Voice and votes--does political decentralisation work for the poor and for women? Empirical evidence from the 2005 Local Government Elections in Pakistan.

Birner, Regina ; Qureshi, Sarfraz Khan ; Akramov, Kamiljon 等

I. INTRODUCTION

Decentralisation is associated with the hope that "bringing government closer to the people" will improve the provision of public services by increasing people's voice in decision-making, and by making the government more accountable to them. Decentralisation is also associated with the hope that disadvantaged groups of society, including the poor and women, will have better possibilities to exercise voice at the local level. Some countries have combined decentralisation with affirmative action, for example, by reserving seats in local councils for women and other disadvantaged groups. Yet the empirical evidence regarding the impacts of decentralisation has been mixed [Bardhan (2002); von Braun and Grote (2002); Jutting, et al. (2004); Steiner (2005)]. In many cases, political decentralisation has not been associated with fiscal and administrative decentralisation, thus limiting the scope of what local governments can actually do. Building capacity at the local level and overcoming coordination problems has been another challenge. Local elite capture has been identified as a major problem that can prevent positive effects of decentralisation for the poor, especially in societies with hierarchical power structures at the local level [Bardhan (2002)]. With regard to gender, there are concerns that decentralisation--even if associated with affirmative action--will not be sufficient to overcome gender-based discrimination. Again, the empirical evidence is mixed [ADB (2004)]. Chatthobadhay and Duflo (2004) found that that women who were elected as village leaders under the reservation policy in the Indian states of West Bengal and Rajastan invested more in those public goods that more closely linked to women's concerns, such as drinking water. Baden (1999) showed that it depends on local power structures and on the availability and competition over resources whether or not women benefit from decentralisation. In view of the mixed results, important knowledge gaps remain regarding the possibilities to promote public service provision for the poor and for women through political decentralisation and associated affirmative action.

This paper examines one major mechanism by which decentralisation can help to improve public service provision for the poor and for women: voting. Three questions are addressed in this paper: (1) To what extent do the poor and to which extent do women use their vote in local government elections to exercise voice regarding public service provision? (2) Which factors influence their decisions to make use of this mechanism? and (3) What are the constraints faced by elected council members to represent the interests of poor constituencies and of women in local governments? The analysis is based on a voter and candidate survey, which the Pattan Development Foundation conducted before and after the 2005 Local Government elections in Pakistan [Pattan 2006).

Pakistan presents an interesting case for analysing local government elections. The country engaged in a far-reaching decentralisation process in 2001. The hopes that decentralisation will improve public service provision have been particularly pronounced, because other strategies to reach that goal had limited success. In fact, Pakistan's track record of public service provision has been rather poor prior to decentralisation [ADB/DFID/World Bank (2004)]. As an indication, growth in Pakistan had been associated with fewer improvements in health and education outcomes than in comparable developing countries. The female-male literacy gap in Pakistan even increased with economic growth while it declined in comparable countries. Likewise, infant mortality declined at a much lower rate in Pakistan than in comparable countries. Moreover, the link between increased funding and improvements service provision was weak as compared to other countries. Thus, neither growth nor increased public spending was likely to improve rural service provision [ADB/DFID/World Bank (2004)]. Against this background, the decentralisation reforms hold considerable promise to address to overcome the long-standing problems of public service provision in Pakistan.

Voting is obviously only one step in an impact chain that links political decentralisation with pro-poor public service provision and gender equity. However, it is an essential step, and--in fact--the major mechanism by which political decentralisation can be expected to lead to improved outcomes. Hence, a better empirical understanding of this mechanism, and the factors that influence how well it works for the poor and for women is an important element in assessing the success of decentralisation reforms.

The paper is organised as follows: Section 2 of presents a conceptual framework that maps the mechanisms that are essential for achieving improved public service provision through decentralisation and locates the role of voting in this context. Section 2 also describes the decentralisation reform in Pakistan in terms of this framework. Section 3 presents the results of the empirical analysis, and Section 4 concludes.

2. DECENTRALISATION AND PUBLIC SERVICE PROVISION

2.1. Conceptual Framework

Figure 1 presents a conceptual framework that maps the mechanisms by which decentralisation can improve the performance of public service provision. Major public services in the provision of which local governments could become involved include, among others, water and sanitation, basic health and education, agricultural advisory and veterinary services, irrigation management and forestry. Likewise, decentralisation may shift responsibilities for the provision of rural infrastructure, such as roads, to local governments. Performance of service provision can be measured by using service-specific indicators, which capture (a) the quality of the services provided, (b) the access of different groups of society to these services, and (e) the efficiency of service provision. As Figure 1 (see box at the right-hand side of the diagram) indicates, improved performance and gender equity in public service provision can be considered as an important factor in influencing overall development outcomes. To analyse the factors that influence the performance of public service provision, it is useful to distinguish between (a) the capacity and the incentives of the service providers to finance and supply these services to the poor in a gender-equitable way (supply-side factors), and (b) the ability of citizens--independently of income statues and gender--to demand these services and hold service-providers accountable (demand-side factors). As shown in Figure 1, both factors are influenced by the extent of political, administrative and fiscal decentralisation.

[FIGURE 1 OMITTED]

The term political decentralisation refers to the shift of authority to elected local governments. Administrative decentralisation describes the transfer of administrative staff to local governments. Fiscal decentralisation refers to the level of authority that local governments have in raising revenues and making autonomous expenditure decisions regarding their own revenues and regarding transfers from higher levels of government.

Figure 1 indicates that next to decentralisation, there are other approaches that can contribute to improved public service provision. On the demand-side, measures to increase voice and accountability may include affirmative action, such as the reservation of seats for women and other disadvantaged groups in local councils, as mentioned above. Other demand-side measures include participatory planning methods, social auditing and the right to information. The effectiveness of these approaches depends on their suitability to reduce obstacles to political participation that are associated with the characteristics of the local population, such as the prevalence of gender discrimination and patron-client relationships ("good fit" link in Figure 1).

On the supply side, measures that can accompany fiscal and administrative decentralisation may include civil service reforms, improved systems of public expenditure management, outsourcing of service provision, and public-private partnerships. Specific measures to address problems of inequity and social exclusion may include training and awareness creation among the staff of service providers. According to the theory of representative bureaucracy, increasing the percentage of female staff and staff representing poor and disadvantaged groups may also help. The effectiveness of these approaches depends on the extent to which they are able to address the specific problems that account for low performance and social exclusion of public service provision ("good fit" link in Figure 1).

This framework indicates that voting is an essential factor, though not the only factor, by which the poor and disadvantaged groups can exercise demand for a better and more inclusive public service provision. Voting is also a major factor by which citizens can hold public service providers accountable. However, voice and accountability will only lead to better service provision, if the public administration has the capacity to improve service provision. Hence, one can hypothesise that the perception of citizens regarding this capacity has an influence on their willingness to vote, and to consider service provision on their voting behaviour. The framework suggests that characteristics of the local communities, for example, the extent to which they consider it appropriate for women to make their own decisions regarding voting, also influences voting behaviour.

2.2. Supply- and Demand-side Approaches to Pakistan's Devolution of Power Policy

The decentralisation reforms that the Musharraf government announced in 2000 ("Devolution of Power" Plan) and enacted with the Local Government Ordinance in 2001 entail both demand-side and supply-side measures. A three-tier local government structure was created, consisting of the Union, Tehsil and District level. There is an elected council at each level, which is headed by a Nazim. The electorate for the union councils comprises all registered voters who are Pakistani citizens and 18 years or older. They directly elect the Union Council members and the Union Nazim on a non-party basis. The elections for the Tehsil and District-level council members and Nazims are indirect. These bodies are elected by the Union Council members. An important demand-side measure is the reservation of 33 percent of seats for women in all tiers of local government. Seats are also reserved for peasants and workers. The government also introduced other demand-side measures, such as social audits [Saleem (2006)].

Elected local governments existed prior to the devolution in 2001, but they did not play any significant role, as they were practically inactive, especially in rural areas [Cheema, et al. (2004)]. Most state functions were carried out by the Provincial administration. As a supply-side measure, the devolution policy transferred the responsibility for most public services that were previously under the local provincial administration to the local governments. This administrative decentralisation process was, however, not uniform across all functions, with significant heterogeneity in its extent across different administrative departments and across different services within a department. As an important measure to create accountability, the bureaucracy is now responsible to elected heads of local government at all three levels. In the previous system, the Deputy Commissioner--a key figure in local government--used to report to the non-elected provincial bureaucracy. Fiscal decentralisation, another supply-side measure, was promoted, as well. The expenditure budget share of local governments in aggregate provincial and local government spending increased from approximately 10 to 30 percent [Cheema, et al. (2004), p. 35]. The financial transfers to local governments are based on a formula, which takes population and backwardness into account.

3. EVIDENCE FROM THE 2005 LOCAL GOVERNMENT ELECTIONS

3.1. Data and Methods

The results presented here are based on a pre-election and a post-election voter and candidate survey of the 2005 Local Government Elections, the second round of elections since the 2001 devolution was enacted. The survey was conducted by the Pattan Development Organisation [Pattan (2006)]. The sample size was 3,792. A three-stage sampling procedure was used. At the first stage, 59 districts were sampled with their probability proportionate to their size. At the second stage, districts were stratified according to urban and rural locations. At the third stage, the respondents were selected from the voter list, using a random start number and the random walk method. Due to the low population density in Balochistan, a booster sample was drawn for this province. For the analysis, the sample was weighted accordingly [Pattan (2006), p. 7]. Pattan also conducted a pre- and post-election survey of the Union councilors to cover the indirect elections. This paper presents descriptive statistics derived from survey, focusing on a comparison of the results by poverty groups and by gender. Thus, the paper provides a basis for subsequent studies, which will employ analytical statistics, such as logistic and multi-level models [Akramov, et al. (2007)].

3.2. Poverty Status of Respondents

In order to analyse to which extent the poor use their vote as voice, it is necessary to disaggregate the survey data by poverty groups. The voter survey featured two questions, from which the poverty status of the respondent can be derived. One question referred to the monthly income of the household. As this question is difficult to answer for respondents that do not receive a monthly wage, such as farmers, the prevalence of missing data is rather high (13.5 percent). The second question grouped the respondents into the categories described in Table 1. Approximately half of the respondents reported that they have only enough income to provide adequate food and shelter for the family (Category 2 in Table 1).

Considering official poverty figures, one can assume that more people opted for this category than would actually classify. In view of the high level of missing data for the question on monthly income, the categories of Table 1 are nevertheless used as a basis for the analysis. Three groups are distinguished: (1) "very poor" (food insecure), corresponding to Category 1 in Table 1, (2) "poor/medium" (Category 2 in Table 1), and (3) 'better-off" (Categories 3, 4, and 5 in Table 1 combined).

3.3. The Decision to Vote

As stated above, participation in local government elections is a necessary, though not a sufficient condition, to be able to use the right to vote as a mechanism to exercise voice regarding public service provision and to hold local governments accountable. According to the official figures of the National Election Commission, voter turnout in the Local Government Elections declined from 52.3 percent in the 2001 elections to 47.4 percent in the 2005 elections [ECP (2005)]. Voter turnout reported in post-election surveys is usually higher than the actual voter turnout, because respondents often feel a responsibility to vote and give an expected answer. This was also the case in the Pattan survey, in which the reported voter turnout was 56.5 percent. Table 2 compares the figures from the Election Commission and the Pattan survey by Province and gender. The highest difference (16 percent) in voter turnout between the Election Commission data and the Pattan data was observed for Balochistan. Still, the differences seem modest and do not suggest a major bias in the data used for this study.

Table 2 shows that women's voter turnout was considerably lower than men's voter turnout in all Provinces. The difference was particularly pronounced in NWFP, followed by Sindh. At the national level, 39.5 percent of the women voted, as compared to 56.8 percent of the men.

Table 4 displays the voter turnout for location, age group, literacy and education, disaggregated by gender. The data indicate that both men and women in rural areas were more willing to vote than people in urban areas. A higher percentage of small and medium-sized farmers voted than any other occupational category. Voter turnout was higher among illiterate men and women than among literate ones.

Participation increased with age, both among men and women. Among the different professional groups, the largest difference in male and female voting behaviour was observed among professionals. The percentage of professional women who voted was particularly low. Answering the question "In your opinion, did the results of the recent local government elections accurately reflect the way people voted or do you think there was rigging?", 52 percent of the respondents felt that the elections reflected the way that people voted, 25 percent were of the opinion that this is not the case, and the rest was undecided. Table 5 displays the differences by income group and gender with regard to this question. Less than half of the very poor and less than half of the female respondents believe that the elections reflect the way that people voted.

[FIGURE 2 OMITTED]

Since the topic of this paper deals with performance of public service provision, it appears useful to present people's perceptions in this regard. Figure 2 displays the satisfaction of the respondents with various aspects of the performance of the previous District Nazim (2001-2005). The figure shows that in general, satisfaction with services is rather low. Law and order received relatively high scores, whereas poverty alleviation and control of corruption received relatively low scores. This result implies that people do not appear to believe that the development work carried out is particularly poverty-focused. Interestingly, there were no major differences between income groups and between men and women in regard to satisfaction with different services.

Table 6 shows the answers to the question: "When the new union council sets development priorities for your area, how much attention do you think they will pay to the problems of ordinary people? Only 8.7 percent of the very poor feel that Union Councils pay a lot of attention to common people's problems, whereas 17.5 percent of the better-off respondents feel that this is the case. There are also gender differences regarding this perception, as shown in Table 6.

Table 7 displays the reasons that people gave for not voting. The table reveals that--to a considerable extent--the very poor and women did not vote because they did not have an ID card, were not registered as voters, or because their name was not in the electoral list. Around 5 percent of women were stopped by their husbands. Lack of possessing an ID card was also a major reason why people could not vote.

The fact that women were stopped by the community or their husbands from voting is related to opposition against women's participation in elections. During the 2001 Local Government elections, thousands of women seats were uncontested due to local opposition to female participation in the political system. In whole districts like Kohistan, Lower Dir and Battagram in NWFP and Dera Bugti in Baluchistan, no female candidates were allowed to contest elections following joint decisions made by the local leaderships of all major political parties [Qureshi (2006)]. Table 8 displays the percentage of respondents who answered that they oppose women's participation in elections, when answering the question: "Do you oppose or support women's participation in election?"

The survey showed that the likelihood of opposition to female participation was greater in case of illiterate women and those that only passed primary school than those with higher educational status. The same was true for male respondents. Similarly, the survey established that the likelihood of opposition to women participation in elections increases with age.

Another problem related to the decision to vote is the electorate's knowledge and awareness regarding the voting system. This question was asked in the pre-election survey, which did not have a comparable poverty group classification than the post-election survey. Table 9 displays the results by gender. Only 9.6 percent of the female voters knew who would vote for Tehsil Nazims and for candidates on reserved seats for District and Tehsil Councils. In contrast, almost half of the men were aware of this important aspect of the voting system. Moreover, in percentage terms, twice as many men than women knew the correct number of ballot papers to be given to voters on polling day.

3.4. Considerations in Voting

While the previous section dealt with the factors associated with the decision to vote, this section is concerned with the considerations that people who voted took into account when deciding for whom they should vote. Table 10 presents the descriptive statistics on the following question: "Now thinking about who you voted for union Nazim, please tell me which of the following factors were important in deciding who to vote for?" Multiple answers were possible. The table indicates that more than half of the voters in each income and gender category consider in their voting decision the plans of the candidate to work on issues that are important to the respondent and his or her family. This parameter is the best available indicator in this data set that citizens use their vote to exercise voice in public service provision. The table shows that next to performance-oriented voting, identity-based voting played an important role, as well. Almost half of the respondents consider the ethnic or family background of the candidate important for their voting decision. Around 70 percent of the respondents state that their own assessment of the candidate is important, but a considerable share of the voters also considered the opinions of others, including the landlord, the community and husband or father as important. The percentage of the very poor who pay attention to the opinion of landlord and his political faction was larger than that of the less poor groups. The integrity and honesty of the candidate was apparently also important across income groups and gender. Since multiple answers were allowed, it is not possible to judge which of the factors listed in Table 10 were decisive for the voters.

3.5. The Role of the Candidates

Using voting to exercise voice for better service provision and for more gender equity will be more effective, if the candidates attach priority to these issues. In the candidate survey, 25 percent of the respondents were women. The pre-election survey of the candidates does not have the same poverty groups as the post-election voters survey quoted above. Therefore, reported monthly household income is used to form income groups, in spite of the problem of missing variables. Two groups of candidates are formed: (a) candidates who reported an income of less than 3,000 Rs (16.6 percent of the candidates), and (b) candidates who reported a higher income. In the pre-election survey of the candidates, Pattan asked the question: "What are the three major local priority issues that you are highlighting in your election campaign?" Table 11 displays the results.

In general, poor and non-poor as well as female and male candidates have similar priorities. Expectedly, poor candidates place more emphasis on poverty reduction. Interestingly, this is also true for female candidates. For all candidates, street pavement, education and water and sanitation have a rather high priority. Rights issues, in contrast, are not high on the agenda. Only 1.5 percent of the male candidates mentioned women issues and violence against women among their three top priorities. In the case of female candidates, the respective figure was 8.9 percent. Expectedly poor candidates place more emphasis on labour rights than other categories of candidates. Health and water, which are often considered to be a higher priority for women, were less frequently among the top priorities of female candidates than of male candidates. Health was surprisingly low on the agenda of the low-income candidates.

Table 12 displays the educational statues of the candidates. More than half of the poor candidates and more than half of the female candidates who were included in the sample of the Pattan survey were illiterate. The overwhelming majority of the women candidates described themselves as housewives.

In addition to challenges in terms of educational background, female councilors face other constraints that may limit their possibilities act effectively. Pakistan's devolution laws lack provisions for the inclusion of women in budgetary committees or for an equal distribution of funds to male and female councilors. Moreover, the law does not ensure female participation as male relatives of female councilors continue to represent them at council meetings. There are no women rights committees at the district level. Female candidates also have to cope with the burdens of security and mobility. In most parts of the country, women continue to face difficulties in moving freely in their constituencies. Moreover, in many areas, women councilors have to cope with the biases and prejudices of their male colleagues. In one instance, the male councilors in Malakand lodged a formal complaint against a woman member for actively participating in council sessions [Naz (2005), p. 57]. Security is another major problem for women councilors. There have been instances where female members have even been murdered e.g., in Kohat, Peshawar and Abbottabad [Qureshi (2006)].

4. CONCLUSIONS

The results indicate that there are considerable obstacles for citizens to use their vote as a mechanism to exercise demand for improved service provision and gender equity and to hold public service providers accountable. In particular, administrative barriers to voting, specifically the lack of identity cards and incomplete registration in election lists prevent citizens from voting. Poor people and women are particularly affected by these problems. Moreover, only half of the respondents believe that the electoral outcomes accurately reflect people's voting decisions. On the positive side, income status and residence in rural areas do not seem to discourage people from voting. NWFP is an exception, however. In this Province, voter turnout among the very poor was particularly low. In general, illiterate citizens and citizens from rural areas were more likely to vote than literate and urban citizens. Expectedly, men showed a higher voter turnout than women. Among the people who voted, one can observe that both identity-based voting and performance-based voting played a role. It appears that women, very poor people and people in rural areas are more somewhat more inclined to use their vote to improve their living conditions. More than 40 percent of the very poor consider the opinion of the landlord and his political faction in their voting decision, which can be considered as an indication of the elite capture problem.

One can expect that the prospects of low-income and of female candidates to be effective in reducing poverty and gender inequity will be limited due to a variety of constraints, including a lack of provisions for their representation in committees. Moreover, in view of the high level of illiteracy among both low-income and female candidates, training may be particularly important to improve their effectiveness. In conclusion, there is a need to make to improve the electoral process so that more people, and especially more poor people and more women, have the possibility to use their vote to exercise voice. There is also a need to improve the working conditions for low-income and for female candidates to become effective representatives of their constituencies once they are elected.

The paper dealt with voting, which is only one mechanism by which decentralisation can lead to better service provision. As the framework presented in the paper shows, there is a need to study voting in the context of other measures that improve the ability of citizens to demand better services, such as social audits. Moreover, the capacity and incentives of decentralised government agencies to improve public service provision with a poverty and gender-focus need to be studied, as well. Future research may focus on explaining differences in the performance of local governments, taking these factors into account. Analysing different demand-side and supply-side measures in an integrated context will hopefully help policy-makers and stakeholders to better understand "what works where and why" in improving local governance.

REFERENCES

ADB (2004) Gender and Governance Issues in Local Government--Regional Report of Technical Assistance in Bangladesh, Nepal, and Pakistan. Asian Development Bank. Manila.

ADB/DFID/World Bank (2004) Devolution in Pakistan. Main Report. Islamabad.

Akramov, K., S. K. Qureshi, R. Birner, and B. H. Khan (2007) Decentralisation, Local Government Elections, and Voter Turnout in Pakistan. International Food Policy Research Institute (IFPRI), Washington, DC. (Manuscript.)

Bardhan, P. (2002) Decentralisation of Governance and Development. Journal of Economic Perspectives 16:4, 185-206.

Bardhan, P. and D. Mookherjee (2000) Relative Capture of Government at Local and National Levels. American Economic Review 90:2.

Chatthobadhay, R. and E. Duflo (2004) Women as Policy Makers--Evidence from a Randomised Policy Experiment in India. Econometrica 72:5, 1409-1443.

Cheema, A., A. I. Khwaja, and A. Qadir (2005) Decentralisation in Pakistan: Context, Content and Causes, Kennedy School of Government. (Faculty Research Working Papers Series, RWP05-034, April.)

ECP (2005) Local Government Elections 2005. Province-wise Turnout. http://www.ecp.gov.pk/content/lgturnout2005.html; accessed 12/10/06.

Inayatullah (1999) Moving towards Change, Savera. Islamabad.

Jutting, J., C. Kauffmann, I. McDonnell, H. Osterrieder, N. Pinaud, and L. Wegner (2004) Decentralisation and Poverty in Developing Countries: Exploring the Impact. Paris, France. (OECD Working Paper No. 236.)

Naz, R. (2005) Training of Women Councilors in Pakistan: A Review, in Pakistani Women in Context: A Companion Volume to the Pakistan Country Gender Assessment.

Pattan (2006) Common Grounds-l: Analysis of Local Government Election 2005. A Report based on Survey of Voters. Pattan Development Organisation, Islamabad

Qureshi, S. K. (2006) Engendering Growth for Poverty Reduction--The Pakistan Experience. Paper presented at the International Conference on "Poverty, Equity and Access to Justice of the Poor in South Asia" at the Centre for the Study of Law and Governance, Jawaharlal Nehru University, New Delhi, November 7-9.

Saleem, M. (2006) Accountability Under Local Government System. National Reconstruction Bureau. Paper Presented at the 22nd Annual Conference of the Pakistan Society of Development Economists, Lahore, December 19-22.

Steiner, S. (2005) Decentralisation and Poverty Reduction: A Conceptual Framework for the Economic Impact. German Overseas Institute, Hamburg, Germany. (Working Paper No. 3.)

Von Braun, J. and U. Grote (2002) Does Decentralisation Serve the Poor? Chapter 4 in E. Ahmad and V. Tanzi (eds.) Managing Fiscal Decentralisation. New York: Routledge.

Waseem, M. (2006) Democratisation in Pakistan, A Study of the 2002 Elections. Oxford University Press.

REGINA BIRNER, SARFRAZ KHAN QURESHI, KAMILJON AKRAMOV, and BILAL HASSAN KHAN

Regina Birner <r.birner@cigar.org> is Senior Research Fellow at the International Food Policy Research Institute, Washington, DC. Sarfraz Khan Qureshi is Chief Executive Officer at Innovative Development Strategies, Islamabad. Kamiljon Akramov is Research Fellow at the International Food Policy Research Institute. Washington, DC. Bilal Hassan Khan is a freelance consultant based at Islamabad.

Table 1

Poverty Groups in Post-election Voter Survey

Question: There are several ways to describe
a household's financial situation. Please Average
tell me which of these statements best Monthly
describes the situation in your household. Percent Income *

1. We sometimes do not have enough income to
 eat three meals a day 12.1 2,750
2. We have enough income only to provide
 adequate food and shelter for the family 50.5 4,680
3. We have enough income to provide food and
 shelter and to buy new clothing from time
 to time 25.9 9,260
4. We have money left over from time to time
 after meeting our basic needs 8.2 12,790
5. We always have money left over after
 meeting our basic needs 2.6 25,230
 Missing 0.6 --

Source: Pattan (2006) Post-Election Survey of voters.

* Reported monthly family income in Rs In 2005, the poverty line
was approx. Rs 930 per adult household member per month. The
survey asked for the number of children below age 15, but not for
the entire household size. Hence, it is difficult to compare the
reported average monthly income of the household with the poverty
line.

Table 2

Voter Turnout by Gender and Province

 Election Commission Pattan Survey

 Male Female Total Male Female Total

Punjab 61.5 44.0 53.5 68.6 55.2 62.2
Sindh 47.6 24.3 37.6 57.9 37.1 50.3
NWFP 57.4 21.1 41.2 72.7 26.6 48.1
Balochistan 39.3 21.5 31.1 52.7 41.5 47.4
Total 56.8 35.9 47.4 65.8 46.3 56.2

Sources: ECP (2005), Pattan (2006): Post-election voters' survey.

Table 3

Doter Turnout by Income Group and Province

 Very Poor
 (Food Insecure) Poor/Medium Better-off

Punjab 59.0 61.8 63.9
NWFP 22.1 45.0 66.2
Sindh 58.3 52.5 41.3
Balochistan 52.3 42.3 51.6
Total 55.8 56.1 57.2

Source: Pattan (2006) Post-election voters' survey.

Table 4

Voter Turnout by Category and Gender

 Female Male
Location
 Urban 39.3 57.3
 Rural 49.7 70.2
Age Group
 18-20 18.3 28.0
 21-50 49.3 68.4
 More than 50 66.3 72.9
Literacy
 Literate 39.1 63.5
 Illiterate 51.0 72.2
Occupation
 Self-employed Shopkeeper 46.9 63.0
 Self-employed Business or Trade 39.3 65.1
 (except shop keeping)
 Government Servant 41.4 55.9
 Private Sector Employee 50.7 45.7
 Manual Labour 47.1 50.6
 Small or Medium Sized Farmer 46.9 63.6
 Large Farmer / Landowner 48.8 48.4
 Professionals (Doctor, Lawyer, etc.) 32.1 75.5
 Retired 53.0 48.6
 Housewife 34.6 --
 Unemployed 46.7 70.3
 Others 46.2 44.8

Source: Pattan (2006) Post-election voters' survey.

Table 5

Perceived Fairness of the Elections by Income Group and Gender

 Income Group Gender

 Very Poor/
 Poor Medium Better-off Female Male

Reflect the Way People 44.9 54.7 50.9 45.9 57.7
Voted
There was Rigging 25.3 21.7 29.2 22.9 26.7
Don't Know 29.8 23.5 19.9 31.2 15.5

Source: Pattan (2006) Post-election voters' survey.

Table 6

Voters Perception Whether Union Councils Take Common People's
Problem into Account when Setting Development Priorities *

 Income Group Gender

 Very Poor Poor/Medium Better-off Female Male

A lot 8.7 14.6 17.5 13.4 16.4
Some 24.3 33.1 31.9 26.4 36.4
Very Little 17.8 18.0 19.7 18.1 18.9
None at All 28.4 20.7 18.6 27.3 14.8
Don't Know 19.9 13.1 11.8 14.1 13.0
Refused 0.9 0.6 0.5 0.7 0.5

Source: Pattan (2006) Post-election voters' survey.

* The question was formulated as follows: "When the new union
council sets development priorities for your area, how much
attention do you think they will pay to the problems of ordinary
people? Do you think they will pay a lot of attention, some
attention, very little attention or no attention at all?"

Table 7

Reasons Not to Vote by Gender and Income Group
(Percent of Respondents)

 Income Group

 Very Poor/
 Poor Medium Better-off

Did not have ID card 40.9 35.4 21.1
Not registered as voter 10.5 12.5 13.0
Went to polling station but my name
 was not on the electoral list 15.6 13.2 13.9
Went to polling station but queue
 too long 0.0 0.4 1.3
No transport 0.5 0.1 0.4
Stopped by community 2.1 1.8 1.5
Stopped by my husband 6.8 2.8 2.4
Fear of violence 0.0 0.3 0.2
Too hot / bad weather 0.6 0.5 0.9
I was out of town/village 7.7 7.8 10.9
Had to work 2.4 4.7 5.8
Didn't want to cast vote 2.2 2.5 6.4
Elections don't matter 3.3 2.4 4.7
No reason in particular 2.1 3.2 6.0
Other 5.4 11.5 10.9
Don't know 0.0 0.3 0.2
Refused 0.0 0.5 0.4

 Gender

 Female Male

Did not have ID card 38.6 19.5
Not registered as voter 12.5 12.2
Went to polling station but my name
 was not on the electoral list 10.6 18.1
Went to polling station but queue
 too long 0.8 0.5
No transport 0.5 --
Stopped by community 2.9 --
Stopped by my husband 5.3 --
Fear of violence -- 0.6
Too hot / bad weather 0.6 0.8
I was out of town/village 4.4 15.4
Had to work 3.0 7.4
Didn't want to cast vote 3.6 4.2
Elections don't matter 3.0 4.1
No reason in particular 4.2 3.8
Other 9.1 12.8
Don't know 0.5 0.2
Refused 0.4 0.4

Source: Pattan (2006) Post-election voters' survey.

Table 8

Percent of Respondents who Oppose Women's Participation in Elections

 Female Male
Location
 Rural 19.3 19.7
 Urban 12.9 18.4
Education
 Illiterate 19.5 25.7
 Primary or Middle 15.6 19.2
 Matric or FA/Fsc 11.1 16.4
 More than FA/Fsc 10.8 10.9
Age group
 18-20 11.6 18.5
 21-30 14.4 17.7
 31-40 17.1 19.1
 41-50 19.3 20.8
 More than 50 22.8 22.0

Source: Pattan (2006) Pre-election voters' survey.

Table 9

Knowledge about the Voting System (by Gender)

(a) Q: "Who Votes for District and Tehsil Nazims and Reserved Seats
for District and Tehsil Councils"

 Female Male

Voters like Myself 39.8% 33.4%
Union Nazimeen and Councilors 9.6% 46.1%
Others 2.6% 0.9%
Don't Know 47.9% 19.7%
(b) Knowledge about Provision of Ballot Papers
 to Each Voter on Polling Day
Incorrect 95.5% 90.5%
Correct 4.5% 9.5%

Source: Pattan (2006) Pre-election voters' survey.

Table 10

Considerations in Voting by Income Group and Gender

 Very Poor/
 Poor Medium Better-off

Personality of the candidate 75.8 89.6 89.5
Candidate's religious views 60.0 68.5 66.6
Candidate's plan to work on issues
 important to respondent and family 65.6 62.5 57.6
Ethnic or family background of candidate 51.6 53.4 49.5
Opinions of husband/father 66.0 59.3 55.1
Opinion of local landlord or his
 political faction 43.1 38.3 27.9
Own opinion or assessment of the
 candidate 71.3 78.1 80.0
Integrity or honesty of the candidate 78.5 83.7 83.3
Joint decision of village /
 neighbourhood 67.4 70.1 60.5
The candidate/s belong to your
 same biradar 34.4 38.2 36.9
The political party of the candidate 43.8 41.8 43.6
Personal relations with candidate 52.5 64.6 62.6
Others 100.0 100.0 88.8

 Female Male

Personality of the candidate 81.8 91.9
Candidate's religious views 70.5 64.8
Candidate's plan to work on issues
 important to respondent and family 66.1 57.9
Ethnic or family background of candidate 51.9 51.4
Opinions of husband/father 84.2 42.2
Opinion of local landlord or his
 political faction 36.1 33.8
Own opinion or assessment of the
 candidate 70.2 82.9
Integrity or honesty of the candidate 79.6 85.1
Joint decision of village /
 neighbourhood 62.9 67.9
The candidate/s belong to your
 same biradar 36.7 37.6
The political party of the candidate 32.4 48.5
Personal relations with candidate 47.7 71.0
Others 85.7 96.3

Source: Pattan (2006) Post-election voters' survey.

Table 11

Campaign Issues Mentioned by Candidates as Priority

 Income Group Gender

 Below Above
 3,000 * 3,000 * Female Male
Economic Issues
 Priority Reduction 22.1 10.7 23.4 11.2
 Employment 5.2 8.4 7.8 8.7
 Inflation 2.5 0.8 2.3 0.7
 Food Security 0.3 0.5 1.2 0.4
Government Services
 Education 34.4 40.4 4.0 39.4
 Health 16.7 28.6 25.7 28.4
 Water 20.3 36.9 28.3 35.4
 Sanitation 30.8 42.1 32.1 43.7
 Street Pavement 58.8 62.6 53.4 63.1
 Electrification 19.4 31.2 18.1 33.0
Rights
 Women Issue/Violence Against 3.0 3.3 8.9 1.5
 Women
 Minority Rights 2.3 2.6 0.7 3.2
 Labour Rights 8.4 5.5 3.6 6.7
 Tenant Rights 4.0 3.0 3.1 3.7
Governance
 Law and Order/Crime Control 5.1 5.4 2.7 6.4
 Tackling Political Corruption 0.0 1.5 0.5 2.0
 Religious Issues
 Strengthening Religion 0.0 0.3 0.0 0.6
Other 22.8 26.0 42.2 24.4

Source: Pattan (2006) Pre-election survey of candidates.

* Candidates who reported a household income of more/less than Rs
3,000.

Table 12

Educational Status of Candidates

 Income Group Gender

 Below Above
 3,000 * 3,000 * Female Male

Illiterate 53.3 13.0 53.3 8.2
Primary 25.4 12.4 15.8 14.0
Middle 8.2 15.7 7.3 16.3
Matric 12.0 26.7 9.7 28.8
F.A/F.Sc 1.0 11.5 6.4 11.5
B.A/B.Sc or more 0.0 11.6 5.1 11.3
M.A or a Pofessional Degree 0.0 8.2 0.8 9.1
Doctorate or Post-Doctorate 0.0 0.5 1.6 0.2
Refused 0.0 0.5 0.0 0.5

Source: Pattan (2006) Pre-election survey of candidates.

* Candidates who reported a household income of more/less than Rs
3,000.