Chapter 7 Bill Models

7.1 Hypotheses

  • Bill Content - More conservative bills will do better in R chambers, more liberal bills will do better in D chambers
  • Party
    • Given issue ownership, Dems want to pass health bills more than Reps
    • Within-party, liberals should be more interested in passing health bills than conservatives
  • Polarization
    • Polarization makes it more difficult to pass legislation

7.2 Putting it all together: Bill Models

Table ?? shows the results of our model for three configurations of legislatures: unified Republican, Unified Democratic, and split.

7.3 Modeling Choices

Because of the complexity in my data, my modeling strategy will to be to use a multilevel model, which can efficiently pool information at multiple levels of observation (Raudenbush and Bryk 2002; Gelman and Hill 2006). They avoid a stark choice between complete pooling or ignoring contextual differences across units, or no pooling (often called fixed effects) which implies ignoring differences across units. The extent of the multilevel model’s partial pooling is dictated by the data.

Varying intercepts for states and years account for unit- and time-specific heterogeneity in levels.

The specific model I use here is a simple linear probability model. While a generalized linear model like logit is typically used for a binary dependent variable like vote, the more modern approach is to use a simpler linear model as the gains in interpretability typically more than make up for the losses inherent in allowing the predicted values outside of the 0-1 range. I ran the generalized models as well, and include them in the appendix. These are qualitatively identical, and thus for simplicity I only show the former in the main results.

7.3.1 Bill level results

  • Bill sponsored by conservatives suffer in Democratic states, and vice versa for Republican states; but the penalty is greater in the former for being ideologically out-of-step.
  • Controversial bills that encounter one or more party unity votes along the way are much less likely to be passed, whereas non-controversial bills that see one or more unanimous roll calls along the way are more more likely to be passed.

7.3.2 Legislature level results

  • Moderate Democratic are much less likely to pass any bill than liberal Democratic legislatures
  • Conservative Republican legislatures are somewhat less likely to pass any bill than moderate Republican legislatures
  • Polarized legislatures are bad for the passage of any legislation, in either Republican or Democratic unified legislatures.

We can plot the marginal effect of our key variables. Figure ?? shows the effect of bill conservatism. Bill sponsored by conservatives suffer in Democratic states, and vice versa for Republican states; but the penalty is greater in the former for being ideologically out-of-step.

Figure ?? shows the effect of legislative ideology. Moderate Democratic are much less likely to pass any bill than liberal Democratic legislatures. Conservative Republican legislatures are somewhat less likely to pass any bill than moderate Republican legislatures.

Figure ?? shows the effect of legislative polarization. Polarized legislatures are bad for the passage of any legislation, in either Republican or Democratic unified legislatures.

References

Gelman, Andrew, and Jennifer Hill. 2006. Data Analysis Using Regression and Multilevel/Hierarchical Models. Cambridge University Press.
Raudenbush, Stephen W., and Anthony S. Bryk. 2002. Hierarchical Linear Models: Applications and Data Analysis Methods. Second Edition. Newbury Park, CA: Sage Publications.