Looking at state-by-state historical bill passage information we use machine learning to identify factors correlated with a bill’s passage at each stage of the legislative process. We customize our models for the US Congress and each state (plus the District of Columbia). Our models are also specific to each of 102 state legislative chambers. This work is done by our team of experts and Professor Mark Crain, a leading authority on quantitative analysis of government activity.
We’ve identified more than 40 factors that have significant forecasting power. These factors include, for example, the partisan composition of the chamber, the relevant legislative committees and past passage rates, the author’s leadership position and committee assignments, the author’s past success rate, and the bill’s current stage in the process.
Legislative Outlook does not take the following into consideration:
- Textual Content of Legislation: although textual analysis is not part of Legislative Outlook, the forecast DOES take into account the subject matter of the bill based on the committee it is referred to. For example a Green New Deal referred to a Progressive Environment Committee in CA may be more likely to pass than similar legislation a conservative state like Texas.
- Sponsor Discretion: political intent and other motivating factors
- Competing Interests: is the state or chamber preoccupied with other priorities