ࡱ> ,.)*+ bjbj İaa~WI?  \\\D Ll lH""^6#6#6#$$$HHHHHHH$JVMCH\W7$$W7W7CH Tt6#6#XHCCCW7\6#\6#HCW7HCCr)GG 6#A<G HnH0HG"N>"NG"N\G($*CC/2X$$$CHCHoCd$$$HW7W7W7W7"N$$$$$$$$$ :  Poverty and Democratic Participation Reconsidered: Evidence from the Local Level in India Anirudh Krishna, Duke University* Abstract Conventional wisdom suggests that a direct relationship exists between greater wealth and higher democratic participation. An original database compiled for 53 local governments in two states of northern India, including interviews with over 2,000 village residents, shows, however, that information, education and gender matter much more for who participates in democracy, while wealth and social status matter relatively little. Policies that enable people to educate and inform themselves better should help raise poor peoples stakes in democracy. Simply crafting democracy from above does not ensure that ordinary citizens, especially poorer ones, can participate in and effectively influence government decisions. Evidence from around the world shows that even in richer and more longstanding democracies poorer citizens participate less often and less vigorously than their wealthier counterparts. In Third World democracies, where the rate of poverty is much higher as high as 68 percent in Zambia and 70 percent in Madagascar participation and influence may be particularly skewed, resulting in a much narrower base of support for democracy. It is important to examine how poor peoples stakes in democracy can be enlarged in such contexts of poverty. The harms of poverty inhibit the political participation of some citizens. It is necessary, however, to investigate more closely what exactly these harms are that inhibit poor people from participating more actively. Is it poverty itself, i.e., lack of income and wealth, that constitutes the root cause of low participation, or are some other individual attributes, usually associated with poverty, and not poverty itself, more readily responsible for participation at low rates? Diamond contends, for example, that participation in democracy can be high even at low levels of economic development if [some] crucial mediating variables are present. Studies undertaken in different parts of the world have shown how literacy can be one such mediating variable. Bratton and Mattes find within six Sub-Saharan countries that education has highly positive effects on peoples attitudes toward democracy. Finkel shows how civic education programs in South Africa and the Dominican Republic have enabled citizens to engage more actively and effectively with democracy in these countries. Results from other studies also attest to the positive effects that the spread of education can have for raising peoples stakes in democracy. Information can be similarly critical for entrenching democracy better. Citizens without information can rarely formulate interests clearly, and they know little about appropriate institutional pathways for expressing these interests, so democracy without information can be exclusionary in effect. However, while the association between democracy and information has been considerably investigated within industrialized democracies, particularly the United States, this link has not been equally well explored in democracies populated by larger numbers of poorer people. In this paper, I look at the respective contributions that information, wealth, education and some other variables have for participation in rural local governments within two states of India. Local governments serve in terms of the classification provided by Eckstein as a least likely case for examining the effects upon participation of information and education in relation to wealth and social status. While the devolution and democratization of power at the local level is widely expected to enhance the effectiveness, responsiveness and accountability of government, grave doubts persist concerning its effects on equitable participation and equal access. The most important caveatis that decentralization can bolster the power of local elites. Numerous studies bear out this fear, suggesting that if barriers to participation by poor people exist at any level of government, they are most likely to occur at the level of decentralized local governments. If information and education can help, therefore, to widen participation at this level of government, they should have at least as much value for participation by poor people at higher levels of government. Location and Methodology Rajasthan and Madhya Pradesh the two Indian states selected for this investigation provide a useful site for examining the separate effects of information and education, on the one hand, and wealth and social status, on the other hand. In general, these two sets of variables have a fairly strong and significant positive correlation, and it becomes hard to isolate their separate effects. In villages of Rajasthan and Madhya Pradesh, however, information and education have become relatively less intertwined with wealth and social status. Literacy has expanded faster in these states compared to all other states in India. Between 1991 and 2001, for instance, an additional 22 percent of the population became literate in Rajasthan and an additional 20 percent became literate in Madhya Pradesh, which are, respectively, the highest and second highest rates of literacy growth in this country (the average figure for all Indian states is 14 percent). While poverty has also declined at the same time, it has not fallen nearly as fast as literacy has increased. Percentage of population in poverty fell from 34 percent to 27 percent in Rajasthan between 1983 and 1993, and in Madhya Pradesh it fell from 49 percent to 42 percent, i.e., a reduction of 7 percent was observed in both states, which is less than one-third of the rate of increase in literacy. A considerable divergence between education and wealth has arisen consequently within villages of these states. While poverty is still widespread in both states, relatively poorer citizens have benefited from the expansion of literacy together with relatively richer ones. These trends are well represented within the 53 villages selected for this study. While only 20 percent of villagers aged 55 years and older have five or more years of school education, as many as 74 percent of villagers aged 18-25 years have achieved at least this level of education, and younger villagers belonging to relatively poorer and lower-status households are functionally literate in numbers that are only marginally lower than the corresponding numbers for richer and higher status households. The correlation coefficient between wealth and education is only 0.14, indicating that the association of education with participation is not confounded to any great degree by the presence of a wealth variable within the framework of analysis. In addition to wealth and education, religion and caste are also important to examine in the Indian context. Rajasthan and Madhya Pradesh have significant proportions of Muslims as well as Hindus, and they also have a full range of caste groups, including large numbers of Scheduled Castes, Scheduled Tribes, and Other Backward Castes. Scheduled Caste refers to the former untouchables, and Scheduled Tribe to what are, loosely speaking, Indias aborigines. These categories are recognized by Indias constitution, which provides schedules listing specific castes and tribes as Scheduled Caste and Scheduled Tribe, respectively. Other Backward Caste is a more recent administrative listing, and it refers to caste groupings that are neither upper caste nor listed in the schedules for Scheduled Castes and Scheduled Tribe. Panchayats (rural local governments) were selected as the sites for this inquiry, and an original database was compiled through field inquiries conducted between May 1999 and August 2000 in a group of 53 villages of Rajasthan and Madhya Pradesh. The selected villages are located within three districts of Rajasthan Ajmer, Bhilwara and Dungarpur and three districts of Madhya Pradesh (Neemuch, Mandsaur and Ujjain). These districts were selected because of three reasons. First, these districts fall within the middle range in each state, i.e., district averages are close to the average for each state in terms of progress in literacy and poverty reduction. Second, these districts are also located adjacent to each other and they share significant cultural and historical characteristics. Third, the choice of these districts allows a more complete comparison of different population characteristics. Significant numbers of Muslims reside within Ajmer district; Scheduled Tribes are concentrated in large numbers within Dungarpur and they are also found, although in smaller numbers, in Mandsaur and Neemuch; and Scheduled Castes as well as Other Backward Castes are found within villages of all six districts. In order to test whether and how well different village characteristics were associated with participation rates, a varied mix of villages was purposively selected, including some that sit astride major roads and others that are relatively hard to access, single-caste-dominant villages as well as villages with mixed caste compositions, and larger villages together with smaller ones. The structure and functions of panchayats need to be explained briefly. There is one position of sarpanch (elected chief of a panchayat), one position of up-sarpanch (deputy chief) and some other elected positions (panch) within each panchayat. Elections to all positions are required to be held every five years, and elected officials are responsible for providing a variety of services and benefits at the local level. National legislation enacted in the early 1990s further strengthened the role played by panchayats: elections to offices of local governments have been held regularly since that time, and the position of sarpanch is reserved in some villages for women, Scheduled Castes, Scheduled Tribes, or Other Backward Castes. Villages where each such type of reservation was in force were selected, and other villages were also selected where this position is not reserved and remains open for all contestants. I worked together with a group of 12 investigators, equally men and women, who are residents of other villages of this region, and together we interviewed a stratified random sample of men and women in each of the 53 study villages. We listed all households in each village, and we classified them into distinct social groups, identified on the basis of caste, religion and tribe. Each Hindu caste group in a village was considered to constitute a separate social group. In addition, Muslims and Scheduled Tribes were considered separate social groups. There are no households in these villages that belong to more than one of these categories, so this procedure ensured that every household in the village was counted once and only once. We selected a random sample of households from each social group in proportion to this groups share in village population. A total of 2,013 households were selected in this manner, and women and men were selected for interviewing alternately from this list of households. We interviewed these 2,013 individuals using a list of questions that had been pilot tested previously. Separately, we also interviewed all 315 persons in these villages who currently hold or who have at some previous time held official positions in panchayats. In the analysis that follows, I look first at participation by ordinary villagers in various activities connected with constituting panchayats and influencing their decisions. Next, I look at participation within council meetings by elected officials. Finally, I examine whether affirmative action or other changes in structure can help raise participation rates among poorer and more disadvantaged citizens or whether it would be more fruitful to invest, instead, in individuals agency, for instance, through expanding information and education. Who Participates? A series of questions related to activities such as voting, campaigning, and attending public meetings was administered to the random sample of 2,013 villagers in 53 selected villages. These questions cover different aspects of participation that villagers here utilize for influencing their village local governments. Self-reported participation in voting to constitute village local governments was uniformly high among all sections of villagers in the 53 selected villages. Ninety five percent of 2,013 villagers interviewed for this exercise reported that they had voted in all or most elections to local government that were held since they had become old enough to vote. Different social groups reported nearly equally high figures in this regard. A number of different reasons can possibly account for these high voting figures. Villagers whom I met mentioned quite often that they had voted at least once for reasons of social solidarity or because they wished to avoid later tension or conflict with the candidates. Such high self-reported voting figures are not reliable indicators, therefore, of high levels of broad political participation or active involvement in the processes of local self-government. In any event, elections occur infrequently, every five years or less often, so they are hardly good instruments for maintaining constant participation and exercising regular control. We need to consider, in addition, some other activities such as campaigning, attending rallies and meetings, and contacting public representatives that are associated with participating on a more regular basis. Participation in political activity falls off sharply when we consider these other activities, as could be expected from similar analyses conducted elsewhere in developing and industrialized countries. Forty-three percent of villagers reported campaigning on behalf of any candidate (or discussing electoral issues with others) during the previous election to the local government; 33 percent attended any rally or meeting on behalf of a candidate; 35 percent made contact with a local government official during the past year; and 11 percent contacted higher-level officials within the structure of rural local government. Factor analysis shows that it is nearly always the same individuals who participate highly in all of these activities. Individuals responses to six different survey questions (related to campaigning, contacting, and attending meetings and rallies) are highly correlated with each other and all of these responses loaded highly on a single common factor. Some villagers are high participators: they campaign vigorously for political candidates, they influence other villagers choices at election time, they attend meetings regularly, and they are in contact frequently with local government officials. But there are some other villagers who are low participators: they tend to be relatively inactive, in general, taking part rarely if at all in different political activities associated with constituting and influencing local governments. What distinguishes more politically active from less active villagers? Do elites capture these processes of decision making, i.e., does high wealth and higher social status mark off highly participating individuals from other villagers, or are there some other traits that correlate more closely with higher participation scores? What roles do information and education play in determining who participates in democracy in these contexts? In order to address these questions, a 100-point Index of Political Activity was constructed by taking an adjusted sum of scores of the six separate survey items. An individuals score on the Index of Political Activity reflects the omnibus effect of diverse modes of participating in local government activities. The higher this score, the greater is an individuals overall level of participation. The most active individuals achieve the maximum score of one hundred points on this index, while the least active individuals have a score of zero points. Table 1 reports the results of regression analysis undertaken to determine how closely different characteristics of individuals are related with their levels of participation in local government. The dependent variable is the Index of Political Activity measured on a 100-point scale. Regression coefficients relate, thus, to the number of percentage points by which the dependent variable increases or decreases as a result of a one-unit change in the respective independent variable. The following independent variables were considered. The variable Gender was coded as one for women and as zero for men. A total of 1,008 men and 1,005 women were interviewed. The variable Religion was coded as zero for Hindus and one for Muslims. No other religion was reported by any respondent. A total of 1,870 Hindus and 141 Muslims were interviewed. In order to assess the effect of caste, separate dummy variables were used for Scheduled Caste, Scheduled Tribe, and Other Backward Caste. The sample of villagers interviewed included 301 Scheduled Castes, 434 Scheduled Tribes, 829 Other Backward Castes, and 447 upper castes. The variable Landholding serves as a measure of wealth. It is measured in terms of hectares of agricultural land owned by the family of the respondent. Income derived from agricultural activities is not taxed in India; consequently, no records of income are available for villagers, who are nearly all dependent upon agriculture for their main source of livelihood. However, land records do exist, and the land records in these two states, particularly Rajasthan, are among the best maintained in India. I also considered other measures of wealth, such as cattle ownership and quality of house construction (brick v. mud). Both these other measures of wealth are closely correlated with the landholding variable, however, suggesting that landholding is not an inappropriate measure of overall wealth in these rural and agrarian contexts. Mean landholding is 1.58 hectares, and standard deviation is 3.03. Education is measured by the number of years that a respondent attended school. On average, these villagers attended schools for 3.3 years. Access to Information is a composite variable that measures the number of information sources, out of seven, that are regularly accessed by a respondent. These seven sources were identified during the initial pilot test as the ones that different villagers consult for politically relevant information, and they include family members, neighbors, leaders, prominent persons in the village, radio, television, and newspapers. The variable Access to Information was constructed by adding up individual scores for the seven different sources. Scores on this variable can vary thus from zero to seven. The observed mean score is four, with a standard deviation of 1.62. Twenty-five individuals among the total of 2,013 had the lowest possible score of zero points, 394 individuals scored one or two points, and another 564 had the highest observed score of six points. No one scored the maximum of seven points. The seven separate sources of information considered here are not all equal in terms of reducing transactions costs associated with participating in political activities. For instance, it could be argued that listening to the radio or reading newspapers can equip an individual much better for dealing with diverse political activities. However, considering radio or television to the exclusion of neighbors and leaders would result in excluding those sources of information that are most often consulted by poorer villagers, thereby biasing the sample in favor of wealthier villagers. Similarly, considering newspapers and not the other sources would have biased the sample in favor of the literate. It seemed preferable, therefore, to include all seven sources within the omnibus variable, Access to Information. While looking below at the results of regression analysis, however, I will also decompose this variable into its individual components. Finally, the variable State is coded as zero for Madhya Pradesh and as one for Rajasthan. A total of 953 respondents belong to Madhya Pradesh, and 1,060 are from Rajasthan. -- Table 1 here -- Wealth is not significantly related with participation in these contexts. The coefficient for the landholding variable did not achieve any significance in any alternative specification of the regression model. The coefficients for religion and caste status are also not significant in association with participation scores. In contrast, education, information and gender are all significantly associated with high participation scores. Because information and education are not closely connected with wealth or social status within these contexts, we can more easily study their separate effects upon participation. Gender is a very significant aspect of participation in these contexts. Controlling for other factors, women score, on average, 24 percentage points lower than men on the 100-point scale of political participation. Social status, as represented by caste, is not significant for this analysis. Among the different caste groups, only Scheduled Tribes have a significantly different participation score. On average, Scheduled Tribes score about six percentage points lower than upper castes, which is a significant difference but is still not as considerable as the difference separating women from men. Scheduled Castes (the former untouchables) and Other Backward Castes do not have significantly lower scores for political activity. Indeed, coefficients for the associated dummy variables are positive showing higher participation scores on average but these coefficients are not statistically significant, indicating that Scheduled Caste and Other Backward Caste villagers have participation scores that are not, on average, any different from those of other villagers. These results can be understood in terms of the particular histories of these different groups. Scheduled Tribes suffer from a location disadvantage that contributes toward and exacerbates the historical deprivations suffered by these groups. Also known as forest dwellers in local parlance, Scheduled Tribes live in villages that are fairly remote and previously forested. Scheduled Castes and Other Backward Castes, on the other hand, live mostly in settlements that are more within the mainstreams of economic and political life in these states. In order to assess the combined effect of gender and caste, I also considered some interactive variables. In particular, I examined whether being Scheduled Tribe and female is likely to reduce participation even further than would be suggested by adding together the individual coefficients of these two variables. The resulting interactive variable (Scheduled Tribe times gender) was significant (at the 0.05 level) and the size of its coefficient was also large (-6.45). Notably, the variable, Scheduled Tribe, loses significance when this interactive variable is brought within the analysis, although the overall gender variable continues to remain significant and high, implying that participation in political activity is low not so much among all Scheduled Tribes, but particularly among women of this social group. While women participate less than men on average in all social groups, participation levels among Scheduled Tribe women are lower by an additional 6.45 percentage points. To further verify the conclusion about the non-significance of caste, I also looked separately at some individual castes. Within the group of upper castes, Brahmins are the traditionally literate castes, and Rajputs have been historically politically dominant in this region (often together with Brahmins). However, neither dummy variable, Brahmin or Rajput, was significant for this analysis. Education is consistently positively associated with higher participation scores. Everything else remaining the same, each additional year of school education adds, on average, 0.24 additional percentage points to an individuals participation score. Access to Information also has a significant and substantial association with participation levels. The addition of one more source of information to a persons repertoire (among the seven considered here) tends to increase participation scores by more than five percentage points. Each of the seven different information sources considered here is also significant when related individually with the Index of Political Activity. These seven separate coefficients are 1.8 (for neighbors), 2.4 (household members), 3.6 (village assembly), 4.1 (village leaders), 5.8 (radio), and 6.2 (television). The coefficient for the newspapers dummy is even larger (8.6). However, the variable Education is no longer significant in analysis when this dummy for newspapers is included, indicating that the dummy variable for newspapers captures the combined effect of education and newspaper readership. It is interesting to find that all seven sources of information are significantly associated with higher participation, including sources such as neighbors and other villagers that are consulted more often by relatively poorer villagers. Sources that individuals actively seek out for information (such as radio, television and the village assembly) are significant along with other sources (such as household members and neighbors) from which information is obtained in the usual course. Age, family size, and religion are not significantly associated with higher political activity scores. Older villagers, villagers belonging to larger households, and Muslim villagers do not participate to any smaller or greater extent, on average, compared to other villagers. Information and education make the critical difference thus, together with gender, in differentiating those who participate actively in local government activities from those who do not. These conclusions were reinforced when the regression model was run once again but this time without including education and information among the independent variables. Previously non-significant variables, including age, length of residence in the village, and most notably, caste, gained significance this time around. It is likely, this analysis suggests, that in situations where education and information are not included within the analysis of participation, variables such as these will tend to capture their independent effects. How can villagers who do not know where to go and what to do take part [in democratic decision making at the local level]? asked one quite active villager whom I interviewed. If they do not know the rules, how can they get any benefit from these rules? Yes, we can have lots of rules [about democracy], but unless people know about these rules and utilize [this knowledge], these rules are of no use to them. Information and education are critical in this regard, and they trump the tendency to dominate that wealth and social status quite often help to confer. This finding runs counter to conventional wisdom about who participates more actively. It is similar, however, to Yogendra Yadavs observation: The textbook rule about political participation is that the higher you are in the social hierarchy, the greater the chances of your participating in political activity, including voting. Contemporary India is[an] exception to this rule. It should be interesting to investigate whether information and education are similarly responsible for enhancing participation by poorer people in other parts of India and the developing world. Participation by Elected Representatives A separate survey instrument was administered to all 315 past and present representatives resident in these 53 villages in order to test separately whether richer and higher status villagers tended to dominate proceedings and decisions in meetings of panchayat officials. Six separate questions were included within this questionnaire relating, respectively, to frequency of attendance at meetings of panchayat officials (two questions), length of time spent at each meeting, degree of involvement (two questions), and self-perceived influence upon the final decision (compared with other members). While individually these questions may not correspond closely to the actual influence exercised by a representative, collectively, they provide a reasonably good approximation. Representatives who have high participation scores in respect of any of these six questions also tended to have high scores with respect to each of the other five questions. I put together a summary measure of participation for representatives, akin to the 100-point Index of Political Activity for ordinary villagers, by combining responses to the six separate questions, and I regressed this Index of Participation by Representatives against a number of demographic and other variables. In addition to the variables considered earlier, such as education, information, caste and wealth, separate dummy variables were also included for elected chief and deputy chief. These variables were included because of the allegations, made by several representatives whom we met, that elected chiefs and deputy chiefs tend to dominate the proceedings of village local governments, thereby curbing the participation of other elected representatives. The variable, Elected Chief, in the following regression table is a dummy variable that is coded 1 if the respondent is an elected chief and as zero otherwise. The dummy variable for Deputy Chief is similarly constructed. Table 2 reports the results of regression analysis. -- Table 2 here -- Wealth, as measured by landholding, once again does not matter for participation. Religion and caste are also not significantly associated with participation by elected representatives. Dummy variables for Scheduled Caste, Scheduled Tribe and Other Backward Caste did not achieve significance in any alternative specification of the regression model. Thus, it cannot be concluded that representatives from these categories participate any less actively, on average, compared to representatives from other social groups. Gender is strongly and negatively associated with participation. Female representatives tend to participate to a considerably smaller degree 15 percentage points less on average compared to male representatives. Education and Access to Information are significantly associated with higher participation by elected representatives. Every additional year of education tends, on average, to raise representatives participation by more than two-and-a-half percentage points. Similarly, higher Access to Information is positively and significantly associated with greater participation by almost three percentage points, on average, for each additional source of information that is consulted. In addition to the seven sources of information considered for ordinary villagers, two additional sources, government officials and political parties, were also added in the case of elected representatives. Two of these nine sources, household members and neighbors, are not individually significant when considered separately in association with the Index of Participation by Representatives. The other seven sources are all significantly associated with this Index, with the size of the coefficient being largest for radios and newspapers (6.4 and 5.8, respectively). Many different types of [government] schemes and programs are in operation, a former panchayat leader informed me. If they do not know about these schemes, then of what use are our leaders? Ordinary villagers do not have the means to know what program benefits exist. Leaders perform these functions [for them] They meet with officials. They [should] know about schemes and programs Otherwise, they can only sit silent and do little in meetings. Information and education matter importantly thus, and wealth and social status do not have any significant association with participation in representatives meetings. In addition, elected chiefs have significantly (and substantially) higher participation score than other elected members. On average, they score 13 percentage points higher than other elected representatives. However, deputy chiefs participation scores are not significantly different from those of other elected representatives. There are substantial similarities thus between factors associated with higher participation by ordinary villagers and by their elected officials: Information, education and gender matter critically for who participates, while wealth and social status do not have any significant association with higher participation in either case. Structure and Agency Education and information can help to enhance participation and equity in democracy. However, structural reforms, particularly affirmative action remedies, have constituted the main policy response to the possibility of inequity and elite capture. Investing in individual capabilities has not usually been considered for this purpose. In this section I look at the relative utility of structure-based versus agency-based approaches. Affirmative action policies are practiced within rural local governments in India through reserving the position of sarpanch in selected village governments for specific categories of individuals (Scheduled Tribe, Scheduled Caste, Other Backward Caste, or women). Up to one-third of all elected positions in village local governments have been reserved for women since 1993. Positions for Scheduled Tribes and Scheduled Castes have been reserved from an earlier time in proportion to their share in each local governments population, and positions for Other Backward Castes were also reserved starting in 1993. I examine below whether reserving elected positions for specific categories of individuals enhances participation rates among these and other categories of villagers. In addition, I look at some other aspects of institutional design. Jurisdiction size is examined as a factor affecting participation in village local governments. On average, three adjoining villages have been combined together in this region to constitute the jurisdiction of a single local government. One among these villages, usually the largest in terms of population, has been designated the official headquarters of the panchayat. Residents of non-headquarter villages have to travel, often over quite significant distances, in order to participate at meetings, and I examine below whether long distances to headquarters deter people from participating more actively. I also look at another group of village-level variables, including variables related to population size, infrastructure availability, and distance to high school, in order to assess how these aspects might affect different individuals participation scores. The 100-point Index of Political Activity (by ordinary villagers), considered earlier, is the dependent variable for this analysis. -- Table 3 here -- Among village-level variables, population size is not significant in any alternative specification of the regression model. Distance to Market was utilized in this analysis as a proxy variable for a villages relative level of commercialization, and it was also found to be not significant. Other variables related to the level of infrastructural facilities available in each village were also considered, including electricity, post and telegraph, roads and transport facilities. However, with the sole exception of high school facilities, reported below, none among these variables was significantly associated with the dependent variable, either individually or in combination with any of the other variables. Among the institutional design variables, Distance from Headquarters is significant in the analysis. Residents of villages located further away from local government headquarters participate, on average, 1.25 percentage points less for every kilometer of distance from headquarters, which is quite significant, given that the distance to headquarters is sometimes as much as 15 kilometers. The variable, Distance to High School, is significant, showing the continued strong effect that education has upon participation in village local governments. Most villages have a primary school located close at hand, and middle schools are also located usually no more than two to three kilometers from most villages in this region. Differences between villages in terms of educational facilities are most apparent, however, when comparisons are made based upon distance to high schools. In the sample of 53 villages considered for this analysis, this distance ranges from zero to 25 kilometers. Villages located closer to high schools have significantly higher participation scores compared to villages located further away from high schools. A set of dummy variables Reservation for Elected Chief was constructed to examine the effect on average participation rates associated with reserving the position of elected chief for women, Scheduled Castes and Scheduled Tribes, respectively. The dummy variable, GENDER, takes the value 1 in villages where the position of elected chief was reserved at the time of survey for a female candidate, and it takes the value zero in other villages. The dummy variables, Scheduled Tribe, Scheduled Caste and Other Backward Caste are constructed similarly. Among these dummy variables, no significance was achieved by GENDER, Scheduled Caste and Other Backward Caste, implying that reserving the position of elected chief for females, Other Backward Castes, or Scheduled Castes does not have any significant association with higher or lower participation scores by ordinary villagers. The dummy variable, Scheduled Tribe, is both negative and significant, indicating that affirmative action may have achieved an effect that is opposite to what was intended; however, this result needs to be interpreted carefully. Separately, analysis carried out for only female villagers showed that their participation scores do not change significantly when the position of elected chief is reserved for women. Larger gains in participation scores might become apparent if these affirmative action policies are continued for a longer period. Reservations for women, for instance, have been in force for only seven years so far, and it is possible that continuing with reservations for a longer period might have the desired effects of enhancing participation through making up for the disabilities suffered presently by women and by Scheduled Tribes. It will be important, however, to supplement affirmative action with other policies, targeted at some other structural and individual-level factors. For instance, structural reforms that have the effect of locating general assembly meetings in all villages, rather than only holding these meetings at the headquarters villages, should help raise participation rates among a larger number of ordinary villagers. Similarly, allocating authority more equally within these elected bodies, and not concentrating it so much with elected chiefs, should also encourage other elected representatives to participate more actively. However, these structural effects tell only one part of the story. Some agency factors have consistently positive associations with participation, and working upon these factors can help to raise participation all around. Information and education are of critical importance in this regard. Gender is negatively related to participation in these contexts, however, information and education also help to reduce the quite severe disabilities suffered by women in this region. Focusing more on these individual-level attributes should help considerably, particularly over the short to medium term. Information is at a relatively low level overall in villages of this region, for example, only 20 percent of rural households in Madhya Pradesh and 27 percent in Rajasthan own a radio; only 19 percent of women in Madhya Pradesh and 15 percent in Rajasthan consult the radio at least once every week; and only nine percent of women (in Madhya Pradesh) and five percent (in Rajasthan) consult newspapers once a week or more often. Considerable scope exists, therefore, to raise participation through higher information flows. Participation and Democracy in Information-Poor Environments This analysis has helped to show how education and information can help overcome the disabilities to participation associated with being poor (or being otherwise disadvantaged, for instance, on account of gender). Educational achievement is increasing in many countries, even as poverty remains widespread. It would be useful to conduct careful empirical examinations in diverse contexts within the developing and transitional world intended to tease out the relative effects on participation of wealth and income, on the one hand, and education and information, on the other hand. Evidence culled from different accounts suggests that there might be considerable utility in these efforts. Investigating participation within regional councils in Namibia, for instance, Forrest found that more actors get involved actively in situations where information provision is better. Gombay shows how informal civil society networks have facilitated information flows and enabled different actors in Uganda to participate more effectively in reviving democracy. Diverse democratic actions geared toward environmental preservation have been widened and improved significantly by enhanced access to information resources. Enhancing poor and disadvantaged peoples participation in democracy should be assisted more widely thus through the provision of more information in the short term and more education in the medium term. Of course, institutional designs will also matter for example, in the Indian case we found that participation could be enhanced by decentralizing village local governments further and by distributing authority more equally between elected chiefs and other representatives. The key, however, is that one need not wait for economic growth to reduce poverty and enhance participation. Just as political institutions need not be linked necessarily with the level of economic development, so it is possible to have basic education and information at high levels even where many people are poor. In the Indian state of Karnataka, for instance, the provision of information through public radio enabled listeners in far-flung villages to participate in numerous activities associated with local democratic governance. In the adjoining state of Kerala, marked out by many for its significantly higher participation rates overall (though not so much for its level of economic development), literacy has reached 91 percentAs a direct result, traditionally marginalized groups, most notably women and dalits [Scheduled Castes], have acquired the basic social skills required for informed participation. Participation and also equitable sharing of benefits and services are both higher among individuals and communities who have higher education and greater access to information. Other things matter for democracy in addition to participation, and other things matter for participation in addition to information. Unionization, party building, and social networks can help to raise participation rates, including among poor people. Education and information are not, therefore, the only means for raising participation in democracy. However, a democracy with poor information is an incomplete system that by its very design tends to promote exclusion and not participation. People react critically to political communication only to the extent that they are knowledgeable about political affairs. It matters little if democratic rights are guaranteed to all by a countrys constitution: if some person does not know what democratic rights she has, or if she does not know what needs to be done in order to have those rights enforced, then for all practical purposes these rights do not exist for this individual. And if there are many such individuals who do not have the information it requires to partake of and participate in democracy, then it follows that their stakes in democracy will be relatively low. Enhancing information availability on a regular and reliable basis will be necessary to establish faith and improve participation in such situations. Else, it is likely that a sham democracy might result, accessible, intelligible and available to only a few, and dominated by them to the expense of all others. Table 1. OLS Regression on Index of Political Activity: 100-point Index of Political Activity is the Dependent Variable CoefficientStandard Error (S.E.)Intercept36.14****6.63Independent VariablesGender-24.45****1.54Age 0.001 0.05Religion-0.573.52Caste  -- Scheduled Caste (dummy)0.860.87 -- Scheduled Tribe (dummy)-5.73**1.79 -- Other Backward Caste (dummy)2.21 1.44Education (years at school)0.24**0.11Landholding 0.10 0.13Family Size-0.130.27Access to Information 5.19****0.48State (dummy)5.01*** 1.35N1,536R20.424Adj-R20.418F-value108.63F-probability<0.0001Note: *p<=.05 **p<=.01 ***p<=.001 ****p<.0001 Note: Mean for the dependent variable = 39.3; Standard Deviation = 22.1; Skewness = 0.49. Multicollinearity is low (condition index = 14.64). Heteroskedasticity not in evidence. Table 2. OLS Regression of Participation by Elected Representatives: 100-point Index of Participation by Representatives is the Dependent Variable CoefficientS.E.Intercept37.18***10.93Independent VariablesGender-14.99***4.16Age 0.170.11Religion6.865.48Caste -- Scheduled Caste (dummy)-1.571.75 -- Scheduled Tribe (dummy)-5.365.61 -- Other Backward Caste (dummy)2.463.24Education (years at school)1.35**0.42Landholding 0.410.38Access to Information 2.70***0.78Elected Chief (dummy)12.37**4.20Deputy Chief (dummy)4.814.13State (dummy)7.2**3.15N190R20.424Adj-R20.396F-value14.15F-probability<0.0001Note: *p<=.05 **p<=.01 ***p<=.001 ****p<.0001 Note: Mean for the dependent variable = 59.6; Standard Deviation = 21.9; Skewness = -0.65. Multicollinearity is low (condition index = 12.24). Heteroskedasticity not in evidence. Table 3. OLS Regression of the Index of Political Activity: 100-point Index of Political Activity is the Dependent Variable CoefficientS. E.Intercept49.29***3.85Independent Variables (a) Village-Level VariablesPopulation0.0040.005Distance to Market -0.030.09Distance to Headquarters -1.25**0.43Distance to High School -0.59**0.21Reservation for Elected Chief -- Scheduled Caste0.822.97 -- Scheduled Tribe-6.39**2.71 -- Other Backward Caste0.641.82 -- GENDER0.531.75 (b) Individual-level VariablesGender-25.23****1.52Social Group Ranking-1.291.56Landholding 0.170.22Access to Information 6.0****0.48State (dummy)5.41*2.44N1,553R20.417Adj-R20.413F-value91.85F-probability<0.0001Note: *p<=.05 **p<=.01 ***p<=.001 ****p<.0001 NOTES     PAGE  PAGE 1 PAGE  PAGE 26 PAGE  PAGE 38 * Earlier versions of this paper were presented at the 2003 Annual Meeting of the American Political Science Association and at the NETSAPPE workshop, Paris, June 28-30, 2004. Pradeep Chhibber, Peter Houtzager, Mary Katzenstein, Judith Kelley, Helen Ladd, David Rueda, Erik Webbels, and three anonymous reviewers provided helpful comments. The usual disclaimers apply.  See Leonardo Avritzer, 2002, Democracy and the Public Space in Latin America, (Princeton, NJ: Princeton University Press); Evelyne Huber, Dietrich Rueschmeyer, and John D. Stephens, 1999, The Paradox of Contemporary Democracy: Formal, Participatory and Social Dimensions, pp. 168-92 in Lisa Anderson, ed., Transitions to Democracy, (New York: Columbia University Press); and Nicolas Van de Walle, 1999, Economic Reform in a Democratizing Africa, Comparative Politics, 32 (1), 21-41.  See, for example, Gabriel Almond and Sidney Verba, 1965, The Civic Culture: Political Attitudes and Democracy in Five Nations, (Boston: Little, Brown); Seymour M. Lipset, 1959, Some Social Requisites of Democracy: Economic Development and Political Legitimacy, American Political Science Review, 53, 65-105; Lipset, 1994, The Social Requisites of Democracy Revisited, American Sociological Review, 59, 1-22; Lipset, Kyoung-Ryung Seong, and John Charles Torres, 1993, A Comparative Analysis of the Social Requisites of Democracy, International Social Science Journal, 45, 155-75; Steven J. Rosenstone and John M. Hansen, 1993, Mobilization, Participation and Democracy in America, (New York: Macmillan); Verba, Norman H. Nie, and Jae-On Kim, 1978, Participation and Political Equality: A Seven-Nation Comparison, (Cambridge, MA: Cambridge University Press); and Verba, Kay Schlozman, and Henry Brady, 1995, Voice and Equality: Civic Voluntarism in American Politics, (Cambridge, MA: Harvard University Press). The last cited work shows how participation rates are higher among relatively richer citizens in the United States (with incomes above $75,000) and relatively poorer ones (incomes below $15,000): voting (86% v. 52%), campaign work (17% v. 4%), and contacting (50% v. 25%).  World Bank, 2001, World Development Report 2000/1, (Washington, DC: The World Bank), p. 281.  Iris Marion Young, 2000, Inclusion and Democracy, (Oxford, UK: Oxford University Press), p. 34.  Larry Diamond, 1992, Economic Development and Democracy Reconsidered, pp. 93-139 in Gary Marks and Larry Diamond, eds., Re-examining Democracy: Essays in Honor of Seymour Martin Lipset, (Thousand Oaks, CA: Sage Publications), p. 125-7.  Michael Bratton and Robert Mattes, 2001, Africans Surprising Universalism, Journal of Democracy, 12 (1), 107-21.  Steven E. Finkel, 2002, Civic Education and the Mobilization of Political Participation in Developing Democracies, Journal of Politics, 64 (4), 994-1020.  See, for example, Jonathan T. Hiskey and Mitchell A. Seligson, 2003, Pitfalls of Power to the People: Decentralization, Local Government Performance, and System Support in Bolivia, Studies in Comparative International Development, 37 (4), 64-88; Robert A. Jackson, 1995, Clarifying the Relation Between Education and Turnout, American Politics Quarterly, 23, 279-99; Jackson, 2003, Differential Influences on Latino Electoral Participation, Political Behavior, 25 (4), 339-66; and Raymond Wolfinger and Steven J. Rosenstone, 1980, Who Votes? (New Haven: Yale University Press).  Information is vital to democracy in myriad ways in the processes by which citizen preferences are formed and aggregated, in the behaviors of citizens and elites, in formal procedures of representation and in the mechanisms of accountability None of these elements of the democratic process can operate apart from the exchange and flow of information (Bruce. A. Bimber, 2003, Information and American Democracy: Technology in the Evolution of Political Power, New York: Cambridge University Press, p. 11)  See, for instance, Bimber 2003; Robert A. Dahl, 1989, Democracy and Its Critics, (New Haven: Yale University Press); Michael X. Delli Carpini and Scott Keeter, 1996, What Americans Know About Politics and Why It Matters, (New Haven: Yale University Press); Shanto Iyengar and Donald R. Kinder, 1987, News that Matters: Television and American Public Opinion, (Chicago: University of Chicago Press); Pippa Norris, 2000, A Virtuous Circle: Political Communications in Post-Industrial Societies, (New York: Cambridge University Press); and Verba, et al. (1995).  Harry Eckstein, 1975, Case Study and Theory in Political Science, pp. 79-137 in Fred I. Greenstein and Nelson W. Polby, eds., Handbook of Political Science, (Reading, MA: Addison-Wesley).  For the opposite viewpoint namely, that decentralization can in certain circumstances have deleterious effects, especially on macro-economic performance see, for example, Jonathan Rodden, 2002, The Dilemma of Fiscal Federalism: Grants and Fiscal Performance Around the World, American Journal of Political Science, 46 (3), 670-87; and Daniel Treisman, 2000, Decentralization and Inflation: Commitment, Collective Action, and Continuity, American Political Science Review, 94 (4), 837-57.  World Bank, 2001, p. 106.  See, for instance, Catherine Boone, 2003, Decentralization as Political Strategy in West Africa, Comparative Political Studies, 36: 355-80; Richard Crook and James Manor, 1998, Democracy and Decentralization in South Asia and West Africa: Participation, Accountability and Performance, (Cambridge, UK: Cambridge University Press); Christopher Garman, Stephan Haggard, and Eliza Willis, 2001, Fiscal Decentralization: A Political Theory with Latin American Cases, World Politics, 53, 205-36; and Milton J. Esman and Norman T. Uphoff, 1984, Local Organizations: Intermediaries in Rural Development, (Ithaca, NY: Cornell University Press). Hiskey and Seligson, p. 83.  Planning Commission of India figures, cited in Ahluwalia (2000: 1640).  These figures are derived from the random sample of 2,013 adult villagers interviewed for this study. Wealth is measured in terms of landholding, as explained later in the text, and education is measured in terms of number of years spent at school.  See, for instance: Francine R Frankel, 1989, Introduction, pp. 1-20 in Francine R. Frankel and M.S.A. Rao, eds., Dominance and State Power in Modern India: Decline of a Social Order, Volume 1, (Delhi: Oxford University Press); Christophe Jaffrelot, 1996, The Hindu Nationalist Movement and Indian Politics, (New York: Columbia University Press); and Ashutosh Varshney, 2002, Civic Life and Ethnic Conflict: Hindus and Muslims in India, (New Haven: Yale University Press).  For a more complete description of panchayats, see: George Mathew, 1995, Panchayati Raj: From Legislation to Movement, (New Delhi: Concept Publishing); and Mathew and Ramesh C. Nayak, 1996, Panchayats at Work: What it Means for the Oppressed, (New Delhi: Institute of Social Sciences).  Copies of this questionnaire and the one reported in the next section can be obtained on request from the author.  See, for example, Michael Bratton, 1999, Political Participation in a New Democracy: Institutional Considerations from Zambia, Comparative Political Studies, 32 (5), 549-588; and Verba, Schlozman and Brady, 1995.  This single common factor has a communality of 3.87, implying that it accounts for more than 65 percent of the combined variance of these six separate elements. Factor loadings are all 0.7 or higher.  Correlations of the six individual items with the Index are all 0.77 or higher. Cronbachs Coefficient Alpha = 0.875.  Standard hectares are used for these measurements. Differences in productive quality between irrigated and un-irrigated land are factored into these numbers.  Apart from one of these sources, newspapers, all of others are accessible equally by uneducated and educated villagers, so education is not closely correlated with this information variable, as is also verified by the data.  District dummies were also employed in the analysis but without achieving any significance.  The correlation coefficient between land ownership and Education is only 0.14. The correlation coefficient between land ownership and Information is even smaller: 0.09. Correlation between Education and Information is 0.445, but this is still not high enough to cause problems of multicollinearity.  It is beyond the scope of this paper to examine why the gender effect should loom so large. Gyarasi Bai, a woman leader of Balapura village, Ajmer district, had the following to say in this regard: In matters outside the house, such as politics, it is the men who have traditionally had most say. Women are becoming more aware now of the world outside, but for generations past their place has been inside their own house. (Interviewed in Balapura, June 2, 2002).  Interactive variables combining gender with other caste and religious categories were also examined, but they were not significant.  The salience of politically relevant information obtained through such social communications among neighbors and fellow citizens is examined theoretically and empirically by R. Robert Huckfeldt, and John Sprague, 1995, Citizens, Politics and Social Communication: Information and Influence in an Election Campaign (Cambridge, UK: Cambridge University Press).  Villagers who live in the state of Rajasthan have participation scores that are, on average, five percentage points higher than those of villagers who live in Madhya Pradesh. However, why this difference between states should arise is not entirely clear from this analysis.  Caste was significant at the 0.01 level in this analysis, while age and length of residence were significant at the 0.1 level.  Interview with Mothulal Vaishnava, Kailashpuri village, Udaipur district, May 22, 2002.  It is not possible to dismiss entirely an alternative explanation suggesting that those who are more politically active also tend to seek information and education more actively. Following Norris (2000: 18), I expect that a two-way interactive process operates, a virtuous circle, with a ratchet of reinforcement [which] moves in a direction that is healthy for democratic participation. It is interesting, however, to note that each of the seven sources of information is separately significant, including sources that an individual makes an effort to consult (such as radio) as well as others from which information becomes available more routinely (e.g., household members). It is not just individuals who are politically active who are also actively looking for information; political activity tends to increase even when information is routinely available to individuals.  Yogendra Yadav, 1999, Electoral Politics in the Time of Change: Indias Third Electoral System, 1989-99, Economic and Political Weekly, Bombay, August 21, 2391-99.  A single common factor underlies responses to the six separate questions. Communality is 4.42 out of a maximum possible score of 6, indicating that more than 70 percent of the combined variance is explained by this common factor.  Removing the components for government officials and political parties does not result in changing the significance of the composite variable, Access to Information, suggesting that endogeneity is not indicated in this relationship between Access to Information and high participation scores.  Interview with Vandana Meena, formerly Zila Pramukh (elected President of the district-level local government), Udaipur, March 10, 1999  Once again, the regression model was run without including the education and information variables, and once again variables gained significance that were not previously significant, including caste (at the 0.01 level), wealth (at the 0.01 level), and age (at the 0.05 level).  Because there exists a close association between the variable Distance to High School and average educational achievement in a village, the individual-level variable, Education, is not additionally considered in this analysis. The variables Distance to High School and Distance to Headquarters are not closely correlated.  The position of elected chief has been reserved for Scheduled Tribes in local governments where Scheduled Tribes form the majority of the population. Because individual Scheduled Tribes tend to have participation scores that are, on average, five to six percentage points lower than other villagers, it can be expected that the average participation score in areas where Scheduled Tribes form the majority of the population will be, in general, less than those for other areas. Reserving the position of elected chief for Scheduled Tribes is not the cause of the observed decline.  See in this context Rohini Pande, 2003, Can Mandated Political Representation Increase Policy Influence for Disadvantaged Minorities? Theory and Evidence from India, American Economic Review, 93 (4), 1132-51).  For instance, a woman who has access to between five and seven information sources is five times more likely to have a participation score within the top three deciles compared to another woman who has access to only three or four information sources.  Data from India National Fertility and Health Survey II, 1988-99.  Evidence from the United States indicates that diminishing returns might set in when levels of education and information become relatively high on average (Richard Davis, The Web of Politics, Oxford, UK: Oxford University Press, 1998; Norman Nie, Jane Junn, and Kenneth Stehlik-Barry, Education and Democratic Citizenship in America, Chicago: University of Chicago Press, 1996). This eventuality is not likely to be of immediate concern in India, however, and in other democracies where poverty remains large and average education levels are still quite low.  Joshua Bernard Forrest, 2000, The Drought Policy Bureaucracy, Decentralization, and Policy Networks in Post-Apartheid Namibia, American Review of Public Administration, 30 (3), 307-33.  Christie Gombay, 1998, Market Vendors, City Government, and the World Bank in Kampala, Uganda, pp. 150-182 in Jonathan Barker, ed., Street-Level Democracy: Political Settings at the Margins of Global Power, (West Hartford, CT: Kumarian Press).  World Resources Institute, 2002, Closing the Gap: Information, Participation and Justice in Decision-Making for the Environment, (Washington, DC: World Resources Institute).  Karen Remmer, 1997, Theoretical Decay and Theoretical Development: The Resurgence of Institutional Analysis, World Politics, 50, 34-61.  The social scientists involved in developing this program have reported on it in a fortnightly newsmagazine (A.P. Kripa and G.S. Ganesh Prasad, The Message of Panchayati Raj, Frontline, April 15-28, 2000).  Patrick Heller, 2000, Degrees of Democracy: Some Comparative Lessons from India, World Politics, 52: 484-519, p. 498.  A parallel analysis of perceived exclusion from the services and benefits provided by panchayats was conducted among the same random sample of 2,013 village residents. Education and information mattered critically (and negatively) for the extent of exclusion from services and benefits.  James L. Gibson, 2001, Social Networks, Civil Society, and the Prospects for Consolidating Russias Democratic Transition, American Journal of Political Science, 45 (1), 51-68; Atul Kohli, 1990, Democracy and Discontent: Indias Growing Crisis of Governability, (Cambridge, UK: Cambridge University Press); Anirudh Krishna, 2002, Enhancing Political Participation in Democracies: What is the Role of Social Capital? Comparative Political Studies, 35 (4), 437-60.  John Zaller, 1992, The Nature and Origins of Mass Opinion, (Cambridge, UK: Cambridge University Press).  This variable reflects the omnibus effect of caste status, and it is calculated by assigning ranks (from 1 to 5) to each caste related to its position within the caste hierarchy. This assignment is somewhat arbitrary, no doubt; however, these results do not change when other caste variables are considered instead.     o p * + +,)*rsHIKLRk""8#9#&&,,,,,,,,*-/-..4/5/77-7.7ؽؽؽؽؽؽؽؽؽؽؽؽؽؽؽؽؽؽظظسظظظظؽج hm@6] hm@] hm@6jhm@0J)U hm@KH hm@5 hm@0J)hm@hm@CJOJQJaJhm@OJQJhM hm@56 hM 5D     Ahij t@`@ $a$ $ hda$$da$##d #$da$gdM tQRk"&R,/477.7q8=?ICE1GIK+OLSWWWh`h@`@.7m8n8p8q888S>T>/@3@_A`AEE]IcInIIIKKK2L3L5LMM+O4OOOJSKSWW#WWW8Z9Z[[AaDaddf!gxkykmmnnoor rrr(t)ttuxxW|X|R}Z}_}d}n}q}   hm@5\ hm@] hm@6 hm@>* hm@6] hm@7jhm@0J)Uhm@RWWW:Z[`df g'hylmjp rttux||||  ҅ɇ@`@hx`hh`h$a$Ѕх.V^5dŒFuvНѝCD\ʪ˪~ծ֮:@STRSĴŴ|}~Ѹ<vQR`aĺ hm@H* hm@aJ hm@KH hm@56 hm@6 hm@\ hm@6] hm@5 hm@]jhm@0J)Uhm@Jɇ.I !56wF\ h`hhx`hx$a$@`@h`h<|}~ $d$Ifa$ d$If d$Ifd$da$@`@ ypdd $d$Ifa$ d$Ifkd$$IfTlF :R Tp0    4 laTѸyp d$Ifkd$$IfTlF :R Tp0    4 laTѸҸٸ $d$Ifa$ d$Ifbkd$$IfTl4>04 laf4Typddd $d$Ifa$ d$IfkdL$$IfTlF :R Tp0    4 laT ypdd $d$Ifa$ d$Ifkd*$$IfTlF :R Tp0    4 laTypdd $d$Ifa$ d$Ifkd$$IfTlF :R Tp0    4 laT<AFynbb $d$Ifa$ d$Ifkd$$IfTlF :R Tp0    4 laTFGhpuypdd $d$Ifa$ d$Ifkd$$IfTlF :R Tp0    4 laTuvyl``` $d$Ifa$ +$If[$\$kd$$IfTlF :R Tp0    4 laTŹ̹ѹyppdd $d$Ifa$ d$Ifkd$$IfTlF :R Tp0    4 laTѹҹ߹ypddd $d$Ifa$ d$Ifkd^$$IfTlF :R Tp0    4 laTypdd $d$Ifa$ d$Ifkd< $$IfTlF :R Tp0    4 laT#(ypdd $d$Ifa$ d$Ifkd $$IfTlF :R Tp0    4 laT()7?@Eypddd $d$Ifa$ d$Ifkd $$IfTlF :R Tp0    4 laTEFHNOypdd $d$Ifa$ d$Ifkd $$IfTlF :R Tp0    4 laTOPSYZypdd $d$Ifa$ d$Ifkd $$IfTlF :R Tp0    4 laTZ[bhiypdd $d$Ifa$ d$Ifkd $$IfTlF :R Tp0    4 laTijryzypdd $d$Ifa$ d$Ifkd$$IfTlF :R Tp0    4 laTz{ypdd $d$Ifa$ d$Ifkd$$IfTlF :R Tp0    4 laTºym $d$Ifa$kdr$$IfTlF :R Tp0    4 laTºúĺyz"vj $d$Ifa$ d$If d$If $@d^@a$$da$bkd^$$IfTl404 laf4T ĺyz"#=SĽ37bcſۿ8RGbFJvw~hGjhGU hm@CJjhm@0J)U hm@6 hm@CJhm@56\] hm@5\ hm@H* hm@KH hm@aJ hm@56hm@56]hm@6CJOJQJ] hm@5hm@ hm@CJ7"#-6<{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laT<=STyl` $d$Ifa$ d$If^kd$$IfTlFlb0    4 laTTU\fkvv $d$Ifa$ d$Ifskd($$IfTl40:04 laf4Tklqv{{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laT{|{rff $d$Ifa$ d$Ifkdr$$IfTlFlb0    4 laT{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laTż{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laTżƼ{rff $d$Ifa$ d$Ifkd.$$IfTlFlb0    4 laT#{nbb $d$Ifa$ +$If[$\$kd$$IfTlFlb0    4 laT#$/AHM{rrff $d$Ifa$ d$IfkdV$$IfTlFlb0    4 laTMN[`e{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laTef}{rff $d$Ifa$ d$Ifkd~$$IfTlFlb0    4 laT{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laTĽɽν{pdd $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laTνϽݽ{rff $d$Ifa$ d$Ifkd:$$IfTlFlb0    4 laT{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laT{rff $d$Ifa$ d$Ifkdb$$IfTlFlb0    4 laT {rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laT {rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laT)12{rff $d$Ifa$ d$Ifkd$$IfTlFlb0    4 laT23a{o $d$Ifa$kd$$IfTlFlb0    4 laTabcW~vmd d$If d$If$da$ $@d^@a$ ($@d^@a$$da$`kdF$$IfTl4V04 laf4T Ŀg^RR $d$Ifa$ d$Ifkd$$IfTl#Fr06'    4 laT $d$Ifa$Ŀſۿul d$Ifkdr$$IfTlFr06'    4 laTۿܿ $d$Ifa$ d$Ifgkd* $$IfTl4a06'4 laf4T ul`` $d$Ifa$ d$Ifkd $$IfTlFr06'    4 laT,27ul`` $d$Ifa$ d$Ifkdh!$$IfTlFr06'    4 laT78RZ_uj^^ $d$Ifa$ d$Ifkd"$$IfTlFr06'    4 laT_`yul`` $d$Ifa$ d$Ifkd"$$IfTlFr06'    4 laTul`` $d$Ifa$ d$Ifkdf#$$IfTlFr06'    4 laTul`` $d$Ifa$ d$Ifkd$$$IfTlFr06'    4 laTul`` $d$Ifa$ d$Ifkd$$$IfTlFr06'    4 laT!ul`` $d$Ifa$ d$Ifkdd%$$IfTlFr06'    4 laT!"49>ul`` $d$Ifa$ d$Ifkd&$$IfTlFr06'    4 laT>?bcdul`` $d$Ifa$ d$Ifkd&$$IfTlFr06'    4 laTdelw|ul`` $d$Ifa$ d$Ifkdb'$$IfTlFr06'    4 laT|}ul`` $d$Ifa$ d$Ifkd ($$IfTlF*A,"06'    4 laTul`` $d$Ifa$ d$Ifkd($$IfTlF*A,"06'    4 laTul`` $d$Ifa$ d$Ifkd`)$$IfTlF*A,"06'    4 laTul`` $d$Ifa$ d$Ifkd *$$IfTlF*A,"06'    4 laTul`` $d$Ifa$ d$Ifkd*$$IfTlF*A,"06'    4 laTul`` $d$Ifa$ d$Ifkd^+$$IfTlFr06'    4 laT sj^^ $d$Ifa$ d$Ifkd,$$IfTl,Fr06'    4 laTul`` $d$Ifa$ d$Ifkd,$$IfTlFr06'    4 laT&,-ul`` $d$Ifa$ d$Ifkd`-$$IfTlFr06'    4 laT-.<DEul`` $d$Ifa$ d$Ifkd .$$IfTlFr06'    4 laTEFuui $d$Ifa$kd.$$IfTlFr06'    4 laTuvwx~&`#$xdgkd^/$$IfTl4a06'4 laf4THIg~ !23m;\Ko}&e@A뻮 hm@6hm@6B*]hphhm@B*hph hm@6]jhm@0J)U hm@0J)hGh6m0JmHnHuhm@ hm@0Jjhm@0JUBH2@j UU($a$$xa$$a$&`#$ATs~DXjk Vq%2UV&UV4B &GHd(:;XY y.<zBhm@6B*]hphhm@B*hphhm@6CJ] hm@CJjhm@0J)CJU hm@6]jhm@0J)U hm@6hm@GUG:XB>;TsU>= $a$($a$$xa$BC>?As]en9:;<2OTUstUV>?z=> xߺߵhm@B*CJphhm@6B*CJ]ph hm@6hm@B*hph hm@6hhm@6]h hm@h hm@6]hm@jhm@0J)U hm@CJjhm@0J)CJU:xMNef +Z>?u#/01We|}BeDƾƴ׬hm@OJQJhm@B*hphhm@6]h hm@h hm@6 hm@6]hm@jhm@0J)Ujhm@0J)CJU hm@CJhm@6CJ]CMe >0|sx$xa$$xa$($a$DaqsthGhm@6CJ] hm@CJjhm@0J)CJUhm@hm@OJQJhm@6OJQJ 8 0 000BP/ =!"#$% ,BP/ =!"#$% ,BP/ =!"#$% ,BP/ =!"#$% $$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#v:V l4>05/ 4f4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / / 4T$$If!vh#vR #vT#vp:V l05R 5T5p/ / / / / / 4T$$If!vh#v:V l405/ 4f4T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b5/ 4T$$If!vh#v:#v:V l405:5/ / / 4f4T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T$$If!vh#v#vb#v:V l055b54T|$$If!vh#vV:V l405V4f4T$$If!vh#v#v:#v:V l#06'555/ 4T$$If!vh#v#v:#v:V l06'555/  / 4T$$If!vh#v:V l406'5a/ 4f4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l,06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v#v:#v:V l06'555/ 4T$$If!vh#v:V l406'5a/ 4f4T^, 2 0@P`p2( 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p8XV~_HmH nH sH tH H`H Normal$da$CJ_HmH sH tH >@>  Heading 1$$@&a$5D@D  Heading 2$$@&a$ 56CJJ@J  Heading 3$$@&a$5CJOJQJH@H  Heading 4$$@&a$ 6htH u:@:  Heading 5$$@&a$D@D  Heading 6$@&5CJOJQJF@F  Heading 7$$@@&^@a$5BB  Heading 8$$@&^a$> @>  Heading 9 $$@&a$5DA`D Default Paragraph FontVi@V  Table Normal :V 44 la (k (No List 2O2 OutlineKHPP Outline1&$ & F hh^h`VOV Outline2' & F ``^``KHVO"V Outline3' & F XX^X`KHVO2V Outline4' & F  PP^P`KH\B\ outlinebullet' & F h>x^`>: @R: Footer$ !a$\S@b\ Body Text Indent 3$`a$ 6WA@&q@ Footnote ReferenceH*4B@4 Body Text$a$J@J  Footnote Text$$da$CJ<P@< Body Text 2$a$5JCJ Body Text Indent$^a$VRV Body Text Indent 2$8h^8`ha$.)@. Page Number8@8 Header  !CJ:: ChapterNumberh>'> Comment ReferenceCJ8@8  Comment Text!CJ>Q@"> Body Text 3" CJOJQJ0>@20 Title#$a$5@OB@ inda$ & F h 5OJQJLOARL indb&% & F hP^`P56Ua6 Hyperlink >*B*phFVqF FollowedHyperlink >*B* ph8+@8  Endnote Text(CJ>*@> Endnote ReferenceH*H@H  Balloon Text*CJOJQJ^JaJR^@R Normal (Web)+$ddd[$\$a$aJPK![Content_Types].xmlN0EH-J@%ǎǢ|ș$زULTB l,3;rØJB+$G]7O٭V$ !)O^rC$y@/yH*񄴽)޵߻UDb`}"qۋJחX^)I`nEp)liV[]1M<OP6r=zgbIguSebORD۫qu gZo~ٺlAplxpT0+[}`jzAV2Fi@qv֬5\|ʜ̭NleXdsjcs7f W+Ն7`g ȘJj|h(KD- dXiJ؇(x$( :;˹! I_TS 1?E??ZBΪmU/?~xY'y5g&΋/ɋ>GMGeD3Vq%'#q$8K)fw9:ĵ x}rxwr:\TZaG*y8IjbRc|XŻǿI u3KGnD1NIBs RuK>V.EL+M2#'fi ~V vl{u8zH *:(W☕ ~JTe\O*tHGHY}KNP*ݾ˦TѼ9/#A7qZ$*c?qUnwN%Oi4 =3N)cbJ uV4(Tn 7_?m-ٛ{UBwznʜ"Z xJZp; {/<P;,)''KQk5qpN8KGbe Sd̛\17 pa>SR! 3K4'+rzQ TTIIvt]Kc⫲K#v5+|D~O@%\w_nN[L9KqgVhn R!y+Un;*&/HrT >>\ t=.Tġ S; Z~!P9giCڧ!# B,;X=ۻ,I2UWV9$lk=Aj;{AP79|s*Y;̠[MCۿhf]o{oY=1kyVV5E8Vk+֜\80X4D)!!?*|fv u"xA@T_q64)kڬuV7 t '%;i9s9x,ڎ-45xd8?ǘd/Y|t &LILJ`& -Gt/PK! ѐ'theme/theme/_rels/themeManager.xml.relsM 0wooӺ&݈Э5 6?$Q ,.aic21h:qm@RN;d`o7gK(M&$R(.1r'JЊT8V"AȻHu}|$b{P8g/]QAsم(#L[PK-![Content_Types].xmlPK-!֧6 0_rels/.relsPK-!kytheme/theme/themeManager.xmlPK-!0C)theme/theme/theme1.xmlPK-! ѐ' theme/theme/_rels/themeManager.xml.relsPK] o*+) r H K84'm0S6_9=2DJKO8RS\xcegjj(lp|}uЕCʢ~զSRĬ    "s]k-  4 re$miEf !"i#$%h&&6**+,x-./22374i6'788[9-::;=>H?K?x:h $$$$$/=====HVVVVVVVVVY.7ĺABxDtWɇѸFuѹ(EOZizº"<Tk{ż#Meν 2aĿۿ7_!>d|-EuU  $+/69=DHORY!!!!!!8@0(  B S  ?w-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~bb{{ddY n n : : n n G!!qDqDEEllll..JJAAllvvZ2==B(22<))2())rrwwN55 MVxxOO       !"#$%&'()*+,-/.0124356798:;<=>?@BACDFEGIHJKLNMOPQRSTUVWXYZ[\]^_`abcdefhgikjlmnpoqrsutvggiie G G s s L!!vDvDEEllll33QQGGqq^;ALL0;FF044$77//ww~~S<<--T\TT   !"#$%&'()*+,-/.0124356798:;<=>?@BACDFEGIHJKLNMOPQRSTUVWXYZ[\]^_`abcdefhgikjlmnpoqrsutv9w*urn:schemas-microsoft-com:office:smarttagsplaceBv*urn:schemas-microsoft-com:office:smarttagscountry-region=t*urn:schemas-microsoft-com:office:smarttags PlaceName=s*urn:schemas-microsoft-com:office:smarttags PlaceType8`*urn:schemas-microsoft-com:office:smarttagsCity9G*urn:schemas-microsoft-com:office:smarttagsState Mwvwtswvvwvvwvwvwvwvwvwv`w`w`wvwvwvwvwvwvwvw`wvw`G`vwts``w`Gwv`w``wtsGwtsw`G``wstwvw`w`vw`w`wvwvwvwvwv`vw`Gw`Gwvw`v`vwts, 2 NV[d""S"["`"g"q${$$$$$$$*%/%=%F%R&\&&&,,w--U]\]k"k#k*knnnn`|i|V^͢Ӣ:@ 0Tfж~ԺպW_ż̼ݼ`e׾ݾjpпֿؿ v|)/@HjpX^_fqvX]lr*309?IeofkmvAHIL`hNVW`bm T_  # X]z r <<I'I^^ dWdkk l lqq $ Twz#~Ժպmp?u{Dr 33333333333333333333333333$$~'~'m0n0AA*G*GGG9N9NJN}Nll0s0s=s>sOsPsss 55nn}~~ԺպHIN\gý 4F̾̾jjппٿ ++,?U"$TUUh>KKNOPW\9::;-TsU<Gtu78 Vf 00|j ~~Ժպ SpAyd9ZGg U6%~y0^`B*CJOJQJo(P^`P.^`.P^`P.  ^ `)@  ^@ `()x^`()H^`()^`()^`() P^`POJQJo(n hh^h`OJQJo(pApApApA~y9ZGgS M m@6mG~@____@UnknownG*Ax Times New Roman5Symbol3. *Cx ArialY CG TimesTimes New RomanABook Antiqua5. .[`)Tahoma;WingdingsA$BCambria Math"1@h*G*GSF'_RG%'Y4G 2qHP ?M 2!xx ;Poor People's Participation in Democracy at the Local LevelSanford Institute7Copy    Oh+'0  8 D P \hpx<Poor People's Participation in Democracy at the Local LevelSanford Institute Normal.dotm7Copy2Microsoft Office Word@@ U@A@A'՜.+,00 hp  Duke UniveristyR_ <Poor People's Participation in Democracy at the Local Level Title  !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~      "#$%&'(-Root Entry FA/Data /1Table>NWordDocumentİSummaryInformation(DocumentSummaryInformation8!CompObjr  F Microsoft Word 97-2003 Document MSWordDocWord.Document.89q