A Computational Approach to Tracing the Progress of Policy Ideas in Legislation

Whose ideas and proposals prevail in Congress? Legislative scholars have traditionally studied important subjects such as legislative effectiveness and agenda control by studying the progress of bills. In chapter 7 of Congress and the Politics of Problem Solving we show that using bills for these purposes has important limitations. Bills are “vehicles” for policy ideas. Often the successful sponsor of a bill that becomes law is a committee or subcommittee chair who claims the credit for a collective policy development process involving “must pass” legislation such as a program reauthorization or a law passed in response to a salient external event.

Where do the ideas in laws come from? In the 111th Congress, HR 3590 as introduced was titled the Service Members Home Ownership TAX Act of 2009. As enacted, HR 3590 was titled the Patient Protection and Affordable Care Act, otherwise known as ObamaCare. The legislative history of HR 3590 is unusual but by no means unique. It illustrates a widely appreciated but rarely studied point: bills evolve as they move through the legislative process. They pick up and shed policy provisions that can range from relatively small adjustments in language to complete bills (in the case of omnibus legislation).


The figure above illustrates a new text-based approach to systematically tracing the progress of policy ideas or provisions (as opposed to bills) in legislation. We use text reuse methods from computer science (think plagiarism detection) to trace similar language wherever it appears in legislation. Figure 1 is based on a comparison of section texts in introduced bills in the 111th Congress to section texts found in the enacted version of the Affordable Care Act (ACA).  A substantial number of ACA sections match sections in bills sponsored by other lawmakers (of both parties). One of these bills (HR 3692) became law after the ACA’s enactment, but apparently not before some of its original provisions found their way into the ACA.

Whose ideas gain currency? The figure below begins to illustrate how the study of policy ideas as opposed to bills might alter current understandings of legislative effectiveness. Scholars have long noted the importance of institutional position as a predictor of bill success. But when we compare the sponsors of successful policy ideas (whether a section of an introduced bill found its way into a different bill that became law) to the sponsors of successful bills (bills that became law), we find that legislative effectiveness is substantially more widespread. We can ask which individual members have the greatest success inserting language into legislation.


Where do ideas get picked up? The final figure (below) begins to illustrate how we might trace the migration of policy ideas from bills under the jurisdiction of one committee to laws emerging from other committees. In this figure, the migration appears to be toward the committees dealing with important  “must pass” legislation such as health care, banking reform, deefense and appropriations.


As the paper emphasizes, many challenges remain.  But by shifting the focus from the bill as the unit of analysis to the evolving substance of bills, the method does seem to open up a wealth of new research opportunities.