The ORE of Politics will be available for subscription in late September. Speak to your Oxford representative or contact us to find out more.

Show Summary Details

Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, POLITICS ( (c) Oxford University Press USA, 2016. All Rights Reserved. Personal use only; commercial use is strictly prohibited. Please see applicable Privacy Policy and Legal Notice (for details see Privacy Policy).

date: 20 September 2017

Incrementalism and Public Policy-Making

Summary and Keywords

Incrementalism is a model of the policy process advanced by Charles Lindblom, who views rational decision-making as impossible for most issues due to a combination of disagreement over objectives and inadequate knowledge base. Policies are made instead through a pluralistic process of partisan mutual adjustment in which a multiplicity of participants focus on proposals differing only incrementally from the status quo. Significant policy change occurs, if at all, through a gradual accumulation of small changes, a process Lindblom calls seriality. For incrementalism to yield defensible policy outcomes, three conditions must be satisfied, all of which are far from automatic: 1) all, or at least most, social interests must be represented; 2) political resources must be balanced sufficiently among groups that no one actor or coalition dominates; and 3) political parties must be moderate and pragmatic, permitting a convergence to an ever-evolving political center.

While Lindblom sees nonincremental policy departures as extremely rare, subsequent research suggests that major policy departures may occur in response to crises or mass public arousal, through the development of a rationalizing breakthrough after many years of experience with policy implementation, or through a process of punctuated equilibrium. While many scholars and policymakers have argued that policymaking can and should be more rational, or that nonincremental alternatives may at times be superior to incremental ones, implementing nonincremental policy departures poses special problems and often gives way to incrementalism in the administrative process as public attention and support for strong action wanes. Nonincremental policy departures are more likely to be both enduring and effective where long experience with an issue leads to consensus on values and an adequate knowledge base, giving rise to a rationalizing breakthrough.

Keywords: incrementalism, partisan mutual adjustment, seriality, public satisfying, punctuated equilibrium, dramaturgical incrementalism, life cycle of issues, rationalizing breakthrough, decrementalism

Incrementalism as an Alternative to the Synoptic Ideal

Incrementalism is a model of policymaking originally advanced by Charles Lindblom (1959, 1965; see also Dahl & Lindblom, 1963, and Braybrooke & Lindblom, 1963). Lindblom advanced it as both a descriptive and a normative model of the policy process: descriptive in the sense that it provides the best account of how policies are actually made under most circumstances, and normative in the sense that it provides a superior strategy for policymaking under normal conditions. In Lindblom’s view, policymakers resort to incrementalism for most issues because it is the best way to make policies.

Lindblom’s normative case for incrementalism always begins with a critique of the rational-comprehensive (or synoptic) ideal. Rational policymaking begins with agreement on objectives and then examines alternatives, choosing the best. Ideally, all alternatives would be considered and the consequences associated with each alternative would be specified accurately, or at least probabilistically. Decision makers make a value-maximizing choice by selecting the alternative with the best consequences. If there is a budget constraint, then the option is chosen that provides the highest affordable level of utility.

Lindblom attacks the rational-comprehensive ideal from two directions. First, agreement on values is typically unattainable. Rationality may be conceived as focused on the attainment of a single goal or value (Dryzek, 1983); viewed in this way, the choice of which objective to pursue is part of the decision process. While examining alternatives to see which one maximizes a single objective may be characterized as purposeful, it is not rational in the stronger sense in which Lindblom defines the term. For Lindblom, agreement on objectives means agreement on how to weight or trade-off multiple values. This implies a system of social indifference curves that express society’s preferences for various values. Kenneth Arrow’s famous impossibility theorem (1963) shows that social indifference curves cannot be constructed reliably for three or more values and three or more decision makers.

Clearly, defining agreement on objectives as agreement on how multiple values should be weighted or balanced makes it much less likely that this precondition for rationality will be satisfied. While focusing on a single objective may result in more purposeful policymaking, it achieves coherence by neglecting competing objectives. Objectives do not go away or become unimportant when they are neglected, and sooner or later ignoring them will result in serious problems that policymakers will need to address. So Lindblom is convincing both in defining consensus on objectives in this more demanding way and in concluding that policymakers are unlikely to achieve a real consensus on multiple values that satisfies his requirement.

The second line of attack focuses on the near impossibility of specifying accurately the consequences for policy alternatives. Forecasting consequences accurately requires a very high level of understanding of cause-and-effect relationships among pertinent variables that is unlikely to be attained for most issues. To the extent that policymakers lack reliable estimates of the consequences associated with each alternative, it is impossible to say which alternative is value-maximizing, just as it is impossible to make a value-maximizing decision where there is no consensus on what value(s) should be maximized. Lindblom sees these two conditions—failure to agree on objectives and inadequate knowledge of policy consequences—as operative for the vast majority of policies. Under such circumstances, in his view, incremental policymaking is both inevitable and superior to misguided attempts at synoptic rationality.

Multiple Meanings of Incrementalism

Unfortunately, incrementalism means different things to different people, which can lead to confusion. First, incrementalism may be defined as a strategy for making the best possible policies under severe constraints (e.g., conflict over values and inadequate information). The primary emphasis here is on making the best possible use of available information by building on past policies, restricting focus to alternatives that differ only incrementally from previous policies, and striving to attain significant policy change through a succession of small steps (a process Lindblom calls seriality).

Second, incrementalism can be viewed as a characteristic of policy outcomes. The tendency for policymakers to focus on incremental alternatives, as described above, in combination with the necessity for bargaining and compromise, virtually guarantees incremental outcomes under most circumstances. In the American context at least, checks and balances built into the political process strongly reinforce this tendency toward incremental outcomes (Hayes, 2001, 2006; Jones, 1974, 1975). While proponents of incrementalism see incremental outcomes as desirable under most circumstances (see Nathan, 1976, for one example) and view large change as possible through seriality, critics in both the political and scholarly realm view incremental outcomes as inadequate and prefer to achieve major policy change through nonincremental policy outcomes (Broyles & Falcone, 1996).

Finally, incrementalism can be understood as a descriptive model of how policies are made under normal circumstances. Focusing on incrementalism as a strategy for more effective policymaking (definition 1) creates the false impression that incrementalism is one approach to policymaking among many—an approach that one can take or leave. By contrast, incrementalism as a descriptive model of policymaking (definition 3) suggests that there is no viable alternative to incremental policymaking under normal circumstances. Whether incrementalism works well or not, and whether it is desirable or not, it is more or less inevitable where conflict over objectives is combined with inadequate knowledge of policy consequences. Moreover, where the conditions for rational decision-making are unmet, as they are most of the time, there is no defensible intellectual basis for saying what the best policy is in any given case. Critiquing incremental outcomes as inadequate assumes that there is some analytical basis for saying that some alternative outcome is preferable, and while a given nonincremental outcome will be preferable to some participants, it will be unacceptable to others and will also be associated with consequences that are unpredictable. Under such circumstances, one can only hope that good outcomes will result from a good process and then try to specify the conditions that must be met for the policy process to function well.

The fact that incrementalism is employed by different people to mean at least three different things can create considerable confusion. In order to clarify our understanding of this important concept, four questions are addressed here. First, what are the characteristics of normal incrementalism as identified by Lindblom, and what conditions must be met for this process to yield “good” outcomes? Second, are incremental outcomes restricted to this process of “normal” incrementalism, or can they occur under a variety of different policy processes? Third, if incremental outcomes are prevalent under a variety of distinct policy processes, under what circumstances, if any, are nonincremental policy departures possible? Fourth, are nonincremental policy departures really desirable, or is Lindblom correct in arguing that better policies result from an exclusive focus on incremental alternatives?

Core Elements of Normal Incrementalism

What are the characteristics of normal incrementalism? As advanced by Lindblom, incremental policymaking has several signature elements (see especially Lindblom, 1959, and Braybrooke & Lindblom, 1963).

  • Remediality. First, policymaking is remedial rather than utopian. Policymakers attempt to mitigate or ameliorate problems, rather than pursuing abstract, positive values.

  • The adjustment of objectives to policies. There is no attempt to secure agreement on objectives before reviewing alternatives. Rather, policymakers consider means and ends simultaneously. Different alternatives represent different value mixes, so the choice of values for each participant occurs through the decision to support one alternative over another, and there is no presumption that all participants will support the same alternative or that those who do support that alternative will all do so for the same reasons.

  • Social fragmentation of analysis. Policymaking is pluralistic rather than rational. No single participant has sufficient information to make a rational decision; to the contrary, information is dispersed throughout the system, and different participants have different self-interests, organizational interests, and perspectives on the public interest.

  • Focus on incremental alternatives. Participants will limit their focus to policy options differing only marginally (or incrementally) from existing policy. Incremental policy proposals will be more feasible politically than nonincremental policies, and the consequences of incremental policy changes will be easier to predict accurately inasmuch as they draw on past experience with existing policies.

  • Marginal analysis or margin-dependent choice. Policymakers do not compare alternatives comprehensively. Rather, they focus on how various alternatives differ at the margins. Here Wildavsky’s analysis of the budget process (1964, 1984, 1992, 2006) provides a good example. In the traditional budget process that operated prior to 1974, agencies could assume their base would be funded; analysis focused on how much of an increment to the base each agency should receive.

  • Partisan mutual adjustment. Policymaking is not a process of analysis followed by decision but rather a process of interaction among multiple actors with different information, interests, and perspectives. Participants anticipate one another’s moves and maneuver to secure the best outcome they can for themselves by adapting to one another’s moves.

  • Seriality. The focus on incremental alternatives combines with the social fragmentation of analysis, margin-dependent choice, and mutual adaptation to virtually ensure incremental outcomes. This does not mean that large policy change is precluded, however. Rather, significant changes in policy may occur over time through subsequent policy cycles. In different works Lindblom terms this process seriality, successive approximations, convergence, and iteration. In a retrospective article on incrementalism twenty years after his original formulation, Lindblom (1979) observed that runners advance faster through a rapid succession of relatively small steps than they would by a series of larger hops or jumps. It follows, according to Lindblom, that the real question for policymakers is not whether to operate incrementally, for there really is no alternative, but rather whether incrementalism operates properly to permit seriality to yield a succession of small steps. Policymaking through trial and error cannot work if error correction is precluded by political inequality, partisan gridlock, or other features.

Conditions for Efficient Operation of Incrementalism

Lindblom argues consistently that incrementalism will operate to yield “good” public policies. While his model is meliorative and not utopian, he sees it as consistently producing better outcomes than would be attainable through an attempt at synoptic rationality. But is he right? What conditions must be met for incrementalism to perform well?

Just like economic markets, incrementalism has its own efficiency conditions that must be met to ensure efficient operation (Hayes, 2006, pp. 23–27). First, all affected interests must be effectively represented; in Lindblom’s memorable phrase, “every important social interest must have its watchdog.” Unfortunately, research on interest groups shows just how naïve this assumption is. The free rider problem (Olson, 1970) makes it especially difficult for large, diffuse groups to mobilize; the more general and public the interest, the less likely it will be to mobilize (Krier & Ursin, 1977, p. 268). Potential members will have a harder time affording membership dues during hard times, when group formation is needed most, and it will be much harder to mobilize the disadvantaged than the affluent (Hayes, 2006; Salisbury, 1969). At best, we can expect many interests to go unmobilized or imperfectly mobilized on most issues.

Second, power must be distributed across groups evenly enough to prevent any one group or coalition from achieving the political equivalent of market power. This does not mean that all interests must have identical resources or even equal resources. It would be more accurate to say, following Dahl (1962) that inequalities in resources should be dispersed rather than cumulative. Where inequalities are dispersed, a group with legitimacy may offset the power of a group with money or a group with expertise may prevail over a group with superior numbers. When, later in his career (1977, 1982), Lindblom argued that corporations occupy a privileged position within capitalist societies, he was arguing essentially that inequalities tend to be cumulative within capitalist societies, as business groups possess advantages in legitimacy, strategic position, and expertise (as well as money) that give them an effective veto over proposals they see as threatening to corporate interests (see also Miliband, 1969, and Hayes, 2006, pp. 63–79).

Where these first two requirements focus on interest groups—all interests must be represented and political resources must be dispersed rather than cumulative—a third condition focuses on the characteristics of political parties. According to V. O. Key, Jr. (1964, pp. 222–227), political parties must be moderate and pragmatic, not ideological and polarized. A two-party system like that operative in the United States works best, in his view, where it rests on an underlying consensus regarding fundamentals—particularly the nature of the political and economic system, but also rules and norms governing how authority will be exercised and how policy changes may be legitimately adopted. Key characterized the American system as one of “dualism within a moving consensus” (1964, p. 222). While the parties are rooted in different constituencies and thus adopt different policy stances, over time they will tend to converge to the same point on any given issue. One party will assume the role of innovator, while the other typically resists change, but new policies that prove effective and popular—no matter how controversial they were initially—will eventually become settled policy. Thus the consensus evolves over time as “the growing edge of the political system moves on to new fields where new problems await solution and invite conflict” (Key, 1964, p. 224). Ideally, this consensus keeps partisan conflict within tolerable limits.

Where this consensus on fundamentals is lacking, by contrast, parties may polarize, taking on extreme viewpoints. While incremental outcomes may be attainable under favorable circumstances, all too often the result will be gridlock and an inability to address pressing public problems—confrontational politics and instability rather than the kind of seriality through which policies gradually evolve, improve, and become accepted by the broader public (Key, 1964, p. 227). Incremental politics is almost surely undermined by the combination of cumulative political inequality and ideologically polarized political parties found in the United States today.

Conditions for Rationality as Variables, Not Constants

As noted earlier, Lindblom’s case for incrementalism always begins with the breakdown of rationality. He sees incrementalism as virtually inevitable because policymakers fail to agree on objectives and lack understanding of policy problems sufficient to forecast accurately the consequences of various alternatives. While Lindblom is probably right that these two adverse conditions are operative most of the time, he is wrong to treat them as constants rather than variables. In this regard, Hayes (2001, 2006) distinguishes between consensual and conflictual objectives as well as between consensual and conflictual knowledge. Contrary to Lindblom, policymakers may achieve consensual objectives for a variety of reasons—for example, a failure to consider alternative goals through a lack of imagination or because a dominant subgovernment or “policy monopoly” is limited to a restricted set of participants with shared goals. Similarly, policymakers sometimes agree on the basic facts of a policy problem—what variables are important and how they are causally related. Rothstein (1984) terms this exceptional situation “consensual knowledge.” The more normal case, where the consequences of various alternatives are contested by policymakers, may be termed “conflictual knowledge.”

These two dichotomous variables may be cross-tabulated to yield four distinct cases. Normal incrementalism, Lindblom’s predominant focus as described in the previous section, is inevitable whenever policymakers lack agreement on objectives and a knowledge base adequate to predict consequences. Where the knowledge base is adequate but policymakers disagree over objectives, the result is a pure problem of value disagreement. A third distinct policy process results when policymakers achieve consensus on objectives but lack the knowledge base necessary to solve problems. This situation may be characterized as a pure problem of knowledge base. The final case combines consensual objectives with consensual knowledge. Here rational decision-making is attainable.

Incremental Processes vs. Incremental Outcomes

While these four distinct policymaking environments give rise to very different patterns of policymaking, three out of the four typically give rise to incremental outcomes, suggesting that it is important to distinguish between incrementalism as a description of the policy process and incrementalism as a characteristic of policy outcomes. While there have been no empirical studies establishing the frequency with which policies fall into the various categories, Lindblom is probably correct in arguing that the conflictual objectives/conflictual knowledge combination is by far the most common, and he is almost surely right to argue that what we have here called “normal incrementalism”—e.g., social fragmentation of analysis, restriction of analysis to incremental alternatives, partisan mutual adjustment, seriality, and so on—is virtually inevitable within this cell.

Where conflictual objectives combine with consensual knowledge, the result is a pure problem of value disagreement. Here we may expect that conflicts will be more intense than those occurring in the realm of normal incrementalism because the adequacy of the available knowledge base removes the uncertainty over consequences that normally operates to temper conflicts. The intensity of conflict over the abortion issue likely stems from the purity of the value disagreement: the issue is not how fetuses develop at various stages but rather whose rights will be paramount—the fetus or the pregnant woman. While conflicts will be considerably more intense here than within the realm of normal incrementalism, policy outcomes will likely be incremental, if indeed it is possible to produce any policy at all.

Policymaking will be very different for pure problems of knowledge base (consensual objectives/conflictual knowledge). Here policymakers agree on objectives (the desire to maintain full employment with price stability, for example) but lack the knowledge base necessary to achieve their shared goals. Hayes (2006) suggests that policymaking within this cell will be cybernetic, as policymakers try to steer policy in a positive direction in response to changing environmental conditions. Outcomes will once again tend to be incremental, reflecting the limited knowledge base, but seriality should operate most efficiently within this cell inasmuch as gradually improving policy over time through trial and error is more likely where all the participants agree on the desirability of the underlying goals. Where, by contrast, at least some participants oppose the objectives of a policy, subsequent policy cycles may focus more on attempts to repeal or weaken a policy rather than improve it.

In the final category, where consensual objectives are combined with consensual knowledge, the policy process will look more like a synoptic analysis and less like a pluralistic process of social interaction (partisan mutual adjustment). Under such circumstances, policy outcomes may be incremental of course, but they need not be. Agreement on objectives and knowledge base frees policymakers to make whatever policy changes they deem appropriate, including nonincremental ones. This does not necessarily mean that Lindblom’s strong conditions for rationality have been met; policymakers may agree on goals without having adequately explored alternative weightings, and consensual knowledge may subsequently prove to have been inaccurate on one or more points (Rothstein, 1984). For example, consensual knowledge may manifest itself in a shared paradigm that eventually leads to policy failures that it cannot explain. Under such circumstances, policymakers may embrace a new paradigm (e.g., shifting from Keynesian to monetarist economic theories). Paradigm shifts of this sort may constitute one form of nonincremental change (Hall, 1993).

As noted earlier, incrementalism can be understood in at least three different ways—as a description of how policies are made, as a characteristic of policy outcomes, and as a strategy for making better policies. Examination of these four policymaking environments shows clearly the importance of differentiating between the idea of incremental outcomes and incrementalism as a policy process. Normal incrementalism is only one of four possible modes of policymaking, but incremental outcomes are the most likely result any time at least one of the conditions for rationality has broken down. Viewed within this framework, nonincremental outcomes are most likely to occur (and most likely to be enduring and effective) when the conditions for rational decision-making are both met.

Defining Nonincremental Outcomes

Under what circumstances are nonincremental policy changes possible? Lindblom saw nonincremental policy outcomes as exceedingly rare. But what qualifies as a nonincremental change exactly? Braybrooke and Lindblom (1963) saw policy change as running along a continuum from incremental to nonincremental; within this scheme, nonincremental changes differ from incremental ones only by degree, and there is no clear point at which such changes can be definitively classified as nonincremental. Analyses of budgetary outcomes as incremental or nonincremental tend to come at the question in this way: relatively small increases or decreases are considered incremental, while larger ones are characterized as nonincremental.

In a subsequent work, by contrast, Lindblom (1977) equated nonincremental change with systemic transformation—for example from capitalism to socialism or democracy to dictatorship. By this reasoning, policy changes short of systemic transformation (like the creation of a single payer national health care system, for example) would merely count as “large increments.” As Hayes (2006) has noted, this definition is self-serving: Lindblom supports his hypothesis that most policy change is incremental by defining what most of us would consider to be significant policy changes as merely “large increments” rather than nonincremental change. Whether we call them large increments or nonincremental, however, some policy changes are clearly more significant than others and their occurrence needs to be explained.

Clearly, equating nonincremental change with systemic transformation lumps together under one incremental rubric policy changes that are really quite distinctive in size and scope. By contrast, the Braybrooke and Lindblom treatment of policy change as a continuum from incremental to nonincremental is adequate for understanding many policies. Proposals to raise the minimum wage, for example, may range all the way from merely restoring value lost from inflation to proposals for a genuine living wage.

However, it may be possible to define policy change with a bit more precision. In a study of policymaking in Dutch departments, Hoppe, van de Graaf, and Besseling (1995) measured the degree of policy change using six criteria: whether the nature of policy goals changed, whether the size of the target group(s) increased or decreased significantly, whether priorities among target groups were reset, whether the size of an agency’s budget increased or decreased substantially, whether the scope of government responsibility increased or decreased substantially, and/or whether two or more of the foregoing occurred simultaneously. In a similar vein, Hogwood and Peters (1983) differentiated among four different forms of policy change: policy maintenance, policy termination, policy succession, and policy innovation. Policies may be incrementally adapted or expanded under policy maintenance, but no new organizations will be created and no changes in legislative purpose will be adopted. Policy termination involves repeal of legislation and/or abolition of administrative organizations. Policy succession occurs when an established policy is replaced by a new one. This may take various forms: termination of an old policy and replacement with a new one, consolidation of existing programs, or splitting an existing program into two or more independent programs, for example. Finally, policy innovation involves creation of a brand new policy; it does not build on existing institutions, programs, our outlays, but rather establishes a new entity with a new and different purpose of its own. This is similar to the distinction Brown (1983) makes between breakthrough and rationalizing policies.

Within the confines of Lindblom’s terminology, we may say that policy maintenance is clearly incremental, replacement of terminated policies with new policies is at the very least a large increment, and policy innovation is clearly nonincremental. Hogwood and Peters see their classification scheme as offering a much more nuanced treatment of the forms of policy change, however. Their scheme does indeed offer a potential advance over the dichotomous distinction between incremental (small or marginal) and nonincremental (large or significant) policy change, and analysts may find it fruitful to shift to their more complex formulation.

Lindblom’s Conditions for Nonincremental Change

In a work co-authored with David Braybrooke (Braybrooke & Lindblom, 1963), Lindblom advanced a four-quadrant model, with nonincremental outcomes confined to one of the quadrants. The quadrants were derived from the interaction of two continuous dimensions. The first dimension ranged from incremental change at one end to nonincremental change at the other extreme. The second focused on the degree to which policymakers understood the problem at hand; this dimension ranged from low to high understanding. Rational decisions were confined to small, technical issues in the high understanding/incremental change quadrant. Normal incrementalism occurred within the low understanding/incremental change quadrant. The quadrant combining high understanding with nonincremental change was empty, according to Lindblom, because large policy changes would necessarily be characterized by low understanding. Major policy departures were confined to the low understanding/nonincremental change quadrant, which he termed the “realm of wars, revolutions, and grand opportunities.” Instances of large change would occur despite the low understanding of the problem at hand because some external event (like a war or revolution) made dramatic change possible or because policymakers felt existing policy was too badly broken to build on. In such cases, policymakers sometimes find it rational to take a “calculated risk,” moving beyond the available knowledge base (Dahl & Lindblom, 1963).

Public Satisfying

In a study of air pollution policymaking from 1941 to 1970, Charles O. Jones (1974, 1975) extended Lindblom’s fourth quadrant to include nonincremental policy responses to an aroused mass public opinion. Throughout most of the period Jones examined, policymaking was characterized by what he called “majority-building incrementalism.” Policymaking was pluralistic and conflictual, as failure to agree upon the public interest led to a conflict among private interest groups. Mass public opinion was relatively quiescent, and policy development was confined to a few interested members of the air pollution policy community. Because of the need to bargain and compromise, policy proposals were “tapered down from the optimal to the acceptable,” and incremental outcomes were the norm.

Between 1967 and 1970, however, mass public concern over air pollution grew dramatically, manifesting itself in both public opinion polls and the formation of a wide variety of new environmental groups. This mass public arousal created what Jones called a “pre-formed majority.” It was no longer necessary to build a majority in order to enact legislation; rather, it was necessary for policymakers to appease a mass public demanding strong action. As a result of this pre-formed majority, the number of active participants expanded dramatically, drawing in legislators who were not part of the policy community, many of whom had taken little interest in the issue before. Rather than tapering down, a process of “policy escalation” took place, as legislation grew stronger (not weaker) at every stage. The end result was a nonincremental policy departure that Jones referred to as “public satisfying speculative augmentation”: a significant expansion in federal capacity and responsibility that went beyond the available knowledge base.

Dramaturgical Incrementalism

Hayes (1987) applied Jones’s public satisfying model to the nuclear freeze issue. In the early 1980s, mass public concern over the level of nuclear armaments and the potential for an arms race grew dramatically. As with the air pollution issue, this mass public arousal manifested itself both through public opinion polls and through the mobilization of a wide variety of new groups opposed to nuclear arms. There was also a comparable expansion in the scope of conflict, as members of the foreign policy and arms control community were joined by a wide variety of new participants interested in capitalizing on this highly emotive issue.

The resemblance between the two issues stopped there, however. Rather than policy escalation, there was instead a classic pattern of tapering down from the optimal to the acceptable, and the ultimate outcome—a very weak freeze resolution passed by only one house of Congress—could hardly be characterized as nonincremental. Rather, the issue conformed to a new model of policymaking, which Hayes (1987) termed “dramaturgical incrementalism.” Enduring conflicts among specialists within the policy community necessitated normal incrementalism, even as mass public arousal precluded the kind of bargaining and compromise incrementalism requires. The result was a dual conflict. The freeze issue was highly public and featured a wide range of participants; the end result was a highly symbolic freeze resolution that permitted both arms control proponents and opponents to claim victory. The second conflict was much more private and featured a narrow range of specialists within Congress negotiating with President Reagan out of public view. That conflict featured a classic horse trade: the president won authorization of the MX missile in return for his pledge of support for a new single-warhead missile (the midgetman) that arms control proponents in Congress wanted. In effect, the highly public and symbolic dispute over the freeze resolution was a classic case of misdirection, drawing public attention away from the more mundane and technical compromises that had to be agreed to by members of the policy community.

Hayes subsequently reexamined the air pollution issue (2001, pp. 72–98), suggesting that the 1970 clean air legislation might be better understood as dramaturgical incrementalism rather than public satisfying as described by Jones. What made dramaturgical incrementalism possible for the freeze issue was the parallel development of the issue on two different tracks. While the freeze issue moved through the House Foreign Affairs Committee, activating the normal foreign policy community, the MX missile was considered as part of the annual defense authorization bill, which fell under the jurisdiction of the House and Senate Armed Services Committees. Although the defense authorization process was the avenue through which new weapons systems, like the MX, received formal approval, it was much less salient for the mass public than the struggle over the nuclear freeze issue. On the air pollution issue, a similar two-track process existed, although the two tracks did not operate simultaneously as they had for the freeze issue. For the air pollution case, the legislative process may be understood as providing symbolic reassurances to an anxious mass public (Edelman, 1964) through passage of a new law that embraced technology-forcing, health-based air pollution standards. Conflict within the environmental policy community necessitated normal incremental bargaining, but such bargaining was largely precluded by the pre-formed majority and the need for public satisfying through the appearance of strong action. The subsequent implementation process took place at the state level (more accurately via negotiations between EPA and each of the individual states) over a very long period of time, long after public arousal had waned. Conflicts between industry and environmentalists—and among energy, environmental, and economic growth imperatives—inevitably resurfaced. Not surprisingly, policy evaluations of the impact of the 1970 Clean Air Act pointed to incremental progress despite the appearance of nonincremental policy change in the legislation itself.

Taken together, these case studies suggest that mass public arousal can transform the legislative process, creating a pre-formed majority, triggering an expansion in the scope of conflict, and necessitating a process of policy escalation resulting in at least the appearance of nonincremental legislation. Whether that legislation ultimately has a nonincremental impact depends on the subsequent implementation process, however, and that administrative process will almost surely not take place within the same atmosphere of public arousal. When mass public quiescence is restored, apparent consensus on the nature of the public interest evanesces and the classic pattern of conflict among private interests reemerges, leading to bargaining, compromise, and incremental progress.

Punctuated Equilibrium

Fenno (1966) and Wildavsky (1964, 1984, 1991, 2006) found the traditional (pre-1974) budgetary process to be quintessentially incremental. Appropriators built on past policy, accepting agencies’ budgetary bases as sacrosanct, focusing on incremental increases, and comparing increments at the margins. Bailey and O’Connor (1975) called these analyses into question, however, finding that a significant fraction of all budgetary outcomes could hardly be characterized as incremental. About a third of all outcomes changed spending between 11% and 30%, while another 15% of outcomes changed spending levels by more than 30%.

More recently, Jones and Baumgartner found budgetary outcomes to be leptokurtically distributed, rather than normally distributed. More specifically, while the fraction of incremental spending changes was even higher than one would expect if outcomes were normally distributed, the tails of the distribution were also “fatter,” implying that genuinely nonincremental spending changes (both up and down) were also higher than one would have expected if outcomes were distributed normally. Jones and Baumgartner (2009) explained this pattern through a “punctuated equilibrium” model of policymaking. In short, new government programs quickly degenerate into policy monopolies, fostering incremental outcomes. When these policy monopolies break down through some combination of external events and media exposure, issues are reframed and politics becomes more pluralistic, leading to nonincremental policy changes. These new policy communities eventually stabilize as well, leading to a new policy equilibrium and incremental policy outcomes, albeit very different ones from those emanating from the original policy monopoly. Baumgartner and Jones provided a convincing mix of qualitative and quantitative evidence in support of this theory (Baumgartner & Jones, 2009; Jones & Baumgartner, 2005; Jones, True, & Baumgartner, 1997), and some scholars have come to see incrementalism as a special case within a broader and more comprehensive punctuated equilibrium model (Howlett & Migone, 2011).

If Baumgartner and Jones are correct in arguing that new policies quickly develop policy monopolies that generate incremental outcomes consistently, with exceptions occurring primarily during periods of punctuated equilibrium, what Lindblom called “disjointed problem-solving” (1965, pp. 178–181) must override pluralism under normal circumstances. While many (albeit not all) interests may have their watchdogs, this representation will occur through the development of a multiplicity of disconnected policy subsystems, and there is no guarantee that the overall pattern of disconnected outcomes will yield a “good” policy mix.

Thus, Baumgartner and Jones find not only that departures from incrementalism through punctuated equilibrium are more common than Lindblom expected, but also that incremental outcomes are typically associated with policy monopolies while nonincremental policy changes result from the kind of pluralism Lindblom consistently associated with incrementalism. An interesting question for further research here would be why policy monopolies—in which the conditions for rational decision-making are met by the exclusion of interests dissenting from shared goals and consensual knowledge—would exhibit only incremental policy outcomes when the potential for spending growth would seem to be unconstrained by the presence of opposition groups. In Hayes’s (2006) view, nonincremental policy changes are more likely to occur when a broad range of groups is already mobilized on an issue, creating a pluralistic pattern of politics, and a disturbance in the group equilibrium (Truman, 1951) leads to the entry of new groups, shifting the balance of forces in a positive direction, making stronger legislation possible. This is essentially what happened between 1967 and 1970 with the air pollution case Jones examined (1974, 1975).

Are Nonincremental Policy Changes Desirable?

As noted earlier, Lindblom argues that focusing exclusively on incremental policy alternatives will ultimately lead to better policymaking. This is Lindblom’s answer to the problem of information costs and the costs of search. Where satisficing policymakers set an aspiration level and search through alternatives randomly until they find one that meets the aspiration level, incrementalism limits analysis to a subset of alternatives that are politically feasible and for which information is most reliable. While large policy changes, for which information on consequences is unreliable, are off the table, significant policy evolution may result over time through the steady accumulation of small steps. While incremental change over time may merely lead to reproduction through adaptation, it may also lead to a significant transformation (Streek & Thelen, 2005).

Some scholars have questioned the superiority of focusing on incremental alternatives. Levmore (2010) defines incrementalism as a stealth or “camel’s nose” strategy in which an initial incremental step is proposed while proponents secretly plan to follow with additional increments, since a package proposal incorporating all the desired elements would be impossible to enact. When the likely reaction of interest groups affected by the initial policy is taken into account, there is no guarantee that the succession of subsequent policy increments will lead to optimal policies. As we have already seen above, group theory is quite adamant in telling us that we can’t expect every important social interest to have its watchdog, contrary to Lindblom’s expectation, so the fact that interest group activity may lead to imperfect policy outcomes comes as no great revelation. To be fair to Lindblom, however, he never advances incrementalism as a utopian model that will always yield optimal outcomes. Rather, he merely argues that incrementalism will consistently yield better outcomes than any attempt at synoptic rationality, which he sees as unattainable most of the time. In this vein, Lindblom would almost surely ask by what criterion Levmore can claim that some outcomes are suboptimal where value conflicts and limited knowledge of policy consequences render rational analysis impossible.

Another critic of incrementalism (Bendor, 2010) employs formal mathematical analysis to identify conditions under which expanding analysis to include nonincremental alternatives may lead to choices that are at least as good as, and sometimes better than, decisions where only incremental alternatives are examined. Bendor focuses first on strategies employed by a single decision maker and then turns his attention to cases of multiple decision makers, differentiating cases where the various decision makers share the same goals from cases where the decision makers disagree on goals. Unfortunately, Bendor treats the latter case (multiple decision makers who disagree on goals) as one in which the different participants care about different aspects of the problem rather than examining what happens when participants sharply disagree over the same goals or how different goals should be traded off. Where different participants merely specialize in different aspects of the problem, logrolling and majority voting will lead to good outcomes; in fact, Bendor asserts (following Condorcet) that the odds of arriving at a good decision will go up consistently as the size of the decision-making group increases. Where participants disagree intensely over the same objectives (pro-choice vs. pro-life, for example) or over how multiple objectives should be weighted—which are the very cases Lindblom emphasizes—it is hard to imagine how the odds of making good decisions (or indeed any decision at all) would go up with the size of the decision-making body. Arrow’s paradox would seem more pertinent here than Condorcet.

Other critics have pointed to potential “sleeper effects” resulting from incrementalism (Weiss & Woodhouse, 1992). Sleeper effects are negative policy consequences that only manifest themselves late in the process when it has become difficult or impossible to change course. In response, it is hard to see how any process (including synoptic analysis) could avoid problems that are, by definition, beyond the available knowledge base (Hayes, 2013). A more telling critique might center on sunk costs: it is hard to walk away from a failed policy that represents a significant investment of resources over time.

Finally, some critics charge that incrementalism has a conservative bias stemming from its focus on past policies and its rejection of radical policy changes. While incrementalism may be employed to pursue liberal policies, as Popper (1994, pp. 91–92) observed, incrementalism can be legitimately characterized as “a process of least steps taken down the path of least resistance” (Krier & Ursin, 1977, pp. 252–257). This process is clearly imperfect: the more truly collective or public a problem is, the harder it will be to mobilize proponents; many interim solutions will be adopted for political reasons even though rational analysis suggests they have no chance of working; the inadequacy of incremental steps will often lead to dramatic performance failures that manifest themselves in crises; and progress will be agonizingly slow at best. However, as Krier and Ursin (1977, pp. 257–263) observe, the problem is not so much with incrementalism as with uncertainty and political inequality; it is just inherently difficult to solve complex public problems under the best of circumstances. Crises serve a positive function by highlighting the inadequacy of previous steps and galvanizing public opinion (Krier & Ursin, 1977, pp. 263–277). More generally, seriality proceeds less through a process of positive learning than via a process they call “exfoliation,” through which interim policies are exposed as inadequate through their very performance failures. In short, as observed above, the breakdown of rationality makes incrementalism inevitable much of the time. While this process works imperfectly, there is simply no alternative to it under normal circumstances.

Implementing Policy Beyond Capability

Jones’s (1974, 1975) analysis of the 1970 Clean Air Act, reviewed earlier, provides a valuable case study of a nonincremental policy outcome. In that case, an aroused mass public led to a bandwagon effect and policy escalation, resulting in a legislative outcome Jones characterized as “public satisfying speculative augmentation”—a significant expansion in federal authority that was ultimately speculative to the extent that it exceeded the available knowledge base. For this reason, Jones also described the policy process in that case as “legislating beyond capability.” In particular, some of the provisions of the new law were explicitly technology-forcing, as they required industries to meet standards that were unattainable through the technologies then available. The eventual result, not surprisingly, was dramaturgical incrementalism, as argued above. Writing shortly after the passage of the law (1975, pp. 272–273), Jones emphasized the difficulty of “implementing beyond capability” in the short run; the first order of business was, necessarily, reorganization of existing institutions, increased staffing, and redirecting organizational energies—all extremely difficult tasks that necessarily took attention away from solving concrete pollution problems. Inasmuch as this organizational transformation would take a very long time, success would hinge ultimately on how long public arousal on the issue persisted, resulting in pressure on both legislators and administrators to produce results that were real and significant rather than merely symbolic. If public pressure waned fairly quickly—the most likely scenario—normal incrementalism would resume as trade-offs between energy and the environment or between environmental protection and economic growth reasserted themselves. This is precisely what happened, as Hayes (2001) demonstrated.

Rationalizing Breakthroughs and the Life Cycle of Issues

The clean air case strongly suggests that nonincremental policy outcomes occurring in response to mass public arousal will revert to normal incrementalism in the implementation phase as public pressure wanes. Much the same could probably be said for large policy changes resulting from presidential elevation of an issue. Hayes (2001, pp. 99–122) showed that presidential leadership typically fails to produce the kind of mass public arousal and policy escalation exhibited in the clean air case. Rather, normal majority building is still required and incremental outcomes are the best that can be expected most of the time. Even where policy change takes the form of a large increment or even a nonincremental change, it is hard to see how erosion of the policy could be avoided in the administrative process, inasmuch as mass public pressure resulting from presidential leadership tends to be weaker than the kind of pre-formed majority Jones found, and public pressure would likely wane with the passage of legislation and the shift of national policymakers’ attention to other issues.

Nonincremental policy changes are most likely to endure, and to produce effective policy outcomes, where the conditions for rational decision-making have been met. In this vein, Hayes (2001, 2006) identified a life cycle of issues that built on Brown’s (1983) distinction between breakthrough and rationalizing policies. Breakthrough policies represent instances in which the federal government takes on a new policy role or responsibility. Rationalizing policies tinker with previously established policies in order to remedy defects that have emerged through experience with implementation. In Hayes’s formulation, the life cycle begins with what he calls federal role breakthroughs. Following Brown, he sees issues as then moving into the realm of rationalizing policies. These rationalizing policies tend to be quintessentially incremental, building on past policies and making small changes in response to feedback and experience. This stage of incremental rationalizing policies may persist for years and may, in fact, be the final stage in the life cycle for many policies. For some policies, however, the steady accumulation of knowledge base through experience with implementation can lead to the emergence of consensual objectives and consensual knowledge. Where the conditions for rational decision-making emerge over time in this way, the result can be a rationalizing breakthrough, which Hayes defines as a new and different approach to an old policy problem—an original and innovative approach that makes it possible to address an enduring problem in a much more effective way. What Jones (1975) referred to as legislating beyond capability—writing legislation that went beyond the available knowledge base—can be understood as an attempt to impose nonincremental change before the conditions for a rationalizing breakthrough have been met.

To summarize, focusing exclusively on incremental alternatives may not always lead to better policies (Bendor, Levmore). However, where the conditions for rational decision-making (consensual objectives and consensual knowledge) are not satisfied, there is no basis by which particular outcomes can be definitively judged to be optimal or suboptimal. Moreover, nonincremental outcomes enacted in response to mass public arousal or presidential leadership call for demanding and unrealistic levels of organizational skill and commitment in the implementation stage, making some form of dramaturgical incrementalism all but inevitable when public attention wanes. Nonincremental policy changes are more likely to endure and to be effective where they constitute rationalizing breakthroughs coming in the final stage of the life cycle of issues. Rationalizing breakthroughs are rare examples of synoptic analysis made possible when long experience with a policy gives rise to consensual knowledge and consensual objectives. If Kingdon (1984) is correct, such rationalizing breakthroughs would emerge from what he calls the “policy stream” and would not reach the decision agenda automatically but rather would require favorable circumstances in his problem and political streams as well.

Policy Contraction

Beginning with the Nixon/Ford proposals for revenue sharing and block grants, policymaking has increasingly become decremental rather than incremental (Dommel, 1974), focusing on policy contraction rather than expansion. Inasmuch as they focus on modifying existing policies, decremental initiatives constitute rationalizing policies. The three primary forms decremental rationalizing policies are disengagement, decentralization, and deregulation. Disengagement involves what Hogwood and Peters (1983) called policy termination: the federal government withdraws from a policy area completely. Deregulation involves fostering competition among private actors instead of regulating prices or practices within an industry through an agency or commission. Decentralization involves devolution of responsibility for decision-making from the federal government to the states and/or localities, typically through block grants.

Budgeting in the decremental age has deviated from normal incrementalism on many points. Most obviously, the budget reforms of 1974 introduced new institutions (House and Senate budget committees, the Congressional Budget Office) and legislative requirements (budget resolutions, reconciliation). Equally important, however, the old process of bottom-up policymaking through appropriations subcommittees has given way to top-down negotiations among major players (the president, party leaders in both the House and Senate) seeking a “grand bargain” that will bring down spending dramatically. The focus is no longer on how well programs work, as evaluated by committee members possessing long experience with their operation, but rather on how large sums of money can be identified and eliminated to achieve large savings. While the 1974 reforms were intended to make budget decisions more rational, what has happened instead is that budget decision-making has become much more conflictual.

One might expect decremental budgeting to be the mirror image of incremental budgeting: where appropriators once accepted agencies’ bases as sacrosanct and focused on increments, now budgetary decision makers might start with an agency’s base and focus on decrements. Through such a process, deficits might be reduced a little bit every year, rather than through periodic budget wars in which some players threaten to drive the nation off a fiscal cliff rather than compromise.

Two explanations have been offered for the pathological quality of decremental budgeting. Wildavsky (2006) stresses the increasing polarization of the parties. In short, the rise of polarized parties reflects a breakdown in the kind of “dualism through a moving consensus” identified decades earlier by V. O. Key (1964). Where the parties disagree fundamentally on the role of government, do not treat established programs as settled policy, and do not even agree that government makes a positive contribution to society, there is no prospect for convergence to the kind of middle ground Key talked about. Under conditions of divided government—increasingly the norm in recent decades—it becomes impossible to write a budget all players will accept. At times it becomes impossible for the House and Senate even to agree upon a joint budget resolution. And, in Wildavsky’s famous phrase, if you can’t budget, you can’t govern (2006, p. 325).

Schick (1983) offers a different explanation. In his view, it is mistaken to view incremental budgeting as primarily about small changes (up or down). Rather, incremental budgeting is primarily about budgetary expansion. When the economy is growing strongly, as the American economy did from the end of World War II to the early 1970s, there is ample room within the economy for government to grow as well. Potential conflicts among agencies over who gets what can be accommodated by focusing on increments and letting everyone grow together. While some agencies may grow more than others in any given year, all are growing; their bases are all sacrosanct and the question becomes how big a given agency’s increment will be.

But when economic growth slows, there is no longer room in the economy to support incremental increases for all participants; budget policymakers are forced to contract spending, and the policy process becomes redistributive rather than distributive. As Lowi (1964) observed, redistributive policymaking is intensely conflictual, and reforming the budget process to make decision-making more rational and informed merely highlights the issue of who gets what. To the extent that budgets determine how scarce resources are allocated, budgeting is inherently redistributive. Where economic growth makes resources abundant, budgeting can be redefined as distributive. With the end of economic growth, slack resources are no longer available within the economy to fund incremental increases for all players and the question of who gets what reasserts itself with a vengeance.

Both explanations contribute to our understanding of the intensely conflictual nature of contemporary budgeting. Incremental budgeting is all about expansionary budgeting, and expansion requires economic growth; decremental budgeting focuses on who will suffer cuts and is thus unavoidably redistributive. When the policymakers confronting this incredibly difficult situation are intensely polarized, and there is no longer any fundamental consensus on the proper rules of government or even on what constitutes legitimate forms of policymaking, budgeting becomes a form of political warfare. The key point here is that decremental budgeting is not and can never be the mirror image of incremental budgeting. Incremental and decremental budgeting are fundamentally different processes featuring very different politics (Schick, 1983).


One’s response to incrementalism reflects underlying assumptions (or worldviews) about what it is possible to achieve in politics through rational analysis. Building on the earlier work of Friedrich Hayek (1948), Sowell (1987, 1999) makes a distinction between two very different visions of rationality. What Sowell calls articulated rationality is the use of synoptic analysis to identify comprehensive solutions to social or economic problems. By contrast, systemic rationality is the mitigation of problems through pluralistic social processes like economic markets or political systems with built-in checks and balances. Hayek (1948) called advocates of articulated rationality “rationalists” and defenders of systemic rationality “antirationalists.” Sowell uses slightly different terms for this same distinction, differentiating between what he calls the unconstrained vision and its more realistic counterpart, the constrained or tragic vision. Rationalists see no limits on the potential of human rationality to solve social problems, even going so far as to replace traditional moral systems with new rules (Hayes, 2001; Spicer, 1995). Because they see reason as a vehicle for problem-solving, rationalists view power as benign and favor centralized forms of decision-making that facilitate the translation of experts’ proposals into legislation. By contrast, antirationalists see man as both fallible and fallen. To the extent that man is fallible, good policies are more likely to result from the interplay of many views, and the information brought to the table by many different participants, than from the articulated rationality of any one individual, however brilliant. And because they view man as fallen, antirationalists see power as subject to abuse unless subject to some system of checks and balances.

Both Hayek and Sowell see rationalists as unrealistic, essentially utopian visionaries in their belief that complex social problems can be solved, once and for all, through rational analysis. Antirationalists, by contrast, view social problems as subject to mitigation rather than final solution and put their faith in good processes to yield good outcomes. Seen in this light, incrementalism is clearly a form of systemic rationality, and Lindblom’s critique of the synoptic ideal (e.g., articulated rationality) makes him a quintessential antirationalist. The criticisms of incrementalism reviewed earlier almost invariably start from the premise that policies can or should be made rationally, or they critique incremental outcomes by holding them up against some preexisting standard (like economic efficiency) that is rooted in rational analysis. But, as noted earlier, where policymakers disagree over objectives or lack the knowledge base necessary to estimate policy consequences accurately, there can be no rational basis for saying what a “good” outcome is.

What can be said with confidence, however, is that for incrementalism to yield good outcomes, certain efficiency conditions must be satisfied: all (or at least most) important social interests must be represented, resource inequalities must be dispersed rather than cumulative, and competing political parties must agree on fundamentals enough to permit convergence to an ever-changing political center. Where some or all of these conditions are not met, we can no longer trust the policy process to yield good outcomes.

As Lindblom wisely noted (1979), however, there is no viable alternative to incremental policymaking much of the time, making the real question whether incrementalism functions properly to permit policy improvements over time through seriality. Just as mainstream economists defend free markets but advocate government intervention to remedy market failures, antirationalist policy analysts (like Lindblom and this author) may defend incremental policymaking while recognizing the need to remedy political market failures. Within the United States, at least, cumulative resource inequalities and ideologically polarized parties threaten the proper functioning of incrementalism. Our best response—as always meliorative rather than utopian—is not to abandon incrementalism but rather to address its defects.

Further Readings

Allison, C. R., & Saint-Martin, D. (2011). Half a century of “muddling”: Are we there yet? Policy and Society, 30, 1–8 [electronic version].Find this resource:

Atkinson, M. M. (2011). Lindblom’s lament: Incrementalism and the persistent pull of the status quo. Policy and Society, 30, 9–18 [electronic version].Find this resource:

Good, D. A. (2011). Still budgeting by muddling through: Why disjointed incrementalism lasts. Policy and Society, 30, 41–51.Find this resource:

Lustik, I. (1980). Explaining the variable utility of incrementalism: Four propositions. American Political Science Review, 74, 342–353.Find this resource:

Saint-Martin, D., & Allison, C. R. (2011). Rationalism and public policy: Mode of analysis or symbolic politics? Policy and Society, 30, 19–27.Find this resource:

Wildavsky, A. (1979). Speaking truth to power: The art and craft of policy analysis. Boston: Little, Brown.Find this resource:


Arrow, K. J. (1963). Social Choice and Individual Values. 2d ed. New York: Wiley.Find this resource:

Bailey, J. J., & O’Connor, R. J. (1975). Operationalizing incrementalism: Measuring the muddles. Public Administration Review, 35, 60–66 [electronic version].Find this resource:

Baumgartner, F. R., & Jones, B. D. (2009). Agendas and instability in American politics (2d ed.). Chicago: University of Chicago Press.Find this resource:

Bendor, J. (2010). A model of muddling through. In Bounded rationality and politics (pp.61–92). Berkeley: University of California Press.Find this resource:

Braybrooke, D., & Lindblom, C. E. (1963). A strategy of decision: Policy evaluation as a social process. New York: Free Press of Glencoe.Find this resource:

Brown, L. D. (1983). New policies, new politics: Government’s response to government’s growth. Washington, DC: Brookings Institution.Find this resource:

Broyles, R. W., & Falcone, D. T. (1996). Health care reform: An American obsession with prescriptive incrementalism. Journal of Health and Human Services Administration, 18, 328–352.Find this resource:

Dahl, R. A. (1962). Who governs? New Haven, CT: Yale University Press.Find this resource:

Dahl, R. A., & Lindblom, C. E. (1963). Politics, Economics, and Welfare: Planning and Politico-Economic Systems Resolved into Basic Social Processes. New York: Harper & Row, Harper Torchbooks.Find this resource:

Dommel, P. R. (1974). The politics of revenue sharing. Bloomington: Indiana University Press.Find this resource:

Dryzek, J. S. (1983). Present choices, future consequences: A case for thinking strategically. World Futures, 19, 1–19.Find this resource:

Edelman, M. (1964). The symbolic uses of politics. Urbana: University of Illinois Press.Find this resource:

Fenno, R. (1966). The power of the purse. Boston: Little, Brown.Find this resource:

Hall, P. A. (1993). Policy paradigms, social learning, and the state: The case of economic policymaking in Britain. Comparative Politics, 25, 275–296.Find this resource:

Hayek, F. A. (1948). Individualism and economic order. Chicago: University of Chicago Press.Find this resource:

Hayes, M. T. (1987). Incrementalism as dramaturgy: The case of the nuclear freeze. Polity, 19, 443–463.Find this resource:

Hayes, M. T. (2001). The limits of policy change: Incrementalism, worldview, and the rule of law. Washington, DC: Georgetown University Press.Find this resource:

Hayes, M. T. (2006). Incrementalism and public policy. Lanham, MD: University Press of America.Find this resource:

Hayes, M. (2013). Incrementalism. In E. Araral Jr., S. Fritzen, M. Howlett, M. Ramesh, & Xun Wu (Eds.), Routledge handbook of public policy (pp. 287–298). London and New York: Routledge.Find this resource:

Hogwood, B., & Peters, B. G. (1983). Policy dynamics. New York: St. Martin’s.Find this resource:

Hoppe, R., van de Graaf, H., & Besseling, E. (1995). Successful policy formulation processes: Lessons from fifteen case experiences in five Dutch departments. Acta Politica, 2, 153–188.Find this resource:

Howlett, M., & Migone, A. (2011). Charles Lindblom is alive and well and living in punctuated equilibrium land. Policy and Society, 11, 53–62.Find this resource:

Jones, B. D., & Baumgartner, F. R. (2005). A model of choice for public policy. Journal of Public Administration Research and Theory, 15, 325–351 [electronic version].Find this resource:

Jones, B. D., True, J. L., & Baumgartner, F. R. (1997). Does incrementalism stem from political consensus or from institutional gridlock? American Journal of Political Science, 41, 1319–1339 [electronic version].Find this resource:

Jones, C. O. (1974). Speculative augmentation in federal air pollution policymaking. Journal of Politics, 36, 438–464.Find this resource:

Jones, C. O. (1975). Clean air: The policies and politics of pollution control. Pittsburgh: University of Pittsburgh Press.Find this resource:

Key, V. O., Jr. (1964). Politics, parties, and pressure groups (5th ed.). New York: Thomas Y. Crowell Company.Find this resource:

Kingdon, J. W. (1984). Agendas, alternatives, and public policies. Boston: Little, Brown.Find this resource:

Krier, J. E., & Ursin, E. (1977). Pollution & policy: A case essay on California and federal experience with motor vehicle air pollution, 1940–1975. Berkeley: University of California Press.Find this resource:

Levmore, S. (2010). Interest groups and the probem with incrementalism. University of Pennsylvania Law Review, 158, 815–858 [electronic version].Find this resource:

Lindblom, C. E. (1959). The science of “muddling through.” Public Administration Review, 19, 79–88.Find this resource:

Lindblom, C. E. (1965). The intelligence of democracy: Decision-making through mutual adjustment. New York: Free Press.Find this resource:

Lindblom, C. E. (1977). Politics and markets: The world’s political-economic systems. New York: Basic Books.Find this resource:

Lindblom, C. E. (1979). Still muddling, not yet through. Public Administration Review, 39, 517–526.Find this resource:

Lowi, T. J. (1964). American business, public policy, case-studies, and political theory. World Politics, 16, 677–715.Find this resource:

Miliband, R. (1969). The state in capitalist society. New York: Basic Books/Harper Colophon Books.Find this resource:

Nathan, R. P. (1976). Food stamps and welfare reform. Policy Analysis, 2, 61–70.Find this resource:

Olson, M., Jr. (1970). The logic of collective action. New York: Schocken Books.Find this resource:

Pal, L. A. (2011). Assessing incrementalism: Formative assumptions, contemporary realities. Policy and Society, 30, 29–39.Find this resource:

Popper, K. (1994). The poverty of historicism. London and New York: Routledge and Kegan Paul.Find this resource:

Rothstein, R. L. (1984). Consensual Knowledge and International Collaboration: Some Lessons from the Commodity Negotiations. International Organization, 38, 733–762.Find this resource:

Salisbury, R. (1969). An exchange theory of interest groups. Midwest Journal of Political Science, 13, 1–32.Find this resource:

Schick, A. (1983). Incremental budgeting in a decremental age. Policy Sciences, 16, 1–25.Find this resource:

Sowell, T. (1987). A conflict of visions: Ideological origins of political struggles. New York: William Morrow, Quill Books.Find this resource:

Sowell, T. (1999). The quest for cosmic justice. New York: Free Press.Find this resource:

Spicer, M. W. (1995). The founders, the constitution, and public administration: A conflict in worldviews. Washington, DC: Georgetown University Press.Find this resource:

Streek, W., & Thelen, K. (2005). Beyond continuity: Institutional change in advanced economies. Oxford: Oxford University Press.Find this resource:

Truman, D. B. (1951). The governmental process. New York: Knopf.Find this resource:

Weiss, A., & Woodhouse, E. (1992). Reframing incrementalism: A constructive response to the critics. Policy Sciences, 25, 255–273.Find this resource:

Wildavsky, A. (1964). The politics of the budgetary process. Boston: Little, Brown.Find this resource:

Wildavsky, A. (1984). The politics of the budgetary process (4th ed.). Boston: Little, Brown.Find this resource:

Wildavsky, A. (1991). The new politics of the budgetary process (2d ed.). New York: HarperCollins.Find this resource:

Wildavsky, A. (2006). Budgeting and governing. New Brunswick, NJ: Transaction Publishers.Find this resource: