Cover image for Shifting States in Global Markets: Subnational Industrial Policy in Contemporary Brazil and Spain By Alfred P. Montero

Shifting States in Global Markets

Subnational Industrial Policy in Contemporary Brazil and Spain

Alfred P. Montero


$44.95 | Paperback Edition
ISBN: 978-0-271-02529-2

272 pages
6" × 9"
5 b&w illustrations/2 maps

Shifting States in Global Markets

Subnational Industrial Policy in Contemporary Brazil and Spain

Alfred P. Montero

“A theoretically ambitious, empirically rich study of regional economic policy in Brazil and Spain. Using case studies based on extensive field research, Alfred Montero demonstrates how political factors such as elite competition can affect economic development. An important contribution to the emerging literature on subnational political economy.”


  • Description
  • Reviews
  • Bio
  • Table of Contents
  • Sample Chapters
  • Subjects
Shifting States in Global Markets contributes to the debates over the political economy of globalization by focusing attention on the increasingly important role of subnational governments in implementing economic policies. Challenging the view that the effects of decentralization are positive or negative uniformly and can be explained by reference to the influence of national political institutions, Alfred Montero uses his comparisons of industrial policy in Brazil and Spain, and between different regions in these countries, to argue that we need to pay attention to political conditions at the subnational level to account for the variation in economic success between regions.

Two crucial conditions are emphasized in Montero’s analysis: how much competition there is among political elites within any region, and how much competition there is between regions for scarce fiscal resources. Lower competition among elites leads to subnational governments delegating more autonomy to public agencies to develop ties with private businesses favoring allocative efficiency and innovation; higher competition between regions provides incentives for political leaders to support involvement in economic development efforts by a greater variety of public agencies, whose cooperation and mutual trust over time create the conditions for long-term success in these efforts. This analysis gives us a much more nuanced understanding of how countries are experiencing the challenges of globalization today.

“A theoretically ambitious, empirically rich study of regional economic policy in Brazil and Spain. Using case studies based on extensive field research, Alfred Montero demonstrates how political factors such as elite competition can affect economic development. An important contribution to the emerging literature on subnational political economy.”

Alfred P. Montero is Assistant Professor of Political Science at Carleton College.


List of Tables and Figures


List of Abbreviations

1. Introduction: The Political Origins of Synergy

2. Dual Transitions in Spain and Brazil

3. Remaking Industrial Policy in Minas Gerais

4. Designing Reindustrialization in the Principado de Asturias

5. Populist Government and Industrial Policymaking in Rio de Janeiro and Andalusia

6. Conclusions: Shifting States in Comparative Perspective

Appendix A: Spanish Interviews

Appendix B: Brazilian Interviews



Introduction: The Political Origins of Synergy

One unexpected parallel to the widely held view that globalization has challenged the utility of national industrial policies is the finding that subnational governments have become more active in spurring industrial investment, higher productivity, and enhancing access to technological innovation. In the state of Minas Gerais in Brazil, an array of public agencies—utility companies, business information institutes, and industrial district companies—have facilitated the rapid expansion of autoparts and assembly firms, chemical industries, and small and medium-sized enterprises in various sectors. In the Spanish region of Asturias, a similar set of public agencies with union and business association partners has promoted productivity-enhancing adjustments in small firms that were previously dependent on the now-eroding public steel and mining sectors. The experiences of Minas and Asturias dovetail a number of other cases around the world in different areas of economic policy (industrial, labor market, agriculture, and infrastructure) that have been reported by political economists.1 Some scholars and policy observers use these "success stories" to contest the popular notion that globalization of trade, production, and capital markets has reduced the importance of states and location of investment as factors in determining development patterns (e.g., Evans 1996, 1997b; Swyngedouw 1992). Moreover, these case studies correct the analytical bias in favor of national institutions and policy-making apparent in much political economic scholarship (Snyder 2001).

These experiences suggest that subnational comparisons can reveal important heterogeneous characteristics in countries many have previously regarded as having certain strictly homogenous, national-level qualities. When these cases are compared to other subnational regions in Brazil and Spain, however, or when they are compared to each other, the successes of industrial policy in Minas and Asturias require additional explanation. For example, the experience of Minas Gerais stands out in a country that has a long history of political clientelism associated with the mismanagement of development policy resources. Brazil’s governors are particularly liable for administering what many scholars of the country see as the most powerful clientelistic networks in Brazil (Abrúcio 1998; Samuels 1998, 2000; Schwartzman 1973). Due to the weakness of national party labels, political institutions are chronically unable to limit the governors’ influence on federal legislators and public firms (Mainwaring 1997, 1999; Ames 1995, 2001; Garman, Haggard, and Willis 2001). In states such as Rio de Janeiro, populist governors have historically developed political dynasties that invoke comparisons to the caciques of Mexican politics or the notorious local party bosses of Tammany Hall and Chicago. If political clientelism is as common as studies of policy-making in Brazil commonly attest (e.g., Weyland 1996; Hagopian 1996), then subnational industrial policy remains vulnerable to these short-term political interests. Subnational party bosses have strong incentives to manipulate industrial reform to enhance their own short-term interests at the cost of the longterm economic performance such policy seeks to enhance.

In other words, the management of subnational industrial policy in Brazil suffers from a kind of delegative dilemma. Self-serving populist governors would be reluctant to waste politically valuable economic policy resources by devolving control over industrial policy to public development agencies. Populists could more easily manipulate economic policy resources by continuing to centralize control over the policy-making apparatus. Without delegation, public agencies and private firms could not coordinate resources in ways that would enhance long-term economic performance. Where Rio de Janeiro reflects these seemingly endemic qualities in Brazilian politics, Minas Gerais does not. In Minas, the governors created a decentralized network of technocratic industrial policy agencies and they delegated politically useful authorities and resources to them from the 1960s through the 1990s. The mineiro agencies proved successful in promoting industrial investment and in enhancing productivity under both a state-led, developmentalist framework from the 1960s to the mid- 1980s and a market-oriented model during the late 1980s and throughout the 1990s. The case of Minas suggests that the experience of subnational economic policy-making in Brazil does not follow a uniform pattern. It takes more of a patchwork, composite shape (Montero 2000, 2001a). We might expect a more uniform pattern in Spain. In contrast to Brazil’s fragmented political party and federal system, Spain’s new democracy during the 1980s and 1990s developed highly disciplined national parties and a well-organized federal state structure (Bermeo 1994; Agranoff 1996). Yet, like Brazil, the pattern of subnational industrial policy- making in Spain is also composite. Different regional experiences with industrial policy are the rule. For example, while Asturias’s regional development agencies avoided the contentious politics of labor union mobilization against national downsizing in the public mining and steel sectors, Andalusia’s regional agencies became devices for collecting labor support for the governing Socialist Party. Although both regions were governed during most of the 1980s and 1990s by the same national party, Andalusia’s industrial policy was plagued by populist intervention and clientelistic exchange; Asturias’s industrial policy was not. Disciplined national parties and a highly organized federal structure in Spain led to the same patchwork distribution of effective and politicized subnational economic programs evident in Brazil. The composite pattern of subnational industrial policy-making in these two very different countries highlights the need for scholars to peer below national-level data and the national unit of analysis to explain withincountry variation. This mandate is particularly challenging to the dominant analytical tendencies in the study of economic policy in federal or decentralized states. Scholars in this research program have divided into two competing camps: those who argue that decentralization enhances economic policy (the optimists) and those who posit that it endangers it (the pessimists). The optimistic view typically sees decentralization as a means for improving the economic performance of markets by limiting central intervention, encouraging policy innovation among competing subnational units, and enhancing allocative efficiency through the local provision of goods.

The pessimists argue that political factors impede the realization of these benefits. Pervasive subnational clientelism, especially in developing countries, leads to the misappropriation of resources.

The benefits of decentralization cannot be realized at the subnational level if patronage-maximizing local incumbents centralize control over policy resources and fail to delegate authority to technocratic agencies. Both competing views suffer from the tendency to treat subnational governments and the process of decentralization as systems with endemic positive or negative attributes. However, these explanations are challenged by the existence of highly composite configurations of subnational policy performance within countries. This finding highlights the role of variables endogenous to subnational politics, including the interests of politicians and the structure of bureaucracies. Rio’s experience with industrial policy differs from Minas’s because of its particular legacy of populist government, conflict over patrimonial politics, and highly centralized economic bureaucracy. Similarly, Andalusia’s industrial policy has from its initiation been governed by populist ideology, labor union politics, and bureaucratic centralization. These comparative experiences challenge both the optimistic and pessimistic views of economic policy in decentralized states. The cases of Rio and Andalusia confound optimistic expectations that decentralization will produce subnational innovation and allocative efficiency. These goals are laudable, but they are, as the pessimists argue, politically vulnerable to clientelistic manipulation. Meanwhile, the pessimistic view of decentralization cannot explain the political sources of subnational policy innovation and allocative efficiency in Minas Gerais and Asturias. These experiences suggest something other than the presence of endemically clientelistic subnational politicians and dysfunctional bureaucracies. Systemic/endemic approaches must be replaced by comparative cross-national and cross-subnational studies of decentralized public sectors to reveal the politically contingent conditions affecting policy performance. This is the main enterprise of this book. The remainder of this chapter outlines the main argument, the theoretical framework, and the format of the empirical study. The Argument The Dependent Variable The empirical finding of composite patterns of subnational industrial policy-making in Brazil and Spain begs a broader question that is central to the focus of this book: Why do some subnational governments produce productivity- enhancing cooperation between public agencies and firms while others do not? This dependent variable is captured by the term "synergy." As Peter Evans (1997b: 3) explains, "the idea of `synergy’ implies [that] . . . the actions of public agencies facilitate [the] forging of norms of trust and networks of civic engagement among ordinary citizens and [the use of] these norms and networks for developmental ends." The presence of synergy reduces the costs of transacting business by facilitating cooperation between state agencies and capitalists. Synergy minimizes the costs of bargaining to smooth out disagreements since all parties can rely on a reservoir of social trust. These conditions limit the opportunity costs of bargainers’ time, the price of surveillance, and the cost of arbitration when cooperation threatens to collapse. Trust will also reduce future costs by lessening uncertainty. Consequently, in the presence of high levels of synergy, firms deem government officials more credible and they demonstrate more confidence in public sector decision-making. Another way to think about synergy is to regard it as highly reciprocal transfers of information between agencies and firms; close collaboration on the setting of goals and the implementation of policy. Elinor Ostrom (1997) illustrates this relationship by first making a distinction between a "regular producer" and a "client" of a public good such as education. Typically, she argues, government agencies are the "regular producers" of education while ordinary citizens are the "clients." This relationship describes an active provider and a passive user. Synergy exists where these distinctions are blurred; where citizens participate actively in the conception and provision of the public good and government agencies are not simply producers but clients of "citizenproduced" information and other resources.

In this study the public good in question is industrial policy, namely, public policies designed to enhance the productivity and growth of firms in the industrial sector. In typical studies of industrial policy, the firm or the industrial sector is conceived of as the "client" of the financing, import protection, research and development programs, and any of the other common instruments of industrial policy that the state provides. Industrial policy, however, that is conceived and implemented through synergistic relations between public agencies and firms is the product of reciprocal exchange of information and resources and close collaboration. Agencies and firms coordinate common resources (e.g., information, financing, etc.) to produce a desired economic outcome (e.g., productivity improvements, positive externalities, growth, etc.). The main purpose of this study is not to explain the causes of positive economic outcomes, but to examine the political causes of the cost-effective forms of public-private cooperation described by Ostrom and Evans as "synergy." My concern is specifically with two questions: (a) Under what political conditions will such associations emerge?; and (b) Under what conditions will these relations be maintained? The study treats synergy as both a dichotomous variable (it exists or does not) and as a variable with more continuous qualities. For example, synergy can be said to be high in cases in which public agencies and firms continually cooperate in the setting of common goals for policy; share information and resources such as finance, technology, and personnel; and work in tandem to produce positive externalities. In these cases, the distinction between the roles of producer of industrial policy and client is blurred. Conversely, sporadic or absent forms of cooperation in which public agencies have few or no contacts with firms and impose policy goals unilaterally are cases of low synergy. These are the experiences that will generate more costly forms of policy implementation. Producing synergy is complicated by a public-goods logic that is magnified in the area of industrial policy. The actors and policies that are responsible for generating synergy operate within a political context. In that context they are subject to the classic problems of public-private interactions. Each intervention by the public sector in private markets creates incentives for business to influence economic policy and therefore to transform the state into an arena of distributive conflict (Rueschemeyer and Evans 1985: 69; Knight 1992). Diverse interests vie for the retention of protection, subsidies, or some other mechanism that defends an allocative outcome favorable to them (Krueger 1974; Buchanan, Tollison, and Tullock 1980; Olson 1982).

This extends most notably to the politician herself/himself. In the short term, a politically optimal strategy (that is, one that assures the political survival of incumbents) may require politicians to manipulate state resources and make patronage appointments to the bureaucracy in return for support. This calculation creates a tradeoff between employing economic policy resources for political purposes in the short term versus risking political survival in the hope of promoting collective economic efficiencies over the long term. Garrett and Lange (1996: 50) describe this tradeoff simply: "Politicians cannot afford to ask what is good for society as a whole in the long run, lest they lose power in the interim." Following Barbara Geddes’s (1994) game-theoretic work on political interests and economic reform, I argue that the conflict between shortrun and long-term political interests constitutes a "delegative dilemma." Incumbents will maximize their chances of political survival in the short term by centralizing their control over policy resources and distributing these goods to cultivate clientelistic support (Bates and Krueger 1993). Delegation of authority over economic policy and resources to agencies would increase the political opportunity costs to incumbents of manipulating these resources. Yet as Geddes and other scholars in the "political survival" literature attest (see Ames 1987; Bates 1989), the incentive structure of short-term versus long-term tradeoffs is not constant. Changing socioeconomic constraints and institutions mediate "between [these] raw preferences and government behavior" (Garrett and Lange 1996: 51). These factors may take many forms. They may include formal institutions such as electoral rules that favor the emergence of large programmatic parties and strategic situations in which all potential political rivals are weak.

These contextual factors alter politicians’ expectations of the costs and benefits accrued from delegation. The chief analytical advantage of highlighting under what conditions political expectations take the shape they do is that this approach links in a generalizable manner the interests of political leaders, the institutions and socioeconomic factors that affect the content of their interests, and policy outcomes (Geddes 1994: 8).

A more specific and universally applicable formulation of this logic focuses on political security. Under conditions in which the tenure of governing elites is not in jeopardy and time horizons can be extended, political leaders are more likely to delegate politically useful resources to technocratic agencies. Where levels of polarization among elite groups are high, incumbents feel vulnerable and are likely to discount future gains from economic policy success.

Elite conflict occurs both in terms of partisan competition over votes and social conflicts involving labor and business. These two dimensions of elite competition can intersect, generating cases in which majority parties/coalitions that might otherwise make incumbents more risk-accepting eschew delegation when faced with significant union and business opposition. Where both partisan and social polarization are muted, incumbents lead what I call delegative governments. Where partisan or social polarization are high or increasing, incumbents face incentives to cultivate personal support in the short term by rewarding their backers through clientelism. Incumbents in this case will not forego potential sources of patronage by delegating control over economic policy to technocrats. They will reduce their political opportunity costs by centralizing their authority over these politically useful resources. These conditions are typical of populist governments. The delegative dilemma is similar to one Barbara Geddes terms the "politician’s dilemma." While industrial policy does not ask politicians to forsake patronage altogether, the kind of industrial policy that generates positive externalities in the market requires that clientelistic exchange be minimized. Geddes argues that this is more likely to occur under conditions in which contending political parties are roughly equal in power. Parity reduces the relative political costs of delegation by requiring all political rivals to pay these costs equally. However, the argument I posit here suggests that parity may not be necessary and that politicians in a more secure position will have incentives to delegate.

Political security adjusts incumbents’ perceptions of the costs of delegation in the short term. Delegative governments, as the empirical cases in this book demonstrate, do not discount the future as heavily as populist governments do. On the contrary, they estimate that the opportunity costs of delegation are low. Therefore, while delegation increases the opportunity costs of using politically useful resources later, incumbents in delegative governments believe that they have made investments and that they will be able to claim credit for improved economic performance over the long haul.

The primary utility of both the partisan parity argument and that of political dominance is that they highlight the importance of low elite polarization. Highly competitive or conflictual political games are more likely to be associated with incentives to pursue clientelistic support than those in which the political opportunity costs of delegation are low. Politically vulnerable subnational incumbents will minimize the opportunity costs of mustering personal support by centralizing control over clientelistic resources, while less vulnerable subnational elites will entertain secondary incentives to decentralize. The primary incentives to delegate, therefore, are determined by the relative level of intraregional conflict. Secondary incentives to delegate operate in the absence of high levels of intraregional political conflict and they are based on the nontrivial benefits of delegation. The act of delegation provides its own political returns, as regional leaders can tell their constituents that they are "doing something" about persisting economic problems and can thus justify the need for additional resources and authority from central government managers. Alternatively, if things go wrong, incumbents can blame the agencies to which they delegated their authority, covering their own responsibility for policy failure under a cloak of ambiguity (Alesina and Cukierman 1990: 846). Doing nothing may become politically costly in the face of eroding economic performance and competition with rival subnational governments over the same pool of fiscal resources. These executives will be motivated to delegate by growing concern for the larger economic welfare of their region, in addition to the expected political benefits accrued from ostensibly addressing these problems. The Intervening Variable: Horizontal Embeddedness Delegation empowers "political technocrats" who have broader reservoirs of technical experience than politicians. These actors are also more politically savvy than careerist technocrats.

Political technocrats are "swing agents" in the bureaucracy. They sit atop key bureaucratic agencies as figures pivotal to the delegation of political authority and the coordination of technical resources (Schneider 1991). As in all principal-agent models of this type, the agent (political technocrat) recommends a strategy that he or she views as the best response to economic problems. Governing politicians (the principals) that are not vulnerable to their political opponents but are less informed about technical responses will support their subordinates’ proposals to avoid the risks of worsening the economic situation through their own inaction or through poorly planned action (Aghion and Tirole 1997). By doing so, in- cumbents will see the delegation of authority and resources to political technocrats as being in their best political interests. Nevertheless, delegation is a necessary but insufficient condition for producing synergy. Two dangers may continue to threaten the emergence of cooperative agency-firm relations over time. First, incumbents may overcome the "delegative dilemma" at one point in time (t0) but face a new set of political opportunities later. Incumbents may become vulnerable to rivals at t1, compelling them to recentralize resources and authorities in ways that weaken the institutions producing synergy. That makes the maintenance of delegation contingent upon the politician’s ex post evaluation at t1 of political costs and benefits. Previous commitments and institutional constraints will certainly reduce the ability of incumbents to renege, but these are not infallible (Samuels and Mercuro 1984). In short, the existence of contingent delegation keeps bureaucracies under the threat of pervasive rent-seeking. Second, delegation gives political technocrats opportunities to advance their own interests. Since these "technopols" "are not marionettes of other politicians" (Domínguez 1997: 5), they may develop an interest in hiding information about their effort levels and performance from supervisors, politicians, and their constituents, thus contributing to "bureaucratic dysfunction."

As in all principal-agent relations, every incentive system allows individuals to pursue their own narrow interests or "shirk." Information asymmetries produce higher "costs of hidden action and hidden information" that facilitate shirking (Miller 1992; Solnick 1996).

Moreover, even when politicians favor delegation, private interests will continue to prod for special protection and subsidies ("rents"), producing incentives for bureaucrats to act in their own selfinterest (Schneider 1993; Bates and Krueger 1993: 464). I argue that the form interagency relations take is a crucial intervening variable in linking political interests with the creation of productive patterns of cooperation between public agencies and firms. The literature on agency-firm cooperation suggests that these relations develop from the creation of strong "vertical ties": associations that produce reciprocity, mutual understanding, and goal-setting. I argue, however, that this is only half the story. Amore important dimension lies above the domain of these vertical associations, at the level of public agencies themselves and, in some cases, decision-making partners such as unions and business associations. The most important aspects of this dimension are the patterns of contacts among these actors and the operation of these ties in conceiving and implementing economic policy.

I call these contacts among policy- making agencies horizontal embeddedness.

Horizontal embeddedness minimizes the costs of political intervention and bureaucratic dysfunction by enfranchising a cross-cutting policy constituency broader than is normally the case in strictly vertical forms of cooperation between bureaucratic agencies and firms. Clientelism can develop more easily within vertical networks, while horizontal embeddedness produces the cross-agency monitoring that raises the costs of bureaucratic shirking. Horizontal embeddedness also raises the costs of intervention by politicians by creating alternative sources of political support through the expansion of the constituency of public agents and private clients involved in industrial policy-making. For example, horizontal ties increase the number of alarm bells that may sound if policy goals are not being met, preventing incumbents from hiding evidence of patrimonial exchange (McCubbins and Schwartz 1984). In sum, subnational governments will implement and maintain industrial policy, generating allocative efficiencies and minimizing clientelism, only when delegative government and horizontal embeddedness are present. Both are necessary and cumulatively sufficient. The Incidence of Delegative Government with Horizontal Embeddedness: The Interregional Determinant The combination of delegative government and horizontal embeddedness emerges in cases where incumbents delegate authority and re- sources to political technocrats but are themselves exposed to incentives to divide the governance of economic policy among an array of different agencies rather than to invest control in just one. Following the work on competitive federalism in American politics, scholars attribute the shape and content of subnational policy to competition among subnational governments for investment. They argue that competition provides incentives for innovation in public policy to attract consumer-voters and businesses (Fosler 1988; Dye 1990).

These conflicts also create political inducements to attract federal funding lest other states acquire relatively larger shares. By creating an array of public agencies to design and implement industrial policy, incumbents can bolster their arguments for more resources from the central government with references to the immensity of the economic crisis and the related need to fund an expansive industrial policy structure. Given that industrial policy includes multiple tasks and policy skills, from utility management to labor retraining, logistical exigencies produce additional incentives for subnational governments to create numerous and connected public agencies to satisfy firm-clients and consumer-voters. Where political incentives to delegate are present, such horizontally embedded structures will be more common than in cases in which populist elites are in command of subnational government. Under populist management, the incentives to centralize control over policy resources will override the inducements created by competitive federalism to decentralize. Elite polarization will continue to tip political institutions in directions that favor clientelistic exchange. Hence, horizontal embeddedness is more likely to emerge where intraregional elite competition is low but interregional elite competition is high. Figure 1.1 illustrates the interplay of these two independent variables and the occurrence of the intervening variable, horizontal embeddedness. Spain and Brazil are decentralized states with relatively high interregional competition. These conditions generate incentives for the development of horizontal embeddedness. In these cases, intraregional competition determines which subnational cases develop both delegative government and horizontal embeddedness. In Chapter 6 I explore change in the same decentralized state over time to illustrate the two other possible combinations. The Mexican states prior to the 1990s faced little interregional competition but also little intraregional elite conflict. Most states were governed by the Institutional Revolutionary Party (PRI). While the opportunity to delegate existed, low levels of interregional competition due to tight federal controls over economic policymaking and spending by the states produced few incentives or resources for the creation of subnational, horizontally embedded economic agencies.

Where delegation did occur, subnational bureaucracies were underdeveloped and tended to follow or be governed by national government mandates. They were, as the figure indicates, "dependent governments." During the 1990s, a more competitive political party system and decentralization reforms created incentives for the governors to use slowly expanding economic policy resources to cultivate their subnational constituencies. Most fiscal and economic policy authorities, however, remained centralized; governed by federal mandates that weakened incentives for subnational executives to create extensive subnational development bureaucracies. Continued control over interregional competition in a polity experiencing greater intraregional elite conflict produced "populist-dependent governments," which, like the "populist" and "dependent" types, developed neither delegative government nor horizontal embeddedness. Once created, horizontal embeddedness protects the decentralized connections between public agencies and private firms that are necessary for producing developmentally successful forms of cooperation. Even if intraregional elite polarization becomes more intense over time, making political leaders more vulnerable, their ability to intervene in these encompassing networks will be diminished. Horizontally embedded bureaucracies are more difficult to disassemble since they can move their resources around with relative flexibility. Public agencies in these networks can also cultivate broader constituencies for industrial policy, creating countervailing political support that raises the costs of intervention for political executives. Horizontal embeddedness also creates cross-cutting networks of surveillance within the bureaucracy that reduce the risk that political technocrats will abuse their authority. Therefore, horizontal embeddedness allows public-private cooperation to persist even after the end of the mandates of the politicians who first made these productive associations possible. Horizontally embedded policy networks are also flexibly reactive. Since information flows are symmetric and frequent, policy-making agencies can solve problems collectively in the most efficient manner possible. Continuous monitoring facilitates continuous experimentation and learning in a manner akin to that identified by scholars of flexible production systems (e.g., Sabel 1995). Consequently, horizontally embedded networks are less likely to become ossified. They are prone to develop responses to economic and political change as these transformations occur. As the cases will demonstrate, delegative governments may produce horizontal embeddedness that varies according to the structure of interagency relations, based upon (a) the frequency of interagency contact and (b) the degree of centralization of the interagency network. The most developed form of horizontal embeddedness is characterized by continuous contacts across agencies that are decentralized but coordinated under a common development plan that specifies targeted sectors or regions and agency responsibilities. The plan establishes a framework for setting goals and timetables for joint action (for high-level functions) and, in practice, requires the sharing of personnel and information as a routine matter. No single agency maintains supremacy over the others, although an interagency council on which all are represented might have priority-setting powers. I label this ideal type "plan-oriented" horizontal embeddedness. The second form of horizontal embeddedness created by delegative government is the "project-oriented" type. The agencies still coordinate high-level functions and oversee each other’s operations, but contact is less frequent than in the plan-oriented system. Strategic planning is coordinated among the agencies, but usually without an interagency council or a development plan supported by a ministry. Planning is done as project opportunities emerge. As in the plan-oriented variety, no agency is supreme or able to see a firm project through to completion without the involvement of the others. As the cases will demonstrate, project-oriented forms of horizontal embeddedness are more likely to exist when politicians have de-emphasized the creation of industrial development plans and have left these functions to the agencies under an extant set of priorities. Populist and populist-dependent governments do not produce horizontal embeddedness, but they do foster the development of two types of interagency relations that are evident in the cases. These are summarized in Figure 1.2 and compared with the types described above. The first is a "hierarchical-functional" type in which interagency relations are highly centralized. Adominant planning agency will devolve operations to more specialized units, but each of these subagencies will have only low-level functional duties (e.g., distributing finance, building infrastructure, etc.) without the authority to participate in strategic planning. Strategic planning will most likely be defined by politicians with command over the dominant agency. Contacts are "continuous" but asymmetrical between the dominant agency and the subagencies. Hence few opportunities exist to generate feedback or produce checks on potential abuses of power at the top. The second type of interagency network evident in populist or populist- dependent cases is the "nonfunctional" type. Like the hierarchicalfunctional system, this type has the concentrated interagency structure favored by populist governments. But in addition to the problems of reduced checks on executive decisions, low levels of interagency contact generate high costs of hidden information. This type maximizes bureaucratic dysfunction and clientelistic misappropriation of resources.

Both the plan-oriented and the project-oriented types of horizontal embeddedness produce coordinated and accountable relations among the agencies, reducing the possibilities for politicization and abuse of resources. These systems are most likely to maintain themselves in the face of subsequent elite polarization and therefore are more likely to sustain cooperative, efficiency-enhancing relations with firms. Contacts in both of these systems will be maintained and they will generate common goals between agencies and firms and collaborative use of resources. The hierarchical- functional and nonfunctional systems, however, are more open to being politicized by incumbents. These are the interagency relations most likely to be favored by populists who seek to minimize the costs of mustering clientelistic resources in the face of threats to their political survival. For populists, the concentration of control over development agencies is the most convenient way of maximizing their control over patronage. Alternative Approaches to Explaining the Causes of Synergy The problem of how to create and sustain cooperation between public agencies and firms has been the focus of a great deal of scholarship in the social sciences. In recent years, much of this work has highlighted patterns of cooperation based on mutual understanding and trust, rather than on market-based efficiency. Whether they use evocative terms such as Mark Granovetter’s (1985) "embeddedness" or Robert Putnam’s (1993) "civicness" and "social capital," these scholars see close trust-based associations between states and firms as developmentally effective. Judith Tendler (1997) illustrates the importance of these close ties in her study of state and municipal programs in preventive health, public procurement, and agricultural productivity in Ceará, Brazil. In addition to the role of favorable publicity and the meritocractic distribution of rewards within the state bureaucracy, Tendler emphasizes the way in which public sector officers and private clients maintained a capacity to listen to each other and adjust goals through the exchange of information. Such demonstrated reciprocity, Tendler claims, explained the effectiveness of Ceará’s innovative economic and social policies. In a similar way, Elinor Ostrom has highlighted the role of close ties between public servants and clients in the creation of systems of "coproduction." In these cases, clients hammer out the details of projects in close consultation with officials, thereby reducing the costs of arbitrating prices and citizens’ complaints after public monies have been committed.

The success of consultation emerges from the proximity of the actors and the frequency of their meetings, conditions that "deepen" their mutual understanding. The emphasis in these cases is on making the representatives of public agencies part of the community in which they work. Close ties are strong ties, and strong ties create confidence. These approaches, however, overplay the importance of "close ties" while underplaying the role of bureaucratic structure. Some do specify the relevant organizational attributes of the public sector, but these tend to highlight factors that deal inadequately with the delegative dilemma and bureaucratic dysfunction. Close vertical ties may enhance economic policy performance, but may also reinforce rent-seeking by recreating dozens of concentrated systems of discretionary power.

Dense associative ties, like those advo- cated by Putnam and Richard Locke (1995), produce their own dangers. In their multicountry test of Putnam’s thesis, Knack and Keefer (1997) found that associational activity built reservoirs of trust, but also facilitated rent-seeking by expanding the organizational base of societal interests. Of course, associations that are broad-based and encompassing may produce cross-cutting cleavages that prevent "Olsonite distributional coalitions" (Schneider 1998: 111; Garrett 1998), but these multisectoral organizations, which are best known in the "democratic corporatist" cases of Western Europe, are either absent or barely coherent in most developing countries. In the context of developing economies with "porous" state structures, encompassing associations are more likely to abuse their market power and exert pressure for rents on the economic bureaucracy.

Associational approaches also suffer from the tendency to sublimate the role of the state (both central and subnational) to one of many relevant sociopolitical actors, or to ignore it altogether by focusing analysis on relations among firms.

Yet in poor regions in developing countries where labor and business groups are often weak, the state is invariably the key actor in providing crucial resources for economic adjustment. The sociopolitical associations that do emerge are likely to be initially created by public officials and political leaders rather than by a preexisting coherent network of well-organized political, economic, and societal actors (Schmitter 1997: 248–49; Fox 1996).

Moreover, how these public agencies organize is an important determinant of agency-firm cooperation. Studies that conflate these variables gloss over this causal relationship. For example, Evans’s (1995) use of the term "embeddedness" includes the linkages between agencies and firms but it fails to disaggregate the interagency dimension. Other works place emphasis on the professional and organizational attributes of the public sector. According to these arguments, professionally staffed, corporately coherent, and "autonomous" bureaucratic agencies guarantee that economic policy will be designed and implemented in an economically rational manner.

Autonomy alone, however, is insufficient for providing state agencies with the information and political resources needed to create synergy. Moreover, bureaucrats may sacrifice economic efficiency in defense of their career interests, particularly when they are induced by politicians and societal actors who have been known to breach the wall of bureaucratic autonomy. Evans (1992, 1995) argues that a mixture of autonomy and embeddedness can produce synergy, but he is unclear about how the proper balance of the two emerges in distinct political settings. The creation of one can sap the other, especially when self-interested actors are involved in the game of institutionbuilding. Available analytical approaches to synergy either "oversocialize" by discounting the role of state interests (politicians and bureaucrats) or they "undersocialize" by ignoring the dangers and insufficiencies of autonomous state intervention.

The prescriptions of associational, social capital, and autonomy arguments remain vulnerable to the problems of the "delegative dilemma" that can cause synergy to fail or not develop in the first place. Even if synergy emerges, these continued problems will threaten its maintenance. How then can the political interests of politicians and bureaucrats be reconciled with the collective action required by synergy? The answer lies in putting off the chimerical search for preexisting networks of civic engagement and taking the political interests and strategies of state managers seriously. The Study Explaining the political causes of synergy has led me to compare subnational regions in Brazil and Spain, two countries literally an ocean apart. My initial expectations for the availability of synergy in either country could not have been more different. In terms of its level of national development, Spain is now, without a doubt, a member of the advanced capitalist world. Although Spanish industrialization followed a state-led path similar to that of Brazil during the postwar period, Spain’s leaders were able to reorient the country’s development model along more marketoriented lines earlier than in Brazil, and with more success. Spain’s fiscal crisis during the 1980s was severe, but not at all on a par with the external debt crisis that decimated Brazil’s fiscal accounts and ended the country’s experiments with "deepening" import-substitution industrialization. By 1986, Spain, along with Portugal, became a member of the European Community (EC) (now, the European Union or EU). EC structural funding and a recovering fiscal capacity allowed the Socialist government to launch a costly and politically risky restructuring of the country’s major industrial sectors. By contrast, Brazil’s industrial adjustment was ad hoc, orchestrated by a gradual opening of the domestic market to foreign competition and complicated by repeated bouts with mega-inflation. Given these differences, it would be no mystery why Spain by the end of the 1990s was able to consolidate its position in global markets as a developed market economy while Brazil continued to struggle with the exigencies of economic adjustment at great cost to domestic industry. Several political factors favored Spain’s ability to adjust. Following the death of longtime dictator Francisco Franco in November 1975 the transition to democracy produced a stable parliamentary system with coherent parties. The Socialist Party (Partido Socialista Obrero Español, PSOE), led by Felipe González, was the most organizationally disciplined and successful. Under González, the PSOE won an absolute majority in the Spanish Cortes (parliament) and held it for two terms (1982–89). Although in 1993 the PSOE would be forced to forge a parliamentary alliance with the Catalonian nationalist party, Convergència i Unió (CiU), to retain a governing majority, the Socialists would remain in power until 1996. Under these conditions, the PSOE government was effective in mediating the numerous economic and political conflicts in the Spanish regions that might threaten the stability of democracy.

Such coherent and stable party rule was completely missing in Brazil, where electoral laws and clientelism continued to erode party identity.

Ad hoc alliances for reform routinely fell apart as parties were unable to maintain a loyal following either among the voters or the "political class" (clase política).

Once again, decentralization played a destabilizing role as personalist party leadership tied to subnational political machines further fragmented parties in the national legislature (Mainwaring 1997: 83–84), compelling Brazilian presidents to exert their considerable authority through decree powers (Power 1991, 1998; Figueiredo and Limongi 1997). Yet even here, policy resulting from these "hyperpresidentialist" efforts was often diluted by the need to distribute patronage to subnational elites in order to consolidate legislative support at the federal level (Smith and Messari 1998). Spain and Brazil could not be more different in their reservoirs of social capital. In their study of social capital in twenty-nine market economies, Knack and Keefer (1997) found that Spain was only slightly below the average score for the level of trust (34.5; the mean was 35.8) and ranked among the advanced industrial countries in levels of civic engagement and associational density. Brazil, by contrast, was ranked dead last on these indicators. Given the tendency for social capital to concentrate in countries with greater and more equal distribution of income, the fact that Brazil remained one of the world’s most unequal countries helps to explain these results. Despite such national differences between Spain and Brazil, both countries maintained a composite subnational pattern of synergy. Synergistic ties emerged in some regions and not in others within both countries. Associative networks became the basis for innovative industrial policy-making in certain regions; these had incumbents interested in delegating authority to subnational political technocracies with developing levels of horizontal embeddedness. Other regions without these factors failed to develop synergy. The presence of more favorable economic conditions, greater fiscal resources, the existence of stronger and more coherent parties, EU membership, and more voluminous reservoirs of social capital in Spain cannot explain the failures. Likewise, the absence of these factors in Brazil cannot account for the emergence of synergy in certain cases. Such composite patterns of subnational adjustment require that scholars put aside analyses of national policy-making and utilize a more precise subnational comparative method (Snyder 2001).

Given the complexity of "composite economies" and the importance of the qualitative features of variables such as horizontal embeddedness and synergy, my chosen method of study utilizes qualitative subnational case study comparisons. While some of the aspects of synergy and horizontal embeddedness are quantifiable (for instance, heightened productivity and frequency of contacts), I agree with Locke and others in maintaining that these quantitative aspects are less important than their qualitative attributes. The emergence of mutual understanding and cooperation among agencies, and between agencies and firms, depends upon human perceptions of complex interactions over time, which are only poorly quantifiable. This does not mean that such objects of study cannot be assessed using a rigorous comparative method and utilizing concrete indicators of the variables.

Having fleshed out my main variables and their indicators in this chapter, I will use the case studies to demonstrate how these variables operate in comparative perspective. I substantiate these variables in the empirical cases through careful use of interviews of politicians, agency directors, middle-level policy-makers, firm managers, labor union organizers, and other actors. The analysis is also based on government documents, internal memos, secondary sources, and periodicals. In this study, I focus on two subnational cases in each country: in Brazil, the states of Minas Gerais and Rio de Janeiro; in Spain, the regions of the Principado de Asturias and Andalusia. Minas Gerais and Asturias developed significant levels of synergy during the 1980s and 1990s while Rio de Janeiro and Andalusia failed to create viable systems of publicprivate cooperation. Although attempts were made in all four cases, the political interests and level of horizontal embeddedness governing subnational industrial policies favored the emergence of synergy in Minas and Asturias. Histories of pervasive clientelism and political polarization in state politics hindered the creation and maintenance of industrial policies in Rio and Andalusia. Selection of these subnational cases was influenced by the complexity of the economic problems they faced. I chose subnational cases that could be considered "hard cases"—secondary industrial regions facing sweeping political and economic change that would present challenges to subnational industrial policy-making. Asturias and Minas Gerais were unlikely places for synergy. Each was a secondary location of industrial investment, unlike the core economies of Madrid/Catalonia and São Paulo. Severe economic crises in both regions challenged the efficacy of subnational policy responses. Moreover, these regions were either ordinary or impoverished in their supply of the ideal-typical cooperative norms of Putnam’s theory. Asturias remained a hotbed of labor conflict throughout the period under study. Minas Gerais’s development continued to be based on a combination of socially undifferentiated and isolated industrial towns and significant levels of labor repression. I subject the role of elite polarization and horizontal embeddedness in explaining the unlikely emergence of synergy in Asturias and Minas Gerais to a more rigorous test by comparing these experiences across both countries. Despite the economic and political characteristics reviewed above that should place Spanish regions in a position superior to that of Brazilian states, the comparison of Asturias and Minas demonstrates that levels of elite polarization and horizontal embeddedness explain the advent of synergy in these two cases. Figure 1.5 illustrates the distribution of the cases, which I preview briefly below. In Chapter 2, I compare the political and economic histories of Brazil and Spain. The analysis focuses on transformations in both countries dating from the 1970s through the mid-1990s. I outline the substantial differences between the two countries in the development and reform of industrial markets, the evolution of economic crises, the management of institutional reforms (particularly democratization and decentralization), and relations with supranational entities. Despite these differences in national institutions, both countries experienced growing interregional competition and composite subnational patterns of synergy. Chapter 3 examines economic policy-making in the Brazilian state of Minas Gerais. In Minas, industrial policy emerged as a focus of the state’s cohesive political class of traditional elites during the 1960s. At that time, the state government developed strong political interests in breaking the economy’s dependency on São Paulo’s larger and more dynamic industrial markets bordering Minas’s south. Amid a fiscal reform initiated by the military, Minas Gerais’s political leadership positioned itself to claim new resources and authorities for the state’s development policy. Intergovernmental conflicts with other states and the federal government encouraged Minas’s leaders to construct an elaborate economic technocracy designed to attract new investment to the state. These public agencies established close relations with numerous large investors, in- cluding the Italian multinational Fiat, which formed the core of the state’s most important industrial sector. Horizontal embeddedness in this period was plan-oriented. Minas’s economic bureaucracy expanded, but the fiscal crisis of the 1980s decimated Brazil’s state-led developmentalist model, threatening Minas’s industrial policy and calling for a revision to the state’s development mission. As these conditions changed, political technocrats seized the opportunity to reorient the state’s industrial policy to support productive changes in the region’s automotive industry. Horizontal ties across the state’s development agencies produced the interdisciplinary knowledge required by automotive projects, but they also established lasting political ties. During the mid-1980s, a populist governor attempted to scavenge fiscal resources from the Minas Gerais agencies. Without the support of a delegative government, Minas’s plan-based horizontal embeddedness became more project-focused. Persisting horizontal ties linking the state development bank, the entrepreneurial information agencies, and the state’s utility companies created a tight-knit constituency within the bureaucracy that helped preserve the industrial policy system. Later, these horizontal ties provided communicative links among agency elites that made possible a dramatic reform in Minas Gerais’s economic bureaucracy and development policy, returning it to a plan-based form of horizontal embeddedness. The state’s industrial policy was reoriented, directed away from developmentalism and toward a market-oriented approach. As market criteria became more important in economic policy-making, horizontal ties proved more valuable in limiting clientelistic exchange. The nature of horizontal ties was both formal and informal. Legislation provided Minas Gerais’s industrial policy agencies with different authorities and resources that called for coordination among them on complex projects. Relations across the state secretariats of planning and the economy created regular executive-level communication among agency administrators. These contacts also operated at an informal level as technocrats communicated routinely and across agency hierarchies at the executive and middle management levels. Given such close and constant ties across agencies, the costs of cross-agency monitoring were low, limiting the opportunities for politicians and bureaucrats to pursue their parochial interests. Moreover, these ties created a lasting level of confidence among firms and agencies that, despite changes in the economic model and the threat of political intervention, cooperation would persist. These conditions produced industrial policies that greatly enhanced the productivity of Fiat and its suppliers during the 1990s, shifting states in global markets.

Chapter 4 analyzes the industrial policy network in the Principado de Asturias in Spain. The Asturian economy depended upon two declining publicly owned industries: mining and steel. Due to the problems of these sectors, labor unions with strong ties to the Socialists remained the most powerful voices in regional politics. This gave the PSOE a solid base of organized support, allowing one Socialist regional president, Pedro de Silva, to govern Asturias for most of the 1980s. In Asturias, the development mission formed around a set of assumptions emerging from decades of public intervention in mining and steel. For much of the 1960s and 1970s, Asturian politicians and the economic bureaucracy accepted the fact that the inefficient mining and steel sectors could not be weaned off national subsidies. Periodic fiscal problems in Spain, and the long, slow decline of public mining and steel in neighboring European countries, convinced the Asturian leadership that this was an unsustainable situation. The government also assumed that the region’s powerful labor unions would not change their preferences in the future; they would attempt to keep public mining alive in Asturias by defending its subsidies and their jobs. Once Spain became mired in a "stagflationary" crisis after 1973 and the pace of political change accelerated with the death of Francisco Franco and the transition to democracy, the Asturian government believed that the time had come to change the direction of the region’s economy. By delegating economic authority and fiscal resources to a handful of bureaucratic agencies, the Asturian leadership believed that it could develop alternative industries in the mining and steel districts. At the same time, building a complex industrial policy and delegating authority to its technocratic management satisfied several political interests of the Asturian leadership. Like their nationalist counterparts in Catalonia and the Basque Country, Asturias’s governing politicians wanted greater political autonomy for the regional government during the transition to democracy. While the Asturian leadership could not credibly use nationalist arguments as the Catalonians and the Basque did in their campaign for regional autonomy, and thus could not access the fiscal authority that would go along with it, the Asturians could compete against other regions for national resources by referring to Asturias’s severe economic crisis. Promoting a regional industrial policy would justify the constitutional devolution of authority and resources to the Asturians, a position that both the Socialists and their political opponents favored. In order to place additional pressure on Madrid, the regional government framed the new Asturian industrial policy as a "compensatory" program for jobs soon to be lost in public mining.

Knowing that the labor unions would resist national restructuring of the public mining firms and that their claims would have to be addressed by national politicians in the new democracy, Asturias’s industrial policymakers deftly used references to "social compensation" to piggyback their interests on those of the region’s unions and position themselves for garnering additional authorities and fiscal resources from the central government. Empowered with substantial resources and political support, the technocrats at the helm of the Asturian development agencies created an array of new programs designed to improve the productivity of small- and medium-sized firms, the region’s chief source of employment. The planbased network matured, producing heightened confidence among the region’s firms in the role of the regional public sector; this was an unlikely outcome in a Spanish region haunted by the destructive role of the national public sector in mining and steel. Even when the Socialists lost control of the regional government in the mid-1990s, ties among the industrial policy agencies, firms, and labor unions preserved these confidence- building effects, although the system became more project-based. Chapter 5 outlines two cases that form a sharp contrast to the experiences of Minas Gerais and Asturias: Rio de Janeiro (Brazil) and Andalusia (Spain). In these two cases, chronic and polarizing political conflicts among elites created strong incentives for governing politicians to centralize their control over economic policy-making. The logic of populist politics led these politicians to manipulate industrial policy to satisfy their constituencies and isolate their rivals. The results for the economy were the generation of severe inefficiencies in the management of fiscal resources and inaction on the needs of firms. In comparison to Minas’s political leadership, Rio de Janeiro’s political class was highly conflictual. It depended upon the rise and fall of numerous populist governors such as Antônio de Pádua Chagas Freitas and Leonel Brizola. These politicians centralized their control over industrial policy agencies and used them to cultivate clientelistic networks of support. As a result, the development mission in Rio was weak and interagency ties were nonfunctional. Vertical ties emerged in some sectors, but these became mechanisms for building a clientelistic constituency. A prominent example were the "industrial districts" managed by the state’s Company of Industrial Development (CODIN). Brizola, and his successor between 1986–90, Wellington Moreira Franco, centralized control over the districts and used them to subsidize land speculation. CODIN also became a clientelistic resource as a center of political appointments in state govern- shifting states in global markets.

As a result, the CODIN could develop few workable linkages with national or state-level agencies, business, or labor groups. Firms continued to distrust the agency, even as a new governorship during the mid- 1990s attempted to salvage the state’s industrial policy system. Andalusia’s governing Socialists were a powerful presence in the region, but early nationalist opposition and periodic confrontations with the region’s chief business association threatened to erode the PSOE’s constituency soon after the region gained statutory autonomy during the Spanish transition to democracy. Socialist leaders were forced to buttress their support base with populist-style distribution. They manipulated industrial policy to build their constituency among the working class during a time of great economic uncertainty in Andalusia. This practice created incentives for regional elites to centralize administration, facilitating the control incumbents could exert over the region’s industrial reforms. The resulting hierarchical-functional form of interagency ties that developed in Andalusia generated high costs in the implementation of industrial policy. Andalusia’s politicians embraced unsustainable big-ticket industrial policies that were fiscally and economically costly. As the Socialists moved to consolidate their electoral constituency through the use of industrial policy, loyalists manipulated the resources of the region’s chief development agency, the Andalusian Institute of Promotion (IFA). Social pressure from the region’s labor unions and chief business association generated incentives for politicians to embrace wasteful distribution of subsidies and public buy-outs. While these actions accrued political benefits to the governing Socialists, they produced few lasting examples of cooperation between regional agencies and private firms. In Chapter 6, I reconsider the causes and incidence of synergy by comparing the data cross-regionally and cross-subnationally and with reference to other subnational cases in Spain, Brazil, and other countries. I then examine the populist-dependent and dependent types with a longitudinal comparison of the Mexican states. Finally, I assess the theoretical significance of these findings for emerging research on the political economy of decentralization.

© 2002 Penn State University