Research forms the heart of the National Center for Digital Government. Our research studies focus on two major questions. The first is: What is the capacity of existing institutions to seize the opportunities that new information technologies offer in digital government? The second broad question reverses the causal arrow to ask: What is the potential impact of the Internet on digital government?
The first major research question focuses attention on the capacity of existing institutions to seize the opportunities that new information technologies offer. Moving beyond the notion of adaptive capacity, Fountain has shown in Building the Virtual State: Information Technology and Institutional Change (Brookings, 2001), a comprehensive study of cross-agency relationships, that government organizations often use, or enact, information technologies in unanticipated, suboptimal ways that reflect organizational arrangements more than system functionality. It is simply not the case that technologies will be adopted by existing organizations and implemented in an “optimal” fashion (even if one could define optimal). Effectively using new technologies often requires a substantial re-allocation of resources and fundamentally changing how business is done within an organization, which will threaten both the existing subcultures and interests. To understand how to seize the opportunities that a technology may offer requires developing a theory that allows one to view that technology through the lens of the organization.
To address the second research question, we are examining existing models that integrate information technology into governance. The web clearly offers the opportunity for dialogue among citizens, citizen input into government policy, and learning, both by the governed and the governing. The extent and type of impact, and how to maximize that impact, are far less clear and where we focus our research. We conduct this research at several levels. Lazer surveys existing policy-area websites, assessing content, usage, and impact. Specifically, he is examining the effect of a highly focused web-based policy forum on information acquisition in a well-defined policy network. There have been a few attempts to develop new models that take advantage of the web, and our research will develop a set of measures of the impact of new technology on participation and the quality of the discourse around rulemaking.
These research activities are vital to the center because the research results are likely to influence government executives in ways that can leverage federal funding for the technologies used in digital government by making clearer to researchers and practitioners the relationship between organizational environments, institutional environments, organizational practices, technical design, and the effectiveness of new technologies in government. As Fountain has written, many IT innovations are tacked on to existing organizational structures. The deeper structural changes made possible by the Internet and related technologies will become the chief changes in government in the decade to come. We study how government actors and those governed are changing their decision making processes and draw out the implications for technological innovation and implementation from these research findings and insights.
1. Institutional Analysis of Digital Government
The open standards and protocols on the Internet potentially allow all computers to be connected resulting in the remarkable connectivity, size, range, and richness of the web. Yet the development of the technical infrastructure for linking the information systems of the government is a necessary but not sufficient step for building digital government. Equally important is the development of institutional infrastructure to support coordinated practices, procedures, cultures, incentives, and a range of organizational, social and political rule systems that guide behavior and structure agencies. Agency autonomy, interagency competition, lack of interoperability, and functional/jurisdictional stovepipes have long hampered coordination, slowed communication, and diminished opportunities for joint policy problem solving in government
The development of the virtual state is not likely to resemble the growth of electronic commerce. Restructuring in firms and industries is qualitatively different from government reform. Ironically, the dramatic efficiency gains driving the development of e-commerce and industry change serve as disincentives for bureaucrats to use the Internet in government. Dramatic efficiency gains and cost savings in the economy are rewarded through profits, promotions, stock price increases, and market share. Similar gains in government are rewarded with budget cuts, staff reductions, loss of resources, and consolidation of programs. In sum, incentives and rewards in the institutions of government are the obverse of those in the market. During the first wave of digital government euphoria, when information and services are beginning to migrate to the web, efficiency gains and their political implications are not apparent. But during the next wave, when government-to-government channel development attempts increase, bureaucratic decision makers will rapidly experience the perversity of incentives for institutional transformation.
Important choices regarding the Internet and government will be forged during the next decade. By clarifying these challenges and their implications through scientific inquiry, not only is the knowledge base for scholars built, but also government practitioners might make more deliberate, informed decisions regarding alternative designs and uses of technology and institutional arrangements.
Fountain’s research incorporates networked computing into social science institutional perspectives on governance and organizations. The technology enactment analytical framework developed during a decade of previous research on government agencies and described in detail in Building the Virtual State (Brookings 2001) extends and refines institutional theory to encompass recent fundamental changes in information technologies. This intellectual territory remains virtually uncharted by institutional theorists in political science, organization theory, or economics.
Institutional analysis of digital government suggests the following research questions: As the growth of information systems use across government agencies and across government and the economy continues, research questions central to the institutional perspective in the social sciences increase in importance. These questions include: How are bureaucratic policymakers actually using networked computing? Are they negotiating new institutional arrangements as a consequence? To what extent and in what ways are they constrained by current institutional arrangements in government (particularly the budget process and oversight processes as well as structural inertia within and across agencies)? What extensions of institutional theory are necessary to take account of fundamental change in organizational communication, coordination and control? Partial answers to such questions would advance theory in this domain. Moreover, by clarifying and extending such social science concepts and relationships to account for contemporary information technology use, this scholarly study also contributes to practice. If organizations, institutions, and technologies carry different and conflicting features, and if the emergence of interagency networks forces a confrontation of these different elements what, exactly, can be said about the details of this new politics? Under what conditions are costs and disappointments likely to be high? What steps can be taken by various actors to improve the situation? If we are to exercise control over our future we must understand our core institutions, their constituent elements, and the mechanisms by which they change with far greater clarity than we now do.
The study of governance has been inextricably linked to institutions from antiquity. Robert Dahl observed: “That the character of a regime and the qualities of its people are somehow related has been a commonplace of political philosophy since the Greeks.” Aristotle argued that effective democratic institutions are intimately connected to the social and economic development of the demos. More recently, interest in institutions has encompassed a range of research programs. Robert Putnam has advanced understanding of the relationship among democratic institutions, politics and social capital. Other researchers have contributed to institutional thought by delineating both rational choice and boundedly rational organizational bases of politics. Robert Keohane and Joseph Nye have argued that realism in international relations fails to account for the effects of complex interdependence, international institutions and the importance of “soft power.” Other political scientists and sociologists have used an institutional lens to examine individual and organizational relationships and behavior in the policymaking process. Researchers in this stream have focused on policymaking as it is actually carried out by individual and organizational actors rather than on more formal models of legislative or interest group behavior. The network variant of this mode of inquiry is perhaps best exemplified in theories of the organizational state, which view policymaking from the perspective of constellations, or networks, of public, private and nonprofit organizations.
The second broad stream of research and theory from which an institutional perspective on digital government might draw is neo-institutional analysis based in organizational theory and sociology. Organizational analysts accounted for similarities in organizational forms and practices within organizational fields not as the result of rational choice but more often as the product of institutional isomorphism, processes by which organizations in a given field conform to normative influences, mimic others, or are coerced by powerful actors in their environments to adopt practices. Others have argued that strategic, purposeful action is embedded in ongoing social structures and social relations. Subsequent research has further developed the antecedents, characteristics, and outcomes of embedded network relationships, explored the mechanisms by which networks and embeddedness influence economic behavior, and explored the links between institutions and networks.
The third major stream of research informing the institutional lens on information technology considers the relationship between information technologies and organizations. Max Weber recognized clearly the rapid development of bureaucracy (the organizational form that that has come to structure the administrative state) in the nineteenth century as a response to the industrial revolution. He explained bureaucracy as a technology of control through its structuring of information into cases and channels, its strict reliance on impersonal relations, and inevitable tendency toward rationalization. Other researchers also have focused on the social and structural mechanisms by which individuals and organizations use new information technologies and on the effects of information technology on organizations and the design of decision making. These three streams of research and theory on institutions underlie the institutional and organizational study of web-based cross-agency relationships, or virtual government agencies.
There is an urgent need for research that correlates organizational and institutional conditions with success or failure of cross-agency partnerships that use the web. The Internet vastly reduces the coordination, control, and communication costs of horizontal relationships across jurisdictional and organizational boundaries. Yet the institutional arrangements and rules that structure the federal government work against horizontal, boundary spanning arrangements. Researchers have focused on the benefits of interorganizational networks, but in both the private and public sectors, the development of such networks has a high failure rate. In some cases, analysts have examined the economic incentives for network development. These incentives include resource pooling, cost sharing, and risk sharing. Other researchers have focused on the benefits of social capital, or the productive capacity of social collaboration and cohesion. While more difficult to quantify, the benefits of cohesion appear to be different from traditional economic benefits of aggregation of resources.
The value proposition for government lies in being able to address problems and challenges through strategic networks that lie across jurisdictions. A wide range of policy problems is not amenable to solutions from single agencies. Therefore, significant economic and policy benefits can be gained from better understanding of how to structure conditions for success of interagency collaboration, particular using the Internet as a source of shared data and information, of communication, and integration.
Fountain’s current research study identifies Bush Administration cross-agency initiatives by examination of government websites and through extensive interviewing of government managers. From this group of more than 30 network initiatives, we will choose a group of about six cases for intensive examination. We will select cases to maximize variance across policy domains. For each case, the researchers will conduct a series of interviews with the key actors involved in each agency in the network as well as in the oversight bodies that structure the incentives for cross-jurisdictional relationships. We will also collect critical incidents from interviews using the critical incident method, a well-known technique in the social sciences for collecting qualitative data. Using thematic analysis of interview data we will develop categories and refine the technology enactment framework developed by Fountain in Building the Virtual State to illuminate the conditions that foster or impede effective development and maintenance of cross-agency networks.
The primary data collection methods are necessarily qualitative relying on in-depth interviews, field research. Data will be gathered through face-to-face interviews with government executives, managers, and specialists in agencies and in Congress. Of particular interest are the mental models and procedural practices used by the institutional overseers for cross-agency relationships in the Office of Management Budget, the General Accounting Office, and in congressional committees. Actors in these institutions structure the incentives that either facilitate or impede development of partnerships across agency boundaries through their rulemaking authority, the budget process, evaluation, and legislation. Data analysis involves the search for patterns that either support or disconfirm working hypotheses, relentless search for alternative explanations to interim findings. This qualitative approach focuses on the mental models of government actors that underlie their decision making and use of information technology.
2. Policy Networks as Informational and Deliberative Structures
David Lazer is leading a team of researchers in a study of the World Wide Web and its effects on deliberation and informational structures in policymaking. Laurence O’Toole has observed that public management “increasingly takes place in settings of networked actors …. Yet the standard writings to which most administrators turn for advice to improve performance devote relatively little attention to acting effectively in such situations.” Networked arrangements in government are prominent and likely to increase resulting in a “virtual state” structured through networked computing and institutional arrangements.
The U.S. governmental system is notable for its decentralization, the result of a political history dictated by concern that no particular part of the government gain overwhelming power. While “information networks” may not have been foremost in the framers’ minds, the result is a rather exceptional one from the perspective of information management: A set of parallel bureaucracies with managers from disparate locations dealing with essentially similar issues. A potentially negative result of this structural feature is the absence of intra-organizational mechanisms that might arise to promote information sharing and learning within and across agencies. On the other hand, a potentially positive result is the experimentation (and thus learning) that might result in a decentralized system.
A variety of informal and formal mechanisms have been developed to serve the informational needs of policy actors. In the U.S. system of decentralized government, most public managers dealing with policy issues are not within the same formal organization. Individual managers may seek out one another informally for information--through accidental meetings at conferences or occasional “cold calls.” Another mechanism by which decentralization is dealt with by professionals is membership in professional organizations. These professional organizations themselves typically use a number of standardized mechanisms by which information is shared, including: (1) newsletters or magazines where the organization processes information for its target community; and (2) sponsoring conferences and other forums for members to get and share information. In addition, the federal government sometimes takes the lead in promoting the spread of information.
As a general matter, the information flow within the institutions involved in public policy is not optimal because the flow of information is generally unidirectional, highly segregated, and not transparent. The institutional arrangements that have evolved for information sharing are optimized for a low-bandwidth world. This is symptomatic of a broader failure of government to seize the opportunities that information technologies offer to improve the connections within government and between government and citizen of digital government. Until recently, there was no cost effective way to share the experiences of all with all. Bureaucratic institutions have developed filtering mechanisms to select out a subset of information that they judge would be most useful to their constituents.
One of the urgent questions faced by government decision makers, and eminently suited for research, is how should our institutions adapt to a high bandwidth world, with networks that carry virtually limitless information from one actor to another. Schumpeter’s notion of “creative gales of destruction” surely describes the Internet and its potential to topple traditional networks and to challenge--if not supplant--them with new relationships made in cyberspace. As Pippa Norris, “It seems likely that the passive web page, where people get vertical access to ‘top down’ information, much as they would from conventional political leaflets, will gradually be superseded by more active formats allowing horizontal communication among networks of citizens and ‘bottom-up’ feedback into the political process.” In order to empirically examine the interplay between the Internet, institutions, trust, and policy networks Lazer will research, design, develop, and manage a web forum designed for a specific policy sub-community, those involved with DNA testing. In the web-based forum, the following information attributes should apply:
Information flow is unlimited: In principle, any experiences from an actor in one part of the system should flow easily to other actors to whom those experiences are relevant. Any actor should be able to deposit materials onto the system that can then be easily accessed by any other actor.
Information flow is bi-directional: It is necessarily bi-directional, where the role of the “hub” shifts from filter to organizer of material. Different actors can deposit information for all to use, other actors can comment on that material.
Information flow is “desegregated”: The hub can break down the clusters that naturally form in social systems—for example, by directly connecting individuals in distant parts of the network. Further, the online forums can be tools for connecting disparate views or for provoking a clash of ideas.
Information flow is transparent: Anyone with web access should be able to access the material and the forums within the website.
The new paradigm of digital policymkaing is potentially vastly superior as measured by the four criteria outlined above. While the Internet makes a new paradigm possible, it is not inevitable that government will move toward such a model for several reasons. As already suggested, institutions seize new opportunities slowly, and, in particular, are reluctant to step back from the filter function, which is a source of power. Thus, many organizations use the web to post materials that previously (and likely still are) mailed to their members. In addition, few organizations are candidates to be neutral arbiters of the complex policy challenges that lie across jurisdictions. Finally, individuals do not naturally seek out information sources that sometimes provide information at odds with their existing views.
During the first stage of the study Lazer will focus on the development of a web-based forum in the issue area of DNA and the criminal justice system that offers the specific functions listed in the figure below. An institution such as the Kennedy School is in a particularly good position to develop this policy forum online because: (1) it has no institutional interest in maintaining a control over the flow of information in this area; (2) it is clearly a neutral institution in these policy areas; and (3) it has an almost unique “power to convene”—to bring together disparate and conflicting points of view. Looking beyond the initial phase of the study, we plan to develop a series of interconnected web-based public policy forums as part of the digital government web platform.
This research study builds on an existing partnership between Lazer and the National Institute of Justice (NIJ) which resulted in an international conference and forthcoming publication of the conference papers and a compendium of state-level DNA database legislation. The Kennedy School and the NIJ have built substantial credibility and visibility in this challenging domain as well as considerable technological, intellectual, and social resources that will be exploited for the web forum and accompanying research on policy networks. There is special value to be added in the case of DNA databases and the national criminal justice policy network because this is an intrinsically multi-disciplinary endeavor, bringing together computer scientists, engineers, political scientists, economists, and sociologists with practitioners from government and the private sector. Harvard provides a neutral forum for discourse in this area.
The development of a web forum aimed at the policy communities connected to the issues around DNA and the criminal justice system is an intervention aimed at members of those communities. In particular, the hypothesized impact of the website is that it would have a substantial effect on the pattern of information acquisition of the members of the policy communities. The ideal research design would be to randomly select members of these communities, assign some to the group to be exposed to the website, and half as controls, who would not be allowed to access the website. This would allow an examination of the impact of the website by comparing the experimental and control groups at fixed points in time in the future. Such an experimental approach is neither practical nor possible because of leakage between experimental and control groups. Instead, a quasi-experimental approach would be used where longitudinal data on information acquisition would be collected in an ongoing fashion from members of the policy communities, starting before the implementation of the web forum. Actors would thus serve as their own controls—where information acquisition before the website would be compared to information acquisition after.
The primary limitation of this research design is the possibility (really, certainty) that there are trends in information acquisition strategies. This study adopts two mechanisms to ameliorate this research design issue. First, a “control” criminal justice issue area unrelated to DNA but still of interest to the particular policy community will be selected for comparison to DNA issues (e.g., community policing, strategies regarding domestic violence). The second mechanism is to use the data collected from the website to evaluate the website’s impact. This is particularly powerful since it will be possible to trace particular actors’ usage of the website and to link this information to a set of questionnaire data that we will gather. It should therefore be possible to include data from usage as intermediate process variables between the key independent variable (inception of website) and the key dependent variable (information acquisition). The survey will first be done before the inception of the website and repeated annually after its inception. This will provide us with longitudinal data on its use and effects. The surveys will include questions about each organization’s resources (such as staff and budget), attitudes and knowledge, and information acquisition (such as professional networks and associations).
Both the technological infrastructure and the study are the anticipated products. The first product is the development of an ongoing web-based policy forum for the issues around the use of DNA in the criminal justice system. The second product is development of a prototype for other web-based forums. The third set of products—in the form of scholarly articles and a book—will include our findings on issues of trust, analyses of how the web might be used to create social capital—i.e., the capacity of individuals to draw on the social system (or network) they belong to in order to become more productive—and our broader common interest in the use of the Internet in ways that affect decisionmaking and policymaking structures in government.