Dynamic responsiveness in the American states : legislators, constituents, and organized interests

Abstract

An independent commission redrew California's electoral map after the 2012 redistricting cycle, inducing large, exogenous shocks to the composition and policy preferences of many districts. The first paper of the dissertation assesses repositioning among members of the state legislature. Did they respond when redistricting led to changes in the policy ideology of their districts? The result speaks to the kind of representation that constituents receive-and the obstacles facing would-be reformers. The paper is the first in the literature to identify the causal quantity of interest using design-based inference; it also improves on previous measures of district ideology. Contrary to prior findings, there is little evidence of responsiveness to shifts in district preferences from redistricting. This result points to the role of strong parties and organized interests in the selection of representatives and legislative activity.

In my second paper, I demonstrate the effectiveness of supervised machine learning methods in recognizing textual references to firms, organized interests, or any other political actors (an application of named entity recognition), and then resolving these references to real-world referents (an entity resolution task). Together, these methods make possible the large-scale measurement of political actors or their activity from sources such as diplomatic cables, transcripts, and administrative or legislative records. Organized interests are embedded in the legislative process in state capitols, writing bills and participating in committee meetings; they contribute stakeholder perspective and testify to the technical points of proposed legislation. Studying exactly which groups participate addresses a minimal standard for democratic governance. The third paper accomplishes this using the measurement strategy described in the second paper.

It reveals how organized interests engage on specific bills (or bill versions) and expands the scope of measurement beyond activities whose disclosure is required under state law. Diverging from typical measurement strategies identifies less-resourced groups, in particular citizen and issue organizations, engaging in undisclosed legislative activities. The paper argues for an alternative view of the distribution of political voice in the states, and the integration of research on dynamic responsiveness and organized interests.