The Local Governance Performance Index (LGPI)1 What is the Local Governance Performance Index (LGPI)? The Local Governance Performance Index (LGPI) provides a new approach to the measurement, analysis and improvement of local governances. The LGPI is a tool that aims to help countries collect, assess, and benchmark detailed information around issues of local and public sector performance and service delivery to citizens and businesses. It is also a methodology, using heavily clustered surveys to uncover important local-level variation in governance and service provision. This information aids policymakers and development specialists in designing specific action plans, provides an initial benchmark from which to measure of progress, and empowers citizens’ and businesses’ voices to influence government efforts on improving quality and access of public service delivery. The tool uses household surveys that gather micro-level data from multiple communities, including data on experience, perception and satisfaction regarding cross-cutting governance issues. Specifically, citizen household surveys that include batteries for health, education, security, voice and participation, and other metrics of governance have been developed as the core of the LGPI instrument. The specificity of the information and its focus on experiences rather than simply perceptions permit an in-depth assessment of institutional quality and capacity, providing a detailed map of institutional weaknesses and strengths from the citizens’ perspective. Such information can be combined with data gathered from civil servants and other providers, to provide a full picture of governance. Such information helps identify specific priority areas for reform, facilitating the policy design and policy implementation improvements at the country and local level (Recanatini 2011). 1 The LGPI summary document was prepared by Ellen Lust, University of Gothenburg; Pierre Landry, New York University – Shanghai; Dhafer Malouche, University of Carthage and Jumana Alaref, World Bank. The LGPI modules described here are part of the comprehensive Local Governance Performance Index that has been developed by a team comprised of Ellen Lust, Pierre Landry, Lindsay Benstead, of Portland State University, and Dhafer Malouche. As part of the activities of the Program on Governance and Local Development (GLD) at Yale University, directed by Ellen Lust and financed by the Moulay Hicham Foundation, the Carnegie Foundation, and Yale University, the team designed the instrument, refined it and implemented a pilot survey in Tunisia. This presentation of the LGPI modules has been developed as a product of the Public Service Delivery Global Solutions Group. Gregory Kisunko, Kimberly Johns, Asmeen Khan, and Francesca Racanatini (World Bank) and Jairo Acuna-Alfaro provided valuable comments. Hana Brixi has led the task at the World Bank. 1 What are the Objectives of the LGPI? By assessing the performance of public administration at the local level, the LGPI provides critical feedback to help government officials, political parties, civil society actors, the public and the international development community pinpoint specific geographical areas and substantive areas where policy reform is needed. The performance of public administration is multi-dimensional, including the quality of service provision, equal access to citizens, and rationalization of procedures in obtaining services. We conceptualize performance from a citizen’s perspective, with an eye on the availability and quality of services as they are actually delivered to the respondents’ respective communities. By providing a mechanism for monitoring and evaluating local governance, based on citizens’ experiences and perceptions, the LGPI can support governments in addressing social, political, and economic challenges facing their country or region. In addition, the data collected provides a basis for effective outreach and dissemination of results to local and central governments, citizens, media, and outreach agencies. By making such information available, the LGPI encourages evidence-based advocacy and policymaking, and also strengthens citizen-state engagement. Indeed, employment of the LGPI can help strengthen local skills and foster a culture of evidence-based reform and implementation. The LGPI implementation in conjunction with local partners builds their skills in data collection, analysis and dissemination. It can also promote policy reform and application that is built upon transparent data collection procedures, replicable analysis and open, evidence-based policy debate. What Tools Informed the LGPI? The LGPI draws upon and extends a tradition of measuring governance and service delivery. Like the Governance and Anti-Corruption (GAC) diagnostic studies and Service Delivery Indicators (SDI) implemented at the World Bank, the Public Administration Performance Index (PAPI) implemented in Viet Nam by the UNDP, and Citizen Scorecards, it aims to provide organized and specific evidence that allows citizens, policymakers and development specialists to assess and address existing problems in governance and service delivery. The LGPI tool builds on the Governance and Anti-Corruption (GAC) country diagnostic surveys that have been implemented and assessed worldwide. GAC surveys have been used as a diagnostic tool that first aimed at unbundling corruption and governance challenges existing at various levels (administrative, state capture, bidding, theft of public 2 resources, etc.) and at identifying weak and strong institutions to determine the source of the problem. The approach further aimed at assessing the costs of corruption and poor governance to different stakeholders, with the objective of identifying concrete and measurable ways to reduce those costs through targeted reforms. GAC surveys have been adapted to sector-specific governance assessments and implemented in fifteen countries in South Asia, Africa, and Latin America and the Caribbean regions, at the request of and in collaboration with multiple local stakeholders. (See the Governance and Anti- Corruption Diagnostics website for more information) The Public Administration Performance Index (PAPI), which was first implemented in Viet Nam, also inspired development of the LGPI. PAPI used experiences with such issues as corruption, participation, transparency and service delivery to develop governance indices aggregated at the provincial level. The index has proven remarkably useful in Viet Nam. Results have spurred provincial officials to request assistance in developing action plans that help improve performance and researchers have observed a marked improvement in public service provision over time. PAPI aids citizens by giving them information to hold officials accountable, providing the information necessary to target reforms and creating incentives for government agencies to improve performance. (See the PAPI website for more information.) Why Focus on the Local Level? The LGPI is based on the premise that local governance matters, and that the drivers that explain local level variation may differ from those that operate at higher levels. The tool aims to uncover this, and thus differs from extant tools by providing information that is representative at the appropriate local level. Even in highly centralized systems, the nature and quality of governance and service provision varies significantly at the local level. Some communities and local leaders find ways to overcome resource deficits, assure transparency and accountability, and provide better services than other communities do. Moreover, decentralization efforts are aimed at extending local rights and responsibilities. Finally, we recognize that communities do not perform equally well in all areas of governance and services provision; for instance, they may provide adequate health services while their schools suffer, while the converse may be true elsewhere. Extant tools that examine governance and service provision tend to be representative at the national level, and are thus unable to detect variation at the subnational level and across different agencies. Simply put, they suffer from aggregation bias. Aggregate outcomes at the governorate or country level often fail to detect significant differences at the local or community level. Such aggregation also leads to erroneous conclusions, 3 making inferences about local level processes by examining outcomes at these higher levels—the ecological fallacy. Instead, we need a tool that can measure the variation both within and across local communities if we are to pinpoint needs, determine drivers of positive outcomes, design relevant programming, and assess outcomes. The LGPI is designed to provide such information. What are the Other Key Characteristics of the LGPI? The LGPI has a number of other key features: First, the LGPI provides a comprehensive picture of governance and service delivery across sectors and criteria, giving clearer insight to their interconnectivity. For instance, it is well-recognized that health and security can affect school performance; unhealthy children or those unable to travel safely to and from school are likely perform more poorly than those who enjoy quality health care and public safety. Examining citizens’ experiences across sectors provides a better understanding of these linkages and the complex root causes of poor performance. Thus, the LGPI provides a tool for programming based on service delivery sectors that both advances and complements information collected with the use of sector-specific tools (e.g., SDI, citizen scorecards). Second, the LGPI emphasizes citizens’ experiences over their satisfaction. Citizen satisfaction is affected by two factors: citizens’ expectations and their experiences. The LGPI focuses on citizens’ experiences, thus providing a better understanding of how access to services, service quality and governance differs across localities and demographic groups. Some GAC tools focus on experience as well; thus, it is not unique in this regard. However, this is an important difference from citizen scorecards, which generally emphasize satisfaction. Third, the LGPI is designed to be scientifically implemented and to include experiences of both users and non-users of services. It is thus distinguished from citizen scorecards, SDI and many GAC tools, which focus on experiences of users through such instruments as student and parent surveys, or patient exit interviews. The experience of users is important, but it is equally important to understand the experiences and decisions of those who turn elsewhere for services. Fourth, the LGPI provides a better understanding of how citizens participate in service delivery and governance. Despite the current concern with participatory planning and implementation, the existing tools tend to limit our understanding of citizen engagement. The SDI, for example, focuses on the role of parents in parent-teacher organizations and 4 contact with the school, but it does little to inform us about the ways parents seek help when their children face problems, to whom they turn to for assistance, and the extent to which they are successful. By asking the problems that citizens often face with regard to education, health, security and other issues, and how they seek to solve these problems, the LGPI draws a better picture of citizen participation and provides insights into entry points for improving participation and accountability. Fifth, the LGPI allows us to examine the role of state and non-state institutions and actors in service delivery and governance. Often, citizens turn to non-state actors to help access public services, enhance community security, and engage in activities that improve community governance and services. Communities also differ significantly in their social composition, including the existence of formal organizations (e.g., trade unions, CSOs, political parties) and the extent and nature of social norms and institutions. It is thus important to look beyond formal institutional arrangements and actors to understand when communities successfully govern themselves and manage to meet their needs. What Are the Main Modules of the LGPI? The LGPI is divided into 6 main modules: 1) Education, 2) Health, 3) Physical Security and Dispute Resolution, 4) Social Assistance and Welfare, 5) Citizen-State Linkages and Corruption, and 6) Social Composition and Culture. Each module, provided in Annex 1, includes batteries of questions that are designed to tap into key issues. These can be inserted into existing surveys, using the additional batteries on demographics and other issues in the analysis. However, the survey is also designed to allow us to combine responses to questions across the batteries in order to form indices that tap into governance dimensions (e.g. transparency, participation, and accountability). Using the full complement of modules, combined with demographic questions, is an optimal because it provides a comprehensive view of local level governance and service provision. Education. The education module allows us to tap into school attendance and drop out rates, education quality, and school governance. For each child under the age of 17, parents are asked if the child attends school and the nature of the school attended (e.g., public vs. private, religious vs. non-religious). Questions target the quality of infrastructure, (e.g., well-built schools, clean bathrooms), service delivery process (e.g. crowded classrooms, teacher absenteeism, favoritism), and governance (e.g., use of bribes, informal payments, existence of parent-teacher organizations). It also includes batteries of questions on how parents seek assistance for school fees and to solve 5 problems with their child’s education, allowing us to determine the extent to which they turn to state actors or instead seek assistance from non-state actors or informal channels. Health. The health module measures self-reported physical and mental health and the ability to access health care (e.g., unmet medical needs, insurance). It also focuses on respondents’ experience with doctors and local clinics, both public and private. It provides assessments of the quality of the health infrastructure, (e.g., clean bathrooms), service delivery process (e.g. patients treated with respect, presence of doctor, posted fees, doctor presence) and governance (e.g., use of bribes, informal payments, referrals to private clinics), satisfaction (e.g., with cost, infrastructure, and treatment). It asks whom respondents turn to for help with medical problems (e.g., pharmacist, local healer, doctor, clinic) and inquires about problems in accessing or financing care. Finally, it taps into expectations about local medical facilities, including doctor presence at the local public clinic and care at various medical providers (e.g., local hospital, private clinic, public clinic). Physical Security and Dispute Resolution. The module on physical security and dispute resolution gauges individuals’ experience with, and perception of, crime and violence. It examines personal experiences with crime and civil disputes and, for each victim, determines whom they turned to for help and their level of satisfaction with the outcome. The module also includes questions that gauge citizens’ experience and perceptions of the police with regard to fairness, responsiveness, honesty, knowledge of the community and effectiveness. Finally, it includes perceptions of security in a variety of places (e.g., at home, in the neighborhood, at the local market) and awareness of crime (e.g., burglary, physical assault, smuggling). Social Assistance and Welfare. The module begins by assessing the utilities and basic public services (e.g., garbage collection, water provision, electricity). It probes personal experiences with food, shelter and clothing shortages, as well as unemployment, asking whom they turn to for help if they experienced problems. It also measures citizens’ experience and perceptions of the availability and fair distribution of welfare assistance, considering the extent to which they believe those in need are served, the need for personal connections (wasta) and informal payments in obtaining assistance, and the relative willingness of various state and non-state actors to provide assistance. Citizen-State Linkages and Corruption. The module on citizen-state linkages is designed to assess governance dimensions, including transparency, participation at the national and local level, and accountability. It comprises experiential questions that examine an individual’s participation in elections, engagement in meetings with local council members or others to discuss needs, and experience with corruption—especially if the 6 person has been victimized by corrupt institutions. This module also includes perception questions regarding the willingness of public officials to address corruption complaints and the relative influence of citizens and other stakeholders. Social Composition and Culture. The social composition and culture module allows us to aggregate information on the types of groups and organizations that are influential in local governance. It also examines the nature of social ties and the informal rules of engagement, particularly focused on norms of obligation and the use of wasta. Wasta is defined as the use of personal connections, usually based on friendship or blood relations, to obtain resources or positions or, alternatively, as the person who performs this mediating function. It can be understood as a form of corruption (e.g., nepotism) but also reflects cultural expectations. Examining these questions, in conjunction with experiential questions in other batteries that examine which actors individuals turn to in times of need, allows us to compare the relative efficacy of formal and informal institutions. What Key Questions Can Be Addressed by the LGPI? The LGPI allows us to examine a range of questions about governance and service delivery. By asking about both citizens’ experiences with and their perceptions of service quality, corruption and other issues, we are able to determine the extent of gaps between perception and revealed behavior. By examining a range of issues (e.g., education, health, dispute resolution, welfare), we can identify the localities that excel in some dimension while falling short in others, and explore how outcomes co-vary across dimensions. With the LGPI, it is also possible to detect sources of inequality. Variation across localities highlights geographic disparities, allowing the considerable differences in the quality of education or health to be better understood, as demonstrated in the results of the pilot study reported in Annex 2. The data also allow for the detection of unequal access to services, levels of participation and other outcomes related to gender, socio- economic status, and age. Finally, it can help us to see differences in the quality of public or private facilities, both experienced and perceived. How Can LGPI Modules be Adapted for Country-Specific Programming? The LGPI modules can, and should, be adapted to country-specific considerations in two ways. First, questions within the modules must be adapted to fit the context—use appropriate terminology, incorporate all possible outcomes and remove inappropriate 7 items. Second, specific issues of interest can be incorporated into the modules. For instance, it may be useful to introduce questions that gauge citizen knowledge of (or engagement with) specific programs. Finally, it may be necessary to eliminate questions that appear to be wholly irrelevant or even culturally offensive in certain country contexts. The questionnaire should be carefully reviewed and revised before implementation. This should be done in collaboration with relevant officials, service providers, civil society activists, other stakeholders and citizens. Such reviews can take place through focus groups to examine key issues and governance problems, collecting annotated comments on the instrument from relevant stakeholders or holding workshops to review the questionnaire. Vignettes can also be used to help calibrate responses across countries. In each case, implementing careful pre-testing can also provide an important check on the relevance and applicability of different batteries. Amendments are possible and often necessary, but they should also be made sparingly. The LGPI can be used both to detect subnational variation and, eventually, to make comparisons across countries. Both comparisons can be useful for government officials, civil society leaders, and citizens. All stakeholders should thus be keenly aware of the fact that even small changes in question wording can make comparisons difficult. They should adapt the module by making necessary changes, but resist making less critical ones. How Can the LGPI be Supplemented or Extended by Other Tools? The LGPI tool, which is a citizen-based approach, can be fielded in conjunction with other instruments. Doing so permits a multi-pronged approach to programming and analysis that can allow for systematic and rigorous measurements of governance and service delivery. It also allows for triangulation of responses to enhance the analysis of existing problems and measurement of programming outcomes. Sector-specific facility-level instruments may be particularly useful in this regard. The Service Delivery Indicators (SDI), developed by the World Bank and adopted mainly in Africa, provide one potential option whereby the quality and performance of health and education services are assessed from users’ and providers’ perspectives against three key indicators: (i) providers’ efforts; (ii) providers’ abilities; and (iii) availability of resources, by administering facility surveys. 8 Additionally, facility-level surveys developed in the health sector as part of the World Bank MDTF in governance and service delivery engagement in Jordan may prove complementary with the LGPI. The MDTF instruments extend beyond the SDI surveys as they examine linkages between service delivery outcomes (as measured by provider effort) and governance and accountability mechanisms at the local levels, both internal (in-facility monitoring and incentive systems) and external (citizen engagement). The MDTF instruments include (i) key informant interviews with directors of health directorates; (ii) Chief Medical Officers’ (CMOs) interviews; (iii) provider surveys; (iv) patient exit interviews; and (v) surveys with chairs of community health committees at the local level. The MDTF instruments have been developed and are to be fielded early August 2015. In addition to facility-level surveys, the LGPI could be further supplemented with targeted municipal and public official surveys. These surveys would further probe local governance challenges, identified in the citizen household survey, that are possibly associated with citizens’ trust and satisfaction levels with services and service delivery outcomes. In specific, such surveys could reveal crucial information on the structure of incentives and actual degree of enforcement within public agencies, the existence and use of accountability mechanisms, the quality of rules and procedures, the degree of transparency of budget and employment decisions, and the overall quality of management. In this regard, a core governance module on municipalities has been developed and pre-tested in Tunisia (funded by the program on Governance and Local Development at Yale University) and may be fielded in conjunction with the LGPI (with a complementary sampling strategy to be identified). The Local Governance Institutions Comparative Assessment (LoGICA) framework is another complementary tool to consider as it analyzes local and intergovernmental systems, processes, and institutions that contribute to effective local governance and service delivery performance, by examining five main domains: (i) the assignment of functional responsibilities; (ii) the existence of local political space and effective local political systems; (iii) the degree of local control over the administration of local services; (iv) local fiscal autonomy and local financial management; and (v) participatory and responsive local service delivery mechanisms. While the LoGICA framework overlaps with the LGPI tool in that the latter touches on some of the mentioned elements, if both tools are utilized together in a targeted manner their distinct features and insights may prove complementary. For example, the LGPI adopts sector-specific modules with respect to local governance and service delivery from the perspective of citizens’ experiences, while the LoGICA framework provides an assessment of the overall local system governing service delivery. In doing so, the LoGICA employs a couple of instruments: (1) LoGICA survey that captures information about the organizational 9 governance structure in the public sector; (2) LoGICA score card that examines the five aforementioned domains on the basis of 40 developed assessment individual indicators for each domain2, which in turn are aggregated into an index; and (3) LoGICA report that in turn provides a narrative assessment of the local governance system based on the indicator-led analysis stemming from the scorecard. Finally, the LGPI may benefit greatly from the Governance and Anti-Corruption (GAC) diagnostic tools that also use experience-based (vs. ‘opinions’) and employ multi- pronged surveys of users of public services/households, business people and public officials, which as previously mentioned permits triangulation of the responses. The GAC diagnostic tools have been adapted to specific sector (health, education, transport etc.) and thematic areas (gender, human rights, security, violence, etc.) depending on policy interest. The approach has been piloted and fielded in over fifteen countries in Latin America, Africa, and South Asia regions. The GAC greatly complements the design of the LGPI, particularly with respect to its emphasis on participatory approach in data collection, analysis, and dissemination to enhance capacity and ownership, in addition to obtaining an initial benchmark of governance and public sector performance at the local level, allowing for monitoring progress on a regular basis, and institutionalizing the tool in projects and country operations. It my offer several valuable complementary tools as an addition the LGPI household survey for consideration, such as public official surveys and business surveys. What Operations can Benefit from the LGPI? The LGPI tool is an essential tool in generating evidence-based, local level data that could be incorporated in upcoming and future World Bank operations to further inform their designs, provide an initial benchmark for measuring progress, revise specific action plans, and involve stakeholders in meaningful dialogue around reform options. The LGPI can be implemented at various levels (e.g., the municipality, village, or neighborhood level). Lower level analyses provide greater insight into how community dynamics impact governance and outcomes, but they increase the costs of implementation. Nevertheless, the results of the Tunisia LGPI pilot reveal considerable variation at the municipal level within governorates, highlighting the importance of examining lower level units. As previously mentioned, the tool is focused on six modules that measure local governance and service delivery performance. We detail the ways in which the LGPI tool could be useful with respect to each program in the following discussion: 2 The 40 assessment indicators are focused on assessing the basic performance of the local governance system based on conceptual norms of effective localization and good local governance as well as existing good international practices. The scores of the individual indicators for each of the five dimensions are aggregated into an overall score that forms an index ranging from zero points to a maximum of 50 points. 10 Participation and Accountability at the Local Level. The LGPI can provide citizens with information about the quality of governance and service provision at the local level, allowing them to monitor providers and officials and to voice their concerns. The tool also measures users’ behavior and modes of engagement with state and non-state actors when seeking public services. It helps gauge state responsiveness, citizen satisfaction and trust, and highlights accountability gaps between citizens and providers. In this regard, household surveys gather data on the experience that citizens have with inappropriate procedures and behavior, the presence of accessible accountability mechanisms (if any) and any informal and non-state mechanisms they rely on for receiving services. The LGPI tool measures participation at the local level through four main sub- dimensions: (1) civic knowledge, which examines citizens’ knowledge of their electoral rights and awareness of institutions that safeguard political and social participation at the grass-roots level; (2) opportunities for service delivery participation, which assesses citizens’ opportunities to participate in the delivery process of public services in various sectors through exercising their voice and choice rights; (3) quality of elections, including citizen ability to voice demands and hold policymakers accountable; and (4) voluntary contributions, which investigates ‘demand-side accountability’ in citizens’ willingness and capacity to participate in planning and to oversee and contribute to local development projects. Control of Corruption and Quality of Institutions. The LGPI provides a diagnostic tool for assessing the extent of corruption experienced by service users. By pinpointing experiences with corruption at the local level and identifying variations across geographic regions, service-provision sectors, or demographic groups, the LGPI helps measure the scope of country corruption. This approach also allows for the identification of weak and strong local institutions, and in doing so, deepens our understanding of the link between institutional factors and corruption. In addition, it allows for the assessment of the costs of corruption to different stakeholders, by looking at the impact of poor governance and corruption on their trust levels in state institutions and consequently, their modes of engagement with the state. Finally, the approach allows for the identification of concrete and measurable ways to reduce those costs through targeted reforms. The LGPI measures corruption at the local level by focusing on four main aspects: (1) pervasiveness of corruption in the public sector within local governments; (2) pervasiveness of corruption in public service delivery; (3) willingness to fight corruption; and (4) presence and effectiveness of anti-corruption and monitoring mechanisms. Vertical Accountability. The assessment of voice mechanisms and participation at local and national levels provides evidence on the strength of vertical accountability. It helps to 11 determine in which sectors and over what issues vertical accountability mechanisms are more or less developed, and provides a diagnostic tool for further reform. This dimension is measured through the inclusion of questions that gauge the levels of citizens’ interactions with local authorities and participation in politics, highlight perceptions on degree of effectiveness of such interactions, and underline factors that may limit governments’ and public officials’ accountability to their citizens in the operationalization of governance functions. Decentralization and Local Governance. Countries in MENA and elsewhere are considering the merits of moving towards decentralized governance and granting more autonomy and decision-making power to local authorities. In this regard, the LGPI could be a somewhat useful tool that generates a quantifiable baseline of the current status, which can be used to assess the new approach and measure progress, highlight any problems and bottlenecks with implementation. In this respect, the tool utilizes questions that probe, from citizens’ experiences, issues around financial autonomy, government responsiveness, and the adequacy and availability of key inputs at the local level in different sectors. Service Delivery. The LGPI produces novel and independent information on health services, quality of education services, levels of personal security, trust and use of local state institutions. It also provides insights into possible correlates between governance and service delivery by looking at incentive structures of providers and identifying weak and strong institutions at the local level to determine the source of the problem. In many cases, countries seek to identify concrete and measurable ways for reforms in specific sectors. In this case, the LGPI, as a cross cutting governance tool, can be adapted to address sector-specific challenges (for example, health and education). Additionally, the tool sheds some light on possible sub-national variation in service delivery performance that may allow for further testing of the relationship between governance at the local level and service delivery outcomes. Incentives for Service Delivery and Policy Implementation. The LGPI can provide a bottom-up and unique perspective on the grassroots incentives that providers and policymakers face. Combining the modules, and particularly the module on social composition and culture with that on citizen state linkages, allows us to understand how citizens view the obligations and responsibilities of various stakeholders, the extent to which they use connections to obtain services, and whom they turn to for help, and whom they view as responsible. By combining the LGPI with the GAC or other facility- and municipal-level studies that gauge the perspectives of providers and civil servants, we can trace the incentive structure and weigh their relative impact. 12 How Can Results be Disseminated to Enhance the Impact of the LGPI? The LGPI tool provides a basis for effective outreach and dissemination of results. As previously noted, the tool allows for a participatory approach in the design, validation, and implementation phase as well as data collection and analysis. As evidenced by previous GAC diagnostic surveys, “if the assessment process is not owned by a broad range of stakeholders, it will not be sustained over time nor will its findings feed back into the policymaking process” (Recanatini 2011). With the aim of empowering governments to develop their own reform strategies based on a more objective and micro- level data, it is also equally important that the analysis and results for the LGPI are disseminated jointly with local stakeholders and policy makers to ensure local ownership and promote long-term, sustainable partnerships between civil society and the government. Joint collaboration with local partners in disseminating findings to stakeholders and the public may include partnering with civil society organizations (CSOs) in the development of websites and outreach strategies to media to promote awareness and use of the data. Workshops, focus group discussions, and a series of policy dialogues with local and central officials, national and foreign agencies can also promote dialogue that helps accurately identify key issues and underlying causes of local weaknesses, and consider potential solutions to remedy them. Such efforts should engage policymakers, local non- governmental actors, and other citizens. Indeed, the PAPI, implemented in Viet Nam, provides a useful model in this regard. The PAPI was undertaken in partnership between the UNDP and a local civil society organization. Importantly, the organization was well respected, committed to doing good work and also had a good relationship with the government. This gave it the inclination to engage constructively with methodologists on the team, the determination to use lessons learned and build capacity over time, and the latitude to undertake the study. Working with the civil society organization, and also engaging other stakeholders in the development of the survey instrument, helped to achieve local ownership. Additionally, the PAPI was disseminated in meetings with the government and local officials, but also through an interactive website that allowed citizens, local civil society organizations, and others to access reports, review the methodology and instruments, and undertake simple data manipulation. (See PAPI.vn and Annex 3 for examples of how data was presented on their website.) The dissemination was furthered as well by small grants provided to local and international researchers, who used the data and prepared working papers 13 addressing critical national issues. These papers formed the basis of workshops and discussions, sometimes sparking intense debate. There is also scope for the LGPI results to be institutionalized and mainstreamed by the various national and sub-national government institutions as part of their strategy and operations. This should be done through an inclusive approach with all relevant actors. In Tunisia, the government, in collaboration with various CSOs, led various efforts following the revolution to institutionalize participatory monitoring approaches. Options for collaboration with many relevant agencies could be leveraged in disseminating the LGPI findings and in incorporating the tool as part of their operations that can further build on existing efforts to improve public service delivery and enhance citizen engagement. Potential government counterparts may include the National Controllers Body for Public Services (Contrôle Général des Services Publics, CGSP) and the Prime Ministry’s Department of Public Administration Reforms (Direction Générale de la Reforme Administrative, DGRA) that jointly implemented a number of reforms and participatory monitoring. Examples include (i) the national score card ("baromètre de qualité et de gouvernance des services publics") that sought public feedback on over ten public services after which results were published online; (ii) the introduction of participatory audits as part of the mandate of the CGSP, to be overseen by a joint government-civil society coordinating committee; (iii) the adoption of international standards for participatory monitoring; (iv) the stipulation that all evaluations be published for reinforcing access to information and accountability; and (v) the clear emphasis on neutrality, objectivity and transparency of the CGSP’s mission (World Bank, forthcoming). A similar government-led approach was also utilized in disseminating the findings of PAPI in Vietnam. The central dissemination strategy included partnering with Ho Chi Minh National Academy of Politics and Public Administration in Hanoi, the Community Part and government-run political school that trained politicians, senior party members, and government officials, in an attempt at leveraging credibility and ownership of research results. Using PAPI results, the academy researchers in turn produced action- based research to support provinces with policy recommendations and drafting action plans. The academy also worked more closely with provinces to issue formal action plans to improve their PAPI scores. The academy researchers further owned the results by conducting case studies that examined in depth why provinces with similar levels of economic development and other indicators received different PAPI scores. Academy researchers afterwards presented their findings to the Provincial People’s Committees (Innovation for Successful Societies Case Study, 2014). 14 If results are disseminated continually and the survey implemented on a regular basis, the LGPI can help inculcate a public culture of demand for good governance and evidence- based policymaking. Once again, the PAPI provides insights. The PAPI has been implemented on an annual basis, in each of the country’s 62 provinces, and the res ults have become increasingly anticipated. In addition, short insights from the results are disseminated through Facebook and other social media on a weekly basis. By providing a short, provocative graph and insight, the PAPI helps to keep the issues of governance in the forefront of public debate. One point of caution to consider is the degree to which the details of the design and the construction of indicators should be shared publicly in recurrent studies. Repeated surveys implemented in the same localities over-time not only lead to survey-fatigue among citizens (possibly resulting in non-response bias), but they can also unwittingly open up opportunities for manipulation when local authorities are eager to rank-higher than then peer ‘competitors’. If localities (e.g., municipalities, provinces) become aware of the specific survey sites, they can strategically deploy resources to potential survey respondents at the expense of non-sampled areas, thus further exacerbating the very inequalities that we seek to measure and reduce. Authoritarian governments may even attempt to ‘mobilize’ citizens in the days preceding the arrival of the survey team in order to minimize the risk that embarrassing findings will travel up the bureaucratic ladder when the data is analyzed. Researchers must also consider the ethical issues regarding the possible tradeoff between the degree of transparency required in the dissemination exercise and the promises made to respondents about the strict anonymity of their answers. For those reasons, we recommend never to reveal information about the last level of clustering in any design, which thus protects both the households and the immediate communities in which they live from possible pressure or even retaliation. Such efforts require extensive attention by local actors and organizations, but they are critical to capitalize fully on the tool’s potential. This can help to establish greater ownership, enhance local capacity, spread awareness of weak governance and service delivery performance, and increase knowledge for policy and research. How Much Does it Cost to Implement the LGPI? The cost of implementing the LGPI depends on a number of factors. One important consideration is whether the LGPI modules are implemented independently or embedded in on-going survey projects. Embedding the modules in a survey is less costly, but it also should be considered carefully. For the LGPI to be 15 effective, it must be implemented in a sampling design that is representative at the local level. Embedding modules in a survey designed to be representative only at the national level will yield little benefit. A second consideration is the sample size of the survey. The sample size depends on the number sampled in each locality, and the number of localities sampled, as well as the overall structure of local administration in a given country, partculualy the number of levels of local government as well as average unit size at each level. The survey would be implemented ideally with local samples of at least 500 respondents, although samples of even 200 respondents may be acceptable. (The problems are a loss of statistical power and less ability to examine sub-population variations at the local level.) The number of localities to be sampled also depends to some extent on the size of the country. Even in very large countries, however, it may be useful to implement the LGPI, targeting regions/provinces of specific interest. Another consideration is the cost of survey implementation. This also varies significantly depending on the country. High-income countries and those that require heavy security (e.g., Libya, Afghanistan) are more expensive than lower income and less fragile countries. The costs will also vary depending on the extent and nature of dissemination. Similarly, supporting local NGO engagement in weekly dissemination requires additional funding, but it also expands the influence of the project. Establishing an interactive website can increase the impact of the LGPI significantly although it is also costly at the outset. Some of these costs may be reduced through cooperative or creative solutions. For instance, forming a consortium or joint effort to disseminate results from several countries on a single platform may reduce some of the costs of the website and dissemination. Finally, the costs decrease as the LGPI becomes institutionalized. The initial fielding incurs start-up costs in adapting the instrument, developing sampling frames, and piloting the survey, and sometimes nurturing local partner organizations in the science and workflow of survey research. Subsequent implementation reduces these costs considerably. Furthermore, by working with local actors and building capacity, costs can be reduced over time as they assume responsibility for implementation. A valuable lesson learned from the PAPI that was fielded in Viet Nam is value of regularized implementation of the tool. It not only helped to inculcate a culture of data-driven analysis and programming, but also built capacity that reduced costs of sustaining the effort. 16 Initial implementation requires resources. As a point of comparison, a full study in Tunisia’s 24 governorates, with 3 municipalities per governorate and a sample of 200 respondents per municipality, would cost approximately $500,000 USD for preparation, training and implementation, analysis and dissemination. (This would be similar to the pilot discussed in Annex 2, but include 24 governorates and 72 municipalities.) Yet, it can yield enormous benefits. The LGPI can play an important role in determining needs and assessing progress at the local level. It can also help answer fundamental questions about governance and service delivery. Today, decentralization is heavily favored but little is known about the variation in and drivers of local governance; the impact of governance on service provision is recognized, but studies of governance often limited to within facility factors; and the strength of non-state actors is valued but we do not yet grasp how and when citizens engage these actors. The LGPI can help answer such questions. 17 ANNEX 1: SURVEY MODULES 1 EDUCATION Now I would like to ask you some specific questions about education: 101. What is your highest level of education? 1. No formal 2. Primary 3. Secondary 4. Higher 96. DK 97. RA education School School education 102. Are there any children 17 years or under living in this household? 1. Yes 2. No 96. DK 97. RA “Children living here” includes all children 17 years and under who normally (or on a regular basis) sleep in this housing unit. Does not need to be the children of the respondent. For example, this can be a cousin or grandchildren if they normally sleep in this household. 103. How many children live in your household? The following was asked for every child separately. 104. Is this child a boy or girl? 1. Boy 2. Girl 105. How old is he/she? |___| years 106. Does this child attend school? 1. Yes 2. No 97. RA 107. Is this school located in this 1. Yes 2. No 96. DK 97. RA municipality? 108. Is this school public or private? 1. Public 2. Private 3. Private religious 97. RA 109. Which of the following are reasons he/she attends this school? [Multiple answers] a. Close to home b. No school choice c. Higher quality teaching d. Low cost school e. School has specific programs f. Only school available other. Other (specify) 110.1. In the last year, have you experienced a problem 1. Yes 2. No (skip 97. RA (skip to 111) with this child’s education at this school (such as to 111) s/he was failing a class, being expelled, etc.)? 110.2. [If yes], did you seek help? 2. No (skip 97. RA (skip to 111) 1. Yes to 111) 110.3. [If yes], who assisted you and your family? [Interviewer: Check all that apply – and follow-up “did anyone else assist you?”] a. Other INGOs/ Local CSOs b. Directorate of education c. Teacher d. Other e. Government official f. Principal g. Refuse to answer h. Business person j. Friend or family member k. Mayor m. Imam n. Local council member p. Religious organization 110.4. (If yes) Did you have to make an informal payment to 1. Yes 2. No 97. RA solve a problem at this school? 18 111.1. Did you get 1. Yes 2. No 96. DK 97. RA government assistance to pay for this child’s education in the last year? 111.2. Did you get non- 1. Yes 2. No (skip 96. Don’t 97. Refuse to government assistance to to 112) know answer (skip to 112) pay for this child’s education in the last year? 111.3. [If yes], who assisted you and your family? [Interviewer: Check all that apply – and follow-up “did anyone else assist you?”] [Check all that apply] a. Principal b. Don’t know c. Government official d. Friend or family member e. Women’s organization f. Directorate of education g. Other h. Refused to answer i. Teacher j. Other INGOs / Local CSOs k. Member of parliament 112.1. Did your child not attend school any day last week 1. Yes 2. No 96. DK (skip to 97. RA (skip although it was in session? 113) to 113) 112.2. If child missed one or more days], what are the main reasons they did not attend school? [Interviewer: Circle all that apply] (Required probe: Are there any other reasons?) a. The child had to work b. It was too hard to get to school c. Need to stay at home to given weather, distance, etc. help with house chores d. The child had problems with the e. The child had problems with f. The child was sick teacher or principal at school other students at school g. The child doesn’t like school h. The child was expelled or i. Other (mention) suspended 96. DK 97. RA 113. I’m going to read to you a series of statements about the school that this child attends. Please tell me if it is true or not. 1. 2. 96. 97. True Not true DK RA a. There are clean toilets □ □ □ □ b. The school has two or more shifts □ □ □ □ c. School is well-built □ □ □ □ d. Teachers are frequently absent □ □ □ □ e. Classrooms have more than 40 students □ □ □ □ f. Boys are safe going to/from school □ □ □ □ g. Girls are safe going to/from school □ □ □ □ j. The school has a Parent-teacher association □ □ □ □ k. Teachers favor in performance evaluations students who □ □ □ □ attend after-school study sessions 114. Do you do any of the following for a child living here 1. Yes 2. No 96. DK 97. RA [If no children living here, check no]: a. Helping a child with homework □ □ □ □ b. Purchasing school supplies for a child □ □ □ □ c. Paying for private lessons □ □ □ □ 19 g. Giving informal payments to teachers or principals at a □ □ □ □ child’s school 2 HEALTH Self-rated general health: 201. In general, how would you rate your overall health–very good, good, fair, poor, or very poor? 1. Very good 2. good 3. fair 4. poor 5. very poor 96. DK 202. Do you suffer from a chronic disease? 1. Yes 2. No 96. DK 97. RA 203.1.Have you 203.2. (If yes,) How old were 203.4. (If yes) Did you 203.3. (If yes) During ever had… you the first time you had the past 12 months, receive any (DX)? treatment for (DX) at any did you still have (DX)? time during the past 12 months? Asked for each of the following: 1. ARTHRITIS OR RHEUMATISM, 2. BACK OR NECK PROBLEMS, 3. FREQUENT OR SEVERE HEADACHES, 4. SEASONAL ALLERGIES, 5. STROKE, 6. HEART ATTACK, 7. HEART DISEASE, 8. HIGH BLOOD PRESSURE, 9. ASTHMA, 10. CHRONIC LUNG DISEASE, 11. DIABETES OR HIGH BLOOD SUGAR, 12. AN ULCER, 13. HIV (INFECTION), 14. EPILEPSY OR SEIZURES, 15. CANCER 16. OTHER（specify） 204. In general, how would you rate your overall mental health–very good, good, fair, poor, or very poor? 1. Very good 2. good 3. fair 4. poor 5. very poor 96. DK 205. The next questions are about how you have been feeling during the past 30 days. 1. All 2. 3. 4. A 5. 96. Most Some Little None DK 1. About how often during the past 30 days did □ □ □ □ □ □ you feel tired out for no good reason —would you say all of the time, most of the time, some of the time, a little of the time, or none of the time? 2. During the past 30 days, about how often did □ □ □ □ □ □ you feel nervous — all of the time, most of the time, some of the time, a little of the time, or none of the time? 3. How often did you feel so nervous that □ □ □ □ □ □ nothing could calm you down? (IF NEC: all, most, some, a little, or none of the time?) 4. During the past 30 days, about how often did □ □ □ □ □ □ you feel hopeless? (IF NEC: all, most, some, a little, or none of the time?) 5. During the past 30 days, about how often did □ □ □ □ □ □ you feel restless or fidgety? (IF NEC: all, most, 20 some, a little, or none of the time?) 607.6. How often did you feel so restless that you □ □ □ □ □ □ could not sit still? (IF NEC: all, most, some, a little, or none of the time?) 7. During the past 30 days, about how often did □ □ □ □ □ □ you feel depressed? (IF NEC: all, most, some, a little, or none of the time?) 8. How often did you feel so depressed that □ □ □ □ □ □ nothing could cheer you up? (IF NEC: all, most, some, a little, or none of the time?) 206. How long has it been since you visited a health clinic or medical doctor for care, either for yourself or accompanying others? 1. Last year 2. Between 1 and 3 years 3. More than 3 years ago (skip to 616) 4. Never (skip to 208) 96. DK 97. RA 207.1. (Speaking of the last time) Were you the patient? 1. Yes 2. No 96. DK 97. RA 207.2. Was this doctor or clinic in this municipality 1. Yes 2. No 96. DK 97. RA 207.3. Was this a: 1. Public clinic or medical center 2. Military center 3. Private doctor or clinic 4. Other 96. DK 97. RA 207.4. For each of the following, is it a reason you went to this doctor or clinic? 1. 2. 96. DK 97. RA Yes No 1. It was the closest facility □ □ □ □ 2. It is known for high quality care □ □ □ □ 3. It is not too expensive □ □ □ □ 4. You know doctors or nurses who work in it □ □ □ □ 5. You know people who helped you access medical care there □ □ □ □ 207.5. As I read you the following statements about that doctor’s office or clinic, please tell me whether it is true or not. 1. 2. 96. 97. True Not true DK RA 1. The restrooms were clean □ □ □ □ 2. The doctor was present □ □ □ □ 3. Patients were treated with respect □ □ □ □ 4. Expenses for the received treatment were reasonable □ □ □ □ 5. Fees were posted and clearly explained ahead of time □ □ □ □ 6. The waiting period between entering the clinic and the □ □ □ □ time you received the treatment was reasonable 21 7. I paid a bribe in order to get a better treatment □ □ □ □ 8. I used wasta/aktef in order to get a better treatment □ □ □ □ 9. The necessary medicines were readily available in the □ □ □ □ pharmacy 10. I was referred to a private clinic for further tests and □ □ □ □ treatment 207.6. For each of the following, would you say you were very satisfied, not very satisfied, somewhat dissatisfied, very unsatisfied? 1. 2. 3. 4. 96. 97. Very Not very Somewhat Very DK RA satisfied satisfied dissatisfied unsatisfied 1. Cost □ □ □ □ □ □ 2. Personal □ □ □ □ □ □ care of patient care 3. Medical □ □ □ □ □ □ services 4. Room and □ □ □ □ □ □ building 207.7. Did you receive assistance in accessing or paying for this health care? 1. Yes 2. No (skip to 208) 96. DK (skip to 208) 97. RA (skip to 208) Financial assistance is material or financial support. 207.8. For each of the following, please tell me if they assisted you and your family? 1. Yes 2. No 96. DK 97. RA 1. Members of your extended family □ □ □ □ 2. Neighbors □ □ □ □ 3. Other members of your tribe or ethnic group □ □ □ □ 4. Local imam □ □ □ □ 5. Religious charity □ □ □ □ 6. Non-religious charity □ □ □ □ 7. Other civil society association □ □ □ □ 8. Trade or professional association □ □ □ □ 9. Leaders of a powerful local family or clan □ □ □ □ 10. Local council member □ □ □ □ 11. State welfare office □ □ □ □ 12.Other group □ □ □ □ 208. In the past year, did you seek medical advice from a …? 1. Yes 2. No 96. DK 97. RA 1. Pharmacist □ □ □ □ 2. Local healer □ □ □ □ 3. The internet □ □ □ □ 4. Friends, family and neighbors □ □ □ □ 22 209. How good do you believe the quality of medical care is at…? 1. 2. 3. 4. Very 96. 97. Excellent Good Poor Bad DK RA 1. The general public hospital nearest you □ □ □ □ □ □ 2. The military hospital nearest you □ □ □ □ □ □ 3. The private hospital nearest you □ □ □ □ □ □ 4. A private doctor or clinic □ □ □ □ □ □ 5. A public clinic nearest you □ □ □ □ □ □ 6. A pharmacist nearest you □ □ □ □ □ □ 7. A local healer □ □ □ □ □ □ 210. If you showed up at the public clinic nearest you, do you think the doctor would be present and see you if the time was…? 1. Yes 2. No 96. DK 97. RA 1. 8:30 am □ □ □ □ 2. Lunch time □ □ □ □ 3. 4:30 pm □ □ □ □ 211. Do you or anyone in your household living here have any health care needs that you are not able to attend to, regarding:… 1. Yes 2. No 96. DK 97. RA 1. Medical health □ □ □ □ 2. Dental health □ □ □ □ 3. Hearing □ □ □ □ 4. Mental health □ □ □ □ 5. Contraception □ □ □ □ 212. Are you covered by medical insurance? 1. Yes 2. No (skip to 214) 96. Don’t know (skip 97. Refuse (skip to 214) to 214) Note: Medical insurance is any private or public coverage that pays all or part of your medical expenses. 213. What kind of medical insurance do you receive? 1. CNAM 2. Private health 3. Chahedat fakr (Poorness coverage certificate) 4. Other (specify)_____ 96. DK 97. RA 214. Are there people in your family who are unable to obtain health insurance coverage? 1. Yes 2. No 96. DK 97. RA 23 3 PHYSICAL SECURITY AND DISPUTE RESOLUTION 301. Please tell me if you would feel very unsafe, somewhat unsafe, somewhat safe, or very safe…? 1. 2. 3. 4. 96. 97. Very Somewhat Somewhat Very DK RA unsafe unsafe safe safe 1. In your home □ □ □ □ □ □ 2. Walking in your neighborhood during the day □ □ □ □ □ □ 3. Walking in your neighborhood at night □ □ □ □ □ □ 4. At the local market □ □ □ □ □ □ 5. At the local mosque □ □ □ □ □ □ 6. At the nearest school □ □ □ □ □ □ (If you don’t know, how safe do you think you would feel?) 302. In the past year, have you heard of the following problems occurring in (NAME MUNICIPALITY)? 1. Yes 2. No 96. 97. DK RA 1. Murder □ □ □ □ 2. Armed Robbery □ □ □ □ 3. Burglary □ □ □ □ 4. Kidnapping □ □ □ □ 5. Car theft and car jacking □ □ □ □ 6. Celebratory shooting □ □ □ □ 7. Assaults □ □ □ □ 8. Sexual violence □ □ □ □ 9. Spousal abuse (Physical, mental, etc.) □ □ □ □ 10. Smuggling of goods □ □ □ □ 11. Illegal gambling □ □ □ □ 12. Prostitution □ □ □ □ 13. Drug use □ □ □ □ 14. Disputes over land □ □ □ □ 15. Disputes over access to water □ □ □ □ 16. Disputes over smuggling □ □ □ □ 303. In the past 12 months, have you been personally the victim of …: 1. Yes 2. No 96. DK 97. RA 1. Assault/ physical violence (other than domestic violence) □ □ □ □ 2.Domestic violence (Physical, verbal, etc.) □ □ □ □ 3. Theft or burglary □ □ □ □ 4. Armed robbery □ □ □ □ 5. Murder of a friend or family member □ □ □ □ 6. Kidnapping of friend or family member □ □ □ □ 7. Car theft and car jacking □ □ □ □ 24 303.1. (if yes) which of the following persons, groups or institutions did you ask for help? [Check all that apply] Questions: 1. Assault/ physical violence (other than domestic 2.Domestic violence (Physical, verbal, etc.) violence) 3. Theft or burglary 4. Armed robbery 5. Murder of a friend or family member 6. Kidnapping of friend or family member 7. Car theft and car jacking Answer choices (for all): 1. Members of your extended family 2. Neighbors 3. Other members of your tribe or ethnic group 4. A local imam 5. Members of your mosque or 6. A religious charity/CSO religious organization 7. A non-religious charity/CSO 8. A trade or professional 9. Leaders of a powerful local family association or clan 10. Local municipal council 11. Other local council member 12. State welfare office (need to head/mayor check specifics) 13. Member of parliament 14. Candidate in parliamentary 15. Other elections 304. In the past 12 months, have you been a party to… 1. Yes 2. No 96. DK 97. RA 1. Dispute over land □ □ □ □ 2. Dispute over water □ □ □ □ 3. Child custody battle □ □ □ □ 4. Contract dispute □ □ □ □ 5. Other dispute □ □ □ □ 304.1. (if yes) Which of the following person, groups or institutions did you ask for help? [Check all that apply.] Questions: 1. Dispute over land 2. Dispute over water 3. Child custody battle 4. Contract dispute 5. Other dispute Answer choices (for all): 1. Members of your extended family 2. Neighbors 3. Other members of your tribe or ethnic group 4. A local imam 5. Members of your mosque or religious 6. A religious charity/CSO organization 7. A non-religious charity/CSO 8. A trade or professional association 9. Leaders of a powerful local family or clan 10. Local municipal council head/mayor 11. Other local council member 12. State welfare office (need to check specifics) 13. Member of parliament 14. Candidate in parliamentary elections 15. Lawyer 16. Other 304.2. Were they able to solve the problem? 1. Yes 2. No 96. DK 97. RA 25 1. Dispute over land □ □ □ □ 2. Dispute over water □ □ □ □ 3. Child custody battle □ □ □ □ 4. Contract dispute □ □ □ □ 5. Other dispute □ □ □ □ 305. Now I’d like to ask you, for each of the following, how much you believe they help to resolve civil disputes, such as conflicts over land, water or family, fairly and effectively and security issues, such as crime against physical property or persons in the area. Please rate on a scale from 0-10, with 0 being absolutely unhelpful and 10 being completely helpful. Let us begin with members of your extended family, how much can they help solve civil disputes? Security issues? (Please repeat for each of the following) 305.1. Civil disputes? 305.2. Security issues? 0-10 Ranking 96/97 0-10 Ranking 96/97 DK/RA DK/RA 1. Members of your extended family 2. Friends 3. Business associates 4. Neighbors 5. Other members of your tribe or ethnic group 6. Local imam 7. Members of your mosque or religious organization 8. Religious charities/CSOs 9. Non-religious charities/CSOs 10. Trade or professional associations 11. Leaders of a powerful local families or clans 12. Local municipal council head/mayor 13. Other local council members 14. State welfare office (need to check specifics) 15. Lawyers and Judges 16. Mediator employed by the local authorities 17. Member of parliament 18. Police 19. Other ……… 306. Would you say that the police in [Name municipality]… 1. Yes 2. No 96. DK 97. RA 1. Can solve security problems □ □ □ □ 2. Respond quickly □ □ □ □ 3. Are honest □ □ □ □ 4. Know the community well □ □ □ □ 5. Are fair □ □ □ □ 26 307. Have you ever been subjected to a traffic fine? 1. Yes 2. No (Skip to 401) 96. DK (Skip to 401) 97. RA (Skip to 4o1) Thinking of the last time you were fined, did he/she clearly explain the reasons for penalty(ies), legal ground and type of applied penalty? 308. did he/she clearly explain the reasons for penalty(ies), legal 1. Yes 2. No 96. DK 97. RA ground and type of applied penalty? 309. engage in abusive acts or language toward the involved person(s)? 310. Did you receive a written ticket and invoice? 311. Were you issued an invoice, but one with a noted amount lower than the actual fine? 312. Were you issued an invoice, but one with a noted amount lower than the actual fine? 313. Did you call a powerful person to intervene? 314. Did you pay a bribe? 4 SOCIAL ASSISTANCE AND WELFARE 401. How often is the garbage picked up by the public works agency in your village/neighborhood? 1. Every day 2. More than once a week 3. Once a week 4. 2-3 times per month 5. Once a month 6. Less than once a month 7. Never, practically never 96. DK 97. RA 402. Do you think that garbage collection is…? 1. More often than 2. As often as 3. Less often than 96. DK 97. RA necessary necessary necessary 403. What source of water do you use for drinking in your neighborhood? [check all that apply] 1. Mineral water (bottled) 2. Water tap to home 3. Shared water tap at a common place 4. Open water well 5. Drilled water well 6. River/Lake water 7. Rain water 8. Other 96. DK 97. RA 404. Within the last week, have you …. 1. Yes 2. No (skip to 407) 96. DK (skip to 97. RA (skip to 407) 407) 1.... not had □ □ □ □ enough food? 2. …experienced □ □ □ □ an electricity cut? 3. …not had □ □ □ □ enough drinking water? 4. …experienced a □ □ □ □ water cut? 27 405. Did you ask any of the following for assistance (charity), including 406. (If Yes to 405), food, clothing and shelter? were they helpful? 1. 2. No 96. 97. 1. 2. 96. 97. Yes (skip to DK (skip RA (skip Yes No DK RA 407) to 407) to 407) 1. Members of your extended family □ □ □ □ □ □ □ □ 2. Neighbors □ □ □ □ □ □ □ □ 3. Other members of your tribe or □ □ □ □ □ □ □ □ ethnic group 4. A local imam □ □ □ □ □ □ □ □ 5. Members of your mosque or □ □ □ □ □ □ □ □ religious organization 6. A charity/CSO □ □ □ □ □ □ □ □ 7. Police □ □ □ □ □ □ □ □ 8. A trade or professional association □ □ □ □ □ □ □ □ 9. Leaders of a powerful local family □ □ □ □ □ □ □ □ or clan 10. Local municipal council □ □ □ □ □ □ □ □ head/mayor 11. Member of parliament □ □ □ □ □ □ □ □ 12. Candidate in parliamentary □ □ □ □ □ □ □ □ elections 10. Other local council member □ □ □ □ □ □ □ □ 11. State welfare office (need to check □ □ □ □ □ □ □ □ specifics) 12. Member of parliament □ □ □ □ □ □ □ □ 13. Candidate in parliamentary □ □ □ □ □ □ □ □ elections 14. other (specify) □ □ □ □ □ □ □ □ ………………………….. 407. Now I would like to talk about assistance to the poor. In your commune/ward, please tell me if each of the following provide assistance for low-income households. 1. 2. 96. 97. Yes No DK RA 1. Neighbors □ □ □ □ 2. Local imam □ □ □ □ 3. Civil society and associations □ □ □ □ 4. Other members of tribe or ethnic group □ □ □ □ 5. Members of mosques or religious organizations □ □ □ □ 6. Leaders of powerful local families or clans □ □ □ □ 7. State welfare office □ □ □ □ 8. Local municipal council members □ □ □ □ 9. Businessmen and organizations □ □ □ □ 10. Trade unions or professional associations □ □ □ □ 28 408. Do you personally feel you are eligible for government subsidies/assistance but have been unable to obtain it? 1. Yes 2. No 96. DK 97. RA 409. Would you say there are none, a few, or many families eligible for government subsidies/assistance in this municipality but have been unable to obtain it? 1. None 2. A few 3. Many 96. DK 97. RA 410. Do you believe that there are households that receive subsidies but who are not poor? Would you say there are none, a few, or many such cases? 1. None 2. A few 3. Many 96. DK 97. RA 411. Did you seek help to find employment in the past year ? 1. Yes 2. No (skip to 96. Don’t know skip to 413) 97. Refuse to Answer skip to 413) 413) 412. (If yes to 411) Which of the following did you seek help from [check all that apply]? 1. Yes 2. No 96. DK 97. RA 1. Friends □ □ □ □ 2. Your family □ □ □ □ 3. Neighbors □ □ □ □ 4. Members of tribe or ethnic group □ □ □ □ 5. Local imam □ □ □ □ 6. Members of mosques or religious organizations □ □ □ □ 7. Civil society and associations □ □ □ □ 8. Trade unions or professional associations □ □ □ □ 9. Leaders of powerful local families or clans □ □ □ □ 10. Local council members □ □ □ □ 11. Businessmen and organizations □ □ □ □ 12. ANETI □ □ □ □ 413. How useful are bribe payments to government officials to: [Read categories 1 to 5] 1. 2. 3. 4. 5. 96. 97. Not Slightl Moderatel Very Essential DK RA useful y y useful useful useful □ □ □ □ □ □ □ 1. Obtain a job in the government sector 2. Obtain a job in the private sector □ □ □ □ □ □ □ 29 414. How useful is wasta (“Aktaf”) for you to: 1. 2. 3. 4. 5. 96. 97. Not Slightly Moderately Very Essential DK RA useful useful useful useful 1. Obtain a job in the □ □ □ □ □ □ □ government sector 2. Obtain a job in the private □ □ □ □ □ □ □ sector 415. How important is a person’s professional qualifications in obtaining a government job relative to other factors? Please indicate whether each is more important, equally important, or less important as a person’s experience and professional qualifications. 1. 2. 3. 96. 97. Less Equal More DK RA important than importance with important than experience and experience and experience and qualifications qualifications qualifications 1. A person’s □ □ □ □ □ political affiliations 2. A person’s □ □ □ □ □ personal network, relation to influential people 3. A person’s family □ □ □ □ □ or tribal affiliation Note: Professional qualifications are things like diplomas, experience, and professional recommendations. 416. How important is a person’s professional qualifications in obtaining at job in the private sector relative to other factors? Please indicate whether each is more important, equally important, or less important as a person’s experience and professional qualifications. 1. 2. 3. 96. 97. Less Equal More DK RA important than importance with Important than experience and experience and experience and qualifications qualifications qualifications 1. A person’s □ □ □ □ □ political affiliations 2. A person’s □ □ □ □ □ personal network, relation to influential people 3. A person’s □ □ □ □ □ family or tribal affiliation 30 417. Please tell me to what extent you believe (your friends) would be to help you with [financial assistance / solving a dispute with a third party]? 417.1. Financial assistance 417.2. Solving a dispute with a third party 1. 2. I 3. 4. 96. 97. 1. 2. I 3. 4. 96. 97. RA Never think Maybe Sure DK RA Never think Maybe Sure DK so so 1. Your friends □ □ □ □ □ □ □ □ □ □ □ □ and family 2. Your □ □ □ □ □ □ □ □ □ □ □ □ neighbors 3. Members of □ □ □ □ □ □ □ □ □ □ □ □ your tribe/ethnic group 4. Members of □ □ □ □ □ □ □ □ □ □ □ □ mosque or religious group 5. Members of □ □ □ □ □ □ □ □ □ □ □ □ a civil society organization you support 6. Leaders of □ □ □ □ □ □ □ □ □ □ □ □ powerful local families or clans 7. Local council □ □ □ □ □ □ □ □ □ □ □ □ members 8. Local □ □ □ □ □ □ □ □ □ □ □ □ businesses 9. Police □ □ □ □ □ □ □ □ □ □ □ □ 10. Other □ □ □ □ □ □ □ □ □ □ □ □ 5 CITIZEN-STATE LINKAGES 501. Did you vote in the second round of the 2014 presidential election? 1. Yes 2. No 96. DK 97. RA 502. Did you vote in the first round of the 2014 presidential election? 1. Yes 2. No 96. DK 97. RA 503. Did you vote for the parliamentary elections that were held in November 2014? 1 Yes 2 No 96. DK 97. RA 504. Do many of the politicians in your community try to buy votes with gifts, money, or access to services? 1. Yes 2. No 96. DK 97. RA 31 504.1. Did any politician or representative offer you gifts, money, or access to services for your vote? 1. Yes 2. No 96. DK 97. RA 505. Did you vote in the elections for the Constituent Assembly in October 2011? 1. Yes 2. No 96. DK 97. RA 506. Do you personally know any members who sat on the Constituent Assembly, elected in 2011? 1. Yes 2. No 96. DK 97. RA 507. Do you personally know any members of the new parliament, elected in 2014? 1. Yes 2. No 96. DK 97. RA 508. In your opinion, what is the most important problem Tunisia is facing today [do not read out categories. Write answer and insert code from Table 3 following interview.] Answer: ___________________ Code: _________ Record exactly what R said. 509. In your opinion, what is the most important problem facing [NAME OF MUNICIPALITY] today? [do not read out categories. Write answer and insert code from Table 3 following interview.] Answer: ___________________ Code: _________ Record exactly what R said. 510. In the last year, have you contacted a member of the municipality council “Majlis Baladiyya” or “Niabba Khassousiya”? 1. Yes 2. No (skip to 511) 96. DK (skip to 511) 97. RA (skip to 511) 510.1. The last time you contacted a member of the local council, was it about a personal or community problem? 1. Personal 2. Community problem 96. DK 97. RA 510.2. (The last time you contacted a member of local council), were you very satisfied, somewhat satisfied, somewhat dissatisfied, or very dissatisfied with the response? 1. Very satisfied 2. Somewhat satisfied 3. Somewhat 4. Very dissatisfied dissatisfied 96. DK 97. RA -999. Not applicable 511. If you had a personal problem in the future that one of the members of the local council could do something about, do you think he or she would be very helpful, somewhat helpful, or not very helpful? 1. Very helpful 2. Somewhat helpful 3. Not very helpful 96. DK 97. RA 512. Do you personally know any local council members in this municipality? 1. Yes 2. N0 96. DK 97. RA 513. Can you please tell me who is the head of the municipal council? 1. Respondent answers with a name or description 2. Respondent does not answer with a name or description 32 514. Do you think that the mayor responds primarily to the: 1. Interests of national government 2. Interests of the community 3. Interests of part of the community as a whole 4. His own personal interests 96. DK 97. RA 515. In general, to what extent do the following groups influence local politics in your area? Not at all, very little, somewhat, a great deal? 1. Not 2. Very 3. 4. A great 96. 97. RA at all little Somewhat deal DK 1. Average citizens □ □ □ □ □ □ 2. Powerful local families □ □ □ □ □ □ 3. Trade unions or professional associations □ □ □ □ □ □ 4. Mosques or religious groups □ □ □ □ □ □ 5. CSOs or associations □ □ □ □ □ □ 6. Political parties □ □ □ □ □ □ 7. The chamber of commerce or business □ □ □ □ □ □ organizations 8. Clans or tribes □ □ □ □ □ □ 516. Have you (personally) engaged in meeting with any of the following groups in order to determine the community needs and plan services? 1. Yes 2. No 96. DK 97. RA 1. Your neighbors □ □ □ □ 2. Your friends □ □ □ □ 3. Family □ □ □ □ 4. A powerful local family □ □ □ □ 5. Trade union or professional □ □ □ □ association 6. Mosque or religious group □ □ □ □ 7. CSO or association □ □ □ □ 8. A political party or parties □ □ □ □ 9. A chamber of commerce or □ □ □ □ business organization 10. Your clan or tribe □ □ □ □ 517. Do you think you are better able to get a local council member to respond to you if the council member is: 1. Yes 2. No 96. DK 97. RA 1. Well educated □ □ □ □ 2. Has lived in this [CITY NAME] his/her whole life □ □ □ □ 3. Is from your family or clan □ □ □ □ 4. Is your friend or from your personal network □ □ □ □ 5. Is a woman □ □ □ □ 6. Is religiously devout □ □ □ □ 7. Is a member of a political party you support □ □ □ □ 8. Is affiliated with a trade union or professional association that you □ □ □ □ 33 support or belong to 9. Is a member of a civil society association you belong to □ □ □ □ 518. Have you been a victim of corruption by a local government official in the past year? (Corruption occurs when an official misuses his/her position for personal financial gain. A local government official is an elected or appointed major or council person or any other civil servant working in your local community, such as a teacher, permit provider, police officer, etc.) 1. Yes 2. No 96. DK 97. RA Note: Corruption occurs when an official misuses his/her position for personal financial gain. 519. Has anyone else in your immediate family been a victim of corruption in the past year? 1. Yes 2. No 96. Don’t know (skip to 522) 97. Refuse to answer (skip to 522) 520. (ASK ONLY IF YES to 518 or 519) Did you report this act of corruption? 1. Yes (skip to 522) 2. No 96. DK 97. RA 521. (If No to 520) Why not? [Check all that apply] 1. It is useless to report 2. The process of reporting 3. I was scared about retaliation corruption corruption is too burdensome 4. I don’t know how to go 5. Other (Specify) ____ 96. DK 97. RA about reporting corruption Note: If R asks for clarification, say, “Whatever it means to you.” 522. When there is a corruption case going on in your governorate, how serious do you think the government deals with the case? 1. Very 2. Quite 3. Not very 4. not 96. DK 97. RA Seriously seriously seriously seriously at all Note: If R asks for clarification, say, “Whatever it means to you.” 523. Have you been a victim of corruption by any other public official in the past year? 1. Yes 2. No (skip to 525) 96. DK(skip to 525) 97. RA(skip to 525) 524. If yes, by which type of official? |__________| 96. DK 97. RA 525. Among the following groups, who do you believe currently has the most impact on (primary and secondary education): central government, the governorate, the local council, imams and religious organizations, the private sector/businesses, unions, large families and clans, and other civil society organizations, who do you think currently has the most impact on for education? [repeat for each] organizations 8. Other CSO Governorate government businesses 4. Imams/ 6. Unions 1. Central 5. Private families/ religious 7. Large 3. Local council sector/ 96. DK 97. RA clans 2. 1. Education (primary □ □ □ □ □ □ □ □ □ □ and secondary schools) 2. Local Economic □ □ □ □ □ □ □ □ □ □ 34 development 3. Health services □ □ □ □ □ □ □ □ □ □ (clinics and hospitals) 4. Public safety □ □ □ □ □ □ □ □ □ □ 5. Social Welfare □ □ □ □ □ □ □ □ □ □ (disability benefits, unemployment benefits, etc.) 6. Solid waste □ □ □ □ □ □ □ □ □ □ collection 7. Drinking Water □ □ □ □ □ □ □ □ □ □ 8. Electricity □ □ □ □ □ □ □ □ □ □ 526. Among the following groups, who would prefer to be responsible for (Education (primary and secondary schools): central government, the governorate, the local council, imams and religious organizations, the private sector/businesses, unions, large families and clans, and other civil society organizations? [repeat for each] organizations 8. Other CSO Governorate government businesses 4. Imams/ 6. Unions 1. Central 5. Private families/ religious 7. Large 3. Local council sector/ 96. DK 97. RA clans 2. 1. Education □ □ □ □ □ □ □ □ □ □ (primary and secondary schools) 2. Local Economic □ □ □ □ □ □ □ □ □ □ development 3. Health services □ □ □ □ □ □ □ □ □ □ (clinics and hospitals) 4. Public safety □ □ □ □ □ □ □ □ □ □ 5. Social Welfare □ □ □ □ □ □ □ □ □ □ (disability benefits, unemployment benefits, etc.) 6. Solid waste □ □ □ □ □ □ □ □ □ □ collection 7. Drinking Water □ □ □ □ □ □ □ □ □ □ 8. Electricity □ □ □ □ □ □ □ □ □ □ 35 527. I’d like to ask you now a little about your own activities. Are you a member or a leader of a … (mosque or religious organization)? [Interviewer: repeat for each] 1. 2. 3. Not 96. 97. RA Member Leader member DK nor leader 1. A mosque or local religious organization? □ □ □ □ □ 2. A civil society organization or association, such as a □ □ □ □ □ charity, women’s group, sports club? 3. A trade union or professional association? □ □ □ □ □ 4. A powerful local family? □ □ □ □ □ 5. A political party □ □ □ □ □ 6. A chamber of commerce or business organization? □ □ □ □ □ 7. A clan or tribe? □ □ □ □ □ 8. Another type of association (Please specify) □ □ □ □ □ 6 SOCIAL COMPOSITION AND CULTURE 601. In your neighborhood would you say… 1. Most people know each other 2. Some people know each other 3. Few people know each other 96. Don’t know 97. Refuse to answer 602. Do you and your neighbors help 603. [If yes,] is this on a daily, weekly, monthly, each other with…? less than monthly basis? (Which is closest?) 1.Yes 2. 96.DK 97. 1. 2. 3. 3. Less 69. 97. No (Skip RA Daily Weekly Monthly than Don’t Refuse (Skip to (Skip monthly know to to 204) to answer 204) 204) .1 Childcare 1 2 96 97 1 2 3 4 96 97 .2 Education 1 2 96 97 1 2 3 4 96 97 .3 Healthcare 1 2 96 97 1 2 3 4 96 97 .4 Finances 1 2 96 97 1 2 3 4 96 97 .5 Obtaining govt 1 2 96 97 1 2 3 4 96 97 services .6 Solving disputes 1 2 96 97 1 2 3 4 96 97 within families . 7 Solving disputes 1 2 96 97 1 2 3 4 96 97 between neighbors .8 Keeping your 1 2 96 97 1 2 3 4 96 97 neighborhood safe from crime .9 Keeping the area 1 2 96 97 1 2 3 4 96 97 clean 36 604. We would like to understand how people interact in your area. How much do you think people are obligated to help others, even if it costs them a day’s wages? Would you say that the person is very obligated, somewhat obligated, or not obligated at all to help, if the person is… 1. Very 2. Somewhat 3. Not 96. Don’t 97. Refuse to obligated obligated obligated at all know answer 604.1. A member of their immediate family (son or daughter, mother or father)? 604.2. A member of their extended family (uncle aunt, grandparent)? 604.3. A neighbor? 604.4. A member of their large family or clan? 604.5. A member of their tribe? 604.6. A resident of their town/village but not a neighbor? 604.7. A fellow work mate? 604.8. Their employee? 604.9. A member of their mosque? 604.10. A Tunisian? 604.11. A non- Tunisian? 605. If you knew only the family name of someone in this area, how certain would you be that you know whether they are among the most or least privileged people in the municipality? Would you say you are 1. nearly certain 2. somewhat certain 3. or don't know whether a person is rich or poor, given just their family name. 606. How useful is wasta (“Aktaf”) for you to: 1 2 3 4 5 96 97 Not Slightly Moderately Very Essential Don’t Refuse useful useful useful useful know to answer 1. Obtain a job in the □ □ □ □ □ □ □ government sector 37 2. Obtain a job in the private sector 607. How important is a person’s professional qualifications in obtaining a government job relative to other factors? Please indicate whether each is more important, equally important, or less important as a person’s experience and professional qualifications. 1 2 3 96 97 Less Equal More Don’t Refuse to important importance important know answer than with than experience experi-ence experience and and and qualifica- qualifica- qualifications tions tions 1. A person’s □ □ □ □ □ political affiliations 2. A person’s □ □ □ □ □ personal network, relation to influential people 3. A person’s □ □ □ □ □ family or tribal affiliation 608. How important is a person’s professional qualifications in obtaining at job in the private sector relative to other factors? Please indicate whether each is more important, equally important, or less important as a person’s experience and professional qualifications. 1 2 3 96 97 Less Equal More Don’t Refuse to importantth importance Important know answer an with than experience experi-ence experience and and and qualifica- qualifica- qualifications tions tions 1. A person’s □ □ □ □ □ political affiliations 2. A person’s □ □ □ □ □ personal network, relation to influential people 3. A person’s □ □ □ □ □ 38 family or tribal affiliation 609. In general, to what extent do the following groups influence local politics in your area? Not at all, very little, somewhat, a great deal? 1 2 3 4 96/97 Not at all Very little Somewhat A great deal DK/RA 1. Average citizens 1 2 3 4 96/97 2. Powerful local families 1 2 3 4 96/97 3. Trade unions or 1 2 3 4 96/97 professional associations 4. Mosques or religious 1 2 3 4 96/97 groups 5. CSOs or associations 1 2 3 4 96/97 6. Political parties 1 2 3 4 96/97 7. The chamber of 1 2 3 4 96/97 commerce or business organizations 8. Clans or tribes 1 2 3 4 96/97 610. Now I would like to talk about assistance to the poor. In your commune/ward, please tell me if each of the following provide assistance for low-income households. 1 2 96 97 Yes No Don’t know Refuse to answer 1. Neighbors 2. Members of tribe or ethnic group 3. Local imam 4. Members of mosques or religious organizations 5. Civil society and associations 6. Trade unions or professional associations 7. Leaders of powerful local families or clans 8. Local municipal council members 9. Businessmen and organizations 10. State welfare office 39 ANNEX 2: PRELIMINARY INSIGHTS FROM LGPI PILOT STUDY, TUNISIA 2015 The Program on Governance and Local Development at Yale University (GLD) designed the pilot LGPI citizen household survey and implemented it in Tunisia in February 2015. The survey was fielded in 6 of Tunisia’s 24 governorates, focusing on 3 municipalities in each governorate.3 (See Table 1 for a list of municipalities.) The pilot study illustrates how the tool can help provide a comprehensive picture of service delivery and governance at the municipal level, identifying key areas of need, highlighting significant inequalities, and shedding light on the relationship between governance and service delivery. Governorate Municipality Bizerte Bizerte Nord Bizerte Sud Menzal Jemil Mahdia Mahdia Ksour Essaf Souassi Monastir Monastir Taboulba Moknine Sfax Sfax Sud Sfax Nord Sakiet Eddair Siliana Siliana Ville Bouarada Gaafour Tunis La Marsa Cité El Khadra Jebel Jloud Table 1. Municipalities included in the LGPI Pilot Survey, Tunisia 2015. Health The Tunisian 2015 pilot study demonstrates how the survey can yield insights into the varying qualities of service across municipalities. It also helps ascertain inequalities based on socio- economic status, gender, and availability of public and private sector services. The study found that across municipalities there were exceedingly low expectations for a doctor to be present in local clinics during usual business hours, suggesting a high rate of absenteeism in public clinics. When asked if they believed that the doctor would be present at 8:30 am, at lunchtime, and at 4:30 pm, respondents overwhelmingly answered that the doctor would not be present. There is some variation at the municipal level, as shown in Figure 1. There is a much greater expectation that the doctor will be present in Gaafour, for instance, 3 Households within each governorate were randomly sampled, using PPS by points of light data (due to the fact that the last census was conducted in 2004 and there appear to be considerable changes in population distribution in the last decade.) Respondents within households were chosen randomly from among those over 18 years old, using Kish tables. 40 than in Siliana Ville, despite the fact that these municipalities are located in the same governorate. Figure 1. Doctor Availability: Questions Included: Do you think there is a doctor present in the local clinic at 8:30 am, lunchtime and 4:30 pm? The answers were coded as a 1 if the answer is “Yes” and as 0 if answer is “No.” This yields a score that ranges from 0 to 3 for each respondent. The sums are averaged across respondents in the municipality. A score of 0 thus would mean that all respondents believed the doctor is never present, and a score of 3 would mean that all respondents believe the doctor is present at all of these times. The study further examined the quality of healthcare, focusing on respondents who had gone to a doctor or clinic in the last two years. Respondents were asked about specific traits of a clinic visit, including facility cleanliness, duration of waiting time, costliness, respectfulness of staff, and availability of medicine. The results again revealed significant variation at the local level, even within the same governorate. For instance, while Mahdia scores among the highest on the Health Quality Index, the other two municipalities sampled in the governorate, Ksour Essaf and Souassi, scored considerably worse. 41 Figure 2. Health Quality Index: Index included 1) clean bathroom, 2) doctor present, 3) reasonable costs, 4) reasonable wait time, 5) respectful treatment and 6) medication availability. Similar variance is seen in the Health Governance Index, which reflects the transparency and accountability of local clinics and doctors. The Index is based on patient use of informal payments or wasta to obtain quality medical care, the clear posting of fees, and referrals to private clinics for necessary care (See Figure 3). We find areas in which poor governance is a clear concern, and, within the index, also find significant variation in the extent to which this is due to low transparency, the use of informal payments and wasta, or the tendency to refer patients to private providers. 42 Figure 3. Health Governance Index. Variables included 1) Were you referred to private clinic/doctor? 2) Did you pay an informal payment for better treatment? 3) Did you use wasta to obtain better treatment? 4) Were fees clearly posted? Furthermore, the LGPI reveals discrepancies between healthcare provided in the public and the private sectors. With the exception of Jbel Jloud, healthcare provided by private clinics far surpassed the quality of that provided by public medical centers. (See Figure 4). For most municipalities even the highest scores given to public medical centers were inferior to those given to the lowest scoring private clinics. 43 Figure 4. Private and Public Clinics: The index is comprised of six variables, including 1) clean bathroom, 2) doctor present, 3) reasonable costs, 4) reasonable wait time, 5) respectful treatment, and 6) medication availability. The results indicate a disparity between the services that patients receive under private care and those that they receive under public care. Public healthcare centers lag in doctor availability, duration of patient wait times, availability of proper medication, and the completion of treatment without the use of bribery or wasta. Education The LGPI also allows us to examine the quality of education service delivery in Tunisia, finding high numbers of students failing to complete their educations. At the national level, 10.4% of the youth aged 8-17 years (or 178,341 Tunisian youth) did not attend school, including 11.9% (107,674) of the males and 8.9% (70,667) of the females. Considering the results at the governorate level, we find the lowest percentage was observed in Sfax, where 3.9% of potential students this age were out of school, and the highest observed in Monastir, with 10.1%. The official MOE statistic (2012-13) is 107,000 (under 16).4 Close examination of the data also reveals gender differences in dropout rates. In Mahdia for example, the dropout rate is relatively low and equal between both male and female students. (See Figure 5.) In Monastir, however, where the dropout rate is the highest overall, the level 4 See more about dropout rates in Tunisia: http://nawaat.org/portail/2014/09/16/abandon-scolaire-en-tunisie-des- chiffres-alarmants-banalises-par-la-crise-economique-et-occultes-par-le-ministere-de-tutelle/ 44 for male students far surpasses that of female students. Siliana, on the other hand, presents a different situation in which there are more female students dropping out of school than male students. A student’s decision to leave school often rests on factors such as financial need, gender dynamics, customs, and available employment alternatives. Thus, the data reflect other fundamental differences in local governance that are ripe for exploration. Figure 5. Drop-out Rates, by Gender and Governorate The LGPI can also help to pinpoint specific needs in schools. For instance, teacher absenteeism appears to vary substantially across municipalities. Few parents report problems of teacher absenteeism in Bouarada, Cite El Khadra or Gaafour, for instance, particularly compared to Moknine, Mahdia or Taboulba. (See Figure 6.) Taken together in the school quality indices, such as those shown in Figure 7, we have an indication of overall school quality across municipalities. The best schools were found in Jbel Jloud and Sfax Sud and the worst in Bizerte Nord and Bizerte Sud. 45 bizerte nord bizerte sud bouarada cite khadhra gaafour .1 .2 .3 .4 0 jbel jloud ksour essaf la marsa mahdia menzel jemil .1 .2 .3 .4 0 moknine monastir sekiet eddayer sfax sud sfax ville .1 .2 .3 .4 0 siliana ville souassi tebolba .1 .2 .3 .4 0 mean of crowded_classroom mean of school_mult_shilft mean of teacher_absent mean of teacher_fav_lessons Graphs by Municipality Figure 6. Education Quality: Crowded classrooms, Teacher favoritism, Multi-shifts per school day, and Teacher absenteeism Figure 7. School Quality Index: Quality of schools across municipalities There are also important disparities in school quality related to socio-economic status. Figure 8 illustrates how school quality increases as socio-economic class scores increase. 46 Figure 8. Socio-economic Status Effects on School Quality The LGPI allows us to assess the difficulties that parents face within the school system and how parents attempt to address these difficulties. Indeed, it is a particularly useful tool in this regard because of our ability to gauge responses both from parents who have and have not found success in enrolling their children in school. That is, the LGPI avoids some of the selection bias problems inherent in the SDI and similar tools. In Tunisia, the survey found significant variation in the problems facing students and parents. Respondents faced the least amount of difficulty in Jbel Jloud, and the most in La Marsa (see Figure 9). Importantly, in Tunisia as a whole, the majority of parents facing difficulties do not seek help. Parents in Siliana reported asking for help the most, as opposed to Bizerte where they reported asking the least (see Figure 10). 47 Figure 9. Students facing difficulty in School, by Municipality Figure 10. Families Seeking Help, by Governorate The survey also sought to understand on whom parents placed the responsibility for education, with the vast majority holding the central government responsible. Moreover, when 48 asked on whom the responsibility for education should rest, parents again cited the central government as best suited for providing education. There appears to be widespread acceptance of Tunisia’s centralized education system. (See Figures 12 and 13.) Figure 12. Who is responsible for education? Figure 13. Who should be responsible for education? 49 Physical Security and Dispute Resolution The LGPI also found a great deal of variation in security and dispute resolution. For instance, when asked how safe they asked in various areas (e.g., the market place, their neighborhood at night, their house), people reported feeling much safer walking around in their neighborhood at night in Ksour Essaf than they did in Moknine. Along with assessing individual quantifiers of security such as feeling safe in one’s neighborhood, the LGPI also measured overall security scores. Siliana Ville reported the highest levels of security followed by Jbel Jloud and Sekiet Eddayer. The results show, once again, that Moknine suffers from insecurity far more than other areas. (See figures 14 and 15.) Figure 14. Neighborhood Safety: How safe do you feel walking around at night in your neighborhood? 50 Figure 15. Security Score by Municipality. The security score is composed of questions asking, “How safe do you feel…?” 1) In your home, 2) Walking in your neighborhood during the day, 3) Walking in your neighborhood at night, 4) At the local market, 5) At the local mosque, and 6) At the nearest school. Somewhat surprisingly, the survey did not reveal a significant gender differences in feelings of insecurity. As shown in Figure 16, women and men perceive the security of their local environment similarly. Moreover, where we do see differences, it is frequently the case that women feel more secure than the men in their area. 51 bizerte nord bizerte sud menzel jemil siliana ville bouarada 0 1 2 3 4 0 1 0 1 0 1 0 1 0 1 Female=1 Female=1 Female=1 Female=1 Female=1 gaafour monastir tebolba moknine souassi 0 1 2 3 4 0 1 0 1 0 1 0 1 0 1 Female=1 Female=1 Female=1 Female=1 Female=1 ksour essaf mahdia cite khadhra jbel jloud la marsa 0 1 2 3 4 0 1 0 1 0 1 0 1 0 1 Female=1 Female=1 Female=1 Female=1 Female=1 sekiet eddayer sfax sud sfax ville 0 1 2 3 4 0 1 0 1 0 1 Female=1 Female=1 Female=1 Graphs by Municipality2 Figure 16. Perceptions of Security by Gender and Municipality. There was, however, important variation in the extent to which people reported that their neighbors helped each other in keeping the area safe from crime. Tebolba reported the greatest amount of cooperation between neighbors, while the least amount of neighbor collaborations for safety was reported from Sfax Sud. (See figure 17.) bizerte.nord sfax.ville bizerte.sud 100 menzel.jemil sfax.sud sekiet.eddayer siliana.ville la.marsa bouarada 0 jbel.jloud gaafour cite.khadhra monastir mahdia tebolba ksour.essaf moknine souassi 52 Figure 17. Neighbor Safety Collaboration: Do you and your neighbors help each other with keeping the area safe from crime? The LGPI also explores attitudes towards police. Municipalities varied in their perceptions of police, with about half of the municipalities reporting that the police are moderately fair, and the other half reporting that they are not very fair. (See Figure 18.) Figure 18. Respondents’ Perception of Police Fairness. The LGPI revealed an interesting relationship between perceptions of the police and respondents’ socio-economic status. Figure 19 shows the socio-economic score against a variety of expectations of police officers. Surprisingly, those of lower socio-economic status appear to have a more favorable view of police officers. They are more likely to report the police as being fairer, more knowledgeable of the community, and more honest, compared to those of higher socio-economic status. 53 Figure 19. Attitudes toward Police Competencies and Socio-Economic Class Social Assistance and Welfare The module on social assistance and welfare examines individual municipalities for problems in social welfare provision as well as the sources of assistance that people turn to in case of need. The LGPI found important differences across municipalities with regard to whom people believed were assisting the poor in their areas. In Siliana Ville, Gaafour, Menzel Jemil, and Taboulba, for instance, the majority of respondents said that neighbors were an important source of assistance for those in need. This was not the case in Bizerte Nord, Bouarada, or Monastir. Far more respondents in Monastir, for example—as well as in Taboulba and Sakkiet Eddayer—believed that businessmen and organizations provided better sources of local assistance. Common among all municipalities, however, was the united belief that local council members were not important points of assistance for the poor. (See figure 20.) 54 bizerte nord bizerte sud menzel jemil siliana ville bouarada .2 .4 .6 .8 0 gaafour monastir tebolba moknine souassi .2 .4 .6 .8 0 ksour essaf mahdia cite khadhra jbel jloud la marsa .2 .4 .6 .8 0 sekiet eddayer sfax sud sfax ville .2 .4 .6 .8 0 Nobody Neighbors Local Iman & Relig. Org. CSO Tribe Powerful Families State Welfare Municipal council Business / business org. Trade Unions / prof. org. Graphs by Municipality2 Figure 20. Sources of Assistance for Those in Need. We find that the vast majority of respondents reported having no sources of assistance when faced with personal problems, such as shortages of food, clothing, or shelter. Those who did seek help turned to others in their personal networks: family, neighbors, and members of their tribe or extended family. Very few, however, turned to formal avenues for personal assistance, including the state welfare office. (See figure 21.) 55 bizerte nord bizerte sud menzel jemil siliana ville bouarada .2 .4 .6 .8 0 gaafour monastir tebolba moknine souassi .2 .4 .6 .8 0 ksour essaf mahdia cite khadhra jbel jloud la marsa .2 .4 .6 .8 0 sekiet eddayer sfax sud sfax ville .2 .4 .6 .8 0 Nobody Family Neighbors Tribe Graphs by Municipality2 Figure 21. Personal Sources of Assistance to Those in Need. Potential sources of assistance included 1) Friends, 2) Your family, 3) Neighbors, 4) Members of tribe or ethnic group, 5) Local imam, 6) Members of mosques or religious organizations, 7) Civil society and associations, 8) Trade unions or professional associations, 9) Leaders of powerful local families or clans, 10) Local council members, 11) Businessmen and organizations, 12) ANETI. Note: We only show results for items with more than 1% positive answers overall. Citizen-State Linkages The module on citizen-state linkages delves into perceptions and experiences with government corruption, transparency, and participation. Many of the questions in this module can be combined with those in previous modules, giving a more comprehensive picture of governance at the local level. The results reveal an important variation of individuals’ perceptions of corruption across municipalities. Respondents in Bouarada, La Marsa and Moknine tended to believe that local corruption was high, while those in Sfax Ville, Siliana Ville and Monastir were less likely to report that corruption exists. (See figure 22.) Within both the municipalities that perform poorly and those that perform well on the corruption index, it is important to note the presence of heavy socio-economic and geographic diversity. 56 Figure 22. Perceptions of Corruption by Municipality. Questions incorporated into the additive, normalized index included 1) Do many of the politicians in your community try to buy votes with gifts, money, or access to services? 2) Would you say there are none, a few, or many families eligible for government subsidies/assistance in this municipality that are unable to obtain it? 3) Do you believe that there households that receive subsidies but who are not poor? Would you say there are none, a few, or many such cases? 4) How useful are bribe payments to government officials to obtain a job in the government sector? 5) 9) When there is a corruption case going on in your governorate, how serious do you think the government deals with the case? Within municipalities, there is also an important disparity between the number of residents who perceive high levels of corruption and those who report experiences with it. For instance, residents of Sfax Ville do not view corruption as prevalent, but they report experiencing corruption frequently. On the other hand, Gaafour perceives corruption to be more prevalent than the reported experience with it would suggest. It is also noteworthy that there is significant variation within some of the governorates. Bizerte Nord and Bizerte Sud have similar experiences, but Sfax Ville and Sfax Sud are very distinct. 57 Figure 23. Experience with Corruption, by Municipality. Questions incorporated into the additive, normalized index included 1) Did any politician or representative offer you gifts, money, or access to services for your vote? 2) Agreement with “I paid a bribe in order to get a better treatment” for patients of municipal clinics or doctors, 3) Have you been a victim of corruption by a local government official in the past year? 4) Has anyone else in your immediate family been a victim of corruption in the past year? It is useful to note that awareness of public meetings and the perception of corruption may be positively correlated. As we see in Figure 24, the greater the percentage of respondents who say that the local council holds public meetings, the higher the perception of corruption. The drivers of this relationship are not entirely clear; it may be that in areas where the perception of corruption is higher, there is also a greater awareness of local meetings in which grievances can be aired. What is important, however, is that this relationship is masked when examining only the governorate level. As shown in Figure 25, when data is aggregated at the governorate level, the relationship between awareness of public meetings and perceptions of corruptions appears slightly negative. 58 Data agregated at Municipal Level 2 sfax ville 1 monastir moknine bouarada siliana ville mahdia souassi la marsa 0 tebolba gaafour bizerte nord sfax sud -1 ksour essaf cite khadhra sekiet eddayer bizerte sud -2 jbel jloud menzel jemil 0 .05 .1 .15 Municipality holds public meetings . Figure 24. Relationship between Awareness of Local Council Meetings and Perceptions of Corruption, Aggregated at the Municipal Level. Data agregated at Governorate Level .5 monastir siliana mahdia 0 sfax -.5 tunis -1 bizerte -1.5 .02 .04 .06 .08 Municipality holds public meetings Figure 25. Relationship between Awareness of Local Council Meetings and Perceptions of Corruption, Aggregated at the Governorate Level. 59 The LPGI measures the extent that economic, social, political and religious actors that play key roles in local governance varies by locality. Although the sample in Tunisia is limited at this stage to 18 delegations, we are able to detect clear local clusters where most actors are perceived to matter a great deal (Siliana Ville) as well as those where they matter very little (Bizerte Sud). (See Figure 26) Notice that even within the same governorate (Siliana), Bouarada and Siliana Ville lie at opposite ends of the distributions. This highlights the importance of designs that allow us to detect variation at the local level. 1 1 monastir siliana ville .5 Informal Social Influence ksour essaf gaafour citesekiet eddayer khadhra monastirsiliana ville tebolba .5 gaafour Informal Social Influence ksour essaf sekiet eddayer cite khadhra la marsa sfax ville 0 jbel jloud mahdia tebolba moknine sfax sud bizerte nord souassi menzel jemil jbel la marsa sfax ville mahdia 0 jloud moknine -.5 sfax sud souassi bizerte nord menzel jemil bizerte sud -.5 bouarada -1 bizerte sud bouarada -1 -1.5 -1 -.5 0 .5 1 -1.5 Formal Non-Economic Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities -1 -.5 0 .5 1 Informal social influence includes: INFL_1 INFL_2 INFL_8 Formal Economic Influence Fornal non-economic influence includes: (1) mosque & religious groups (2) CSOs (3) political parties Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Source: LGPI dataset, Tunisia 2015 Informal social influence includes: (1) regular citizens (2) powerful local families (3) clans or tribes Formal economic influence includes: (1) Trade Union (2) Chamber of commerce or bus. assoc. Source: LGPI dataset, Tunisia 2015 1 monastirsiliana ville .5 gaafour Informal Social Influence ksour essaf citesekiet eddayer khadhra tebolba la marsa sfax ville 0 jbel jloud mahdia moknine sfax bizerte souassi sud nordmenzel jemil -.5 bizerte sud bouarada -1 -1.5 -1 -.5 0 .5 1 Formal Non-Economic Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Informal social influence includes: (1) regular citizens (2) powerful local families (3) clans or tribes Fornal non-economic influence includes: (1) mosque & religious groups (2) CSOs (3) political parties Source: LGPI dataset, Tunisia 2015 Figure 26 Cross-Correlations between Key Measures of Social Influence There is evidence that these local level differences are related to variation in governance. We detect a negative relationship between the density of informal social ties and the perception of corruption in a given locality. It is crucial to keep in mind that these measures should be taken in the immediate geographical context of the respondents daily lives rather than at an (arbitrary) administrative level. Figure 27 displays the relationship between the extent to 60 which informal social actors—matter defined as relatives, powerful local families and clan/tribes—and the perception of corruption. We find that at the delegation level, those areas in which informal social influence is stronger report lower levels of perceived corruption. Importantly, as the Figure 27 illustrates, a relationship that is not detectable at the (higher) governorate level appears far more clear when the data is disaggrated at the municipal level. We recognize that with only six governorates included the survey, a ‘true’ signi ficant correlation that would be detected if all governorates were taken into account is unlikly to appear when the LGPI Tunisia is used at the governorate level. It is thus especially important to design and analyze data at the most appropriate level. The LGPI reveals an important contrast between answers to questions regarding the perception of corruption in the local area and the individuals’ experience with corruption. In delegations where local stakeholders (e.g., political parties, trade unions, powerful families) have a strong influence, respondents percieve a lower incidence of corruption; yet, the reported experience with corruption is constant across localities, regardless of the strength of local actors. These relationship of local non-economic actors with perceived corruption and experienced corruption are demonstrated in Figure 28. Non-economic actors include mosques and religious groups, CSOs, and political parties. We find a virtually flat line – indicating no apparent relationship – when we examine the relationship with experienced corruption. Yet, there is a negative slope and narrower confidence interval with regard to perceived corruption. One can question whether this indicates response bias, wherein individuals are less willing to report the perception of corruption in areas with strong local stakeholders, or whether it reflects a perception of security and order, in which those living in areas with strong local stakeholders perceive the sitaution to be better than those who live in areas without such stakeholders. More work needs to be done to adjudicate between these explanations. 61 2 1 TunisMonastir 0 Bizerte Siliana Mahdia Sfax -1 -2 -1 -.5 0 .5 1 Informal Social Influence 2 bouarada la marsa moknine 1 tebolba souassi bizerte nord gaafour 0 menzel jemil bizerte sud sfax sud jbel jloud sekiet eddayer cite khadhra 3 mahdia essaf ksour monastir -1 2 bouarada siliana ville la marsa sfax ville moknine 1 tebolba -2 souassi bizerte nord gaafour 0 -1 bizerte sud -.5 menzel jemil 0 .5 1 jbel jloud sfax sud Informalsekiet eddayer mahdiacite khadhra Social Influence ksour essaf monastir -3 -2 -1 Indices computed by MCA. Regression and sfax ville siliana Confidence Interval ville based on 18 Municipalities Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Informal social influence includes:(1) regular citizens (2) powerful local families (3) clans or tribes -1 -.5 Source: LGPI dataset, 0 Tunisia 2015 .5 1 Formal Legend Non-Economic Influence 95% CI Fitted values Bizerte Siliana Monastir Mahdia Tunis Sfax Indices computed by Figure 27 Informal MCA. Regression Social and Confidence Influence Interval and based on 18 Perceived Corruption at Municipalities the Governorate and Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment Municipal (3) corruption control (Delegation) (4) perceived failure Levels: to provide social An assitance Example when due of Aggregation Bias (5) providing assistance to those not in need Source: LGPI dataset, Tunisia 2015 62 3 2 sfax ville monastir 1 moknine siliana ville bouarada mahdia souassi la marsa 0 tebolba gaafour bizerte nord sfax sud -1 ksour essaf cite khadhra sekiet eddayer bizerte sud -2 jbel jloud menzel jemil -3 -1 -.5 0 .5 1 Formal Non-Economic Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Source: LGPI dataset, Tunisia 2015 3 2 bouarada la marsa moknine 1 tebolba souassi bizerte nord gaafour 0 menzel jemil bizerte sud jbel jloud sfax sud sekiet eddayer cite khadhra ksour essaf mahdia monastir -1 3 siliana ville sfax ville 2 bouarada -2 la marsa moknine 1 tebolba souassi bizerte nord gaafour -3 0 bizerte sud menzel jemil jbel jloud sfax sud sekiet eddayer cite mahdia khadhra ksour essaf monastir -3 -2 -1 -1 -.5 0 .5 1 sfax Non-Economic Formal ville siliana ville Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need -1 -.5 Source: LGPI 0 .5 dataset, Tunisia 2015 1 Formal Non-Economic Influence Legend 95% CI Fitted values Bizerte Siliana Monastir Mahdia Tunis Sfax Indices computed by Figure and Confidencethe 28. Contrasting MCA. Regression Relationship Interval between based on 18 Municipalities Non-Economic Actors on Perceived vs. Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment Experienced (3) corruption control Corruption (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Source: LGPI dataset, Tunisia 2015 63 However, as Figure 29 demonstrates, the relationship holds as well between the extent to which other stakeholders influence local governance and perceived corruption. Perceived corruption is also negatively related to the influence of economic stakeholders, which include trade unions and chambers of commerce and business associations. The same holds for informal social influences, which includes ordinary citizens, powerful local families, and clans and tribes. 3 3 2 2 bouarada bouarada 2 3 Perception Corruption Perception la marsa la marsa moknine moknine 1 1 tebolba tebolba souassi souassi bizerte nord gaafour bizerte nord gaafour 0 0 menzel jemil menzel jemil 1 Corruption bizerte sud jbel jloud bizerte sud jbel jloud sfax sud sekiet eddayer sfax sud sekiet eddayer cite khadhra cite khadhra -3 Perception ksour essaf mahdia ksour essaf mahdia monastir monastir -1 -1 bouarada siliana ville siliana ville sfax ville sfax ville la marsa moknine -2 -2 tebolba souassi bizerte nord gaafour 0 menzel jemil -3 bizerte sud Corruption jbel jloud sfax sud sekiet eddayer cite mahdia khadhra -1 -.5 ksour essaf 0 monastir .5 1 -3 -2 -1 -1 -.5 0 .5 1 Formal Non-Economic Influence Formal Economic Influence siliana ville Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities sfax ville Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities (3) corruption control (4) perceived failure to provide social assitance when due Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (5) providing assistance to those not in need (3) corruption control (4) perceived failure to provide social assitance when due Source: LGPI dataset, Tunisia 2015 (5) providing assistance to those not in need Formal economic influence includes: (1) Trade Union (2) Chamber of commerce or bus. assoc. -1 -.5 0 2015 Source: LGPI dataset, Tunisia .5 1 3 Formal Non-Economic Influence Legend 2 Corruption Perception bouarada la marsa moknine 95% CI Fitted values 1 tebolba souassi bizerte nord gaafour Bizerte Siliana 0 menzel jemil bizerte sud sfax sud jbel jloud mahdia sekiet eddayer cite khadhra essaf ksour monastir Monastir Mahdia -1 sfax ville siliana ville Tunis Sfax -2 Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment -3 -1 -.5 0 .5 corruption control (3) 1 (4) perceived failure to provide social assitance when due Informal Social Influence (5) providing assistance to those not in need Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Source: LGPI dataset, Tunisia 2015 Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Informal social influence includes:(1) regular citizens (2) powerful local families (3) clans or tribes Source: LGPI dataset, Tunisia 2015 Figure 29. Perception of Corruption and Economic, Non-Economic and Informal Forms of Local Influence Importantly we also find a similar relationship if we parse local influences by the extent to which they reflect vertical relations (e.g., clientelist networks driven by political parties, trade unions or powerful families) or horizontal ones (e.g., level, mutually reinforcing relationships between ordinary citizens and clans and tribes.) In both cases, the negative relationship between perceived corruption and social ties remains. (See Figure 30.) 64 3 2 bouarada la marsa moknine 1 tebolba souassi bizerte nord gaafour 0 menzel jemil bizerte sud jbel jloud sfax sud sekiet eddayer cite khadhra mahdia ksour essaf monastir -1 sfax ville siliana ville -2 -3 -1.5 -1 -.5 0 .5 Vertical Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Vertical influence includes: (1) powerful families (2) trade unions (3) political parties Source: LGPI dataset, Tunisia 2015 3 2 bouarada la marsa moknine 1 tebolba souassi bizerte nord gaafour 0 menzel jemil bizerte sud jbel jloud sfax sud sekiet eddayer cite khadhra mahdia ksour essaf monastir -1 sfax ville siliana ville 3 -2 2 bouarada la marsa -3 moknine 1 tebolba souassi -1.5 bizerte -1nord -.5 gaafour 0 .5 0 bizerte sud menzel jemil jbel jloud sfax sud citeHorizontal sekiet Influence eddayer khadhra ksour mahdia essaf monastir Interval based on 18 Municipalities Indices computed by MCA. Regression and Confidence Corruption perception includes: (1) vote-buying (2) bribes obtain gov. employment to ville siliana sfax ville (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Horizontal Influence includes: (1) relatives (2) tribe) Source: LGPI dataset, Tunisia 2015 -1 -.5 0 .5 1 Formal Non-Economic Influence Legend 95% CI Fitted values Bizerte Siliana Monastir Mahdia Tunis Sfax Indices computed by Figure 30. Perception MCA. Regression and Confidence of Corruption Interval and Horizontal based on 18 Municipalities and Vertical Forms of Local Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control Influence (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Source: LGPI dataset, Tunisia 2015 65 The LGPI also reveals a relationship between the reported necessity of wasta to get a government job and the prevalence and need for informal payments. As shown in Figure 31, there is a positive correlation between the use of wasta and perceived corruption. Once again, there is no apparent relationship with experienced corruption. The LGPI pilot study also suggests an interesting relationship between the social composition of localities and citizen engagement. In Tunisia, we find a negative relationship between attendance at public meetings and the strength of stakeholders, whether defined by the type of stakeholder or vertical influence. (See Figure 32.) This suggests that in areas where stakeholders are powerful, people do not appear compelled to attend them. This occurs despite the fact that meetings are held in such areas; indeed, the data reflect a slightly positive relationship between knowing that meetings are held and the power of stakeholders. The results have important implications for programs designed around participatory processes and citizen engagement. We should expect their impact to vary according to the social composition of the locality. 2 bouarada la marsa moknine 1 tebolba souassi gaafour bizerte nord 0 menzel jemil bizerte sud jbel jloud sfax sud sekiet eddayer cite khadhra mahdia ksour essaf monastir -1 siliana ville sfax ville -2 2.4 2.6 2.8 3 3.2 3.4 Perceived importance of wasta for getting a government job Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Source: LGPI dataset, Tunisia 2015 66 2 sfax ville monastir 1 siliana ville moknine mahdia bouarada souassi la marsa 0 gaafour tebolba bizerte nord sfax sud -1 ksour essaf cite khadhra sekiet eddayer bizerte sud -2 jbel jloud menzel jemil 3 2.4 2.6 2.8 3 3.2 3.4 2 bouarada la marsa moknine Perceived importance of wasta for getting a government job 1 tebolba souassi bizerte nord gaafour 0 bizerte sud Indices computed by menzel jemil Factor jbel sfax sud Analysis jloud (pcf)eddayer citesekiet khadhra Regression and Confidence ksour essaf Interval mahdia based on 18 Municipalities. monastir Corruption experience includes: (1) Vote buying attempt on respondent sfax ville siliana ville (2) Being personally a victim of corruption by local government officials (3) Relative is a victim of corruption by local government officials (4) Having to bribe doctors at the local clinic (5) Bribing a policeman seeking to fine you -1 -.5 0dataset, Tunisia 2015 Source: LGPI .5 1 Formal Non-Economic Influence Legend 95% CI Fitted values Bizerte Siliana Monastir Mahdia Tunis Sfax Indices computed by Figure and Confidenceand 31. Corruption MCA. Regression the Interval Use based ofMunicipalities Wasta on 18 Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Source: LGPI dataset, Tunisia 2015 67 1.4 1.2 1.2 Attend public meetings Attend public meetings 1 ksour jbel bouarada bizerte souassi sud tebolba jloud essaf bizerte nord sfax sud gaafour moknine 1 bouarada bizerte sud souassi bizerte ksour jbel tebolba sfax jloud essaf sud nord gaafour moknine .8 cite khadhra .8 cite khadhra la marsa mahdia monastir mahdia la marsa monastir .6 .6 siliana ville siliana ville .4 .4 -1 -.5 0 .5 1 -1 -.5 0 .5 1 Formal Non-Economic Influence Formal Economic Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Formal non-economic influence includes: (1) mosque & relig. groups (2) CSOs (3) political parties Formal economic influence includes: (1) Trade Union (2) Chamber of commerce or bus. assoc. Source: LGPI dataset, Tunisia 2015 Source: LGPI dataset, Tunisia 2015 1.4 1.4 1.2 1.2 Attend public meetings Attend public meetings 3 1 1 bouarada bizerte sud bizerte souassi sfax sud moknine nordjbel jloud tebolba gaafour ksour essaf bouarada bizerte jbelsud jloud sfax bizerte souassi sud nord tebolba ksour essaf gaafourmoknine 2 bouarada la marsa .8 .8 moknine 1 tebolba cite khadhra cite khadhra souassi mahdia bizerte nordla marsa gaafour monastir la marsa mahdia monastir 0 bizerte sud menzel jemil .6 .6 jbel jloud sfax sud sekiet eddayer mahdia cite khadhra ksour essaf monastir -3 -2 -1 siliana ville siliana ville siliana ville .4 .4 sfax ville -1 -.5 0 .5 1 -1 -.5 0 .5 1 Informal Social Influence Vertical Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Informal social influence includes: (1) regular citizens (2) powerful local families (3) clans or tribes Vertical influence includes: Source: LGPI dataset, Tunisia 2015 Source: LGPI dataset, Tunisia 2015 -1 -.5 0 .5 1 Formal Non-Economic Influence Legend 95% CI Fitted values Bizerte Siliana Monastir Mahdia Tunis Sfax Indices computedFigure 32. Civic by MCA. Regression and engagement and Confidence Interval Social based Influence on 18 Municipalities Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need dataset, the Source: LGPIFinally, pilot Tunisia study 2015 in Tunisia uncovers interesting potential relationships between local social organization and the quality of service delivery. The study suggests that, if anything, there is a negative relationship between the significance of local stakeholders (here, economic stakeholders) and the quality of education. The same holds true with the relationship between education quality and density of social ties. (See Figure 33) One potential explanation for this relationship may be the dominance of public education in Tunisia and its centralized provision. This does not fully explain why we observe a negative relationship, rather than a non-relationship, but it may help explain disparities in the service quality of services in education versus health within localities. 68 .4 bouarada jbel jloud .2 ksour essaf bizerte sud cite sekiet khadhraeddayer menzel jemil bizerte nord sfax sud 0 gaafour souassi monastir siliana ville la marsa sfax ville mahdia -.2 tebolba -.4 -.6 moknine -.5 0 .5 Formal Economic Influence Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Formal economic influence includes: (1) Trade Union (2) Chamber of commerce or bus. assoc. Source: LGPI dataset, Tunisia 2015 sfax sud .7 cite khadhra jbel jloud la marsa .6 sfax ville bizerte nord .5 sekiet eddayer souassisud menzel jemilbizerte .4 monastir mahdia 3 siliana ville bouarada gaafour .3 ksour essaf 2 bouarada la marsa moknine tebolba moknine 1 tebolba .2 souassi bizerte nord gaafour 0 bizerte sud menzel jemil jbel jloud sfax sud sekiet eddayer cite mahdia khadhra 1.6 1.8 essaf ksour 2 monastir 2.2 2.4 2.6 -3 -2 -1 sfax ville siliana ville Extent to which people know each other in this area Indices computed by MCA. Regression and Confidence Interval based on 18 Municipalities Source: LGPI dataset, Tunisia 2015 -1 -.5 0 .5 1 Formal Non-Economic Influence Legend 95% CI Fitted values Bizerte Siliana Monastir Mahdia Tunis Sfax Indices computed by Figure 33. Quality MCA. Regression of Education, and Confidence Influence Interval based of Economic on 18 Municipalities Stakeholders and Social Density Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need We observe Source: LGPI dataset, very different a 2015 Tunisia relationship between local social influences and health quality. As shown in Figure 34, for instance, the more people know each-other, the better the quality of 69 health services in clinics and doctors’ offices. One explanation may be that local clients are able to put greater pressure on health care providers than they are on education providers, given the higher level of private provision and consequently less dominant state-centralized provision in the health sector. More remains to be done to examine these relationships, both by expanding the sample size and combining the LGPI with other tools (as discussed previously.) bouarada .5 bizerte sud mahdia gaafour siliana ville monastir bizerte nord menzel jemil ksour essaf 0 tebolba moknine souassi sekiet eddayer la marsa sfax ville cite khadhra jbel jloud -.5 sfax sud 3 -1 2 bouarada 1.6 1.8la marsa 2moknine 2.2 2.4 2.6 1 tebolba souassi bizerte Extent nord tomenzel which jemil peoplegaafour know each other in this area 0 bizerte sud jbel jloud sfax sud sekiet eddayer cite mahdia khadhra ksour essaf monastir Interval based on 18 Municipalities -3 -2 -1 Indices computed by MCA. Regression and Confidence silianaof HQI includes: (1) clean restrooms in the clinic (2) presence a doctor at the clinic ville sfax ville (3) reasonable expenses (4) reasonable waitng period at clinic (5) patients treated with respect (6) necessary medicines readily available Source: LGPI dataset, Tunisia 2015 -1 -.5 0 .5 1 Formal Non-Economic Influence Legend 95% CI Fitted values Bizerte Siliana Monastir Mahdia Tunis Sfax Indices computed by Figure 34. Health MCA. Regression Quality and Confidence and Interval Social based on 18Influence Municipalities Corruption perception includes: (1) vote-buying (2) bribes to obtain gov. employment (3) corruption control (4) perceived failure to provide social assitance when due (5) providing assistance to those not in need Source: LGPI dataset, Tunisia 2015 In short, the LGPI provides an important tool for examining local level variation in governance and service delivery. Even where systems are highly centralized, there are significant variations in implementation and needs at the local level. The tool allows us to assess these needs, to better gauge local drivers of governance, and to develop locally appropriate programming. There are often no straightforward answers to what troubles local communities and how best to serve those concerns, but the LGPI can help to pinpoint problems, identify the major players engaged in developing solutions, and create better informed and more effective policies. 70 ANNEX 3: Dissemination Guidelines5 Many tools exist for disseminating survey findings and communicating messages clearly and in a simplified manner to the public and the broader audience for information and future use. Among the many tools are: o Holding Public Awareness Campaigns Running a public awareness campaign requires a conscious effort to define the problem, shape a message and get the message to the community in an effective inclusive manner. Common methods for campaigns include Media, advertisements, internet, public institutions, and direct outreach. An effective public awareness campaign will have:  A well-defined issue;  A clear message;  An identified target audience;  Multiple communications strategies- including use of volunteers for peer-to-peer  o Holding small group discussions A small group discussion is a conversation between a small group of people on a specific topic. It is different from a meeting because it is moderated or managed by a person who asks specific questions on a topic. Small group discussions allow citizens to find out a lot of information and different perspectives from the municipality in a single meeting. For effective discussions, it is important to:  Have a structured discussion, with well-planned questions that will guide the discussion;  Reach beyond the usual circle of colleagues and actors to reach more vulnerable and marginalized groups;  Take time to test the small group questionnaire before formally implementing the discussion. o Developing interactive websites that allowed citizens, local civil society organizations, and others to access reports, review the methodology and instruments, and undertake simple data manipulation 5 For a broader discussion on the tools and guidelines for enhancing communication with the public and disseminating results, please see “Resilient Municipalities: A Resource for Countries Affected by the Syria Crisis. Module 4: Increasing two-way communication” UNDP. 71 The Vietnam PAPI results were extensively promoted through various social media outlets, such as Facebook and twitter, and through an interactive website to help citizens review and understand the results for future results. Below are some snapshots captured from the website (http://papi.vn/en/): 72 73 74 Reference Innovation for Successful Societies. (2014). Measuring Citizen Experiences: Conducting a Social Audit in Vietnam, 2009-2013. Princeton, NJ: Princeton University. Recanatini, F. (2011). Country-Specific Diagnostic Assessments: An Alternative Approach for Policy Reform. Washington, DC: The World Bank Group. UNDP (forthcoming). Resilient Municipalities: A Resource for Countries Affected by the Syria Crisis. Module 4: Increasing Two-Way Communication. New York: UNDP. World Bank (forthcoming). Raising the Bar: Building Citizen Feedback Loops for Service Delivery in Tunisia. Policy Note. Washington, DC: The World Bank Group. 75