Women's autonomy and happiness: the case of Pakistan.

Ali, Syed Mubashir ; ul Haq, Rizwan

It is generally believed that "autonomy" brings happiness and satisfaction in women's lives. In this study we examine whether or not the established autonomy indicators are a source of "happiness" for Pakistani women. By using the nationally representative data, only two indicators, i.e., "women's education" and "decisionmaking authority", prove to be important factors in finding "very happy" status in women's life. Additionally, "possession of assets" also proves to be an important factor in providing the "very happy" status in a women's life. However, the "possession and utilisation o[" assets" and "going alone outside the house" are not important indicators of a "very happy" status in women's life in Pakistan and "Labour force participation" is indicative of unhappiness.

The results of this study show that not all established indicators of" autonomy bring about happiness in the lives of Pakistani women. This is because Pakistani society differs from other societies, in particular the western society, and hence the concept of "autonomy" in bringing about "happiness" in the lives of Pakistani women yields effects different from those in other societies. Thus, there is a need to focus on the advocacy of" only those autonomy variables which lead to happiness in a woman's life, which is the end-goal for women, who form a vital part of the society.

INTRODUCTION

An important human right of both men and women is pursuit of happiness and satisfaction in life. It is generally believed that "autonomy" brings happiness and satisfaction in women's lives [Ghuman (2003); Jejeebhoy (2002); Siddiqui, et al. (2006)]. The fact remains that happiness and satisfaction in life are subjective qualities and hence women living in different circumstances with or without autonomous status may be happy and satisfied with their lives. According to a recent world value survey of people in 65 countries, the world's happiest people lived in Nigeria, a volatile, poverty-stricken country. The prosperous Americans were placed a dismal 16th on the list [Richard (2004), p. 16].

Pakistan is a patriarchal society, bound by age-old traditions of a subservient and subordinate role of women. A girl child in Pakistani society often grows up learning to serve and obey men. Most of the time, decisions relating to her career, marriage, number of children to be borne etc. are to be made by a male--her father, her brother, and then her husband after her marriage. According to the Gender Development Index (GDI), Pakistan has been rated the second poorest (0.471) among South Asian countries, where the average index is 0.554 [MHDC (2004)]. Various surveys have shown that Pakistan has the second highest gender gap in school enrolment rates in the world [cited in Nasir (2002)]. Also the population sex ratio in Pakistan is 108 males per 100 females Pakistan (2000) which is opposite to that of most countries in the world where women outlive men. In Pakistan, even at older ages the sex ratio is substantially high [Ali and Sultan (2003)]. A partial explanation for excess males in the older ages is that the female life expectancy remained lower than male at later ages [Mahmood (2003)]. In view of the discriminatory attitude of males towards females, many researchers working in this area have emphasised the need for increased autonomy and improved social status for women in Pakistan [Sathar and Kazi (1997); Siddiqui, et al. (2006)].

In the backdrop of the above stated empirical evidence, it is worth investigating the relationship between women's autonomy and their happiness in life--a topic that has been rarely researched in the socio-structural context of Pakistan. This study therefore, attempts to address whether Pakistani women's happiness in life is associated with the established autonomy indicators. Factors such as women's education, employment status, decision making power in the household, access and utilisation of assets, and the ability to leave home alone, are taken as proxies for women's autonomy. Variables such as abuses faced in the house, illness, poverty status, household sanitation status, and region of residence are included in the model so as to control for factors that may influence the level of happiness among Pakistani women. The analytical framework given below explains the relationship between the above stated variables.

[ILLUSTRATION OMITTED]

DATA AND METHODOLOGY

The analysis is based on data from the Pakistan Socio-economic Survey (PSES) Round-2 named MIMAP-II. The survey is nationally representative and was carried out in 2001 in all of Pakistan except the Federally Administered Tribal Areas (FATA), Federally Administered Northern Areas (FANA), military restricted areas and district of Kohistan, Chitral, and Malakand. The population of these excluded areas is around 4 percent of the population of Pakistan. The sampling frame in the rural areas constituted all the villages. The large villages were subdivided into more than one PSU. Urban areas were divided into Enumeration Blocks, each containing on average 250 households.

It may be mentioned here that PSES Round-2 was carried out in the same households visited two years earlier in PSES Round-1. The total number of households visited in PSES Round-1 was 3564. Since PSES Round-1 was based on the sampling frame of 1981, another 1170 households were added in the sample of PSES Round-2 to make it representative of the 2001 population at national and urban-rural level. Altogether, in PSES Round-2, 4021 households were surveyed, of which 2577 were rural and 1444 were urban. The attrition rate in the panel households was about 20 percent.

Each survey team consisted of male and female interviewers headed by a supervisor from the Pakistan Institute of Development Economics. Two separate questionnaires--one for male and the other for female respondents--were administered in each household. The section on gender issues was included in the female questionnaire to elicit detailed information on the autonomy of female respondents in the household aged 15 years and above.

The analysis is carried out on the weighted sample and is based on simple crosstabulation and multivariate techniques. Cross-tabulation is used to show the unadjusted relationship of independent variables with the dependent variable. Logistic Multinomial Regression was applied to test the net impact of variables denoting autonomy and some other variables which may influence the level of happiness of a woman. The Logistic Multinomial Regression is mathematically defined as:

ln p/1 - p = [alpha] + [SIGMA] [[beta].sub.i][X.sub.i] + [u.sub.i] ... ... ... ... (1)

Where p is the probability of happiness in life; [alpha] is the intercept; [[beta].sub.i], are the estimated regression coefficients; [X.sub.i], are the characteristics of women, and [u.sub.i] is the error term.

The dependent variable is named happiness, formed from the answer to the question "Compared with the past, do you feel your present life is very happy, quite happy, or not happy?". The dependent variable is coded 1 for "Not happy", 2 for "Quite happy" and 3 for "Very happy".

As suggested by Hoffman (2004), the reference category of the dependent variable in a Multinomial Logistic Regression Model should be the one with the most common occurrence, which in this case is the category of quite happy. Since the analysis is carried out using SPSS, which treats the last category as the reference category, we recoded the category of quite happy from '2' to '3' and not happy from '1' to '2'. The very happy category is transformed from '3' to '1'.

The independent variables such as going out of home; decision-making power in the household; abuses faced in the family; current employment; illness; sanitation condition (in terms of availability of water, latrine etc. inside the house): urban-rural residence; and poverty status of the household are entered into the mode[ as dichotomous variables, whereas, the age of a woman is entered as an interval variable. Educational status was categorised as illiterate, primary or less, middle and secondary (6 to 10 classes passed), and above secondary. The possession and use of assets is defined as: having no assets, having assets but cannot use them, and having assets and can also use them. The correlation matrix shows that there is no case of high correlation between independent variables (see Appendix Table 1).

RESULTS

As stated earlier, happiness in life is subjective and hence women living in different circumstances may have different levels of happiness. Table 1 portrays a picture of women's present happiness level by their autonomy, socio-economic, cultural, and area of residence status.

It is encouraging that the vast majority (71 percent) of Pakistani women say they are generally happy with their present life. Whereas, 29 percent of women were "not happy". This shows that Pakistani women generally are complacent by nature and hence are satisfied in life. This internalised complacency could be because of the teachings of Islam which instruct its followers to be patient and contented in life. About 96 percent of Pakistanis are Muslim [Pakistan (2000)].

It is generally believed that an important indicator of women's autonomy is her freedom of movement outside her home. Generally, Pakistani women have limited mobility outside their homes. In a study based on the data of rural Punjab-the largest of the four provinces of Pakistan--Sathar and Kazi (1997) find that over two-thirds of women require someone's permission to visit relatives in the village and 90 percent of women require someone's permission to go to the next village. Moreover, when women leave their homes, they generally are escorted by someone who is usually a close relative. The same study estimates that only 12 percent of women could travel alone to the next village, whereas 28 percent of the women could go unescorted to a health centre. A national level survey conducted by the National Institute of Population Studies in 1996-97 reported that 18 percent of women have travelled alone outside their locality or village and about a quarter of women could go to hospital/clinic unescorted [Hakim, et al. (1998)].

The question is whether or not this important indicator of autonomy has any association with the women's happiness level in life. An insignificant chi-square test value (.102) shows that women's happiness level does not have a significant association with women's freedom of movement.

Another important indicator of women's autonomy is her decision-making authority in the household. Decision-making ability is a source of strength and hence a matter of self respect and esteem which may provide happiness in the life of a woman. In Pakistan, women's participation in decision-making varies by the types of decision. For example, when a child is sick, 28 percent of women reported to have decision-making power about the treatment of the child. With regard to buying clothes, 49 percent of women reported to have the final say in such decisions [Hakim, et al. (1998)]. But many a time important and vital decisions are taken by a male member who in most cases is the head of the household. (1)

Interestingly, the happiness level of women has a significant association with the grant of authority or with no authority of decision-making in the household. Overall, the chi-square test value confirms this relationship, as a highly significant association (0.000) is found between the responses on decision-making and happiness.

Education is a well-established indicator of socio-economic status of women. Education exposes a woman to the outside world which may bring about a change in her attitude, thinking and perception about life. That change may also lead to more egalitarian and autonomous behaviour. But demographers like Caldwell (1986) caution that female education could be rendered irrelevant to women's position unless it is contrasted with men's education.

In the case of Pakistan, there is a clear positive relationship between women's educational levels and their being happy in life. Overall, 34 percent of the women who have completed 11+ classes were 'very happy', compared to 12 percent of those who have no education.. The variation in women's educational level and happiness status is also reflected in the chi-square test value of 0.000-a highly significant value.

It is widely believed that women's labour force participation enhances their financial autonomy which in turn leads to happiness and satisfaction in life. However, there is also literature suggesting that women's domestic power depends heavily on social context. For example, Cain, et al. (1979) found that Indian women who participate in income earning work seem to have more domestic autonomy than Bangladeshi wives have. Jain (1979) found that overseas Indian women who were working as wage labourers on Malaysian rubber estates have very little domestic power. He argues that in Malaysia, male dominance was so strong that wives more or less automatically turned over all their wages to the husband, thereby giving him control of the family's most important material resource.

In the majority of the cases, the situation of Pakistani working women is not very much different than their counterparts in Bangladesh or overseas Indian women working in Malaysia. It is probably because of this reason that a lower proportion of working women (66 percent) than non-working women (73 percent) are found to be happy. The differentials observed between the two attributes are found to be significantly associated with each other.

Another dimension of women's autonomy is access and utilisation of assets and resources. It is generally observed that many wealthy women who have access and control over assets and resources gain autonomy and hence have more equal relationship with their husbands.

Table 2 shows that most likely to be happy are women, who have access to assets, whereas the least happy are those women who did not have any assets. A study assessing the magnitude of women's autonomy in relation to access to assets observed that having access alone does not provide a woman with autonomy; rather it is the access and control over these assets which give a woman a real sense of autonomy [Sathar and Kazi (1997)]. The chi-square test value (0.000) shows a significant inter dependence between the two attributes. The data from this study show that for deriving happiness in life, access and control over these assets do not seem to be a sufficient condition. Nevertheless, a rather clear picture of the relationship appears when the net impact of this variable is measured in the next section.

The variable 'abuses faced inside the house' has significant association with the happiness status of the women. For example, the proportion of 'very happy' women is almost double where there is no abuses faced in the house than among women who face abuses inside the house. The variables of 'sickness' and 'sanitation' are also highly significantly associated with happiness. The results show that those women who are not sick and living in houses where the sanitation condition is safe are relatively happier than those who are either sick or living in sanitarily unsafe houses.

Place of residence also shows high association with the happiness status of women. The results show that the women living in urban areas are happier than those living in rural areas. More poor women are found to be 'not happy' as compared to non-poor women. Particularly, the proportion of 'very happy' women was more than double in the non-poor category as compared to poor women. This relationship is statistically highly significant.

Multivariate Analysis Results

We have so far tried to establish the association between women's happiness level and various indicators of women's autonomy as well as her socio-economic, cultural and place of residence. In order to measure the net impact of each autonomy variable on women's happiness we applied Multinomial Logistic Regression technique. The Model-1 includes autonomy variables only. In the second step (Model-2) some socio-economic, regional and cultural variables were added to Model-l. The inclusion of these as control variables was necessitated because of their role in explaining women's happiness. This way the results will portray a more exact picture of autonomy variables in explaining women's happiness in life.

Job Status

The majority of the variables show expected results. One exception is Women's labour force participation, which was entered in the model as a dichotomous variable where non-working women are the reference category. Women who work are hypothesised to be more autonomous than those who do not work. It is further conjectured that these autonomous women are happy with their life. Interestingly, multivariate analysis results show an inverse relationship of women's work participation with their happiness level. In other words, women's work participation makes them dissatisfied and unhappy in life (Table 2, Equation 1). This relationship becomes statistically significant and even stronger in case of the equation 'very happy' relative to 'quite happy' (see Table 2 Equation 2). The direction of relationship and the statistical significance does not change with the addition of socio-economic and cultural variables--hereafter called the 'control variables'--in the Model-2 (Table 3).

The possible explanation for this contrary to expected relationship is that many Pakistani women with their low socio-economic status and poverty experience financial constraints and seek work out of need. They are mostly are involved in low-income menial jobs [Shah (1986); Ali, Siyal, and Sultan (1995); Nasir and Kiani (2005)]. Ironically, the money earned by them is usually taken over by husband or spent to cater for the needs of other household members with almost no money left for themselves that could enhance their status. Moreover, most Pakistani women, besides their gainful work, have to bear the double burden of household and work responsibilities single handedly, which is likely to make them feel unhappy. Due to the above stated reasons, working women report of not having a 'very happy' status in life.

Education

Education is another indicator of autonomy, the attainment of which is conjectured to be a source of happiness among women. The results of the multinomial regression analysis (Table 3) show that in general, educational attainment is related to happiness in a woman's life. Interestingly, educational attainment up to primary level brings about a maximum change in their happiness status. The reason for declining odds of 'very happy' status at higher levels of education could be that the level of consciousness and expectations of those with high education is higher as compared to the ones with primary education, and it is likely that highly educated women do not receive the returns to education as expected. The inclusion of socio-economic, cultural and place of residence variables in the Model-2 has a depressive effect on the magnitude of coefficients at all the three levels of education. Yet the relationship remains statistically significant and in the same direction (Table3, Equation 2). To sum up, one may conclude that attainment of education by women, especially primary level is related to happiness.

Freedom of Movement

A woman's happiness in life is also measured here, by the level of her freedom of movement outside home--an indicator of women's autonomy. This variable shows a weak inverse relationship with the dependent variable, i.e., women's happiness in life (Table 2, Equation 1 and 2). The addition of control variables in Model-2 (Table 3, Equations 1 and 2) reverses the relationship. However, it remains unimportant as the relationship is statistically not significant.

Decision-making Power

Women's decision-making power in the household conforms to the expected outcome, that is, the decision-making power of women is associated with 'very happy status' in their lives relative to 'quite happy status' (Equation 2 in Tables 2 and 3). The relationship remained statistically highly significant and the value of coefficients and odd ratios remains almost unchanged when controlling variables are added. However, this variable remains statistically insignificant in both the models for its relationship with 'not happy' relative to 'quite happy'.

Access and Control of Assets

The degree of women's access to material resources such as land, house/flat, vehicle, jewellery, bank deposit and consumer durables and control over their utilisation is conjectured to provide not only socio-economic status to women, but in turn generate happiness in their lives. This relationship when tested through logistic regression model conforms to this hypothesis. The relationship is not only statistically highly significant at a less than 1 percent level but the magnitude of these coefficients is also quite large (Table 2, Equation 1). However, interestingly in Equation 2 where as the possession of assets remained an important factor in explaining the relationship between "very happy" relative to "quite happy" status of women; the ability and control over asset utilisation loses its statistical significance. The addition of controlling variables in Table 3 does not bring much change except for the relationship between 'very happy' relative to 'quite happy' status of women (Table 3, Equation 2). This relationship loses the statistical significance for the women who have not only the possession but control over the utilisation of assets.

In Pakistani culture, the supremacy of women is generally not tolerated by men, so all women who possess and have control over these assets have to face more pressures and criticism from men folk. This attitude of men may make these women less happy than those who possess the assets only.

Age of Women

The age of women, entered in the model at interval scale shows an inverse relationship with women's happiness in life. This inverse relationship suggests that as women grow older, mounting pressures of responsibilities result in more anxiety, rendering them relatively discontented and less happy in life. Although the magnitude of the coefficient is small, it is statistically significant (Table 3, Equations 1 and 2).

Poverty

In recent years, there has been an increase in the poverty level in Pakistan, from 17 percent in 1987-88 to 33 percent in 2001 [Qureshi and Arif (2001)]. The increase in the levels of poverty undoubtedly affects the lives of people particularly women who are mostly dependent on the earnings of male members of the household. The variable of poverty is included in the model for two reasons: firstly, to measure the relationship of this variable with happiness of women and secondly, to control the effect of poverty so as to test the net affect of autonomy variables.

As expected, the poverty level denoted here in terms of "above poverty line" and "on or below poverty line" emerged as not only a significant variable to explain the happiness level of women but in terms of magnitude of the coefficient, it also brings about a substantial change in the dependent variable i.e., 'not happy' relative to 'quite happy' as well as 'very happy' relative to 'quite happy' status of women (Table 3, Equations 1 and 2). The result clearly suggests that financial well-being generates happiness in the lives of women.

Region of Residence

There is a great divide in urban and rural areas. The region of residence has been included in the analysis, so as to control the impact of the differences in the urban and rural areas. The results of the analysis show that as compared to rural women, urban women are 'quite happy' relative to 'not happy' women (Table 3, Equation l). But this relationship is not statistically significant. Likewise, urban life also brings about a 'very happy' status relative to 'quite happy status among women (Table 3, Equation 2). However, the relationship is not statistically significant here as well.

Conflict/Abuses Faced in the House

Facing conflict and abuse is, unfortunately, common in Pakistani households [Sathar and Kazi (1997)]. Our analysis shows that conflict in the house affects women's happiness. In other words, women who do not face conflict or abuse in the house are more likely to be happy. For example, women who are not facing abuses inside the household are 14 percent less likely to be 'not happy' relative to 'quite happy' as compared to those who are facing abuses. The coefficient is significant at the 10 percent level of significance (Table 3, Equation 1). Similarly, women not facing abuses are 66 percent more likely to be 'very happy' relative to 'quite happy' compared to those who are facing abuses but the coefficient is statistically not significant (Table 3, Equation 2).

Sickness Status

Although good health may not guarantee happiness but bad health certainly makes one feel unhappy and depressed. Pakistani women who are discriminated against men in almost every sphere of life get sick more often than their male counterparts [Mahmood and Ali (2002)]. Illness affects the wellbeing of the population at large but a woman is affected the most as because she has to do not only the household chores single-handedly but has to take additional responsibilities of bearing and rearing the children. This is reflective in the results of this study which shows that as compared to sick women, healthy women are 'quite happy' in life. This relationship is very strong as evidenced by odds ratios as well as statistical significance level at less than 1 percent (Table 3, Equation 1).

But somehow, sickness status of women does not yield a significant effect on the 'very happy' status of women (Table 3, Equation 2). This means that whereas good health guarantees a 'quite happy' status for Pakistani women, it does not necessarily bring about" a 'very happy' status for women in Pakistan.

Sanitation

The provision of potable water and improved sanitation is an important basic right of the population at large. But somehow various national level surveys as cited in of Pakistan, (2000) indicate that coverage and access to water supply facilities in Pakistan range between 50 to 80 percent and sanitation facilities between 40 to 55 percent.

The sanitation variable in this study is formulated on the basis of the availability status of flush or pour flush latrine and water in the house. If these facilities are available inside the house, a good sanitation condition is assumed. Having water and latrine facilities available inside the house is a matter of comfort for women as traditionally, fetching water from a source outside the house is a woman's job. Moreover, she also faces hardship if a toilet facility is not available inside the house. This is reflective in the results of this analysis as sanitation facilities proved to be an important factor in yielding 'quite happy' status for women in Pakistan (Table 3, Equation 1)

DISCUSSION AND CONCLUSION

This study has been undertaken to examine whether or not the established autonomy indicators are a source of 'happiness' for Pakistani women. Only two indicators i.e., "women's education" and "decision-making authority" proved to be important factors in soliciting 'very happy' status in women's life. Additionally, "possession of assets", also proved to be an important factor in providing 'very happy' status in a woman's life. However, the "possession and utilisation of assets" and "going alone outside the house" were not important indicators of a 'very happy' status in women's life in Pakistan and "Labour force participation" is indicative of unhappiness.

The results of this study show that not all established indicators of autonomy bring about happiness in the lives of Pakistani women. This is because Pakistani society differs from other societies, in particular the western society and hence the concept of 'autonomy' in bringing about 'happiness' in the lives of Pakistani women yields different effects than in other societies.

In order to better understand this relationship, one may have to study the contextual effects of Pakistani society which is influenced by a blend of the South Asian, Arab, and Western cultures. The contentment, tolerance, happiness, self-denial, and sacrifice are generally the characteristics of people at large. The religious injunctions and rituals followed by a vast majority form a national character. For example fasting in the month of Ramadan, practiced by an overwhelming majority, results in staunch self-control and self-denial. A blend of the above stated influences with the patriarchal and male dominated nature of our society make a Pakistani woman to gain protection, comfort, and happiness living within their families without the need of adopting many western style autonomy indicators. Thus there is a need to focus on the advocacy of only those autonomy variables which lead to happiness in a woman's life. After all, the end goal should be to understand and promote the well-being and happiness of Pakistani women who form a vital part of society. Appendix Table 1 Correlation Matrix Going Abuses out Education Faced Alone Education 1 Abuses Faced 0.121 1 Going out Alone 0.021 0.015 1 Decision-making 0.001 0.03 0.21 Job 0.0587 -0.011 0.113 Sickness 0.042 0.008 -0.084 Sanitation 0.329 0.083 -0.057 Area of Residence 0.335 0.026 -0.012 Poverty 0.265 0.06 -0.049 Assets 0.216 0.037 0.024 Age 0.226 -0.002 -0.198 Decision- making Job Sickness Education Abuses Faced Going out Alone Decision-making 1 Job 0.06 1 Sickness 0.035 0.009 1 Sanitation 0.017 -0.153 -0.001 Area of Residence 0.038 -0.11 -0.032 Poverty 0.005 0.089 -0.041 Assets 0.142 0.076 -0.000 Age -0.133 -0.021 0.168 Correlation Matrix Sanitation Area of Assets Age Residence Poverty Education Abuses Faced Going out Alone Decision-making Job Sickness Sanitation 1 Area of Residence 0.489 1 Poverty 0.243 0.241 1 Assets 0.185 0.161 0.161 1 Age 0.008 -0.019 0.013 0.064 1

REFERENCES

Ali, Syed Mubashir, and Mehboob Sultan (2003) Age and Sex Distribution of 1998 Census: An Evaluation. In A. R. Kemal, Mohammad Irfan, and Naushin Mahmood (eds.) Population of Pakistan: An Analyses of 1998 Population and Housing Census. Pakistan Institute of Development Economics/UNFPA, Islamabad.

Ali, Syed Mubashir, H. B. Siyal, and Mehboob Sultan (1995) Women's Empowerment and Reproductive Choices. The Pakistan Development Review 34:4, 1137-1150.

Cain, et al. (1979) Class Patriarchy and Women's Work in Bangladesh. Population and Development Review 5, 405-438.

Caldwell, John C. (1986) Routes to Low Mortality in Poor Countries. Population and Development Review 12, 171-220.

Ghuman, Sharon J. (2003) Women's Autonomy and Child Survival: A Comparison of Muslims and Non-Muslims in Four Asian Countries. Demography 40:3, 419-436.

Hakim Abdul, et al. (1998) Pakistan Fertility and Family Planning Survey 1996-1997. National Institute of Population Studies and London School of Hygiene and Tropical Medicine.

Hoffmann, John P. (2004) Generalised Linear Models: An Applied Approach. Brigham Young University.

Jain, Raviandra K. (1979) South Indians on the Plantation Frontier ill Malaya. New Haven, Connecticut, Yale University Press.

Jejeebhoy, Shireen J. (2002) Convergence and Divergence in Spouses' Perspectives on Women's Autonomy in Rural India. Studies ill Family Planning 33:4, 299-308.

Mahmood, M Arshad (2003) Estimates of Mortality. In A. R. Kemal, Mohammad Irfan, and Naushin Mahmood (eds.) Population of Pakistan: An Analyses of 1998 Population and Housing Census. Pakistan Institute of Development Economics/UNFPA, Islamabad.

Mahmood, Naushin, and Syed Mubashir Ali. (2002) The Disease Pattern and Utilisation of Health Care Services in Pakistan. The Pakistan Development Review 41:4 Part II, 745-757.

MHDC (2004) Human Development in South Asia: The Health Challenge. Islamabad, The Mahboob ul Haq Human Development Centre. Table-6, 255.

Nasir Shabnam (2002) "Poor Pakistani Women" the daily The News. Islamabad, September 20.

Nasir, Zafar Moeen, and F. M. K. Kiani (2004) The Effect of Economic Changes on Fertility in Pakistan. 5th Annual Research Conference Proceedings, Population Association of Pakistan, 14-16 December, Karachi.

Pakistan, Government of (2000) Pakistan Integrated Household Survey Round 3, 1998-1999. Federal Bureau of Statistics.

Pakistan, Government of (2001) Census Report Pakistan, 1998. Statistics Division, Pakistan Census Organisation.

Qureshi, S. K., and G. M. Arif (1999) Profile of Poverty in Pakistan 1998-99. PIDE, Islamabad. (MIMAP Technical Paper Series No. 5).

Richard, Emberger Jr. (2004) Behind the Smile. Newsweek Magazine Weekly, Newsweek Inc. Times Printers Ltd. New York.

Sathar, Zeba A., and Shehnaz Kazi (1997) Women's Autonomy, Livelihood and Fertility--A Study of Rural Punjab. Islamabad: Pakistan Institute of Development Economics.

Shah, Nasra. M. (1986) Female Employment: Trends, Structure, Utilisation and Constraints. In Nasra M. Shah (ed.) Pakistani Women: A Socio-economic and Demographic Profile. Pakistan Institute of Development Economics, Islamabad/East-West Population Institute, Honolulu, Hawaii.

Siddiquie, et al. (2006) Gender and Empowerment Evidence from Pakistan. Pakistan Institute of Development Economics, Islamabad.

(1) However, there are some female headed households where women need to take major decisions, or there are situations where the view of the older women (mother-in-law, grandmother, mother/aunt) are likely to play a crucial role in decision-making within the family--such as sending girls to school or college, girl's marriage, etc. At the same time, men usually have more say in decisions relation to sale/purchase of property, livestock or other financial matters.

Syed Mubashir All and Rizwan ul Haq are Senior Research Demographer, and Staff Demographer, respectively, at the Pakistan Institute of Development Economics, Islamabad. Table 1 Percent Distribution of Women by Level of Happiness and Socio-economic, Cultural, and Autonomy Status Levels of Happiness Not Quite Very Dependent Variables Happy Happy Happy N 934 1823 488 All Women 28.8 56.2 15.0 Going Out No 26.7 58.1 15.2 Yes 30.1 55.0 14.9 Decision No 30.1 61.3 8.6 Yes 28.6 55.2 16.2 Education No Education 33.1 55.9 11.0 Less than 6 22.6 56.3 21.1 6 to 10 16.4 58.2 25.4 11+ 11.9 54.2 33.9 Current Employment No 27.2 56.3 16.5 Yes 34.2 55.9 9.9 Assets No 35.8 54.2 10.0 Yes but can't Use 17.8 64.1 18.1 Yes and can Use 19.7 56.7 23.7 Abuses Faced Abuses Paced Inside the House 33.7 56.8 9.5 No Abuses Faced 26.7 55.9 17.4 Sickness Yes 38.1 47.8 14.1 No 26.9 58.0 15.1 Sanitation Unsafe 34.3 54.8 11.0 Safe 24.4 57.3 18.2 Place of Residence Rural 30.8 56.8 12.4 Urban 25.3 55.2 19.5 Poverty Poor 34.7 56.1 9.1 Non Poor 25.1 56.2 18.7 Levels of Happiness Total Chi-square Dependent Variables Test N 3245 All Women 100.0 0.000 Going Out No 100.0 0.102 Yes 100.0 Decision No 100.0 0.000 Yes 100.0 Education No Education 100.0 0.000 Less than 6 100.0 6 to 10 100.0 11+ 100.0 Current Employment No 100.0 0.001 Yes 100.0 Assets No 100.0 0.000 Yes but can't Use 100.0 Yes and can Use 100.0 Abuses Faced Abuses Paced Inside the House 100.0 0.000 No Abuses Faced 100.0 Sickness Yes 100.0 0.000 No 100.0 Sanitation Unsafe 100.0 0.000 Safe 100.0 Place of Residence Rural 100.0 0.000 Urban 100.0 Poverty Poor 100.0 0.000 Non Poor 100.0 Source: Original data file of Pakistan Socio-economic Survey 2001. Model-1 Table 2 Logistic Regression Coefficients and Odds Ratios of Predictor Variables on Women's Happiness Equation 1 Not Happy Relative to Quite Happy Predictor Variable Coefficients Odds Ratios Constant -0.448 Job Status No Job (a) Doing Job 0.126 1.134 Education No Education (a) 1-5 -0.805 0.447 6-10 0.642 *** 0.526 11+ -0.298 ** 0.743 Going out Behaviour Can't Go Alone (a) Can Go Alone 0.098 1.102 Decision-making No (a) Yes 0.038 1.039 Assets No Assets (a) Yes but can't Use -0.557 *** 0.573 Yes and can Use 0.79 *** 0.454 -2 Likelihood 596.8 Chi Square 293.15 N 3244 Equation 2 Very Happy Relative to Quite Happy Predictor Variable Coefficients Odds Ratios Constant -2.307 Job Status No Job (a) Doing Job -0.473 *** 0.623 Education No Education (a) 1-5 0.943 *** 2.576 6-10 0.647 *** 1.91 11+ 0.554 *** 1.741 Going out Behaviour Can't Go Alone (a) Can Go Alone -0.036 0.965 Decision-making No (a) Yes 0.666 *** 1.947 Assets No Assets (a) Yes but can't Use 0.575 *** 1.777 Yes and can Use 0.306 * 1.358 -2 Likelihood Chi Square N Source: Original data file of PSES 2001. (a) Reference category. * Significant at .1 level. ** Significant at .05 level. *** Significant at .01 level. Model-2 Table 3 Logistic Regression Coefficients and Odd Ratios of Predictor Variables on Women's Happiness after controlling for Socio-Economic Variables Equation 1 Not Happy Relative to Quite Happy Predictor Variable Coefficients Odds Ratios Constant -0.515 Job Status No Job (a) Doing Job 0.121 1.128 Education No Education (a) 1-5 -0.474 * 0.622 6-10 -0.362 ** .696 11+ -0.096 0.908 Going out Behaviour Can't Go Alone (a) Can Go Alone -0.027 0.974 Decision-making No (a) Yes -0.052 0.95 Assets No Assets (a) Yes but can't Use -0.53 *** 0.589 Yes and can Use -0.68 *** 0.507 Age of the Women Poverty Status Below or on Poverty Line (a) 0.0204 *** 1.021 Above Poverty Line -0.177 ** 0.837 Region of Residence Rural (a) Urban -0.139 1.149 Abuses Faced Yes (a) No -0.156 * 0.855 Sickness Yes (a) No 0.436 *** 0.647 Sanitation Bad (a) Good -0.24 ** 0.786 2 Likelihood 5233.7 Chi-square 426.696 N 3244 Equation 2 Very Happy Relative to Quite Happy Predictor Variable Coefficients Odds Ratios Constant 2.092 Job Status No Job (a) Doing Job -0.454 *** 0.635 Education No Education (a) 1-5 0.523 ** 1.687 6-10 0.343 ** 1.403 11+ 0.338 ** 1.410 Going out Behaviour Can't Go Alone (a) Can Go Alone 0.045 1.046 Decision-making No (a) Yes 0.753 *** 2.213 Assets No Assets (a) Yes but can't Use 0.508 *** 1.662 Yes and can Use 0.221 1.248 Age of the Women Poverty Status Below or on Poverty Line (a) -0.0224 *** 0.978 Above Poverty Line 0.468 *** 1.597 Region of Residence Rural (a) Urban 0.106 1.112 Abuses Faced Yes (a) No 0.505 1.657 Sickness Yes (a) No 0.210 0.811 Sanitation Bad (a) Good 0.013 0.923 2 Likelihood Chi-square N Source: Original data file of PSES 2001. (a) Reference category. * Significant at .1 level. ** Significant at .05 level. *** Significant at .01 level.