills with even one cosponsor from the other party become law 5 times as often as one-party bills

Bills and joint resolutions, 113th to 118th Congress. 5% of bills with a cross-party cosponsor became law, against 1% of the rest.

Exhibit 1Share of bills that became law, by cosponsor count, percent of bills introduced, 113th–118th

One party onlyAt least one cross-party cosponsor

So what

Cosponsor count alone predicts little. Volume of support is not breadth of support: one-party bills with many cosponsors pass less often than bipartisan bills with a few.

Breadth carries weight in our model too. Cross-party cosponsorship is one of the passage model's inputs, alongside whether the sponsor is in the majority.

Majority sponsors still win more. 3.2% of majority-party bills became law, against 2% for the minority.

Exhibit 235 vetoes since 2013, 2 overridden, one mark per vetoed bill

Exhibit 3Days from introduction to law, median and middle half, by Congress

Exhibit 4Passage model's track record on the 118th Congress, predicted vs. actual share enacted, rated at introduction; trained on 113th–117th only [4]

So what

The model ranks well. Given a random enacted bill and a random failed one, it rates the enacted one higher 77% of the time (AUC 0.77).

Its percentages run high. In every band, fewer bills passed than predicted. This is why the directory shows bands, not percentages. The live model is retrained each night on every completed Congress.