This art icle was downloaded by: [ T&F I nt ernal Users] , [ Mr Susan Cullen] On: 24 May 2012, At : 11: 47 Publisher: Rout ledge I nform a Lt d Regist ered in England and Wales Regist ered Num ber: 1072954 Regist ered office: Mort im er House, 37- 41 Mort im er St reet , London W1T 3JH, UK Philippine Political Science Journal Publicat ion det ails, including inst ruct ions f or aut hors and subscript ion inf ormat ion: ht t p: / / www. t andf online. com/ loi/ rpsj 20 The ties that do not bind: party affiliations and the delivery of devolved health services in the Philippines Joseph J. Capuno a & Marian Panganiban a a Universit y of t he Philippines, School of Economics, Diliman, Quezon Cit y, Philippines Available online: 24 May 2012 To cite this article: Joseph J. Capuno & Marian Panganiban (2012): The t ies t hat do not bind: part y af f iliat ions and t he delivery of devolved healt h services in t he Philippines, Philippine Polit ical Science Journal, 33: 1, 63-80 To link to this article: ht t p: / / dx. doi. org/ 10. 1080/ 01154451. 2012. 684517 PLEASE SCROLL DOWN FOR ARTI CLE Full t erm s and condit ions of use: ht t p: / / www.t andfonline.com / page/ t erm s- and- condit ions This art icle m ay be used for research, t eaching, and privat e st udy purposes. Any subst ant ial or syst em at ic reproduct ion, redist ribut ion, reselling, loan, sub- licensing, syst em at ic supply, or dist ribut ion in any form t o anyone is expressly forbidden. The publisher does not give any warrant y express or im plied or m ake any represent at ion t hat t he cont ent s will be com plet e or accurat e or up t o dat e. The accuracy of any inst ruct ions, form ulae, and drug doses should be independent ly verified wit h prim ary sources. The publisher shall not be liable for any loss, act ions, claim s, proceedings, dem and, or cost s or dam ages what soever or howsoever caused arising direct ly or indirect ly in connect ion wit h or arising out of t he use of t his m at erial. Philippine Political Science Journal Vol. 33, No. 1, June 2012, 63–80 The ties that do not bind: party affiliations and the delivery of devolved health services in the Philippines Joseph J. Capuno* and Marian Panganiban Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 University of the Philippines, School of Economics, Diliman, Quezon City, Philippines Using a panel of province-level data, we investigate the effects of political party affiliations of local chief executives on the financing and delivery of devolved health services, where arguably the opportunities for and potential gains from inter-local governmental unit (LGU) cooperation abound. Despite these potential gains, the proportion of mayors that belong to the same party as the governor are found not to have any direct, independent and statistically significant effects on the local chief executives’ ability to secure additional resources from the national government and other external sources, mobilize greater spending on local health services or improve select health service outputs. However, the re-election status of mayors and governors is found to have a direct, independent and positive impact on some of these indicators. These results support the view that narrow electoral objectives more than party platforms remain the dominant influence in local fiscal decisions under decentralization. Keywords: party affiliations; local chief executives; local services; Philippines Introduction This article aims to contribute to the empirics of party politics at the local level in the Philippines. Specifically, it investigates whether the party affiliations of municipal and city mayors and provincial governors have effects on the financing, provision and delivery of devolved health services. From the policy standpoint, the research is relevant since health services constituted the bulk of national government functions devolved to local governments in 1991 (World Bank 1993) and health care access and outcomes since then have improved, but only marginally and unevenly across regions (Asian Development Bank [ADB] 2005; Kraft et al. 2011). Further, more potent policies are needed as the government gives itself modest chances of attaining some health targets as part of its Millennium Development Goals by 2015 (National Economic and Development Authority [NEDA] 2010). Finally, the issue bears on the continuing debate regarding the importance to the realization of the promises of decentralization of accountability measures, including political and electoral institutions, in making local governments more responsive (Bardhan 2002; Bardhan and Mookherjee 2006). Instead of political parties, however, more recent studies resonate with that of Lande (1965) in that factions, clans and classes still dominate the local political economy in the Philippines (e.g., Gutierrez, Torrente, and Narca 1992; Lacaba 1995; Rivera 1999; Sidel 1999; De Dios 2007; Solon, Fabella, and Capuno 2009). While providing important *Corresponding author. Email: jjcapuno@up.edu.ph ISSN 0115-4451 print/ISSN 2165-025X online q 2012 Philippine Political Science Association (PPSA) http://dx.doi.org/10.1080/01154451.2012.684517 http://www.tandfonline.com Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 64 J.J. Capuno and M. Panganiban insights, most of these studies use an inductive approach based on selected cases. Generalizing from these case studies is difficult since members of local factions or political clans also join political parties in the country.1 The effects of clan membership or party affiliation on, say, local fiscal decisions can be easily mistaken for the other since political parties and clans often espouse populist programs. Also, some younger members of political clans have shown development orientation while in office, perhaps as much as can be expected from dedicated members of progressive political parties. Therefore, a more statistical analysis of party affiliations may add the needed grist to the mill. It is also widely accepted that political parties in the Philippines are weak and unlike those found in developed countries (e.g., Banlaoi and Carlos 1996; Montinola 1999; Teehankee 2002; Hutchcroft and Rocamora 2003; Co and Tigno 2005). Providing more objective evidence, Kasuya (2009) further finds that party switching remains a regular feature of elections in the country and more intensely so since 1987. While previous analyses have focused more on the shifts in party alignments of candidates for national positions (e.g., presidents and members of congress), some have noted as well that candidates for local positions align themselves with the parties of presidential candidates (or presidentiables).2 Without loyalty, members are made to toe the party line more through the use of carrots (e.g., promises of pork barrels or campaign finances) than sticks (e.g., expulsion or censure). Unfortunately, the party leader, usually the incumbent president, can only distribute pork barrel funds while in office. Worse, even before his or her term expires, the promise of pork barrel funds by presidentiables is often more attractive to potential turncoats. Notwithstanding the weaknesses of the country’s party system, it may be asked what elected party mates do in between elections and whether it is easier for them to collaborate to address a common concern. In the next two sections, we present a priori reasons for cooperation in the provision of devolved health services among local government units (LGUs). Briefly, with the fragmentation of local health systems, much of the gain from economies of scale, control of epidemics or health information sharing is lost and can only be recovered with interLGU cooperation. While most local health personnel are competent and dedicated civil servants, and are continually guided and trained by the Department of Health, they are unlikely to solve health problems without their local chief executives’ support. Where devolved health services are improved, there is usually a supportive mayor or governor and innovative health staff working together (Pineda 1998; Quimpo 1996). This is because mayors and governors, like the Philippine President, have the power of the purse. He or she can determine the timing or amounts of cash allocations to all departments including the local health office, which does not generate its own budget or keep whatever revenues it can raise. Moreover, the local chief executives (LCEs) have powers to hire, fire and deploy personnel. According to the Civil Service Commission (2005), LGUs in 2004 had a total of 370,227 employees, of which 268,740 were career officials and 101,487 were non-career officials. Of the latter, 20,514 were elective officials, 8085 were occupying positions coterminous with those of the appointing officials and 6448 were contractual workers. These figures do not yet include personnel on job-order status (i.e., on short-term engagements) who are paid using the LGUs’ budget for maintenance, operating and other expenses (MOOE).3 In the fourth section, we discuss the dataset and the econometric model used to tease out the effects of party affiliations. The dataset consists of a panel of province-level factoids for 2003 and 2008. We chose these non-election years to avoid confounding the effects of party affiliations with extraneous, election-related factors. While it is true that most incumbents are always preparing for their next term or office, we assume that party Philippine Political Science Journal 65 mates are less concerned about campaigning in between election years and therefore may coordinate their provisions of devolved health services. We present and discuss the empirical results concerning the effects of party affiliations in the fifth section. The final section concludes the article. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 Legal and economic framework This section reviews the legal and economic frameworks for inter-LGU cooperation mediated through party affiliations or by some other means. The legal framework effectively reduces some of the possible legal or administrative impediments to cooperation, while the economic framework identifies the potential gains from the coordinated or consolidated provision of some local public services. The point to be made is that local governments can collaborate if they want to and could realize benefits if they do so. Providing the legal framework is the Local Government Code of 1991 (Republic Act 7160, 10 October 1991), specifically its Section 33: Section 33. Cooperative Undertakings among Local Government Units. Local government units may, through appropriate ordinances, group themselves, consolidate, or coordinate their efforts, services, and resources for purposes commonly beneficial to them. In support of such undertakings, the local government units involved may, upon approval by the sanggunian concerned after a public hearing conducted for the purpose, contribute funds, real estate, equipment, and other kinds of property and appoint or assign personnel under such terms and conditions as may be agreed upon by the participating local units through Memoranda of Agreement. Consistent with this provision, some LGUs banded together to address their common problems or pursue shared goals, such as in industry (e.g., Metro Iloilo-Guimaras Development Council), urbanization (e.g., Metro Cebu Development Council) and health (Inter-Local Health Zones). These inter-LGU collaborations range from simple coordination in the planning and delivery of public services to formal cooperation involving the joint use and financing of public goods, and to de jure consolidation of devolved services under the administration or supervision of the province or a single authority created for the purpose. Besides political parties, there are in practice other formal and informal structures and procedures that can facilitate, if not enforce, collaborations among local governments. Some of the existing administrative structures would be the regional development councils, the provincial development councils, and even the provincial chapters of the League of Municipalities or League of Cities. The informal structures would include political families or clans whose members occupy several local or national positions, elective or otherwise. Arguably, both formal and informal linkages can foster communication, knowledge sharing or peer-to-peer mentoring among local chief executives and other officials. Some of these cooperative arrangements can even effectively enforce joint decisions through peer pressure, moral suasion or some form of threat or promise from national government officials. While the Code reduces the transaction costs, where do the potential benefits of interLGU cooperation come from? Here, an economic analysis of the country’s decentralized fiscal set-up could help identify the sources. Under devolution, many of the assigned functions (especially in health) do not match the political jurisdiction, fiscal capacity or institutional competence of the recipient LGUs (Manasan 2004; Lieberman, Capuno, and Minh 2005). The inefficiencies arising from these mismatches can be improved by amending the Local Government Code to reassign the devolved functions or through interLGU cooperation. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 66 J.J. Capuno and M. Panganiban To show the benefits of cooperation, consider Figure 1 (adapted from Oates 1972 and Wellisch 2000), which illustrates the economic advantages and disadvantages of a decentralized provision of certain public services. The vertical axis measures the benefits and costs of devolved public services. The horizontal axis measures the degree of decentralization from the point of origin (which indicates centralized provision). The benefits of decentralization come from the provision of public services customized to the specific needs of local populations. Examples of such customer-specific services are office hours or outreach programs that accommodate the schedule of working mothers or schoolchildren, or health services that respect cultural sensitivities or beliefs. Presumably, these idiosyncratic factors are more known to local officials who are also more directly accountable to the service clients. In contrast, a central authority is likely to provide a “generic” service to suit a typical or average user. However, decentralization can also lead to higher social costs of service provision. The reason is that typically each LGU will want its own facility (such as a hospital) and personnel regardless of the actual number of facility users. Thus, there will be duplication or gaps in services provision across areas, and even excess capacity in some places. In contrast, a central agency (or a less decentralized set-up) can spread the costs to more service clients (i.e., economies of scale).4 Depicted as D* in Figure 1, the ideal level of decentralization then is one that maximizes the net benefits (i.e., the biggest difference in the total benefits and total costs associated with decentralized provision). The problem of course is that D* may not correspond to any level of local government, i.e., it may be too big or too small to be contained in one political jurisdiction. In such cases, the cooperation of two or more adjacent LGUs is warranted. Presumably, such cooperation is more likely or easier among incumbent LCEs that belong to the same political parties since they would also benefit from exploiting the potential gains from a consolidated provision of local services. To be sure, many inter-LGU collaborations have been documented since 1991 (see Gonzales 2004; Galing Pook Foundation 2009). For example, the provincial governments of Negros Oriental and Negros Occidental in the early 1990s jointly operated barangay health facilities for their constituents living in the mountains that separate the two provinces. Since the term of former President Joseph Estrada, the Department of Health has encouraged all LGUs around a district or provincial hospital to form Inter-Local Health Zones to facilitate their resource sharing, coordinate local health plans and to ensure an effective hospital referral system (Capuno 2008). Total costs, benefits from customized services Total costs Total benefits from customized services D* Level of decentralization Figure 1. Total benefits from customized services and total costs of decentralized provision. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 Philippine Political Science Journal 67 The case for LGU collaboration in health This section further discusses some features of local health systems in the Philippines that motivate, if not demand, LGU collaboration. These health features provide the opportunities or justifications for cooperating local governments, or LCEs who belong to the same party, to improve their health financing, service provision or outputs. However, they would require a greater level of LGU cooperation than their usual compliance to existing procedures, such as the LCE participation in provincial development councils or their submission of their local government’s annual investment plans to the provincial government. If there is strength in numbers, then a coalition of LGUs may receive more attention or support from the central authority. In health, the national government support would include co-sponsoring the social health insurance coverage of indigent families and extraordinary grants or transfers (both cash and in kind). These grants or transfers are special and not part of the LGUs’ internal revenue allotments (IRA), which constitute their regular, mandated share in the national internal revenues. Under the National Health Insurance Act of 1995, LGUs are mandated to identify and enroll the indigent families among their constituents with the Philippine Health Insurance Corporation (PhilHealth) that runs the country’s social health insurance program. The sponsoring LGU shares with the national government the premium contribution for each indigent family based on the LGU’s income class and years of participation in the program. To boost the coverage of poor families under PhilHealth’s Sponsored Program, the national government in 1994 aimed to enroll about five million families under the socalled Plan 5 million (Plan 5M). This was continued as OPlan 2.5 million (OPlan 2.5M) in the following years, but with only half the original target of indigent families for coverage. Under these initiatives, the national government paid for the entire annual premium contribution (amounting to 1200 pesos per beneficiary family) while the local officials identified the eligible families and delivered to them their PhilHealth membership cards. Besides the national government, provincial governments can also share with their component cities or municipalities in the premium payments.5 Given the variations in PhilHealth coverage of the poor across provinces, it may be asked if those with higher enrolments are also those with a high proportion of LCEs who are party mates. Besides securing greater external support, cooperating LGUs can also align their health spending by integrating their health plans, jointly procuring drugs or medical supplies, or resource sharing. An integrated health plan would be ideal to control epidemics, hazardous wastes or the spread of communicable diseases across municipal boundaries. Bulk purchasing can also lead to lower unit costs since participating LGUs would be able to bargain harder and the winning supplier can afford to give discounts due to savings on delivery or transaction costs. Cooperating LGUs can also share ambulances, expensive medical equipment and even critical health personnel. Thus, to the extent that LCEs who are party mates also collaborate, then party affiliation will have an impact on local health spending. Further improvements in certain health service outputs are only possible with LGU collaboration. For example, the control of epidemics would require the coordinated monitoring of cases among LGUs. To this extent, the health statistics that local health personnel regularly collect are most useful when shared with other LGUs. When reinforced by all LGUs, health education, information dissemination and advocacy activities also become more effective. To contain the spread of pathogens or disease vectors, immunization campaigns must be undertaken by LGUs at the same time. The Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 68 J.J. Capuno and M. Panganiban efficient use of hospital services calls for a referral system where simple cases are handled at the lowest facility levels (usually rural health units) and only the more complicated cases are referred to district or provincial hospitals. This will require LGU cooperation since the rural health units are devolved to municipalities, while the hospitals are devolved to provinces. Without a functioning referral system, higher-level facilities will be overcrowded, while lower-level facilities will be under-utilized. The variations in health service outputs again provide the opportunity to tease out the possible effects of party affiliations. While the efficient level of decentralization (or consolidation of devolved services), depicted as D* in Figure 1, may correspond to a configuration of LGUs smaller than a typical province but bigger than a typical municipality or city, we use the province in our empirical analysis for two reasons. Firstly, the province in many places is the highest level of local government and has administrative jurisdiction over most lower-level LGUs. Hence, the province as a political-administrative unit provides a natural or default grouping of local political leaders. Secondly, the province as a geographical unit is often large enough to exploit whatever possible economies of scale or synergy can be had from inter-LGU cooperation and yet still close enough and directly accountable to the service clients to provide for their specific needs. In addition, the assignment of some fiscal functions builds on the hierarchical structure of LGUs. In health, for example, the lowest level of facilities that provide primary care services are devolved to municipalities and cities. The provision of secondary or tertiary health care services is assigned to provinces. However, the three levels of health care are complementary and must be administered as integral parts of a functioning provincial health system. Arguably, provincial health systems work better where the governors and mayors work closely together. Data and methodology Data To tease out the effects of party affiliations, we assembled a province-level dataset obtained from the Bureau of Local Government Finance, the Department of Health, the Commission on Elections, the Department of the Interior and Local Government, PhilHealth and the National Statistical Coordination Board. Specifically, our dataset constitutes a panel of 73 provinces for the years 2003 and 2008. The list of provinces excludes those in the Autonomous Region of Muslim Mindanao because of the region’s unique administrative structure, and three provinces with missing data for 2003 (Batanes, Zamboanga Sibugay and Dinagat Islands). Table 1 shows the descriptive statistics of the regression variables used. The key variable of interest is the proportion of incumbent mayors that belong to the same political party as the incumbent governor. Party affiliation is constructed from the reported local political parties of the elected governors and mayors in the 2001 and 2007 local elections. In cases where the elected candidate listed his or her membership in a local political and a national political party, we used the local party to construct the party affiliation variable. In cases when a governor is listed as independent, the variable party affiliation is assigned a value of zero. Around 38% and 36% of the elected mayors in 2003 and 2008, respectively, had the same party affiliation as the elected governors. The effect of party affiliation on the devolved health services is measured using several dependent variables. As proxy measures of national support (other than the IRA), we use the proportions of indigent families sponsored by the national government in PhilHealth’s Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 Table 1. Summary statistics. 2003 Variable Min. Max. Mean Standard Deviation Min. Max. 0.034 0.084 0.00 0.68 0.88 0.72 0.00 6.25 0.00 0.00 0.00 0.00 0.078 0.16 0.00 1.10 0.15 0.45 0.00 3.89 37.88 157.77 0.00 1223.24 163.35 113.62 135.90 50.19 0.00 38.41 910.68 296.16 138.90 367.00 161.98 1602.46 3.38 1.00 979.02 12805.93 0.081 0.086 0.00 0.469 0.027 0.172 0.00 1.53 0.080 0.812 0.002 0.469 0.178 0.239 0.00 1.514 0.30 0.83 0.18 0.12 0.00 0.39 0.76 1.18 0.15 0.80 0.06 0.12 0.05 0.45 0.46 1.24 0.38 0.47 0.47 744.56 119.24 0.27 0.19 0.50 729.99 254.59 0.00 0.00 0.00 188.33 5.89 1.00 1.00 1.00 6020.68 1791.33 0.36 0.33 0.64 784.36 87.54 0.25 0.19 0.48 1017.49 372.18 0.00 0.00 0.00 24.62 0.00 1.00 1.00 1.00 5655.28 2851.51 251.23 31.547 0.42 2.51 0.869 0 872.21 13.196 0.59 1.36 2.393 0 1959.00 64.6 4.52 8.47 20.339 0 284.86 22.960 0.36 2.88 0.604 1 0 3.40 0.00 0.00 0 0 924.89 19.367 0.35 3.20 0.820 0 0.00 0 0.01 0.07 0.070 1 5578.60 63 2.80 21.41 5.484 1 69 Explanatory variables Proportion of mayors that belong to same party as governor Proportion of mayors re-elected Governor is re-elected Internal revenue allocation per capita, provincial goverment Internal revenue allocation per capita, component cities and municipalities Internal revenue allocation per capita, independent cities Poverty incidence Doctors per 10,000 population Midwives per 10,000 population Nurses per 10,000 population Year 2008 Standard Deviation Philippine Political Science Journal Dependent variables Proportion of indigent families sponsored by national government in PhilHealth Proportion of indigent families sponsored through Plan 5M or OPlan 2.5M in PhilHealth Total extraordinary receipts or aid received by province and component cities and municipalities, per capita Health spending per capita of provincial government Health spending per capita of component cities and municipalities Proportion of indigent families sponsored by local government in PhilHealth Proportion of indigent families sponsored by local government in PhilHealth, less those sponsored by others Proportion of facility-based deliveries Proportion of fully immunized under-5 children Mean 2008 Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 70 J.J. Capuno and M. Panganiban Sponsored Program, or through its Plan 5M/OPlan 2.5M, and the total extraordinary receipts or aids (i.e., not the internal revenue allocation) received by the province and its component cities and municipalities from the national government. In 2003 and 2008, the national government on average enrolled 3.4% and 8.8% of the indigent population in each province. In 2008, Plan 5M/OPlan 2.5M on average enrolled around 7.8% of the indigent families per province. The value for this variable is zero in 2003 since these initiatives were started only in 2004. The average extraordinary receipts amounted to around 38 pesos per capita in 2008, up by a significant amount from less than a peso in 2003. To measure the coalition’s ability to coordinate their fiscal spending, we use the health spending per capita (in 2000 prices) of the provincial government and of the component cities and municipalities of the province. In 2008, the mean provincial health spending per capita was around 139 pesos, down by about 23 pesos from 2003. In contrast, health spending per capita of the component LGUs was 367 pesos in 2008, about three times more than they were spending on average in 2003. Since the provincial government on the one hand and the component cities and municipalities on the other hand are supposed to provide complementary health services, their health spending should not have negative effects on the other, more so when local political leaders coordinate their fiscal decisions. Finally, we assess the party’s impact against several health service delivery outputs. The first set of output indicators pertains to the proportion of indigent families that the province and its component municipalities and cities themselves enrolled in PhilHealth’s Sponsored Program. The corresponding mean proportions are around 8% and 2.7% in 2003 and 2008. The lower proportion in 2008 is due to the increased sponsorship by the national government under Plan 5M/OPlan 2.5M. However, when we net out these national government-enrolled indigent families, the average proportions of LGUsponsored indigent families are now 8% in 2003 and 17.8% in 2008. The other set of output variable includes the proportion of facility-based deliveries and the proportion of children aged below 5 years old who are fully immunized. In 2003 and 2008, the average proportions of maternal deliveries that were performed in public or private hospitals and clinics are 30% and 15%, respectively. The average proportion of under-5 children who have completed immunization against polio, diphtheria, tetanus and other childhood diseases was about 80% in both 2003 and 2008. Like the other sets of indicators mentioned above, these maternal and child health indicators are presumably higher in places where the LCEs are united and can thus share resources or personnel, other things being constant. We also included other variables to account for the political incentives for the incumbent LCEs to perform, and other province-level indicators of service needs and of fiscal or service delivery capacity. To account for electoral incentives, we used the proportion of mayors that were re-elected in 2003 (47%) and 2007 (33%) and the re-election status of governors in the same years. About 47% and 64% in 2003 and 2008, respectively, were re-elected governors. Presumably, the re-elected LCEs already have some experience running their local governments and some may still look forward to another term. The local population’s need for local services is captured with poverty incidence, which on average was 31.54% in 2003 and 22.96% in 2008. Fiscal capacity is indicated by the IRA per capita of the province and of the component cities and municipalities. For the provincial government, the average IRA per capita (in 2000 prices) was 745 pesos in 2003 and 784 in 2008. For the component cities and municipalities, the corresponding figures were about 119 pesos and 88 pesos. To account for possible spending spillovers from Philippine Political Science Journal 71 independent cities within some provinces, we also include the IRA per capita (also in 2000 prices) of these cities, which amounted to about 251 pesos in 2003 and 285 pesos in 2008. Finally, the numbers of doctors, nurses and midwives per 10,000 population are included as measures of service delivery capacity. In both 2003 and 2008, on average there was less than one doctor, one nurse and around two midwives per 10,000 population. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 Estimating equation Using the province-level dataset, we estimated the following model using the panel-data estimation technique (see Greene 2008, 183). We assume that the dependent variable Y, which indicates a specific health impact (financing, spending, output) is linked to party affiliation and a host of other possible explanatory variables as follows: Y it ¼ ai þ bPit þ dX it þ uT t þ 1it ; ð1Þ where the subscripts i and t refer to the ith province and year t, respectively, P is an indicator of party affiliation, X is a vector of province-level characteristics, T is the dummy variable for year, e is the error term, a is a constant term, and u, b and the vector d are regression coefficients. An estimate of b would capture the marginal (or own independent) effect of party affiliation on Y, holding other factors constant. Using stata’s xtreg module, Equation (1) is estimated as a panel data model with province-level fixed effects. Specifically, the term ai in Equation (1) captures the effects of the omitted or unobserved, province-specific characteristics such as topography, culture and physical capital or infrastructures. The coefficient ai also captures the possible effects of political clans, which is found to be a critical factor in previous studies but is not included in our list of explanatory variables because of data unavailability. While our specification allows us to tease out the independent effects of party affiliations, it is not possible to tease out the effects of political clans from those other unobserved factors captured by ai. Further, the model allows for year fixed effects or those shared characteristics of the provinces in a given year but which change through time, such as the prevailing national political situation, macroeconomic or global economic conditions. Analysis of results Effects on securing national government support Table 2 presents the estimated effects of party affiliation on securing some national government support for health. The regression results show that the proportion of mayors that belong to the same party as the governor has no statistically significant effect in terms of securing a higher proportion of their poor constituents extended insurance coverage by the national government or through the Plan 5M/OPlan 2.5M programs (in 2008 only). It also does not appear to help generate additional funds or aid from the national government and other external sources. Likewise, the proportion of re-elected mayors or re-elected governors does not show any statistically significant effects. This means that re-elected LCEs are not necessarily favored over first-term LCEs. Each peso of IRA per capita received by the province or the component cities and municipalities increases by around 20 –25 centavos the extraordinary receipts or aid that the LGUs additionally get. These results suggest that extraordinary receipts are allocated Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 72 Table 2. Regression results on the effect of party affiliation on national government support. Independent variables Proportion of mayors that belong to same party as governor Proportion of mayors re-elected IRA per capita, province IRA per capita, component city and municipality IRA per capita, independent cities Poverty incidence Doctors per 10,000 population Midwives per 10,000 population Nurses per 10,000 population Proportion of indigent families covered by national government Proportion of indigent families covered by national government (inc. Plan 5M/OPlan 2.5M in 2008) Total extraordinary receipts/aid, per capita 2 0.056 (0.455) 0.096 (0.371) 0.023 (0.565) 2 0.00004 (0.631) 0.00001 (0.835) 0.000001 (0.825) 0.0005 (0.693) 0.151 (0.537) 0.005 (0.830) 2 0.096 (0.443) 20.025 (0.432) 0.010 (0.799) 0.017 (0.284) 20.00007 (0.337) 20.000009 (0.745) 0.000 (0.886) 20.0002 (0.897) 0.061 (0.651) 0.0187 (0.254) 20.001 (0.977) 20.034** (0.021) 0.017 (0.826) 2 0.106 (0.103) 0.105 (0.200) 0.017 (0.578) 2 0.00017 (0.476) 2 0.000009 (0.856) 0.00001 (0.444) 2 0.0036 (0.271) 2 0.186 (0.492) 0.021 (0.514) 0.708 (0.546) 0.536** (0.068) 0.017 (0.269) 16.316 (0.546) 20.553 (0.550) 12.221 (0.360) 0.251** (0.000) 0.214*** (0.000) 2 0.003 (0.664) 0.111 (0.936) 9.151 (0.936) 2 42.730** (0.003) 2 2.308 (0.963) 50.343*** (0.000) 2 118.272* (0.064) Year 2008 Constant 0.042 (0.594) Number of groups R-squared Prob . F 73 0.0587 0.95 Note: Figures in parentheses are p-values. ***p , 0.01, **p , 0.05, *p , 0.10. 73 0.171 0.360 73 0.143 0.534 73 0.812 0.000 J.J. Capuno and M. Panganiban Governor is re-elected Proportion of indigent families sponsored through Plan 5M or OPlan 2.5M (in 2008 only) Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 Philippine Political Science Journal 73 based on population density or equity consideration (as are used in the IRA formula), or transferred as matching grants tied to the uses of the IRA funds. The other statistically significant factors are the ratio of midwives to 10,000 population and the year dummy variable for 2008. The first of these factors shows that it is associated with lower extraordinary receipts or aids (247.73 pesos), which implies that this external assistance goes more to LGUs with low existing health human resource capacity. Finally, national government insurance coverage of the indigent population, coverage under Plan 5M/OPlan 2.5M and the amount of extraordinary receipts are all higher in 2008 than in 2003. The model is able to explain 5% of the variation across provinces in the proportion of indigent families sponsored through Plan 5M/OPlan 2.5M (in 2008 only); 17% of the variation in the proportion of indigent families sponsored by the national government; 14% of the variation in the proportion of indigent families enrolled by the national government and under Plan 5M/OPlan 2.5M; and 81.2% of the variation in total receipts and aids received by the province and component LGUs per capita. The F-test for the regression on proportion of indigent families sponsored through Plan 5M/OPlan 2.5M and total extraordinary receipts or aid per capita tells us that we can reject the null hypothesis that all coefficients of independent variables are equal to zero at the 10% level. However, we are unable to reject this hypothesis for the regression on the proportion of indigent families sponsored by the national government (including or not Plan 5M/OPlan 2.5M), which means that the econometric model as a whole does not very well explain the patterns in these dependent variables. Effects on local spending If LCEs that are affiliated to one party are unable to obtain additional support from the national government, are they at least able to coordinate and increase their health spending? The regression results in Table 3 show that this is not the case. The proportion of mayors that belong to the same party as the governor has no statistically significant effect on the level of health spending of either the provincial government or the component cities and municipalities. Moreover, the re-elected mayors and governors also do not appear to be associated with a higher level of health spending. The only statistically significant variables pertain to the available fiscal resources, health service capacity and the dummy variable for the year 2008. Specifically, a peso increase in the provinces’ IRA or in that of the component cities and municipalities tends to increase the latter’s health spending by about two pesos. However, the effect is not symmetric. A peso increase in the IRA of the component LGUs does not have any statistically significant effect on the province’s own health spending. Interestingly, the IRA of independent cities has a positive, statistically significant although small impact (around 20 centavos) on provincial health spending. This suggests possible spillovers to neighboring jurisdictions, such as residents of the independent cities crossing political boundaries to make use of provincial hospitals or rural health units of neighboring towns. Again, the ratio of midwives to 10,000 population and the dummy variable for the year 2008 are both statistically significant. In particular, a unit increase in the midwives-to10,000-population ratio leads to a 47-peso rise in provincial health spending and a 441peso reduction in the health spending of the component LGUs. Since most midwives are employed in rural health units and city health offices, the results suggest that provinces could be spending on drugs and medical supplies that complement the services of the midwives, while cities and municipalities may be constrained to spend more on the same things the more health personnel they already employ. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 74 Table 3. Regression results on the effect of party affiliation on local health spending. Independent variables Proportion of mayors that belong to same party as governor Proportion of mayors re-elected Governor is re-elected IRA per capita, component city and municipality IRA per capita, independent cities Poverty incidence Doctors per 10,000 population Midwives per 10,000 population Nurses per 10,000 population Year 2008 Constant Number of groups R-squared Prob . F Note: Figures in parentheses are p-values. ***p , 0.01, **p , 0.05, *p , 0.10. Component city and municipality health spending, per capita 2 20.180 (0.411) 32.257 (0.303) 2 3.295 (0.785) 2 0.079 (0.139) 2 0.034 (0.138) 0.019** (0.012) 2 1.542 (0.221) 2 119.620 (0.248) 47.100*** (0.000) 2 18.427 (0.680) 2 34.841*** (0.002) 185.554*** (0.002) 204.844 (0.428) 178.981 (0.585) 95.752 (0.447) 2.584*** (0.000) 2.060*** (0.000) 2 0.048 (0.534) 0.735 (0.955) 141.914 (0.896) 2441.238*** (0.001) 2 8.664 (0.985) 373.738*** (0.002) 21116.828* (0.067) 73 0.589 0.000 73 0.824 0.000 J.J. Capuno and M. Panganiban IRA per capita, province Provincial health spending, per capita Philippine Political Science Journal 75 Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 The R-squared statistics indicate that econometric models are able to account for 59% of the variation in provincial health spending per capita and 82% of the variation in health spending per capita. Further, the F-test for the regressions on health spending per capita indicates the joint significance of the coefficients of the independent variables.6 Effects on service delivery outputs Finally, Table 4 shows the estimated impact of party affiliation, with the expectation that LCEs in the same parties will work together to improve selected service delivery outputs, especially those that affect the health of mothers and children. As in the previous tables, however, the proportion of mayors that belong to the same party as the governor still has no statistically significant effects on the proportion of indigent families extended insurance coverage by the LGU, facility-based deliveries or fully immunized under-5 children. These results are consistent with those reported in Table 3 since health spending should translate to service outputs. However, they also suggest that party affiliations per se do not help improve the administration (or the technique) of health services delivery to make up for lower spending. Interestingly, the underlying incentives and competence of the LCEs matter. Recall the assumption that re-elected LCEs already have some experience with actual public administration. Indeed, the higher the proportion of mayors re-elected, the greater the proportion of indigent families that the LGUs enroll in PhilHealth (between 22 and 24 percentage points) and of fully immunized under-5 children (8 percentage points). On the other hand, re-elected governors are associated with a higher proportion (6.6 percentage points) of facility-based deliveries. The component cities and municipalities with high IRA per capita also tend to extend insurance coverage to a higher proportion of their poor families. The proportion of indigent families sponsored by their own LGUs (and not by the national government or other sponsors) also tends to be higher in poor areas. A lower proportion of facility-based deliveries is associated with higher midwives-to-10,000-population ratio, presumably because midwives can deliver babies at home. Finally, all four output indicators are higher in 2008 than in 2003, indicating possible shifts in the overall economy or national government policies that affected all provinces. The models explain around 48% of the variation in the proportion of indigent families sponsored by the local government and in the proportion of indigent families sponsored by the local government (net of those sponsored in other programs). The model’s explanatory power is slightly higher at 52% with facility-based deliveries but is only 26% with proportion of children fully immunized. The F-test for all the regressions on service delivery outcomes allows us to reject the null hypothesis that all coefficients of independent variables are equal to zero and therefore statistically significant. Conclusion In sum, our empirical results confirm that party affiliations do not matter to the provision of devolved health services. More precisely, the proportion of incumbent mayors that belong to the same party as the incumbent governor has no direct, independent and statistically discernible effect on their ability to secure additional support for health from the national government, on increasing health spending of the provincial government or of the component cities and municipalities, or on improving selected health service delivery outputs such as health insurance coverage of the poor, maternal deliveries in health Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 76 Table 4. Regression results on the effect of party affiliation on selected service delivery outputs. Independent variables Proportion of mayors that belong to same party as governor Proportion of mayors re-elected IRA per capita, province IRA per capita, component city and municipality IRA per capita, independent cities Poverty incidence Doctors per 10,000 population Midwives per 10,000 population Nurses per 10,000 population Year 2008 Constant Number of groups R-squared Prob . F Proportion of indigent families covered by local government, less those sponsored in other programs Proportion of facility-based deliveries Proportion of fully immunized under5 children 20.118 (0.119) 0.241** (0.014) 0.003 (0.936) 20.0001 (0.114) 0.0004*** (0.012) 0.00004 (0.063) 20.008** (0.038) 0.388 (0.222) 20.018 (0.625) 20.102 (0.455) 0.150*** (0.000) 0.064 (0.717) 20.103 (0.151) 0.218** (0.018) 0.00008 (0.998) 0.0002 (0.100) 20.0001*** (0.009) 20.007* (0.069) 0.377** (0.045) 20.018 (0.209) 20.018 (0.615) 20.112 (0.385) 0.138*** (0.000) 0.057 (0.729) 0.028 (0.626) 0.047 (0.528) 0.065** (0.027) 2 0.00005 (0.093) 0.00006 (0.093) 2 0.00001 (0.353) 0.003 (0.289) 0.201 (0.414) 2 0.029*** (0.324) 2 0.136 (0.205) 0.143*** (0.000) 0.164 (0.231) 0.057 (0.159) 0.081* (0.078) 2 0.013 (0.466) 2 0.00001 (0.852) 0.00001 (0.722) 2 0.00007 (0.944) 0.0008 (0.658) 0.029 (0.849) 0.003 (0.861) 2 0.087 (0.186) 2 0.033** (0.046) 0.827*** (0.000) 73 0.475 0.000 Note: Figures in parentheses are p-values. ***p , 0.01, **p , 0.05, *p , 0.10. 73 0.480 0.000 73 0.524 0.000 73 0.262 0.047 J.J. Capuno and M. Panganiban Governor is re-elected Proportion of indigent families covered by local government Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 Philippine Political Science Journal 77 facilities and basic immunization of infants and toddlers. Given that we specifically chose the devolved health services where the potential gains from inter-LGU collaboration are clear and realizable, the results only confirm (objectively so in this case) that party affiliations in the Philippines count for less than they should. In contrast, there is some evidence that the re-election status of mayors and governors has a direct, independent and positive impact on the same health service delivery outputs. Perhaps the re-election status reflects the incumbent LCEs’ prior experience in local administration and their desire to perform well to secure another. The two results suggest that unless political parties are strengthened to make them responsive and accountable to voters, candidates will continue to change alliances to win elections and their decisions once elected will be motivated more by narrow electoral objectives than party platforms. Reform advocates contend that through a strong party system, successful candidates will espouse more long-term, strategic objectives (e.g., Montinola 1999). The internal revenue allocations per capita, midwives-to-10,000-population ratio and, in some cases, poverty incidence are the factors that consistently show some effects. These suggest fiscal resources, existing health personnel and the health needs of the local population are stronger determinants than party affiliations. Moreover, external assistance, local health spending and service outputs are also better in 2008 than in 2003, which perhaps reflects more favorable government policy or macroeconomy in the latter year. It is important to note, however, that while party affiliations do not have an effect on the devolved health service, the results do not mean that party affiliations may not have an impact on the financing or delivery of other public services. It is possible as well that party affiliations matter more for local governments in the same congressional district. If elected, representatives align themselves to the incumbent president or to presidentiables to secure pork barrel allocations. They can then use their allocations to secure the political support of the mayors within their districts. This topic and the following ones can be pursued in future studies toward a more comprehensive assessment of political parties in the context of decentralization. Aside from helping each other during elections, party affiliates who are elected may also work together to secure other forms of national government support, such as the establishment of special economic zones or free port zones or the construction of airports or seaports in their jurisdictions. In addition, they could be pursuing sub-provincial goals involving only a few of their party mates so that the gains from cooperation are greater and the enforcement costs are lower. This could explain, for example, why Metro Naga Development Council and the Banate Bay Coastal Resource Management Council, Inc. appear to be successful. A province-wide voluntary collaboration of LGUs runs into a collective action problem, i.e., some will inevitably free-ride on the efforts of others, which then eventually discourages cooperation. However, where party affiliations seem to work, they do so perhaps only because most members belong to the same political clans or dynasties. Hence, it is important to disentangle the effect of party membership (which is presumably more ideology or platform driven) from that of clan membership (which caters more to narrow or special interests) by having these two indicators in one regression model. While we effectively controlled for clan presence here using a fixed-effects model, we cannot tease out its own effects from other unobserved factors to compare it with that of party affiliation. It would also be worthwhile to investigate where party members draw the line between their party’s positions in regard to their clan interests, and why political clans persist despite opportunities to build or strengthen existing political parties (see De Dios 2007 and Querubin 2010 for some interesting insights). To push the analysis further, future research 78 J.J. Capuno and M. Panganiban may compare the effectiveness of the de jure or de facto rules for candidate selection, division of the “spoils,” and the enforcement of discipline among members of political parties and clans. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 Acknowledgements The authors gratefully acknowledge the two anonymous referees for comments and suggestions on an earlier version of the article, the participants in the 11th National Convention on Statistics held on 4 – 5 October 2010 at the EDSA Shangri-La Hotel, Ortigas Center, Mandaluyong City, and those in the 12th International Conference of the East Asian Economic Association held on 2 – 3 October 2011 at Ewha Women’s University Seoul, Korea. This study is a spin off from another report by the authors submitted to the UPecon-Health Policy Development Program. The contents of this study however are the sole responsibility of the authors and do not necessarily reflect the views of USAID, the United States Government or UPecon Foundation, Inc. Notes 1. 2. 3. 4. 5. 6. Gutierrez, Torrente, and Narca (1992) observe that “political parties in the Philippines are coalitions of political clans.” Coronel et al. (2007) list party and family alignments in the 12th House (2001– 2004) which include the Dys in Isabela, the Floirendos in Davao del Norte, and the Zubiris in Bukidnon for Lakas; the Cojuangcos and Teodoros in Negros Occidental, Pangasinan and Tarlac for the National People’s Coalition; the Angaras in Aurora, the Remullas in Cavite, and Lobregats and Zamboanga for the Laban ng Demokratikong Pilipino; and the Abads in Batanes for the Liberal Party. After the 2010 elections, 24 congressmen left Lakas-Kampi-CMD, the party of the previous president, Gloria Macapagal-Arroyo, and were sworn into the Liberal Party, the party of Benigno Aquino III who was to be inaugurated a few days later. More representatives and local chief executives would follow. One of those who joined the Liberal Party said that they would be “supporting the present administration but it does not mean all the way in time for the 2013 elections” (see Diaz and Padua 2010; Managbanag 2011). Charging against the MOOE allows the LGUs to get around the cap on expenditures on personnel services specified in the Local Government Code. Surgeons in health facilities are an example of trade-offs between customized service provision under a fragmented set-up and exploiting economies of scale in a more centralized configuration. There are no regular surgeons serving in barangay health stations because their capacities are under-utilized in this facility. Visits from patients who need medical attention from a surgeon are not very common. The services of surgeons are most needed in district hospitals where they can serve a broader group of patients who require their attention. However, having surgeons serve in district hospitals or other higher-level facilities implies a loss in service specificity for patients who go to barangay health stations who may immediately need the services of a surgeon. Such instances may be rare, but the unavailability of the kind of service at the specific time and place needed by a patient constitutes a loss. See Lavado (2010) for a lengthier discussion on this. A similar analysis of the shares of health in total spending (not shown) yields F-test statistics that do not allow us to reject the null hypothesis that the coefficients are simultaneously equal to zero. However, we maintain that the levels are adequate for the purposes of the study. Notes on contributors Joseph J. Capuno is an Associate Professor of Economics in the University of the Philippines School of Economics, where he is also developing a graduate-level course on local public economics. His original research interest in the devolution of health services in the Philippines has led him to further work in the economics and politics of decentralization. He has published articles in both economics and political science journals and edited books on local fiscal decisions and political clans, innovations in local services, governance and local development, and local government performance ratings and civic participation. He earned his PhD from the School of Economics of the University of the Philippines. Philippine Political Science Journal 79 Marian Panganiban is a BS and MA graduate of the University of the Philippines School of Economics. Her recent work on health policy has been featured in The Lancet, the Global Health Research Forum’s Young Voices in Research for Health, and in Asian Politics and Policy. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 References Asian Development Bank. 2005. Poverty in the Philippines: Income, assets and access. Manila: Asian Development Bank. Banlaoi, R.C., and C. Carlos. 1996. Political parties in the Philippines: From 1900 to present. Makati City, Philippines: Konrad Adenauer Foundation. Bardhan, P. 2002. Decentralization of governance and development. Journal of Economic Perspectives 16, no. 4: 185– 205. Bardhan, P., and D. Mookherjee. 2006. Decentralization and local governance in developing countries: A comparative perspective. Cambridge, MA: The MIT Press. Capuno, J. 2008. A case study of the decentralization of health and education services in the Philippines. Report submitted to the Philippine Human Development Network. UP Campus, Diliman, Quezon City. Civil Service Commission. 2005. 2004 inventory of government personnel. http://excell.csc.gov.ph/ cscweb/2004IGP_stat.pdf (accessed February 15, 2012). Co, E., and J. Tigno. 2005. Philippine democracy assessment: Free and fair elections and the democratic role of political parties. Quezon City: Ateneo de Manila University. Coronel, S., Y. Chua, L. Rimban, and B. Cruz. 2007. The rulemakers: How the wealthy and the wellborn dominate congress. Quezon City: Philippine Center for Investigative Journalism. De Dios, E. 2007. Local politics and local economy. In The dynamics of regional development: The Philippines in East Asia, ed. A.M. Balisacan and H. Hill, 157– 203. Quezon City: Ateneo de Manila University Press. Diaz, J., and R. Padua. 2010. 24 House members join Noy’s Liberal Party. The Philippine Star, June 26. Galing Pook Foundation. 2009. Kalusugan muna! Policy forum on health. http://www.galingpook. org/main/images/stories/news/HealthForumMay09Documentation.pdf (accessed September 12, 2010). Gonzales, G. 2004. Metro Cebu: A metropolitan area in need of coordinative body. PIDS Discussion Paper 2004– 49, Philippine Institute of Development Studies, Makati, Philippines. Greene, W.H. 2008. Econometric analysis. 6th ed. Upper Saddle River, NJ: Pearson Prentice Hall. Gutierrez, E., I.C. Torrente, and N.G. Narca. 1992. All in the family: A study of elites and power relations in the Philippines. Quezon City: Institute of the Popular Democracy. Hutchcroft, P., and J. Rocamora. 2003. Strong demands and weak institutions: The origins and evolution of the democratic deficit in the Philippines. Journal of East Asian Studies 3: 259– 92. Kasuya, Y. 2009. Presidential bandwagon: Parties and party systems in the Philippines. Pasig City: Anvil Publishing. Kraft, A., P. Mariano, S. Kault, E. Jimenez-Soto, and K.-H. Nguyen. 2011. Philippines equity report: Investment case for financing equitable progress towards MDGs 4 and 5 in the Asia-Pacific region. UP School of Economics Discussion Paper 2011–15, University of the Philippines School of Economics, Diliman, Quezon City, Philippines. Lacaba, J.F., ed. 1995. Boss: 5 case studies of local politics in the Philippines. Pasig, Metro Manila: Philippine Center for Investigative Journalism. Lande, C. 1965. Leaders, factions and parties: The structure of Philippine politics. Monograph Series 6, Yale University Southeast Asia Program, New Haven. Lavado, R. 2010. Is PhilHealth’s sponsored program reaching the poorest of the poor? The Filipino Child Policy Brief 2, UNICEF-PIDS, Philippines. http://www.unicef.org/philippines/ brief02_fnl.pdf. Lieberman, S.S., J.J. Capuno, and H.V. Minh. 2005. Decentralizing health: Lessons from Indonesia, the Philippines and Vietnam. In East Asia decentralizes: Making local government work, ed. R. White and P. Smoke, 155– 78. Washington, DC: The World Bank. Managbanag, N. 2011. Misamis Oriental execs join Aquino’s party. Sun Star Cagayan de Oro, September 6. Manasan, R. 2004. Local public finance in the Philippines: In search of autonomy with accountability. PIDS Discussion Paper Series 2004– 42, Philippine Institute for Development Studies, Makati City, Philippines. Downloaded by [T&F Internal Users], [Mr Susan Cullen] at 11:47 24 May 2012 80 J.J. Capuno and M. Panganiban Montinola, G. 1999. Parties and accountability in the Philippines. Journal of Democracy 10, no. 1: 126– 140. National Economic and Development Authority. 2010. Philippines progress report on the millennium development goals 2010. Pasig City, Philippines: National Economic and Development Authority. Oates, W. 1972. Fiscal federalism. New York: Harcourt Brace Jovanovich. Pineda, V.S. 1998. Health management: Strategies from selected cities. PIDS Discussion Paper 98-36, Philippine Institute for Development Studies, Makati City, Philippines. Querubin, P. 2010. Family and politics: Dynastic persistence in the Philippines. Mimeo, Massachusetts Institute of Technology. http://econ-www.mit.edu/files/6365 (accessed September 20, 2011). Quimpo, B.A. 1996. Devolution matters: A documentation of post-devolution experiences in the delivery of health services. Manila: Department of Health, Local Government Assistance and Monitoring Service, Manila. Rivera, T. 1999. How fare the clans? Poverty and elite reproduction in the Philippines. Paper presented at the annual conference of the Philippine Political Science Association, July 23 – 24, in UP Campus, Diliman, Quezon City. Sidel, J. 1999. Capital, coercion and crime: Bossism in the Philippines. Stanford, CA: Stanford University Press. Solon, O.J., R.V. Fabella, and J.J. Capuno. 2009. Is local development good politics? Local development expenditures and the re-election of governors in the Philippines in the 1990s. Asian Journal of Political Science 17, no. 3: 265– 84. Teehankee, J. 2002. Electoral politics in the Philippines. In Electoral politics in Southeast and East Asia, ed. A. Croissant, G. Bruns, and M. John, 149– 202. Singapore: Friedrich-Ebert-Stiftung, Office for Regional Co-operation in Southeast Asia. Wellisch, D. 2000. Theory of public finance in a federal state. Cambridge, UK: Cambridge University Press. World Bank. 1993. Philippines devolution and health services: Managing risks and opportunities. Population and Human Resource Operation Division, East Asia and Pacific Region Office. Report No. 12343-PH. Washington, DC: The World Bank.