Comparing the status of never-marriage after age 40 of males and female in urban and rural areas in 2009. The results show that, delayed marriage is more frequent in urban areas than in rural areas in all four age groups of both sexes. The proportion never-married among males in the age group 40–49 in urban areas is about three times higher than in rural areas (5% versus 1.7%), and the proportion never-married among females in urban areas is about 1.7 times higher than in rural areas (7.9% versus 4.6%). This corresponds to the general pattern that delayed marriage or never marriage is becoming more common in regions with higher levels of economic development and industrialization. Figure 7 presents the maps for the proportion never married for both sexes among people age 40 and older in all provinces in Viet Nam in 2009. It is likely that the pattern of “delayed marriage becoming more common in areas with higher levels of economic development and industrialization” is more consistent with the situation in provinces from Da Nang and further south. The proportion never married among the population aged 40 and older is highest in more industrialized provinces such as Da Nang, Ho Chi Minh City, and Binh Duong. In the North, the situation is different when the highest proportions delaying marriage are not found in Hanoi or Hai Phong, but in Ha Giang (for males) and Thai Binh, Ha Nam, Nam Dinh, and Ninh Binh (for females). Thus, it can be concluded that delayed marriage in Viet Nam is not only related to the level of industrialization but also depends on other socio-cultural factors.
Table 20 presents the regression model of probability of never-marriage among the population aged 40–69 in Viet Nam in 2009. The age group 70+ is excluded from the model because of the small proportions never married and strong influence of mortality. The dependent variable is delayed marriage status of individuals (never-married=1, ever-married=0). The independent variables are the same as earlier but include also four types of disability status: vision, hearing, walking, and memory. Other technical specifications and interpretations are similar to the previous regression model. In general, the results indicate that, in comparison to the previous model on early marriage, the independents variables in this regression model explain smaller part of the variation of the dependent variable, especially in female group. The reason is that delayed marriage may be more strongly related to other omitted variables. Below can be found the analysis of the relationship between each independent variable and the probability of never-marriage among the population aged 40–69. The results show that among males and females aged 40–69, the probability of never-marriage is higher than the Northern Midlands and Mountains for all five remaining regions, especially for the Southeast and Mekong River Delta. Only the regression coefficient for females in the Central Highlands is negative. Thus, holding constant the other independent variables in the model, females in the Central Highlands and males in the Northern Midlands and Mountains are most likely to be married by age 40 in comparison with other regions. On the other hand, for both males and females, the probability of delayed marriage in the Southeast was significantly higher than in other regions.
Table 20: Viet Nam (2009) - Logistic regression of probability of never marriage among population aged 40–69 Male Female
Northern Midlands and Mountains - -
Red River Delta 0.714 0.718
North and South Central Coast 0.519 0.563
Central Highlands 0.446 - 0.185
Southeast 1.545 1.080
Mekong River Delta 0.944 0.688
Urban (Rural =0) 0.917 0.438
40-44 - -
45-49 - 0.556 - 0.045
50-54 - 1.216 - 0.136
55-59 - 1.794 - 0.309
60-64 - 2.446 - 0.709
65-69 - 3.044 - 1.495
In-migrant 0.216 0.147
Ethnic minority 0.141 - 0.077
Religious adherent 0.220 0.321
Below Primary - -
Below Lower Secondary - 1.314 - 0.608
Below Upper Secondary - 1.518 - 0.712
Upper Secondary - 1.443 - 0.651
Post-Secondary - 1.573 - 0.564
Working - 1.238 - 0.038
Vision Disability 0.399 0.722
Hearing Disability 0.087 0.166
Walking Disability 0.084 0.363
Memory Disability 1.676 1.150
Constant - 2.279 - 3.081
Source: Viet Nam (2011): Table A.20
Regarding urban and rural areas, the regression model one again confirms the results analysed earlier. Population aged 40–69 in urban area is more likely to be never-married than in rural areas, and the difference is clearer for males than for females. Third, regarding age, the probability of being never-married decreases quickly as age increases, especially for males. That means as age increases, the proportion delaying marriage decreases because many individuals get married after they turn 40 (not because old people can get married more easily than the young). Only for females, the difference between the age group 40–44 and 45–49 is not statistically significant. Forth, regarding migration status, the results show that for males and females aged 40–69, in-migrants have a higher probability of delaying marriage than non-migrants. Combined with the results in the regression model in Table20, it can be concluded that migration is relevant to both early and delayed marriage of females in contemporary Viet Nam.
Fifth, regarding ethnicity, it is interesting that, the probability of delayed marriage among ethnic minority males is higher than among Kinh males. In contrast, the probability of delaying marriage among ethnic minority females is lower than for Kinh females, holding other variables constant. The difference is small but it is statistically significant. One of the possible reasons is that the sex ratio among the young and middle-aged people in the ethnic minorities is lower (more balance) than in the Kinh population.
Sixth, concerning religion, the probability of being unmarried among both males and females who are religious adherents is higher than in the non-religious groups. This seems reasonable as some people do not marry because they are religious adherents, while some people become religious because they are unable to get married.
Regarding educational attainment, people with higher educational levels are less likely to delay marriage compared to those with less than primary education, and the difference is stronger for males than for females. Thus, low educational levels may be the direct or indirect cause of delaying marriage for people aged 40–69, especially for males. However, the regression coefficients do not vary much between the level at “less than lower secondary school” and the higher levels, especially for females. This shows that the probability of delaying marriage among people aged 40–69 is not significantly related to educational achievement, except for the group “less than primary school” that are more likely to be married at the age 40-69. If high educational attainments leads to later marriage, it must be very high educational achievement such as post-university, not the educational levels considered in the regression.
The results on working status show that there is a significant difference between males and females. The probability of delaying marriage among working males is significantly lower than for nonworking males. However, the working status of females aged 40–69 is not significantly related to their probability of being unmarried. This result corresponds with the general view that working males can more easily get married than unemployed males and vice versa, married males are more responsible than unmarried males so they find jobs in order to be the breadwinners for their families.
And last but not least are the results on disability status. As predicted, people with disabilities have a higher probability of delaying marriage than people without disabilities. The highest probability of delaying marriage is for people with memory disability (difficulty with memory and concentration), followed by people with vision disability (difficulty in seeing even with glasses). Males with walking disabilities (difficulty in moving around) and hearing disabilities (hard of hearing) are more likely to delay marriage than people without disabilities, but the differences are small. Compared to the female model, the male model reports a higher coefficient for memory disability, lower coefficients for vision and walking disabilities, and a similar coefficient for hearing disability.
In short, delayed marriage (defined as being unmarried among the age group age 40 to 69) is most correlated to low educational attainment, disability (especially memory and vision disability), religious adherence, in-migration status, and residence in the Southeast and the Mekong River Delta.”
252. By using appropriate multivariate regression techniques, one may underpin the marital status-education-work relationship for women. The basic question in this relationship is whether the marital status of a woman has a direct effect on her labour force participation, after controling for other intervening factors. The following logistic regression, based on the Aruba 2010 Population and Housing Census, was used to study the relationship. The dependent variable in the analysis was a dichotomy: whether the woman worked or not. The analysis was restricted to women aged 25-50 years, because below age 25 many women are still in school and above age 50 many women on Aruba withdraw from the labour market and most mothers have grown up children. The predictors used in the analysis were: age, number of children ever born (CEB), household income excluding that of the women in question, country of birth, educational attainment, marital status and a variable which indicated if the woman was living together with a partner or not. Country of birth was included in the analysis as many foreign women come to work in the Aruban hotel sector. Next to marital status, the variable indicating if the woman was living together with a partner on a durable basis was added, because on Aruba, consensual unions are very common. The variable on household income (excluding that of the women in question) is included to control for the economic necessity of the female respondent to work or not.
253. The results of the analysis are presented in Table 21. As the analysis is based on census data – and not a survey - the standard errors and significance levels of the regression coefficients are irrelevant and left out of the table. Among the categorical variables, the following reference categories were used: Aruba (country of birth), less than primary/no education (educational attainment), never married (marital status) and living together. The values in the exp(B) column show the odds ratios. These ratios are computed by raising e to the power of the regression coefficient.
254. The logistic regression shows some interesting results in terms of the position of women and their labour force participation. First, the odds ratio for CEB (0.935) shows that on Aruba the odds of being at work for a woman is 6.5 per cent lower for each additional child she gave birth to. Second, participation in the labour force varies quite significantly by country of birth. The highest participation is among women from the Dominican Republic and the lowest among women from the USA. Note that no coefficient is entered for Aruba, as this is the residual category against which all the others are measured. Also, the higher a woman’s educational attainment, the higher her chances of being at work. Note the very low value of women with a PhD, which group consists only of a few women. Living together with a partner has some effect on the chances of having a job, but not substantially (1.08). However, marital status plays a much more important role to determine the work status of a woman. The odds of being at work for married women on Aruba is only 0.653 that of never-married women, after controlling for all other predictors. Divorced women, on the other hand, have higher odds (1.158), while widowed women score lower (0.704).
Table 21: Aruba (2010) - Logistic regression of the probability of working for women aged 25-50, by selected explanatory variables Explanatory variable Category B exp(B)
Constant -0.114 0.892
Age 0.021 1.021
Number of children ever born -0.067 0.935
Total income of other household members 0.000 1.000
Country of birth Aruba
Colombia -0.294 0.746
USA -0.899 0.407
Dominican Republic 0.196 1.217
Venezuela -0.804 0.447
Curaçao -0.100 0.905
Netherlands -0.273 0.761
Other -0.101 0.904
Educational attainment None/Less than primary
Primary 0.490 1.633
Lower vocational 0.765 2.150
High school (4 year cycle) 1.089 2.971
High school (5 year cycle) 0.978 2.659
High school (6 year cycle) 0.697 2.008
Intermediate vocational 1.401 4.059
Higher (Bachelor) 1.656 5.236
Higher (Master's) 1.474 4.368
Higher (PhD) -0.142 0.867
Marital status Never married
Married -0.426 0.653
Divorced/legally separated 0.147 1.158
Widowed -0.352 0.704
Living together Yes
No 0.077 1.080
Source: Population and Housing Census Aruba 2010
7. Interpretation, Policy and Advocacy 255. When interpreting data on marital status for the purpose of gender analysis, it is important to remember that “being married” may not mean the same thing to women and men in terms of lived experiences. Particularly in countries where laws governing married status differ by religious denomination, “being married” diverges even in its legal meaning. For instance, the level of difficulty involved in passing on religious denomination and nationality or securing custody for their children differs for Muslim, Christian and Druze women in Lebanon.
Text Box 11: Marriage and Divorce from a Gender Perspective Gender advocates have struggled for decades to make divorce an option for women. While one reason for this is the possibility to escape an abusive relationship, another is that the mere possibility of divorce provides women with leverage to gain a more equal status within marriage (Yodanis, 2005).
One example in support of this view (i.e. the possibility of divorce leads to better marriages) is Indonesia: Here, divorce rates have been declining not as a consequence of conservative gender ideologies, but due to increased free choice in marriage, educational expansion, delayed marriage, urbanization, increasing employment before marriage, and legislative change (Heaton et al, 2001).
This example also illustrates the importance of contextual information in analysing the data. Thus, a decline in divorce rates can be interpreted in different ways. Additional research and qualitative studies are often useful to correctly interpret the findings.
256. At the time of writing, divorce is legal in all countries globally, except for Filipino non-Muslims. However, in many Muslim-majority countries, obtaining a divorce is significantly more difficult for women than for men.
257. Polygamy is a contentious issue in many societies, with all countries influenced by Islamic law, except Tunisia, permitting polygamy. However, some countries restrict polygamy by requiring court permission (Syria, Morocco, Iraq), or, in the case of Pakistan, the permission of an arbitration council. Also, Jordan has enacted legislation permitting a wife at the time of marriage to include a stipulation in her contract that gives her the right to divorce her husband if he marries another woman (Mashhour, 2005). Polygamy, typically polygyny by nature and practice, confers power, status and privilege to a man over and above that of a woman. Hence, the CEDAW Committee in its General Recommendation 21 notes, “Polygamous marriage contravenes a woman's right to equality with men, and can have such serious emotional and financial consequences for her and her dependents that such marriages ought to be discouraged and prohibited.”20
258. A high prevalence of girls married under the age of 18, when their male peers remain single, is an indicator for gender inequality in that country. Of note, many countries set legal marriageable ages that differ from the internationally agreed benchmark of age 18 for both women and men. Moreover, governments set different marriageable ages for females and males (e.g. Senegal, 20 for men and 16 for women; state of Ohio in the US, 18 for men and 16 for women; Bangladesh, 21 for men and 18 for women) and some countries such as Kenya, Jordan and Paraguay set the minimum age for marriage below 18 years for both sexes.
259. Child marriage takes place almost exclusively within the context of poverty and gender inequality and has important social, cultural and economic dimensions. While impoverished rural parents may believe that child marriage will protect their daughters, it in fact results in lost development opportunities and limited life options. Often, child brides are pulled out of school, depriving them of an education and meaningful work, and increasing their dependency on their husbands (Manda and Meyer, 2005). Early widowhood is also associated with child marriage as many girls are married to older men, and men’s life-expectancy is lower than that of women in most countries.
260. As is always the case, ultimately the reduction of child marriage can be brought about by providing better alternatives to women. For example, the Government of Malawi has decided to enhance educational spending on girls nationwide in order to curb the negative social and economic consequences of child marriage. In parallel, Malawi has introduced targeted programmes in some regions to boost women’s employment and support family planning services (Manda and Meyer, 2005). For countries with important ethnic cleavages, the targeting of child marriage prevention efforts might be refined to focus on girls from communities most at risk of marrying their girls as children, be they majority or minority groups.
261. Uganda recently unveiled its long awaited proposed revisions to the Marriage and Divorce Bill. Seems the updates have been about 4 decades in the making. The bill, which gives women the right to divorce an impotent husband, also establishes equitable distribution of property between spouses upon divorce, providing co-habiting couples with the same rights to property as married people. For the first time in Uganda, it also establishes marital rape as a crime. The wide ranging piece of legislation outlaws the practice of widow inheritance and makes it an offense to demand the return of bride price upon dissolution of a marriage. The most important thing is that the laws in Uganda finally recognize the “non-monetary” contributions of aggrieved women to a broken marriage, i.e. childrearing, and finally gives them means of redress in a divorce. It prohibits widow inheritance, in conformity with Article 32(2) of the Ugandan Constitution stating that laws, cultures, customs and traditions which are against the dignity, welfare and interest of women or which undermine their status, are prohibited by the Constitution. It reforms and consolidates the law relating to civil, Christian, Hindu, Bahai and customary marriages in terms of marital rights and duties, separation and divorce legislation. Data from the Ugandan Bureau of Statistics (UBOS) was used to show how the principles of equality and non-discrimination are violated with the current state of affairs: Women from certain backgrounds (less educated, certain ethnic groups) are affected by polygamy more than others.
262. Showing the effect of child marriage on girls’ education, economic status and other indicators of women’s wellbeing can help highlight the loss for a national economy and the consequences on public health. All of the tabulations and multivariate analyses described above can be reproduced at local geographical level, which allows identifying areas such as rural areas or regions in the country that should be targeted by specific measures or campaigns.