Narratives, Arguments, and Institutional Processualism:

Learning about Implementing Presidential Priorities from Brazil in Action

Dr Michael Barzelay Reader in Public Management

Interdisciplinary Institute of Management London School of Economics and Political Science

M.Barzelay@lse.ac.uk

Paper delivered at Conference on Smart Practices Toward Innovation in Public Management

IPSA Research Committee on the Structures and Organization of Government University of British Columbia

Vancouver, B.C., Canada June 15-17, 2004

Research conducted with the generous financial support of the Institute of Complex Strategic Studies (ICSS), Moscow, (www.icss.ac.ru). Collaborative research assistance provided Evgeniya Shvets, Researcher, ICSS, shvets@icss.ac.ru. Research cooperation provided by Professor Armando Cunha, School of Public and Business Administration, Getulio Vargas Foundation, Rio de Janeiro. acunha@fgv.br.

2

The responsibility of public managers includes designing processes that effectively operationalize policy mandates and contextual goals. The field of public administration has a long tradition of research and commentary on how to fulfill this responsibility. The classical administrative theorists fashioned a doctrine of organizational design that set forth such principles as unity of command and scalar chain. This doctrinal approach was roundly criticized by Herbert Simon (1946) nearly 60 years ago. Simon instead proposed that public managers engage in administrative analysis. This cognitive procedure included describing the functioning of particular organizations in terms of information processing and decision-making. It also included diagnosis, or identifying factors limiting organizational performance, analogous to specifying the binding constraints on the output of a physical system. He encouraged learning from experience by experimenting with structures and procedures intended to relax the constraints. He later admitted that administrative experimentation, in the literal sense, was rarely feasible (Simon 1976). Other scholars, such as James March (1999), have vigorously pursued the question of how else organizations can learn from experience.

A recent variation on this theme is contained in Eugene Bardach’s (1998)

Getting Agencies to Work Together, from which the title of the present conference is drawn. In that volume, Bardach examined experience with developing inter-organizational collaboration in social program delivery in the U.S. He also sketched a wider position about public management that he called craftsmanship theory. As indicated by the conference theme, this approach deserves serious consideration.

The book displays a number of distinctive merits. First, it provides a way to

describe and analyze functioning systems of organizations. Second, it provides an analysis of the developmental dynamics of such systems. Third, it develops analytical generalizations about the development and operation of processes by which agencies collaborate in delivering public services. Fourth, it suggests how public managers should make use of such knowledge. Finally, it can be read as a stimulating general argument about how to study public management.

The problem is that Bardach’s approach is difficult to parse and assess.

Presenting theoretical and methodological ideas is awkward in the confines of an empirical, policy-oriented book published by the Brookings Institution Press.1 The book’s general argument about learning from experience in public management requires restatement, even at the risk of oversimplification. It would also help if the book’s articulation of its analytical approach were more explicitly linked to familiar research paradigms within organization studies, management, and political science. In addition, the value of the work could be more fully appreciated if used to add subtlety to academic discussions of lesson drawing and policy-transfer, which pertain to the larger issue of learning from experience.

This paper is a critical, empathetic examination of the Bardach approach. To

reduce the risk of excessive abstraction, the discussion goes hand-in-hand with the analysis of a case. The case of “Brazil in Action” – to be introduced momentarily -- is not uniquely well suited to this purpose, but is appropriate nonetheless. It embodies what could plausibly be claimed is a smart practice under Bardach’s criteria. The

1 The same goes for Barzelay and Campbell (2003).

3

candidate smart practice – a presidential-priority action program -- is a way to perform the function of management control of program delivery using a process styled on project management. This process --which involves line agencies, contractors, central coordinating agencies, and the office the president -- is possibly a smart practice for a number of reasons. First, this type of process is possibly workable even in varied operating conditions. Second, it may take clever advantage of an underlying potential in public sector delivery situations – in a sense, obtaining results on the cheap. Third, it may pose relatively small risks of negative side effects.

Part I presents a narrative history of the Brazil in Action experience. Part II presents a case analysis. The analysis involves both argumentation analysis and discerning what it was about the Brazil in Action program that made it work. In the latter connection, I push beyond the immediate boundaries of Bardach’s analysis and bring to bear theoretical tools from a social science approach I call “institutional processualism.” The result is, in effect, an analytic narrative that is intended to help political officials and public managers craft imaginative and workable solutions to well-defined administrative design problems, within a certain range. Overall, the paper represents a point of view about how to learn from vicarious governmental experience along with a detailed example of how to put that view into practice.

Part I: A Narrative History of Brazil in Action Brazil in Action emerged on the scene in a year and a half after Fernando

Henrique Cardoso’s inauguration as president in January 1995. As revealed publicly in August 1996,2 the program designated 42 priority presidential projects, all with development overtones. About half were infrastructure projects, including major highways, port facilities, waterway transport systems, railroads, long-distance electrical transmission lines, and an international gas pipeline. The rest were major initiatives in the policy fields of health, education, housing, and sanitation. These various projects were grouped together, identified personally with the president, and publicized under the common banner of the Brazil in Action program.

This program represented a mid-point correction in the substance and style

Cardoso’s presidency. Until this point, the president’s policy agenda was widely perceived as extending little beyond maintaining the hard-won macroeconomic stability that had been ushered in with the so-called Real Plan, introduced in 1994 while Cardoso was minister of finance in the government of Itamar Franco. Polls showed that the electorate expected that the country would finally move forward under the new president, after a long hiatus of severe economic and political difficulties. In the words of a government pollster, “Brazilians expected the new government to address the desire that Brazil develop, that it forges ahead, and that it no longer stagnates or worse, as had been the case before Fernando Henrique became president.”3 Brazil in Action thus represented a modest but visible shift towards a developmental policy agenda.

The program also represented a shift in the government’s administrative

processes. Until this point, the central coordinating agencies were only involved in 2 [check date] 3 Interview with Fernando Barros, President, Propeg, Salvador, Bahia, August 25, 2003.

4

major projects through performing planning, budgeting, and financial management functions. In 1995, during the first several months of the Cardoso presidency, the Ministry of Planning coordinated the formulation of a four-year plan for public investment and spending, in accord with constitutional requirements. That plan was comprehensive, including hundreds more projects than later selected for Brazil in Action. The budget secretariat of the same ministry coordinated the annual spending plan, which earmarked funds for the Federal share of infrastructure projects and social programs. The treasury secretariat of the ministry of finance decided when, if ever, to disburse funds for projects as for the rest of Federal activities. While these roles and responsibilities were not disrupted by Brazil in Action, the program deepened the involvement of central coordinating agencies, as well as the presidential staff, in the process of managing projects.

The outcome of this episode is difficult to sum up in a few lines. Suffice it to

say at this point that the center of government – especially the ministry of planning – became capable of overseeing the implementation of a large number of major presidential priority projects. Underneath this capability was a functioning management system, built around project management principles, and augmented by subtle changes in budget execution procedures. Reflecting this capability, a large fraction of the priority projects were concluded according to plan. The president was delighted with the political dividends he derived from the program, which became the principal symbol of his first term of office. Finally, the episode set the stage for two further major efforts to improve public management, broadly defined. It did so, in part, by changing beliefs about the limiting factors on government performance in Brazil. Looking back on this episode in the light of the subsequent initiatives, a veteran senior official in the planning ministry observed: “I think that by implementing principles of management, Brazil in Action showed that the lack of results in the public sector, in large measure, does not stem from the lack of resources, but rather from the lack of management.”4

Origins of Brazil in Action The conception, design, and launching of Brazil in Action was a result of

several independent influences coming together at the same time. During his first year as president, Cardoso gave much thought to basic issues about the role of the state. These issues arose because successive governments had abandoned the strategy of import substituting industrialization in favor of trade liberalization, reduced subsidization of selected industries, and selective privatization. Many of the most intensively used instruments of microeconomic policy – such as tax breaks and credit subsidies – had been discarded. Yet, Brazil remained a country with prodigious social needs and regional inequalities. For many, these conditions seemed to call for a national project of development. Such a project had been completely off the agenda while the overriding concern of politicians and technocrats was to reestablish macroeconomic stability. With inflation suddenly coming under control in the run up to Cardoso’s election in October 1994, the big questions about the state and development resurfaced.

4 Interview with Ariel Pares Garces, Brasília, D.F., August 2003.

5

Cardoso wrestled with these questions in frequent dialogue with long-standing confidants, whose ties dated back to the president’s earlier career as an internationally prominent sociologist and a left-of-center public intellectual.5 By the end of 1995, Cardoso had reached closure: While priority would be given to stability, Brazil’s public sector needed to use its limited resources effectively to compensate for the shortcomings of market processes. These limitations were most convincing, he concluded, in infrastructure provision as well as health, education, and sanitation. Cardoso was strongly disposed to pursue an activist public policy in these areas, but was equally disinclined to fashion a national developmental project of the sort symbolized by the five-year National Development Plans of the 1970s and early 1980s. He was later known to say, “We have a national project, but it is a national project for a new reality. A deterministic national project was possible in the past, in a closed economy where the state was the principal investor, captured the country’s savings. and implemented its plan. That is no longer possible in a climate of stability, fiscal discipline, and international financial interdependence. Today, the national project is more a national process of development than a national project.”6

During his first year in office, Cardoso enjoyed great popularity, with approval

ratings exceeding 60 percent. Sustaining such a favorable position in public opinion was not taken for granted, however. None of the recent, democratically elected presidents had left office with much of their popularity in tact. To gain some insight into how to sustain presidential popularity, the president’s team cast back more than 35 years to the time of Juscelino Kubitschek, who served as president from 1955-60.

President Kubitschek was best remembered for transferring the Federal capital

from Rio de Janeiro to the newly constructed planned city of Brasília in 1961. However, Brasília was only the capstone of a numerous projects, many of them involving infrastructure. As a presidential leader, Kubitschek was admired for his success in communicating with the mass public. A particularly effective message was that his government’s efforts, including the construction of Brasília, were unified in a Plano de Metas (Plan of Goals). The idea caught on. In a book of memoirs, a finance minister under Kubitschek recalled a fishing trip during which the owner of a shop he visited made a point of showing him a notebook on which the phrase “Plano de Metas” was written in pencil across the cover. The shop owner told the former finance minister, “The president made a Plano de Metas for Brazil, and I made a plan de metas for my shop.”7

Kubitschek’s Plano de Metas provided an important starting point for

Cardoso’s team as they sought to work out a governing strategy. The historical analogy led to the idea that Cardoso needed to be personally identified with development projects. Fernando Henrique would have to be seen visiting ongoing or completed projects, just as Juscelino had done, at that time usually traveling by car.

5 Interview with Ambassador Sergio Amaral, former Minister of Communications, Presidency, Paris, October 30, 2003. For background on Cardoso’s thought, as it evolved through engagement in partisan politics, see Resende-Santos (1997). 6 Interview with José Paulo Silveira, Rio de Janeiro, RJ, August 27, 2003. 7 Interview with Ignácio Muzzi, Brasília, August 22, 2003.

6

The historical analogy also led to the idea that the current president’s priorities should fall under a common, catchy name.

The historical analogy with the late 1950s and early 1960s was considered far

from perfect, however. To carry out his projects, including Brasília, Kubitschek had bypassed the regular ministries of the Federal government. In effect, a parallel universe of non-ministerial bodies was established. These arrangements were broadly labeled “indirect administration.” However, that model of public management had become discredited by the time the country’s democratic elites wrote the 1988 Constitution. The archipelago of non-ministerial bodies had subsequently been reorganized, putting many of them out of existence. In general terms, the Cardoso entourage did not wish to challenge the constitutional bias in favor of direct administration.8 Indeed, circumventing the ministerial bureaucracy would have been blatantly inconsistent with his government’s well-publicized “managerial reform of the state apparatus” (Gaetani 2004). The challenge for his government was therefore to carry out something like the Plano de Metas through the structures and people of the Federal bureaucracy.

Designing Brazil in Action The idea that Cardoso should become identified with a development agenda

built around implementing priority projects took shape early in the president’s second year in office. This idea coalesced after the political heavyweight, José Serra, resigned as minister of planning in June 1996 to run for mayor of São Paulo. Serra was succeeded by Antonio Kandir, who at the time was representing São Paulo in the lower house of Congress. Before entering electoral politics, Kandir had achieved distinction as a macroeconomic researcher and a sometime policy-maker. His ties to the president extended back to the 1970s, when Kandir worked within Cardoso’s research institute, CEBRAP, on the economics of inflation. As the incoming planning minister with responsibility for the budget, Kandir’s mandate included keeping the lid on Federal spending. However, his mandate did not stop there. As Kandir later recalled, “When Fernando Henrique invited me to serve as planning minister, he asked that I move forward all the projects that were in operation, and he told me that I was welcome to make new proposals.”9

The Policy Stream and the Planning Secretariat. From the start of the Cardoso presidency, the ministry’s planning secretariat had been brimming with ideas, many of them concerned with crafting a suitable form of development planning within the Federal government.10 The secretary of planning was Mauro Marcondes, a senior career official on loan from the National Bank for Economic and Social Development (BNDES), the federal government’s 50-year-old instrument of economic policy implementation and principal source of domestic long-term credit. A chemist and economist by education, Marcondes’s specific background within BNDES included infrastructure development and strategic planning. At the ministry

8 They did, however, agree to support, albeit tepidly, a constitutional amendment that would allow departments to create “social organizations.” 9 Interview with Antonio Kandir, São Paulo, SP, August 2003. 10 During this period, the secretariat was entitled, Planning and Evaluation. Its acronymn was SPA.

7

of planning, he and his staff were dissatisfied with conventional thinking about how to perform the planning role at the center of government. According to Marcondes, “When I became planning secretary, I already had a vision that the secretariat should not just conduct studies. It had to be a secretariat that took a position on important issues, proposed actions, and interacted vigorously with the sectoral ministries.”11

In pursuing this idea, Marcondes’s assembled team decided to identify actions

that could have a dramatic impact on production and development on a regional scale. Such actions would include big infrastructure projects, such as developing major inland waterways for transport purposes in several regions of the country. Many of these projects had been featured in the campaign manifesto on which Cardoso had run for the presidency. The secretariat also wanted to articulate a planning philosophy that was coherent with its substantive emphasis on infrastructure as well as responsive to widespread skepticism about national planning.

Within a few months of Cardoso’s inauguration in January 1994, Marcondes’s

secretariat had settled on the idea of planning “axes of national development and integration.” The axes idea was already gaining ground in elite circles and in BNDES, thanks to the steady efforts of its highly visible progenitor, Eliezer Batista, former president of the innovative state-owned mining and logistics company, Companhia Vale do Rio Doce (CVRD), who had served as secretary of strategic issues in the government of Fernando Collor. An axis of development was a variation on Batista’s initial concept of forging transport corridors between inland locations of economic activity and major ports. This concept became broader at the hands of a senior member of the planning secretariat, Ariel Pares, who wished to give national planning a strong territorial dimension, as in France, where he had studied for five years. This approach, as it developed, was politically advantageous, as well, since it did not conjure up the specter of sectoral economic planning, akin to Brazil’s National Development Plans of the 1970s. Even so, enthusiasm for the embryonic axes approach to planning apparently did not extend to the planning minister, José Serra.

On the heels of this reflective period, the planning secretariat turned its full

attention to preparing the government’s proposed multi-year expenditure (or investment) plan for 1996-1999. Under terms of the 1988 Constitution, Article 165, the president was obliged to submit such a plan (known as the Plan Pluri-annual or PPA) to Congress, where it would be debated and voted upon. The intense process of formulating the PPA involved more than 350 people from across government. The proposed goals and actions were initially formulated by groups of technical specialists, advisers, and managers who worked in the sectoral ministries. These inputs were then vetted by cross-ministerial working groups, convened by the planning secretariat, whose task included selecting out goals and actions that would have a significant policy impact from those that would play a supporting role. The result was an articulation of the government’s major guidelines. As a consequence of giving priority to forming consensus on main goals and actions from a whole-of-government perspective, the planning secretariat did not formulate an outline of the Federal budget for the coming four years, as some in Congress had expected. Instead, the official version of the PPA, sent to the Congress, identified goals and

11 Interview with Mauro Marcondes, Washington, D.C., September 24, 2003.

8

actions to deal with public policy problems, without reference to expenditure levels or the ministries responsible. The PPA also identified “strategic investments” – not necessarily to be funded by the Federal budget. These strategic investments were grouped according to axes of national development and integration, despite Serra’s lack of enthusiasm for the underlying idea. “If we hadn’t had this critical mass of reflection and had not begun to view planning from a territorial point of view,” according to Marcondes, “we would not have done the PPA in the way we did.”

With the PPA complete, Marcondes began to think about the role that the

secretariat could play in implementing the strategic investments. One idea was to construct a database of projects. The task was to compile information about all current and possible future investment projects, with a view to helping to select ones that would be implemented. The information system developed at the hands of secretariat staff members, who were on loan from the state-owned commercial bank, the Banco do Brasil, which had embarked on introducing information technology into its managerial and operational activities years earlier. Marcondes also began to work towards the concept of “a collection of projects that would have a management system, monitoring, and prioritization within the budget, so that they would be implemented by the end of the president’s term.” He failed to sell the idea to Serra. When Serra left the ministry to run for mayor of São Paulo in June 1996, Marcondes shared the same thoughts with his successor. “Kandir picked up the idea,” he later observed.

Picking up the Idea, Running Fast. Kandir entered office convinced that the

government would face political difficulties if it continued to be perceived as entirely preoccupied with short-term macroeconomic concerns. Substantively, he became committed to addressing infrastructure problems, because of their symbolic importance and potential impact on development. His immediate agenda included the practical concern of implementing projects in a timely and efficient manner. On his first day as minister of planning he asked for status reports on the projects that were supposed to be in operation, given their inclusion in the multi-year plan. The information served up the next day struck him as patchy, out of date, and obtained from dubious sources. Kandir’s summary assessment was that the ministry possessed a “mountain of totally useless data.”12 What they needed, he asserted, was up-to-date information, provided by individuals with project responsibility, on a selection of important undertakings.

It was common knowledge that getting projects done on time in Brazil was

immensely difficult. Many projects required consistent action by more than one line ministry, one or more state-owned enterprises, one or more state governments, private firms, and the treasury secretariat of the finance ministry. Coordination was not necessarily spontaneous. In terms of information flows, the bureaucracy operated in a compartmentalized manner. Officials’ principal loyalties attached to their ministries. Ministers, meanwhile, did not necessarily form a cohesive government, since they were drawn from different parties.

The government’s fiscal predicament was another well-known source of

difficulty. Annual budgets operated as a ceiling on spending, but were rarely

12 Interview with Antonio Kandir, São Paulo, Brazil, August ?, 2003.

9

executed as authorized. The level and timing of disbursements depended on such volatile variables as the level of tax receipts, interest rates, and fiscal targets. Uncertainty about disbursements, brought on by such factors, exacerbated the coordination problems and tended to cause delays. In Kandir’s words, “The resources available for the public sector were not so much scarce as extremely variable. What’s really troublesome for a project is not to have certainty that it is going to progress through all of its stages. This lack of confidence was the source of many problems in implementing projects.” This situation reinforced Kandir’s inclination to give special administrative status to the Brazil in Action projects. The president would have to say that his priority projects would not suffer from budget cutbacks. In sum, Kandir settled on the view that three conditions were necessary for implementing projects: “To have money guaranteed, mechanisms of communication, and people who thought about their projects 24 hours a day.”

Relations with Coordinate Authorities. As planning minister, Kandir shared

responsibility for coordinating ministerial activities with the Minister-Chief of the Civil House (Casa Civil) of the Presidency of the Republic, Clóvis Carvalho. Carvalho and Kandir easily found common cause in fashioning a way to move the priority projects forward expeditiously. They were both from São Paulo, educated as engineers, had long been active in the Brazilian Social Democratic Party (PSDB), and had working experience in both government and the private sector. Known informally as a kind of prime minister, Carvalho had set up a system of cabinet committees, in such areas as economic policy and infrastructure, to achieve some alignment among government ministers, many of whom belonged to Cardoso’s broad center-right governing coalition but not to the PSDB.

Carvalho was also deeply involved with the managerial reform of the state

apparatus, an initiative being vigorously pursued by the minister for federal administration and reform of the state, Luiz Carlos Bresser Pereira (Gaetani 2004). On this issue, Carvalho was of two minds. On the one hand, he favored a managerially oriented public administration; on the other hand, he did not think that Bresser Pereira’s proposed structural changes would lead directly to results-oriented public management practice. Complementary changes in people and processes were, in his vision, required.13 From where Carvalho sat, Kandir’s effort could be made to dovetail with Bresser’s managerial reform. Indeed, he thought that specific efforts were needed to foster on results-oriented management in practice.

Recruiting Talent. In preparing to head the ministry, Kandir tapped into his extensive professional network, asking José Paulo Silveira to serve as his secretary for state-owned enterprises. Kandir came to know Silveira in the early 1990s when he had been secretary of economic policy in the government of Fernando Collor. At the time, Silveira was directing a massive Brazilian Program of Quality and Productivity that involved thousands of firms trying to adapt to the country’s rapid trade liberalization. In this role, Silveira widened his reputation as a highly skilled leader of major project-like undertakings, previously earned while employed by the state-owned oil major. During his long career at Petrobrás, Silveira had reengineered the company’s procurement systems, directed the research and development department when the company was studying how to drill offshore at unprecedented depths, and

13 Interview with Clóvis Carvalho, São Paulo, SP, August 2003.

10

headed the strategic planning unit when the state-owned enterprise was preparing for the end of its legal monopoly in the upstream domestic petroleum sector. As head of the Brazilian Program of Quality and Productivity, Silveira made an impression on Kandir for what the minister considered “his clarity of vision, from the point of view of the planning process – his organized thinking….Perhaps because of our common engineering background, I immediately felt a strong identity with Silveira.” Contact between Kandir and Silveira continued after both left government, in connection with Cecrisa, a firm in the ceramics industry, where Kandir served as a non-executive director and Silveira as director of development.

Silveira demurred on the offer to be secretary of state-owned enterprises.

Meanwhile, he participated informally in discussions over project selection for what was to become Brazil in Action. According to a close observer, Silveira found himself in immediate agreement with the view that planning should involve selecting strategic projects and also strategic management within government itself.14 In this context, Silveira pressed the argument that the government needed to introduce a genuine managerial process in order to achieve project results. More specifically, he argued that:

Each project should be organized according to the principles of project management. Each should have a stated purpose, objectives, physical programming, and financial programming. Each project should have a manager, an explicitly identified individual. The ministry of planning should monitor these projects. And, finally, this model would have to be seen as the president’s program and not that of the ministry of planning. Otherwise the structure of power would not permit it to operate.15

As it happened, Mauro Marcondes accepted a position at the Inter-American

Development Bank in Washington, D.C., within weeks of Kandir’s arrival at the planning ministry in June 1996. Revising his initial offer, Kandir asked Silveira to succeed Marcondes, with specific responsibility for creating a management system for the presidential priority projects. He replied, “Now you’re talking about something that really interests me.”16 Like Clóvis Carvalho, Silveira was attracted to “the fantastic challenge” of “transforming a bureaucratic style of management into a managerial style oriented to results, that is, to introduce entrepreneurship into the public administration.”17 He thought project management principles could be applied across the entire Federal government, making Brazil in Action a pilot experience. Silveira accepted Kandir’s offer.

14 Interview with André Amaral de Araújo, Rio de Janeiro, RJ, September 3, 2003. Amaral was director for strategic investments within the secretariat of planning, under both Mauro Marcondes and José Paulo Silveira. 15 Interview with José Paulo Silveira, Rio de Janeiro, RJ, August , 2003. 16 Interview with Antonio Kandir. 17 Idem.

11

The Proposal Stage and Beyond. During the June-September period, events moved quickly. At the end of June, Kandir, joined by Silveira, presented the proposed program to the president during a low-profile session at Cardoso’s home in São Paulo. As presented, the program would consist of no more than 50 presidential priority projects to be administered through a system of project management. These projects would be the president’s projects, not just projects of the government. About half would be infrastructure projects, with the other half social programs. Priority infrastructure projects would be ones that promised a significant economic impact, while satisfying political sensitivities concerning geographic distribution. Once selected, priority projects would be sheltered from budgetary cutbacks, barring a major fiscal crisis. The presidency’s media and social communications operation would seek to position the program in the public mind as a concerted effort by the president to direct the action of government to develop the country in tangible ways. Administratively, each project would be headed by a project manager, a conceptually new role within the ministerial bureaucracy. The role of project managers would include not only monitoring the status of projects, but also taking the initiative in removing foreseeable obstacles impeding their on-time completion. Project managers would report regularly on project status and issues requiring attention and resolution. The ministry of planning would closely monitor projects on behalf of the president. If its executives failed to resolve coordination problems, they would be able to elevate them to the presidency – to Clóvis Carvalho at Casa Civil, in the first instance. As Kandir recalled, “the president liked the project.” By the end of the meeting, Kandir had received the president’s go-ahead to proceed.

With the conceptual design stabilized, the work during July focused on filling

in the next level of detail. This effort proceeded with a sense of urgency for several reasons. First, public opinion polling suggested the president’s popularity was beginning to wane. Second, Kandir’s political instincts told him that his own political capital would tend to diminish as time passed. Third, Kandir wanted to differentiate the Brazil in Action projects from other planned expenditures in the 1997 Federal budget. Under the constitution, the president was obliged to submit the following year’s budget to Congress by the end of August. The sectoral ministries needed to be told of the president’s priorities some weeks before this deadline.

Putting Brazil in Action into Operation Selecting Priorities. An immediate task was to make a selection of projects.

In doing so, they built on prior efforts undertaken by the planning ministry in coordination with Clóvis Carvalho to segregate genuinely strategic investments from the hundreds that had found their way into the four-year plan. As recalled by one director within the planning secretariat, Francisco Batista da Costa, “We started with ten projects, the list then grew to 60, was cut to 30, and finally settled down at 42.”18 In preparing Brazil in Action, specifically, some projects seemed obvious candidates, given the fundamental criterion that “each should make a difference to development by catalyzing other investment decisions, whether by the private sector or the states.”19 An example was the Bolívia-Brazil natural gas pipeline, which was to 18 Interview with Francisco Batista da Costa, Rio de Janeiro, RJ,. September 5, 2003. The central actors in this process included Kandir, the executive secretary, Andrea Calabí, and Marcondes. 19 Idem.

12

provide a source of energy to São Paulo and stimulate industrial development along its 3,100 km path. This project was already moving towards the construction phase, with the project finance arrangements nearly in place. It enjoyed a demonstrable consensus among elites in São Paulo and its wide area of economic influence as well as full acceptance by Petrobrás, which became its champion. Another example was the building of a railroad, Ferronorte, which would link the Center-West’s expanding grain-growing region to the distant port of Santos in São Paulo state. Within this larger undertaking was a plan to use money from the Federal budget to build a railroad bridge across the wide Paraná River. A number of road widening and rehabilitation projects were also quickly identified, including the route between São Paulo and the city of Belo Horizonte in the neighboring state of Minas Gerais. Various projects that would benefit the Northeast, North, and South regions of the country were also pinpointed. As an example, in the Northeast, a set of projects intended to create a more robust tourist infrastructure, including expanded airports in major regional cities, made the list. The ultimate decision-maker was the president himself, who took an active interest in finalizing the selection. He took advice from state governors and regional elites represented in the Congress and made some adjustments in the set of “undertakings” (emprendimentos), as the projects came to be called in the official discourse. However, the president limited his choices to those that had been served up from the planning process, linked to the PPA.

Choosing a Name. Meanwhile, in July, the idea of formulating a trademark,

analogous historically to the Plano de Metas, came into focus in discussions with the president’s communications team, including Ana Tavares and Sérgio Amaral. They initially toyed with the words, Investe Brasil. However, a different name was suggested by a public relations firm, Propeg, working on a contract with the presidency and the planning ministry. According to Propeg’s president, Fernando Barros, “I felt that Brasil em Ação fit the bill as a way to characterize the whole range of actions that provided a lever for the country’s development across its diverse territory and society. ‘Brazil in Action’ encompassed the various areas of governmental action, and it conveyed the Federal government’s association with other governments and the private sector.”20 Brazil in Action’s symbols – apart from the president himself -- were to include a logo and, more importantly, a map of the country showing the location of the various projects.

Going Public. On August 9, 1996, less than a month after the president

approved the program, Brazil in Action went public. The occasion was a meeting of all ministers, to which reporters were invited. Kandir delivered a speech presenting the program outline and the 42 projects. The planning minister also explained that each project would have a manager, and he informed the ministers that their proposed 1997 budgets should specifically include lines for the Brazil in Action projects.21 The president, nonetheless, stole the show. As Silveira recalled, “During the meeting the president spoke a number times. One of these times, he stood up and went to screen to comment on the importance of certain projects. The following day the image of FHC and a Brazil map indicating the Brazil in Action projects was displayed on the 20 Fernando Barros, President, Propeg, Salvador, Brazil, August 25, 2003. 21Prepared remarks by both Cardoso and Kandir, along with an outline of the projects, were published in, “Brasil Em Ação: Investimentos para o Desenvolvimento,” (Brasília: Presidência da República, 1996).

13

front page of main Brazilian newspaper.” As Clóvis Carvalho remarked about the same event, “Fernando Henrique mentioned, for instance, ‘Highway 163 between Cuiabá and Santarém…’ He has an excellent sense of spatial geography and a fantastic memory for numbers. Nothing was written, but he knew perfectly each one of these things.”

Hitting the Road. Within two weeks of this event, which resulted in a media

splash, the president participated in his first ceremonies related to particular Brazil in Action projects. In Manaus, the capital of the state of Amazônia, Cardoso presided over the signing of an agreement between three state-owned enterprises -- Petrobrás, Electrobrás, and Eletronorte -- to build a natural gas pipeline from Manaus to Urucu and then to Porto Velho. The project would involve providing fuel for electricity generation in the Amazon region and elsewhere in the north. The same day, Cardoso flew west to the city of Porto Velho, where he presided over the signing of a BNDES loan to finance barges as part of developing a waterway transportation system linking the west and north of the country.

Finding Managers. While the president was creating personal identification

with Brazil in Action projects, following in the distant footsteps of Juscelino Kubitschek, his team was trying to implement the administrative approach of working with the ministerial bureaucracy. At the August 9 meeting, ministers were informed that lead responsibility for projects would be assigned to a single ministry or state-owned enterprise (SOE) and that each project would have a manager. In preparing to unveil Brazil in Action, it had been decided that project managers should be situated organizationally within the sectoral ministries or SOE’s responsible for the project, rather than, say, the ministry of planning. The fear was that furious turf battles between the planning ministry and the sectoral ministries would otherwise break out, impeding the very progress in execution that was the raison d’être of Brazil in Action. Besides, with a staff of fewer than 30 professionals, the planning secretariat hardly had the capacity itself to manage 42 presidential priority projects on its own. Once it was decided that project managers would be situated organizationally in the sectoral ministries or SOE’s, the president’s team were essentially committed to letting ministries choose who would be the project managers, lest they be seen as undermining normal personnel practice and making a power grab.

To make the best of the situation, face-to-face discussions with more than a dozen ministers and their executive secretaries were held during August. The agenda was to explain the fundamentals of the Brazil in Action program in a direct and personalized manner. Kandir, Marcus Tavares (who had succeeded Calabi as executive secretary), and Silveira went to all of these meetings. In many cases, they were joined by Clóvis Carvalho and Eduardo Jorge, the general secretary of the Presidency, to underscore Cardoso’s personal commitment to implementing Brazil in Action. The presidentially backed delegation tried to persuade ministers that, whatever a manager’s formal hierarchical position, direct access to the ministry’s executive secretary (who, as the undoubted number two authority, controlled budgetary and other resources) was a must. The profile of a manager as someone who was both technically competent and adept at interacting effectively across organizational boundaries was also discussed. Specific names were considered at these high-level sessions. As Silveira recalled, “The minister of environment, who was responsible for two projects, mentioned two names early in the meeting. After

14

we talked, he remarked, ‘No, I’ve changed my mind. I’m going to appoint two other people now that I understand better what’s needed.’”

Developing a Prototype Project Management Information System. While these

sensitive discussions were taking place, Silveira directed his planning secretariat staff to stop building the database of investment projects, which had been one of his predecessor’s priorities, and to turn their full attention to designing a “real-time management information system” for the Brazil in Action program. From an organizational standpoint, this system was meant to compensate for the lack of a direct hierarchical relationship between the ministry of planning and project managers. With good management information, the planning ministry would be able to apply timely pressure on ministries to implement projects and assist them in resolving problems as soon as they arose. “We came to believe that transparency by itself would begin to induce managers to take more responsibility, insofar as it would be possible for people to follow what was happening with more clarity. Poor results would reflect badly on managers,” in the words of one member of staff. The IT-oriented staff in the planning secretariat, largely drawn from the ranks of those seconded by the Banco do Brasil, started to develop a prototype management information system (MIS) right away. “We couldn’t design a model and present it in a year or a year and a half. The requirement was immediate,” recalled Mauricio Albuquerque.22 The staff decided to translate the desire for a real-time MIS into a prototype design for a web-based system, considered a relative novelty in 1996. Each project was to be represented by a single web page that described the main elements of the project, including goals, investment levels, physical programming, and financial programming. The director in charge of strategic investments, André Amaral, recalled being told by Silveira that the prototype system needed to be on the president of the republic’s desk within forty days. Amaral recalled, “We were a group of ten people. It was crazy. I stayed up many nights without sleeping.”23

Recruiting More Talent. The planning secretariat staff was bolstered during

this intense period, although its numbers only grew by a handful. Silveira established a role for an individual to direct day-to-day work on Brazil in Action – in particular, to work closely with project managers, monitor projects, and speed up execution. Finding a qualified person was not necessarily going to be easy, since the presidential priority projects ranged from infrastructure projects, like gas pipelines and electrical power grids, to social projects, like one for in-service training of school teachers and restructuring the system of school finance. One place he looked for candidates was BNDES, and the name that several contacts mentioned was that of Aluysio Asti. A lawyer by background, Asti was unusual among BNDES staff because he had worked in the social area during the relatively brief period when the bank took a strong institutional interest in such aspects of economic development. Between 1982 and 1989, he had worked to finance small rural producers and help them to access low income urban markets. When BNDES closed down its socially oriented operations, Asti returned to the core activities. After heading the energy department, he became

22 Interview with Maurico Albuquerque, Brasília, August 2003. 23 Interview with André Amaral de Araújo, Rio de Janeiro, RJ, September 3, 2003. This interview was conducted for this project by Professor Armando Cunha of the Getúlio Vargas Foundation, Rio de Janeiro.

15

superintendent for the broader area of infrastructure in 1993. In August 1996, he received an invitation from Silveira to head up the Brazil in Action program. He recalled, “I never had any contact with him previously. But, as I had significant background in the bank in both infrastructure and the social area, I was a sort of rarity there. So he sought me out. Very soon after I arrived at the ministry in September, I realized that I was working with someone with enormous skill and vocation for project management – this man, Silveira. His competence and a capacity for work were far from ordinary.”

Socializing Managers. As soon as project managers began to be named,

Silveira and his staff met with them personally. They sought to inculcate the idea that managers were responsible for achieving results. The managers were reminded that the Brazil in Action projects would be sheltered from budget cuts. As Kandir recalled the message: “Don’t worry about financial resources. Instead of spending a third of your time working political contacts to secure resources, concentrate on the project. Negotiate the best form for the project, lower costs, raise the goals. You’ll have the money guaranteed.” The planning secretariat representatives conceded that managers did not have sufficient formal authority to be responsible for projects in a conventional administrative sense. However, Silveira and his team insisted that if a project went badly and the manager had not made the planning secretariat aware of the difficulties in time to overcome them, they would be considered personally at fault.24 In this connection, the managers were told of their responsibility to provide accurate and up-to-date information to the planning secretariat through the envisioned management information system.

Going Public with More. Within 40 days of the unveiling of Brazil in Action

to the government and media on August 8, the budget had been submitted to Congress, managers had been selected, and a prototype MIS had been built. The time had come to go public with the management system for Brazil in Action. The ministry of planning organized a media-oriented event, held on September 20. The codified fundamentals of the management model were presented. More strikingly, all of recently designated project managers – about 35 in number, as some were responsible for multiple projects – were on hand and formally introduced. The press were given more than these officials’ names and project responsibilities: they received a few biographical facts, such as about their field of specialization and age, as well contact details, such as telephone and fax numbers. In effect, the media were invited to regard the managers as sources of information about the presidential priority projects. The glare of publicity further underscored that managers were personally responsible for their projects. At the same event, the planning secretariat presented the conceptual design of the web-based management information system, for use by sectoral ministries, the ministry of planning, Casa Civil, and the president himself. Since the managers had yet to start working, the prototype did not include actual data.

24 Interview with Aluysio Asti, Rio de Janeiro, RJ, August , 2003. Since ideas about accountability, blame, and responsibility are culturally and linguistically specific, we quote Asti verbatim as follows:. “Em qualquer uma dessas instituições, havia uma pessoa que era o endereço do projeto para nós, a gente cobrava deles o desempenho. E ele tinha que apontar problemas. Se ele não apontasse problemas antecipadamente, ele era o responsável por aquilo lá. Essa era a idea.” In English, “In all these institutions, there was an individual who was the address of the project as far as we were concerned, and we held these people accountable. This person had to point out problems. If he didn’t point out problems in a timely way, he was responsible for that. This was the idea.”

16

Still, it was made clear how each project would be represented. (Two months later, however, at a similar event, the media were invited to see an operational version of the management information system, complete with actual data on the projects.)

Perceiving Risks of Bureaucratic Warfare. In parallel with this effort to apply

project management principles to Brazil in Action, through socializing project managers, developing a management information system, and attracting media attention, Kandir was attending to the Achilles heel of the whole administrative scheme. He and his people were on record saying that the presidential priority projects would be sufficiently funded, as we have seen. However, Kandir could not rely on his formal authority to secure resources for Brazil in Action projects. While Kandir’s influence over the formulation of the Federal budget was evidently strong, the planning ministry shared responsibility for budget execution with the ministry of finance’s treasury secretariat. The division of labor in the execution process was as follows. The planning ministry’s secretariat of budget and finance translated the annual budget into a program of monthly allocations, which was handed off to the treasury secretariat. The treasury secretariat would credit the accounts of the sectoral ministries on a monthly basis, with the amounts depending on many factors, including the financial programming and the pace of tax receipts. The secretary of the treasury had a relatively free hand in deciding how much to credit the accounts of sectoral ministries, but usually cared about the aggregate volumes rather than the allocations, the politics of which were usually handled by the planning ministry.

With this situation in mind, Kandir moved to establish a good working

relationship with the treasury secretary, Eduardo Guimarães, who happened to come into office shortly after his own move to the planning ministry. The two were not strangers to one another, as Guimarães served as president of the national statistical institute (IBGE) when Kandir had been secretary of economic policy. Kandir admired Guimarães: “He is a professor of industrial organization and had a micro view of things…a terrific person, super intelligent.” When Kandir held a press conference to present the 1997 budget that introduced special line items for the priority projects, Guimarães appeared with him -- for a specific reason: “The message was absolutely clear, that the ministry of planning and the treasury were in complete agreement about this aspect of the budget.”

Privately, however, Guimarães shared deep concerns with Kandir about how

the budget execution process would actually operate. The treasury secretary pointed out that Brazil in Action could easily fall victim to opportunism on the part of the sectoral ministries, a familiar kind of bureaucratic warfare known as the “policy of inverted priorities.” As he recalled,

Brazil in Action is a program that involves large expenditures and a priority program for the president, so it would be clever for the sectoral ministries to take the funds that treasury released to them and spend them on everything except for Brazil in Action projects. Then, the ministries could bring pressure on us directly and through the presidency to release

17

more money to them so that Brazil in Action programs wouldn’t stop.25

The source of this opportunism could be traced, in part, to the institutional

rules and routines for financial management in the federal government. Once the treasury credited a sectoral ministry’s accounts, neither the planning ministry nor the treasury could exercise any direct administrative control over how funds were spent. The only check on whether funds were spent in accordance with the budget was a financial control unit in the finance ministry (Federal de Controle) and the national audit body (Tribunal de Contas da União). The planning ministry could apply pressure, of course, but doing so would involve interacting with each sectoral ministry – a time consuming and possibly fruitless activity.26 What Kandir wanted to do was find a way to ring-fence the budget for Brazil in Action projects within the execution process. However, it was not obvious how that desire could be met.

Ring fencing, Symbols and Power. A practical solution to this problem

emerged during the subsequent six months, with close cooperation between the planning ministry and the treasury. From the start, part of the solution appeared to lay in modifying existing administrative procedures. As mentioned above, a routine task of the planning ministry’s budget and finance secretariat (Secretaria de Orçamento e Finanças - SOF) was to translate the annual budget into a program of monthly allotments to particular ministries. The idea was for SOF to program separate allotments for the Brazil in Action projects. To underscore the point, the treasury would execute this financial program by crediting sectoral ministries’ accounts with funds for Brazil in Action projects on a different day of the month than for other budget items. This method of ring-fencing money for priority projects was considered attractive, but largely symbolic, since sectoral ministries could still practice the policy of inverted priorities, if they wanted to do so.

In working out a fuller solution to this problem, Aluyisio Asti suggested that

standard loan administration practices followed by BNDES and most commercial banks could serve as a model for the federal government. As he recalled the thought, “When I finance a company – for instance, one building a steel plant – I pace the release of resources in tune with the project’s execution. If the project is not going well, I stop the flow of credit.” Building on this idea, Asti suggested a further alteration in budget execution procedures. The treasury would not execute the financial program for Brazil in Action projects unless the planning ministry was satisfied with their progress. To obtain the monthly allotment associated with these projects, the sectoral ministries would have to provide a formal request to the planning ministry. As a matter of practice, the request would be considered by the secretariat of planning and strategic investments, which was following the projects closely as part of operating the management system for Brazil in Action. These procedural adjustments would make it much harder for sectoral ministries to follow a policy of inverted priorities and would similarly provide incentives to keep the president’s priority projects on track.

25 Interview with Eduardo Guimarães, Rio de Janeiro, RJ, September 8, 2003. This interview was conducted for this project by Professor Armando Cunha. 26 Interview with Aluysio Asti.

18

The idea easily gained the support of the top people in the planning ministry – and of Guimarães, as well. This modest procedural adjustment was formalized in the 1997 version of the standard decree that gave the government’s budget execution plan legal effect and official presidential backing. As the treasury secretary recalled, “We didn’t change much from our standpoint, we continued to pass the money to the sectoral ministry. The difference was just that the money we allotted was only what Silveira indicated….This measure really gave a lot of influence to personnel at the planning ministry, because they would now determine when, if ever, the sectoral ministries would receive the money.” The significance was equally apparent within Silveira’s secretariat. In Asti’s words, “This was a gigantic alteration in the practice of administering projects in the federal public administration. This banality changed everything.”

Brazil in Action, Up and Running The Brazil in Action program enjoyed relatively smooth sailing for most of its

first calendar year of operation. In the course of 1997, the president’s travel schedule included monthly visits to Brazil in Action projects. The management information system became more sophisticated as the staff responded to user feedback. Many of the project managers began to perform their roles with considerable skill. Key projects moved forward, generating a stream of favorable news. The president was clearly pleased with the positive image that Brazil in Action was giving to his government, especially as he prepared to advocate a constitutional amendment permitting presidents to serve two consecutive terms. In this section, we examine briefly how the Brazil in Action program operated during its heyday.

Political Management and Social Communication. The public face of Brazil

in Action was managed cooperatively by the president’s ministerial-level social communication unit, headed by a senior career diplomat, Sérgio Amaral. This operation organized all the president’s travels in relation to Brazil in Action. The idea was clearly to enable the president to benefit politically from the whole effort, but also to keep up pressure on the ministerial bureaucracy. As Kandir put it, “all the time we worked with the symbolic presence of the president, emphasizing that this was his project of development, joined with the public, too.” The president’s own behavior helped in this regard. As recalled by Fernando Barros, the public relations executive, “the president was demonstratively positive and very active whenever there were media events related to the program.”

The political management of Brazil in Action did involve some differences of

opinion. An issue was whether the president should visit ongoing projects or only attend ceremonial ribbon-cutting events when projects were completed. The view taken by the media relations director within the planning ministry, Inacio Muzzi, was that the president should be seen visiting ongoing projects, just as Juscelino Kubitschek had done. However, he was overruled, because the president’s own communications team did not want Cardoso associated with incomplete infrastructure projects.

Media communications represented an important ongoing activity. According

to the director for strategic investments under Silveira, André Amaral, the president’s media operation made nearly daily requests to the planning secretariat to provide

19

written information about Brazil in Action projects. Working in concert with the media relations units in the sectoral ministries, the planning secretariat sought extensive coverage in newspapers, magazines, and television. The planning ministry’s press office, headed by Inacio Muzzi, organized a workshop for the press officers of all the participating ministries and state-owned enterprises to brief them on the whole effort. Muzzi drew together trained journalists to edit a bulletin containing up-to-date information on projects, which was fed directly to national and regional media outlets. This media team also worked with the ministries and state-owned enterprises to leverage their own publicity efforts. Toward the end of 1997, the government organized a television broadcast to publicize the anniversary of Brazil in Action. Appearing on the program were the president and all of the ministers involved in the program, with 700 people in the audience. The program, filmed in the Information Center of the Banco do Brasil, was transmitted live to state-level TV channels in six regions of the country. Overall responsibility for planning this event fell to Silveira’s director for strategic investments. André Amaral recalled, “I suffered a lot.”

Achieving Transparency. A lower-profile, but equally demanding activity was

to build up Brazil in Action’s management information system (MIS) – and to use this process as a vehicle to improve project management in the sectoral ministries. In terms of information inputs, an important step was to clarify the goals and actions within the projects in the social area. This conceptual design work for projects was undertaken by Asti in conjunction with the sectoral ministries. A similarly important step was to gain the cooperation of project managers in providing up-to-date information through the formal MIS system. Mauricio Albuquerque, who was directly responsible for gathering and analyzing the management information, recalled that many project managers asked him whether they really had to “fill out another set of forms”. The planning secretariat responded by selling the advantages of doing so. The standard message was,

No, put in the information if you’d like, but we strongly believe that if you don’t do so, you’re going to have problems. You’re going to have people who don’t want your project to go ahead. If you don’t put the correct information in the hands of those who want your project to move ahead, these interested parties won’t be in a position to help deal with those who oppose the project. And, remember, the ministry of planning is able to stimulate other actors, who are not directly part of the project, to resolve the problems you face.

The inputs to the formal information system, combined with the knowledge

gleaned from extensive face-to-face contact, was sufficient for the planning secretariat’s immediate purposes. It helped to pinpoint steps that needed to be taken to move forward particular projects. These issues were discussed at Silveira’s daily staff meetings, which began at 9:00 sharp and broke up promptly at 10:00.

As 1997 unfolded, the on-line information system became increasingly

sophisticated. One additional feature was a series of windows reflecting the management philosophy of identifying and eliminating constraints. One window identified problems that could come up in the near-term; another identified what could

20

be done to avert those problems. Access to the on-line MIS was limited to officials who needed to know, and the planning ministry counted the “hits”. Asti recalled that, “the president sometimes accessed the system once or twice a week.” Demand for access to the management information system grew rapidly as people became aware of it. For instance, international organizations, like the World Bank, started to see the system as a way to obtain information about the status of projects and programs. Suggestions for additional functionality were commonly received from these and other potential or actual users. The IT team in the planning ministry developed a new version of the on-line MIS every eight months.27

As a supplement to the on-line system, the planning secretariat issued a full

written report on the status of Brazil in Action projects on a bi-monthly rhythm. One section of each project’s entry outlined all the constraints inhibiting the near-term progress of a project, together with measures that were to be taken in overcoming them. This full report was sent up the line to the planning minister, the presidential staff, and the president himself. A summary report, stripped of detail on problems, was made available to the public. In time, word came down that the president and minister wanted more detailed information about some projects, for instance, information on social area projects at the regional level. The president also requested a monthly page-length report on each project, which he was known to take with him on long flights across Brazil.

Socializing Managers Some More. Training and socialization of project

managers was a major order of business. At first, Brazilian and international consultants specializing in project management were invited to give lectures. Later, specialist courses were offered. Specific training was offered to operate the management information system. However, the planning secretariat performed much of the task of socialization by modeling the role of a project manager. This group of professionals was helped in this by how Silveira performed his own role as secretary. In the words of his director of strategic investments, “He was precisely the maximum representation of what we wanted to see in a manager.” In Silveira’s own words, “the key to this process is not technique, it is a new attitude and posture – that of entrepreneurship. That is the essence of managerial change.”

Eliminating Constraints. A core routine activity of the planning secretariat became solving problems with particular projects. For the most part, the secretariat managed to find satisfactory resolutions to issues without elevating them to higher levels. As Kandir recalled, “Silveira, with huge patience and much competence, did this heavy work. He brought me only those problems where I had to talk to talk with the ministers.” Some of the coordination was conducted within the planning ministry – in particular, between the planning and budgeting secretariats. When a project was running into difficulty, a common practice was to get the discordant parties together in the same room. An example was an electricity distribution project, which was unable to proceed in the absence of an environmental license. The planning secretariat arranged a meeting between the federal environmental agency (IBAMA), the ministry of mines and energy, the project manager, and the secretariat itself. They were able to adjust the plan for the project so that the environmental agency was willing to issue the license. The planning secretariat came to believe increasingly in

27 Interview with André Amaral do Araújo.

21

the project management approach as they had come to practice it. “The problems became known because of the management methodology of Brazil in Action,” recalled Asti. “That’s not to say that things were perfectly marvelous – far from it.”

The planning staff felt they were on the right track for all sorts of reasons,

including the president’s demonstrative interest and consistent signals from the presidential staff and ministerial leadership and measured progress with particular projects. Other kinds of feedback were also important. A guiding idea of the program was that the projects – including the commitment of federal budget funds – would be induce further investment, especially in the infrastructure areas. Partly due to political pressure, the state-owned enterprises, state pension funds, and BNDES increased their financial role in the projects.28 Seeing an opportunity, the Inter-American Development Bank doubled the number of Brazil in Action projects it had initially sponsored. Japan’s Export-Import Bank provided credits to eight projects, up from their original two. In September 1997, Kandir was willing to write that, “We can comfortably say now that total investment in Brazil in Action projects for the year will reach R$ 31.7 billion, of which 13 percent are provided from the federal budget.”29

Part II. Case Analysis

The time has come to provide an analysis of the Brazil in Action experience,

so that it speaks the conference theme. I will pursue the idea that Brazil in Action can be analyzed as an innovative public management practice, related specifically to public management policies and the role of central coordinating agencies. The case commentary will satisfy two requirements for creating usable knowledge about innovative practices of public management: first, by providing a novel policy argument and, second, by providing insight into what organizational and managerial processes work -- and how. Viewed as a novel policy argument, Brazil in Action claims that it is important and possible for central coordinating agencies to use a management system, styled on project management ideas, to implement priority projects within a process of national development, provided that such a management system is adapted to the governmental system’s characteristics. Viewed as a functioning organizational process, the analysis of Brazil in Action offers practical insights to political leaders and public managers who wish to turn this policy argument into an actual innovation. In sum, by analyzing Brazil in Action as an argument and functioning organizational process, this part of the paper seeks to craft usable knowledge about public management policy.

Some Definitions. This statement of intent includes some terms that ask for definition. “Innovation” is both a nominalized noun and a verb. As a noun, innovation refers to a novel system-in-motion, considered as a functioning whole. As a verb, innovation refers to the process through which such a system develops and is stabilized. A “practice” is a limited-purpose practical theory of action that exists

28 See Kandir book p. 63. 29 “Discurso no I Seminário “Brasil em Açao,” Brasília, September 4, 1997.,” reprinted in Kandir (1998), p, 63.

22

within an “ecology of vicarious organizational learning” (Levitt and March 1986).30 A theory of action is prescriptive, but warranted by descriptive understandings. Practical theories of action – to be truly worthy of the label -- should meet appropriate academic standards of sound reasoning and presentation.31 By “limited purpose,” I mean that its subject matter is narrower than that of constitutional theory or management theory generally.32 “Public management policies” are government-wide institutional rules and organizational routines, whose administrative functions include guiding, motivating, and controlling the core public sector. Analytically, public management policies fall into several categories: expenditure planning and financial management, civil service and labor relations, procurement, organization and methods, and audit and evaluation (Barzelay 2001, 2003).

Relation to Public Management Policy and NPM. The Brazil in Action case program exhibits a set of public management policy choices. For instance, the management system -- which included the role of project managers, the management information system, and routines for resolving anticipated problems with project execution -- constituted public management policy choices related to “organization and methods” and “auditing and evaluation.” The delineation of priority projects within both the formulated budget and the programmed flow of allotments, along with the provisions for matching disbursements to project progression, constituted public management policy choices related to “expenditure planning and financial management.” Indeed, it could be said that the routines utilized to identify priority projects affected this same realm of public management policies.

Brazil in Action’s set of public management policies is different from familiar

cases of New Public Management (NPM). For instance, the Brazil in Action approach involves hands-on, results-oriented management control, rather than the relatively hands-off style of performance management using tools of management accounting and control employed in divisionalized corporations. The Brazil in Action approach does not attempt to change the way whole organizations operate. Rather, it 30 In public management, as in other domains of management, many practices are described in terms of administrative instruments or techniques, such as program budgeting and benefit-cost analysis. Other practices are keyed to performing types of organizational functions, such as strategic visioning (Barzelay and Campbell 2003), performance planning and resourcing (Thompson and Jones 2000), management control (Behn 2003), and production (Kelman 1986). Another category of practices is keyed to types of individual roles within governmental systems, such as the presidency (Neustadt 1960, Heifetz 1991), peak-level agency officials (Moore 1995), and regulatory inspectors (Bardach and Kagan 2002).30 A final category of practices to be mentioned is keyed to types of institutions, such as entire systems of public bureaucracies (Simon, Smithburg and Thompson 1950, Aucoin 1995) or types of organizations within such systems, such as mass adjudication agencies (Mashaw 1983), central agencies (Barzelay 1992), and auditing bodies (Barzelay 1996). Public management policies cuts across and combines these distinctions, as it is concerned with performing many of organizational functions mentioned above and especially with central agencies as types of institutions.. 31 For an interesting critique of a practical theory of management – specifically, the balanced scorecard -- as departing from such standards, see Norreklit (2003). 32 One genre of such practical theories of action are doctrinal arguments, a good example of which is Moore’s (1995) Creating Public Value. Another genre of practical theories of action are design theories, nicely illustrated by the core of Bardach and Kagan’s (2002) discussion of flexible regulatory enforcement in Going by the Book and by Mashaw’s (1983) discussion of mass adjudication in Bureaucratic Justice.

23

seeks to change the way selected activities are performed. The Brazil in Action approach seeks to make government work in a more joined up fashion, rather than allocating accountability for delivery along departmental lines. Brazil in Action involves direct central agency engagement in improving the management skills of public officials in sectoral ministries, rather than just providing an environment to permit or stimulate such development (e.g., generalized relaxation of government-wide institutional rules or creating awards programs, respectively).33 Finally, Brazil in Action involved changing organizational processes rather than reorganizing the machinery of government.

Two Cultures of Management Research? The question arises, how to do research that creates usable knowledge about public management policy or, for that matter, other aspects of public management? The introductory paragraph of this section indicates my position. This position chains together two rather different approaches. The first – viewing Brazil in Action as a novel policy argument – reflects the argumentative turn in policy analysis (Fischer 2003), which is typified by Majone (1989) in policy studies and by Hood and Jackson (1991) and Hood (1998) in public management. This perspective accepts that management practices in government should, and often do, result from practical reasoning and judgment. Such reasoning and judgment involves drawing debatable inferences from questionable presumptions and, inevitably, from unstated presuppositions (Toulmin 1958, Walton 1992, Simons 2001). The presumptions and presuppositions will naturally concern such matters as good and responsible government, the realities of organizational life, and empirical tendencies of particular governmental systems (Mashaw 1983, March and Olsen 1989, Moore 1995, Aucoin 1995). People will often disagree about these considerations and their bearing on public management practices. However, it is possible to deliberate rationally about such matters. Opinions may change and, in any event, bases for disagreement will become clearer.

The second approach is essentially to figure out what works and how. This

approach is usually associated with Herbert Simon (1946), and Bardach’s recent work is situated in this lineage. Bardach’s metaphors are mainly taken from the design sciences of engineering and architecture, rather than from philosophy, law, and rhetoric. The central question is why do types of organizational processes, which appear to work, actually do so. The subsidiary questions are, first, how are such “designs” clever in taking advantage of potentialities within situations and, second, how robust are the results with respect to variation in operating conditions. Smart practice analysis is styled, specifically, on reverse engineering. For these reasons, it is fair to assert that Bardach’s research paradigm plays down the idea that practices are plausible arguments.34

However, this characterization overlooks the introductory chapter, which lays

out what Bardach terms “craftsmanship theory.” Public managers are portrayed as working out sound arguments about what would be appropriate and effective to do in particular circumstances, with regard to the design of processes for performing the

33 On the U.S. experience, see Moynihan (2003). This description also fits the case of Italy’s central government, according to work in progress by Valentina Mele. 34 The general idea that policies are plausible arguments is presented in Majone (1989) and (Fischer 2003)

24

organizational function of delivery. Bardach’s practices analysis is designed to provide insights that managerial craftsmen should take account of as they fashion made-to-measure interventions that they believe will create public value. At least the managerial craftsman has to work out practical arguments as part of their work. Still, Bardach appears to focus on the “what works” question when examining his case studies.

Why Choose if Aim is Usable Knowledge? These two approaches –

exemplified by Hood (1998) and Bardach (1998), respectively – are clearly in competition in some sense.35 However, it seems to me that these approaches, while representing different scholarly outlooks and requiring different analytical tools, are highly complementary. For concreteness, consider the Brazil in Action case. An understanding of how this innovation functioned will remain irrelevant to the “ecology of vicarious organizational learning” unless the innovation is appreciated as a nuanced, even novel, argument. However, once the argument is so appreciated and attention turns to how it might be operationalized, Bardach’s style of practice analysis becomes exceedingly relevant. For these reasons, this section of the paper will try to give equal emphasis to providing critical and creative reflection about extant practical theories, on the one hand, and to Bardach’s method of providing a causal analysis of innovations as novel systems-in-motion, on the other.36

The remainder of this paper, in sum, sets forth two interpretations of the Brazil

in Action case. The first represents the innovation in terms of a novel argument about public management policy, which is particularly relevant to jurisdictions pursuing an activist developmental policy agenda through the core public sector in a governmental system with extensive institutional checks and balances within the executive. The second represents the innovation in terms of a functioning system-in-motion that appears effective in translating governmental priorities into coherent efforts to implement projects in an effective and timely manner.

Case Analysis: Brazil in Action as Policy Argument Viewed as a policy argument, Brazil in Action relates to extant practical

theories of public management policy. The best-in-breed commentaries on public management policies are Aucoin (1995) and Schick (1996). The authors present arguments in the following sense: they make claims about what public management policies are appropriate and potentially effective in defined contexts, and those analyses are rooted in broader understandings about government and management. What makes these arguments large-scale is that such understandings include public philosophies of government, empirical knowledge of governmental systems (including public organizations), and traditions of management thought (Barzelay

35 See Hood and Jackson’s (1991) seeming dismissal of the Simon tradition of scholarship and Bardach’s (1999) fairly negative review of Hood (1998), in the International Public Management Journal. 36 In my view, two impressive works in public management strike the balance correctly: Mashaw’s Bureaucratic Justice (1983) and Tendler’s Good Government in the Tropics (1997). However, these works make great demands on the analytically-oriented reader, are challenging to teach, and difficult to emulate. This latter claim can be made of Bardach’s book, as well. I think more scholars will get the balance right more of the time if they use the kinds of analytic and presentational tools used here.

25

2001: pp). These texts happen to be concerned with experience of the United Kingdom, Australia, and, especially, New Zealand. In the context of this discussion, the Brazil in Action argument covers relatively fresh ground. Project management approaches to management control are less common than accounting-based approaches. The argument considers how to take account of the tenacity of prerogatives of sectoral ministries in personnel and budgetary matters, which are stabilized by partisan power-sharing arrangements that, in turn, accommodate the country’s fragmented party system and strong executive branch. These considerations are not unique to this case (e.g., they apply in some measure to Germany), but differ from the situation in the NPM benchmark cases. The Brazil in Action argument also includes the proposition that in developing countries (at least), the state should lead a developmental process. A subsidiary argument is that the head of government, central agencies, and sectoral ministries hold substantial responsibility for carrying out such a process, albeit with non-governmental partners. These arguments are not common in the NPM benchmark countries, and it is not at all clear that they are endorsed by the most vocal commentators on public management policies in developing countries, such as public management specialists at the World Bank (Tendler 1997). Establishing Contact with Large-Scale Arguments. To prepare the ground for deliberation on these matters, let us try to model the Brazil in Action policy argument. In doing so, I bring to bear the tools of argumentation analysis that I refined in an article and book chapter on, “How to Argue about the New Public Management” (Barzelay 2001a, 2001b). In that context, I maintained that the form of large-scale policy arguments about public management policies in the best-in-breed works is:

(1) CPMP = A (PPG, KG, MAN), where CPMP is the set of claims and the warrants for those claims are PPG, KG, and MAN. The term, PPG refers to public philosophies of government. KG refers to empirical knowledge of governmental systems. MAN refers to conceptions of management practice. These terms can be specified in various ways, just like in mathematical formulae (which this is obviously not). For instance, I suggested that Aucoin’s PPG is defined more in terms of institutional properties than Schick’s, which is defined more functionally, in terms of creating public value. Similarly, Aucoin’s management ideas come more from principal-agent theory and Brodtrick’s category of well-performing organizations, while Schick’s come from mainly strategic management thought and theories of management accounting and control systems. The Claims. If Brazil in Action is an argument, it can be understood – in part – by identifying its characteristic claims, that is, specifying the term, CPMP. My interpretation is that the claims are a chain of five statements, as follows:

C1 Central agencies should choose public management policies that enable them to discharge their responsibilities for implementing a developmental policy agenda

26

C2 Such responsibilities can be discharged by operating a coherent management system for a limited number of priorities

C3 Such priorities should and, in principle, can, be established by a

process that brings to bear sound analysis and the authority of the head of government

C4 A coherent management system should be designed and operated so

that it works around institutional checks and balances built into the executive branch and the wider governmental system

C5 The specific design of the management system should reflect the

guiding ideas and techniques of project management

As suggested above, these claims are distinctive within contemporary discussions of public management policies, even those concerned with developing countries.

From Claims to Argument. In understanding Brazil in Action as an argument,

it is important to look at the rationale. As anyone who knows about practical argumentation will attest, the complete argument cannot be written down. Some presuppositions will not be stated (hence, there will be enthymemes). Not all inferences can be stated. So, in examining the rationale, I will simply identify the warrants for the claims. I will proceed in two steps. First, identifying the warrants, by specifying PPG, KG, and MAN. Second, associating the warrants with claims.

Specifying ‘Management.’ In analyzing Aucoin and Schick, I identified three relevant schools of thought about management (MAN): strategic management (SM), management accounting and control (MAC), and well-performing organizations (WPO).37 In the interest of policy dialogue, let me formulate the “right-hand side” of the Brazil in Action argument in accord with this same pattern of argument. Two basic propositions about SM, which can be found in any textbook on the topic, are clearly part of the Brazil in Action argument:

SM1 Peak-level executives should define organizational priorities SM2 Organizational structures and management processes should be

adapted to fit their functions and the environment The school of thought called management control (MAC), as applied to

projects, can be brutally reduced to the following:

MAC1 Effective and efficient delivery of projects requires a management system for projects:

MAC1a individuals with personal responsibility for project schedule,

cost, and performance;

37 In Aucoin, principal agent theory played a role in PPG and in MAN. Within MAN, principal agent theory is mainly associated with management accounting and control (MAC).

27

MAC1b a management information system for projects;

MAC1c procedures for diagnostic control and problem-solving; and

MAC1d resource allocation procedures that incentivize and reward individuals and groups for meeting or exceeding project targets

This doctrine can be extended to include the theory of constraints principle

that the effective and efficient management of projects requires an ethic of identifying and eliminating constraints on performance in an anticipatory manner (MAC2).

The looser school of thought of well-performing organizations (WPO) can be

specified in terms of project management as well:

WPO1 Effective and efficient management of projects requires employee commitment, interpersonal trust, and organizational decompartmentalization

WPO2 Commitment can be elicited by the attractiveness of missions, meaningful performance feedback, and a sense of personal responsibility for achievement

WPO3 Trust can be established through the design of management processes and interactive performances by leaders and followers

WPO4 Decompartmentalization can be established through planned and spontaneous flows of information across organizational boundaries

Specifying ‘Public Philosophies of Government.’ It is widely accepted that generic management ideas do not suffice as warrants for public management practice. For one reason, they are deafly silent about what government is for and about matters of public accountability. Aucoin’s book is outstanding in discussing public philosophies of government (PPG) in dialogue with “traditional public administration” and arguments rooted in the new institutional economics. Brazil in Action’s argument – patterned on the political ideas of the Cardoso presidency -- does have something to say on these issues, too:

PPG1 Good government provides for macroeconomic stability and leads a national process of development

PPG2 Good government requires choosing projects and programs that solve collective problems that the market cannot be expected to solve

PPG2a Collective problems in Brazil include the poor life chances of a large fraction of the population

PPG2b Collective problems in Brazil include the structure of costs that make exporting goods difficult or insufficiently profitable

PPG3 Good government requires implementing such projects and programs in an effective and efficient manner

PPG4 Good government requires professional, knowledgeable, and experienced individuals in positions of authority within the core public sector

28

PPG5 Good government requires involvement of institutions outside the core public sector to perform various functions of policy implementation and “delivery”

Specifying ‘Knowledge of Government.’ A second reason why generic

management ideas do not suffice as warrants for public management practice is that they are not codified from experience where they operated in actual governmental systems. Not only is government different from commercial enterprise, but governmental systems differ among each other in important respects, such as form of the state, form of government, party systems, bureaucratic culture, and so on. My interpretation of Brazil in Action’s argument includes most of the following observations about government and politics in Brazil.38 I use the symbol, KG, to denote “knowledge of government.”

KG-1 Brazil’s state is a federal republic.

KG-2 Brazil’s form of government is presidential at the federal level

KG-3 The party system is fragmented, with relatively fluid party identification and only one mass membership party

KG-4 The federal government’s executive branch exhibits institutional checks and balances, owing to constitutional arrangements, dynamics of building coalitions capable of governing, and the culture of (public) bureaucratic organizations

KG-4-1 Coalition building dynamics include logic of appointing ministers from parties other than that of the president; therefore the ministers’ loyalty to the president may be attenuated

KG-4-2 Coalition building dynamics usually include placing members of the president’s party at the top of central agencies (including ministry of finance and ministry of planning); therefore, the president can rely on these organizations to exert administrative influence

KG-4-3 Constitutional arrangements include the lack of requirement for the government or ministries to execute the Federal budget as formulated; therefore, many decision rights in the budget execution process are assigned to the treasury (ministry of finance) and top budget holders in sectoral ministries

KG-4-4 The culture of bureaucratic organizations includes norms of hierarchical loyalty and subordination; therefore, liaison devices between a given ministry and other parts of the executive are difficult to build and sustain

KG-5 The federal government’s personnel regime is variegated, with highly professionalized career systems in a number of institutions and domains (e.g., diplomatic corps, central bank, National Bank for Economic and Social Development, and, increasingly, planning and budgeting).

38 For analyses of Brazil’s governmental system, see, for instance, Roett (1999).

29

Visualizing the Argument as a Whole. The figure below indicates how the various claims making up CPMP are backed up by specific considerations of public philosophy of government, knowledge of government, and systematic management thought. Except for C1, the claims are also backed up by earlier claims, since this argument is a compound one, involving five units of argument linked together. When a term appears without a subscript, it refers to all claims of that type; for instance, SM refers to both SM1 and SM2. Claims Key Warrants

C1 Central agencies should choose public management policies that enable them to discharge their responsibilities for implementing a developmental policy agenda

SM2

PPG1 , PPG2, PPG3 , PPG4

C2 Such responsibilities should and can be discharged by operating a coherent management system for a limited number of priorities

SM, MAC1

C1

KG-4-1

C3 Such priorities should, and can, be established by a process that brings to bear sound analysis and the authority of the head of government

SM C1

KG-2, KG-4-2

C4 A coherent management system should be designed and operated so that it respects and works around institutional checks and balances built into the executive branch and the wider governmental system

SM2, MAC C2 , C3

KG-1, KG-4

C5 The specific design of the management system should draw its guiding ideas and techniques from systematic thought about the practice of project management

MAC, WPO C4

KG-5

Table 1: Brazil in Action as Policy Argument, Summarized

The point of this exercise is as follows. First, it should be noticed that Brazil in Action exhibits or reflects a large-scale policy argument that is based on much more than generic management ideas. Second, this pattern of argument is different from familiar arguments about public management policy. It may be a more relevant argument for developing countries than most NPM arguments about public management policy. It may also be relevant to governmental systems – even in developed countries -- with some of Brazil’s characteristics, such as those listed under KG-4. Third, modeling arguments makes it easier – and faster -- to compare, combine, and assess them.39 In this way, we can make full use of the human capacity to engage in practical reason according to appropriate standards when considering matters of public management. We can easily locate the basis for disagreement or even change reasoned opinions. Equally, the exercise shows that thinking of practices as arguments is not entirely satisfying by itself. Thus, it is time to change the frame and consider how Brazil in Action functions as a practice that puts this argument into operation.

39 Imagine how much easier it would be to parse the arguments in such highly crafted, discursive texts in public management as Mashaw (1983), Tendler (1997), and Bardach (1998) if the arguments had been written up in this fashion and presented as such in an appendix.

30

Brazil in Action as Analytic Narrative of a Functioning Organizational Process

In this section, I develop an analytic narrative about the Brazil in Action program.40 At least two analytic narratives could be crafted about this experience. In one, innovation would be conceived as a verb. The case outcome would be the Brazil in Action management system and the associated “routines” of presidential leadership.41 The episode’s trajectory would represent the innovation’s development.42 As mentioned above, this analytic narrative is not developed here for reasons of space.43

The Case Outcome. In the other analytic narrative, innovation is conceived as

a noun. The case outcome to be explained is the pattern of action in operating the management system (defined broadly). This outcome is of analytic interest not because it is a puzzle: the outcome is intriguing and pleasantly surprising, but it is not puzzling in relation to existing positive theory and knowledge. It is rather of analytic interest because the Brazil in Action “system” worked rather well -- and under conditions where clinical wisdom or preconceptions suggest it would not be expected to do so. In other words, the outcome is of policy rather than theoretical interest.44

The Research Tradition. This outcome is to be explained using a mode of

analysis that is consistent with Bardach’s practices analysis, while more broadly utilized by management and political science researchers. I will call this approach “institutional processualism,” although it could be called “systemic processualism.” This theoretical approach is a variant on processual analysis as described by Pettigrew (1997). In political science, institutional processualism is exemplified by March (1994), Steinbruner (1974), and Hall (1992). Institutional processualism seeks to understand the process dynamics of political and organizational phenomena, such as policy decisions (Kingdon 1984), organizational change (Campbell and Wilson 1995), strategy development (Barzelay and Campbell 2003), leadership interventions (Heifetz 1991), public service delivery (Tendler 1997), regulation (Mashaw and Harfst 1987), and various forms of contentious politics (McAdam, Tarrow, and Tilly 2001).45 This research “tradition” is processual in the spirit of Cyert and March’s (1963) behavioral theory of the firm, but also “systemic” in its concern with the effects of institutional (and other kinds of) context on the trajectories and outcomes of

40 The term comes from Bates, et. al. (1998) 41 On the concept of case outcome, see Ragin (1987). 42 On the concept of episode trajectory, see McAdam, Tarrow, and Tilly (2001). 43 An example of such an analytic narrative can be found in Barzelay and Campbell (2003), chapter 5. 44 A similar justification for choosing case outcomes – and for that matter, the cases themselves – is evident in Tendler (1997) and Barzelay (1993). 45 Speaking autobiographically, I would place the following works under this header: Barzelay (1986), Barzelay (2001b: chapter 3), Barzelay and Fuechtner (2003), Barzelay and Thompson (2003), Barzelay and Campbell (2003), and Barzelay (2003). I thank Jostein Askim for suggesting the term processual institutionalism and Jerry Mashaw and Mikhail Pryadilnikov for suggesting the more suitable term, “institutional processualism” during a recent series of informal seminars at LSE.

31

dynamic political and organizational phenomena.46 Narrative as against statistical explanation is its stock in trade (Abbott 1992). Institutional processualism is less cultural in orientation than much of the new institutionalism in organizational sociology, more actor-centered than much of historical institutionalism, and less choice-theoretic than much of rational choice institutionalism.47 Institutional processualism is grounded in such sociological accounts of individual action as the logic of appropriateness (March and Olsen 1989, March 1994), or such social psychological accounts as “sense-making” (Weick 2001).48 Institutional processualism also examines social phenomena at the collective level. In doing so, it often employs social mechanisms as tools of causal explanation (Hedström and Swedberg 1996, McAdam, Tarrow, and Tilly 2001).49

My interpretation of Bardach’s version of institutional processualism is as follows. Patterns of action are flows of effort. Any discrete effort can be attributed, in part, to an actor’s interpretation of an appropriate fit between his or her identity, on the one hand, and the situation at hand, on the other (March and Olsen 1989, March 1994). This logic of appropriateness operates with an actor’s repertoire of identities as given. The situation is not given, however. The situation is constituted, in part, by actors’ interpretations of other actors’ past or anticipated efforts and, in part, by actors’ interpretation of the context, whether this context is stable or in motion. Explanation of action under the logic of appropriateness involves verstehen or construing, through empathy, how actors have interpreted and acted upon their identities and the situation. However, institutional processualism seeks to say more – from an observer’s perspective -- about how the situation facing actors comes to be constituted. It is at this point that social mechanisms become relevant. Social mechanisms help explain the effects of actions at one time on the situation facing actors at a later time. Social mechanisms can also help explain how relatively aggregate properties of the context are causally linked to actions and efforts.

Ordering Data. What is the situation facing actors in cases like Brazil in Action, considered as a functioning organizational process? One aspect of the situation is the design of the organizational process. For expositional convenience, let’s call this aspect of the situation, “the process design.” In the Brazil in Action case, the process design includes the project manager role, the management

46 The distinction between processual and systemic approaches to organizational theorizing is set forth in Whittington (1993). 47 These broad-brush claims derive their categories from Hall and Taylor (1996) and Lichbach (1997). 48 The following quote, from McAdam, Tarrow, and Tilly (2001: 56), exhibits a combination of the March/Olsen and Weick approaches: “Actors, in our view, are not neatly bounded, self-propelling entities with fixed attributes, but socially embedded and constituted beings who interact incessantly with other such beings and undergo modifications of their boundaries and attributes as they interact. Contentious politics does not simply activate pre-existing actors but engages actors in a series of interactive performances that proceed through incessant improvisation within broadly defined scripts and organizational constraints.” 49 The specific approach employed here combines an analytic framework evident in chapter 5 of Barzelay and Campbell (2003) and the more explicit explanatory strategy set out in my synthesis piece for a symposium on public management policy-making published in the International Public Management Journal (Barzelay 2003).

32

information system, and routines for responding to identified constraints, communicating with the press, and scheduling the president’s public appearances related to Brazil in Action. A second aspect of the situation facing actors is the environment within which the process design operates. I use the term “process context factors” to identify relevant properties of this (often changing) environment. What is eligible to be considered as process context factors, in the present study, are aspects of the situation that fit such political science categories as institutions, ideas, related policy developments, and the political stream.

A process design is a complex artifact. I find it helpful to view the inner characteristics of such a system through four conceptual lenses. The first views the process design as a body of guiding (and legitimating) ideas, which include the accepted beliefs about means-ends relationships and, possibly, terminal values. The second views the process design as governance arrangements, which set out roles, information rights, decision rights, and the like. The third views the process design as a flow of experience, interpreted as linked, discrete events. The fourth views the process design as a repertoire of accepted cognitive techniques used to gather and interpret information. In sum, a process design is a system of interacting parts that can be disaggregated and re-assimilated into four categories: guiding ideas, governance arrangements, structured events, and cognitive techniques.

Introducing Social Mechanisms. To understand or explain a case outcome, we need to see how a great number of process design features, operating in a context, affect the actions and efforts of such people as officials in the planning secretariat and the managers in the sectoral ministries. More than that: for purposes of providing usable knowledge, the explanation of the case outcome needs to provide more theoretical insight than would be offered by a close-to-the-ground analytic narrative. For this purpose, I emulate the methodological approach of McAdam, Tarrow, and Tilly (2001). This team of social scientists seeks to provide theoretical insight into their subject – contentious politics – by examining the workings of various kinds of social processes and social mechanisms.

What counts as a social mechanism is not entirely agreed in social theory and political science. Among those who work within institutional processualism, the criteria for being an apt social mechanism (for purposes of analysis within institutional processualism) appear to include:

The causal story is a well-defined, recurring tendency within social psychology, sociology, economics, and/or political science

The causal story involves the effect of belief on action and a feedback effect of

action upon belief

The causal story can be fitted into a logic of appropriateness or sense-making explanation at the individual level

The causal story’s influence over the pattern of action is contingent on

institutional or other kinds of context factors.

33

Some social scientists, such as Elster (1996), tend to emphasize the first and second of these criteria, but not the last. By contrast, institutional processualists, such as McAdam, Tarrow, and Tilly, tend to consider all four criteria. As a consequence, institutional processualists often explain case outcomes by reference to “relational processes,” which can be disaggregated into a number of component social mechanisms.50 The activation and operation of such compound social mechanisms as relational processes is linked causally to process context. The building blocks of explanation are often these compound mechanisms. A good example of such a compound mechanism in the public policy literature is the policy spillover effect (Walker 1977, Kingdon 1984). This compound mechanism is a scenario that indicates a causal influence of policy developments in one domain upon subsequent policy developments in another. The policy developments in the first of these domains are a “context factor.” The influence is felt upon the agenda setting or alternative-specification process in the second domain. The more primitive mechanisms include diffusion of ideas, conceptual shifts, and inferring means-ends causal relationships from interpretations of vicarious experiences of policy change.

To explain the Brazil in Action case outcome, I examine the causal influence

of three “compound social mechanisms” or relational processes in the terms of McAdam, Tarrow, and Tilly’s framework. Two of these building blocks come directly from their book, Dynamics of Contention, while the third comes from processually oriented organization theory (Greve 2003) and draws from Cyert and March (1963). The chosen compound mechanisms are actor certification, attribution of opportunity, and performance feedback.

Anticipating Prescription. Before discussing these compound mechanisms in

detail, let me pause briefly to indicate why my analysis goes to the trouble of analyzing the Brazil in Action case in this fashion. Put aside the reasons having to do with creating a common methodology for the generation of usable knowledge about public management innovations and general arguments for transparency and rigor of explanation. Suppose that a political leader or public official is prepared to buy the argument presented in the previous section of this part of the paper. Suppose that they proceed to formulate an administrative design problem of the sort: how to choose a set of public management policies for implementing presidential (or prime ministerial) priority projects in a timely manner without increasing government spending. In formulating a solution to this administrative design problem, it is possible to regard Brazil in Action as a design exemplar (Schön 1983). If the intention is to emulate Brazil in Action, in the sense of copy it, then all that has to be described is its system of process design features. However, copying is a risky mode of administrative innovation, if for no other reason than the likelihood that the process design’s operation depends on the process context, which may differ between the source and target situations (Rose 1993). Therefore, the design exemplar should be theorized so that in crafting a solution to the administrative design problem for the target situation, the managerial craftsman can apply knowledge of the causal texture of the exemplar.

50 A background justification of this approach is couched in these terms by McAdam, Tarrow, and Tilly (2001: 30): “We employ mechanisms and processes as our workhorses of explanation, episodes as our workhorses of description. We therefore make a bet on how the social world works: that big structures and sequences never repeat themselves, but result from differing combinations and sequences of mechanisms with very general scope.”

34

I believe this is the essence of Bardach’s position about case research on administrative innovation.

Standing on the shoulders of Bardach and self-conscious institutional

processualists, I suggest that the causal texture of the exemplar can be apprehended – in a way that is relevant to formulating solutions to administrative design problems -- if the analyst examines the workings of compound social mechanisms, such as the ones listed above. In other words, managerial craftsmen can proceed with the idea that their designed system should seek to activate the social mechanisms of certification, attribution of opportunity, and performance feedback. In proceeding with this idea, managerial craftsmen might model their system of process design features on those observed in Brazil in Action, but there is no reason to copy them.

Three Social Mechanisms Defined. Having now previewed the punch line, let

us return to the main flow of discussion by briefly commenting upon how the social mechanisms of this study are defined. McAdam, Tarrow, and Tilly (2001: 121) define certification in terms of “the validation of actors, their performances, and their claims by external authorities.” Attribution of opportunity “is an activating mechanism responsible in part for the mobilization of previously inert populations” (43). Performance feedback entails the production, handling, and interpretation of information about effort outcomes, in the light of previously established aspirations and goals. The concept of actor certification needs to be stretched slightly from its moorings in the study of contentious politics to analyze stable organizational processes. The mechanism involves more than validating actors’ claims. It involves making up people as agents of abstract responsibilities, whether thought of as a broadly defined function or a more elaborate programmatic idea (Goodin 1992, Miller and Rose 1990). For instance, people can be made up as “public managers” with responsibility for performing the function of delivery in an effective manner, with effectiveness defined in terms of achieving results. The concept of actor certification can be related to the logic of appropriateness. When this mechanism operates, it can affect actors’ identities. Further, it can affect perceptions of these actors within a collectivity that has witnessed the events through which certification has publicly occurred. In this way, certification can influence the situation as well as the identity of actors.

The concept of attribution of opportunity is of central importance to studies of

policy change, especially to Kingdon’s idea that policy entrepreneurs respond with intense effort to situations where they perceive that the window of opportunity may open. This social mechanism is no less relevant to the study of administrative innovation, where this idea is viewed as a verb. It does seem to play a role in the study of administrative innovation, viewed as a noun (as here). One can imagine that an actor is certified as a public manager and operates within a management system, but does not engage in a pattern of sustained, creative effort to perform the function for which he or she has been certified. One reason for the effort shortfall could be the attribution of the likely futility or perversity (Hirschman 1991) of the efforts. In this context, the attribution of opportunity mechanism relates to a belief that the outcome of the efforts involved will be consistent with aspirations and certified responsibilities.

35

The concept of performance feedback, as a mechanism, probably needs no special comment here. However, it should be mentioned that the process design features of a typical project management system are precisely intended to activate this mechanism. A more general point worth mentioning here is that each of these compound social mechanisms – actor certification, attribution of opportunity, and performance feedback -- is no more than a building block of the analysis. They can be seen as combining in actual situations, to produce a synergistic effect. This general perspective on social mechanisms as tools of social theory building is expressed powerfully by Gambetta (1996) and is reflected in the work of many institutional processualists.51 In McAdam, Tarrow, and Tilly’s (2001: 27) words: “Mechanisms seldom operate on their own. They typically concatenate with other mechanisms into broader processes. Processes are frequently recurring causal chains, sequences, or combinations of mechanisms.”

The Analytic Narrative in Substance. Returning now to Brazil in Action,

considered as a functioning organizational process, we can proceed to build a usable structure on top of these theoretical foundations. The basic causal claim is that the case outcome is explained by the activation of three social mechanisms, which operated in combination: actor certification, attribution of opportunity, and performance feedback. These mechanisms were individually and severally activated by a mix of process design features and process context factors. The process design features included such guiding ideas as a program for implementing presidential priority projects; such governance arrangements as establishing the planning secretariat as a clearance point in the project-related branch of the budget execution process; such structured events as set out in the management system; and such cognitive techniques as borrowed from canonical practices of project management. The process context factors included such institutions as the presidency, central coordinating agencies, and sectoral ministries; such ideas the importance of infrastructure to development; such related policy developments as the managerial reform of the state; and such currents in the political stream as sustained media attention to the president’s engagement with Brazil in Action.

A relatively simple representation of this analysis is Table 2. The rows are

defined in terms of the social mechanisms of actor certification, attribution of opportunity, and performance feedback. Each row includes two columns. One lists the mechanism’s activating and sustaining process design features, while the other lists its activating and sustaining process context factors. A brief discussion of this table follows, as the reader can fill in the details by referring back to the narrative historical account of Brazil in Action.

Table 2 about here

Actor certification. The case outcome can be partially attributed to the

activation and operation of the actor certification mechanism. The basic idea is that actors’ efforts were influenced by their evoked identities as results-oriented public managers on whose shoulders rested responsibility for making Brazil in Action work. Planning secretariat officials and managers’ identities as agents of results-oriented implementation of presidential priority projects were evoked by Brazil in Action’s

51 Gambetta refers to the “concatenation of mechanisms.”

36

structured events (such as public release of managers’ names), guiding ideas (such as the Brazil in Action rhetoric), and governance arrangements (such as the structural position of these agents in central coordinating agencies or sectoral ministries).

The effect of these process design features was strengthened by process

context factors. For instance, the planning secretariat officials and managers in sectoral ministries received some certification by virtue of the their positions’ placement in a scalar chain leading to the office of the president of the republic. The identity of a results-oriented public manager was present in the repertoire of many of these actors because of the career systems within Federal organizations, such as the National Bank for Economic and Social Development (BNDES) and Petrobrás. These institutions not only provide career incentives that are conducive to a results-orientation, but have historically been sites for producing cadres of individuals certified as agents of national development. These people were accessible as certified agents of results-oriented management, in part, because of personnel rules that allowed for ease of movement between such governmental organizations and the ministry of planning. These institutional context factors – positions linked to the president via a scalar chain, careers, and secondment provisions – all strengthened the effects of the process design features on actor certification.

The actor certification mechanism was also strengthened by three other types

of process context factors: ideas, related policy developments, and the political stream. The idea of project management (Graham and Englund 2004) is widely known as a practical theory of management control for project-like activities. For this reason, it was possible to invite international and national experts in project management to training and motivational events during the Brazil in Action program. The existence of this model of management (Guillén 1994) as an easily traveling idea (Sahlin-Andersson 2002), both domestically and internationally, added force to the program’s own rhetoric of project management. The collective representation that Brazil was embarking on a course of public policy-making apart from conserving macroeconomic stability and involving an appropriate governmental role in the development process surely contributed to actor certification. The managerial reform of the state, centered in the ministry of federal administration and reform of the state (Gaetani 2003), helped stabilize identities around special responsibility for improving public management within the core public sector of Brazil. The sustained media attention of the issue of development and of the president’s involvement in Brazil in Action did much to augment the influence of the guiding ideas, governance arrangements, and institutions on actors’ efforts.

Attribution of opportunity. This mechanism involves a two-stage link: first,

from process design and context factors to actors’ perception that Brazil in Action could actually work and, second, from this perception to their pattern of effortful action to make it work. The demonstration of the possibility that it could work included the guiding idea that budget funding for priority projects would be guaranteed (despite the ever present possibility of reductions in federal spending). Without this credible commitment, actor certification for attaining results would presumably have been diluted to some extent. Another important process design feature was the modification of budget execution procedures for priority projects, so that the planning secretariat became a clearance point (Pressman and Wildavsky 1974) in this process. As we have seen, one planning secretariat official wryly

37

observed that “this banality changed everything.” One way it changed everything was to eliminate the perception that Brazil in Action would simply play into sectoral ministries’ policy of inverted priorities, the result of which would be that projects would not be fully implemented and public spending would actually exceed finance ministry intentions. In a sense, this process design feature inoculated Hirschman’s reactionary rhetorics of futility and jeopardy for Brazil in Action. A final process design feature was the implicit rule that disagreements between the ministry of planning and the sectoral ministries could be elevated to the presidency (specifically, the Civil House), overcoming the futility argument that the former’s efforts would be cancelled by the latter’s possible intransigence.

The attribution of opportunity mechanism was also activated by process

context. Unquestionably, an important such context factor was the idea that the government should and could do something in the development arena since hyperinflation had been eliminated. In a hyperinflation environment, no one could plan on anything, let alone the implementation of a complex, risky multi-year undertaking. For one reason, the federal budget was a meaningless document, with the treasury the only effective player in the resourcing game.

Another importance context factor was “closure” (Power 1997) about the basic

assumptions built into Brazil in Action: first, that infrastructure was a priority area for governmental efforts and, second, a management system was needed so that the federal bureaucracy would operate effectively in implementing the president’s priorities. I have listed these assumptions under the header, “related policy developments.” The source of such closure was, in part, the events through which Brazil in Action came to be established. I treat these matters as context, because the consensus on these questions was a given with respect to the operation (as distinct from the creation) of Brazil in Action.

The political stream contains a number of significant contextual influences of

relevance to the attribution of opportunity. First, the rhetoric about presidential priorities and guaranteed funding would have rung hollow if the top echelons of the finance and planning ministries were not known to be loyal to the president. Hence, if the president’s quota of ministerial positions in his coalitional arrangement had been smaller, an attribution of opportunity could easily have been undercut. Second, the president was in a strong position in terms of public opinion, with good prospects for winning approval of a constitutional amendment permitting re-election to a second term. Third, media attention, picking up on the government’s own communication efforts, made the president particular committed to following through on Brazil in Action.52 This political context clearly reinforced the favorable attributions of opportunity stemming from the process design features, given the delivering on the guiding ideas depended on a cohesive presidential team situated within the most relevant institutions.

Performance feedback. This mechanism is basically activated by those

process design features that are most closely associated with stereotypes of project management, as can easily be seen from Figure 2. Just two points are to be stressed

52 Parenthetically, this last statement indicates how process design can come to influence process context.

38

here. First, this mechanism affects the case outcome via two different channels. One is that by enacting the practice of project management, actors are conducting themselves appropriately in light of their certification. The other is that they themselves receive performance feedback, as when they meet or fail to meet expectations for providing accurate information crucial to taking timely corrective action. It was also the case that some managers were relieved of their duties when failing to meet the standards.

In sum, the Brazil in Action program – considered as an administrative

innovation – worked effectively as a functioning organizational process in implementing presidential priority projects because its process design features took advantage of opportunities presented by the process context. The combination of these features and the context activated and sustained three social mechanisms, actor certification, attribution of opportunity, and performance feedback. These three mechanisms, working in concert, shaped actors’ identities and situations in such a way as to produce a flow of actions and efforts to identify and eliminate obstacles to project implementation. The analysis suggests that would-be designers should choose process design features – guiding ideas, governance arrangements, structured events, and cognitive techniques -- that will activate this same system of social mechanisms. The specific process design features should be appreciated for their functional role in activating these mechanisms and, ultimately, shaping actor efforts. But these process design features should not be unthinkingly copied, if for no other reason than the probability that, without modification and realignment -- they will be less than congruent with the process context factors in the target situation.

Limitations to this analysis. I would like to conclude this section by pointing

out some limitations to the analysis of Brazil in Action as a functioning organizational process. First, this is not a comparative study. Second, I have not tried to provide a fully satisfactory explanation about the progression of any individual project. My focus has been on Brazil in Action as an administrative innovation, not a substantive or technical innovation in the policy domain of infrastructure creation. Such could be pursued by a similar methodology, but would require analyzing a different aspect of the episode than covered here. Third, and relatedly, I have not considered ways in which a scheme like Brazil in Action could create adverse consequences. For instance, I have not questioned the presumption that the projects met all appropriate tests for public value creation, within the context of Brazilian democratic institutions and processes. As many interviewees pointed out, an efficient system to implement bad projects does little good. Fourth, I have not tried to account for how Brazil in Action came into existence as a functioning organizational process, in favor of focusing on the innovation as a noun. However, the methodology for explaining this other case outcome would be similar to that employed here. Indeed, the certification, attribution of opportunity, and performance feedback mechanisms are relevant, as are most of the process context factors.

Fifth, the analysis largely omits an important “variable,” namely the

reputations and interactive performances of individuals such as José Paulo Silveira. In the words of the planning secretariat’s senior staff:

Silveira had a fundamental role. He wasn’t a puppet, very much to the contrary. He was exactly the maximal expression

39

of what we wanted in managers….He obliged us to accept the same discipline we wanted to see in the managers….This was very good, very tiring and stressful, but very good so that we could go on maturing the process.

And to quote another senior staff member: Silveira is a very disciplined person, apart from being extremely qualified. He’s tireless. He is one of those workaholics that has a commitment with something….He undertook personally to call to here, call to there, fight here, fight there. I remember battles for budgets and such, he was always present. I think that he was “a big difference.”

The concession about an omitted “variable” can be overstated, however. This

“variable” is not at all omitted from the narrative history presented in Part I, only from the analytic narrative in Part II. It shows the importance of being able to adopt a critical perspective on analytic narratives based on more fully presented ordinary historical narratives. Moreover, the “individuals” variable is actually captured in part by the actor certification mechanism. Sure, Silveira was effective in part because he was tireless, disciplined, and skilled at project management. Sure, these qualities inspired his team. But he and his team were committed, in part, because of the relational processes or social mechanisms of actor certification, attribution of opportunity, and performance feedback. Further, the analysis of process context factors shows the importance of institutions that produce, promote, and share people with some of these same qualities: institutions like BNDES, Petrobrás, and other elements of the federal establishment (and some state governments, such as São Paulo’s), as indicated above.

Having said this, one must mention presidential personality as an omitted (but

includable) individual-level “variable.” Actor certification surely does not operate in the same way for presidents as for officials, so the causal texture of presidential participation in an organizational process is different, under this analysis. What should we make of this lacuna in the analysis, specifically in the case of Brazil in Action? We can take as an exogenous, if beneficent, factor, Cardoso’s disposition as a pragmatist, who accepted that a coherent management process was to be valued and supported personally by the president. Still, it must be pointed out that the president’s own conduct in the program was not an entirely exogenous variable, as plainly suggested by Antonio Kandir’s own lessons from this experience:

You have to design a management process in function of the characteristics of your system of power and the leader. You have to present the idea so that he feels comfortable with the fact that other people will perceive that the effort is his…. Fernando Henrique Cardoso is a professor. He was not a person very connected with management ….So we put together a program where he could say, look, here’s the map, this here is national integration, we’re going to have managers to do this….In other words, you have to make political adaptations looking at the style of whichever president.

40

Conclusion This paper has attempted to create usable knowledge about public

management policies on the basis of an in-depth analysis of Brazil in Action, a program of implementing presidential priority development projects that operated in 1997-99. This program did an impressive job of performing the function for which it was intended, by Brazilian standards and probably international standards. It deserves the label “administrative innovation.” The paper’s central issue is how to create usable knowledge about administrative innovations of Brazil in Action and may other functioning organizational processes.

This study exhibits the following, partial answer to that question. First,

produce an historical narrative of the experience. Second, consider the innovation as a policy argument with fresh nuances, and then model the argument in terms of its claims and warrants derived from public philosophies of governance, empirical knowledge of governmental systems, and systematic management thought in such areas as strategic management, management accounting and control, and well-performing organizations. Compare the argument’s claims, warrants, and flow of reasoning with received wisdom and assess its worth in that light. Third, consider the innovation as a functioning organizational process, relatively stable in time. Construct an analytic narrative of a cycle of operation. Choose an appropriate case outcome. Organize the data in terms of process design features and process context factors. Array process design features into guiding ideas, governance arrangements, structured events, and cognitive techniques. Array process context factors into institutions, ideas, related policy developments, and the political stream. Adopting the scientific outlook of institutional processualism, explain how these factors influence actors’ identities and situations and thereby actions and efforts. In doing so, make use of theoretical constructs such as the social mechanisms (relational processes) of certification, attribution of opportunity, and performance feedback. These fruits of these three forms of interpretation of vicarious organizational experiences – narrative history, revealed patterns of argument, and analytic narrative of the institutional processualist sort -- should be of use to managerial craftsmen as they formulate administrative design problems and proceed toward imagining, shaping, and winning allies to intelligent solutions to them.

Social Mechanism

Process Design Features

Process Context Factors

Actor certification

Guiding Ideas: presidential priorities, management system for core public sector, project management approach, project managers’ responsibilities, results orientation in government, theory of constraints Governance arrangements: structural positions of planning secretariat and managers in sectoral ministries Structured events: Appointing program directors, selecting project managers, public release of manager names, socialization occasions

Institutions: presidency, central coordinating agencies, sectoral ministries, state-owned enterprises; career systems within economic area of Federal government, personnel rules Ideas: Project management, national process of development Related policy developments: managerial reform of the state Political stream: sustained media interest in the program

Attribution of opportunity

Guiding ideas: presidential priorities, “guaranteed” funding for projects, pilot project in results-oriented public management Governance arrangements: planning secretariat as clearance point for release of budgeted funds, presidency as venue for resolving disputes

Ideas: Development agenda can be pursued, given macroeconomic stability attained Related policy developments: selection of infrastructure as a priority area, closure on qualities of the management system and on actual project priorities Political stream: president’s attentiveness to program symbols and substance, program seen as source of presidential popularity, prospect of re-election, presidential quota includes central coordinating agencies

Performance feedback

Guiding ideas: managers responsible for projects, management system’s operation is important to president, Governance arrangements: roles and responsibilities for management information system Structured events: defining projects, negotiating targets, monitoring project status, reporting formally and informally, identifying constraints, taking corrective action, disseminating information to media, sending progress reports to president on monthly basis, replacing non-performing managers, annual review of program and project performance Cognitive techniques: various quality and project management techniques

Institutions: presidency, central coordinating agencies, sectoral ministries, state-owned enterprises; career systems within economic area of Federal government Ideas: Project management, national process of development Political stream: sustained media interest in the program

Table 2. Structure of the Analytic Narrative

42

References

Abbott, Andrew (1992). “What do Cases do? Some Notes on Activity in Sociological Analysis,” in What is a Case? Exploring the Foundations of Social Inquiry, ed. Charles C. Ragin and Howard S. Becker. Cambridge: Cambridge University Press.

Aucoin, Peter (1995). The New Public Management: Canada in Comparative Perspective. Montreal, IPPR.

Bardach, Eugene (1994). “Comment: The Problem of Best Practice Research,” Journal of Policy Analysis and Management, 13, 2: 260-68.

Bardach, Eugene (1998). Getting Agencies to Work Together: The Practice and Theory of Managerial Craftsmanship. Washington, D.C.: Brookings Institution Press.

Bardach, Eugene (1999). “Book Review: The Art of the State,” International Public Management Journal, 2, 2(A): 389-392.

Bardach, Eugene (2004). “Presidential Address -- The Extrapolation Problem: How Can we Learn from the Experience of Others,” Journal of Policy Analysis and Management, 23, 2: 205-220.

Bardach, Eugene, and Robert A. Kagan (2002). Going by the Book: The Problem of Regulatory Unreasonableness, 2d ed. New Brunswick, NJ: Transaction Books.

Barzelay, Michael (1997). “Central Audit Institutions and Performance Auditing: Comparative Analysis of Organizational Strategies in the OECD,” Governance, 10, 3: 235-260.

Barzelay, Michael (1992), Breaking Through Bureaucracy: A New Vision for Managing in Government. Berkeley: University of California Press.

Barzelay, Michael (1991). “Managing Local Development: Lessons from Spain,” Policy Sciences, 24, 271-290.

Barzelay, Michael (2001a). “How to Argue about the New Public Management,” International Public Management Journal, 2, 2(A), pp. 183-216.

Barzelay, Michael (2001b), The New Public Management: Improving Research and Policy Dialogue. Berkeley: University of California Press.

Barzelay, Michael (2003). “Introduction: The Process Dynamics of Public Management Policy-Making,” International Public Management Journal, 6, 3: 251-281.

Barzelay, Michael, and Natascha Fuechtner (2003). “Explaining Public Management Policy Change: Germany in Comparative Perspective,” Journal of Comparative Policy Analysis, 5,1: 7-27.

43

Barzelay, Michael, and Fred Thompson (2003). “Efficiency Counts: Developing the Capacity to Manage Costs at Air Force Materiel Command,” Arlington, VA: IBM Center for the Business of Government.

Barzelay, Michael, and Colin Campbell (2003). Preparing for the Future: Strategic Planning in the U.S. Air Force. Washington, D.C.: Brookings Institution Press.

Bates, Robert H., et. al. (1998). Analytic Narratives. Princeton, NJ: Princeton University Press.

Behn, Robert D. (2003). “Why Measure Performance? Different Purposes Require Different Measures,” Public Administration Review, 63,5: 586-606.

Campbell, Colin, and Graham K. Wilson (1995). The End of Whitehall: Death of a Paradigm? Oxford: Blackwell.

Cyert, Richard, and James G. March (1963). A Behavioral Theory of the Firm. Englewood Cliffs, NJ: Prentice-Hall.

Elster, Jon (1996). “A Plea for Mechanisms,” in Social Mechanisms: An Analytical Approach to Social Theory, ed. Peter Hedström and Richard Swedberg. Cambridge: Cambridge University Press.

Fischer, Frank (2003). Reframing Public Policy: Discursive Politics and Deliberative Practices. Oxford: Oxford University Press.

Gaetani, Francisco (2003). “Public Management Policy Change in Brazil: 1995-1998,” International Public Management Journal, 6,3: 327-341

Gambetta, Diego (1996). “Concatenations of Mechanisms,” in Social Mechanisms: An Analytical Approach to Social Theory, ed. Peter Hedström and Richard Swedberg. Cambridge: Cambridge University Press.

Goodin, Robert E. (1992). Green Political Theory. Cambridge, Eng.: Polity Press.

Graham, Robert J, and Randall L. Englund (2004). Creating an Environment for Successful Projects, 2d ed. San Francisco, CA: Jossey-Bass.

Greve, Henrich (2003). Organizational Learning from Performance Feedback: A Behavioral Perspective on Innovation and Change. Cambridge: Cambridge University Press.

Guillén, Mauro F. (1994). Models of Management. Chicago: University of Chicago Press.

Hall, Peter A. (1992), “The Movement from Keynesianism to Monetarism: Institutional Analysis and British Economic Policy in the 1970’s,” in Structuring Politics: Historical Institutionalism in Comparative Analysis, ed. Sven Steinmo, Kathleen Thelen, and Frank Longstreth. Cambridge: Cambridge University Press.

Hall, Peter A., and Rosemary C.R.Taylor (1996). “Political Science and the Three New Institutionalisms,” Political Studies 44: 936-57.

44

Heifetz, Ronald (1993). Leadership without Easy Answers. Cambridge: Harvard University Press.

Hirschman, Albert O. (1991). The Rhetoric of Reaction: Perversity, Futility, Jeopardy. Cambridge, MA: Belknap Press of Harvard University Press.

Hood, Christopher (1998). The Art of the State: Culture, Rhetoric, and Public Management. Oxford: Oxford University Press.

Hood, Christopher, and Michael Jackson (1991). Administrative Argument. Aldershot, Gower.

Kelman, Steven (1986). Making Public Policy. New York: Basic Books.

Kingdon, John (1984). Agendas, Alternatives, and Public Policy. Boston: Little, Brown.

Levitt, Barbara and James G. March (1998). ‘Organizational Learning’, Annual Review of Sociology, 14: 319-40.

Lichbach, Mark I (1997). “Social Theory and Comparative Politics,” in Comparative Politics: Rationality, Culture, and Structure, ed. Mark Irving Lichbach and Alan S. Zuckerman. Cambridge: Cambridge University Press, pp. 239-276.

Majone, Giandomenico (1989). Evidence, Argument, and Persuasion in the Policy Process. New Haven: Yale University Press.

March, James G. (1994). A Primer on Decision making: How Decisions Happen. New York: Free Press.

March, James G. (1999). The Pursuit of Organizational Intelligence. Malden, MA: Blackwell.

March, James G., and Johan P. Olsen (1989). Rediscovering Institutions. New York: Free Press.

Mashaw, Jerry L. (1983). Bureaucratic Justice: Managing Social Security Disability Claims. New Haven: Yale University Press.

Mashaw, Jerry, and David Harfst (1987). The Struggle for Auto Safety. Cambridge, MA: Harvard University Press.

McAdam, Douglas, Sidney Tarrow, and Charles Tilly (2001), Dynamics of Contention. Cambridge: Cambridge University Press.

Miller, Peter, and Nicholas Rose (1990). “Governing Economic Life,” Economy and Society, 19, 1: 1-31.

Moore, Mark H. (1995). Creating Public Value: Strategic Management in Government. Harvard: Harvard University Press.

45

Moynihan, Donald P. (2003). Public Management Policy Change in the United States during the Clinton Era,” International Public Management Journal, 6,3: 371-394.

Neustadt, Richard E. (1960). Presidential Power. New York, Wiley.

Norreklit, Hanne (2003), “The Balanced Scorecard: What is the Score? A Rhetorical Analysis of the Balanced Scorecard,” Accounting, Organizations, and Society, 28, 6, (August): 591-619.

Pettigrew, Andrew (1997). “What is a Processual Analysis?” Scandinavian Journal of Management, 13, 4, pp. 337-348.

Power, Michael (1996). “Making Things Auditable,” Accounting, Organizations, and Society, 21, 2/3: 289-315.

Presidência da República (1996). “Brasil Em Ação: Investimentos para o Desenvolvimento,” Brasília: Presidência da República.

Pressman, Jeffrey, and Aaron Wildavsky (1974). Implementation. Berkeley: University of California Press.

Ragin, Charles (1987). The Comparative Method. Berkeley: University of California Press. Resende-Santos, João (1997). “Fernando Henrique Cardoso: Social and Institutional Rebuilding in Brazil,” in Technopols: Freeing Politics and Markets in Latin America in the 1990’s, ed. Jorge I. Dominguez. University Park, PA: Penn State University Press.

Roett, Riordan (1999) Brazil: Politics in a Patrimonial Society, 5th ed. Westport, CT: Praeger.

Rose, Richard (1993). Lesson-Drawing in Public Policy. Chatham, NJ: Chatham House.

Sahlin-Andersson, Kerstin (2002). “National, International, and Transnational Constructions of New Public Management,” in New Public Management: The Transformation of Ideas and Practice, ed. Tom Christensen and Per Lægreid. Aldershot, Eng.: Ashgate.

Schick, Allen (1996). “The Spirit of Reform: Managing the New Zealand Sector in a Time of Change.” Wellington, NZ: State Services Commission. (http://ssc.govt.nz)

Schön, Donald (1983). The Reflective Practitioner. New York: Basic Books.

Simon, Herbert A. (1946). “The Proverbs of Administration,” Public Administration Review, 6: 53-67

Simon, Herbert A. (1976). Administrative Behavior, 3d ed. New York: Free Press.

Simon, Herbert, Donald Smithberg, and Victor Thompson (1950). Public Administration. New York: Knopf,

46

Simons, Herbert (2001). Persuasion in Society. Thousand Oaks, CA: Sage.

Steinbruner, John D. (1974). The Cybernetic Theory of Decision: New Dimensions of Political Analysis. Princeton: Princeton University Press.

Tendler, Judith (1997). Good Government in the Tropics. Baltimore: Johns Hopkins University Press.

Thompson, Fred, and Larry R. Jones (2000). “Responsibility Budgeting and Accounting,” International Public Management Journal, 3, 2: 205-227.

Toulmin, Stephen (1958). The Uses of Argument. Cambridge: Cambridge University Press.

Walker, Jack L. 1977. “Setting the Agenda in the U.S. Senate,” British Journal of Political Science, 7 :423-445.

Walton, Douglas N. (1992). Plausible Argument in Everyday Conversation. Albany: SUNY Press.

Weick, Karl E. (2001). Making Sense of the Organization. Oxford: Blackwell.

Whittington, Richard (1993). What is Strategy – and Does it Matter? London: Routledge.