The Political Economy of Bureaucracy

Steven Richardson

• Publisher: Routledge
• Year: 2012
• ISBN:
  9781136868771
• Edition: 1
• Topic: Betriebswirtschaft
• Subtopic: Business allgemein

The Political Economy of Bureaucracy applies Public Choice theory and a complex systems view of government institutions to analyze policy implementation as an economic process. It addresses the common and vexing question of why managing federal agencies for results is so difficult by challenging traditional assumptions of institutional design and policy analysis. Using creative methods that focus on relationships that constrain the choices of executives and managers in a political hierarchy, the author reveals control and coordination as goals that are imperfectly achieved and often conflicting with one another.

Despite decades of intense study, serious reform efforts and impressive technological advances, the U.S. government remains a typical bureaucracy that fails to meet citizens' expectations. Clearly, policy analysis is missing something. The problem may rest with "machine" models of government. Rules, especially those governing expenditures, are assumed to be feasible and effective. Analysis of the federal government as a complex system of relationships between semi-autonomous agents helps explain the disconnect between policy and results. The solution is to roll back micro-management of ends and means; policymakers should focus on objectives and facilitate implementation by selectively relaxing constraints that prevent experimentation needed to determine the most effective methods.

This book devotes unusual attention to the interaction between executive and legislative branches of government and between political appointees and career civil servants. Most studies of government policy take existing institutional structure for granted. Different conclusions emerge from this analysis by virtue of the systems view that accepts status quo hierarchies but questions the effectiveness of the rules that govern policy implementation. This book will be of interest to postgraduates and researchers focussing on Economic Theory, Public Choice, Institutional Economics and Political Science, as well as to those working in the public sector interested in Public Administration, Public Policy, and Organizational Behavior.

1 The Research Question

How can US government agencies more effectively implement legislative and executive policies? Clearly, the scope of this study is quite broad; accordingly, the hypothesis is general: Accountability for results can improve the effectiveness of federal programs only if accompanied by greater flexibility to allocate resources. Governments at all levels have made progress in utilizing technology to gather and disseminate information, including data on performance of their own programs. However, accountability is not reporting; it adds understanding and responsibility – for explaining what happened, how experience compares to expectations, and how the new knowledge can be used to inform future decisions.

Planning, measuring and managing for results are distinct and critical components of effective organizations. I will argue that planning is incomplete if institutional effects are not explicitly addressed. I do not mean to suggest that the federal government will become effective if agencies are less constrained. That is a question with at least as many answers as there are citizens (effective at what, and for whom?). I am claiming that we cannot expect to learn whether policies actually work – nor can we expect to improve upon them – unless the agencies responsible for their administration have authority and incentives to implement them in the most efficient and effective manner. Structural reform is necessary but insufficient for effectiveness. In other words, we must attend to system capability before we can reasonably expect any sort of causal relation between inputs, policies, and outcomes. This follows from recognition of the federal government as a network. To be clear, the research question isn’t about ultimate effectiveness which, as explained in the next section, is subjective. Rather, it deals with effectiveness of the network of federal agencies and programs.¹

According to Kickert et al. (1997: 1–9, 171–2) failures of government to “steer” society led to experiments in decentralization that represented a retreat of the public sector. The concept of policy networks has emerged as an approach that reconciles limitations of central planning with the loss of coordination associated with autonomous policy making. The existence of policy networks cannot be denied, so the challenge is to improve our understanding of how they operate and to use that knowledge to improve the quality of interactions.

Similarly, this effort takes the existing US system as given – meaning it is the only available starting point.² Questions of network effectiveness are worth exploring independent of whether one agrees with existing policies. It matters whether government works because otherwise formulation and execution of policies – lacking evidence of their feasibility and effectiveness – may be an enormous waste of resources and opportunities. President Obama declared in his inaugural address:

The question we ask today is not whether our government is too big or too small, but whether it works – whether it helps families find jobs at a decent wage, care they can afford, a retirement that is dignified. Where the answer is yes, we intend to move forward. Where the answer is no, programs will end.

(Obama 2009)

That question will be answered, one way or another. Citizens will be better served if it is an informed decision; improving dynamic operation of the network that implements federal government policies will improve the quantity and quality of relevant information.

What is Effective Government?

It almost goes without saying that effectiveness requires coordination because our government, like all modern organizations, is highly specialized. While the need is apparent, what many do not appreciate is that “people must be motivated to carry out their parts of the cooperative activity” (Milgrom and Roberts 1992: 25). Even given a reasonable degree of coordination, we encounter additional questions such as whether the policies are effective and how results could be further improved. Coordination is important not just for the sake of efficiently executing political directives, but for testing policy propositions and learning from experience.

A polycentric arrangement that relaxes prescriptive resource allocation and focuses oversight on cost-effectiveness would facilitate programs’ responsiveness and potential for positive impact. For example, suppose Congress passes a bill increasing the minimum wage to help low-skilled workers. Presumably, since this is not entirely new legislation, implementation is straightforward – meaning those agencies responsible for enforcement already have rules and procedures in place. However, other programs are affected. Raising the price of low-skilled labor causes many employers to lay off workers, reduce their benefits or reconsider hiring plans. This increases demand for public assistance programs such as unemployment benefits and Medicaid, and it is not likely that Congress made specific funding allowances even if they recognized these effects. Unless the social service agencies have flexibility to shift resources between programs and localities, some rationing will occur. By the time funding formula change requests could work their way through the political process, it may no longer matter.

Thus, even when coordination problems are known, structural constraints prevent operation of mechanisms that could mitigate their impact. The budget cycle is too cumbersome to respond to many important performance indicators. Agencies are expected to improve quality of life for the populations they serve, despite having their resources tied to obsolete definitions of those populations and their needs. Programs may be reauthorized just once per decade, while participants’ challenges and opportunities change continuously. Unless decision rules shift focus from organizations to results, better information about the interaction of complex policies and programs is unlikely to improve matters. This is because accountability in Congress is highly dispersed – by design. Legislators favor ad hoc controls (asset specificity) that give them lots of power over discrete functions; they avoid broader controls that require sharing power. Agencies and programs are designed strategically to prevent opponents from unraveling or radically altering objectives. Tactics include rigid constraints on mission and decision processes, imposing deadlines, and limiting oversight/enhancing independence. “[Congress] willingly builds, piece by piece – however grotesque the pieces, however inconsistent with one another – the kind of bureaucracy interest groups incrementally demand in their structural battles. This ‘congressional bureaucracy’ is not supposed to function as a coherent whole” (Moe 1990: 136–7, 140, 142). “[O]rganizational boundaries often mirror the jurisdiction of legislative committees and subcommittees, and sharing power among legislative jurisdictions is a feat of supreme difficulty. . . . That political fragmentation makes it increasingly hard to ensure administrative coordination” (Kettl and Kelman 2007: 13).

Robert Behn explains the persistence of “stovepipes” in government, despite knowledge that networks can be more effective:

Legislators have very specific ideas about what government should do. And, to ensure these tasks get done, they assign each one to a specific, bureaucratic stovepipe. This fixed, formal organizational arrangement provides legislators with the ability to make a specific assignment to a specific unit. And, having made such an assignment, legislators then know whom to ask for an account of progress on the assignment. Finally, if the results are not achieved, the legislators know whom to punish.

(Behn 2006a)

Congressional oversight focuses on the assignment, which is to spend money as directed – not whether that activity led to the intended outcome. To look beyond the intermediate “result” of spending would be to question the policy decision that was so difficult to achieve and so proudly celebrated when the legislation passed. There are plenty of implementation failures to fill any Congressional committee hearing schedule; why would legislators invite debates over the soundness of their own policy? The research question is intended to call Congress’ bluff. Laws with multiple objectives are passed with relative certainty that some of them will be achieved (such as spending targets) and a high degree of uncertainty regarding achievement of the noble outcome goals that are presented to voters as primary justification for the legislation. Let’s grant politicians the benefit of the doubt regarding intentions. Still – if they are serious about their own policies – they should be interested in whether their stovepiped system and strategies are achieving stated outcome goals. Inattention to programs’ effectiveness is unacceptable, especially in the Information Age.

It is important to clarify the distinction between efficiency and effectiveness. Put very simply, efficiency is doing things right, while effectiveness is doing the right things. There are, of course, tradeoffs between the two. For example, public assistance is usually designed to assist those most in need. Therefore, cost-cutting is perceived as a perverse incentive to serve those who can help themselves at the expense of those supposedly targeted by the program. Furthermore, how do we gauge “right” for either measure?

Efficiency in public expenditures is not a prior fact out there waiting to be discovered‚ but is an artifact to be worked out. . . . CBA [cost-benefit analysis] or any alternative requires political‚ ethical judgment and the key issue is whether these are hidden as technical issues or are explicitly invited political inputs.

(Schmid 2004: 435–6)

The cost of any given function isn’t easily determined because there may be wide discretion in factor/cost assignment and allocation. Definition and measurement of outcomes and benefits is clearly a policy issue.

Legislators and bureaucrats protect and expand their property rights in policy and administration, respectively, by placing every decision in what is perceived as their domain. However, virtually nothing in government can be completely defined by one perspective or the other. Every policy has implications for administration, and vice versa. Therefore, it can be difficult, if not impossible, to separate the two. Conversely, they are not one and the same thing. Policies make value judgments that identify desirable ends; administrators exercise knowledge of processes to supply the means of achieving those ends (provided, of course, they have been granted the necessary resources). Regardless of who makes which decisions, though, in any complex system it is not possible to specify ends and means simultaneously – as Congress increasingly attempts to do. Means come first and are far easier to verify, so they come to dominate authorities’ attention. By default, ends become the dependent variable. Results are hardly sought, however, since they were virtually assumed to follow the prescribed plans.

The question of what the “right things” are is a continuous source of conflict within government. Property rights are fuzzy, there is plenty of room for honest disagreement about specific ends and means in any conceivable public choice, and it is not easy to distinguish decisions motivated by selfish and partisan interests from sincere attempts to promote common welfare. And often, abundant expertise actually hinders the effort: “Too many people, indeed, and especially too many experts, devote their lives to finding the best way of doing something that should not be done at all” (Boulding 1966: 10–11).

Clearly, then, effectiveness is subjective.³ So which outputs and outcomes should be measured? Government has many contextual goals that cannot be summed up by a single measure (Wilson 1989: 315–16). For example, the Job Corps measures placement of disadvantaged youth in employment or secondary education, but clearly seeks to prevent delinquency as well. Another problem is that government institutions are designed to separate values from facts. Traditionally, ethical judgments are policy decisions reserved for legislators or political appointees. Administrators are responsible for producing the outputs that have been determined appropriate by elected officials (Simon 1976: 45–60, 66–77).⁴

At all levels, decision-makers must evaluate which among several alternatives will better promote a given policy. This is why bureaucracies are rule-bound; expressing official preferences for every activity is intended to remove doubt and prevent manipulation regarding policy. It will be no small feat to convince politicians that such constraints only guarantee one thing – that the agency will be incapable of efficiently implementing any policy. Equally important is an explicit recognition that we choose actions, not outcomes. Policies are justified by judgments of social welfare and conjectures regarding which strategies will realize outcome goals. Actual progress is an empirical question with two parts: (a) Was the policy implemented as directed? and (b) Did those activities produce the desired result?

For administrators, to be “effective” is to execute directives of the President and the Congress, who are in turn directly accountable to voters for overall effectiveness – positive impact of policies and programs on social welfare. Agencies are formally organized under the Executive but authority to spend is controlled by legislative appropriations. Agencies exist to implement policies; they serve no other purpose. Program effectiveness, then, is a function of both policy and administration. Effective implementation of policy is a prerequisite to obtaining information about policy effectiveness because policies can only be tested in an operating context. Quality information about the consequences of fully implemented government policies may lead to better decisions.⁵

Organizations justify error and promote current activities by adopting vague objectives that enable substitution of actual consequences for original goals. Analysis, on the other hand, seeks to discover error as a means of promoting changes that create knowledge (Wildavsky 1987: 35–6).

Always we must be prepared to learn that we are wrong. For, as Popper tells us, “It is through the falsification of our suppositions that we actually get in touch with ‘reality.’ It is the discovery and elimination of our errors [read, policy failures] which alone constitute that ‘positive’ experience which we gain from reality.”

(Wildavsky 1987: 57–9)

Popper defined science as the “systematic criticism of error,” which Martin Landau (1973b: 541–2) says implied that “in the absence of criticism, there is mere change; where there is criticism, there is progress and the solution of problems.” Landau called for scientific management – systematic testing of policies as theories – as an antidote to bureaucratic “trained incapacity to think.” It is insufficient to measure historical performance, he says. Competing policies applied to the same problem areas present more relevant and potentially valuable challenges. “The institutionalization of criticism . . . would permit our stock of knowledge to rise if only because of the insistence that programs be constructed as to permit effective and continuous testing.” The interactive process of discovering and testing new knowledge is also known as catallactics – a Greek term for exchange. “Competition has, therefore, to be understood as a process in which . . . individuals dedicate resources peering into the fog of ignorance” (Kasper and Streit 1998: 222–5).

In many cases, even if good data systems and knowledge of program logic are in place, evidence of the impact of new policies and strategies on results will not be available until terms of elected office or appointment have expired. However, to avoid the issue is to invite cynicism and reversal. It is incumbent on policy makers to demand evidence of policy implementation and impact because commitment to measurement boosts credibility. Otherwise, opponents will justifiably conclude that it was just the preference du jour; the bureaucrats will resume doing what they please and successors will righteously replace it with a creation of their own (perhaps just as ill-conceived). A record of achievement, on the other hand, presents a challenge to demonstrate improvement. Even a benchmark has value in forming expectations that may support a minimum level of performance.

Performance measurement can become just another tool for ideologues if we forget that, in the political arena, everyone has an agenda – and their perspectives will affect the questions and answers. Key questions about data are “Who wants to know?” and “What for?” Some parties are indeed seeking a rational basis for change, but others are on a witch hunt or acting as advocates. Some are concerned about resources and others about process, output or outcomes. Each case may call for a different measure. It is impossible for program administrators to identify a single set of expectations that define efficiency or effectiveness because some programs are designed to achieve “symbolic” action rather than measurable outcomes. One reason for the reluctance of Congress to accept PART and other GPRA offerings is that they represent executive branch filtering of competing values from enabling legislation and budgets. For example, agencies’ outcome goals may conflict with mandates regarding methods of service delivery and policy implementation that can even specify types of staff involved and which resources should be expended. Performance measurement may have another unintended consequence; instead of expanding choices, it has led to a pseudoscientific management because it was implemented by traditional planners, who tend to deal with uncertainty by amplifying the rigor and details of their plans⁶ (Radin 2006: 93–7, 132, 189–92).

Kickert et al. state the same point in their d...