Panel Paper: Diagnosing Diffusion: Untangling Inter-Municipal Effects on Local Regulation

Friday, November 8, 2019
Plaza Building: Lobby Level, Director's Row I (Sheraton Denver Downtown)

*Names in bold indicate Presenter

Gwen Arnold, University of California, Davis and Le Ahn Nguyen Long, University of Münster


How jurisdictions influence one another’s policy decision-making is an enduring debate in policy
and political science. While scholars have theorized about how and the conditions under which one
jurisdiction’s policy choices influence another’s, specific empirical tests of these hypotheses are
relatively scarce. And while a number of studies of policy adoption and diffusion identify neighbor
and peer effects as significant drivers (e.g., a jurisdiction is more likely to adopt a policy when a
neighbor does so first), these studies rarely identify the specific nature of these inter-municipal
effects. Our project tackles these gaps in scholarship.
The primary mechanism of policy diffusion among jurisdictions are competition, imitation, learning,
and vertical coercion (Shipan and Volden 2006, 2008). Further, jurisdictions are expected to engage
in the first three of these behaviors with geographic neighbors and peers. We propose specific
measurements for each mechanism and engagement dynamic and, in dyadic analysis, test their
relative impact on (a) policy uptake and (b) regulatory similarity.
Using a novel, full-text database of municipal laws concerning high-volume hydraulic fracturing
(fracking) in New York, we create a dyadic dataset containing all possible pairs of jurisdictions in the
state. Jurisdiction dyads are described by variables quantifying the extent to which dyad partners are
similar on socio-economic, political, and demographic measures which we link to neighbor and peer.
For example, geographic proximity between two dyad partners measures neighbor effects, while
similarity of past land use and environmental policymaking measures peer effects. We regress these
indicators (using quadratic assignment) on policy outcome data, anticipating that dyads with greater
proximity and peer similarity will be more likely to experience policy adoption.
After identifying the cases wherein policy diffusion occurred (both dyad partners adopted a fracking
policy), we then explore whether that diffusion was driven by competition, imitation, or learning.
Our tool for untangling these drivers is policy textual similarity. Dyad partners in an imitative
dynamic will have the most similar policies while those in a learning dynamic will have the least,
since policy learning facilitates jurisdiction-specific policy tailoring. (Competitive dynamics may
foster low or high similarity.) Using automated content analysis, we extract from the policies key
regulatory features and compile these into dyadic measures of regulatory similarity between all
adopters. Dyadic similarity scores are regressed, using quadratic assignment, on socio-economic
indicators representing economic competition (similarity between dyad partners in size and industrial
profile), imitative dynamics (similar jurisdiction type and similar citizen political partisanship), and
learning dynamics (participation in common venues for policy discussion), controlling for
geographic proximity and peer status. This modeling effort will help us understand whether our
indicators of diffusion type appear predictive and the types of diffusion undergirding the spread of
anti-fracking policies across New York.