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.
Research
Issues:
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 (see the conference website at www.ksg.harvard.edu/dna) and
forthcoming publication of the conference papers and a compendium of
state-level DNA database legislation (see www.ksg.harvard.edu/dnabook).
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.
Acknowledgment and Disclaimer - This material is based upon work supported by the National Science Foundation under grant numbers 0131923 and 0630239. Any opinions, findings, conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation (NSF).