From Social Theory to Policy Design

Author(s): Stephen H. Linder and B. Guy Peters

Source: Journal of Public Policy , Aug., 1984, Vol. 4, No. 3 (Aug., 1984), pp. 237-259
Published by: Cambridge University Press

Stable URL: [Link]

REFERENCES
Linked references are available on JSTOR for this article:
[Link]
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@[Link].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
[Link]

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access
to Journal of Public Policy

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
Jnl Publ. Pol., 4, 3, 237-59

From Social Theory to Policy Design

STEPHEN H. LINDER School of Public Health

University of Texas
and
B. GUY PETERS Department of Political Science
University of Pittsburgh

ABSTRACT

Little attention has been given in policy analysis to the creative process of
designing solutions to public policy problems. There are a number of
difficulties in applying macro-level theories - whether from economics,
sociology, philosophy or macro-systems theory - in the policy process.
Any macro-level theory will tend to provide inadequate guidance in one
or more of three aspects of policy-making: a model of causation, a model
for evaluating alternatives and outcomes, and a model of how interven-
tions operate. Our current knowledge about which policy strategies work
best under which conditions is at best rudimentary. Academic disciplin-
ary perspectives focus on a narrow repertoire of policy instruments. What
is required is a design focus which draws on instruments associated with a
range of disciplines and professions. A design perspective involves both a
systematic process for generating basic strategies and a framework for
comparing them. Such an approach will require at least the following
elements: (i) the characteristics of problems (scale, collectiveness,
certainty, predictability, independence); (2) characteristics of goals
(value-laden, operational, process of goal-setting); (3) characteristics of
instruments (suitability of different instruments).

i. Introduction

Governments have been in the business of making public policies for as

long as they have existed, but frequently, or perhaps usually, they were
making policy choices in areas about which they had very little real
knowledge. Even in the now well-established policy areas such as educa-
tion, crime, and international affairs there is little coherent theory for
explaining how government could intervene in the most efficient and
effective manner. This has resulted in policy-making by trial and error, or
by analogy with other policies and programs, or simply by intuition. In

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
238 Stephen H. Linder and B. Guy Peters

the literature on decision-making in organizations this form of decision-

making has been characterized as the 'garbage can' model, where all
inputs into the decision are thrown into the garbage can and the decision-
makers decide which bits to retrieve as and when it suits them (Cohen,
March and Olsen, I972). In such a decision-making environment the
results of any round of choices are scarcely predictable.
There are two major options for moving policy-making out of this
garbage can; both require some more explicit attention to theory. One
approach would be to employ social theory more explicitly in making
public policies, and particularly to employ macro-level social theory as a
guide for action. Policy analysis is, to some extent, applied social science
(MacRae, I976) and as such could apply the theoretical approaches of the
social sciences to solving social problems. The most familiar case of this is
the utilization of Keynesian economics as a guide for national economic
policy-making. It has been argued, for example, that what is needed for
policy-makers in the waning portion of the Twentieth Century is a
comprehensive theoretical approach to the problems which beset many
industrialized nations; without such an understanding at a very
fundamental level, policy-making is indeed muddling through
(Benjamin, I983; Goodin, I982). This call for a new Weber/Marx/
Keynes for the emerging post-industrial world is intellectually appealing
in many ways, as it would provide as (Benjamin points out) a par-
simonious means of comprehending the world as well as a means of
evaluating any proposed changes. As appealing as such an approach
would be, we find a number of potential problems with the acceptance of
such an approach as it might be applied through policy analysis, even
leaving aside the normative implications of the social engineering which
would be required by the approach and the more practical problem of
locating the new Messiah.
The alternative to a more comprehensive theoretical understanding of
the social world within which policy analysis functions is a better
understanding of policy-making and policy analysis. If we better under-
stand the characteristics of policy itself, and the characteristics of the
policy process, we may be able to intervene more effectively into the social
world even without a Keynes or Marx to guide us. Thus, we will be
arguing that policy and the policy process can and should be understood
without agreement on a social paradigm. (Asinus asino, sui suspulcher.)
Such an approach need not be the 'muddling through' of incremental-
ism, as such an approach would have a firm theoretical base; that base
would be in policy analysis itself rather than broader social or political
theory. As such it can seek to provide answers to social questions without
fully comprehending all the characteristics of the social process. This has
risks (Goodin, I982, 2o-8) but so does waiting for the right social theory.

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 239

We will be adopting the perspective of policy analysis as a distinctive

approach to social problems for several reasons. First, the 'emerging,
post-industrial world' is one of great complexity, interdependence and
indeterminacy. There will be a greater fusion of domestic and interna-
tional polities and economies to the extent that national decision-makers
will have less control over their own destinies. In addition, the increasing
interdependence of policy areas and programs will mean that any
intervention into one area may have significant consequences for many
other policies and programs. This may be the very sort of world that may
make the development and utilization of a comprehensive social theory
approach difficult. This will be especially true if the type of theory
envisioned is of a deductive and closely ordered nature (Alexander, I 965).
Paradoxically, then, the complexity of the post-industrial world being
described creates a great demand for a comprehensive theoretical under-
standing of social processes at the same time that it may make such an
approach less useful for policy-making, especially policy-making at a
mundane, program level. Thus, what may be needed as much as a theory
of the post-industrial world is a theory of policy design which allows the
policy analyst to deal with the complexity of the world in a more
intelligent and contingent fashion (Dryzek, I 983).

2. Policy design

The application of ideas from the design sciences to solving social

problems first appeared in the planning literature. Architects and others
in the applied disciplines connected with urban design turned their
attention to the organizational problems of the modern city. The basic
idea was to externalize the process of creating solutions to ensure both its
rationality and manipulability. Techniques for quantifying this design
process and adding computer assistance were incorporated from opera-
tions research (which continues to play a controversial role in architec-
tural design) and were coupled with an engineering perspective on
problem solving. The most notable proponent of this approach to design
has been Christopher Alexander (i 964; I 965).
Not only could the process of designing be divorced from the thing
designed, but it was no longer limited by the capacities of a single
designer. Techniques to aid the design process could convert the designer
from 'magician' working with intuition and insight to 'systems engineer'
capable of decomposing a problem and systematically searching for an
optimal set of possible solutions. These efforts to remove the individual
designer's internal constraints and to widen the area of search drew the
attention of Herbert Simon. His contribution to a design approach
focused on the logic of the design process. He applied several of the

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
240 Stephen H. Linder and B. Guy Peters

notions from his work in artificial intelligence to developing an algorithm

for conducting problem-motivated design (1978). His emphasis on the
hierarchical structure of problems and strategies for their decomposition
linked the design focus to decision problems in the management science
literature.
Early writers from this group concerned with design include Ernest
Alexander (I982) (no relation), who studied the inadequacies of the
policy formulation process with the framework inspired by Simon. In the
last few years several authors have applied the design theme to problems
of policy formulation in the Federal government; these include Mosher
(i 980), Salamon (i 98 I), Wolman (i 98 I) and Dryzek ( 983) . In each of
these instances, the emphasis has been on systematizing the process of
policy formulation to overcome both policy-makers' biases and the
preferences of most analysts for ex post evaluation.
As noted earlier, our attention to policy-making has been skewed in
favor of evaluating the consequences rather than the origins of specific
alternatives. While subsequent effort is devoted to analyzing data on a
given problem, little systematic attention is given to fashioning the means
of providing a solution. Those involved in actually developing alterna-
tives within the policy area tend to put aside training in systematic
analysis in favor of ajudgemental and intuitive approach. And with good
reason, since little attention is given to design training in professional
policy schools.
Several recent texts in policy analysis (Quade, Dunn, Mood) attempt
to summarize in several sections the design experience of professional
analysts. However, these commentaries typically offer advice on what not
to do and what pitfalls to avoid in applying one's intuitions, rather than
offering even a conceptual framework for undertaking design tasks. We
are left with two views of design: an interactive and creative process with
few rules and guidelines, or a winnowing process of heavily constrained
search governed by a concern for feasibility. In the former case, designing
is idiosyncratic and chancy, while in the latter, it appears deterministic
and inertial. The utilization of a macro-theory as the background to
design would tend to push design towards the more deterministic
approach.
Even if we were not dealing with the complexity of the post-industrial
world as outlined, however, there are a number of difficulties in the
application of a macro-level theory in the policy process. As discussed by
MacRae and Wilde (1979) policy-making requires: a definition of the
problem, models and alternatives, criteria for the evaluation of alterna-
tives, and an assessment of political feasibility. Stated another way, we
will argue that to engage in policy-making one needs a model of causation,
a model of evaluation and a model of interventions.

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 241

The model of causation allows the analyst to associate certain outcomes in

the society with certain initial states; disease results from germs. There
may be alternative models of causation - disease coming from an
imbalance of humors in the body or from evil thoughts - and these
alternative models may be associated with world outlooks such as
Marxism.
The model of evaluation allows the analyst to map the outcomes of the
policy process, or the initial state of society, into a set of normative
premises about policy and society. These premises may be very broad,
e.g. justice, or they may be more policy specific, e.g. efficiency.
Finally, the model of interventions guides the analyst in the selection of the
locus and form of intervention into ongoing social and economic processes
in order to produce a desired outcome. Thus, this would be a theory of at
what stage government should intervene, and which of the tools in its
toolkit (Hood, I984) it should select.
We will argue that any macro-level theory will tend to provide very
inadequate guidance in each of these three aspects of policy-making, and
that again there is a need for a theory of policy design to aid in the
formulation of policy, such a theory to contain within it guidance for
understanding causes, evaluating outcomes, and guiding interventions.

3. Politics in the postindustrial world

Before returning to the fundamental problem discussed above, we should

also point out the political difficulties which may be associated with the
development of a more comprehensive vision of policy-making in the
emerging post-industrial world. Two features of politics in the majority of
industrial/post-industrial states would appear to make the utilization of a
comprehensive approach to policy-making problematic. The first is the
sectorization of policy-making. Whether the sectors are conceived of as
'iron triangles', 'issue networks' or 'woolly corporatism' it is clear that
most policies are now made in semi-autonomous sectors, each with its
own sets of actors, issues, and concepts (Jordan, I98I). Although the
problems with policy-making in this manner are well known (Habermas,
I976, 62; Peters, I98I), there are also decided political advantages. In
particular, fragmented political decision-making tends to be more con-
sensual and depoliticized than decision-making in more open political
arenas (Richardson, I982) and it keeps difficult decisions confined within
a narrow framework of the 'journeymen of issues' (Heclo, 1978, i o6), who
if they do not always agree on the solutions to problems certainly will
agree on the problems.
The second political feature of politics in the post-industrial world is
somewhat related to the first. This is the increasing demands for partici-

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
242 Stephen H. Linder and B. Guy Peters

pation in decision-making (Inglehart, 1977). Participation has come to be

valued as an end in itself, as well as for its instrumental value in affecting
the distribution of public resources. This in turn means that policy-
making in post-industrial political systems tends to be by bargaining and
negotiation rather than by imposition or control. It also means that
policy-making in these systems will be rather untidy - perhaps the price
which must be paid for high levels of democratic participation.
Richardson (1 982) has combined two aspects of policy-making similar
to these to describe what he calls 'policy styles' in Western democracies.
The two dimensions he employs are an active-reactive dimension, and a
bargaining versus imposition dimension. Although there are differences
among the six countries discussed using this framework, and among
policy areas within those countries, the general finding is that these
societies all tend toward the consensual/reactive cell of the four-fold table.
This is, therefore, diametrically opposed to the active/imposition ends of
the dimensions which would be the logical place for the political utiliza-
tion of a macro-level theoretical approach. Also, in at least one of the
countries where a more radical policy approach has been successful
previously, Sweden, Ruin (I982) finds that this form of policy-making is
less acceptable. Thus, the political style of the post-industrial world may
make the use of a macro-level theory for policy guidance less acceptable to
members of the society, and to members of an increasing number of
organized groups, who may have theories of their own. The utilization of a
macro-level theory would tend to be extremely centralizing. This may be
useful given the nature of some of the problems in the post-industrial
world which have been outlined. It would, however, run counter to many
of the decentralizing trends seen in contemporary politics and, as
Benjamin (I980) has pointed out, decentralization may not be costless.
As in most issues of importance, there is substantial disagreement with
the characterization of government in the post-industrial world as being
highly dependent upon its environment and upon agreement of the
affected actors. Nordlinger (I982), for example, argues quite strenuously
that the modern democratic state has substantial (and by implication)
increased autonomy. He finds that these governments can indeed make
decisions even when their policy conceptions diverge from those of the
population. But even in this case the discussion appears to be more of a
case-by-case exercise of authority rather than the imposition of a compre-
hensive framework of action. Even in perhaps the most single-minded
government in a democratic country in recent memory - Mrs Thatcher's
Great Britain - there have been some notable U-turns (Keegan, I984).

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 243

4. Macro-theory and policy-making

To return to the main line of argument, the shortcomings of applying a

macro-level theory to the policy process might be illustrated best by
briefly considering the well-known case of macro-economic theory. Put-
ting aside for the moment the widespread disagreements over the
appropriate premises of the theory, many of the problems connected with
its policy relevance are a product of its scope - in a sense it is too macro a
theory for generating policies with any precision. While aggregation has
permitted theorists to explain the behavior of a complex economy by
means of a few basic relationships and a small number of economic forces,
it has also clouded the impact of policy interventions on economic actors
and immersed policy development in guesswork.
Aggregate indicators of economic performance, summarizing the
behavior of prices, output and unemployment for example, serve as
proxies for the outcomes of a variety of micro-level adjustment processes.
Relying on these proxies as target variables for constructing and monitor-
ing policy instruments requires more than a courageous ceteris paribus
assumption; the known and unknown adjustment processes must behave
in a uniform, reliable manner. When adjustments are not largely uniform
in space and time, or when their performance is idiosyncratic, the
information summarized by an aggregate indicator is seriously com-
promised. This can bias policy targets, at times signalling the need for
intervention where none exists, while hiding a genuine need when one
arises.
Much of the policy debate in the past few years has centered on
stabilization policy, the use of various instruments to counteract short-
term fluctuations in economic performance. Advocates of smoothing
fluctuations in aggregate demand (Tobin, I 981 ) generally call on monet-
ary and fiscal instruments; more recently, those concerned with disrup-
tions in aggregate supply (Okun, I98I) have turned to incomes and
manpower policies. The task of design in this context is not so much one of
fashioning new instruments as of finding the plausible interventions. As
generic policy strategies, the available instruments, taxing and spending
for example, are highly institutionalized and are widely viewed as the
most potent sources of potential, policy-induced changes in the economy.
Despite their power, these instruments can rarely be wielded with any
precision.
On the one hand, design from a macro-theory warrants caution.
Because the details of the micro-level adjustments in the economy are
open to dispute we can expect lags of indeterminate duration in the
responses of policy targets to any particular intervention. Furthermore,
we cannot be certain about either the location or the magnitude of the

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
244 Stephen H. Linder and B. Guy Peters

eventual effect of an intervention. Designing an instrument to have a

precise effect of a given magnitude is practically impossible. Moreover,
many of the available instruments work in an antagonistic fashion. Using
one instrument to reach a selected policy target may actually move us
farther away from other targets. Raising output toward full employment
through an expansionary policy that applies fiscal instruments will
disrupt the trade balance, for example. To keep both targets in line we
need an additional instrument to reverse the negative effects of the first,
and so on. As two noted theorists (Dornbusch and Fischer, i982) have
noted, we normally require as many instruments as there are targets.
Every time one instrument is used it spoils the balance achieved by
another. If we design the right instruments, these displacements should
diminish in size; otherwise the increasing interdependence of portions of
the economy will mean that they will grow.
On the other hand, design may be viewed as altogether ill-advised.
Some theorists eschew these instruments, viewing the products of any
design effort as typically too blunt and the prospects of success in using
them as too remote. Theorists such as Sargent and Wallace (I 976) would
prefer to rely on market forces to restore the level of equilibrium to the
economy and view most policy instruments as an additional source,
rather than cure, of disruptions. Those advocating interventions find the
self-adjusting propensities of the economy too sluggish or largely ineffec-
tual. Unfortunately, the resolution of these views must await a more
extensive articulation of the behavioral foundations of the theory at a
micro-level. If we fix our focus on the macro-level, we are left wondering
whether to design at all; a curious implication for a macro-theory with
such an elaborate scheme of causation. Unless the micro-level assump-
tions were well-articulated the theory would have little hope in providing
guidance for designing effective and efficient public policies.
The experience with macro-policy theory in sociology has been little
better. Attention to social processes has produced a rather elaborate set of
causal propositions regarding institutional change and social develop-
ment, but little in the way of refined policy instruments. The work of
Weber and his followers concentrates primarily on the role of various
background factors - Protestant ethic, leadership style, organizational
climate - in accounting for the development and performance of social
institutions. Changing these features, say to improve performance on
some criterion, requires substantial leverage. In effect, one has to undo
the result of an evolutionary process, overcoming built-in equilibriating
and defense mechanisms, as well as neutralizing counterforces in the
environment. Unlike macro-economics, macro-sociology has few tools
available for reaching any given policy target; as a result, theory is as
likely to support the shotgun or the sledgehammer as it is the scalpel

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 245

(Scott and Shore, I 979; Sieber, I 98 I). The approac

sociology supports this view. Remedying social problems, in their view,
requires fundamental changes not only in social policies, but in the
institutions and people that make them. The potential for error and the
social costs of intervention on this scale are what frightened Popper (I 945)
and Hayek (1 944) so much about centralized planning. However, social
planning or the use of theory in policy-making per se is not the basic
problem. The problem rests with the affinity of macro-theory for blunt
instruments.
Although enamored of macro-theory, Etzioni (I968) is nonetheless
concerned about the blunt instrument problem. Rather than reverting to
the atheoretical position of the hardline i-ncrementalists (Lindblom, i 965)
he advocates various forms of a mixed-scanning approach - theoretical in
its choice of interventions but incremental in their adjustment and
revision. This, of course, is basically the same process of 'fine-tuning' used
by advocates of stabilization policy to correct potential instrument errors.
Once again, the difference lies in the relative availability of levers and
concrete targets (Hogwood and Peters, I985). Etzioni is left with the
problem of how to guide social processes. While he can specify the
dynamic underlying these processes in different contexts, his choice of
intervention points is guesswork at best.
The difficulty faced by sociologists and economists in developing
refined instruments from macro-policy suggests that the limitation resides
not in the particular disciplinary approach or in the details of the theory,
but in the nature of macro-theory itself. While one of the principal ways of
differentiating macro from micro theories is scale - macro-theories adopt
aggregates as their unit of analysis and focus on systems-level phenomena
- the shortcomings of macro-policy approaches can be tied more closely to
the implicit tradeoff between precision and inclusiveness that an increase
in scale demands. We cultivate precision in our theories by controlling, or
at least understanding, many of the individual sources of variation that
macro-theories typically regard as sources of error. Unfortunately, as we
expand our theories to be more inclusive of causal relations, values, or
instruments, our ability to maintain precision rapidly diminishes. Main-
taining both precision and inclusiveness is the hallmark of elegance
among theories in the natural sciences, but is as yet too complex a task for
the social and behavioral sciences. Levels of aggregation correspond to
levels of acceptable error in dealing with human behavior. Thus, to a large
extent, the transition from a micro to a macro-level explanation involves
the implicit choice between alternative mixes of precision and inclusive-
ness. The available set of choices is constrained by complexity.
Consider the choice space depicted in Figure i. To develop a service-
able design theory we need to maximize precision without sacrificing a

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
246 Stephen H. Linder and B. Guy Peters

great deal of inclusiveness. This takes us to the fabled middle-range

(Merton, I957) at the intersection of macro and micro approaches. Here
we can find the optimal mix to satisfy each of the three dimensions of
design theory: causality, evaluation and instruments. In the case of
instruments, for example, not only is a broad range of different kinds of
instruments needed, we also need precision in our mapping of instru-
ments into problems. Imprecision, of course, means that we would likely
have both sleeper and spillover effects, many of which might exacerbate
the problem of concern (Goodin, 1982, 22-4). Macro-theory alone,
despite its capacity for accommodating a greater array of causes, values
and instruments, cannot provide the necessary precision. Moreover, as
we will see, the inclusiveness of macro-theory generates its own, internal
contradictions.

FIGURE I: Tradeoffs between precision and inclusiveness

Precision

r2 \ Design Theories
Teories /

-. ;Macro Theories

Inclusiveness

5. Toward a theoty of evaluation

Evaluation is widely recognized as an integral part of the policy-making

process. It is at times a rather simple mechanical exercise of applying
statistics to the outcomes (or presumed outcomes) of policies and asses-
sing the extent to which the policy reached its goals (or presumed goals).
If, however, we begin to think about evaluation in a broader theoretical
sense, and from the perspective of policy design outlined previously, some
important problems arise.
The problem with making a claim to incorporate fundamental values
(e.g. justice, fairness, equality) into any macro-policy theory is that we
need a macro-level theory of evaluation for doing so. Ifwe disavow macro-
theories because of their weaknesses in generating propositions about
instruments (see below), we may have to stick with an operational theory

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theoty to Policy Design 247

of evaluation as well. Consider Rawls' (I97I) the

elaborate macro-theory of evaluation, he has only a rudimentary theory of
instrumentation and not much of a theory of causality. Macroeconomics,
on the other hand, has a well-developed theory of causality, a rudimen-
tary theory of instrumentation, but little theory justifying its normative
premises. As a general rule, the stronger the theory of causality at the
macro-level, the weaker the theory of evaluation and vice-versa. This is
largely the produce of the incompatibility between the instrumental logic
of our causal assertions and the non-instrumental nature of most of our
societal overriding values.
Let us consider the constituent of theoretical 'strength' most closely
associated with macro-policy theory, inclusiveness, and reexamine this
general rule. To aid our reexamination, as well as to illustrate our earlier
points, we will construct a heuristic device involving the analytic
representation of macro-theories as points in a three dimensional space.
Each dimension will correspond to one of our three design ingredients -
causes, values and instruments - and positions along these dimensions
will represent degrees of inclusiveness. An example appears in Figure 2.
As we move to the right in Figure 2, we will find theories accommodating
increasingly greater numbers of causal variables and relationships. Mov-
ing upward vertically signifies an increase in the array of values, and
moving outward increases the types of instruments. From our earlier

FIGURE 2: Three dimensions of design theoiy

VALUES

S b~~~~

/ *-> : . ~~~CAUSES

/. . . . .. . . . * .00. * ...O

INSTRUMENTS

Note: Dimensions are ordered according to increasing

inclusiveness.

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
248 Stephen H. Linder and B. Guy Peters

statements about inclusiveness as a definitional attribute, we can show

macro-theories as falling some distance from the origin in this space. In
addition, the optimal mix of precision and inclusiveness proposed earlier
as a distinguishing feature of design theory can be represented in this
space. Balancing the need for precision against the demands of inclusive-
ness must take place for each of these dimensions. As a result, we would
expect design theories to fall somewhere along the middle portion of the
diagonal from the origin, labelled 'a'. The curve labelled 'b' in Figure 2
illustrates the logical contradiction suggested by the general rule stated
above. Capturing a greater number of causal relations demands an
increasingly narrow set of values in order to refine the focus of one's
explanations. Conversely, attention to greater numbers of values eventu-
ally leads one beyond causal relations to consider relations of other types
(e.g. responsibility, obligation).
Consider the extreme example of a complex macro-policy approach
which has highly inclusive theories of both causality and instruments.
Macro-systems theory (Kozmetsky, 1980), the approach underlying our
largest scale engineering projects (NASA's Project Apollo, the Jubail
Petrochemical Complex, etc.), represents the best the design sciences have
to offer. Its practitioners, moreover, are quite optimistic about the
approach's applicability to policy-making: 'macro-engineering must be
co-joined [sic] with social engineering at the macro-systems level so that
newer institutional arrangements - economic, political, social - can be
formed that solve today's and tomorrow's human and social needs'
(Kozmetsky, I 980). This carries the endorsement of the American Associ-
ation for the Advancement of Science (Davidson et al., I980). The
advantages of this macro-systems approach over the other macro-theories
based in social science is its capacity for reaching a higher degree of
precision for any given level of inclusiveness. In effect, it is a theory which
can be applied at either micro or macro levels, spanning the space shown
in Figure i.
However, its capacity for accomplishing this is based on the careful
exclusion of behavioral and social phenomena. Performance is equated
with the ability to meet certain mechanical and physical stresses within
fiscal limits. The one caveat acknowledged by the proponents of this
approach is the often troublesome problem of 'cultural impact'. The
products of macro-systems design are typically large-scale projects that
have profound socio-cultural impacts on the societies that choose to
undertake them. Solutions of the technological and managerial aspects of
the problem can be accomplished within the macro-theory. The issue of
social values remains.
Figure 3 fills in Figure 2 with examples of the four classes of theories
introduced in the preceding sections. Macro-systems theory, while the

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theoty to Policy Design 249

most inclusive of instruments and causal relations, is largely exclusive of

social values. In a sense, this represents asocial engineering; thus, it is not
the ideal prototype for policy design that some say it is. Macro-economic
and macro-sociological theories appear at a lower level of causal and
instrumental inclusiveness, with the latter accommodating far fewer
instruments. Finally, the macro-philosophical theories appear higher
than the other theories on values but lower on the remaining dimensions.
Notice the suggestion of a pattern resembling the curve labelled 'b' in
Figure 2. Notice as well the absence of any cases close to the diagonal
labelled 'a'. Filling in this diagonal is one way of defining the task facing
the developers of a theory of policy design. It is not a task which can be
accomplished from the macro perspective alone.
Moreover, as a practical matter, industrialized societies typically avoid
direct dealings with fundamental values for policy purposes because of the
potential conflicts this might create. Value-laden matters are generally
settled through some type of mechanism which can sustain pluralistic
competition while fashioning temporary resolutions on specific policy
issues. The design of effective policies, then, is better served by focusing on
the instrumental values which play a prominent role in problem recog-
nition and definition. Instrumental values not only define desired per-
formance at an operational level and reveal any gap left by existing

FIGURE 3: Macro theories and design dimensions

VALUES

MACRO C
PHILOSOPHY

[-. . . . . .V.
(e.g. Rawis)

MACRO .
SOCIOLOGY , _ . _ * _ __ _

(e.g. Etzioni) 0 CAUSES

/ -- . < MACRO
9..5%. ECONOMICS

MACRO
INSTRUMENTS SYSTEMS

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
250 Stephen H. Linder and B. Guy Peters

practices, but also identify the underlying attributes of a desirable

instrument.
The designer's efforts should eventually yield a match between a
generic strategy or set of instruments and a specific policy goal, but need
go no further. Sorting out conflicts between fundamental values or
implicitly weighting the claims of different individuals are better left to the
decision-makers authorized to make the ultimate choice among plausible
alternatives. The designer need only reach what the decision-maker
considers a 'satisfactory' degree of approximation in a suggested match
between goal and instrument. In this context, the designer's criteria are
non-decisive, generating an agenda of plausible alternatives representa-
tive of certain generic strategies. Restraint by the designer, resisting the
temptation to interject certain values or artificially delimit choices, not
only disciplines the design process but recognizes the legitimate role of
qualitativejudgement in reaching policy decisions.
Efforts to systematize the decision process or to represent implicit
values in quantitative terms should remain logically distinct from the
design process. Although one may consider the fashioning of choice
models or mechanisms a proper design problem, it is a second-order task
which must presume that the design of alternatives has already taken
place. Separating the design of instruments from the eventual selection
process does not rule out the iterative refinement of instruments until one
converges on a single 'best' choice. Rather, we are arguing for an analytic
division of labor which will insulate the design process from choice-
connected biases that lead to premature closure. In sum, the evaluation of
designs, from the building of generic strategies through the screening and
sorting of substantive alternatives, need not employ criteria as demanding
as those applied to choices. By emphasizing the instrumental values over
non-instrumental ones, we can avoid the pitfall common to many evalua-
tion methods, including cost-benefit analysis, of claiming too much in
view of the poorly developed nature of one's theory of evaluation.

6. Why a theory of interventions?

One of the basic lessons of the Great Society's policy experiments during
the I 96os is that our most pressing problems can not be solved simply by
spending enough federal dollars (Aaron, 1978). While concern for the
scale of policy solutions has prompted a reassessment of federal efforts in
many policy areas, there have been only minor revisions of the solutions
themselves (Schulman, I980). For the most familiar problems facing
society - crime, pollution, urban blight - there is little understanding of
the factors which contribute to a policy's eventual success or failure.
Knowledge about which strategies work best under which conditions is at

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theogy to Policy Design 25I
best rudimentary. Yet, without this knowledge society must rely upon
trial and error for developing the solutions to its policy problems. Under
these circumstances, only the most familiar strategies receive attention as
possible solutions. Strategies producing even minimal success are guaran-
teed not only longevity but many reincarnations. For example, social
insurance schemes, despite their fiscal limitations, have been recycled
from their applications to problems of labor markets to health care and
are viewed by some as holding promise for product safety problems.
Earlier strategies, including previously discredited ones, constantly
reappear as new solutions (Hogwood and Peters, I983, 26I-4). Hoover's
Reconstruction Finance Corporation, for example, has reemerged as a
progressive strategy to promote economic development.
Informed by only fragmentary evidence, the formulation of most
strategies proceeds on an ad hoc basis. Components of other strategies
which are presumed to be successful are incorporated in a haphazard
fashion. Moreover, there is a tendency to choose policy instruments on the
basis of how they work, without seriously considering how well they may
perform given what needs to be accomplished. We routinely identify
problems with the method used to deal with them, e.g. regulation, to the
point of naming a problem by its corresponding method. This practice is
especially evident in our treatment of legal remedies, but also shapes our
view of market failures.
Also, because of an absence of an adequate theory of interventions,
there is a similar tendency to label problems with nominal, functional
titles, e.g. health, education or defense. Such a labelling scheme tends to
lump together problems requiring (probably) very different forms of
intervention. For example, the 'health' problems of quality assurance
probably should be treated by very different mechanisms than problems
of cost control. Unfortunately, we do not have an adequate guide for
choosing what methods of intervention would be most appropriate for
either. We especially are not sure why one would be more appropriate
than another for either problem.
Unfortunately, defining a problem in terms of a strategy for solving it
reinforces the bias toward favored strategies and deceptively simplifies
the formulation process. Poor performance, then, is attributed not to the
choice of the wrong policy instrument, but to the symptoms of poor choice
such as waste and mismanagement. The failure of a poorly fashioned
program is too often blamed upon the implementor, under the mistaken
presumption that success is a product of the program's adaptability in the
field, rather than its underlying strategy.
The artificial separation of social science theory from the formulation of
practical policy inhibits the application of insights from research to the
design of concrete solutions. Apart from economics, the social sciences

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
252 Stephen H. Linder and B. Guy Peters

typically resist the formulation of novel alternatives for improving policy

(MacRae, 1976). Although acknowledging some value bias, efforts to
combine judgements about values with empirical propositions to
influence policy are viewed with suspicion. While economists have con-
tributed readily to policy advice, an emphasis on one value - economic
efficiency - over others has restricted both the variety of proposals and the
basis of evaluation. To the extent that economic instruments are restric-
ted to altering market incentives and changing information and owner-
ship conditions, other disciplines have a complementary contribution to
make. Sociology focuses attention on changing norms and organizational
arrangements, psychology on changing preferences and beliefs,
anthropology on changing symbols and social relations, and political
science on changing policy processes and political institutions. The
variety of disciplinary perspectives can expand the range of political
instruments and the values which they serve.
Although the design of physical forms to solve certain problems has
long been a central focus for the applied sciences, such as engineering and
architecture, no comparable focus on the design of solutions for public
problems has emerged within the policy profession (but see Alexander,
1982). The emphasis on fashioning solutions which once characterized
the planning profession with its emphasis on social reform has given way
to a limited focus on evaluating the short- and long-term consequences
of alternative policies. Policy analysts, trained primarily in the social
sciences, also deemphasize the design of solutions, preferring instead to
concentrate on the comparison and evaluation of given alternatives. Left
to the political process, designs emerge less as the result of creative,
systematic effort than as a product of experiences, precedent and
expedience. Analytic skills simply are not devoted to the formulation of
alternatives. The professional analyst typically accepts the set of alterna-
tives as given, the product of some advocate's entrepreneurship and of
political compromise. The analyst's role then is not to remake the
alternatives but to predict their impact and, to an increasing degree, to
evaluate the prospects for enactment and trouble-free implementation.
However, if the design of policy instruments is left to intuition, and the
details worked out informally, the analyst's emphasis on evaluation is
misplaced. The outcome of policy will be shaped more by the agenda of
alternatives and the values underlying it, than by the results of the
analyst's formal evaluation. For example, once the strategies for dealing
with social risks are limited to various forms of direct controls on private
activity, subsequent evaluation of the controls' performance is virtually
irrelevant. Alternative means that might prove less costly and perhaps
more effective have been foregone. The design and formulation of strate-
gies must be given systematic attention if we are to overcome institutional

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 253

and conceptual obstacles to innovativeness in our policy solutions. At

present, no middle-range theory for guiding the design of specific policy
strategies exists. Without even a rudimentary theory, there is no way of
assigning a policy instrument to a problem with any confidence of the
outcome. Similarly, there is no direction for developing new instruments
or adapting existing ones to different purposes.
A focus on the design of policy instruments shifts attention from the
evaluation of different substantive programs to a concern with the basic
strategy implicit in each alternative. What kind of instruments, if any,
should the government employ for a particular problem, and why? More
importantly, how do we define an instrument as the appropriate choice for
a particular task, and what values are at issue? Research on the design of
policy should address these questions in the course of developing a theory
that will accommodate familiar instruments, such as taxation and price or
quantity controls, and suggest new ones. The instruments associated with
the perspective of any one discipline constitute only a limited set of cases.
The study of designs, then, should range from the legal profession's focus
on changing entitlements or liability rules to the economist's concern with
pricing schemes.
Whether the problem is an architectural, mechanical or administrative
one, the logic of design is fundamentally similar. The idea is to fashion an
instrument that will work in a desired manner. In the context of policy
problems, design involves both a systematic process for generating basic
strategies and a framework for comiparing them. Examining problems
from a design perspective offers a more productive way of organizing our
thinking and analytical efforts. Once the set of basic instruments is
developed, attention can be devoted to the substantive details of each
alternative. The analyst must ensure that an appropriate range of designs
is advanced during the formulation stages of policy-making. Systematic
attention to design will not only enhance the performance of the alterna-
tive eventually chosen, but also expand the opportunities for serious
consideration of innovative strategies.
Efforts to develop a theoretical approach to interventions should have
several goals. The first is to adopt a coherent approach to policy problems
which stresses the central role of design. This involves a reorientation of
conventional policy analysis away from ex post evaluation and a preoccu-
pation with the substantive details of particular programs to the develop-
ment of generic strategies for solving problems. Consider the issue of
consumer protection. While there is no general theory of regulatory
design identifying certain strategies for consideration, we can propose a
range of strategies, which emphasize different criteria, and then examine
their implications. We might consider various information instruments
(such as mandatory disclosure of consumer education), direct controls

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
254 Stephen H. Linder and B. Guy Peters

(such as design or performance standards), and legal instruments

(including liability and negligence systems). Moreover, each of these pure
strategies might be combined with others to form hybrids and implemen-
ted in ways that stress selected values, perhaps emphasizing equity over
cost minimization, or discretion over accountability (Bardach and
Kagan, I982).
The second goal of policy design is to fill in the gaps of our knowledge
about what might work and to broaden our view of possible strategies. In
the process, various obstacles to innovation in policy design - disciplinary
prejudice favoring the elaboration of causal theories, biases toward
familiar strategies, reliance on ad hoc formulations, parochial views of
feasibility - must be confronted (Scott and Shore, 1979; Sieber, I98I).
Following conventional practice in the design professions, we can
enhance the productivity of our efforts by carefully organizing both the
search among existing strategies and our efforts to create new ones.
Establishing a logical procedure for designing policy instruments not only
reduces the likelihood of errors but also makes explicit the thinking that
goes into the development of each design. Complex problems, then, can
be reduced to manageable proportions by clarifying basic design require-
ments and developing plausible strategies for their solution.

7. Notes towards a theogy of design

As we stated in section 2, policy analysis requires an understanding of

causation, of evaluation, and of intervention. The approach to policy
design which we have been sketching here will require at least the
following elements to deal with the complexity of design problems in the
complex world in which policy is now fashioned.

7.1 Characteristics ofproblems

Rather than dealing with policy problems in nominal categories (health,

education, or regulation, i.e. categories bearing either the name of a
'policy area' or of a tool), we should develop some attributes of policy
problems which will be useful for designing interventions. These might
include such things as:
i. Scale. As discussed by Schulman (1 980), scale implies that some policies
cannot be effective unless some threshold size is attained. Obviously,
designing interventions for policies of this type will be different than for
policies which can be managed more incrementally.
2. Collective consumption goods. Goods of this sort have special features which
may influence the manner in which they must be treated in the policy
process (Benjamin, I980).

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 255

3. Certainty. In designing policies, we can be certain - or relatively certain

- that some environmental changes have occurred and less certain about
others. For example, we are quite sure that major demographic changes
will be occurring in almost all industrialized societies. We are somewhat
less certain of the occurrence, extent, and nature of changes in the natural
environment.
4. Predictability. Certain occurrences can be predicted on a 'if x, then y'
basis. This does not appear to be the case for much of social change. As in
the Moon and the Ghetto analogy, some seemingly difficult problems can
be engineered more readily than others.
5. Independence. In the contemporary post-industrial world, virtually all
programs and policies are impacted by international forces. Even seem-
ingly domestic issues such as social security programs may be heavily
impacted by external events, e.g. interest rates in the international
market. The degree of independence of a program is, however, a variable,
and decision-makers must understand the degree to which they can, or
cannot, determine the outcomes in a particular policy area.

7.2 Characteristics of goals

We have to understand better the nature of the goals which may be

posited for public programs. As noted in section 2, these may be broken
down first into the heavily value-laden goals such as justice, equity, etc.,
and the more operational goals such as the types of economic criteria used
to justify governmental intervention into the economy (Wolf, 1979).
Goals are among the least well-understood components of policy-making,
and a great deal of additional work is required to be able to bring them
into the policy design process in a more operational manner.
As noted in section 5, the importance of goals in policy analysis also
implies some attention to, and respect for, the nature of the process
through which goals are set. Unlike the asocial approach to policy design
exemplified by the macro-systems approach, an approach to policy design
workable within the context of democratic policy-making systems will
require some attention to both goal-setting and goal clarification through
the political process. This does not mean, however, that the policy analyst
and the 'policy organization' must be only reactive to goal statements
produced elsewhere; they can and should be quite active in attempting to
modify goals. They must, however, remember that technical solutions do
not necessarily imply the goals that would justify them.
At the present state of development of value theory in policy analysis,
perhaps the best the analyst can hope for is to inject goal statements into
the analysis in a contingent fashion. That is, the analyst may be able to tell
the decision-maker that ifXis his or her principal goal then x should be the

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
256 Stephen H. Linder and B. Guy Peters

policy adopted within the context of the range of alternatives available. It

may well be that there may be no alternative readily available that scores
highly on that particular goal; without the more explicit attention to goals
and the design process the decision-maker might not be made aware of
that weakness in the range of alternatives considered.

7.3 Characteristics of instruments

As noted in section 6, this will be a major component of any theory of

policy design. Some preliminary work has already been done on this
problem. For example, Hood (I 984) has classified the 'tools' available to
government as NATO (nodality, authority, treasure, and organization).
Again, a great deal more must be done to gain an understanding of the

TABLE I: Rudimentay appraisal criteria

Criteria Model Instrument

I. Relevancy Does the model describe Does the instrument relate

the policy content? directly to the policy
problem?

2. Distortion Is there a bias between Will the instrument cause

the model and the distortion of other social or
reference system? economic processes?

3. Stuctural Is the model design based Is the instrument based on

integrity on internally consistent internally consistent
principles? principles?

4. Reproducability What is the model's 'track Has the instrument been

record' for replicating successful in similar
historical data? situations?

5. Tractability Is the model easily used? Is the instrument easily used?

6. Accessibility Are the model's input and Is information available to

output familiar and make the instrument work and
intelligible? to monitor its effects?

7. Flexibility Is the model design capable Can the instrument work in

of undergoing change? a changing environment?

8. Common sense Are the model's forecasted Does intuition tell one the
results offensive to basic instrument should be
intuition? effective?

9. Credibility Is there consonance between Do those administering the

the model-builder and the instrument believe it will
policy-makers? be effective?

io. Efficiency What are the costs What are the costs
associated with operating associated with operating
the model? the instrument?

Adapted from Hall (1975).

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 257

nature, political, social and economic, of the weapons which governments

can wield.
One checklist which may be useful in evaluating the utility of instru-
ments for particular settings was developed by Hall (I975) for the
evaluation of policy models. Table i contains the ten criteria developed by
Hall, along with the descriptors he applied for policy models and
descriptors we have developed for policy instruments. In some of the cases
the criteria applied to a model may not be obviously applicable to a policy
instrument. For example, the criterion of distortion implies rather dif-
ferent things when discussing possible bias in a computer simulation and
the distortion which a policy instrument may cause in other elements of
the policy environment. This checklist does, however, provide a place at
which to begin thinking about the types of questions which must be asked
every time governments begin to intervene, and begin to select the
instruments they will use to intervene. In a similar vein, May (i 98I ) has
assembled some 'hints' for crafting alternative policies. As with our
checklist and earlier admonitions about reasoning by analogy, May
warns against misclassifying problems and applying stock solutions to
difficult, complex, and perhaps novel problems. All these ideas and
approaches merely reinforce the basic premise that there is as yet very
little in the way of a theory of instrumentation in policy design, and that
there is a need for such a theory. Such a theory would, of course, need to
meet the criteria of inclusiveness and precision which have been advanced
earlier.

8. Summary

In summary, we have attempted in this article to lay the groundwork for a

more theoretical approach to policy-making through a concept of policy
design. This work is as yet rudimentary, but hopefully it will allow us to
address the difficult problems in a way that will produce effective and
efficient public interventions, rather than the almost random interven-
tions which we would argue characterize much contemporary policy-
making. We have argued that such a theory of policy design must contain
means of including causation, evaluation and instrumentation. We have
further argued that concentrating on such a theory of design which is
specifically policy related will be a more efficient means of improving
public policy-making than will be waiting for the development of more
robust and precise macro-level social theories.

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
258 Stephen H. Linder and B. Guy Peters

REFERENCES

Aaron, H. (1978) Politics and the Professors: The Great Society in Perspective. Washington DC: The
Brookings Institution.
Alexander, C. (i964) Notes on the Synthesis of Fom. Cambridge, Mass.: Harvard University Press.
Alexander, C. (I965) A city is not a tree, Architectural Forum, 122.
Alexander, E. R. (i982) Design in the decision-making process, Policy Sciences, I4, 279-92.
Bardach, E. and R. A. Kagan (I982) Going by the Book: The Problem of Unreasonableness. Philadelphia:
Temple University Press.
Benjamin, R. (I980) The Limits of Politics. Chicago: University of Chicago Press.
Benjamin, R. (I983) Memo to members of policy roundtable. Mimeo, University of Pittsburgh.
Cohen, M. D.,J. G. March andJ. P. Olsen (I972) The garbage can model of organizational choice,
Administrative Science Quarterly, I 7, I-25.
Davidson, F. P. et al. (I980) How Big and Still Beautiful: Macro-Engineering Revisited. AAAS Selected
Symposium 40. Boulder, Colorado: Westview.
Dornbusch, R. and S. Fischer (I982) Macroeconomics, 2nd ed. New York: McGraw-Hill.
Dryzek, J. (I983) Don't toss coins into garbage cans: a prologue to policy design, Journal of Public
Policy, 3, 345-67.
Dunn, W. (i98i) Public Policy Analysis. Englewood Cliffs, NJ: Prentice-Hall.
Etzioni, A. ( I 968) The Active Society. New York: Free Press.
Goodin, R. (I 982) Political Theory and Public Policy. Chicago: University of Chicago Press.
Habermas,J. (1976) Legitimation Crisis. London: Heinemann.
Hall, 0. P. (I 975) A policy model appraisal paradigm, Policy Sciences, 6, I85-95.
Hayek, F. A. (I 944) The Road to Serfdom. London: Routledge and Kegan Paul.
Heclo, H. (I978) Issue networks and the executive establishment. In A. King (ed.), The New American
Political System. Washington DC: American Enterprise Institute.
Hogwood, B. W. and B. G. Peters (I983) Policy Dynamics. New York: St. Martins.
Hogwood, B. W. and B. G. Peters (I985) The Pathology of Public Policy. Oxford: Oxford University
Press.
Hood, C. (I 984) The Tools of Government. London: Macmillan.
Inglehart, R. (1977) The Silent Generation in Europe. Princeton: Princeton University Press.
Jordan, A. G. (1 98 I) Iron triangles, woolly corporatism and elastic nets: images of the policy process,
Journal of Public Policy, 1, 95-I 23.
Keegan, W. (X984) Mrs. Thatcher's Economic Experiment. London: Lane.
Kozmetsky, G. (I980) Evaluation of macro-systems: models and case analysis, in Davidson et al.
(i 980).
Lindblom, C. (1 965) The Intelligence of Democracy. New York: Free Press.
MacRae, D. (1976) The Social Function of the Social Sciences. New Haven: Yale University Press.
MacRae, D. andJ. A. Wilde (I 979) Polity AnalysisforPublic Decisions. North Scituate, Mass.: Duxbury.
May, P.J. (I98I) Hints for crafting alternative policies, Policy Analysis, 7, 227-43.
Merton, R. (I957) Social Theory and Social Structure, 2nd ed. Glencoe, II.: Free Press.
Mood, A. (I983) Introduction to Policy Analysis. New York: North-Holland.
Mosher, F. (I980) The changing responsibilities and tactics of the federal government, Public
Administration Review, 40, 54I-3.
Nordlinger, E. (I 982) On the Autonomy of the Democratic State. Cambridge, Mass.: Harvard University
Press.
Okun, A. (I98I) Prices and Quantities. Washington DC: The Brookings Institution.
Peters, B. G. (1 98 I) The problem of bureaucratic government, Journal of Politics, 43, 5682.
Popper, K. (I 945) The Open Society and Its Enemies. London: Routledge and Kegan Paul.
Quade, E. (I982) Analysisfor Public Decisions. New York: North-Holland.
Rawls, J. (I97I) A Theoy ofJustice. Cambridge, Mass.: Harvard University Press.
Richardson, J. (ed.) (I982) Policy Styles in Western Europe. London: Allen and Unwin.
Ruin, 0. (I 982) Sweden. InJ. Richardson (ed.), Polity Styles in Western Democracies. London: Allen and
Unwin, I02-34.
Salamon, L. (1 98 1) Rethinking public management - 3rd party government and the changing fo
of government action, Public Policy, 29, 255-75.
Sargent, T. and N. Wallace (I 976) Rational expectations and the theory of economic policy,Journal of
Monetary Economics, 2, I69-83.

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]
 From Social Theory to Policy Design 259

Scott, R. A. and A. R. Shore (I 979) Why Sociology Does Not Apply. New York: Elsevier.
Schulman, P. R. (I980) Large-Scale Policymaking. New York: Elsevier.
Sieber, S. D. (I98I) Fatal Remedies. New York: Plenum.
Simon, H. A. (I972) The Architecture of Complexity. Cambridge, Mass.: MIT Press.
Simon, H. A. (I978). On how to decide what to do, BellJoumal, 9, 494-507.
Tobin, J. (I98I) The monetarist counterrevolution today, EconomicJournal, 9, 29-42.
Wolman, H. (1 98 I) The determinants of program success or failure,Journal of Public Policy, i, 433-64.
Wolf, C . (I 979) A theory of non-market failure: framework for implementation analysis,Journal of Law
and EconomicS, 22, I07-40.

This content downloaded from

[Link] on Mon, 16 Dec 2024 [Link] UTC
All use subject to [Link]