CHAPTER ONE

The New Governance and the Tools of Public Action: An

Introduction

Lester M. Salamon

In economic life the possibilities for rational social action, for

planning, for reform—in short, for solving problems—depend not
upon our choice among mythical grand alternatives but largely
upon choice among particular social techniques… techniques and
not ‘isms’ are the kernal of rational social action in the Western
world.1

—ROBERT DAHL AND CHARLES E. LINDBLOM, 1953

Far-reaching developments in the global economy have us revisiting

basic questions about government: what its role should be, what it

can and cannot do, and how best to do it.2

—WORLD BANK, 1997

I. INTRODUCTION: THE REVOLUTION THAT NO ONE

NOTICED
A fundamental rethinking is currently under way throughout

the world regarding how to cope with public problems.3

Stimulated by popular frustrations with the cost and
effectiveness of government programs and by a newfound faith
in liberal economic theories, serious questions are being raised
about the capabilities, and even the motivations, of public-
sector institutions. Long a staple of American political
discourse, such questioning has spread to other parts of the
world as well, unleashing an extraordinary torrent of reform.4
As a consequence, governments from the United States and
Canada to Malaysia and New Zealand are being challenged to
be reinvented, downsized, privatized, devolved, decentralized,
deregulated, delayered, subjected to performance tests, and
contracted out.
Underlying much of this reform surge is a set of theories
that portrays government agencies as tightly structured
hierarchies insulated from market forces and from effective
citizen pressure and therefore free to serve the personal and
institutional interests of bureaucrats instead.5 Even defenders
of government concede that we are saddled with the “wrong
kind of governments” at the present time, industrial-era
governments “with their sluggish, centralized bureaucracies,
their preoccupation with rules and regulations, and their

hierarchical chains of command….”6

Largely overlooked in these accounts, however, is the extent

to which the structure of modern government already embodies
many of the features that these reforms seek to implement. In
point of fact, a revolution has taken place in the “technology”
of public action over the last fifty years, both in the United
States and, increasingly, in other parts of the world.
The heart of this revolution has been a fundamental
transformation not just in the scope and scale of government
action, but in its basic forms. A massive proliferation has
occurred in the tools of public action, in the instruments or
means used to address public problems. Whereas earlier
government activity was largely restricted to the direct delivery
of goods or services by government bureaucrats, it now
embraces a dizzying array of loans, loan guarantees, grants,
contracts, social regulation, economic regulation, insurance, tax
expenditures, vouchers, and more.

What makes this development particularly significant is that

each of these tools has its own operating procedures, skill
requirements, and delivery mechanism, indeed its own
“political economy.” Therefore, each imparts its own “twist” to
the operation of the programs that embody it. Loan guarantees,
for example, rely on commercial banks to extend assisted credit
to qualified borrowers. In the process, commercial lending
officers become the implementing agents of government
lending programs. Since private bankers have their own
worldview, decision rules, and priorities, left to their own
devices they will likely produce programs that differ markedly
from those that would result from direct government lending,
not to mention outright government grants.

Perhaps most important, like loan guarantees, many of these

“newer” tools share a significant common feature: they are
highly indirect. They rely heavily on a wide assortment of
“third parties”—commercial banks, private hospitals, social
service agencies, industrial corporations, universities, day-care
centers, other levels of government, financiers, and construction
firms—to deliver publicly financed services and pursue publicly
authorized purposes. The upshot is an elaborate system of third-
party government in which crucial elements of public authority
are shared with a host of nongovernmental or other-
governmental actors, frequently in complex collaborative
systems that sometimes defy comprehension, let alone effective
management and control. In a sense, the “public administration
problem” has leaped beyond the borders of the public agency
and now embraces a wide assortment of “third parties” that are
intimately involved in the implementation, and often the
management, of the public’s business.

Take, for example, the system for delivery of publicly

financed mental health services in Tucson, Arizona. Funding for
such services comes from a variety of federal and state
government programs. However, no federal or state bureaucrat
ever comes in contact with any mentally ill person. Indeed, no
federal or state bureaucrat even comes in contact with any local
government official or private agency employee who actually
delivers services to the mentally ill. Rather, the entire system is
operated at two and three steps removed. The state of Arizona
not only contracts out the delivery of mental health services, it
also contracts out the contracting out of mental health services.
It does so through a “master contract” with a private, nonprofit
local mental health authority called ADAPT Inc. ADAPT, in
turn, handles all dealings with more than twenty other local
agencies that deliver mental health services in the Tucson area
with funds provided by state and federal programs.7 While this
may be an extreme case, the pattern it exemplifies has been a
central part of public-sector operations for well over a
generation now.

What is involved here, moreover, is not simply the

delegation of clearly defined ministerial duties to closely
regulated agents of the state. That is a long-standing feature of
government operations stretching back for generations. What is
distinctive about many of the newer tools of public action is
that they involve the sharing with third-party actors of a far
more basic governmental function: the exercise of discretion
over the use of public authority and the spending of public
funds. Thanks to the nature of many of these tools and the
sheer scale and complexity of current government operations, a
major share—in many cases the major share—of the discretion
over the operation of public programs routinely comes to rest
not with the responsible governmental agencies, but with the
third-party actors that actually carry the programs out.

This development has proceeded especially far in the United

States, where hostility to government has long been a staple of
political life, and where the expansion of governmental
programs consequently has had to proceed in a highly
circuitous way.8 Contracting arrangements invented to fight the
Revolutionary War and later elaborated to handle the far more
complex tasks of product development during World War II
were thus quickly expanded in the aftermath of that war to
fields as diverse as agriculture, health, space exploration, and
social services. Grants-in-aid, loan guarantees, social
regulations, insurance, and other indirect instruments have
expanded as well. As Donald Kettl has reminded us, “[E]very
major policy initiative launched by the federal government
since World War II—including Medicare and Medicaid,
environmental cleanup and restoration, antipoverty programs
and job training, interstate highways and sewage treatment
plants—has been managed through public-private
partnerships.”9

Reflecting this, a study of a cross section of U.S.

communities carried out by the present author in the early
1980s found that the majority of the government-financed
human services available at the local level was already being
delivered by private nonprofit and for-profit organizations as of
that date, and this was well before the advocates of
“privatization,” contracting out,” and “reinventing government”
had proposed it. In particular, as shown in Table 1-1,
government agencies delivered only 40 percent of these
publicly funded services, while private agencies—both
nonprofit and for-profit—delivered 60 percent.10
 Instead of the centralized hierarchical agencies delivering
standardized services that is caricatured in much of the current
reform literature and most of our political rhetoric, what exists
in most spheres of policy is a dense mosaic of policy tools,
many of them placing public agencies in complex,
interdependent relationships with a host of third-party partners.
Almost none of the federal government’s more than $300
billion annual involvement in the housing field, for example,
bears much resemblance to the classic picture of bureaucrats
providing services to citizens. Rather, nearly $190 billion takes
the form of loan guarantees to underwrite mortgage credit
extended by private commercial banks; another $114 billion
takes the form of tax subsidies that flow to homeowners
through the income tax system; and more than $20 billion takes
the form of housing vouchers administered by semiautonomous
local housing authorities to finance housing provided by private
landlords (see Table 1-2).

TABLE 1-1 Share of Government-Funded Human Services

Delivered by Nonprofit, For-Profit, and
Government Agencies in Sixteen Communities,
1982 (Weighted Average)*
 More generally, as reflected in Table 1-3, the direct
provision of goods or services by government bureaucrats
accounts for only 5 percent of the activity of the U.S. federal
government. Even with income transfers, direct loans, and
interest payments counted as “direct government,” the direct
activities of the federal government amount to only 28 percent
of its activities. Far larger in scale are other instruments of
public action—contracting, grants-in-aid, vouchers, tax
expenditures, loan guarantees, government-sponsored
enterprises, insurance, and regulation, to name just a few.
Including just the $376 billion in net additions to outstanding
deposit insurance in 1999 and not the far larger amounts of
pension, crop, and disaster insurance, a rough estimate would
put the total monetary value of these activities in the
neighborhood of $2.5 trillion as of fiscal year 1999, two and a
half times higher than the roughly $1 trillion in direct activities
in which the federal government is engaged, and one and a half
times higher than the amounts recorded as outlays in the
federal budget that year. This highlights another interesting
feature of many of these more indirect tools: they often do not
show up on the government’s budget, which further helps to
explain their attractiveness.

TABLE 1-2 U.S. Federal Housing Programs by Type of

Tool, Fiscal Year 1999
TABLE 1-3 Scale of U.S. Federal Government Activity, by
Tool of Public Action, Fiscal Year 1999
 This reliance on third parties to deliver publicly funded
services is not an exclusively American phenomenon, however.
It has also been a classic—if largely overlooked—feature of the
European welfare states, at least outside Scandinavia. In the
Netherlands, for example, a fierce conflict between secular and
religious communities in the late nineteenth century over
control of public education was resolved early in the twentieth
century by a compromise under which the state was called on
to finance elementary and secondary education but to leave the
actual provision in the hands of private schools, many of them
religiously affiliated. As government was enlisted to assist in
the provision of health care, social services, and even
humanitarian assistance overseas, this same model was
replicated in these other spheres, producing a widespread
pattern known as “pillarization” under which state resources
are used to finance services delivered by private institutions
organized along religious and, later, ideological lines.11 A
similar phenomenon is also apparent in Germany, where the
Catholic doctrine of “subsidiarity” has been enshrined in basic
law, obliging the state to turn first to the “free welfare
associations” to address social needs before enlisting state

institutions.12 Belgium, Ireland, Israel, and other nations also

exhibit a similar pattern. Even France, long known for its
centralized governmental structure and highly developed state
welfare provision, dramatically increased its reliance on
government contracts with private nonprofit institutions during
the 1980s to implement a major decentralization of social

welfare functions.13 The upshot is that many countries in

western Europe have nonprofit sectors quite a bit larger than
that in the United States, financed largely through grants and
contracts from the state, as reflected in Figure 1-1 below.14
II. THE NEED FOR A NEW PARADIGM

The proliferation of these new tools of public action has created

new opportunities to tailor public action to the nature of public
problems. It has also made it possible to enlist a wide
assortment of different actors—governmental as well as
nongovernmental—in meeting public needs. At the same time,
however, this development has vastly complicated the task of
public management. Instead of a single form of action, public
managers must master a host of different “technologies” of
public action, each with its own decision rules, rhythms, agents,
and challenges. Policymakers must likewise weigh a far more
elaborate set of considerations in deciding not just whether, but
also how, to act, and then how to achieve some accountability
for the results. And the public at large must somehow find ways
to make sense of the disparate actions that are then taken on
their behalf by complex networks of public and private actors.
One of the central conclusions of the new field of
“implementation studies” that emerged during the 1970s, in
fact, was that the convoluted structure of many public programs
was the source of many of the problems causing public

programs to fall short of their promise.15

Regrettably, however, existing concepts of public
administration and public policy offer little help in coming to
terms with these dilemmas. Traditional public administration
remains preoccupied with the internal operations of public
agencies—their procedures for staff recruitment, budgeting, and
task accomplishment. Indeed, a cardinal tenet of the field has
been that the management of public affairs is best left to
neutral professionals organized in public agencies that are
arrayed in hierarchical fashion and therefore able to achieve
the needed specialization of functions so crucial to effective
operations and democratic control.16 Such concepts leave little
room for the proliferation of new forms of public action
featuring the wholesale surrender of key elements of
discretionary authority over the exercise of public authority and
the spending of public funds to a host of nongovernmental or
other-governmental actors. “Much of the time, when
‘government’ does something, it is the [government] employees
who really take action,” one recent text thus notes,
conveniently overlooking the fact that in the current era it is
mostly government’s third-party partners that take the action

instead.17
FIGURE 1-1 Nonprofit share of total employment, by
 country; 22 countries; 1999. Source: Lester
M. Salamon et al, Global Civil Society:
Dimension of the Nonprofit Sector
(Baltimore: Johns Hopkins Institute for
Policy Studies, 1999), p. 14.

Nor does the new field of policy analysis that recently has
gained prominence offer much help. The central preoccupation
of this field has been the application of sophisticated techniques
of microeconomics to the analysis of public problems. Of far
less concern has been the nitty-gritty of actual program
operations. Indeed, the implementation of public programs has
long been the “missing link” in the policy analysis world-
view.18

Even the “new public management” and the “reinventing

government” movement that it helped spawn have failed to
improve much on this record. To be sure, this line of thinking
has made the use of alternative instruments a major goal of

public sector reform.19 However, to justify this prescription, as

we have seen, reinventing enthusiasts have embraced a
caricature of current government operations that overlooks the
extent to which such instruments have already been adopted. In
the process, they downplay the immense difficulties that these
instruments entail and the strong possibility that the reforms
they are espousing may be the source, rather than the cure, for
the problems they are seeking to remedy.
What this suggests is that government does not need to be
“reinvented,” as the new public management has suggested.
That process is already well advanced. The great challenge now
is to find a way to comprehend, and to manage, the reinvented
government we have produced. For that, however, a new
approach is needed, one that acknowledges the existence and
likely persistence of “third-party government,” and that focuses
more coherently and explicitly on the distinctive challenges
that it poses.

Fortunately, some progress has been made in developing

such an approach. A half century ago, for example, Robert Dahl
and Charles Lindblom called attention to the rapid innovation
in techniques of social intervention already in evidence,
referring to it as “perhaps the greatest political revolution of
our times.”20 Frederick C. Mosher returned to this theme
during the early 1980s, emphasizing our failure to take
sufficient account of the extent to which the federal
government in the United States had changed its role from one
of doing to one of arranging.21 The present author at around
the same time proposed a wholly new focus for public
management training and research concentrating on the
distinctive tools or instruments through which the public sector
increasingly operates.22
Despite some useful progress in formulating such a “tools

framework,”23 and the further proliferation in the use of

diverse policy tools, however, most of our political rhetoric and
much of our public administration training remains dominated
by the image of the centralized bureaucratic state, as a recent
survey of public administration textbooks makes clear.24

The purpose of this book is to remedy this situation, to bring

the new tools of public action that are now in widespread use
to the center of public and professional attention. To do so, the
discussion builds on an earlier volume that first elaborated on
the concept of tools of government.25 Where that volume
focused on only six tools, however, this one extends the analysis
to many more. In addition, the present volume supplements the
discussion of individual tools with an analysis of some of the
overarching issues that the proliferation of tools of government
action raises.
 In the process, this book suggests a new approach to public
problem solving for the era of “third-party government” in
which we find ourselves. I call this approach “the new
governance” to underline its two defining features. The first of
these, signified by use of the term “governance” instead of
“government,” is an emphasis on what is perhaps the central
reality of public problem solving for the foreseeable future—
namely, its collaborative nature, its reliance on a wide array of
third parties in addition to government to address public

problems and pursue public purposes.26 Such an approach is

necessary, we will argue, because problems have become too
complex for government to handle on its own, because
disagreements exist about the proper ends of public action, and
because government increasingly lacks the authority to enforce
its will on other crucial actors without giving them a
meaningful seat at the table. The second feature, signified by the
use of the term “new,” is a recognition that these collaborative
approaches, while hardly novel, must now be approached in a
new, more coherent way, one that more explicitly
acknowledges the significant challenges that they pose as well
as the important opportunities they create.
The balance of this introduction outlines this approach in
more detail and introduces some of the basic concepts on which
it rests. To do so, the discussion falls into three major sections.
The first section introduces the major features that form the
heart of this “new governance” paradigm and shows how they
relate to existing conceptualizations in the field. The second
section then spells out some of the basic analytics of the
approach—what is meant by a “tool” of public action, how
tools can be assessed, and what dimensions of tools are
consequently most important. Finally, the third section
examines the implications that flow from this analysis and
explains the format of the rest of the book.

III. THE NEW GOVERNANCE PARADIGM

Like any new approach to a topic as old as public

administration, the “new governance” is hardly entirely novel.
Rather, it builds on a rich history of past thinking, changing
emphases, and incorporating new elements, but hardly
replacing all that has gone before. The result, however, is a new
synthesis, a new paradigm, that brings prevailing realities into
better focus and consequently makes more sense of some of the
central dynamics at work. In particular, five key concepts form
the core of this approach, as outlined in Table 1-4 below. In this
section we examine these five concepts and show how they
relate to existing approaches in the field.

From Agency and Program to Tool

At the heart of the new governance approach is a shift in the

“unit of analysis” in policy analysis and public administration
from the public agency or the individual public program to the
distinctive tools or instruments through which public purposes
are pursued. As we have seen, such instruments have
mushroomed in both number and scale in recent decades. A
central argument of the “new governance” is that this has
altered the nature of public management and the pattern of
public problem solving in rather fundamental ways, but ways
that are only partly acknowledged in existing theories and
approaches.

This focus on the tools or technologies of public action

differentiates the new governance both from classical public
administration and from the more recent “implementation”
school that emerged during the 1970s. For the former, the
central focus of public administration is on the operation of
governmental agencies. This reflects the origins of the public
administration field in the Progressive-era effort to legitimize
government action to cope with the increasingly apparent
shortcomings of the unfettered market system. As formulated by
Woodrow Wilson, Max Weber, Frederick Taylor, Luther Gulick,
and others, the classical theory posited a new type of
institution, the democratic public agency, that would overcome
the three major problems long associated with government
bureaucracy in the American mind—that is, excessive
administrative discretion, special-interest capture, and
inefficiency. This was to be achieved through three principal
devices: first, the restriction of executive agencies to
administration rather than policymaking; second, personnel
recruitment on the basis of technical competence rather than
political influence; and third, a set of “scientific” management
principles designed to ensure the efficient conduct of
administrative work.27 Although subsequent work has refined
and elaborated on these ideas, the basic principles have
remained largely intact, fixing on public administration
thinking a focus on the public agency as the basic unit of
analysis, a sharp distinction between the public and private
sectors, a separation between policy and administration, a
preference for clear lines of administrative responsibility and
control, and an emphasis on the skills of command and control.
 TABLE 1-4 The New Governance Paradigm

While these ideas have provided a workable framework for

the development of a relatively successful administrative
apparatus in the American context, however,28 they take as
given that the funding and provision of public services are
typically carried out by the same public entity. As a result, they
apply most clearly to only one of a range of possible forms that
public action can take (i.e., direct government). However, as we
have seen, this is no longer the dominant form of public action
at the present time.

This point became clear as early as the 1970s as efforts were

made to explain why the Great Society social programs of the
1960s were not living up to their promise. The answer, a new
school of implementation studies concluded, was not that the
classical theory was wrong, but that the American political
system was failing to supply the conditions necessary for it to
work.29 Instead of clear specification of program objectives,
sufficient authority to put programs into effect, and reasonable
attention to the management challenges that programs entailed,
studies of program implementation revealed that administrators
were often set adrift with only vague or conflicting guidance
about program purposes, insufficient authority to act, and little

attention to the administrative tasks that programs involved.30

Especially problematic was the highly indirect character of
many of the Great Society initiatives. The reason public
programs were failing, students of implementation therefore
concluded, was not that America adhered too closely to the
Progressives’ ideal and built too centralized an administrative
state, as “privatization” advocates now contend, but that it
departed too extensively from this ideal and created programs
that resembled Rube Goldberg cartoons instead, with multiple
actors linked together in often implausible decision sequences.

To remedy this, implementation theorists proposed to shift

the unit of analysis in policy work from the public agency to
the individual public program and to encourage clearer
specification of program objectives and greater attention to
program management. Far less clear, however, despite
numerous case studies, was what improved management really
entails and how this might vary systematically among the many
types of programs that exist.31

The “new governance,” by contrast, takes a significantly

different approach. Rather than seeing programs as sui generis,
the new governance finds commonalities flowing from the tools
of public action that they employ. It thus shifts the unit of
analysis from the individual program or agency to the
distinctive tools or technologies that programs embody.
Underlying this approach is the notion that the multitude of
different government programs really embody a more limited
number of basic tools or instruments of action that share
common features regardless of the field in which they are
deployed. Among other things, these tools define the set of
actors who will be part of the cast during the all-important
implementation process that follows program enactment, and
they determine the roles that these actors will play. Since these
different actors have their own perspectives, ethos, standard
operating procedures, skills, and incentives, by determining the
actors the choice of tool importantly influences the outcome of
the process. Thus, this focus builds on the insight of the
implementation studies that the division between policy and
administration assumed in the classical theory does not seem to
work in practice, and that the process of program design does
not end with legislative enactment but rather continues into the
implementation phase as well. Under these circumstances, it
makes sense to focus attention on the decisions that shape
which actors have significant roles in this stage of the process,
and this is precisely what the “tools focus” of the new
governance does. By shifting the focus from agencies or
programs to underlying tools, therefore, the new governance
provides a way to get a handle on the postenactment process
that the implementation literature identifies as crucially
important. Tool choices significantly structure this process and
therefore affect its results.

Because of this, however, tool choices are also not just

technical decisions. Rather, they are profoundly political: they
give some actors, and therefore some perspectives, an
advantage in determining how policies are carried out. This is
especially critical given the degree of discretion that the
implementation literature suggests is left to this stage of the
process. The choice of tool thus helps determine how this
discretion will be used and therefore which interests will be
most advantaged as a result. For this reason, the choice of tool
is often a central part of the political battle that shapes public
programs. What is at stake in these battles is not simply the
most efficient way to solve a particular public problem, but also
the relative influence that various affected interests will have in
shaping the program’s postenactment evolution. Indeed, it may
well be the case that the need to involve particular actors is
what importantly determines which tool is chosen.

Such choices are also importantly shaped by cultural norms

and ideological predispositions, and they, in turn, affect public
attitudes toward the state.32 A strong promarket bias underlies
tool choices in the United States, for example, whereas western
Europe is much more wary of the market and much more
favorably inclined toward the state. At the same time, such
cultural norms are hardly immutable. To the contrary, debates
over the appropriate techniques of social intervention—over
block grants vs. categorical grants, direct government vs.
contracting out, public enterprise vs. economic regulation—
forms the core of much of our political discourse.
If tool choices are fundamentally political choices, however,
they are also operational choices with significant implications
for the management of public affairs. Different tools involve
different management tasks and therefore require different
management knowledge and skills. The operation of a grant-in-
aid program is significantly different from the operation of a
regulatory program and this differs, in turn, from the operation
of a voucher. Whatever generic skills of public management
may exist, they must be supplemented by skills peculiar to the
various tools being employed if public programs are to be
effective. However, this requires a body of literature and a type
of training that is geared to the characteristics of the different
tools, which is precisely what the new governance seeks to
provide.

From Hierarchy to Network

In shifting the focus in public problem solving from agencies

and programs to generic tools, the new governance also shifts the
attention from hierarchic agencies to organizational networks.
The defining characteristic of many of the most widely used,
and most rapidly expanding, tools, as we have seen, is their
indirect character, their establishment of interdependencies
between public agencies and a host of third-party actors. As a
result, government gains important allies but loses the ability to
exert complete control over the operation of its own programs.
A variety of complex exchanges thus come into existence
between government agencies and a wide variety of public and
private institutions that are written into the operation of public
programs. Under these circumstances, the traditional concerns
of public administration with the internal operations of public
agencies—their personnel systems, budgetary procedures,
organizational structures, and institutional dynamics—have
become far less central to program success. At least as
important have become the internal dynamics and external
relationships of the host of third parties—local governments,
hospitals, universities, clinics, community development
corporations, industrial corporations, landlords, commercial
banks, and many more—that now also share with public
authorities the responsibility for public programs operations.

Not only does this broadening of the focus from public

agencies to “networks” of organizations differentiate the new
governance from traditional public administration, it also
differentiates it from the “privatization” and “reinventing
government” perspectives that have surfaced in recent years.
Both of these schools of thought acknowledge the
importance of indirect forms of government action. More than
that, they both advocate it, the former as a way to replace
government and the latter as a way to incentivize it.

In neither case, however, is the use of third parties viewed

as particularly problematic. Privatization theories, for example,
actually view reliance on the private sector to deliver public
services as more likely to serve public interests than reliance on
public agencies themselves. This is so, privatization advocates
argue, because the civil service protections designed to insulate
bureaucrats from political pressures insulate them as well from
the citizens they are supposed to serve and consequently free

them to pursue their self-interests instead.33 Under these

circumstances, “the key to effective government” becomes
“privatization”—reducing the size of the public sector, shifting
responsibilities to the private sector, and establishing “private
sector alternatives that are more attractive to the current

supporters of government programs.”34

The reinvention school and the “new public management”

of which it is a part take a different tack. For these theories,
contracting out and other forms of indirect government are less
ends in themselves than a means to improve internal agency
management by forcing public managers to compete.35
Reinventers thus have an incentive to downplay the extent to
which such indirect devices are already being used and to
minimize the difficulties to which they give rise. An internal
contradiction thus creeps into the new public management
prescription because managers are simultaneously encouraged
to take more responsibility for the results of their activity and
obliged to surrender significant shares of the authority for
achieving those results to third-party implementers.

The “new governance,” by contrast, shifts the focus of

attention much more explicitly from the internal workings of
public organizations to the networks of actors on which they
increasingly depend. While acknowledging the advantages such
networks can bring, however, it also acknowledges the
considerable challenges they pose. As such it builds on two
other bodies of theory: “principal-agent theory” and “network
theory.”

Principal-agent theory is part of a broader body of concepts

designed to explain the existence of organizations in a market
system.36 What is relevant for our purposes here is the insight
this theory provides into one of the central paradoxes that
arises in relationships between principals and agents in
contractual or other third-party arrangements of the sort that
third-party government entails. Despite the apparent influence
that the principals in such relationships wield by virtue of their
control of the purse strings, it turns out that the agents
frequently end up with the upper hand. This is so, principal-
agent theory explains, because the agents in such relationships
typically have more information than their principals about
what they are doing with the discretion that is inevitably left in
their hands. They therefore have significant opportunities to
“shirk” their duties and subject the principals to the “moral
hazard” of having to rely on agents whose competence and
diligence the principal cannot fully know. The only way for
principals to avoid this is to secure better information about
how the agents are performing, but this involves costs.
Therefore, every principal has to find an equilibrium between
the level of control it would like and the level it can afford.
Moreover, the more disparate the goals and characteristics of
the principal and the agent, the more information will be
needed and the more costly a given equilibrium is likely to be.
Under these circumstances, “who pays the piper” may not really
“call the tune” at all, at least not without considerable effort.

What network theory adds to this insight is the observation

that the principals in such relationships may have difficulty
getting their way even when the agents share their basic goals.
This body of theory was developed to explain the complexities
of policymaking in many modern democracies, where power is
splintered among numerous divergent groups. However, it can
also help explain the challenges of policy implementation as
well, especially where indirect tools are used. In such
situations, network theory argues, the standard relationship
among actors is one of interdependence. As a consequence, no
single actor, including the state, can enforce its will. This is
especially true, network theory emphasizes, because of four
crucial attributes that commonly characterize policy networks,
making the tasks of network management in general, and the
tasks of managing indirect tools in particular, especially

demanding:37

• First, their pluriformity—the fact that they engage a diverse

range of organizations and organizational types, many of
which have limited experience cooperating with each other
and limited knowledge of each other’s operating styles;

• Second, their self-referentiality—the fact that each actor has

its own interests and frame of reference and therefore
approaches the relationship with a different set of
 perspectives and incentives;

• Third, their asymmetric interdependencies—the fact that all

the actors in a network, including the state, are dependent
on each other but rarely in a fully symmetrical way. Even
when all the parties want the same thing, therefore, they
may still not be able to cooperate fully because they may
not all want it with the same urgency, in the same
sequence, or at the same time; and

• Finally, their dynamism—the fact that all of these features

change over time even as the network seeks to carry out its
mission.

Far from automatically sharing the same objectives, as the

privatization and reinventing paradigms tend to assume, the
actors brought into the operation of public programs through
indirect tools thus typically have goals, operating styles, skills,
worldviews, incentives, and priorities that, even with the best
of intentions, often differ widely from each other. As a
consequence, the task of securing concerted action becomes a
major administrative challenge. Under these circumstances, the
hopeful assumptions of the reinventing government school that
government can move easily from a “rowing” to a “steering”
role are far from assured.38
What the “new governance” and its “tools approach” add to
this network theory is a clearer understanding of the
commonalities of various network arrangements. In a sense,
tools significantly structure networks: they define the actors
that are centrally involved in particular types of programs and
the formal roles they will play. When policymakers choose a
loan guarantee, for example, they choose a network that
involves a structured interaction between a public agency and
the commercial banking system. When they select a grant-in-
aid, by contrast, they choose a different network that engages
state and local governments. By shifting the focus from
hierarchies to networks and specifying more precisely the kind
of network a program embodies, the “tools approach” of the
new governance thus can offer important clues about the kinds
of management challenges that particular programs will
confront.

From Public vs. Private to Public + Private

In moving the focus of public management and policy analysis

from the program and the agency to the tool and the network,
the new governance also brings a new perspective to the
relationship between government and the other sectors.
Traditional public management posits a tension between
government and the private sector, both for-profit and
nonprofit. The public sector is distinguished, in this view, by its
monopoly on the legitimate use of force, which it acquires by
virtue of its responsiveness to the democratic will of the people.
Public agencies thus are imbued with sovereignty, the power to
act on behalf of the public.39 Many of the central precepts of
classical public administration flow from this central premise
and are designed to ensure that the administrative officials so
empowered do in fact respond to the public’s will and not the
partial will of some private group. Without this clear
differentiation, accountability for the spending of public funds
and the exercise of public authority becomes impossible and the
public sphere polluted by the intrusion of private interests.
Keeping private interests and private organizations at arm’s
length thus becomes a central motivation of organizational
design.
This notion of a sharp divide between the public and private
sectors also figures prominently in the privatization theories.
Here, however, it is the protection of the private sphere from
the intrusion of the state that is the object of concern. In this
view, the expansion of the state inevitably comes at the expense
of the private sector, both for-profit and nonprofit. The best
way to preserve a healthy market system and private voluntary
sector therefore is to shrink the state and allow the private
sector to take up the slack.40

Many of the new tools of public action defy these precepts

rather fundamentally, however. Instead of a sharp division
between the public and private spheres, they blend the two
together. This is not to say that sectoral differences are blurred,
as is often suggested. A central precept of network theory, after
all, is that the participants in a network retain important
elements of their individuality. However, collaboration replaces
competition as the defining feature of sectoral relationships.
Rather than seeing such collaboration as an aberration or a
violation of appropriate administrative practice, moreover, the
new governance views it as a desirable byproduct of the
important complementarities that exist among the sectors,
complementarities that can be built upon to help solve public
problems.41 For example, the state enjoys access to resources
that are often critically needed by private, nonprofit groups. For
their part, nonprofit groups are often already actively involved
in fields that government is newly entering. By combining the
actions of the two, utilizing the state for what it does best—
raising resources and setting broad societal directions—while
using nonprofit organizations for what they do best—delivering
services at a human scale and innovating in new fields—
important public advantages thus can be gained.

Similar synergies exist, moreover, with the private business

sector.42 So long as due attention is given to the management
challenges they entail, cross-sectoral partnerships thus can yield
important dividends in terms of effective public problem
solving. Rather than viewing such interaction as a “fall from
grace” that undermines the purity of the respective sectors, the
“new governance” views it as a source of opportunity instead.

Front Command and Control to Negotiation and Persuasion

In emphasizing the shift from programs run by public agencies

to cooperative action orchestrated through complex networks,
the “new governance” also underlines the need for a new
approach to public management. In this also it differs from both
traditional public administration and the new privatization
theories.

Traditional public management, with its focus on the

operation of public agencies, emphasizes command and control
as the modus operandi of public programs. This assumes that
public action is carried out by hierarchically organized agencies
whose central spinal chord is the chain of command. Such
centralized control is, in fact, vital to the preservation of
democratic accountability. Much of traditional public
administration thus is preoccupied with clarifying lines of
control and centralizing authority.
The privatization school, by contrast, downplays the need
for administrative management altogether. Instead, it posits the
market as a superior mechanism for achieving coordination and
advancing public goals. Market competition, in this view,
replaces public decisionmaking and obviates the need for

administrative control.43
The “new governance” rejects both of these approaches and
suggests a third route for achieving public purposes in the
world of third-party government that now exists. Unlike the
privatization school, it emphasizes the continued need for
public management even when indirect tools are used. This is
so because private markets cannot be relied on to give
appropriate weight to public interests over private ones without
active public involvement. “Government’s relationships with
the private sector are not self-administering,” one expert on
privatization has thus noted; “they require, rather, aggressive

management by a strong, competent government.”44 Even the

World Bank, long known for its market-oriented economic
policies and endorsement of privatization, has had to
acknowledge recently that “Institutions Matter,” as the title of a

recent World Bank publication puts it.45 “An effective state,”

the World Bank noted in the 1997 edition of its influential
World Development Report, “is vital for the provision of the
goods and services—and the rules and institutions—that allow
markets to flourish and people to lead healthier, happier lives.
Without it, sustainable development, both economic and social,

is impossible.”46 In fact, even the process of privatization itself

has been found to require “strong political commitment and

effective public management.”47

While stressing the continued need for an active public role,

however, the new governance acknowledges that command and
control are not the appropriate administrative approach in the
world of network relationships that increasingly exists. Given
the pervasive interdependence that characterizes such
networks, no entity, including the state, is in a position to
enforce its will on the others over the long run. Under these
circumstances, negotiation and persuasion replace command and
control as the preferred management approach, not only in the

setting of policy but in carrying it out.48 Instead of issuing

orders, public managers must learn how to create incentives for
the outcomes they desire from actors over whom they have
only imperfect control. Indeed, negotiation is even necessary
over the goals that public action is to serve since part of the
reason that third parties are often cut into the operation of
public programs is that such clarity cannot be achieved at the
point of enactment.

All of this suggests a new body of administrative “doctrine”

that makes collaboration and negotiation legitimate
components of public administrative routine rather than
regrettable departures from expected practice. Reconciling such
an approach with longstanding prohibitions against excessive
administrative discretion will be no easy task, but interesting
examples of how this can be done are already apparent in such
approaches as negotiated regulation and cooperative
contracting, as subsequent chapters of this book will show.

From Management Skills to Enablement Skills

Finally, because of the shift in emphasis from command and
control to negotiation and persuasion, the world of third-party
government necessitates a significantly different skill set on the
part of public managers and those with whom they interact.
Both traditional public administration and the “new public
management” emphasize essentially management skills, the
skills required to manipulate large numbers of people arrayed
hierarchically in bureaucratic organizations. For traditional
public administration, these are essentially the control skills
summarized nicely by Luther Gulick in the classic
administrative acronym POSDCORB—Planning, Organizing,

Staffing, Directing, Coordinating, Reporting, and Budgeting.49

The new public management moves the emphasis considerably
from control to performance, but it remains preoccupied with
internal agency management and with the manager as the key
to success. Under this body of thought, the path to successful
public-sector performance is to introduce business management
practices into the public sector, freeing managers to manage but
subjecting them to increased competition and holding them
accountable for results.50

Unlike both traditional public administration and the new

public management, the “new governance” shifts the emphasis
from management skills and the control of large bureaucratic
organizations to enablement skills, the skills required to engage
partners arrayed horizontally in networks, to bring multiple
stakeholders together for a common end in a situation of
interdependence. Three rather different skills thus move into
the center of attention as a consequence of this shift:

Activation Skills
In the first place, the new governance requires activation skills,
the skills required to activate the networks of actors

increasingly required to address public problems.51 Many of

the new governance tools create opportunities for third parties
to take part in public problem solving but do not mandate that
these opportunities be taken. Public managers therefore must
perform a mobilization and activation role, marketing the new
opportunities and encouraging the potential partners to step
forward and play their roles. Thus, competent contractors must
be identified and encouraged to bid in purchase-of-service
programs; banks must be convinced to participate in loan
guarantee schemes; and private individuals and corporations
must be made aware of tax expenditures. In none of these cases
can participation be taken for granted. Rather, it must often be
coaxed and cajoled. One of the great challenges in purchase-of-
service contracting, for example, has been to ensure an
adequate supply of vendors willing to compete on the
government’s terms,52 and similar problems have confronted
loan guarantee programs as well. Those ultimately responsible
for program success therefore often find themselves in the
unaccustomed position not of withholding desired support but
rather of trying to mobilize appropriate partners to accept it.

Moreover, the task of activating networks for public

problem solving is not an exclusively governmental function.
Other actors can also often take the initiative. In some cases,
these are nonprofit organizations or community groups
mobilized by grassroots activists who bring the other

stakeholders to the table.53 Increasingly, private foundations

have played this role in the United States, either on their own
or in cooperation with corporate and community partners.
Rather than wait for government to act, in other words, private
institutions are taking the initiative instead. This proliferation
of a sense of responsibility for activating problem-solving
networks is, in fact, one of the more hopeful facets of the “new
governance.”
Orchestration Skills 17

In addition to activating networks, the new governance requires

managers who can then sustain them. This calls for orchestration
skills, the skills required of a symphony conductor. Essentially, a
conductor’s job is to get a group of skilled musicians to perform
a given work in sync and on cue so that the result is a piece of
music rather than a cacophony. Clearly, the conductor cannot
do this by playing all of the instruments. Rather, he or she must
tease the music out of the musicians, setting the tempo and
conveying an interpretation, but nevertheless remaining within
the bounds set by the physical capacities of the instruments
(and the musicians) not to mention the melody prescribed in
the score. The conductor thus is an enabler rather than a doer,
but his or her interpretation and skill can nevertheless
determine whether a given orchestra plays poorly or well.

Orchestration, therefore, does not mean command and

control, nor is the orchestrating role an exclusively
governmental one any more than is the activation one. Indeed,
in major systems acquisition projects, government contracts out
the orchestrating role to a general contractor who then
mobilizes subcontractors to produce the components of the
system. In recent years, this model has been applied as well to
human service contracting. In fact, defense contractors such as
Lockheed Martin have drawn on their experience in
orchestrating the production of complex weapons system to bid
successfully on contracts to orchestrate the complex networks
of day care, drug abuse counseling, mental health service, job-
search, health, job placement, and related service providers
required to move welfare recipients into jobs and keep track of
the results.54 Beyond this, however, other actors can also lift
the baton even without this kind of governmental imprimatur.
What is needed to be effective is not simply command of
resources—whether financial or legal—but also the intangibles
of knowledge, vision, persuasiveness, and community respect.

Modulation Skills
Finally, the new governance requires the sensitive modulation
of rewards and penalties in order to elicit the cooperative
behavior required from the interdependent players in a complex
tool network. Urban economic development specialists have
referred to this as enoughsmanship—the provision of just enough
subsidy to get private parties to make investments in run-down
areas they might avoid, but not so much that it produces
windfall profits for doing what the developers would have done
anyway. Inevitably, as we have seen, third-party government
leaves substantial discretion over the exercise of public
authority and the spending of public funds in the hands of a
variety of third parties over which public officials have at best
limited control. Under these circumstances, the central
challenge for public managers is to decide what combination of
incentives and penalties to bring to bear to achieve the
outcomes desired. Excessive use of authority can clearly
backfire if partners choose not to “play” or to disguise their
activities in ways that “principal-agent theory” predicts. On the
other hand, insufficient accountability can invite complete
disregard of public goals. Public managers in the era of the new
governance are consequently perennially confronted with the
dilemma of deciding how much authority or subsidy is
“enough” and how much is too much.

Eugene Bardach and Robert Kagan recognized this point

clearly in their classic analysis of the problem of regulatory
enforcement. Rather than the classic “tough cop,” Bardach and
Kagan suggest regulatory enforcement may actually be more
successful if it promotes the concept of the “good inspector,”
the inspector who understands when forbearance rather than
rigid enforcement best achieves regulatory compliance, and
who has the discretion to adjust regulatory enforcement
accordingly.55 Similar notions are also evident in endorsements
of new types of contracting stressing cooperation as opposed to
classic competitive bidding.56 Instead of narrowing the range
of administrative discretion left to the “street-level
bureaucrat,”57 in other words, the “new governance” calls for
broadening that discretion and equipping the public official
with the skills and understanding needed to exercise this
discretion in a way that advances program objectives.

The growing use of entire “suites” of tools in particular

programs only accentuates the need for this modulating,
enoughsmanship approach to program implementation and
enforcement. With rich medleys of instruments at their
command, public managers can assemble highly targeted blends
of incentives and disincentives specially tailored to the
circumstances at hand. While this opens opportunities for
abuse, it also creates the potential for truly effective
management of public programs. To be effective, however, this
approach requires site-level managers who can cope with the
discretion involved, and who have a well-developed feel for
what constitutes the appropriate mixture of penalties and
rewards required to get a given job done.
 Therefore, as with other facets of the new governance, the
enablement skills required will vary with the type of tool being
used. The task of securing the concurrence of industrial firms
with the operation of an air pollution control program is likely
to differ markedly from the task of enlisting financiers to take
advantage of a tax credit for low-income housing. This points
up again the importance of tool-specific knowledge to the
operation of the third-party arrangements that now exist.
However, it also underlines the fact that the new tools of public
action, far from reducing the demands on public management,
may increase them instead, necessitating more sophisticated
management skills, requiring greater exercise of discretion, and
calling for better information on performance and results. All of
this suggests not the withering away of public administration,
as privatization theories tend to assume, but its transformation
and refinement instead.

Summary

In short, the proliferation of tools of public action has

necessitated a new approach to public problem solving, a new
governance that recognizes both the collaborative character of
modern public action and the significant challenges that such
collaboration entails. Central to this new governance is a shift
in the basic paradigm guiding action on public problems.
Instead of focusing exclusively on public agencies or public
programs, the new governance moves the focus of attention to
the distinctive tools or technologies used to address public
problems. Underlying this shift is a recognition that different
tools have their own characteristic features that impart a
distinctive twist to the operation of public programs. Tools
importantly structure the postenactment process of policy
definition by specifying the network of actors that will play
important roles and the nature of the roles they will perform.
Under these circumstances, the whole character of public
management has to change. Instead of command and control, it
must emphasize negotiation and persuasion. In place of
management skills, enablement skills are increasingly required
instead. Far from simplifying the task of public problem
solving, the proliferation of tools has importantly complicated it
even while enlarging the range of options and the pool of
resources potentially brought to bear. All of this makes the
development of a systematic body of information about the
dynamics and characteristics of the different tools of public
action all the more urgent.
IV. BUILDING THE KNOWLEDGE BASE: BASIC ANALYTICS

The new governance thus calls attention to the new world of

public problem solving that has been ushered in by the
proliferation of tools of public action over the past half century
or more. Rather than resisting this trend, like the traditional
public administration, or uncritically celebrating it, like the
reinventing government school, however, the new governance
calls for the development of a systematic body of knowledge
that can help policymakers, public managers, and others
engaged in the increasingly collaborative business of public
problem solving take advantage of the special opportunities and
cope with the special challenges that these new tools entail. In
the process it directs our attention to the characteristic features
of the different tools and at the often complex networks of
interaction on which many of them depend.

But which features of the different tools are most important?

How can tools be analyzed and compared? Which facets are
likely to have the biggest effects? And which effects are most
important? Clearly, if the “new governance” and the “tools
framework” on which it rests are to be more than mere
metaphors, they must offer meaningful answers to these
questions. Therefore, it is necessary to turn from the rationale
for the “new governance” and the general features that
characterize it to a more detailed exploration of its analytical
core.

Definition and Classification: The Basic Building Blocks

Basic Definition
As a first step in this direction, it may be useful to specify more
precisely what is meant by a “tool” or “instrument” of public
action.

This is no simple task since tools have multiple features and

can be defined at any of a number of levels of abstraction. For
our purposes here, however, the most basic descriptive level
seems most appropriate. Therefore, as used here a tool, or
instrument, of public action can be defined as an identifiable
method through which collective action is structured to address a
public problem.58 Several features of this definition are
particularly notable:

BOX 1-1 A Tool of Public Action

A tool of public action is an identifiable method through which

collective action is structured to address a public problem.
• In the first place, each tool is assumed to have certain
common features that make it “identifiable.” This is not to
say that all tools of a particular type share all features. In
addition to their common, or defining, features, tools also
have design features that can vary from one embodiment of
the tool to another. For example, all grants-in-aid involve
payments from one level of government to either another
level of government or a private entity, but different grant
programs can vary in the level of specificity with which
they define eligible purposes, in the range of eligible
recipients, in how funds are distributed, and in many other
features.

• Second, tools “structure” action. What this means is that the

relationships that tools foster are not free-form or transient.
Rather, they are institutionalized. Thus, tools are
“institutions” in the sense emphasized by students of the
“new institutionalism” (i.e., they are regularized patterns of

interaction among individuals or organizations).59 They

define who is involved in the operation of public programs,
what their roles are, and how they relate to each other.
Thus, they importantly shape the set of considerations that
effectively come to bear in the all-important
 implementation phase of policy.

• Finally, the action that is structured by tools is “collective

action” aimed at responding to “public problems.” This is
different from saying that tools structure only government
action. Other entities are also often involved in the action
that is structured by the tools of public action.

Given this definition, it is possible to distinguish tools from

both programs and policies, two other concepts commonly used
to discuss policy action. Tools are more general than programs.
Programs thus embody tools, applying them to the
circumstances of a particular field or problem. A single tool
therefore can be used in many different programs in many
different fields. Typically, a program embodies a single tool,
although increasingly, as we will see below, programs are
coming to embody entire suites of tools. A central premise of
the tools approach is that particular tools impart similar
pressures and have similar operating requirements wherever
they happen to be applied.
If tools are typically more general than programs, they are
typically less general than policies. Policies are essentially
collections of programs operating in a similar field or aimed at
some general objective. The programs comprising a policy can
all utilize a single tool (e.g., multiple grants-in-aid) or multiple
tools. An interesting question that tools analysis raises is
whether some tools are more appropriate for some policy
objectives than others, an issue we return to below.
One other distinction worth making is that between internal
tools and external tools. Internal tools refer to the procedures
that governments use to handle their own internal operations.
Included here would be basic procedures for personnel
recruitment, human resource management, budgeting, and
procurement for the supplies that government needs to operate.
External tools, by contrast, are those used to affect society at
large, not just the government. The focus of this book, and of
the “new governance” approach, is on the latter type of tools,
those that seek to affect society and not just the internal
workings of government.

Tools as Bundles of Attributes

From what has been said, it should be clear that while the
concept of a tool of public action is relatively straightforward,
in reality tools are often quite complex. Any given tool is really
a “package” that contains a number of different elements. These
include:

• A type of good or activity (e.g., a cash or in-kind payment, a

restriction or prohibition, the provision of information);

• A delivery vehicle for this good or activity (e.g., through a

loan, an outright grant, a voucher, the direct provision of a
service, or the tax system);

• A delivery system, that is, a set of organizations that are

engaged in providing the good, service, or activity (e.g., a
government agency, a nonprofit organization, a local
government, a for-profit corporation); and

• A set of rules, whether formal or informal, defining the

relationships among the entities that comprise the delivery
system.

These multiple facets naturally complicate the task of sorting

and describing tools, as we will see more fully below. Tools can
be classified according to any of the different facets—the nature
of the good or service, the delivery vehicle, the nature of the
delivery system. This means that no single classification of tools
is possible. Classification schemes will differ depending on
which facet is used as the basis. Table 1-5 illustrates this point
by portraying how some of the most commonly used tools
compare to each other descriptively in terms of these four
features. Thus, for example, loan guarantees provide cash
delivered through a loan by commercial banks operating
according to a set of rules that stipulate the conditions under
which the government will reimburse the bank if the loan
becomes uncollectable. By contrast, direct loans provide cash
through loans delivered by a government agency.

The Challenge of Classification

This multidimensionality of policy tools naturally complicates

the task of describing and sorting them. This is particularly true
in view of the fact that unlike tools in the physical world, such
as hammers, saws, and screwdrivers, the tools of public action
rarely appear in pure form. Rather, they come bundled in
particular programs, many of which combine more than one
tool, and all of which bring different approaches to the design
issues that each program must address. Beyond this, there is
occasionally ambiguity about which features of a tool are truly
the defining features and which are the design features that can
vary with particular manifestations. For example, some
observers treat “block grants,” a form of grant-in-aid that
defines eligible purposes fairly broadly, as a separate tool from
“categorical grants,” which define eligible activities more
narrowly. Other observers, however, consider this distinction
inconsequential.60

Coupled with the considerable ingenuity that has

characterized the design of public action in recent years, the
multidimensionality of individual tools has made it difficult to
reach consensus even on the number of tools that exist. Thus,
Savas identified ten different arrangements that can be used
just for the provision of public services, the U.S. Office of
Management and Budget’s Catalog of Federal Assistance
identifies sixteen distinct tools, Osborne and Gaebler recorded
thirty-six, and E. S. Kirschen of the Netherlands identified no

fewer than sixty-three.61

TABLE 1-5 Common Tools of Public Action: Defining

Features
 Complicating matters further is the fact that tools are often
mislabeled, sometimes deliberately. For example, President
Roosevelt insisted on including a symbolic employee
contribution in the Social Security program so that this program
could be characterized as “insurance,” which was easier to sell
politically, even though it lacks most of the defining features of
insurance (current recipients receive their benefits from current
wage earners not from their prior contributions to the trust
fund). This mislabeling, whether deliberate or inadvertent, can
play havoc with efforts to characterize tools and analyze their
consequences.
 All of this makes it difficult to reach clear consensus about
the types of tools that exist. Several different classifications are
available in the literature, but each uses a slightly different tool
dimension as the basis for its grouping. Thus, Hood, in one of
the earliest schemes, sorted tools in terms of two major
dimensions: (1) the role of government for which they are used
(i.e., detecting vs. effecting); and (2) the governmental resource
they enlist (i.e., nodality, treasure, authority, or
organization).62 McDonnell and Elmore focused instead on the
strategy of intervention that government uses, producing a
fourfold division of tools into (1) mandates, (2) inducements,
(3) capacity building, and (4) system changing.63 Schneider
and Ingram elaborated on this with a classification that focuses
on the behaviors that programs seek to modify, leading to a
fivefold distinction among: (1) authority tools, (2) incentive
tools, (3) capacity tools, (4) symbolic or hortatory tools, and (5)
learning tools.64 Finally, Evert Vedung returned recently to a
scheme first developed by F. C. J. van der Doelen and identified
three classes of tools—carrots, sticks, and sermons—based on
the extent of force that each involves.65 Given this diversity,
some analysts have begun to question whether the concept of a
policy tool is rigorous enough to support any serious
analysis.66
Our approach, by contrast, is to recognize this diversity not
as a drawback of the tools approach, but as a strength. The fact
is that tools have multiple dimensions in terms of which they
can be compared and contrasted, and particular tools may be
alike along some dimensions and different along others. This
means that multiple classifications of tools are entirely
appropriate since different classifications will highlight
different facets. Thus, tools can be sorted in a two-step process:
first, basic descriptive features can be used to define different
tools; and second, various dimensions can then be identified in
terms of which various tools so defined can be grouped together
for analytical purposes.

But which dimensions are the most appropriate to use?

Since the tools approach argues that various tool dimensions
have significant consequences for how programs operate and
what results they produce, the answer to this question depends,
first, on which outcomes are of particular interest to us; and
second, on which tool dimensions our theories suggest might
affect them. Our approach to sorting tools therefore must be to
focus on these two factors.
Evaluating Tools: The Criteria

So far as the first step in this process is concerned, the field of

policy analysis has identified three criteria in terms of which
public interventions are typically assessed: effectiveness,
efficiency, and equity. The policy implementation and political
science literature suggest two other criteria that also seem
highly germane: manageability and political legitimacy. Taken
together, this gives us five criteria in terms of which the
consequences of tools can be assessed. Let us look briefly at
each of these.

Effectiveness
Effectiveness is the most basic criterion for gauging the success
of public action. It essentially measures the extent to which an
activity achieves its intended objectives. Although
considerations of cost can enter into this judgment,
effectiveness judgments are typically made independent of
costs. Using this criterion, the most effective tool is the one that
most reliably allows action on a public problem to achieve its
intended purposes.

Gauging the effectiveness of public action is far from easy,

however. For one thing, as we have seen, program purposes are
often quite ambiguous, either because precise indicators are
technically difficult to locate or because conflicts exist about
what really is the principal purpose. Indeed, such ambiguity is
almost chronic in fragmented political systems like that in the
United States, where multiple perspectives have ample
opportunities to influence the definition of program objectives.
This makes the choice of tool all the more important because
ambiguity at the point of enactment pushes the specification of
program purposes into the implementation process, where the
choice of tool can have an even more decisive impact.

The effectiveness of different tools also varies with the

circumstances. Not just the nature of the tool, but also the
nature of the circumstances therefore must be considered when
making tool choices. One of the major tasks of the tools
approach, in fact, is to specify the circumstances under which
particular tools are likely to be most effective. The tool of
contracting has great advantages, for example, where a
competitive market exists for the goods and services that
government wants to buy. However, this is often not the case,
so that the adoption of the contracting tool in such
circumstances can lead to great disappointments. Since other
considerations are often involved in tool choices, the new
governance can hardly avoid such dilemmas. However, it can at
least clarify the risks and point up the tradeoffs involved.

Efficiency

Where effectiveness focuses exclusively on results, a second

criterion—efficiency—balances results against costs. The most
efficient tool may not be the most effective one. Rather, it is the
one that achieves the optimum balance between benefits and
costs.

The costs that are relevant to a judgment about the

efficiency of a tool are not only the ones that show up on the
ledger of the government that authorizes the program,
however. The costs imposed on nongovernmental institutions
are also relevant, and for some tools these are far more
immense. Regulation, for example, places heavy compliance
costs on private businesses that never show up in the balance
sheet of government. Indeed, with severe fiscal pressures on
governments, there is a strong incentive to utilize tools that
have precisely this effect. This suggests the need for a “double
balance sheet” to assess the efficiency of various tools, one
focused on the costs to government alone and one focused on
the costs to other social actors as well.
Equity

A third crucial criterion in terms of which the consequences of

tools can be judged is equity. The criterion of equity has two
different meanings, however. The first of these involves basic
fairness—the distribution of benefits and costs more or less
evenly among all those eligible. A tool that facilitates the
distribution of program benefits evenly across the country thus
can be considered equitable in this “fairness” sense.
However, equity also has a different connotation relating to
“redistribution,” to channeling benefits disproportionately to
those who lack them. Achieving such redistribution is, in fact,
one of the principal rationales for public action. In this view,
government exists in part to remedy past inequalities and
ensure equal opportunity and access to all. Students of policy
thus distinguish between distributive programs, which
essentially distribute benefits evenly among a class of
recipients; and redistributive programs, which tilt the benefits

toward the disadvantaged.67 Some tools might be more likely

to serve such redistributive goals than others.

Manageability

In addition to the classic economic criteria of effectiveness,

efficiency, and equity, recent research on program
implementation suggests the importance of manageability, or
“implementability,” as an additional criterion in terms of which
to assess tools. Implementability refers to the ease or difficulty
involved in operating programs. The more complex and
convoluted the tool, the more separate actors are involved, the
more difficult it is likely to be to manage. Some tools are more
cumbersome to operate than others. While they may promise
great efficiency and effectiveness in theory, they are unlikely to
deliver it in practice because of the managerial difficulties they
pose. It was for this reason that Jeffrey Pressman and Aaron
Wildavsky identified implementability as the “first rule” of
program design.68 Generally speaking, this presumably means
choosing simpler, more direct tools.

Legitimacy and Political Feasibility

Finally, tool choices can also affect the political feasibility and
perceived legitimacy of public action. They do this, in the first
instance, by helping to determine which actors, and hence
which interests, get to shape program implementation, and
therefore which are most likely to support or oppose program
passage. Clearly, no matter what the prospects for effectiveness,
a program that cannot win political support cannot make
headway.

Beyond this, tool choices can also affect broader public

perceptions of the legitimacy of public action. As we have seen,
some approaches are considered more legitimate than others in
particular national settings regardless of their technical
advantages.69 Quite apart from such national styles, the choice
of tool can affect the perceived legitimacy of public action in
other ways as well. For one thing, some tools may facilitate
accountability for the exercise of public authority or the
spending of public funds better than others, a matter of some
importance in a democratic society where such accountability is
highly valued. So, too, the choice of tool can affect the extent to
which the public can perceive a link between the taxes they pay
and the services they receive. The more this link is attenuated
or broken, the greater the degree of alienation between
government and citizens and the greater the risk to democratic
participation.70 Tool choices thus can affect the overall sense of
legitimacy that government enjoys in the eyes of citizens.

Key Tool Dimensions

Armed with this set of criteria, it is possible to identify more
precisely which tool dimensions are likely to be most
important, and therefore how best to classify tools for analytical
purposes. Rather than focus on a single dimension that can
work for all purposes, however, the discussion above suggests
the need for a range of dimensions in terms of which tools can
be compared and contrasted. Tools can differ from each other
along one dimension and be similar along others. Only in this
way will it be possible to clarify the full matrix of choices that
policymakers face and the significant tradeoffs that exist among
them.

More specifically, five key tool dimensions seem most likely

to have implications for the kinds of consequences identified
above. These are not, of course, the only tool dimensions that
might be important. Nevertheless, they usefully illustrate the
analytical power that the “new governance,” and its tools
framework, possess.

Degree of Coerciveness
Perhaps the most salient of these dimensions has to do with the
nature of the activity that a tool embodies, and particularly
with the degree of coercion that it utilizes.
 Essentially, this dimension measures the extent to which a
tool restricts individual or group behavior as opposed to merely
encouraging or discouraging it.

This coerciveness dimension is probably the most common

basis for classifying tools in the literature.71 Economists in
particular consider this dimension important since it essentially
measures the extent to which a tool involves a deviation from
reliance on the market as a mechanism to allocate resources
and settle social roles. Such deviations are commonly viewed by
economists as inappropriate except where “market

imperfections” make them imperative.72

BOX 1-2 Coercion

Coercion measures the extent to which a tool restricts

individual or group behavior as opposed to merely encouraging
or discouraging it.

The coerciveness of tools is also of concern to political

scientists. This is so because coercion has implications not only
for the operation of the market, but also for the operation of the
political system, and especially for the preservation of
democracy. Of particular concern is the degree of infringement
on individual liberty that a tool entails. In a political
democracy, all such infringements are viewed with skepticism
and are expected to be undertaken only with clear popular
authority. As we have seen, much of the classic theory of public
administration, with its stress on the distinction between
politics and administration, took shape in response to this
concern to root administrative authority clearly in democratic
decisionmaking. The more coercive the tool, the greater the
infringement on individual liberty, the greater the potential
threat to political legitimacy, and therefore the greater the
burden of proof on those advocating the program embodying it.

Although almost all government action involves at least

some degree of coercion, there are considerable differences
among tools in the extent to which they rely on it. This is
apparent in Table 1-6, which groups the various tools of public
action in terms of the degree of coercion they utilize. Thus:

• At the low end of the coerciveness scale are tort liability,

tax expenditures, and public information campaigns. All of
these essentially rely on the voluntary cooperation of
individuals and groups for their effects, although as
Chapter 7 shows even information tools can involve
considerable coercion if information crosses the border into
 indoctrination.

• In a “medium” category are a variety of tools that deliver

subsidies of various sorts. The least coercive of these are
vouchers, which deliver subsidies directly to consumers
and leave it to them to do (or not do) what the program is
seeking to encourage. Somewhat more restrictive are
grants-in-aid, loan guarantees, direct loans, and
contracting, which tend to exact more requirements in
return for the subsidies they offer. On the outer border of
this category are mandatory labeling and corrective fees
and charges, which impose potential burdens on those who
fail to comply. These fees are still in some sense voluntary,
however, since the citizen is still permitted to engage in the
penalized behavior but has to pay a fine or tax on it.

• Finally, in the “highly coercive” category are social and

economic regulations, both of which impose formal
limitations on activities considered undesirable.

Based on the implementation literature reviewed earlier, it

seems reasonable to hypothesize that, other things being equal,
the more coercive the tool, the more effective it is likely to be,
and the more likely to yield redistributive results, as shown in
Table 1-6. These consequences flow from the clearer authority
these tools give governments to act, the limited leeway they
allow private actors to deviate from specified program
purposes, and the limited costs that governments incur in
operating them since much of the burden is imposed on
external actors. Because of this, these tools are also more likely
to generate political support among those most eager to engage
government in a particular form of social action.
These features may help to explain why the consumer and
environmental movements of the 1970s in the United States
insisted on command-and-control regulatory arrangements,
even though most academic economists cautioned against the
use of this tool. After years of political struggle against often-
powerful entrenched interests, advocates of protection typically
wanted tools that provided the maximum certainty that the
goals they sought would actually be achieved. Less coercive
tools, even when backed by sophisticated economic theories,
were often not able to provide this assurance.
The problem, however, is that coercive instruments
purchase these advantages at a relatively high price, as Table 1-
6 also shows. For one thing, they often entail a loss of efficiency,
for society at large if not for government. This has been a
central theme of economic critiques of social regulation: that
the apparent efficiency this tool enjoys from the point of view
of government is misleading since it focuses exclusively on the
government’s costs, which are trivial, and overlooks the far
more substantial costs such regulations impose on the private
sector. Critics argue, in fact, that these social costs are likely be
higher than necessary under regulation because by replacing
market decisions with administrative ones regulation surrenders
the market’s efficiencies. The solution, economists like Charles
Schultze therefore have argued, is not improved regulatory
management but a change in the basic tool being used: in
particular, a shift to less coercive tools that utilize marketlike
incentives and thus make “public use of private interest.”73

TABLE 1-6 Policy Tools Grouped by Degree of

Coerciveness
 Coercive tools can also be more difficult to manage since
they impose on administrative agencies the difficult job of
keeping abreast of literally thousands of decisions made by
hundreds of private entities in widely disparate settings. As
Schultze has put it, under social regulation “[s]ocial
intervention becomes a race between the ingenuity of the
regulatee and the loophole closing of the regulator, with a
continuing expansion in the volume of regulations as the
outcome.”74

Finally, because they restrict human freedom, coercive tools

are presumptively suspect in liberal political regimes and
therefore are vulnerable to political attack. As the political
movements leading to the enactment of these tools subside, as
they frequently do, therefore, the agencies administering them
often find themselves face-to-face with hostile vested interests
determined to use the full panoply of legal protections available
to them to rein in public authority. To avoid being totally
hamstrung, agencies often find it prudent to reach some modus
vivendi with the affected interests. The result is the well-known
phenomenon of “agency capture” by those it is seeking to
control.75

Directness

While the degree of coerciveness is by far the most common

basis for differentiating policy tools in the literature, it is by no
means the only possible basis. To the contrary, the
implementation literature of the 1970s and 1980s points our
attention to another dimension that may be equally, or more,
important, as the discussion above has already suggested. This
dimension has to do with the nature of the delivery system that a
tool utilizes, and particularly its degree of directness.
Directness measures the extent to which the entity
authorizing, financing, or inaugurating a collective activity is
involved in carrying it out. Underlying this concept are two
crucial observations: first, that any effort to cope with a public
problem is really made up of a number of separate activities;
and second, that these different activities need not be carried
out by the same entity. Thus, for example, it is possible to
distinguish between the financing of a public service and its
delivery. Moreover, each of these can be handled either
publicly or privately. This creates a minimum of four possible
combinations, as shown in Table 1-7: (1) public finance and
public delivery; (2) public finance and private delivery; (3)
private finance and public delivery; and (4) private finance and
private delivery. The first of these, depicted in Cell A in Table
1-7, represents the stereotypical view of how government
operates: government raises revenues through taxes and uses
them to support the delivery of services to citizens by a
government agency. As we have seen, however, this turns out
not to be the most common pattern at all, certainly not at the
national level in the United States. For one thing, even when
the public sector is involved in both finance and service
delivery, it is often not the same level of government that
performs both functions. Rather, the federal government may
raise some or most of the revenues, but it then often shifts them
to state or local governments to finance the actual delivery of
the services. Thus, Cell A itself becomes subdivided into four
subcells. Alternatively, the public sector—either national or
local—can raise the revenues but then contract with the private
sector to deliver the services (Cell B). What is more, the private
entities involved can be either for-profit or nonprofit firms.
Finally, any of these delivery mechanisms can be connected to a
private system of finance. Thus, for example, a public agency
can charge a user fee for the services it provides, in which case
the finance is private but the delivery public (Cell C of Table 1-
7). Alternatively, special tax advantages can be provided for
private purchases of services such as day care (Cell D of Table
1-7). All of these are forms of public action in the sense that
they engage governmental authority, but each utilizes this
authority in a different way and for a different part of the
process.

TABLE 1-7 Patterns of Public Problem Diving

 Even this does not exhaust the range of combinations that is
possible, however, since raising revenues and providing services
are hardly the only actions that public problem solving can
involve. Some tools—such as regulations—do not involve
services or finances at all, but rather restrictions. In others, the
services themselves are financial (e.g., the provision of
mortgage finance for housing purchase). Imposing charges,
creating inducements, providing information, delivering
benefits—all of these as well can be used to promote public
purposes. As a result, an extraordinary range of possibilities
exists for combining public and private institutions in public
problem solving.

Given these possibilities, it should be clear that “directness”

is a matter of degree and that different tools can vary greatly in
the degree of directness they embody. Generally speaking, the
more the various functions involved in the operation of a public
activity are carried out by the same institution, the more direct
the tool. Thus, a direct tool is one in which authorization,
funding, and/or delivery are all carried out by essentially the
same governmental entity. Indirect tools parcel these various
functions out to various other parties—semiautonomous
agencies, other levels of government, community groups,
nonprofit organizations, commercial banks, hospitals, and
others. The more extensively functions are performed by “third
parties,” the more organizationally distinct and autonomous
these third parties are from the authorizing body, and the
greater the discretion the third parties enjoy in the conduct of
their functions, the more indirect the tool. Thus, for example,
tax expenditures are typically more indirect than contracts since
they leave more discretion in the hands of citizens; however,
grants are more indirect than tax expenditures because they
surrender authority to other sovereign units of government, and
these typically have greater powers to resist. All three of these
are more indirect than service provision by government
agencies, however.76

Table 1-8 illustrates this point by ranking tools in terms of

their relative degree of directness. Thus, at the low end of the
directness continuum are tort liability and grants, while at the
high end are direct government service provision, government
corporations, and information campaigns that governments
conduct themselves. In between are tax expenditures (which
leave considerable choice to recipients but nevertheless are
administered by the enacting government), contracts, and, in
the American context, federal social regulations that make
extensive use of state and local governments.

As the grouping of tools here suggests, there is some overlap

between the degree of coerciveness BOX 1-3 Directness and the
degree of directness of a tool. This is so because the more
coercive tools are difficult to implement through indirect
delivery systems. However, this overlap is far from complete.
For example, information campaigns, one of the least coercive
tools, are typically operated directly while social regulation, the
most coercive, often leave ample opportunity for involvement
by lower levels of government. Clearly, these two tool
dimensions tap different facets of tool operations.

BOX 1-3 Directness

Directness measures the extent to which the entity authorizing,

financing, or inaugurating a public activity is involved in
carrying it out. A direct tool is one in which authorization,
funding, and execution are all carried out by essentially the
same entity.

In classical public administration, a distinction between

direct and indirect tools makes little sense since it is taken for
granted that a publicly authorized and funded program should
be carried out by a duly constituted, and staffed, public agency.
Yet, as we have seen, much of the growth of government action
over the last half century, especially in the United States, has
taken place through indirect tools—such as grants, loan
guarantees, tax expenditures, vouchers, and indirect regulation.
The result, as noted earlier, is an elaborate system of “third-
party government” that vests a substantial portion of the
discretionary authority over the spending of public funds and
the operation of public programs in the hands of a variety of
third-party partners. Indeed, many tools that operate directly in
other countries take a more indirect form in the American
context. Thus, for example, many European countries rely on
public enterprise to handle the natural monopolies that often
exist in public utility industries (e.g., electricity, telephones),
whereas the United States tends to leave these businesses in
private hands and subject them to economic regulation. The
United States has also used more indirect approaches in its
social regulatory programs in such areas as the environment,
worker safety, and health. While establishing national
standards, these programs leave much of the responsibility for
implementation in the hands of state and local governments
and those being regulated.

TABLE 1-8 Policy Tools Grouped by Degree of Directness

 One reason for the popularity of third-party government
appears to be the political advantages that indirect tools enjoy.
In particular, indirect tools provide important opportunities to
cut affected interests into a “piece of the action” when
government programs threaten to infringe on their fields. The
more fragmented political power is in a country and the more
controversial the issue, therefore, the more likely it will be that
indirect devices are used. Thus, for example, in recent American
experience:

• By using a cost-based reimbursement “voucher” whose

 proceeds flowed to existing hospitals, it was possible to
defuse the medical community’s opposition to the creation
of the federal Medicare program in the 1960s;

• By relying on loan guarantees instead of direct loans, it was

possible to neutralize commercial bank opposition to
federal involvement in home mortgage lending in the
1930s; and

• By using grants and purchase-of-service contracts, it was

possible to enlist research universities to support the
expansion of federal involvement in scientific research, and
private nonprofit organizations to support the expansion of
federal involvement in social services for the poor.

Federal constitutional structures also contribute importantly

to the widespread use of indirect forms of action. For much of
American history, for example, the federal government’s
authority to act on a wide range of domestic policy issues has
been contested thanks to constitutional provisions limiting the
federal role and a political structure and system of
representation firmly anchored at the state and local levels.77
Use of indirect tools—particularly the grant-in-aid—has thus
often been a political and constitutional prerequisite for any
federal involvement. State and local officials have frequently
resisted federal involvement unless that involvement is
channeled through them, and a meaningful degree of discretion
is left to them in the definition of the policy substance. Interests
opposed to federal involvement have often used their influence
at the state and local level to insist on a significant state and
local role as a way to retain some degree of influence over the
implementation of policies with which they disagree. Use of
indirect tools thus becomes the basis for political compromise,
shifting the battle over the definition of policy from the
enactment stage to the implementation stage where state and
local officials, and the interests that are more powerful at the
state and local level, can play a more meaningful role. The use
of indirect tools for federal environmental and welfare policy is

probably attributable in substantial part to this factor.78

These political advantages of indirect tools are hardly

unique to the American context, however. Reliance on indirect
instruments of public action is increasingly common in other
countries as well, driven by historical traditions of
“subsidiarity,”79 by a growing diversification of social and
political power, by deepening doubts about the capabilities of
state action alone to cope with complex social and economic
problems, and by a resulting inability of governments to secure

sufficient authority to act on their own.80

The political advantages of indirect tools are not their only

benefits, however. At least three other benefits are often
claimed for them:

• First, indirect tools can inject a useful degree of competition

into the provision of public services, breaking the
monopoly of governmental agencies and thereby
potentially improving service quality and “customer
orientation.”

• Second, indirect tools can provide access to talents and

resources that are desperately needed to cope with complex
public problems, but that public agencies may not
command. These include technical talents (e.g., university
researchers, private service providers, loan officers) as well
as financial and physical resources (existing facilities,
charitable contributions). Indirect tools therefore can
extend the reach of public agencies, making it possible for
them to avoid costly start-up problems and maximizing the
energies that can be brought to bear on public problems.

• Finally, indirect tools offer a greater degree of flexibility,

 making it easier for government to experiment, to change
course when needed, and thus to remain responsive to new
needs. This is so because the authorizing government does
not have to create the entire administrative structure to
operate an initiative.

While indirect tools may have important political and

operational advantages, however, they also carry with them
offsetting liabilities. For one thing, as Table 1-8 also notes, they
can be far less effective and far more difficult to manage. This
certainly seems to be one of the central conclusions of the
implementation literature of the 1970s and 1980s. In his 1979
study of the implementation of three human service programs
in Massachusetts, for example, Stuart Chase found that the most
serious implementation problem was the presence “of some
player or players in the implementation process whom the
program manager does not control but whose cooperation or
assistance is required.”81 Jeffrey Pressman and Aaron
Wildavsky similarly found that even clear specification of goals
and general concurrence on desired outcomes are no guarantee
of success when multiple actors are involved in executing a
program. The sheer mechanics of securing agreement at each
stage of the process can still inject debilitating delays. No
wonder “more direct means for accomplishing … desired ends”

is their recommended “first rule in program design.”82

Experience with indirect tools of public action thus has
ironically provided new reason to value bureaucracy. “The costs
of bureaucracy—a preference for procedure over purpose or
seeking the lowest common denominator—may emerge in a
different light,” Pressman and Wildavsky thus note, “when they
are viewed as part of the price paid for predictability of
agreement over time among diverse participants.”83 By
internalizing transactions, minimizing the legalisms involved in
complex contractual negotiations with external actors, and
providing a more stable framework for bargaining, direct
government offers distinct advantages for accomplishing
complex tasks.84

These conclusions find considerable support, moreover, in

the new economic theories of organization. According to these
theories, the “principal-agent” problems that inevitably arise
within organizations are even more severe in cross-
organizational relationships. This is so because having all of the
factors of production in a single entity creates certain
advantages:85
 • It permits a more creative reward structure to induce agents
to pursue the principal’s objectives.

• It helps convey the expectation that all involved should

work to a common purpose.

• It may help diminish the losses associated with breakdowns

and delays in bargaining.

When multiple organizations are involved in a given task, the

chances increase that the interests and values of the principal
and the agents will diverge. The more dispersed the authority,
therefore, and the less the coincidence of interests and
perspectives between principals and agents, the greater the risk
of goal displacement and principalagent difficulties. Not just the
extent of indirectness but also the type of third-party partner a
tool engages thus can affect the extent to which public purposes
are achieved. Public-sector managers of human service
programs thus have traditionally shown a preference for
nonprofit contractors over for-profit ones when service
contracting has been employed on the grounds that nonprofits
are more likely to share the objectives of the public sector.
Where principals and agents lack a shared set of values or
world-views, the task of ensuring that the principal’s objectives
are being served grows more complex and more problematic.

Not only does the directness of tools have implications for

the overall effectiveness of programs, it also may have
particular implications for their ability to promote equity and
redistribution goals. This is especially the case where the partners
brought into the operation of a public program by a tool lack
incentives to achieve these equity goals. Yet, this is often the
case with private businesses. As management theorist Regina
Herzlinger has pointed out, “when resources are given to
providers who in turn have the discretion to allocate the goods
and services they produce … [the] providers will try to attract

consumers who will improve their measurable performance.”86

In the process, however, redistributive goals may be sacrificed
as producers engage in “creaming” to attract better-off clients.
While public agencies themselves are hardly immune from
these pressures, the risks appear greater with indirect tools.
Finally, while enjoying important immediate political
advantages, indirect tools also suffer from certain longer-term
political limitations. In particular, they weaken the perceived
link between citizens and government by channeling services
financed by public revenues to recipients through private
intermediaries or other levels of government. In the process, the
connection between the taxes citizens pay and the services they

receive can become dangerously attenuated.87

In short, despite their advantages, indirect tools are

especially difficult to manage. Far from easing the public
management problem, as is often supposed, they significantly
complicate it instead.

Automaticity
A third key dimension in terms of which policy tools can be
differentiated is the level of automaticity they embody.
Automaticity measures the extent to which a tool utilizes an
existing administrative structure for its operations rather than
creating its own special administrative apparatus.

Tools that utilize the market, for example, are highly

automatic. This would include corrective fees and charges or
the “tradable permit” system authorized by the 1990 Clean Air

Act Amendments.88 Vouchers are another example of a market-

based tool: by placing purchasing power in the hands of
program beneficiaries rather than the institutions that serve
them, vouchers equip these beneficiaries to make use of the
market rather than an administrative mechanism to select the
quantity and quality of services they will receive.

BOX 1-4 Automaticity

Automaticity measures the extent to which a tool utilizes an

existing administrative structure to produce its effect rather
than having to create its own special administrative apparatus.

The market is not the only existing system that can be

mobilized to carry out public purposes, however. Others
include the tax system, the private credit system, the court
system, and, to a lesser extent, the networks of local
governments and private, nonprofit agencies. Where these
existing systems are operational, important options exist for
structuring public interventions in ways that build on them
rather than having to establish separate administrative
structures. Therefore, a certain overlap exists between the
automaticity dimension and the directness dimension of tools.
However, not all automatic tools are indirect, and not all
indirect tools are automatic. For example, tax expenditures are
automatic but not wholly indirect, whereas contracting is
indirect but far from wholly automatic.

Table 1-9 below arranges various tools of public action in

terms of their reliance on automatic, nonadministered
processes. As this table shows, tools embodying fees and
charges, vouchers, or tax expenditures are relatively automatic,
as is the use of the existing tort law system to control
environmental damage or ensure workplace safety. By contrast,
social regulation, direct service programs, and government
information campaigns are at the low end of the automaticity
spectrum. In between are tools such as grants and contracting,
which have some automatic features but operate within
essentially administered systems.

Like the other tool dimensions we have examined, there is

reason to believe that the automaticity dimension has
significant implications for the performance of programs.
Indeed, in a 1972 book on the preconditions of public program
success, economist Robert Levine identifies this dimension as
the single most important determinant of public program
success. Programs are most likely to fail, Levine argued, when
they rely on “highly administered systems.” Instead, more
reliance should be placed on “marketlike and bargaining
systems that combine the workable features of decentralization,
self-administration, personal economic or political motivation,
and the gross application of public policy rather than systems
administered in detail by public officials to private clienteles
according to plans laid out in detail by public planners.”89

Economist Charles Schultze reached a similar conclusion in

his pioneering 1977 analysis of regulatory programs, criticizing
prevailing regulatory approaches as inherently inefficient
because they utilize command and control techniques rather
than relying on the automatic mechanisms of the market to
promote public objectives. Instead of prescribing what
antipollution devices polluters must install to clear the nation’s
rivers, for example, Schultze recommends making “public use of
private interest” by imposing effluent charges that would give
polluters an economic incentive to find the lowest-cost way to
meet environmental goals.90

TABLE 1-9 Policy Tools Grouped by Degree of

Automaticity
 Because they make use of existing mechanisms, such as the
market, automatic tools can also be expected to be more
manageable. In a sense, they reduce the amount of public
management that is necessary, substituting for it the control
systems already built into these existing systems—the market,
the tax system, the court system, or the private banking system.

For all their appeal, however, automatic tools have proved

in practice to fall significantly short of their promise. For one
thing, there is reason to question how effective they are. The
great advantage of automatic tools is that they make it possible
to enlist existing systems in the pursuit of new objectives.
However, this advantage is also the source of serious problems
since these systems typically have their own objectives and
dynamics. There is always a question, therefore, whether the
public objectives will redirect the existing system or the
existing system will co-opt, and distort, the public objectives.
After all, it is the failure of the existing systems that often
necessitates public involvement in the first place. For example,
the demands for environmental and safety regulation grew
directly out of disappointments with the effectiveness of tort
law and the court system to handle consumer and
environmental problems.

The effectiveness of automatic tools thus depends on

identifying the incentives that can turn existing systems to
desired public purposes. In practice this has proved to be more
difficult than often assumed. Economist Robert Levine, for
example, acknowledged “the hard barrier of our lack of
knowledge about what incentives work for officials of local

government. …”91 The incentive structures of for-profit

businesses are presumably easier to fathom, but even here
complications arise. Contracting, for example, is assumed to be
a fairly automatic tool because of its reliance on the private
market. However, this assumes that a competitive market
actually exists for the goods and services needed to address
public problems. In fact, however, this critical prerequisite is
often lacking in government contracting since government is
often in the position of purchasing goods and services that are
not generally available on the open market (e.g., military
aircraft) in markets where the number of suppliers is highly
constricted. Complicating matters further is the fact that the
desired outputs in public programs are often difficult to specify
and then to achieve (e.g., getting the most disadvantaged
welfare recipients into permanent jobs). As management
specialist Regina Herzlinger has noted, this makes it extremely
difficult to “structur[e] enforceable contracts with the private
sector” and to exercise the control function that such contracts
require.92

Similar problems have arisen with voucher programs. The

effectiveness of vouchers depends critically on the
responsiveness of markets to the kind of demand that voucher
recipients will make and on the ability of voucher recipients to
make wise decisions. Both of these are often problematic,
however, making vouchers potentially a source of windfall
profits for providers without making recipients significantly
better off.
 Perhaps because of these problems, automatic tools also
have political problems. On the one hand, their reliance on
existing structures lends them a certain political legitimacy.
However, because they enlist institutions and processes that
have somewhat different objectives, such tools can rarely
attract the enthusiastic support of those pushing for a policy
initiative. Environmental advocates thus have been reluctant to
embrace the concept of tradable pollution permits for fear that
this would legitimize the right to pollute and create “hot spots”
of heavy pollution in less desirable neighborhoods. Consumer
advocates have similarly resisted the idea of weighing the value
of a human life against the cost of protection in structuring
approaches to workplace or consumer safety. Therefore, despite
their claims to greater efficiency, automatic tools have often
lacked a political constituency.

Finally, experience with automatic tools has raised some

serious questions about how easy they are to manage.
Seemingly automatic tools turn out to be far more cumbersome
to administer than advocates assume. Tradable permitting
schemes, for example, still require the establishment of initial
threshold pollution levels, the estimation of pollution charges
that are consistent with prevailing technology and industry
incentive structures, the collection of detailed information on
actual pollution levels, and the maintenance and operation of a

market in pollution rights.93 As with the other tool dimensions,

therefore, this one too involves difficult tradeoffs in terms of
the criteria outlined earlier.

Visibility
The fourth tool dimension that seems likely to be important is
the degree of visibility a tool exhibits in the normal policy
review processes, particularly the budget process. Obviously,
this dimension is highly sensitive to the structure of these
processes. Thus, for example, countries that do not utilize a
capital budget, like the United States, tend to put direct lending
programs at a competitive disadvantage by requiring that the
full value of a loan show up on the operating budget as an
expenditure the year in which the loan is made.
Until changes were made in 1990, this gave a real
advantage to loan guarantee programs over direct lending
programs since the value of loan guarantees shows up on the
budget only if and when they go into default. Similarly, until
the 1970s in the United States no official record was kept of tax
expenditures, making them largely invisible in the annual
budget process.94 Not until the 1990s, moreover, were such tax
expenditures considered in the normal budget decisionmaking
process.

BOX 1-5 Visibility

Visibility measures the extent to which the resources devoted to

a tool show up in the normal government budgeting and policy
review processes.

While the visibility of tools may be affected by the

accounting practices in place, however, there are still some
general structural features of tools that affect their standing
along this dimension. Table 1-10 thus offers a tentative
grouping of tools in terms of their degree of visibility. As shown
there, insurance and regulatory tools are still relatively
invisible, whereas direct government, grants, contracts, and
vouchers tend to be more visible. Loan guarantees and tax
expenditures are examples of tools that were once largely
invisible in normal budget processes in the United States but
have become more visible in recent years as a result of
accounting changes designed to bring them into better view.
Similar changes have been under way for more than a decade
to increase the visibility of social regulatory programs by
requiring economic impact analyses before regulations go into
effect.

Visibility has perhaps its greatest impact in the political

realm. At a time of budgetary stringency, invisibility is a
tremendous political asset. Invisible tools therefore are the
easiest to pass. This may explain why regulation, loan
guarantees, and tax subsidies grew so massively in the 1970s
and 1980s. Although different in many respects, all of these
tools shared a low level of visibility in the normal budgetary
process. In the case of loan guarantees, for example, until
passage of new credit budgeting procedures in 1990, only the
projected losses from loan guarantee defaults were carried on
government budgets, while the face value of the contingent
liabilities were relegated to a special annex. A similar procedure
is used with insurance programs. In the case of regulation, only
the direct cost of the regulatory agency personnel show up in
the government’s budget, whereas the indirect costs imposed on
businesses and households are largely invisible. Finally, in the
case of tax expenditures, until the adoption of the Budget and
Accounting Act in 1974, these were largely invisible as well in
the budget process, and it was not until the early 1990s that
lawmakers were required to take explicit account of such tax
expenditures in making annual budget decisions.

This dimension may also explain why corrective fees and

charges have made rather limited headway as vehicles for
environmental control despite the advantages claimed for them
as efficient mechanisms of public action. Unlike tax
expenditures, which are essentially invisible, corrective taxes
and fees are highly visible and therefore harder to enact.
The very feature that makes invisible tools so attractive
politically, however, makes them problematic along other
dimensions. Most obviously, the less visible the tool, the more
difficult it is to hold accountable. This can have implications for
the efficiency of programs embodying this tool. One of the
central criticisms of regulatory programs, for example, is that
by keeping their true costs hidden, they impose burdens on the
economy that are far greater than are needed to accomplish
their purpose. Similarly, tax expenditures are sometimes
accused of delivering windfall gains to taxpayers who would
engage in a particular activity even in the absence of the
subsidy. Programs embodying tax subsidies therefore may be
highly inefficient, paying unnecessarily for behavior that would
have occurred anyway. This same concern applies to insurance
programs. Far from preventing activities that entail risk—such
as locating houses in flood plains—insurance programs may
inadvertently encourage them. However, their relative
invisibility keeps the inefficiencies of these tools from being
recognized and addressed.

TABLE 1-10 Policy Tools Grouped by Degree of Visibility

 Because of these accountability problems, those opposed to
public spending tend to resist the use of invisible tools. On the
other hand, those on the receiving end of public largesse
naturally prefer to have their benefits delivered in the least
visible form. What this means in practice is that the stronger
the constituency, the less visible the tool it is likely to be able to
use for any benefits it receives. This may explain why low-
income welfare recipients receive their benefits through highly
visible grants-in-aid, whereas middle-class homeowners receive
theirs through far less visible tax expenditures.
 The visibility of tools may also have implications for the
extent to which they are used to pursue equity goals because of
the legitimacy attached to equity goals in the political arena.
Therefore, the more visible the tool a program uses, the more
likely the program will be to serve redistributive goals.
Conversely, the more special subgroups of the population, such
as oil-well owners or large investors, are being targeted for
benefits, the more attractive it will be to use less visible tools.

V. FROM ANALYTICS TO ACTION: RESOLVING THE

PARADOX OF THIRD-PARTY GOVERNMENT

The four dimensions identified above hardly exhaust the bases

for classifying different tools of public action and analyzing
their effects. Further fruitful distinctions can be drawn, for
example, between tools that deliver their benefits in the form of

cash versus those that deliver them “in kind”95; and between
those that operate through producers and those that deliver

their benefits directly to consumers.96 What is more, at this

stage of research the relationships between tool dimensions and
tool consequences are more in the nature of plausible
hypotheses than proven facts.
Even with these caveats, however, it should be clear that the
“new governance” approach, and the tools framework on which
it rests, has considerable analytic power as a source of insights
into the challenges of public problem solving in the era of third-
party government. Perhaps most fundamentally, the discussion
has pointed up a critical paradox that seems to characterize
contemporary efforts to respond to public problems. That
paradox, very simply, is this: policymakers seem to be under
increasing political pressures to select those tools of public action
that are the most difficult to manage and the hardest to keep
focused on their public objectives.

More specifically, a variety of factors—the growing

fragmentation of political power, the increased complexity of
public problems, the recent skepticism of government, the
preoccupation with efficiency as the major criterion for public
action—have put a premium on tools that are indirect,
invisible, and automatic. Such tools have the advantage of
defusing political opposition to governmental action, recruiting
new talents and resources to the tasks of public problem
solving, and avoiding the enlargement of the public sector. At
the same time, however, they have the disadvantage of vastly
complicating the tasks of public management and risking the
subversion of public purposes. In a sense, we seem caught in a
vicious circle in which disappointment with public action yields
forms of such action that seem most likely to further disappoint.
Clearly, the future of collective efforts to respond to public
problems will remain gloomy unless this paradox can be
resolved.
For this to be possible, however, it will be necessary to
move beyond slogans and address the three critical challenges
associated with the rise of third-party government:

The Management Challenge

The first of these is the management challenge. Contrary to the

hopeful assumptions of some, third-party government poses
immense management challenge. Contrary to the hopeful
assumptions of some, third-party government poses immense
management challenges, perhaps far more immense than those
posed by traditional public administration. With power
dispersed and numerous semiautonomous entities involved in
the operation of public programs, even straightforward tasks
become difficult. Indirect tools require advanced planning of far
more operational details than is the case with more direct tools.
Matters that could be dealt with internally on an ad hoc basis in
direct government have to be settled in advance through legally
binding contracts under “third-party government.” Similarly,
incentives have to be devised sufficient to induce desired
behavior but not so substantial as to yield windfall gains;
concurrence has to be secured at numerous points in complex
decision chains; and disparate organizations have to be forged
into effective networks capable of integrated action. Each of
these tasks requires not only extensive programmatic
knowledge, but also considerable diplomatic skill as well as
detailed knowledge of the operational parameters of the
different tools and the internal dynamics of the entities that the
tool engages. Also necessary is a sophisticated appreciation of
the context in which the tool is being deployed and how this
compares to the conditions required for the tool to function
optimally.

The Accountability Challenge

Side-by-side with this management challenge is the

accountability challenge that third-party government poses. As
noted, many of the newer tools of public action vest substantial
discretionary authority in entities other than those with
ultimate responsibility for the results. What is more, these other
entities have their own autonomous sources of authority that
allow them to operate with considerable independence of the
authorizing body: they include sovereign state and local
governments, private commercial banks, independent nonprofit
organizations, profit-seeking companies, and universities,
hospitals with powerful governing boards. Each of these enters
its relationship with governmental authorities on its own terms,
with its own expectations, objectives, and bottom line. What is
more, as we have seen, the choice of the instrument that
structures these relationships is often dictated as much by
political considerations as by the appropriateness of the
instrument for the purpose at hand. Under these circumstances,
classical notions of democratic accountability may need to be
loosened and more pluralistic conceptions developed. However,
this will require extensive education of all involved, new
decisionmaking procedures, and new attitudes.

The Legitimacy Challenge

Finally, and perhaps most significantly, for all its political

appeal third-party government may ultimately pose even more
serious challenges to popular support of government than did
the bureaucratic model before it. Fundamentally, third-party
government threatens to fray the link between citizens and the
services they receive in return for the taxes they pay. It does so
by vesting much of the responsibility for delivering these
services in the hands of institutions other than those that voted
the programs and raised the revenues for them. Under these
circumstances, it is not surprising that citizens might begin to
wonder where their taxes are going and what they receive in
return.

VI. OBJECTIVES OF THIS BOOK

The purpose of this book is to address these challenges and thus

help to resolve the paradox that now confronts public problem
solving. To do so, it seeks to develop three bodies of knowledge
that are critical to the “new governance” that is now needed:

• First, tool knowledge, that is, knowledge about the operating

characteristics of the different tools, about the players they
engage, and about how they structure the play;

• Second, design knowledge, that is, knowledge about how to

match tools to the problems being addressed in light of the
objectives being sought and the political circumstances that
exist; and

• Third, operating knowledge, that is, knowledge about how

 best to operate the new instruments to achieve these
objectives in the most effective fashion.

Structure of the Presentation

To do this, this book is divided into three broad sections:

Overview
In the first place, this introduction and a subsequent concluding
chapter are designed to put the “new governance” approach
into perspective, to identify its central features, and to explain
how it relates to other approaches to public problem solving.

Tool Chapters

The heart of the book consists of a series of chapters focusing

on particular tools of public action now in widespread use, both
in the United States and around the world. Altogether, fifteen
such tools are examined in depth here. As noted in Table 1-11
below, this includes direct tools such as direct government
service provision, government corporations, direct loans,
economic regulation, and information campaigns; as well as
indirect tools such as grants, contracts, tax expenditures, loan
guarantees, insurance, social regulation, vouchers, fees and
charges, government-sponsored enterprises, and tort law.

TABLE 1-11 Tools of Government Action Covered in This

Book

For each such tool, the discussion below offers a detailed

analysis prepared by a leading authority and focusing on a
common set of topics. These topics include:

• The defining features of the tool, how the tool compares to

others in terms of the key tool dimensions identified above
(coerciveness, directness, automaticity, and visibility), and
 what major design features and resulting variants of the tool
exist;

• The extent and pattern of tool use, including recent trends,

both in the United States and elsewhere;

• The mechanics of tool operations, that is, the tasks that the
tool entails, the actors it engages, and the roles these actors
are typically called on to play;

• The dynamics of tool selection, including the circumstances

for which the tool is most appropriate and the political
considerations that affect whether it is likely to be selected;

• The major management challenges the tool poses and the way
they can be handled; and

• The overall advantages and disadvantages of the tool for

various purposes.

The result is a more comprehensive and thorough body of

information on the major tools of government action than has
heretofore been available, presented in a readable format, and
designed to be accessible to scholars and practitioners alike.

Crosscutting Chapters
In addition to the overview material and the individual tool
chapters, this book also includes a set of chapters examining the
crucial crosscutting issues that the proliferation of new tools of
public action and the growth of third-party government have
posed. These include:

• The special management challenges of indirect government;

• The problem of cost accounting in third-party arrangements,

where fixed costs have to be allocated among multiple
activities;

• The more general accountability challenge that indirect

government poses and the ways it can be addressed;

• The politics of tool choice;

• The impact of third-party government on democratic

governance and citizen attachment to the political system;
and

• The international experience with alternative tools of public

action.

Tools Workbooks

Finally, to supplement the discussion here, a series of

Workbooks has been prepared on the major tools. These
workbooks contain documentary materials that help illustrate
how particular tools operate. Each focuses on a particular
program embodying a tool and includes materials such as the
following:

• The basic authorizing language for the program

• Key facets of the legislative history surrounding the

program

• The regulations issued to implement the program

• Key program management documents (e.g., sample

Requests for Proposals, proposal rating sheets)

• Study questions to direct attention to key decisions that had

to be made in structuring the tool for use in the program

Access to this workbook material is available through the

Internet at http://www.jhu.edu/~ccss/toolsworkbooks.

VII. CONCLUSION

A new era of public problem solving has dawned in the United

States and many other parts of the world. Instead of relying
exclusively on government to solve public problems, a host of
other actors are being mobilized as well, sometimes on their
own initiative, but often in complex partnerships with the state.
In this new setting, traditional notions of public and private
responsibilities are being turned on their heads and traditional
conceptions of public administration rendered largely obsolete.

To cope with this new reality, a new paradigm, a new

conceptualization, is needed, one that acknowledges the
complex networks of interaction that now characterize our
efforts to deal with public problems, that appreciates the
strengths these networks can mobilize, but that also recognizes
the challenges they entail.
We have argued that what we have called the new
governance provides such a conceptualization. The new
governance focuses our attention on the wide array of tools
now being used to address public problems and on the diverse
collection of institutions being activated in the process. In doing
so, it alerts us to the increased substitution of complex networks
of organizations for the rigid hierarchies of old to solve public
problems and to the resulting need for enablement skills rather
than simple management skills to cope with the resulting
interdependencies. Far from simplifying the tasks of policy
management, the new governance thus emphasizes the
increased difficulties it now entails.
 The new governance thus is a realistic framework for public
decisionmaking even while it is an optimistic one. It celebrates
the proliferation of tools of public action and the resulting
activation of new partners for “public work,” while squarely
acknowledging the challenges this creates. More than that, it
offers concrete insights into the operational requirements that
these various tools impose. The result, we hope, will be a better
basis both for public understanding of the way the public sector
works and for improving the effectiveness with which we
address public problems. That, at any rate, is our goal.

NOTES

1. Robert A. Dahl and Charles E. Lindblom, Politics,

Economics, and Welfare: Planning and Politico-Economic Systems
Resolved into Basic Social Processes (New York: Harper and Row,
1953), 6, 16.
2. World Bank, World Development Report 1997: The State in
a Changing World (New York: Oxford University Press, 1997), 1.
3. World Bank, World Development Report 1997: The State in
a Changing World (1997), 1–3; Marcel Masse, “Economic,
Political, and Technological Pressures Shaping Public Sector
Reform,” Proceedings of the Canada-South East Asia Colloquium
on Transforming the Public Sector (Ottawa, Canada: Institute on
Governance, 1993); Donald Kettl, “The Global Revolution in
Public Management: Driving Themes and Missing Links,”
Journal of Policy Analysis and Management 16, no. 3 (1997):
446–462; Tim Plumptre, “Public Sector Reform: An
International Perspective,” Proceedings of the Canada-South East
Asia Colloquium: Transforming the Public Sector (Ottawa, Canada:
Institute on Governance, 1993); C. Pollitt, Managerialism and the
Public Service (Oxford: Basil Blackwell, 1990).
4. Shahid J. Burki and Guillermo E. Perry, Beyond the
Washington Consensus: Institutions Matter, World Bank Latin
American and Caribbean Studies Viewpoints (Washington, D.C.:
World Bank, 1998).
5. Included here is the “new institutionalism” associated
with the work of Oliver Williamson and the public choice
theories associated with the work of Gordon Tullock. See Oliver
Williamson, Markets and Hierarchies (New York: Free Press,
1975); Gordon Tullock, The Politics of Bureaucracy (Washington,
D.C.: Public Affairs Press, 1965). For excellent summaries of
these theories and their application to public bureaucracy see
Terry Moe, “The New Economics of Organization,” American
Journal of Political Science 28 (November 1984): 739–777; and
Gerald Garvey, Facing the Bureaucracy (San Francisco: Jossey-
Bass, 1993), 25–35.
6. David Osborne and Ted Gaebler, Reinventing Government:
How the Entrepreneurial Spirit Is Transforming the Public Sector
(Reading, MA: Addison-Wesley, 1992), 48.
7. H. Brinton Milward and Keith G. Provan, “Measuring
Network Structure,” Public Administration 76 (summer 1998):
387–407.
8. For an interesting analysis of the impact of America’s
“nonstate” tradition on the development of public
administration in the United States, see Richard J. Stillman,
Preface to Public Administration: A Search for Themes and
Direction, 2d ed. (Burke, VA: Chatelaine Press, 1999).
9. Donald Kettl, Sharing Power: Public Governance and Private
Markets (Washington, D.C.: The Brookings Institution, 1993), 4.
10. Lester M. Salamon, Partners in Public Service: Government-
Nonprofit Cooperation in the Modern Welfare State (Baltimore:
Johns Hopkins University Press, 1995), 88.
11. For a description of the Netherlands case, see Ralph
Kramer, Voluntary Agencies in the Welfare State (Berkeley:
University of California Press, 1981), 19–36.
12. For a comparison of the German and U.S. systems, see
Lester M. Salamon and Helmut K. Anheier, “The Third Route:
Government-Nonprofit Collaboration in Germany and the
United States,” Private Action and the Public Good, eds. Walter
W. Powell and Elisabeth S. Clemens (New Haven: Yale
University Press, 1998), 151–162.
13. Claire Ullman, “Partners in Reform: Nonprofit
Organizations and the Welfare State in France,” in Private Action
and the Public Good, eds. Powell and Clemens (1998), 163–176.
14. Lester M. Salamon et al., Global Civil Society: Dimensions
of the Nonprofit Sector (Baltimore: Johns Hopkins Institute for
Policy Studies, 1999), 14.
15. See, for example, Jeffrey L. Pressman and Aaron B.
Wildavsky, Implementation (Berkeley: University of California
Press, 1973); D. Kettl, Sharing Power (1993), 4–5.
16. For a useful summary of this conventional wisdom, see
Vincent Ostrom, The Intellectual Crisis of Public Administration
(University: University of Alabama Press, 1973).
17. Barton and Chappell, Public Administration: The Work of
Government (1985), quoted in Richard Stillman, Preface to Public
Administration (1999), 150.
18. Erwin Hargrove, “The Missing Link: The Implementation
Challenge in Policy Research,” working paper (Washington,
D.C.: Urban Institute Press, 1975). A considerable body of
“implementation” literature has emerged in more recent years,
but, as noted below, it has not systematically come to terms
with the array of new tools of public action either. See note 28
below.
19. See, for example, Mohan Kaul, “The New Public
Administration: Management Innovations in Government,”
Public Administration and Development 17, nos. 13–26 (1997);
Andrew Massey, Managing the Public Sector: A Comparative
Analysis of the United Kingdom and the United States (Aldershot,
U.K.: Edward Elgar, 1993); Osborne and Gaebler, Reinventing
Government (1992); Owen E. Hughes, “New Public
Management,” in International Encyclopedia of Public Policy and
Administration, ed. Jay M. Shafritz (Boulder: Westview Press,
1998), 1489–1490.
20. Robert Dahl and Charles E. Lindblom, Politics, Economics,
and Welfare (1953), 8.
21. Frederick C. Mosher, “The Changing Responsibilities and
Tactics of the Federal Government,” Public Administration
Review AO (November/December 1980): 541–548.
22. Lester M. Salamon, “The Rise of Third-Party
Government,” Washington Post (29 June 1980); and Lester M.
Salamon, “Rethinking Public Management,” Public Policy 29,
no. 1 (summer 1981): 255–575.
23. See, for example, Christopher C. Hood, The Tools of
Government (Chatham, NJ: Chatham House Publishers, 1983);
Stephen H. Linder and B. Guy Peters, “From Social Theory to
Policy Design,” Journal of Public Policy 4, no. 3 (1984): 237–
259; Lorraine M. McDonnell and Richard F. Elmore, “Getting
the Job Done: Alternative Policy Instruments,” Educational
Evaluation and Policy Analysis 9, no. 2 (summer 1987): 133–152;
Stephen H. Linder and B. Guy Peters, “Instruments of
Government: Perceptions and Contexts,” Journal of Public Policy
9, no. 1 (1989): 35–58; Donald Kettí, Government by Proxy:
(Mis?) Managing Federal Programs (Washington, D.C.:
Congressional Quarterly Press, 1988); Lester M. Salamon, ed.,
Beyond Privatization: The Tools of Government Action
(Washington, D.C.: Urban Institute Press, 1989); Anne
Schneider and Helen Introduction Ingram, “Behavioral
Assumptions of Policy Tools,” Journal of Politics 52, no. 2 (May
1990): 510–529; Marie-Louise Bemelmans-Videc, Ray C. Rist,
and Evert Vedung, eds., Carrots, Sticks and Sermons: Policy
Instruments and Their Evaluation (New Brunswick, NJ:
Transaction Publishers, 1998). Beyond this, some individual
tools have attracted considerable attention. For example, the
tool of grants-in-aid has been thoroughly examined in a series
of publications produced by the Advisory Committee on
Intergovernmental Relations during the 1970s and early 1980s
and in U.S. Office of Management and Budget, Managing Federal
Assistance in the 1980s (Washington, D.C.: 1983). “Regulation”
has been examined as a tool of public action in a number of
studies, including Charles Schultze, The Public Use of Private
Interest (Washington, D.C.: The Brookings Institution, 1977);
and Neil Gunningham, Henry Grabosky, and Darren Sinclair.
Smart Regulation: Designing Environmental Policy (Oxford: Oxford
University Press, 1998).
24. Beverly Cigler, “A Sampling of Introductory Public
Administration Texts,” Journal of Public Affairs Education 6, no.
1 (January 2000): 48, 51. For evidence that some headway is
being made, see David L. Weimer and Aidan R. Vining, Policy
Analysis: Concepts and Practice, 3d ed. (Upper Saddle River, N.J.:
Prentice-Hall, 1999), 196–252; and H. G. Frederickson, The
Spirit of Public Administration (San Francisco: Jossey-Bass,
1997), 4–11, 78–96.
25. Salamon, Beyond Privatization (1989).
26. I am indebted to George Frederickson for his suggestion
to use the term “governance” to depict what I earlier termed
the “tools approach” to public problem solving. Frederickson
uses the term “governance” to refer to a broader array of
phenomena than is intended by the term here—namely, the
processes of policy formation as well as implementation. The
central idea of multiple stakeholders involved in the task of
governing remains the same, however. As noted below, I have
added the term “new” to the term “governance” to suggest a
greater consciousness about the consequences of choices among
tools and accompanying actors. See George Frederickson, The
Spirit of Public Administration (1997), 78–96.
27. For a succinct summary of the classical theory see
Ostrom, The Intellectual Crisis in Public Administration (1989),
20–41; Garvey, Facing the Bureaucracy (1993), 18–23. Similar
concerns also lay behind the European development of
administrative theory, as reflected in the work of Max Weber;
Stillman, Preface to Public Administration (1999), 109–123.
28. For a vigorous defense of the American administrative
state, see Charles Goodsell, The Case for Bureaucracy: A Public
Administration Polemic, 3d ed. (Chatham, N.J.: Chatham House
Publishers, 1994). For a discussion of the continuing vitality of
“direct government” as a tool of government action, see
Chapter 2 by Christopher Leman.
29. This line of argument also was evident in the public
administration literature more generally. See, for example,
Ostrom, The Intellectual Crisis of American Public Administration
(1989).
30. See, for example, Pressman and Wildavsky,
Implementation (1973); Hargrove, The Missing Link (July 1975);
Donald S. Van Meter and Carl E. Van Horn, “The Policy
Implementation Process: A Conceptual Framework,”
Administration and Society 6, no. 1 (February 1975), 447–474;
Eugene Bardach, The Implementation Game: What Happens after a
Bill Becomes a Law (Cambridge, Mass.: MIT Press, 1977); Walter
Williams, The Implementation Perspective: A Guide for Managing
Social Service Delivery (Berkeley: University of California Press,
1980); Robert T. Nakamura and Frank Smallwood, The Politics
of Policy Implementation (New York: St. Martin’s Press, 1980);
Daniel Mazmanian and Paul Sabatier, Implementation and Public
Policy (Glencoe, III.: Scott, Foresman, 1983).
31. Implementation studies have generated numerous lists of
factors thought to influence program success but with limited
progress in cumulating these findings into a more generalizable
body of theory. One attempt to summarize this literature
identifies no fewer than seventeen such factors that have so far
been identified and need to be taken into account. Included
here are such factors as the clarity of the law, the adequacy of
the causal theory embodied in it, the multiplicity of decisions
points, the characteristics of the implementing agencies, the
presence of an implementation entrepreneur, and the adequacy
of external review. Mazmanian and Sabatier, Implementation and
Public Policy (1983). For a general discussion of the limited
success of implementation studies in generating testable theory,
see Helen Ingram, “Implementation: A Review and Suggested
Framework,” in Public Administration: The State of the Discipline,
eds. Naomi B. Lynn and Aaron Wildavsky (Chatham, N.J.:
Chatham House Publishers, 1990), 463; Calista, “Policy
Implementation,” in Encyclopedia of Policy Studies, ed. Stuart
Nagel (New York: Marcel Dekker, 1994), 118.
32. Michael Howlett, “Policy Instruments, Policy Styles, and
Policy Implementation: National Approaches to Theories of
Instrument Choice,” Policy Studies Journal 19, no. 2 (spring
1991): 1–21.
33. For a discussion of these theories, see Moe, “The New
Economics of Organization” (1984), 762–768; Garvey, Facing
the Bureaucracy (1993), 26–33.
 34. E. S. Savas, Privatization: The Key to Effective Government
(Chatham, N.J.: Chatham House Publishers, 1987); Stuart
Butler, Privatizing Federal Spending: A Strategy to Eliminate the
Deficit (New York: Universe Books, 1985).
35. S. J. Burki and G. E. Perry, Beyond the Washington
Consensus, 125. Osborne and Gaebler, Reinventing Government
(1998), 20. For a general discussion of the new public
management agenda of which the reinventing government
perspective is a part, see Kaul, “The New Public
Administration” (1997), 13–26; Larry Terry, “Administrative
Leadership, Neo-Managerialism, and the Public Management
Movement” 58, no. 3 (May/June 1998): 194–2000; Hughes,
“New Public Management,” (1998), 1490; Chrisopher Pollitt,
Managerialism and the Public Services: Cuts or Cultural Change in
the 1990s (London: Blackwell, 1993).
36. Moe, “New Economics of Organization” (1984), 749–
757; John W. Pratt and Richard J. Zeckhauser, “Principals and
Agents: An Overview,” in Principals and Agents: The Strategy of
Business, eds. John W. Pratt and Richard J. Zeckhauser
(Cambridge, Mass.: Harvard Business School Press, 1991
[1985]).
37. J. A. de Bruijn and E. F. ten Heuvelhof, “Instruments for
Network Management,” in Managing Complex Networks:
Strategies for the Public Sector, eds. Walter J. M. Kickert, Erik-
Hans Klijn, and Joop F. M. Koppenjan (London: Sage
Publications, 1997), 122–123.
38. W. J. M. Kickert, E.-H. Klijn, and J. F. M. Koppenjan,
“Introduction: A Management Perspective on Policy Networks,”
in Managing Complex Networks, eds. Walter J. M. Kickert, Erik-
Hans Klijn, and Joop F. M. Koppenjan (1997), 33.
39. Ronald Moe, “Exploring the Limits of Privatization,”
Public Administration Review (November/December 1987): 453–
460.
40. As Ronald Reagan put it in 1981, “We have let the state
take away the things that were once ours to do voluntarily.” For
further discussion of this perspective as it applies to the
nonprofit sector, see Lester M. Salamon and Alan J. Abramson,
“The Nonprofit Sector,” in The Reagan Experiment, eds. John L.
Palmer and Isabel V. Sawhill (Washington, D.C.: Urban Institute
Press, 1982), 223–224.
41. For a discussion of the theoretical basis for government-
nonprofit cooperation in these terms, see Lester M. Salamon,
Partners in Public Service (1995), 33–49.
42. See, for example, James Austen, The Collaboration
Challenge: How Nonprofits and Businesses Succeed through
Strategic Alliances (San Francisco: Jossey-Bass, 2000); Reynold
Levy, Give and Take: A Candid Account of Corporate Philanthropy
(Cambridge, Mass.: Harvard Business School Press, 1999).
43. See, for example, Schultze, The Public Use of Private
Interest (1977).
44. Kettl, Sharing Power (1993), 6. See also Robert W.
Bailey, “Uses and Misuses of Privatization,” in Prospects for
Privatization: Proceedings of the American Academy of Political
Science, ed. Steve Hanke, 36, no. 3 (1987), 150.
45. Burki and Perry, Beyond the Washington Consensus:
Institutions Matter (1998).
46. World Bank, World Development Report (Washington,
D.C.: World Bank Group, 1997), 1.
47. Dennis Rondinelli, “Privatization, Governance, and
Public Management: The Challenges Ahead,” Business and the
Contemporary World 10, no. 2 (1998), 167.
48. Negotiation and persuasion operate within
administrative agencies as well, of course. In the new
governance, however, they are clearly the dominant form of
management action.
49. Luther Gulick, “Notes on the Theory of Organization,” in
Papers on the Science of Administration, eds. Luther Gulick and
Lyndall Urwick (New York: Institute for Public Administration,
1937), 13.
50. See, for example, Hughes, “New Public Management”
(1998), 1489–1490; Christopher Pollitt, Managerialism and the
Public Service (1993), 6–10.
51. W. J. M. Kickert and J. F. M. Koppenjan, “Public
Management and Network Management: An Overview,” in
Managing Complex Networks, eds. Kickert et al. (1997), 50.
52. See, for example, Ruth Hoogland DeHoog, “Human
Service Contracting: Environmental, Behavioral, and
Organizational Conditions,” Administration and Society 16
(1985), 427–454.
53. For a discussion of such civic initiatives, see Carmen
Sirianni, Civic Innovation in America: Community Empowerment,
Public Policy, and the Movement for Civic Renewal (Berkeley:
University of California Press, 2001).
54. Edward Skloot, “Privatization, Competition, and the
Future of Human Services,” Unpublished paper prepared for
delivery at the Council on Foundations Conference, New
Orleans (21 April 1999).
55. Eugene Bardach and Robert A. Kagan, Going by the Book:
The Problem of Regulatory Unreasonableness: A Twentieth Century
Fund Report (Philadelphia: Temple University Press, 1982),
123–151. Peter May develops this point more fully in Chapter
5.
56. See, for example, Ruth Hoogland DeHoog, “Competition,
Negotiation, or Cooperation: Three Models for Service
Contracting,” Administration and Society 22, no. 3 (November
1990): 317–340.
57. Michael Lipsky, Street-Level Bureaucracy (New York:
Russell Sage Foundation, 1980).
58. This definition is quite similar to that suggested by Evert
Vedung, who defines public policy instruments as “the set of
techniques by which governmental authorities wield power in
attempting to ensure support and effect or prevent social
change.” Evert Vedung, “Policy Instruments: Typologies and
Theories,” in Carrots, Sticks, and Sermons, eds. Bemelmans-Videc
et al. (1998), 21.
59. This usage is close to that suggested by the “new
institutionalism,” particularly in economics. As economic
historian Douglas North puts it, institutions are “regularities in
repetitive interactions … customs and rules that provide a set of
incentives and disincentives for individuals.” Douglas North,
“The New Institutional Economics,” Journal of Institutional and
Theoretical Economics 142 (1986): 231. For a broader discussion
of the “new institutionalism,” see Walter W. Powell and Paul J.
DiMaggio, “Introduction,” in The New Institutionalism in
Organizational Analysis, eds. Walter W. Powell and Paul J.
DiMaggio (Chicago: University of Chicago Press, 1991), 1–40.
60. For an argument that the difference between “block
grants” and “categorical grants” is not sufficientiy great to
warrant treating them as different tools, see Paul E. Peterson,
Barry G. Rabe, and Kenneth K. Wong, When Federalism Works
(Washington, D.C.: The Brookings Institution, 1986), 21–23.
61. E. S. Savas, Privatization: The Key to Effective Government,
62; Osborne and Gaebler, Reinventing Government, 21; E. S.
Kirschen et al., Economic Policy in Our Time (Amsterdam: North
Holland Publishing, 1964); Gaebler and Osborne, Reinventing
Government (1992). For a general discussion of the difficulties
of comprehensive and authoritative lists of policy tools, see
Linder and Peters, “The Design of Instruments for Public Policy:
Groundwork for Empirical Research,” in Policy Theory and
Policy Evaluation, ed. Stuart Nagel (Westport, Conn.: Greenwood
Press, 1990), 103–119.
62. Christopher Hood, The Tools of Government (1983).
 63. McDonnell and Elmore, “Alternative Policy Instruments”
(1987), 12.
64. Schneider and Ingram, “Behavioral Assumptions of
Policy Tools” (1990), 514–521.
65. Evert Vedung, “Policy Instruments: Typologies and
Theories,” in Carrots, Sticks, and Sermons, eds. Bemelmans-Videc
et al. (1998), 21–52. In the earlier version of this scheme, F. C.
J. van der Doelen divided tools into three “families” based on
the type of intervention the public sector uses: (1) the legal
family, (2) the economic family, and (3) the communications
family. See F. C. J. van der Doelen, Instrumenten voor
energiebesparing (Enschede Netherlands: Universiteit Twente,
1993) cited in J. A. de Bruijn and E. F. ten Heuvelhof,
“Instruments for Network Management,” in Managing Complex
Networks, eds. Kickert et al. (1997). On the general popularity
of classifications using the degree of coercion as the central
criterion, see Linder and Peters, “The Design of Instruments for
Public Policy,” 114.
66. Hans DeBruijn and Hans A. M. Hufen, “The Traditional
Approach to Policy Instruments,” in Public Policy Instruments:
Evaluating the Tools of Public Administration, eds. B. Guy Peters
and Frans K. M. van Nispen (Cheltenham, U.K.: Edward Elgar,
1998).
67. Theodore Lowi, “American Business, Public Policy, Case
Studies, and Political Theory,” World Politics 16 (1964): 677–
715; Peterson, Rabe, and Wong, When Federalism Works, 15–20.
68. Pressman and Wildavsky, Implementation (1973), 143.
69. Howlett, “Policy Instruments, Policy Styles, and Policy
Implementation,” Policy Studies Journal 19, no. 2 (1991): 1–21.
70. Anne Larson Schneider and Helen Ingram, Policy Design
for Democracy (Lawrence: University Press of Kansas, 1997), 5–
7, 129–135; Stephen Rathgeb Smith and Michael Lipsky,
Nonprofits for Hire: The Welfare State in the Age of Contracting
(Cambridge, Mass.: Harvard University Press, 1993), 207–211.
71. Linder and Peters, “The Design of Instruments for
Policy,” 114; Vedung, “Policy Instruments: Typologies and
Theories” (1998), 35–36.
72. For a detailed discussion of the types of market failures
that serve as rationales for public intervention, see Weimar and
Viner, Policy Analysis (1999).
73. Charles Schultze, Public Use of Private Interest (1977). For
a fuller discussion of these market-oriented approaches, see
Chapters 5 and 8.
74. Schultze, Public Use of Private Interest (1977), 57.
 75. On the phenomenon of regulatory capture, see Merver
Bernstein, Regulating Business by Independent Commission
(Princeton: Princeton University Press, 1955).
76. Differences in degree of directness also exist within tool
categories. Thus, for example, block grants are more indirect
than categorical grants. Similarly, contracts for major military
systems are more indirect than contracts for the purchase of
easily specified, off-the-shelf products since more discretion has
to be left to the contractor in major systems acquisitions.
77. Daniel Elazar, American Federalism: The View from the
States (New York: Thomas Y. Crowell, 1973), 3.
78. On environmental policy, see Nancy Kubasek,
Environmental Law (Englewood Cliffs, N.J.: Prentice-Hall, 1990),
124–141. For a discussion of the role that Southern
conservative opposition played in the choice of a grant-in-aid
with substantial state discretion for the nation’s basic welfare
program, Aid to Dependent Children, see Lester M. Salamon,
Welfare: The Elusive Consensus (New York: Praeger, 1976).
79. Salamon and Anheier, “The Third Route: Government-
Nonprofit Collaboration in Germany and the United States”
(1998), 151–162.
80. See, for example, L. J. O’Toole, K. I. Hanf, and P. L.
Hupe, “Managing Implementation Processes in Networks,” in
Managing Complex Networks, eds. Kickert et al. (1998), 137.
81. Gordon Chase, “Implementing a Human Services
Program: How Hard Will It Be?” Public Policy 27 (fall 1979):
385–435.
82. Pressman and Wildavsky, Implementation (1973), 143.
83. Pressman and Wildavsky, Implementation (1973), 133.
84. Christopher Leman, “Direct Government: The Forgotten
Fundamental,” in Beyond Privatization: The Tools of Government
Action, ed. Lester M. Salamon (1989), 3–38. For an update of
this analysis, see Chapter 2.
85. John W. Pratt and Richard J. Zeckhauser, “Principals
and Agents: An Overview” (1991), 11. On the importance of
coincidence of “worldviews” in potentially minimizing
principal-agent problems, see William G. Ouchi, “Markets,
Bureaucracies, and Clans,” Administrative Science Quarterly 25
(March 1980): 129–140.
86. Regina Herzlinger, A Managerial Analysis of Federal
Income Redistribution Mechanisms: The Government as a Factory,
Insurance Company, and Bank (Cambridge, MA: Ballinger, 1979).
87. Smith and Lipsky, Nonprofits for Hire (1993), 207–211;
Schneider and Ingram, Policy Design for Democracy (1997), 5–7,
129–135. For further elaboration of this point, see Chapter 20.
88. Under the tradable permit system, a target is set for the
total amount of pollution permitted in an air quality district,
but companies are free to purchase rights to emit pollutants in
excess of their fair share from other businesses in the same
district. For a more detailed discussion of tradable pollution
permits and corrective fees and charges as tools of public
action, see Chapter 8.
89. Robert A. Levine, Public Planning: Failure and Redirection
(New York: Basic Books, 1972), 17, 23.
90. Schultze, The Public Use of Private Interest (1977).
91. Levine, Public Planning (1972), viii.
92. Herzlinger, Managerial Analysis, 111.
93. For an analysis of the administrative challenges entailed
in the application of market-based tools to environmental
protection, see National Academy of Public Administration, The
Environment Goes to Market: The Implementation of Economic
Incentives for Pollution Control (Washington, D.C.: National
Academy of Public Administration, July 1994).
94. Until very recently, in fact, tax expenditures did not
show up in the regular budget documents or figure prominently
in budget debates. This changed in the early 1990s when
budget agreements stipulated that both tax and spending
decisions had to be taken into explicit account when budgetary
decisions were made.
95. Most economists would argue, for example, that cash
benefits are more efficient than in-kind benefits since they
allow beneficiaries to utilize resources where they value them
the most. However, in-kind tools such as vouchers and loan
guarantees have substantial political advantages since they can
often mobilize producer interests in support of particular
programs and neutralize opponents who fear that recipients will
squander benefits on purposes other than those intended.
96. Consumer-side subsidies such as vouchers and tax
expenditures are thought to be more efficient because they
allow consumers to shop for the best combination of service
and cost. However, producer-side subsidies such as grants and
contracts retain the political edge because they are more likely
to stimulate producer-side political support.