Principal-agent vs. Local Context: Immigration Policy and Executive Control

• In Event: Diversity and Bureaucratic Behavior

Abstract

Through the decades, public administration and policy scholars have emphasized the tools available to executives for controlling or influencing the behavior of government agencies (e.g., Balla 1998; Moe 1982). These principal-agent models have illustrated the power and limits of how chief executives can shape the implementation of policy. Yet, prior research has overlooked an important factor for influencing bureaucratic discretion. We argue that local context, specifically socio-political conditions, can mitigate the effects of executive-level guidance. We test our thesis by examining the county-level deportation decisions of ICE agents from the years 2013 through 2018. Using dynamic panel modelling, we predict local deportations starting with the early Secure Communities program, continuing under Obama’s two-year Priority Enforcement Program (PEP), and up through Trump’s zero-tolerance policy. We find that Obama-era executive guidance, advising agents to target only dangerous criminal immigrants, does lead to a significant national decline in total removals. However, local context matters, as conservative localities with strong anti-immigration sentiment continue to see high deportation numbers, even during PEP. This study adds nuance to the greater scholarly understanding of political control over the bureaucracy. Based on our analysis, it appears street-level agents are most responsive to chief-executive direction in the absence of local-level opposition to top-down demands.

Authors

• William Dean Schreckhise, University of Arkansas, Fayetteville

• Daniel E. Chand, Kent State University