This essay will discuss the starting points, premises, and relative utility of the top-down and bottom-up approaches within policy implementation. I will account when the top-down and bottom-up approaches should be utilized depending on the nature of the policy and the task environment in which the policy is being implemented. In general, Top-down implementation is the carrying out of a policy decision—by statute, executive order, or court decision; whereas the authoritative decisions are “centrally located” by actors who seek to produce the “desired effects” (Matland, 1995, 146). The bottom-up implementation approach initiates with the target groups and service deliverers, because they find that the target groups are the actual implementors of policy. Prior to divulging the relative utility of the top-down and bottom-up approaches amid the nature of varying policy and task environments, let us further explore the characteristics of the top-down and bottom-up approaches.
The top-down implementation approach is a clear-cut system of command and control—from the government to the project, which concerns the people. The top-down system showcases: (1) clear and consistent goals—articulated at the top of the hierarchical environment, (2) knowledge of pertinent cause and effects, (3) clear hierarchy of authority, (4) rules established at the top and policy is aligned with the rules, (5) resources / capacity to carry our the commands from the top (Elder, 2011, lecture). The top-down approach is the rational comprehensive approach to planning. It is consistent with overhead democracy, whereas elected officials delegate implementation authority to non-elected public servants (civil service) who are accountable to the democratically elected officials. However, DeLeon and Deleon (2001) point out that top-downers may implement [EPA] policy with standards that citizens do not understand, which might also circumvent their rational preferences. When this happens, top-down becomes a “tactic” and not a strategy for implementation (484).
Top-downers are quite often guilty of three criticisms. First, top-downers often initiate their analysis with the statutory language, which “fails to consider the significance of actions taken earlier in the policy-making process” (Matland, 1995, 147). Thus, implementors often engage cues from various groups, which differ in intensity and history, none of which may be reflected in the statutory language. For example, Matland described O’Toole’s analysis of water treatment plants; hence, the top-down approach revealed that privately owned treatment plants out-performed public treatment plants. However, once broader issues of affirmative action, Davis-Bacon labor laws, and technology were integrated into the analysis, then public treatment plants outperformed private plants (Ibid). Second, top-down implementors ignore or eliminate the political aspects of implementation (Matland, 1995). For instance, top-downers set clear goals for a policy, while the legislation “often requires ambiguous language and contradictory goals” in order to gain enough votes for passage (Matland, 1995, 147). Thus, a Weberian approach may be desirable in theory, but its practice may result in “policy failure” (Matland, 1995, 148). Finally, top-down implementors see the “statute framers as key actors,” however, local officials and people impacted by the policy could more reasonable be considered as the key independent variable of analysis.
Bottom-up designers begin their implementation strategy formation with the target groups and service deliverers, because they find that the target groups are the actual implementors of policy (Matland, 1995, 146). Moreover, bottom-uppers contend that if local bureaucrats [implementors] are not allowed discretion in the implementation process with respect to local conditions, then the policy will “likely fail” (Matland, 1995, 148). Accordingly, goals, strategies, and activities must be deployed with special attention to the people the policy will directly impact. Thus, evaluation based upon the street-level bureaucrat would be the best practice (Matland, 1995, 149). For example, Matland discussed Hjern’s findings that central initiatives poorly adapted to local conditions failed, and, that success depended greatly on the local implementer’s ability to adapt to local conditions (Ibid).
Discretion by agents is the underlying premise of the bottom-up approach (Elder, Lecture, 2011). Discretion may be a very good thing, especially when it uses expertise of people impacted by the policy to increase the likelihood of success and approbation. In bottom-up, one size doesn’t fit all cases, and so discretion may enable implementors to activate more useful practices or to ignore policy that will hamper the goal of the program. For example, the safe water drinking act became prohibitively expensive for smaller water systems. Of the 350,000 municipal water systems in U.S., only the conglomerates had the financial resources to quickly live up to the law without federal / state aid. However, the size of the water systems meant that a visit from a regulator was unlikely, so the great financial burden that would have occurred with a visit from a regulator was ignored. On the other hand, a regulator using the bottom-up approach in Chicago may overlook minor code violations for a bribe—since the mom and pop shop is out of code but is not a real threat to safety. The bottom-up model is thus a challenge to administration due to the reality of delegated authority, to the discretion allowed to different agents, which invariably causes a measurable variance of goal achievement. The bottom-up approach thus creates ambiguous goals.
Bottom-uppers are at times guilty of two criticisms. First, street-level bureaucrats are usually not accountable to the people. In this case, the local agents may intentionally subvert the elected officials’ policy goals and engage personal sub goals (Matland, 1995, 150). Second, bottom-uppers ignore the fact that many policies are created in a top-down manner, and likely in a manner which reinforces top-down authority. For example, Matland describes Sabatier’s analysis of environmental regulation in the United States, whereas the federal designers of the federal act integrated the necessary clauses to allow for class and individual lawsuits (150). Overtime, it was these lawsuits that adjusted the rule-of-law, not the local implementors.
DeLeon and DeLeon (2001) find that bottom-uppers are more likely to be reflective of community interests, while top-downers are more likely to impose policy narrowly upon focused interest groups (478). They conclude that bottom-up implementation is “more realistic and practical” and much more “democratic” than the top-down approach (Ibid). Further, if the policy is indeed meant to coerce people’s behavior, then the bottom-up approach may go beyond informing people of the proposed legislative action to manipulate behavior. In fact, bottom-uppers may garner the consent of the target group before their representatives’ vote for the law (Ibid).
Matland (1995) synthesized the top-down and bottom-up approaches, which illuminated the situations when each would be most useful. Briefly, top-down designers use central bureaucratic actors who manipulate policy implementation from the central level (146). These actors are analytical, and will attempt to aggregate their data into “generalizable policy advice,” predicated on patterns from articulated policy goals and policy outcomes (147). This structure seeks to minimize actors, limit change, and locate sympathetic agencies to implement policy (Ibid). Top-downers are likely to present “prescriptive advice” (Matland, 1995, 147). Conversely, bottom-uppers locate and describe the level of policy success in relation to the policy goal. While top-downers clearly explain the goal and seek to discover failure or success, bottom-uppers may fiddle with the implementation and encourage a policy that will at least partially achieve its goal.
Researchers must account for the differences in the top-down and bottom-up approaches when discovering implementation results. The top-down approach is often labeled a policy a failure because the implementation deviated dramatically from the original design—even if the goal was somewhat achieved. The variance from the hierarchical structure may showcase that the implementation process stumbled at multiple steps. On the other hand, bottom-up researchers might find less variance with the same policy, considering that the target population was allowed to alter the implementation of the policy due to local circumstances—in order to achieve the desired policy outcome. Clearly, the basic premises of the top-down and bottom-up approaches differ in structure and, indeed, utility.
Matland (1995) creates “structure” for researchers to use when interpreting the appropriate use of the top-down or bottom-up approach (146). This structure may integrate forward [top down] and backward mapping [bottom-up] [proposed by Elmore, 1982] (Matland, 151). The former is a precise and explicit statement of policy objectives, means-ends schemes, and outcome criteria (Ibid). The latter is a precise and explicit statement of local level behavior, the operations that shall cause change, and then the path to the central level (Ibid). Alternatively, Matland uses Sabatier’s movement from top-down analysis through multiple variables to the “advocacy coalitions” unit of analysis as an example of a potential middle ground between the top-down / bottom-up approaches (Ibid). DeLeon and DeLeon (2001) contend that bottom-uppers are more aligned with participatory democracy than top-downers, primarily because local officials must be responsive to their constituents and hence instruct bureaucrats on implementation processes (479-480). DeLeon offers a different structure: “Implementation should follow democratic procedures (and preferably in the most direct democracy practices) unless prior analysis demonstrates that another model (e.g., a top-down, or command, implementation) is superior” (2001, 488).
West (1984) articulates appropriate ways to structure and utilize the top-down and bottom-up approaches and comments on political styles. First, conservative elected officials are time and again critical of “excessive and unreasonable regulation” (342). On the other side of the aisle, liberals have “complained that agencies are overly solicitous of regulated interests” (Ibid). Either way, congressmen generally prefer the top-down approach because it allows them to “serve narrow clientele groups and to perform constituent service,” which contributes to incumbency advantage (341). Furthermore, the structure of the bureaucracy has grown beyond what legislators are able to direct. The bureaucracy has grown beyond congressional control; however, the judicial sphere has provided as a check against “arbitrary and capricious” legislation (353). Moreover, West found that the courts have abandoned the top-down traditional model, or, the “administration as a rational pursuit of objective goals” (348). Rather, the courts act upon pluralistic interests. Indeed, “institutionalized pluralism” within the congress and courts illuminates responsiveness by top-downers (348).
The relative utility of the top-down and bottom-up approaches to policy implementation, accounting for the nature of the policy and the task environment of the policy being implemented, differs from researcher to researcher. For example, Matland (1995) integrates Dunsire (1978) and Saetren (1983) whereas the top-down prospective is “more appropriate in the early planning stages, but a bottom-up view is more appropriate in the later evaluation stage” (152). Second, Matland discovers that Berman (1980) found that there are different and specific instances when each approach would be best utilized. When the policy goal seeks incremental change, exhibits stable technology, a stable environment, low conflict within the goal, and a tightly coupled institutional setting; then, top-down is the appropriate implementation strategy. Conversely, uncertain technology, goal conflicts, and an unstable / loosely coupled environment should cause the utilization of the bottom-up approach (1995, 152-153).
However, the starting points, premises, and utilization of the top-down and bottom-up approaches may be more complicated than I have so far explained. For example, Civil Rights legislation revealed a conflicted target population amidst a major policy change. The literature predicts that Civil Rights implementation would analytically benefit from the bottom-up approach, since the policy addressed abrupt change in an environment that would be hostile to implementing the policy change. On the contrary, the bottom-up approach enabled local leaders to engage personal goals “contrary to the policy objectives” (153). Indeed, the hierarchy of the military—beginning with the Commander-In-Chief—was necessary for President Eisenhower to deploy the 101st Airborne Division to Little Rock, Arkansas, and guarantee that 9 Black children would be integrated into a previous all-White school. Thus, the top-down approach may force a goal at odds with the relative status quo to be achieved through force, and may be the best alternative once the bottom-up approach fails [as opposed to altering the bottom-up approach]. The top-down approach may be the best approach to cause radical change in an unstable environment amidst high conflict.
There are numerous concerns regarding the effectiveness of the top-down approach. First, those at the top of the hierarchy may engage (unknowingly) authority leakage, whereas less and less control is allowed as you move down the hierarchy of implementation. In this manner, the local officials may not believe that they have the authority to implement the policy. Another concern within the top-down approach is shirking, which is the inability to control subordinates non-productive behavior. Shirking provides an opportunity for individuals to ignore their duty, and may be particularly potent when the shirkers are not stakeholders in some way to the policy. Moreover, people without a stake in the policy may engage counterproductive behavior, such as opportunism. Opportunism, or self-serving behavior [as opposed to serving the goals of the central authority], may thwart the success of the program. Opportunism as self-serving behavior certainly may lead to goal displacement. Finally, agency capture may spoil successful top-down implementation, because the implementors go native; that is to say, the implementors sympathize with those that they are supposed to regulate and then do not regulate them. In this case, the implementors become subservient to the very people they are supposed to monitor or control. In order to alleviate these concerns, the top-down approach often integrates compliance reporting and performance monitoring. Accordingly, “we call the former red tape” (Elder, Lecture, 2011).
The top-down model prides itself on structure and hierarchy, which are designed to provide ample accountability. However, many scholars find that the design lacks evaluation strategies, which would enable researchers to more readily outline the implementation failure/success. For example, Wood and Waterman (1991) inspected seven different public bureaucracies for responsiveness to political tools. They find that “leadership” of a bureaucratic agency is most likely to alter behavior, but urge: “Federal agencies should report selected outputs into a central tracking system on a regular basis, where they would be available to policy analysts and academics” (832). Since results would be publicly published, “tracking movements in the pluralist equilibria” of the U.S. government would make “politicians and bureaucrats more accountable and informed” (823). Indeed, Girdwood and Girdwood (2011) found that 17% of Iraq Reconstruction dollars (11 billion) was unaccounted for from 2004-2006. Moreover, they were unable to find how the contractors spent the received money. Thus, the top-down system is at times not amenable, or designed, to impose accountability—regardless of red tape.
Participation of stakeholders affords opportunities for administers to contribute in the rational planning model / top-down approach. Though this may decrease shirking, it may invite ritualistic behavior, creaming, waste, and goal displacement. Respectfully, ritualistic behavior may be illuminated when participants are obliged to engage in an ongoing planning process that has ongoing goals; such as public hearings for stakeholders and potential participants, but the process simply becomes ritualistic—going through the motions—because it is required, and thus does not fulfill its purpose. Creaming is observed when the servicers choose to accommodate those most easily served (e.g. job training programs for the car wash instead of a skilled art). Waste is regularly reported by the Government Accountability Office. And goal displacement, as described in prior paragraphs, is oftentimes advanced as greater participation of stakeholders increases the diversity of interests (Elder, Lecture, 2011).
Policy conflict levels may illuminate which approach should maximize opportunities for success. Matland sets forth that actors will increase conflict in accordance with the perceived stakes (1995, 157). Increased conflict will result in “more aggressive behavior” (Ibid). Top-downers assign conflict as an endogenous variable, and find that conflict may be mitigated once the task is given to a “sympathetic agency” (Ibid). However, the sympathetic agency may still encounter fierce conflict from opposition parties. In general, top-downers will attempt to limit conflict in order to progress with a rational decision-making strategy, while bottom-uppers find conflict to be a natural tendency in a pluralistic environment, which is to be commended (Ibid). Thus, bottom-uppers will expect the varying values of participants and see conflict as natural in an attempt to activate a more judicious normative policy (i.e. its structure). On the other hand, top-downers often actively attempt to diminish the importance of varying values and seek rational goals that all parties may agree upon in order to reduce conflict.
Policy ambiguity of goals and means may illuminate which approach could maximize opportunities for success. Like conflict, top-downers like goal clarity, or limited ambiguity. However, Matland finds that clearer goals are more likely to increase conflict and ambiguous goals often reduce conflict, particularly when clear goals reveal infringements onto stakeholders turf (1995, 158). Ambiguity in means may be observed when goals are unreachable due to a lack of technology, lack of role clarity, and a complex environment that complicates which tools should be utilized to proceed with implementation (Ibid). Conversely, this ambiguity is often necessary to acquire compliance from legislators (or stakeholders) and may accelerate opportunities for experimentation of new methods and technology (Ibid). Bottom-uppers again find that ambiguity is a natural part of the normative policy process and encourage “toleration for ambiguity” (Matland 1995, 167). Thus, top-downers would like to limit ambiguity, but find middle ground with bottom-uppers, since there may be a “fixed parameter” of ambiguity necessary to pass legislation so that the policy may be implemented (159). Once that legislation is passed, bottom-uppers may best solve the ambiguous issues at the street level.
In conclusion, this essay has provided an in-depth analysis to starting points, premises, and utilization regarding the top-down and bottom-up approach. Upon reflection, I ponder whether or not implementors may contrive a design that engages the best practices of each approach. For example, top-downers could merge with bottom-uppers if they: (1) consult the target population [perhaps through focus groups or random sampling], (2) predict the cost-benefit analysis based upon citizen participation and street-level implementers’ analysis, (3) organize and implement a randomized before and after test of the policy—impact analysis, (4) calculate the actual cost-benefit analysis after implementation occurs, (5) review the impact, results, and necessary revisions with the participants and street-level bureaucrats, (6) provide lee-way in different districts for second stage implementation, (7) pair similar districts into a larger analysis and encourage collaboration, and (8) publish raw data and sponsor social scientists to analyze the data. This might encourage evolving solutions within the implementation process.
******************** Works Cited
deLeon and deLeon, “What Ever Happened to Policy Implementation? An Alternative Approach,” Journal of Public Administration Research and Theory, 12, 4 (2002), pp. 467-492.
Girdwood and Girdwood, “Iraq Reconstruction: Investing in Multi-national Corporations?” Midwest Political Science Association, (2011) Presentation.
Matland, “Synthesizing the Implementation Literature: The Ambiguity-Conflict Model of Policy Implementation,” Journal of Public Administration Research and Theory, 5 (April, 1995), pp. 145-174.
West, “Structuring Administrative Discretion,” American Journal of Political Science, 28 (May, 1984), pp. 340-360.
Wood and Waterman, “The Dynamics of Political Control of the Bureaucracy,” American Political Science Review, 85 (Sept., 1991), pp. 801-828.