Decentralization and Municipal Governance: Suggested Approaches for CrossRegional Analysis* Eduardo J. G6mez The study of decentralization and municipal governance has captured much scholarly attention. This article highlights the importance of factors that have been generally ignored, and, in the process, suggests dimensions that facilitate comparison, including at the cross-regional level. First, regarding the creation and reform of decentralization policy, scholars may compare cases based on the horizontal and ex-post vertical political processes of reform. Second, cases can be compared based on the degree of center-state policy fluctuation, i.e., the institutions and incentives generating continual policy change and delayed outcomes, over time. Finally, I encourage scholars to scale down to the municipal level, comparing cases based on the following variables: historical state-municipal fiscal relations, institutional innovations, and the policy-making process. I close by explaining the various benefits associated with these approaches and the new research questions and challenges that they pose for comparative scholars. ecentralization and d e m o c r a c y are e n g a g e d in an intricate dance. Over the past two decades, nascent democracies in the developing world have designed and worked with new decentralization programs at the subnational level ~ in order to increase participatory governance, social welfare, and economic growth. In attempting to assess how effective these new initiatives have been in achieving these objectives, scholars have begun to unravel and explain the political logic of decentralization (Grindle 2000; Samuels and Montero 2003; Willis et al. 1999); its economic, political, and p o l i c y - m a k i n g consequences (Burki et al. 2000; Montero 2000; Samuels 2003); and the institutional and b u r e a u c r a t i c obstacles limiting d e c e n t r a l i z a t i o n ' s e f f e c t i v e n e s s
D
Eduardo J. G6mez is a doctoral candidate in the Department of Political Science at Brown University. His research interests are in federalism, decentralization, and the politics of economic and welfare reform in developing nations. Studies in Comparative International Development, Fall 2003, Voi. 38, No. 3, pp. 57-80.
58
Studies in Comparative International Development / Fall 2003
(Litvack et al. 1998; Rodden et al. 2003). New studies have also emerged addressing these issues at the municipal level: that is, the obstacles and incentives for state and municipal politicians to increase electoral competition, accountability, and economic growth (Manor 1999). That many of these publications have emerged within the past five years-and many are in progress--is evidence that studies in decentralization and municipal governance are an important current topic of research. International research associations and institutions are focusing on these issues in order to establish a dialogue between scholars and policymakers. In 2000, for example, the Latin American Studies Association created a new section, "Decentralization and Subnational Governance," in response to a substantial increase in panel submissions on these issues. Meanwhile, the World Bank recently created the "Decentralization Network," a website dedicated to sharing knowledge and empirical data through World Bank and external scholarly reports (World Bank, 2000a). In light of growing interest in decentralization, this article suggests some strategies for studying these policies at the cross-regional level. Given that many developing nations are experiencing similar challenges in creating and subsequently revising decentralization policy, this article offers approaches for comparing cases. To date, most of the work on these issues has focused on particular regions, e.g., Latin America, and single case studies. Yet, this approach fails to partake of the rich empirical and theoretical benefits associated with comparing different nations sharing similar institutional challenges and the consequences of decentralization. Moreover, researchers have conducted comparative analyses among similar regime types within particular regions and detailed small-N case analyses of subnational developments. Nevertheless researchers have failed to consider and propose alternative strategies for broadening our comparison of the decentralization process. Furthermore, they have not paid enough attention to the benefits this approach offers for redefining governments, refining data, and building theories. In this article I develop new approaches that can help us address these shortcomings. With regard to the creation and reform of decentralization policy, I conduct comparisons between different regime types (e.g., Brazil, Russia, and India) by focusing on the following underlying independent variables. First, I examine the horizontal and ex-post vertical bargaining relationships that occur between executives, political parties, and subnational governments, relying less on formal regime structures as predictors of policy outcomes than on the similar underlying horizontal and vertical relationships between actors controlling the reform process. Second, we can compare cases based on the variable of center-state policy fluctuation: that is, the degree to which decentralization/re-centralization policies undergo frequent change, in turn reducing the long-term efficacy of decentralization. After c o m p a r i n g cases based on the presence of policy fluctuation, cross-regional comparisons can be accomplished by focusing on the underlying causes of this variable: 1) the presence of juridical laws and informal relationships that are either open or closed in permitting future amendments to decentralization policy; 2) the sequencing of decentralization; and 3)
Gomez
59
the economic conditions shifting the bargaining leverage between national and subnational governments. Using these criteria provides a set of institutional, sequential, and economic variables that allow us to compare institutional sequences of policy fluctuation within and between nations and to explain variations in decentralization outcomes between sectors. Finally, this article suggests ways scholars can conduct cross-regional analysis at the municipal level. Studies analyzing the political process and consequences of decentralization at the municipal level are often case-specific rather than comparative. In order to push the research agenda forward, I encourage scholars to consider comparisons based on the following variables: 1) historical state-municipal fiscal relations and autonomy; 2) new institutional innovations in response to decreased municipal autonomy; and 3) mayoral-legislative relations and the policy-making process. Choosing cases based on these criteria allow us to re-define and re-code the efficacy of decentralization policy at the subnational level while providing insight into how we can explain variations in municipal autonomy. I close this article with some considerations for researchers interested in recoding and re-classifying decentralized regimes based on variations in underlying causal relationships. Lastly I reevaluate methods in cross-regional m u n i c i p a l analysis, highlighting recent limitations in properly coding subnational governments while suggesting alternative approaches.
The Underlying Mechanisms of Decentralization: Approaches to CrossRegional Analysis A new consensus has emerged among academics and policy-makers that cross-regional comparisons of the processes and consequences of decentralization offer new insights relevant to other regions not included in the initial cross-regional analysis (Hutchcroft 2001; G6mez and Smoke, forthcoming). For example, some studies concentrate on how certain types of electoral systems---e.g., open versus closed-list--lead to inefficient fiscal decentralization outcomes (Willis et al. 1999). Lessons drawn from these findings have been applied to different nations, such as Brazil and Spain (Montero 2002), the Philippines and Argentina (Eaton 1998), and several Latin American countries (O'Neill 1999). On the other hand, studies of institutional approaches to decentralization are yielding important cross-regional lessons. Rodden, Eskeland, and Litvack's (2003) recent work on decentralization and soft-budget constraints illustrates, for example, how weak federal institutions and electoral systems in Brazil, Russia, and India tend to prolong soft-budget constraints and threaten macroeconomic stability. While these lessons in institutional design and outcomes are important, there is an emerging consensus that no direct causal relationship exists between the structure of political institutions and policy outcomes. Indeed, recent research suggests that strictly focusing on federal governance structures, electoral systems, and party system design does not provide the theoretical leverage needed to explain decentralization outcomes (Eaton 2002; G6mez 2001; Eaton and Dickovick, forthcoming; Escobar-Lemon 200l). Unfortunately the common
60
Studies in Comparative International Development / Fall 2003
trend in comparative decentralization is to focus exclusively on comparisons of similar political party (electoral systems) and presidential-legislative relations as predictors of policy outcomes (Willis et al. 1999; Rodden et al. 2003; Burki et al. 2000). Yet unexpected outcomes abound, suggesting, for example, that the degree of political party fragmentation and polarization (Haggard and Kaufman 1995) does not delay the passage of new fiscal decentralization and re-centralization measures---e.g., Brazil's new Fiscal Responsibility Law. Furthermore, Echeverri-Gent (1998) finds that in the case of India, Haggard and Kaufman's (1995) position that political party polarization thwarts the reform process does not hold, as excessive polarization at times facilitates the decentralization process, especially when opposition parties are so paralyzed that they cannot organize an effective coalition in opposition to reforms (EcheverriGent 1998). In Argentina, on the other hand, low levels of party polarization within traditionally stable, closed-list electoral systems, once hailed for creating loyal political party members and facilitating the decentralization process (Dillinger and Webb 2001a), have recently thwarted new decentralization policies due to conflicts of interest over personal political ambitions between leaders in the governing party and the president (Corrales 2001). The discovery of these kinds of paradoxes suggests that we should search for alternative explanations of decentralization policy. The approach I offer breaks from the emphasis on regime type, and focuses instead on the underlying political relationships driving this process. Therefore I build on and support Karen Remmer's (1990) position that "the political similarities and differences among nations appear to be far too complex and multifaceted to be captured by simple distinctions among types of regimes. As a result, knowing that a regime is civilian rather than military, democratic rather than authoritarian, or even exclusionary rather than inclusionary establishes only a limited basis for making predictions about policy outcomes" (Remmer 1990: 316). Instead one should focus on "underlying relationships," which may consist of institutions or less formal structures. In this regard, two factors are particularly important. First, the horizontal relationships between the executive (i.e., presidents and prime ministers), political parties, and institutions (committees and/or bureaucracy) responsible for creating decentralization policy. And second, the ex-post vertical relationships (bargaining) between executives and subnational governments and between federal legislators and subnational governments over changes to decentralization policies. While vertical relationships co-exist with horizontal relationships, they vary in degree and influence: that is, they are weakest at the ex-ante policy-making phase and most effective--and visible--during the ex-post consequences of decentralization. During the ex-ante phase, subnational politicians try to influence the details of decentralization policies by funneling requests and resources (such as campaign finance) to supportive legislative members (Abrucio and Samuels 1999). Yet because the outcomes of decentralization are uncertain during the ex-ante phase, rarely do we see vehement, well-organized opposition to reforms. In contrast, the ex-post consequences of decentralization often motivate subnational governments to collectively organize in opposition to
Gomez
61
reforms. This is most prominent when the decentralization policies implemented reduce sub-national governments' economic and policy-making autonomy, or when the center decides to re-centralize policy after the consequences of decentralization unfold. Because the ex-post consequences are more likely to lead to policy change, I concentrate on the ex-post vertical relationships present after the first wave z of decentralization. In this section, I stress two components of ex-post vertical relationships that can be used for cross-regional analysis. These are corrective reintegration and re-centralization bargaining relationships--top-down measures initiated by the executive--and subnational pressures of accommodation and adjustment-bottom-up measures used by subnational governments to justify alterations to decentralization policy.
Horizontal and Ex-Post Vertical Relationships Political science offers several approaches for cross-regional analysis through the horizontal and vertical processes of creating and reforming decentralization policy. To date, most of the work in this area has been regime- and regionspecific. Recently, for example, a number of Latin Americanist scholars, including Haggard and Webb (2003), Willis et al. (1999), O'Neill (1999), Eaton (1998), Samuels and Montero (2003), Penfold-Becerra (2001), and EscobarLemon (2001), have been tackling these issues. In this article, horizontal relationships are defined as executive-political party relationships and decision-making regarding the devolution of economic and social policy. Executives and political party members are always debating how to decentralize responsibilities and how and when to enforce federal regulations governing these measures. In some cases, for example, we may find that executives are stronger players in their relationships with political parties--an executive-led process. This situation leads to the timely devolution and enforcement of policy, either because the executive wields uncontestable decree authority or because it is able to craft a governing coalition that is stable and supportive-for example, the cases of Argentina (until 1998), Colombia, and Hungary? Though prevalent in several Latin American democracies, executive-led processes can also be found within different regime types. Susan Shirk's analysis of the decentralization process in China provides a good example (Shirk 1993). Indeed, Shirk begins her analysis with the following statement: "My starting point is the unusual idea that we can study policy-making in communist countries much as we would in non-communist countries by looking at the patterns of competition among politicians who operate in an institutionalized political setting" (Shirk 1993: 7). Thus, notwithstanding the absence of a formally fragmented and polarized political party system, as seen in democratic regimes, Shirk shows how provincial interests were represented and contested between pro-reform Central Committee members, conservative communist party members, and the premier over the decentralization of tax (specifically VAT) revenue (Shirk 1993: 15). The key lesson from Shirk's analysis is that there was a substantial amount of subnational representation and contestation over fiscal decentralization, as seen in several Latin American countries. Contrary to what
62
Studies in Comparative International Development / Fall 2003
many scholars would assume, Shirk shows that decentralization was very much a representative political process that relied to a great extent on the premier's leadership and ability to organize coalitions in favor of reform. On the other hand, executives are weaker players in horizontal relationships when political party members continually oppose decentralization policies, typically because they ideologically oppose decentralizing authority. In this legislative-led process, party members in federal legislatures have career incentives to support subnational interests by either sharing more fiscal revenue or opposing measures limiting subnational governments' fiscal and hence policy-making autonomy. Supporting subnational interests in turn increases the political party and/or party members' electoral prospects at the national and subnational level. My research on Brazil and Russia, for example, shows how autonomous legislatures can hamper the executive's ability to create and enforce policies limiting the amount of money subnational governments can borrow (G6mez 2001). In both cases, federal constitutions delegated complete authority to the Senate and Federal Council, respectively, over budgetary reform issues, especially limitations in subnational borrowing. Because senators in Brazil and Federal Council representatives in Russia do not have career incentives to impose borrowing controls, they often disagree with the executive's call for reform and hence delay policy implementation (on Brazil, see Haggard and Webb [2003]; on Russia, see Martinez-Vasquez and Boex [2001]; Sutela [1999]). Furthermore, in both cases, fiscally constrained governors and mayors inheriting large stocks of debt and social expenditure responsibilities appealed to federal representatives to block borrowing controls. In doing so, they acted within a long tradition in Brazil and Russia of governors and mayors financing the campaign of congressmen and senators and appointing them to lucrative posts within state and municipal administrations (on Brazil, see Abrucio and Samuels [1999]; on Russia, see Kirkow [1996]). Ex-post vertical relationships, on the other hand, represent the bargaining that takes place after decentralization. In this case, bargaining occurs between the executive and subnational governments, on the one hand, and between the legislature and subnational governments, on the other. Ex-post vertical bargaining is a static process that occurs immediately after the devolution of fiscal and administrative responsibilities, thus falling under the second wave of decentralization. As I discuss in the next section, however, the nature of institutional systems representing subnational interests, when combined with the effects of sequencing in decentralization policy and economic crisis, adds more momentum to the ex-post vertical bargaining process and ultimately leads to problems of center-state policy fluctuation. Just as horizontal relationships vary depending on who is the stronger player, ex-post vertical bargaining relationships may be led by either executives or subnational actors. The first type, executive-led processes, may be classified as corrective reintegration or re-centralization mechanisms. Entailing different measures, both are used when the executive realizes that decentralization policies were implemented too quickly, and that the institutions needed to properly regulate and enforce polices were never properly installed. Executives
Gomez
63
initiate corrective reintegration measures by changing the internal bureaucratic structure of decentralized administrative branches (e.g., changing wage-payment policies, where the center pays local officials rather than local governments in order to reduce corruption, increase autonomy and loyalty to the center). This, of course, is met with staunch resistance from local governments accustomed to manipulating administrative decision-making and controlling the distribution of resources for political gain. Dali Yang's (2001) analysis of the vertical reintegration of Central Banks in China provides a good example of corrective reintegration measures. Following the rapid decentralization of banking industries in China, the governors' control of banking policies through the staffing and financing of salaries increased corruption and informal lending, ultimately contributing to a major banking crisis. In an effort to avoid further problems and enhance financial regulation, in 1998 Central Bank governor Zhu Rongji succeeded in re-integrating the wage-payment system, thereby truncating the governors' influence while reducing the number of central bank offices from thirty-two to nine regional banks (Yang 2001). Similar corrective reintegration measures are taking place in Vietnam, under the president's new PAR Master Program4 (Wescott, forthcoming). And in Brazil, former President Fernando Henrique Cardoso instituted new administrative reform programs reintegrating control over financial auditing policies at the state and municipal level through new staffing procedures within the Tribunal de Contas, equivalent to the United States' General Accounting Office (Nascimento 2002). On the other hand, executives may enact a vertical re-centralization process, that is, an attempt to reassert greater control over particular decentralization policies, such as fiscal. This typically occurs when the executive realizes that subnational governments were not prepared to concomitantly handle new economic and social policies, and that by re-centralizing fiscal policy, they can better control local spending in order to provide social services more efficiently. Several re-centralization strategies may include: 1) re-centralizing the distribution of fiscal revenue (Eaton and Dickovick, forthcoming); 2) imposing hard-budget constraints and borrowing controls (Rodden et al. 2003); and 3) seeking the privatization of state-owned enterprises and banks (Eaton and Dickovick, forthcoming). In contrast to corrective reintegration and re-centralization (both initiated by the executive), under a subnational-led vertical bargaining process, state and municipal governments approach the center with complaints about the consequences of specific policies devolved under previous administrations. As seen in Brazil and China, subnational governments approach the center because of a need for accommodation and adjustment: that is, they typically resist the initial design of decentralization policies after realizing that they cannot afford to adhere to newly imposed fiscal and expenditure requirements, thus requiring more time to accommodate and adjust (on China, see Shirk [1994]; on Brazil, see G6mez [2000]). This is most prominent when governors and mayors inherit outstanding debts from outgoing administrations that constrain their ability to finance expensive social expenditure requirements, as seen in Brazil (Dillinger and Webb 2001a). Alternatively, leaders of large mu-
64
Studies in Comparative International Development / Fall 2003
nicipalities may object to inherited debts, but may ask to receive the proper amount of revenue guaranteed to them through the constitution in order to service debts in a timely manner, as seen in India (McCarten and Vyasulu, forthcoming). As discussed later in this article, any subsequent change in fiscal and social policy as a result of these accommodation and adjustment measures will be shaped by the presence of either open or closed institutions, informal relationships, the sequencing of decentralization, and the extent of economic crisis. Haggard and Webb's (2003) position that these vertical relationships coexist with and are shaped by horizontal relationships is well taken. A harmonious vertical relationship (which occurs only rarely) exists when central and local governments not only agree to the design of decentralization programs, but also to the subsequent pace and depth of these measures. The center's role in its vertical relationship with subnational authorities is strongest when horizontal mechanisms are strong, that is, when the executive and political parties agree to recentralize measures, despite subnational opposition (O'Neill 2001). Conversely, the center is a weaker player in the vertical relationship when subnational governments influence decentralization policy by bribing legislative representatives and marshalling coalitions that oppose reform. In sum, notwithstanding formal differences in regime type, because developing nations experience similar horizontal and ex-post vertical political processes, scholars should consider comparing cases based on these variables in order to conduct cross-regional analysis. Comparisons based on these variables not only facilitate cross-regional analysis by helping scholars hone in on particular causal variables--especially in a context where multiple actors at different tiers of government are trying to influence the decentralization process. They also strengthen our theoretical and empirical understanding of the general causal factors determining the pace and depth of decentralization and re-centralization, in turn revealing how and to what extent decentralization is actually working in the developing world.
Policy Fluctuation Center-state policy fluctuation provides yet another important dimension for comparison. I define policy fluctuation as the frequent, rapid change-over time in the decentralization and re-centralization of policies, in turn leading to inefficient policy outcomes. Thus policy fluctuation is distinguished from static vertical relationships because it illustrates continuous change in policies and outcomes, and the institutional and economic conditions contributing to this process. As a result, because policies are always changing, the relevant actors do not have firm expectations about the location of power, that is, who is making the decisions and their capacity to implement policies. Consequently, decentralization policy is often ephemeral in nature, limited in its ability to meet policy objectives. To begin to understand the institutional conditions leading to these outcomes, we can focus on three underlying causes of policy fluctuation: 1) the presence of juridical laws that are either "open" or "closed," and informal
Gomez
65
relationships facilitating future amendments to decentralization policy; 2) the sequencing of decentralization; and 3) economic conditions that shift the bargaining leverage between national and subnational governments. The presence of formally open or closed juridical laws provides a good point of departure for understanding the institutional conditions leading to inefficient policy outcomes. Building on Alfred Montero's (2000) work explaining changes in decentralization policy after the first wave of decentralization in Brazil, Venezuela, Spain, and Mexico, we can distinguish formal laws based on their degree of openness and closure with regards to subnational interests. Both in Montero's work and in this article, openness does not suggest that federal legal institutions are flexible, that is, consistently changing in response to federal/subnational interests. Rather, they are open because they permanently represent the interests of subnational politicians. Indeed, Montero shows that institutions are "open-ended" when national laws and associations permanently represent the interests of the governors, which in turn allow them to periodically influence and change decentralization policy (Montero 2000). In Venezuela, for example, Montero (2000) and Michael Penfold-Becerra (2001) show that the central government's creation of the Association of Venezuelan Governors (AGV), housed in the newly formed Ministry of State for Decentralization, allowed the governors, by law, to collectively approach the center and pressure for changes to intergovernmental fiscal transfers. This law in turn allowed Venezuela's governors to negatively influence changes in the distribution of fiscal revenue and expenditure responsibilities to the states (Montero 2000: 12). By contrast, when institutions are "closed," subnational interests are not legally represented at the federal level. Under these circumstances executives are able to completely control the decentralization process, periodically decentralizing and recentralizing as they see fit. We find similar legal processes occurring within different regime types. In Russia, for example, Kathryn Stoner-Weiss's (1999) work illustrates how open laws and representation allowed governors to periodically amend the federal budget. Through the Federal Council (the Senate's upper chamber), the constitution gives the governors complete discretion over budgetary matters. Because governors met periodically in the Federal Council and voted on policies affecting their regions, amendments to budgetary laws, expenditure responsibilities, and fiscal distribution (especially the VAT) were continually changing (Kirkow 1996; Sutela 1999). Like Brazil, moreover, Federal Council and Duma (congressional) members were disposed to take advantage of these legalities and periodically amend budgetary policy because many aspired to other state offices, or wanted to get re-elected, or retire into industry (Kirkow 1996). Hence, as in Venezuela, the governors' legal right to question and amend budgetary policy led to a burgeoning fiscal crisis at the subnational level (Martinez-Vasquez and Boex 2001). On the other hand, researchers have ignored how informal relationships contribute to policy fluctuation. As Douglas North (1989) maintains, informal customs of behavior lead to law-like expectations of mutual behavior. Similarly, governments periodically engage in informal relationships with subnational governments. Moreover, formal laws are often amended because binding, in-
66
Studies in Comparative International Development / Fall 2003
formal expectations motivate politicians to change them. Expectations driving institutional reforms in the area of center-state fiscal relations take the form of periodic soft-budget constraints, financial bailouts, and debt workouts, found and analyzed comparatively in several nations (Rodden et al. 2003). It is important to note that these mechanisms provide both direct (e.g., soft-budget constraints) and indirect (e.g., debt workouts) fiscal transfers benefiting national and subnational politicians. Moreover, while both measures provide immediate financial relief, there are more payoffs with debt workouts. These bailouts are intended to cover a longer period of time as debt repayment contracts signed in the past are continuously extended in response to new economic conditions, thus providing more time for governors to marshal the resources needed to payoff debts. When the center is accustomed to providing these measures periodically, local political actors' expectations of federal intervention often cause them to periodically collectivize and pressure the government for greater relief. Let us examine the cases of Brazil and India. My research on Brazil shows that subnational expectations carry the same weight as formal laws, and that these expectations periodically influence members of congress and senators to either avoid the imposition of hard-budget constraints or amend formal laws on subnational borrowing and debt repayment (Grmez 2000). A typical pattern ensues as legislative representatives (especially in the senate) are motivated to bail out banks and provide debt workouts when facing national elections or seeking lucrative employment at the subnational level, which, in turn prolongs subnational fiscal crisis (G6mez 2000). Similarly, in India, McCarten (2003) shows how the 1993 fiscal austerity program induced indebted chief ministers (governors) to appeal to accommodating bureaucrats in the National Finance and National Planning Commission for the forgiveness of public debt (McCarten 2003). Meeting every five years and recommending policies to parliamentary budgetary finance committees, commission bureaucrats have periodically postponed the repayment of debts owed to the center. As in Brazil, McCarten (2003) shows how informal expectations often permitted--and incited--chief ministers to continuously approach the Finance and Planning Commissions in expectation of receiving periodic bailouts. The next question we must address, however, is how do open juridical institutions and informal relationships bolster the bargaining power of subnational governments? Are these factors necessary and sufficient for bargaining strength, or do we need to include other variables in our analysis? In the realm of center-state policy fluctuation, open institutions in and of themselves provide necessary but insufficient conditions. Rather, we need to consider how the sequencing of decentralization affects center-state bargaining relationships and policy preferences and to what extent economic crisis bolsters the bargaining power of subnational governments. These two variables further affect the bargaining process and lead to inefficient policy outcomes. To date, scholars have not paid enough attention to the issue of sequencing in decentralization policy. Most researchers have concentrated on the political process and the economic and institutional consequences of decentralization without addressing how the sequencing of decentralization--that is, the tim-
Gomez
67
ing (when) and process (how)--influence these developments in the first place. Recently, however, the work of Anwar Shah and Teresa Thompson, and Kent Eaton and Tyler Dickovick, is starting to address these issues. Through a comparative analysis, Shah and Thompson (2002) and Eaton and Dickovick (forthcoming) explain that the reasons why decentralized fiscal and welfare policies often fail is because subnational governments are given too much discretion (borrowing, expenditure, and administrative) either too quickly or too slowly. For the cases of Brazil and Argentina, Eaton and Dickovick (forthcoming) find that because the federal constitution hastily and completely decentralized financial borrowing and taxation privileges to state and municipal governments, 5 presidential efforts to stabilize the macroeconomy were delayed because of the prolonged haggling over changes to the revenue-sharing process. Hence, decentralizing too much fiscal discretion without the proper expenditure responsibilities in place will immediately increase the bargaining power of subnational governments, thus complicating macroeconomic management. The sequencing of decentralization is important because it provides the basis for understanding why subnational governments have the bargaining power that they do and to what extent high degrees of autonomy, over time, solidify expectations and customs, providing subnational governments with the ability to continuously influence decentralization policy. Furthermore, it is important to note that sequencing also generates new preferences and incentives for subnational governments to safeguard policies. If comprehensive fiscal policies are immediately devolved, for example, local governments will prefer to keep them and adamantly resist any subsequent re-centralization efforts (e.g., Brazil)6; this leads to perpetual renegotiation and conflict, in turn complicating the center's ability to manage and reform policy. An important ex-post consequence of the sequencing problem is the degree and influence of economic crisis. In periods of subnational fiscal crisis, for example, we often find that open institutions permitting local governments to approach the center will increase their bargaining power. Prior to elections, fiscal crisis increases local politicians' bargaining power because national representatives are seeking political advancement and are thus more inclined to support constituent needs. Second, fiscal crisis in addition to open institutions convinces local politicians that the costs of collectivizing in opposition to policy are low and that the chances of influencing policy are greater. For example, Michael Penfold-Becerra's (2001) research shows that even in Venezuela's closed-list electoral system, governors collectivized in opposition to reforms because they perceived that they could influence policy because of a) the prevalence of local deficits and debts and b) formal laws and institutions allowing them to voice their opposition--for example, through the Association of Venezuelan Governors (Penfold-Becerra 2001). Crisis and open institutions thus enabled governors to continuously bargain with the center. In conclusion, in order to understand the causal factors leading to decentralization's rapid policy fluctuations, scholars should consider conducting cross-regional analysis based on the institutional, sequential, and economic variables that influence the efficacy of decentralization policy. Notwithstanding differences in regime type, these causal variables exist within many devel-
68
Studies in Comparative International Development / Fall 2003
oping nations. Because of this I encourage scholars to employ and compare these variables in order to better understand why governments are still not partaking of the fruitful benefits associated with permanent, effective decentralization policies.
Scaling Down Even Further: Cross-Regional Municipal Analysis Municipalities often influence the world in significant ways. Yet comparativists tend to overlook the role that municipalities play in development. Especially within decentralized federations, scholars tend to focus on state governments' impact on economic reform, political parties, and electoral competition (Mainwaring and Samuels 2003; Abrucio and Samuels 1999; Kirkow 1996), without considering municipalities' contribution to this process. Why is this the case? Some argue that the neglect of municipalities has to do with the historic role that municipalities played in development. Andrew Nickson, for example, maintains that the transition to authoritarianism led to the de-municipalization of political influence and autonomy (Nickson 1995). That is, dictators re-centralized control over municipal governments in order to suppress political competition, even if municipalities were constitutionally recognized and played decisive roles in political development after independence--e.g., Brazil. Similarly, within post-socialist systems, such as Russia and India, dictatorships maintained control, setting developmental agendas for municipalities until the transition to democracy created new demands for grassroots democracy and municipal autonomy. Thus, centralization tendencies within developing nations t h r o u g h o u t most of the twentieth century marginalized municipalities politically and economically, in turn limiting their role in national policy-making. Scholarly attention to their role in development has thus been similarly limited. Nevertheless, the transition to democracy and the re-emergence of federalism and decentralization as key components of the re-democratization process have rejuvenated municipalities and increased their influence. In contrast to unitary systems, which are often heavily dependent on the center for revenue and expenditure responsibilities, the introduction and/or reform of federal constitutions provides new, more reliable sources of revenue for municipalities, while new demands for social services and increased accountability has transformed them into powerful political centers and springboards for the most prominent positions in government, such as the presidency (Myers and Dietz 2002). In contrast to unitary systems, moreover, parties created at the municipal level are often seen as key stepping-stones into national politics, as new candidates are introduced and coalitions formed at the municipal level. In addition, powerful urban centers are often viewed as lucrative posts for federal legislators after their tour of duty, as retired legislators often aspire to become powerful governors or mayors (Samuels 2003). Hence, because they influence the trajectory of politics within federations, municipalities have not only captured the attention of federal legislators, but also of comparative scholars interested in learning more about municipalities' burgeoning influence in democratic consolidation and policy reform.
Gomez
69
A problem with recent studies in municipal political economy, however, is that they tend to be case-specific rather than comparative. Most of the innovative work on these issues stem from policy reports published by international organizations, such as the World Bank, Inter-American Development Bank, and the Asian Development Bank. The World Bank in particular has published several case studies addressing varying degrees of municipal autonomy in fiscal and welfare policy. Nevertheless, World Bank experts are finally realizing that we need to start thinking comparatively in order to better understand the causes for these variations in autonomy and policy outcomes. Three factors seem particularly salient: 1) historical state-municipal fiscal relations and autonomy; 2) new institutional innovations in response to low levels in municipal autonomy; and 3) mayoral-legislative relations and the policy-making process. The first variable, historical state-municipal fiscal relations, can be seen as affecting the second, institutional innovations. The latter in turn has an impact on mayoral-legislative relations and municipal policy-making.
Historical State-Municipal Fiscal Relations Cases where municipalities have experienced historically low levels of fiscal autonomy provide good examples of a particular value 7 of the state-municipal fiscal relations variable that can be used for cross-regional analysis. That is, identifying cases where municipalities never received adequate fiscal revenues from the state and central government because of dominant party interests in suppressing local political competition allows the investigator to use statemunicipal relations as a control variable. As the next section explains, this approach may illuminate the common incentives and strategies that municipalities use to increase their fiscal (and hence policy-making) autonomy. For instance, a comparison between two dominant party regimes, e.g., Mexico's PRI (Partido Revolucionario Institutional) and India's Congress (I), 8 reveals that in both cases dominant parties have always had political incentives to limit the transfer of fiscal revenues to municipalities, where most of the opposition parties are formed (Edmonds-Poli and G6mez 2001). What is more, this process continued despite subsequent constitutional amendments requiring states to increase the flow of fiscal revenues to municipalities. Thus in both cases states sought to suppress municipalities' fiscal autonomy, in turn generating similar incentives for municipalities to find ways to increase their fisc a l - a n d eventually political--autonomy. In addition to historically low levels of fiscal autonomy, one should address the question of de facto fiscal decentralization. In developing nations there is often a large disparity between what the federal constitution says municipalities should receive in fiscal revenue and the actual revenue received. Nascent democracies with strong traditions of single-party dominance at the federal and state level often withhold fiscal transfers to municipalities controlled by the opposition. Moreover, this typically occurs in a context where the constitution explicitly mandates the periodic sharing of fiscal revenues between states and municipalities. We can find these disparities in de facto versus de jure
70
Studies in Comparative International Development / Fall 2003
decentralization within several developing nations (McCarten and Vyasulu, forthcoming; Edmonds-Poli and G6mez 2001). The fact that municipalities are experiencing similar disparities behooves us to compare cross-regionally in order to determine if there are potential differences in the tactics that dominant state governments use to limit municipalities' fiscal autonomy, such as continuous revisions to state constitutions and/or the periodic postponement of state finance commissions responsible for distributing revenue to municipalities (e.g., India; see McCarten and Vyasulu, forthcoming). Cross-regional assessments of state suppression tactics can provide insight into why municipalities are still not benefiting from decentralization policies and why the transition to democracy at the municipal level is occurring at a much slower pace when compared to federal transitions (G6mez 2002). Institutional Innovations
In response to low levels of fiscal autonomy, municipalities may use similar institutional innovations to increase their autonomy. Recent studies confirm the fact that municipalities in different nations share similarities in their approach to increasing municipal autonomy through new institutional innovations; this is especially the case when suppressive dominant parties historically control states. Researchers should consider two mechanisms: the development of municipal capital markets and increases in administrative efficiency. In Mexico and India, for example, Edmonds-Poli and G6mez (2001) explain how opposition political parties at the municipal level created new bond markets and enhanced local tax administration in order to increase their revenue base, popularity, and autonomy from opposing state governments. In Mexico, Edmonds-Poli explains how the opposition mayor for the PAN, Olvera Nieto, in the city of Tlalnepantla, used the modernization of tax-collecting institutions and the creation of new employment programs in order to get his party reelected in 1999. In India, on the other hand, I show how Ahmedebad Mayor Lilji Parmar of the opposition BJP within the traditionally Congress (O)-governed state of Gujurat created new bond markets in order to increase foreign direct investment and, with the help of the USAID, create new urban infrastructure projects, such as new sewage and water treatment infrastructure. 9 Related developments are also occurring in Indonesia. Rasyid (2002) explains how municipal and local governments, constrained by ambiguous fiscal and licensing laws implemented by the central government's decentralization programs in 1999, are trying to work around these laws by creating illegal licensing contracts in the fields of mining, forestry, fishing, and trade, while ignoring taxation policies. Although formal markets have not yet been developed, Rasyid (2002) argues that mayors and local leaders are responding to the central government's lack of attention by working around ambiguous laws in order to increase their municipal autonomy (Rasyid 2002: 7-9). Mexico and India show similarities in innovations between two different nations. But of course, even though different nations may exhibit similarities in historical state-municipal fiscal relations, they may not, due to a lack of technical capac-
Gomez
71
ity and/or political will, possess the conditions ripe for institutional innovations, thus increasing their dependence on the center. A further step will be to examine the conditions negatively influencing variations in institutional innovations, such as state-suppression tactics and the absence of state capacity, and what municipalities are doing to overcome these challenges.
Mayoral-Legislative Relations and Policy-Making While mayors may prove effective at collectivizing and pressing for policy change at the federal level, how effective are they in getting policies passed at the municipal level? Do they always enjoy the majority support, or do we see a legislative-led process of reform, where, as mentioned earlier, parties resist the implementation of decentralization measures based on personal career ambition and constituent needs? With the exception of Myers and Dietz (2002), we have yet to see any systematic comparison---especially cross-regional--of executive-legislative relations and policy-making at the municipal level. Several case studies suggest that intra-and cross-regional comparisons can be achieved, however. Based on these findings, I suggest the following variables for the purposes of comparing cases: First, mayor strength in policy reform, measured by the degree of executive decree authority and support within legislatures; and second, the extent of legislative-led opposition to reform, where parties are unwilling to support the mayors' policies because of either personal career ambitions or loyalty to constituents. Latin American scholars are now starting to unravel the municipal politics of reform. A special emphasis has been placed on the nature of mayor-legislative relations and the policy-making process (Myers and Dietz 2002). Akin to arguments made about the politics of policy reform at the federal and state level, the reform process at the municipal level has been shaped by the mayors' ability to implement and enforce new decentralization programs or by the legislatures' staunch resistance to reforms, t~ Myers and Dietz's (2002) recent volume explains, for example, wide variation in mayoral legislative strength, that is, mayoral ability to implement policy via executive decree and/or strong legislative support for reforms. Myers and Dietz (2002) find that mayors are often capable of passing legislation either when they act as powerbrokers (having secured the loyalty and support of presidents and political parties), or when they possess strong executive decree authority (as part of a governing coalition). On the other hand, recent evidence challenges Nickson's (1995) position that mayors have traditionally dominated policy-making through executive decree. For instance, Gilbert and D~ivila (2002) find that in the case of Bogotfi, Colombia, the mayor's ability to implement policy was often hampered by the Council of Bogot~i (legislature), which was riddled with party members who aspired to more prestigious positions in the federal government, including the presidency, notwithstanding their proclaimed loyalty to the mayor's party (Gilbert and D~ivila 2002). Since the constitutional provision of municipal elections in 1998, the Council's opposition has delayed the passage of several
72
Studies in Comparative International Development / Fall 2003
budgetary amendments and the appointment of new bureaucrats to administrative posts. Detailed case studies of local governments in Eastern Europe also reveal growing tensions between elected mayors and legislatures. In Poland and Hungary, for example, Regulski and Kocan (1994) and Maurel (1994) explain that municipal legislatures are extremely fragmented, polarized, and often unwilling to endorse the mayor's policies. This has been extremely problematic for the efficacy of decentralization programs (Bennett 1994). In Poland, the sources of legislative gridlock stem from deeply inculcated traditions of controlling the administrative process as direct agents of the central government. Notwithstanding the formal election process, policy legacies have induced municipal council members to reject budgetary and administrative reform policies initiated by the mayor (Regulski and Kocan 1994). What is more, Regulski and Kocan maintain that because of this, mayors are often afraid to resist opposition out of fear that council members will replace them if they do not adhere to the council's wishes (Regulski and Kocan 1994: 60-61). Similar experiences in mayoral-legislative relations and the policy-making process in Latin America and Eastern Europe suggest that there is ample room for cross-regional analysis. Comparing cases based on these experiences allows us to find the institutional rules and conditions accounting for variations in municipal legislatures' ability to implement decentralization policies. This, in turn, gives us a better indication of the effectiveness of decentralization policies at the subnational level and the future challenges that municipalities may encounter when adjusting to new circumstances.
Concluding Notes for Future Analysis The study of decentralization policy is a messy and complicated issue. In contrast to macro-level policy-making, decentralization entails the participation and influence of political actors at several tiers of government. Moreover, the fact that decentralization policies and outcomes are continuously changing presents further challenges. In response, the task of this article has been to propose new dimensions for comparative analysis, in turn facilitating comparisons of these complex processes. A good point of departure is to compare cases based on similar underlying horizontal, ex-post vertical relationships and sequences of policy fluctuation. Doing this not only facilitates cross-regional analysis, but also provides us with a new analytical approach based on the horizontal, ex-post vertical, and policy fluctuating political process of reform. This stands in contrast to the classic World Bank "tool kit" of indicators which focuses exclusively on economic and administrative similarities between nations (World Bank 2000b). In addition, these types of comparisons permit the discovery of subsequent causal differences between nations after these relationships are controlled for. In turn, this approach provides for more accurate descriptions of federal and unitary systems. If researchers select cases based on similarities in horizontal relationships, ex-post vertical relationships, and sequences of policy fluctuation, thereby
Gomez
73
controlling for these important influences, they may uncover other differences among cases that explain divergent outcomes. For example, countries such as China and Brazil exhibit differences in subnational bureaucrats' willingness to truncate ties with local politicians after the center has reintegrated wages and administrative practices. The critical difference may eventually lie in bureaucratic incentives to remain popular with local communities (Tendler 1997) and thus community-based pressures and incentives to oppose re-integration, as seen in Brazil. The converse may hold for China, where bureaucrats may never have had the chance to establish these types of traditions and incentives. These differences in bureaucratic incentives may in turn explain the subsequent pace and depth of reintegration--fast in China, slow in Brazil--which is a promising area of future research for scholars interested in examining the politics of reintegration after the first wave of decentralization. Regarding policy fluctuation, after observing striking similarities in open institutions, we may find that in the future these institutions generate few incentives for governors to collectively oppose the center as new presidents emerge with interests in re-centralizing policy and discounting the role of federalism. For example, although Russia and Venezuela do possess open institutions legally representing the interests of the governors, eventually new presidents may want to curb the impact of these institutions and hence reduce the governors' role in policy-making. President Vladimir Putin, for example, entered office wanting to curb the Federal Council's role in policy change, while proposing new policies reducing subnational autonomy. While scholars contend that he has not made much progress (Stoner-Wiess 2001), these efforts may in the future negatively affect the governors' perceptions that open institutions can effectively represent their interests. In Venezuela, on the other hand, we may find that the president, notwithstanding high levels of political instability, is still committed to the Association of Venezuelan Governors, and that the governors still perceive this institution as an effective way of pressing for policy change. Hence, after controlling for similarities in processes of policy fluctuation, we may find that after further research, new differences emerge, triggered by changes in the views and interests of the governing elites. Similarly, after controlling for horizontal relationships, we may formulate new theories by focusing on differences in causal variables between cases. For example, a variable that may emerge after controlling for horizontal relationships is the career incentives of the executive and legislative members. Although Brazil and Mexico share commonalities in horizontal relationships, executive and legislative career incentives may differ between them, thus explaining why, for example, Brazil does not share revenue immediately or impose hard-budget constraints, as seen in Mexico. These differences in career incentives offer more insight into the reasons why federations often intervene in the fiscal and administrative affairs of state and municipal governments, and why some federations still view decentralization as a means to increase their political power. In turn, recoding regimes based on differences and similarities in career incentives may provide us with a more accurate description of federations and
74
Studies in Comparative International Development / Fall 2003
unitary systems. Recent scholarship suggests, for example, that the career incentives of presidents and political party members to decentralize and/or recentralize reflects politicians' strategies to either remain in office or run for office at the subnational level. These career incentives can be found, for example, within Bolivia's unitary and Brazil's federal systems (O'Neill 1999; Samuels 2003, respectively), thus highlighting similar political strategies within different regime types. In addition, there is ample room for cross-regional recoding based on the careerism variable. My recommendation accords, for example, with Jonathan Rodden's (forthcoming) suggestion that scholars should consider recoding federal and unitary systems based on similarities and differences in career patterns and incentives. In the future, comparing cases based on these variables may provide more theoretical leverage in our explanation of the differences in policy outcomes, notwithstanding similarities in policy design. Focusing on policy fluctuation also allows for strengthening the linkage between data and theory, as well as recoding and reclassifying regimes more accurately. Rodden, for example, suggests using the concept of policy dynamism (synonymous with fluctuation), but he does not provide a clear definition. Rather he suggests only that scholars assign scores to particular dimensions of federalism, such as "formal representation in the legislatures, based on strong upper chambers" (Rodden, forthcoming: 22). But what precisely is policy fluctuation and how can scholars go about assigning values to this variable? While Rodden does an excellent job of addressing the existing gaps between theory and evidence, he does not adequately address these issues, leaving us only with "guides for measuring federalism" and "promising avenues" for future research (Rodden, forthcoming: 21, 23, respectively). In contrast to Rodden's work, this article explains what policy fluctuation is and explores its unique causes, such as the open institutional, sequential, and economic conditions that ultimately lead to the ephemeral, delayed efficacy of decentralization policies. Furthermore, this essay has shown how we can use these causes of policy fluctuation as criteria to compare cases. We can then r e c o d e nations and reclassify them accordingly, having achieved a clearer measure and description of how the decentralization process is unfolding. We can, for example, assign rankings based on the temporal consistency and efficacy of these institutional, sequential, and economic variables. Selecting policy fluctuation as a criterion for comparison also encourages theory building. Fiscal data, for example, constantly changes during sequences of policy fluctuation, thus requiring careful attention when matching data to theory. This rapid change reveals that there is a continuous, interactive dialogue between case studies and empirical evidence and suggests that theories need to be continuously refined and new hypotheses formed due to the ephemeral nature of empirical data (Geddes 1990). Because decentralization outcomes are continuously changing within and between sectors, by comparing the trajectories of policy success and failure over time we can learn more about particular cases and sectors while finding the most crucial and consistent causal variable, such as closed versus open institutions.
Gomez
75
Hence, by comparing cases based on the independent variable of centerstate policy fluctuation, we can develop another useful approach to the subnational comparative method (Snyder 2001). Researchers have paid scant attention to this variable in large part because it is a new empirical phenomenon, emerging only recently in the late 1990s. Nevertheless, future research will need to explore the policy consequences of center-state policy fluctuation, and how this is used as a justification for marshaling the political will and support necessary to implement more permanent and effective decentralization policies. The reemergence of municipal governments as key players in the democratic consolidation and policy-making process should also encourage scholars to consider the theoretical and methodological benefits associated with cross-regional analysis. In this article, I have suggested the comparison of cases based on the following variables: historical state-municipal fiscal relations and autonomy; new institutional innovations in response to constraints in municipal autonomy; and mayoral-legislative relations in the policy-making process. Comparing cases based on these variables may provide several empirical benefits. First, by focusing on variations in autonomy and how state-municipal fiscal relations delay this process, we can begin to unmask the various state government tactics used to continuously suppress municipal autonomy. Moreover, this has the potential to explain why subnational authoritarian withdrawals (that is, the removal of state-level dominant party regimes, previously linked to national-level dominant parties that had made the transition to democracy by agreeing to increase inter-party competition, as in Mexico and India) are delayed at the municipal level and hence why municipalities are not developing democratically (G6mez 2002). Second, institutional innovations can reveal how fiscally constrained municipalities are responding through the creation of new bond markets and greater administrative efficiency. And finally, by focusing on mayoral-legislative relations we begin to unravel the reasons why municipalities vary in their ability to implement and subsequently modernize decentralization policies. The fact that all three of these mechanisms can be found in several developing nations should encourage scholars to consider cross-regional analysis in order to better understand the institutional logic accounting for variations in municipal autonomy and policy performance. In addition to avoiding the problems of inter-state (in this case, inter-municipal) diffusion (Snyder 2001), H comparing municipal performance across nations allows us to more accurately code state-level performance in decentralization policy. Just as Snyder (2001) describes how scaling down to subnational units helps scholars avoid national biases and recode nations more accurately based on differences between national and subnational units, in many cases the same can be said for differences between state and municipal governments. The state of S~o Paulo in Brazil, for example, is always coded as fiscally irresponsible, deluged in public sector deficits and debts (Dillinger and Webb 2001a). Yet one only needs to look at the different levels of economic performance within S~o Paulo's vast municipalities to notice wide variations in performance levels between them and the state capital. Moreover, even
76
Studies in Comparative International Development / Fall 2003
greater variation m a y be f o u n d b e l o w the municipal level in c o m m u n a l and village g o v e r n m e n t s , where some scholars have found the greatest d e m a n d s for fair and free elections, adequate welfare distribution, and hence more eff e c t i v e d e c e n t r a l i z a t i o n p r o g r a m s ( M a n o r 1999). T h u s , in order to p r o v i d e more accurate descriptions o f the decentralization process and to recode states accordingly, scholars should consider comparing municipal, village, and communal g o v e r n m e n t s at the cross-regional level. 12 In conclusion, in addition to analyzing the increasingly complex, multifaceted political process o f decentralization at the federal, state, and municipal levels, recurrent problems in recoding and reclassifying regimes and finding appropriate data for the study o f policy fluctuation's slow m o v i n g dependent variables suggests that the study o f comparative decentralization is qualitative in nature, benefiting from sound qualitative methods and well organized, foc u s e d cross-regional analysis. T h e suggestions p r o v i d e d in this essay about how to go about comparing cases and the various empirical benefits involved suggests that there is ample r o o m - - a n d n e e d - - f o r future studies addressing these issues. As developing nations continue to confront similar institutional challenges and policy outcomes, I can only surmise more of a need for sound methodological guidance and suggestions, ill turn p r o v i d i n g new insight and lessons for international policy-makers and comparative scholars. Indeed, this is virgin territory, and forms part o f what appears to be many new and ongoing efforts to properly s y n c h r o n i z e and measure the intricate dance between decentralization and democracy. Notes
1would like to thank James Mahoney, Richard Snyder, Melanie Cammett, Jonathan Rodden, Kent Eaton, Tyler Dickovick, Steven Webb, Stephan Haggard, Philip Oxhorn, William McCarten, Thomas Bossert, Javier Corrales, and four anonymous reviewers for excellent comments and suggestions. 1. 2. 3. 4. 5.
6.
I define subnational governments as state and municipal governments that are fully elected by the citizenry. Briefly, the first wave refers to the initial process of decentralizing fiscal, administrative, and welfare policies, while the second wave refers to the ex-post consequences of decentralization and the efforts to reform policy and institutions in response to these challenges. On Argentina, see Dillinger and Webb (2001a); on Colombia, see Dillinger and Webb (200t b); and on Hungary, see Wetzel and Papp (2003). As Clay Wescott (forthcoming: 16) explains, the PAR program seeks to: "reform provincial and other subnational administrations, to redefine their relations with one another and the center, and to undertake salary reform for public employees." As they explain, in Brazil the hasty devolution of new fiscal revenues and expendilnre responsibilities in health and education occurred under the 1988 constitution, whereas in Argentina, the devolution of fiscal revenues occurred under the 1988 Co-participation law, forcing the government to transfer automatically to the provinces more than half of the tax revenues it collected; expenditure autonomy was also devolved in 1993 (Eaton and Dickovick, forthcoming: 10). Indeed, the recent adoption of the Fiscal Responsibility Law (May 2001) in Brazil is an effort to reduce the financial and administrative autonomy that governors and mayors have had since the 1988 constitutional reforms. Several states and municipalities initially opposed these measures because it reduced their autonomy and historic access to resources, which provided many political benefits---especially access to borrowing and increased payroll expenditures, both of which are
Gomez
77
limited under the FRL at levels that make it difficult to adjust to crisis and civil servant demands for more pay. Similar initiatives are taking place in India, under the newly implemented Fiscal Responsibility and Budget Management Bill (2003), which plans to reduce the amount of fiscal transfers guaranteed through the National Finance Commission (NFC) and the National Planning Commission (NPC), while imposing new guidelines for state-level fiscal management; the passage of this bill, initially known as the Fiscal Responsibility Act, took three years to pass, illustrating the high degree of subnational resistance to a reduction in revenue and expenditure autonomy initially provided by the NFC and NPC (MeCarten 2003). 7. Comparing cases that are matched on this particular value--historically low levels of municipal fiscal autonomy--is but one of many strategies that can be used as a basis for cross-regional analysis. Alternatively, investigators may compare two different values of this variable in two separate cases, e.g., low versus high levels of municipal fiscal autonomy, in order to see if this variable has an impact on policy autonomy and subsequent responses, such as institutional innovations. Low levels of municipal fiscal autonomy were used in this article in order to provide a suggested example of how we may go about conducting cross-regional analysis. 8. The Congress (I) stands for the "Indira" party (named after Indira Gandhi), which was the first party to govern India immediately after political independence from the British in 1950. Led by Mahatma Gandhi, the Congress (I) was originally known as the "Congress Party." After Indira's sudden death in 1984, however, the Congress officially adopted the "I" in Congress (I). With the exception of only two brief periods in the 1970s and 1980s, the Congress (I) dominated all national and subnational elections, recently losing power, however, to the BJP (the Hindu-nationalist Bharatiya Janata Party) in 1998. The Congress (O), on the other hand, stands for "Organization" and represents the more traditional, hierarchical party organization of the Congress (I); ideologically, the O party is not different from the I party, though it runs as a separate party at the subnational level. 9. See Suresh Krishnamurthy, "India: Ahmedebad Municipal Corporation," Business Line, January 20, t998. 10. Even though mayors shape the policy-making process at the municipal and federal level, it is important to note that they also work with federal legislators to influence policy at the state level, especially when governors call the shots over revenue sharing. Indeed, while not the focus of this study, future research should explore the historical and contemporary conditions increasing mayors" abilities to form coalitions at the federal level, and how these coalitions, perhaps organized under colonial and authoritarian regimes, in turn strengthen the mayors' ability to influence policies at the state level. While the mayors' influence over a large population may increase their bargaining power with the governors, this may not be sufficient; rather, we should look at this in conjunction with mayor-federal legislative ties, and how they are used to influence governors' policies. 11. Snyder's (2001) recommendation of comparing non-contiguous units at the state level can also be applied to the municipal level. This helps us to overcome KKV's warnings against inter-state (and municipal) comparisons due to problems of diffusion via "cross-boarder commuting, residential mobility, and similar socio-economic status" (King, Keohane, and Verba 1994: 222) and to establish independent observations and tests of causal variables within mad between nations (Snyder 2001: 97). 12. Of course, one should not continually scale down to the hundreds of micro-village communities in a state, especially when there is no supportive empirical evidence alluding to the fact that citizens are demanding more services and political accountability. Rather, the unit of analysis should be limited to cases in which there is ample empirical evidence illustrating that the citizens' demands are being voiced. References
Abrucio, Fernando and David Samuels. 1999. "A 'Nova' Politica dos Governadores: Polltica Subnacional e Transicao Democr~itico no Brasil." Lua Nova 40/41:137-66. Bennett, Robert J., ed. 1994. Local Government and Market Decentralization: Experiences in Industrialized, Developing, and Former Eastern Bloc Countries. New York: United Nations University Press.
78
Studies in Comparative International Development I Fall 2003
Burki, Shavid, Guillermo Perry, and William Dillinger. 2000. Beyond the Center: Decentralizing the State9 Washington, D.C.: The World Bank Group9 Corrales, Javier. 2001. "The Political Causes of Argentina's Recession." Paper presented at Woodrow Wilson Center for International Scholars, Washington, D.C., March 1. Dillinger, William and Steven Webb. 2001a. Fiscal Management in Federal Democracies: Argentina and Brazil9 Washington, D.C.: The World Bank Group. 9200lb. Decentralization and Fiscal Management in Colombia. Washington, D.C.: The World Bank Group. Eaton, Kent. 1998. The Politics of Tax Reform: Economic Policy Making in Developing Presidential Democracies9 Ph.D. Dissertation, Department of Political Science, Yale University9 92002. "Fiscal Policy-Making in the Argentine Congress: Party versus Province." Pp. 287314 in Legislative Politics in Latin America, eds. Scott Morgenstern and Benito Nacif. New York: Cambridge University Press. Eaton, Kent and Tyler Dickovick. Forthcoming. "Decentralization and Re-Centralization in Argentina and Brazil." In Decentralization in Comparative Perspective: Lessons from Latin America and South Asia, eds. Eduardo J. G6mez and Paul Smoke. Edmonds-Poll, Emily and Eduardo J. G6mez. 2001. '`The Politics of Decentralization and Municipal Autonomy in Mexico and India9 Paper presented at the Annual Meeting of the American Political Science Association, San Francisco, August 29 - September 2. Escobar-Lemon, Mafia. 2001. "Fiscal Decentralization and Federalism in Latin America." Publius: The Journal of Federalism 31,4 (Fall)9 Echeverri-Gent, John. 1998. "Weak State, Strong Reforms? Globalization, Partisan Competition, and the Paradox of India's Economic Reforms." Paper presented at the Annual Meeting of the American Political Science Association, Boston, September 3-6. Geddes, Barbara. 1990. "How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics9 Pp. 131-50 in Political Analysis, vol. 2, ed. James A. S timson. Ann Arbor: University of Michigan Press. Gilbert, Alan G. and Julio D. D~ivila. 2002. "Bogots Progress within a Hostile Environment9 Pp. 2964 in Capital City Politics in Latin America: Democratization and Empowerment, eds. David Myers and Henry Dietz. Boulder: Lynne Rienner Press. G6mez, Eduardo J. 2000. "A Political Economy of Brazil's Debt Crisis." Paper presented at The Political Economy of Reform in Latin America, David Rockefeller Center for Latin American Studies, Harvard University, November 15. 9 2001. "Presidents, Coalitions, and Subnational Borrowing Controls in Developing Nations." Paper presented at the Fiscal Affairs Department, the International Monetary Fund (IMF), Washington, D.C. . 2 0 0 2 . "Subnational Authoritarian Withdrawals: Method and Evidence." Unpublished manuscript, Brown University. G6mez, Eduardo J. and Paul Smoke, eds. Forthcoming. Decentralization in Comparative Perspective: Lessons from Latin America and South Asia. Grindle, Merilee, ed. 1997. Getting Good Government: Capacity Building in the Public Sector in Developing Countries. Cambridge: Harvard University Press. 92000. Audacious Reforms: Institutional Invention and Democracy in Latin America. Baltimore: Johns Hopkins University Press. Haggard, Stephan and Robert Kaufman. 1995. The Political Economy of Democratic Transitions9 Princeton: Princeton University Press. Haggard, Stephan and Steven Webb. 2003. "Transfers and Incentives: Inter-governmental Fiscal Relations in Latin America," Pp. 235-270 in Decentralization and Democracy in Latin America, eds. David Samuels and Alfred Montero. Notre Dame: University of Notre Dame Press. Hutchcroft, Paul. 2001. "Centralization and Decentralization in Administration and Politics: Assessing Territorial Dimensions of Authority and Power." Governance: An International Journal of Policy and Administration 14, 1 (January): 23-53. King, Gary, Robert Keohane, and Sidney Verba. 1994. Designing Social Inquiry: Scientific Inference in Qualitative Research. Princeton: Princeton University Press9 Kirkow, P. 1996. "Distributive Coalitions, Budgetary Problems, and Fiscal Federalism in Russia." Communist Studies and Economic Transformation 8.
Gomez
79
Litvack, Jennifer, Junaid Abroad, and Richard Bird. 1998. Rethinking Decentralization. Washington, D.C.: The World Bank Group. Mainwaring, Scott and David Samuels. 2003. "Federalism and Constraints on Economic Reform in Brazil." Pp. 85-130 in Federalism and Democracy in Latin America, ed. Edward Gibson. Baltimore: Johns Hopkins University Press. Manor, James. 1999. The Political Economy of Democratic Decentralization. Washington, D.C.: The World Bank Group. Martinez-Vazquez, Jorge and Jameson Boex. 2001. Russia's Transition to a New Federalism. Washington, D.C.: The World Bank. Maurel, Marie-Claude. 1994. "Local Government Reforms and the Viability of Rural Communities in Eastern Central Europe." Pp. 92-112 in Local Government and Market Decentralization: Experiences in Industrialized, Developing, and Former Eastern Bloc Countries, ed. Robert Bennett. New York: United Nations University Press. McCarten, William. 2003. "Soft Budget Constraints in India." Pp. 249-86 in Fiscal Decentralization and the Challenge of Hard Budget Constraints, eds. Jonathan Rodden, Jennifer Litvack, and Gunnar Eskeland. Cambridge: MIT Press. McCarten, William and Vinod Vyasulu. Forthcoming. "Experiments in Indian Democratic Decentralization: The Case of Madhya Pradesh." In Decentralization in Comparative Perspective: Lessons from Lat#~ America and South Asia, eds. Eduardo J. G6mez and Paul Smoke. Montero, Alfred. 2002. Decentralization and Subnational Industrial Policy in Brazil and Spain. University Park: Pennsylvania State University Press. . 2 0 0 0 . "After Decentralization: Patterns of Institutional Change in Argentina, Brazil, Spain, Venezuela, and Mexico." Paper presented at the Congress of the Latin American Studies Association, Miami, March 16-18. Myers, David and Henry Dietz, eds. 2002. Capital Oty Politics in Latin America: Democratization and Empowerment. Boulder: Lynne Rienner Press. Nascimento, Edson. 2002. "Dos Anos de Lei Responsabilidade Fiscal." Unpublished manuscript, Ministry of Finance, Brasilia, Brasil. Nickson, Andrew. 1995. Local Government in Latin America. Boulder: Lynne Rienner Press. North, Douglas. 1989. Institutions, Institutional Change, and Economic Performance. Cambridge: Cambridge University Press. O'Neill, Kathleen. 1999. "Tugging From the Purse Strings: The Politics of Fiscal Decentralization in the Andean Region of Latin America." Paper presented at the Annual Meeting of the American Political Science Association, Atlanta, September 1-5. .2001, "Return to Center? Fiscal Decentralization and Re-centralization in Comparative Perspective." Paper presented at the Annual Meeting of the American Political Science Association, San Francisco, August 29 - September 2. Penfold-Becerra, Michael. 2001. "Institutional Electoral Incentives and Decentralization Outcomes: State Reform in Venezuela." Paper presented at the Congress of the Latin American Studies Association, Washington, D.C., September 6-8. Rasyid, Ryaas. 2002. "The Policy of Decentralization in Indonesia." Paper presented at "Can Decentralization Help Rebuild Indonesia?" Georgia State University, Atlanta, Georgia, May 1. Regulski, Jerzy and W. Kocan. 1994. "From Communism Towards Democracy: Local Government Reform in Poland," Pp. 41-66 in Local Government and Market Decentralization: Experiences in Industrialized, Developing, and Former Eastern Bloc Countries, ed. Robert J. Bennett. New York: United Nations University Press. Remmer, Karen. 1990. "Democracy and Economic Crisis: The Latin American Experience," World Politics 2:315-335. Rodden, Jonathan, Jennifer Litvack, and Gunnar Eskeland, eds. 2003. Fiscal Decentralization and the Challenge of Hard Budget Constraints. Cambridge: MIT Press. Rodden, Jonathan. Forthcoming. "'Comparative Federalism and Decentralization: On Meaning and Measurement." Comparative Politics. Samuels, David and Alfred Montero. 2003. Decentralization and Democracy in Latin America. Notre Dame: University of Notre Dame Press. Samuels, David. 2003. Ambition, Federalism, and Legislative Politics in Brazil. Cambridge: Cambrirge University Press.
00
Studies in Comparative International Development / Fall 2003
Shah, Anwar and Theresa Thompson. 21302. "Implementing Decentralized Local Governance: ATreacherous Road with Potholes, Detours, and Road Closures." Unpublished manuscript, The World Bank Group. Shirk, Susan. 1993.1"he Political Logic of Economic Reform in China. Berkeley: University of California Press. Snyder, Richard. 2001. "Scaling Down: The Subnational Comparative Method," Studies in Comparative International Development 36, 1 (Spring). Stoner-Weiss, Kathryn. 1999. "Central Weakness and Provincial Autonomy: Observations on the Devolution Process in Russia." Post Soviet Affairs 15, 1: 87-106. 92001. "The Regional Sources of Russia's Stalled Reforms." Unpublished manuscript, Princeton University. Sutela, Pekka. 1999. "Fiscal Federalism m Russia." Pp. 151-172 in Economic Institutions, Markets, and Competition: Centralization and Decentralization in the Transformation of Economic Systems, eds. Bruno Dallago and Luigi Mittone. New Hampshh'e: Edward Elgar publishers. Tendler, Judith. 1997. Good Government in the Tropics. Baltimore: Johns Hopkins University Press. Wescott, Clay. Forthcoming. "'DecentralizationPolicy and Practice in Viet Nam: 1991-2001 ." In Decen tralization in Comparative Perspective: Lessonsfrom Latin America and South Asia, eds. Eduardo J. G6mez and Paul Smoke. Wetzel, Deborah and Anita Papp. 2003. "Strengthening Hard Budget Constraints in Hungary." Pp. 393,128 in Fiscal Decentralization and the Challenge of Hard Budget Constraints, eds. Jonathan Rodden, Jennifer Litvack, and Gunnar Eskeland. Cambridge: MIT Press. World Bank.2000a. Decentralization Network, . .2000b. Decentralization "Toolkit, "" . Willis, Eliza, Christopher da C.B. Garman, and Stephan Haggard. 1999. "The Politics of Decentralization in Latin America," Latin American Research Review 34, 1: 7-56. Yang, Dali. 2001. "Economic Transformation and State Rebuilding in China." Paper presented at the Annual Meeting of the American Political Science Association, San Francisco, August 29 September 2.