The Impact of Open Government: Assessing the Evidence

"[O]pen government proponents must design transparent procedures that take into account who is likely to be in a position to respond to the information made available by transparency, and to offset the existing inequalities of power and engagement."

In this Brookings Institution paper, authors Vanessa Williamson and Norman Eisen review the empirical and theoretical literature examining the international impact of open government to the end of answering the question, does open government actually work? They develop a rubric that organises what we know about the impact of open government and identifies the contexts in which open government efforts are likely to be successful. They emphasise that even accessible, important, well-publicised information is not enough to create change if people do not have channels of influence. Finally, they identify a burgeoning body of research that examines how open government initiatives can strengthen the channels of influence necessary to their success, offering recommendations for policymakers and an agenda for further research on the subject.

Williamson and Eisen employ a broad definition of open government, focusing on 3 governance processes that allow the perspectives, needs, and rights of citizens - including the most marginalised - to be addressed: (i) initiatives to increase transparency; (ii) interventions intended to expand public engagement and participation; and (iii) efforts to improve responsiveness and accountability. In assessing whether open government "works" or is "effective", Williamson and Eisen look for those interventions that the evidence shows cause critical improvement in people's lives (e.g. by improving health care, reducing corruption, increasing voting rates, and so on). They bring together research on open government reforms, from right-to-information laws in India, to grassroots monitoring of health care in Uganda, to federal audits in Brazil, to election monitoring in Ghana, to school-quality report cards in the United States.

Based on an analysis of the literature, the authors express these 6 features as a series of questions that proponents of such programmes should pose:

1. Have the proponents identified the specific principals (e.g., segments of the public, civil society, media, and other stakeholders) intended to benefit from the new open government initiative?
2. Is the information revealed by the initiative important to the principals?
3. Is the information accessible and publicised to the principals?
4. Can the principals respond meaningfully as individuals?
5. Are governmental agents supportive of the reform effort?
6. Can the principals coordinate to change their governmental agents' incentives?

Where open government initiatives have been effective, Williamson and Eisen find, proponents have clearly identified the principals they were trying to reach - instead of imagining a unitary "public", they identified the specific principals who are expected to respond to and benefit from the new information the initiative is intended to provide - and publicised information that was important and accessible to those principals. For instance: "Open government proponents must ensure that their interventions actually reach their intended audience. Accessibility concerns confront both top-down and bottom-up approaches to open government. Open data projects should be adaptable for use not only by individuals but by media, academics, and civil society organizations. They should also be amenable to aggregation to the level of official accountability. Those considering direct monitoring as an approach to transparency should ensure that the information in question is truly accessible to the monitors. In addition, there is a critical role for a free and active press in reporting open government information. Here, as throughout our analysis, open government works best when the institutions and mechanisms of democracy are strong."

Williamson and Eisen stress that (footnote number removed): "Open governance proponents should seriously consider whether the information made available is actually of interest to the principals....But it is also important to consider the possibility that local priorities might not align with those of academics, public officials, or civil society organizations....A striking example of divergent local and national priorities comes from the persistence of polio vaccine noncompliance in Nigeria. Naively, it might seem that a community would resist polio vaccination only because of fear or ignorance of the vaccine's effects, and that an information campaign would be the best solution to ensure polio eradication. In reality, however, vaccination noncompliance in Nigeria is part of a 'strategic move by citizens to gain the government's attention' and extract concessions in the form of better infrastructure and schools. The success of government interventions in this area is dependent on a clear understanding of local priorities....The manner in which information is delivered can influence perceptions of the information's importance, but equally or more essential is a respect for the local context in which principals must choose their priorities."

In addition, at least one of the following conditions held in effective open government initiatives (discussed in depth in the report): either the principals could respond meaningfully on their own, or they could do so with the support of government officials, or they could do so through a coordinated effort by the principals to change the behaviour of their representatives in government. In the section on the latter (coordination), approaches that have shown promise in the literature include: recognise the social component of collective action; seek communications mechanisms that resist capture; privilege local knowledge; build new points of access to decision makers; and delegate actual decision-making power. "In sum, there is clear evidence that open government initiatives thrive when recipients of new information about government have access to channels of influence, such as competitive elections and robust grassroots organizations. Where these avenues do not yet exist, there is some suggestive evidence that channels of influence can be built within an open government framework."

Having assessed what we know about where and when open government works, the authors suggest some avenues for future study. Williamson and Eisen begin by highlighting the need for more rigorous evaluation, drawing on both qualitative and quantitative methods of analysis; for instance, they recommend. expanded use of randomised controlled trials (RCTs) to fill crucial gaps that remain in the literature. Then, they offer 3 primary recommendations for the next generation of research: to confront selection bias in open government research, to increase attention to the political mobilisations that make open government possible, and to ensure that research addresses both the initial implementation and the sustainability of open government successes.

"Ideally," Williamson and Eisen conclude, "open government becomes the status quo. In the coming years, we look forward to more research that examines not only where open government initiatives show early success, but where open government becomes institutionalized."