Association for Public Policy Analysis & Management

Since President Obama created the open data initiative in January of 2009, governments across the United States at every level have adopted and implemented open data systems. These systems are a hybrid of internal-facing IT systems and external-facing “e-government” components. Viewed in a historical context, open data is the latest in a series of attempts to reform informational institutional arrangements in American government that stretch back to the mid 20^th century (Bimber, 2001, 2003; J. A. Musso & Weare, 2005; Verba, Schlozman, Brady, & Brady, 1995). The motivations for open data vary, as it is expected to improve transparency and engagement, reduce administrative costs, and support performance management systems in a similar manner to previous e-government initiatives (Baldwin, Gauld, & Goldfinch, 2012; Kim & Lee, 2012; J. A. Musso & Weare, 2005; Tolbert & Mossberger, 2006). Open data requires public service agencies to surrender control of the analysis and interpretation of their data; this paper posits that departments may perceive this disruptive organizational change as a potential threat to their professionalization and maintenance. Examining implementation at the level of the municipal department, the paper analyzes the institutional, organizational, and supply- and demand-side factors that promote or inhibit implementation. It utilizes data on all departmental posting of records on open data management systems gathered using Python from all US cities with populations over 100,000 that have adopted open data systems (n = 76). The model analyzes the association of city- and departmental political, fiscal, and institutional factors with implementation, measured in two different ways. The first measure is the level of contribution of machine-readable data to open data systems by different departments/bureaus within a city government. The second uses qualitative coding to differentiate between usable, open data and files that, while available, do not constitute genuine open data. A two-stage model is employed to control for bias by first determining the factors associated with the likelihood of open data adoption. This analysis is central to the study of the open data phenomenon because it speaks to how the system is implemented in practice – a stage of the innovation process widely recognized as crucial (Andersen, 2008; Damanpour & Schneider, 2006, 2009; Fernandez & Wise, 2010; Hansen, 2011; Jun & Weare, 2011; Nelson & Svara, 2011; Torsten Oliver Salge & Vera, 2012; Walker, 2008).