Jump to ContentJump to Main Navigation
Towards Gender Equity in Development$

Siwan Anderson, Lori Beaman, and Jean-Philippe Platteau

Print publication date: 2018

Print ISBN-13: 9780198829591

Published to Oxford Scholarship Online: November 2018

DOI: 10.1093/oso/9780198829591.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 17 September 2019

Bride Price and the Well-Being of Women

Bride Price and the Well-Being of Women

Chapter:
(p.117) 6 Bride Price and the Well-Being of Women
Source:
Towards Gender Equity in Development
Author(s):

Sara Lowes

Nathan Nunn

Publisher:
Oxford University Press
DOI:10.1093/oso/9780198829591.003.0006

Abstract and Keywords

Bride price, a payment from the groom to the bride’s family at the time of marriage, is a common cultural practice in many African societies. We examine the relationship between the bride price amount and a range of outcomes using a sample of 317 couples from the Democratic Republic of the Congo. Motivated by common concerns associated with high bride price, we examine whether payment of a higher bride price is associated with earlier marriage and higher fertility; a greater acceptance of domestic violence; decreased ability of the wife to leave her husband; lower-quality marriages; and lower levels of happiness for the wife. We find evidence that women for whom a high bride price was paid are less accepting of domestic violence and are happier.

Keywords:   bride price, marriage customs, happiness, violence, fertility

1. Introduction

The practice of paying a bride price—which is a payment from the groom or groom’s family to the bride’s family—at the time of marriage is a custom that is widespread throughout sub-Saharan Africa. Among the African societies represented in the Ethnographic Atlas, 83 per cent report having bride price practices. Historically and today, the magnitude of the bride price is often significant. It is not uncommon for bride price transfers to be in excess of a year’s income and sometimes as large as seven or eight times annual income (Anderson 2007).

In recent years, this practice has come under criticism, particularly in Africa. A number of objections have been raised in both the media and in political discourse. Recent examples of articles from African newspapers criticizing the practice include Kelly (2006), IRIN News (2006), and Eryenyu (2014). The objections stem from the view that the practice is transactional in nature and, therefore, results in the commodification of women, which has adverse consequences. Husbands may feel they that because they have paid for their wives, they can mistreat them, leaving women in marriages prone to physical violence and conflict. The Ugandan women’s rights group Mifumi has reported cases where men say ‘I am beating my cows’ when they hit their wives, where women are denied ownership of property, and where women may be expected to be sexually available to their husbands at any time and without protection (Eryenyu 2014). In response to these potentially negative (p.118) effects of high bride price, Kenya’s most recent round of marriage laws legislates that a token bride price must be counted as sufficient to meet the needs of the custom (Dudley 2014). The Zambian government has recently spoken out to discourage families from requesting exorbitant amounts for their daughters (Tembo 2014).

In many customs, the woman’s parents are required to return the bride price if the woman leaves the marriage, particularly if she has not yet had any children. Thus, it is possible that the practice of bride price results in women being locked in the marriage because parents are unwilling or unable to repay the bride price. Due to this concern, Ugandan courts have outlawed the requirement for the bride price to be paid back upon divorce (Government of Uganda 2001; Mwesigwa 2015). The stated rationale for this legal change was that it ‘would make it easier for women to leave abusive relationships’ (Biryabarema 2015).

Another concern centres around the incentives that the bride price generates for parents. It has been argued that parents may have an incentive to ‘sell’ their daughters early to obtain the bride price payment, resulting in early marriage and higher rates of lifetime fertility. For example, Hague et al. (2011) report accounts from Uganda of parents taking children out of school so they can be married early in return for a bride price. In the words of one focus group participant, the ‘selling [of] a human being because the family wants wealth, [and] selling your daughter at a tender age’ are common. This is because ‘people prefer to get wealth at the expense of their daughters’ education’ (p. 556). Consistent with this concern, Corno and Voena (2016) and Corno et al. (2016) find that adverse shocks to family income increase a woman’s chance of early marriage among societies that practise bride price. Families appear to use early marriage, and with it the receipt of bride price, to smooth consumption. To combat early marriage due to bride price, the local government in Laikipia county, Kenya, has instituted a programme to give cows to parents whose daughters graduate from high school.

These views are not universal, however. While the view of the bride price as a purchase price of a wife is common in the (Western) media, this is very different from the general interpretation of the practice made by anthropologists. For example, Vroklage (1952) explicitly rejects the idea that a bride price is the price paid for the purchase of a woman. Interviewees told him, ‘a bride is not a buffalo’ and ‘a bride is not an animal’. Vroklage (1952: 135) instead describes it as ‘a compensation for the expense, the care and trouble spent on the bride’s upbringing…It is compensation for the complete loss of a worker as a bride withdraws from her own kindred and henceforth belongs to her husband’s.’ Bride price is particularly common among groups that practise patrilineal descent, and is considered as a compensation payment for the (p.119) bride’s future children, who will no longer belong to her parents’ family. In fact, in many groups, marriage is equated with the payment of bride price. In their work on the Kikuyu in Kenya, Adams and Mburugu (1994: 162) write that bridewealth (another term for bride price) is the primary indicator of marriage, with one respondent saying: ‘There was no ceremony, but traditionally I am married because I paid the bridewealth.’ With regard to the Sebei of Uganda, anthropologist Goldschmidt (1974: 312) notes that without the transfer of bride price there is no marriage and any children will not belong to the father’s lineage.

In this chapter, we contribute to a better understanding of the effects of bride price by studying the relationship between the bride price amount and a range of outcomes. Motivated by the most common concerns that are associated with high bride price, we examine whether a higher bride price paid at marriage is associated with earlier marriage and higher fertility; a greater acceptance of violence within the home; decreased ability of the wife to leave her husband; lower-quality marriages; and lower levels of happiness for the wife. Our analysis also examines the closely related question of whether the custom of having to pay back the bride price upon divorce causes wives to be trapped in unhappy and low-quality marriages.

We contribute to answering these questions with survey data collected in Kananga, a provincial capital in the Democratic Republic of the Congo (DRC). We collect information from 317 marriages, for a total of 634 individuals. In this setting, the practice of the payment of bride price is widespread. Thus, our focus is on the value of the bride price payment and how this is related to different characteristics of the marriage. This can be contrasted to other studies that focus on the presence or absence of a bride price tradition (e.g. Bishai and Grossbard 2010 and Ashraf et al. 2016).

According to our estimates, there appears to be no evidence that a larger bride price payment is associated with earlier marriage or with higher fertility. We also find that larger bride price payments are associated with better-quality marriages as measured by beliefs about the acceptability of domestic violence, the frequency of engaging in positive activities as a couple, and the self-reported happiness of the wife. We also examine the correlates of the requirement for the bride price to be paid back upon divorce. Contrary to general concerns about this aspect of the custom, we find no evidence that this requirement is associated with women being less happy in their marriages. In fact, we find a positive association, although the coefficient is statistically insignificant. However, we do find that if the value of the bride price paid is very high (over US$1,000), then the requirement is, in fact, negatively associated with the happiness of the wife. Thus, the combination of a very high bride price and a requirement to pay back the bride price upon divorce is associated with wives being less happy.

(p.120) Overall, our estimates do not provide overwhelming evidence in support of the concern that bride price has detrimental effects on the well-being of married women. In fact, the practice generally appears to be correlated with good outcomes. The one exception is that the combination of a very high bride price and the requirement for the bride price to be paid back upon divorce is negatively associated with the wife’s happiness.

Although these estimates are conditional correlations and not causal estimates, and so should be treated with the necessary amount of caution, we do feel that they are informative. At the very least, these findings, combined with the dearth of other estimates of the correlates of bride price, suggest that much more research into the effects (or correlates) of bride price is needed, especially given the calls to amend or abolish the practice in many countries within Africa.

The rest of this chapter is structured as follows. The next section provides a description of the Congolese setting. Section 3 explains the data collection procedure, the sample, and our bride price measure. Section 4 reports our regression equations and our estimated relationships of interest, and Section 5 discusses the significance of these results and compares them to related evidence from other African samples. Section 6 concludes.

2. The Congolese Setting

In the DRC, bride price is referred to as la dot (the French word for dowry, though the payment is made from the groom’s family to the bride’s family) or biuma in Tshiluba, the language spoken in the south-central part of the DRC, where we collected our data. Although, historically, there was variation in marriage payment customs, today bride price is practised among all ethnic groups in this part of the DRC. Bride price also functions as legal proof of marriage, and a couple are not considered married until a bride price is paid in full. Therefore, bride price is also important for inheritance and determining the lineage of any children of the marriage since, if a husband dies, it allows a wife to prove that they were officially married. Chondoka (1988: 158) writes that traditionally ‘marriages were all legalized on delivery of the “main” payments’.

The modern practice of bride price has its roots in pre-colonial customs. At that time, as in many other parts of Africa, the practice was widespread and common. The Ethnographic Atlas has data on sixty-four pre-colonial ethnic groups that are located within the DRC. Of these, approximately 90 per cent practised the payment of bride price and none had dowries. The remaining groups that did not practise payment of bride price tended to have token bride price (small symbolic payments) or bride service (where the husband performs (p.121) work for the bride’s family). Using more detailed data from Vansina (1966) on 300 ethnic groups within the DRC, we have coded historical bride price practices at a more disaggregated ethnicity level. These practices ranged from no payment of bride price, bride price payments of varying economic value, bride service, or payment of dowry. The spatial distribution of these practices is shown in Figure 6.1. The picture that emerges from the finer data from Vansina (1966) is broadly similar to that from the Ethnographic Atlas, although with a bit more variation and nuance. Of those groups represented in Vansina (1966), approximately 80 per cent practised some form of payment to the bride’s family at the time of marriage. However, there has existed great heterogeneity in the size of these transfers, who is involved in the payments, and the terms under which bride price must be repaid.

Bride Price and the Well-Being of Women

Figure 6.1. Spatial variation in historical bride price practices

Source: Authors.

Interviews and focus groups conducted by the authors with men and women in the DRC suggest that bride price is still very important for marriages. For both men and women, bride price payment signals honour and respect for the wife. As one Congolese woman explained, ‘Bride price is important for all African women, but for Congolese women in particular…The bride price is an official custom that expresses the love a husband has for his wife. For the parents of the wife, the bride price symbolizes a reward and an honour.’ Without payment of the bride price, marriages are not recognized. In fact, some women believe it is better to live with a friend than to live with a man that does not want to pay the bride price. When asked about the role of (p.122) bride price, a Congolese man noted: ‘The bride price is how a man honours his wife.’ However, he also says that the bride price serves as ‘a guarantee that prevents the [woman’s family] from taking her back when there is a dispute’. These interviews suggest that although the bride price is customarily associated with a man’s commitment to his wife and is a signal of respect, it may also hinder the woman’s options in the case of marital disputes.

3. Description of the Data and the Sources

3.1. Sampling Procedure and Data Collection

The surveys for the project were administered between June and October 2015 in Kananga, DRC.2 Kananga is an ethnically diverse city of over one million people, and is the capital of Kasai Central province. The most populous ethnic group in the city is the Luluwa; however, there are dozens of other ethnic groups represented in the city. Figure 6.2 is a map of the DRC with the city of (p.123) Kananga denoted by a large dot and the villages of origin of the individuals in the sample in smaller black dots.3

Bride Price and the Well-Being of Women

Figure 6.2. Villages of origin of the sample and ethnic group boundaries

Source: Authors.

A screening survey was administered across the city to identify individuals in monogamous marriages. We selected 442 individuals by this means and asked them (and their spouses) to participate in the study. Ultimately, 320 of these 442 individuals agreed to participate in the study, yielding a sample of 640 individuals. Three men reported either not having paid bride price or not knowing the value of the bride price they paid. These couples were excluded from the analysis. The final sample consists of 634 individuals from twenty-eight different ethnic groups. Summary statistics for the sample are available in Lowes and Nunn (2017).

Couples were visited at their homes three different times by a team of one male enumerator and one female enumerator. The male enumerator met with the husband, while the female enumerator met with the wife. Thus, the husband and wife both undertook the surveys in private and away from their spouse. In the first visit, participants completed a long survey. This survey had questions on demographics, economic activities, land ownership, family history, and a child roster. During the second and third visits, individuals completed shorter surveys that asked questions about their views on gender norms and on characteristics about their marriage. The surveys were conducted in either French or Tshiluba, which are the languages spoken in this area of the DRC.

3.2. Bride Price and an Exploration of its Determinants

We first describe our primary variable of interest. In our surveys, both men and women were asked: ‘At the time of your own marriage, what was the total value of the bride-price that was paid? Please include the cost of all of the goods and cash payments given as a part of the bride price.’ The bride price can be paid in many forms, and usually involves some combination of goats, money, food, and other household items. Thus, our question explicitly asks the respondents to include the estimated value of all non-monetary items as well.

Although both the husband and wife were asked separately, we expect the men to have better information on the amount of the bride price paid because they were the person who made the payment and who was directly involved in the transfer of funds (sometimes with the help of their families). In our data, around 80 per cent of men report having contributed to the bride price payment. Additionally, 40 per cent of men report receiving help from their (p.124) fathers and 30 per cent report receiving help from their mothers. By contrast, because the payments are received by the wife’s parents, she is not directly involved in the transfer of money and therefore is less likely to have detailed knowledge of the amount.4

Since the bride price is often paid in a variety of goods, it is difficult to assess an exact dollar value of the bride price. Therefore, we presented respondents with categories of the total value of all money and items paid as part of the bride price, and asked respondents to choose one of the categories. These categories, and the proportion of respondents with a bride price payment within each category, are reported in Figure 6.3. The modal bride price payment was between US$250 and US$500, a significant sum given that for a number of decades the average real per capita gross domestic product (GDP) of the DRC has been stagnant at around US$250 per person. Larger payments in excess of US$500 (two years of per capita GDP) are also common.

Bride Price and the Well-Being of Women

Figure 6.3. Distribution of bride price payments

Source: Authors.

We begin by examining the correlates of bride price payments. The estimates are reported in Table 6.1. In Column 1, we examine the wife’s education, age, and age squared as correlates. In Column 2, we add year of marriage, (p.125) an indicator variable for whether the couple were married in Kananga (rather than in the village), and the interaction of these two variables (which allows the effect of year of marriage to be different inside and outside Kananga). In Column 3, we include the husband’s education, age, and age squared.

Table 6.1. Wife’s characteristics and bride price

Dep. var.: bride price amount

(1)

(2)

(3)

Wife’s age

0.011

−0.016

−0.021

(0.025)

(0.037)

(0.042)

Wife’s age squared

−0.000112

0.000074

0.000168

(0.000269)

(0.000337)

(0.000388)

Wife’s years education

0.091***

0.096***

0.090***

(0.026)

(0.027)

(0.027)

Year married

−0.011

−0.025

(0.013)

(0.016)

Married in Kananga

−4.604

−3.637

(25.625)

(25.792)

Year married*married in Kananga

0.002

0.002

(0.013)

(0.013)

Husband’s age

0.014

(0.046)

Husband’s age squared

−0.00032

(0.00048)

Husband’s years education

0.008

(0.026)

Observations

317

317

317

Mean dep. var.

5.722

5.722

5.722

Notes: Robust standard errors in parentheses. Years education is the number of years of education. Year married is the year the couple were married as reported by the husband. Married in Kananga is an indicator variable equal to 1 if the couple lived in Kananga at the time of marriage. Bride price amount is a variable from 1 to 9 corresponding with various bride price values, where (1) is equal to US$0–25 and (9) is equal to US$1,500. *p<0.1; **p<0.05; ***p<0.01.

Source: Authors.

Interestingly, we find that the wife’s education is the only robust predictor of the bride price payment at the time of marriage. This finding is consistent with previous studies that have found that a wife’s education is a strong predictor of the value of the bride price in Kenya (Borgerhoff Mulder 1995), Zambia, and Indonesia (Ashraf et al. 2016). In our sample, the mean value of the wife’s education is 8.48 and the standard deviation is 3.65. Thus, a one-standard-deviation increase in the wife’s education is associated with a 3.65×0.09 =0.33 point increase in the bride price index. This is fairly sizeable given that the standard deviation of the index is 1.44 (the mean is 5.72).

Our estimates show that the amount of the bride price, conditional on education, is similar inside and outside the provincial capital of Kananga. Also, we find no evidence that the value of bride price payments has been systematically increasing over time. One concern in interpreting the coefficient on ‘year married’ is that it may be collinear with the wife’s age and the (p.126) husband’s age. However, we obtain similar estimates regardless of whether these covariates are included in the estimating equation (estimates not shown here, but they are available upon request).

4. Estimating Equations and Empirical Results

4.1. Estimating Equations

For the analysis that follows, we estimate the following specification:

Yi,h,w=α+βBridePriceAmounti,h,w+XiΩ+XhΠ+XwΓ+εi,h,w
(1)

where i indexes marriages that comprise a husband h and a wife w. Yi,h,w denotes our outcome of interest, which, depending on the specific measure, will measure a characteristic of the marriage, for the husband or for the wife. Bride Price Amounti,h,w is the amount of the bride price payment received by woman w’s parents from husband h at the time of the marriage. Individuals recall the total value of the bride price and select the appropriate category among the options available (see Figures 6.3 and 6.4). We convert the categories into a scale that ranges from 1 to 9, increasing with the value of the bride price payment. Xh and Xw denote vectors of the following characteristics of the husband and wife, respectively: age, age squared, and educational attainment (measured in years). Xi denotes a vector of marriage characteristics, including the year of the marriage, an indicator that equals 1 if the marriage was in Kananga (the provincial capital where the surveys were conducted), and the interaction of the two.

4.2. Bride Price, Age at Marriage, and Fertility

We begin by estimating the relationship between the bride price payment and the wife’s age at marriage. A potential concern with the tradition of bride price is that it induces parents to arrange for their daughters to be married at a younger age so that they can obtain the bride price payment. This might occur, for example, if the parents are credit constrained, and therefore need funds immediately, even if it comes at the cost of their daughter’s future welfare. Therefore, we test for the presence of a relationship between the value of the bride price payment and the age at marriage by estimating equation (1) with the wife’s age at the time of marriage (in years) as the dependent variable. The average age of marriage for women is nineteen years. By contrast, the average age of marriage for men is twenty-seven years.

The results are reported in Table 6.2. In this table, and all subsequent tables, the coefficients are not reported, but these tables are available in the related working paper (Lowes and Nunn 2017). Column 1 reports the estimates of (p.127) equation (1) with a more parsimonious set of covariates that only includes the wife-level covariates, Xw. Column 2 adds the marriage-level measures, Xi, and Column 3 adds the husband’s characteristics, Xh. In all specifications, we estimate a negative but statistically insignificant relationship between the bride price payment at marriage and the age at marriage. In addition to the coefficients being insignificant, their magnitudes are also small. For example, according to the estimate from Column 3, a one-standard-deviation increase in the bride price index (equal to 1.44) is associated with a decline in the age at marriage by 0.10×1.44=0.14 years. We also find that as we add more covariates to our estimating equation, the magnitude of the point estimate declines noticeably. This suggests that even the small estimated effects are potentially spurious. The point estimates appear to be converging towards zero with the inclusion of additional controls.

Table 6.2. Bride price, wife’s age at marriage, and fertility

Dep. var.: wife’s age at marriage

Dep. var.: number of children

(1)

(2)

(3)

(4)

(5)

(6)

Bride price amount

−0.289

−0.177

−0.102

0.125*

0.095

0.065

(0.213)

(0.155)

(0.141)

(0.075)

(0.067)

(0.061)

Wife-level controls, Xw

Y

Y

Y

Y

Y

Y

Marriage-level controls, Xi

Y

Y

Y

Y

Husband-level controls, Xh

Y

Y

Observations

317

317

317

317

317

317

Mean dep. var.

19.37

19.37

19.37

4.208

4.208

4.208

Notes: Robust standard errors in parentheses. Wife’s age at marriage is the age of the wife when the couple were married. Number of children is the number of living children a woman has. *p<0.1; **p<0.05; ***p<0.01.

Source: Authors.

We next turn to fertility, which is an outcome that may be related to the age at marriage. If bride price leads to earlier marriage, then this may, in turn, cause higher fertility rates. Thus, we also examine the number of living children a woman has.5 We estimate equation (1), with the number of children as the outcome of interest. The estimates are reported in Columns 4–6 of Table 6.2. The three columns report the same three specifications as in Columns 1–3. We estimate a positive relationship between bride price payments and fertility, although only the estimate in Column 1 is significant, and this only at the 10 per cent level. In addition, we find that as we add additional covariates the magnitude of the estimated effects converge towards zero. Lastly, the estimated magnitudes are also small. According to the point (p.128) estimates from Column 3, a one-standard-deviation increase in the bride price index is associated with the following increase in the number of children: 0.065×1.44=0.09.

Overall, we do not find evidence of an association between the value of the bride price and either a woman’s age at marriage or the number of children that she has.

4.3. Bride Price and Attitudes About Domestic Violence

We now turn to an examination of whether bride price payments are associated with different attitudes about domestic violence on the part of husbands. In particular, a common concern is that the payment of a high bride price causes men to feel that because they have ‘paid’ dearly for their wife, they therefore have the right to mistreat her. We test for this possibility by estimating versions of equation (1) where the dependent variables are measures of men’s self-reported views about the acceptability of domestic violence.

We measure the husband’s attitudes about domestic violence using survey questions that we take from the DHS modules on domestic violence attitudes. Husbands are asked to report whether they believe that domestic violence is justified in five different situations. The situations are: (1) if the wife goes out without the husband’s permission; (2) if she neglects the children; (3) if she argues with her husband; (4) if she refuses to have sex; (5) if she burns the food. For each scenario, the husbands chose an integer on a 1–5 scale, where 1 indicates that he ‘Strongly Disagrees’ (with domestic violence being justified in the situation) and 5 indicates that he ‘Strongly Agrees’ (with domestic violence being justified). We take the average of the answers in the five scenarios to obtain an index that ranges from 1 to 5 and that is increasing in the extent to which the husband believes that domestic violence is justified.

Estimates of equation (1), with the index of the husband’s perceived acceptability of domestic violence as the outcome, are reported in Panel A of Table 6.3 in Columns 4–6. We estimate a positive effect in all three specifications. However, the estimates are generally not significant and are small in magnitude. According to the estimates of Column 6, a one-standard-deviation increase in the value of the bride price amount, results in an increase of 0.094×1.44=0.14 in the domestic violence index. This is equal to 6 per cent of the mean of the domestic violence acceptability index and 0.11 standard deviations. Although these effects are not completely trivial, they are fairly modest in addition to being imprecisely estimated.

Table 6.3. Bride price and domestic violence, marriage quality, and happiness

Panel A: Bride price and views about domestic violence

Dep. var.: positive views of domestic violence index

Women

Men

(1)

(2)

(3)

(4)

(5)

(6)

Bride price amount

−0.153***

−0.163***

−0.172***

0.064

0.067

0.094*

(0.035)

(0.036)

(0.036)

(0.050)

(0.050)

(0.050)

Wife-level controls, Xw

Y

Y

Y

Y

Y

Y

Marriage-level controls, Xi

Y

Y

Y

Y

Husband-level controls, Xh

Y

Y

Observations

317

317

317

317

317

317

Mean dep. var.

2.470

2.470

2.470

2.322

2.322

2.322

Panel B: Bride price and marriage quality

Dep. var.: frequency of activities index

Women

Men

(1)

(2)

(3)

(4)

(5)

(6)

Bride price amount

0.076**

0.076**

0.075**

0.119***

0.119***

0.114***

(0.031)

(0.031)

(0.032)

(0.027)

(0.027)

(0.026)

Wife-level controls, Xw

Y

Y

Y

Y

Y

Y

Marriage-level controls, Xi

Y

Y

Y

Y

Husband-level controls, Xh

Y

Y

Observations

315

315

315

315

315

315

Mean dep. var.

4.999

4.999

4.999

5.285

5.285

5.285

Panel C: Bride price and self-reported happiness

Dep. var.: happiness

Women

Men

(1)

(2)

(3)

(4)

(5)

(6)

Bride price amount

0.188***

0.193***

0.192***

0.025

0.026

−0.005

(0.037)

(0.038)

(0.038)

(0.033)

(0.033)

(0.035)

Wife-level controls, Xw

Y

Y

Y

Y

Y

Y

Marriage-level controls, Xi

Y

Y

Y

Y

Husband-level controls, Xh

Y

Y

Observations

317

317

317

317

317

317

Mean dep. var.

2.681

2.681

2.681

2.700

2.700

2.700

Notes: Robust standard errors in parentheses. Positive views of domestic violence index takes the average response to the following questions: Domestic violence is justified if wife (1) goes out without husband’s permission; (2) neglects children; (3) argues with husband; (4) refuses sex; (5) burns food; all questions answered with 1 = strongly disagree to 5 = strongly agree. The index sums the responses to each of the questions and divides by 5 to generate the average response. Frequency of activities index takes the average response to the following questions: How frequently do you and your spouse (1) laugh together; (2) work on a project together; (3) receive a gift; (4) walk together; (5) talk about your day; (6) discuss your feelings; all questions answered with 1 = never to 6 = more than once a day. The index sums the responses to each of the questions and divides by 6 to generate the average response. Happiness asks respondents to rate how happy they are on a scale of 1 = very unhappy to 5 = very happy. Bride price amount is a variable from 1 to 9 corresponding with various bride price values, where (1) is equal to US$0–25 and (9) is equal to US$1,500. *p < 0.1; **p < 0.05; ***p < 0.01.

Source: Authors.

We next examine the potential effects of bride price amounts on women’s attitudes about domestic violence. The concern with high bride price payments is that wives may also believe that husbands can legitimately treat them badly if they have paid a high bride price. Thus, in Panel A of Table 6.3 in (p.129) (p.130) Columns 1–3, we present the estimates with the wife’s view on the acceptability of domestic violence in the household as the outcomes variable. We find evidence that a higher bride price paid at the time of marriage is associated with the wife being less accepting of domestic violence. All three coefficients of interest are negative and highly significant. According to the estimates of Column 3, a one-standard-deviation increase in the bride price payment index is associated with a decrease in the women’s views of the acceptability of domestic violence that is equal to 0.172×1.44 = 0.248. This is a large effect that is equal to about 10 per cent of the mean of the domestic violence index and 0.43 standard deviations.

Overall, we find weak evidence that bride price amounts are associated with a greater reported acceptance of domestic violence for husbands, although these estimates are generally not significant and are small in magnitude. In contrast, the estimates show a significant and robust negative relationship between bride price amounts and the acceptability of violence for wives.

4.4. Bride Price and Marriage Quality

Another potential consequence of bride price is that it may affect the quality of the marriage. Therefore, we examine whether a higher bride price is correlated with marriage quality as measured by the frequency with which couples have positive interactions with each other. Husbands and wives were asked a series of questions about how often they did each of the following six activities with their spouse: laugh, work together on projects, receive gifts, walk together, talk about the day, and discuss their feelings. For each activity, the respondents chose from the following six response options: (1) never; (2) less than once a month; (3) once or twice a month; (4) once or twice a week; (5) once a day; and (6) more than once a day. We assign the reported numeric values to each response, and create an index equal to the average response value across each of the six activities. The final index ranges from 1 to 6 and increases with the frequency with which the couple engage in positive interactions. Most couples report engaging in positive activities fairly regularly, with the mean value being around 5.

Estimates of equation (1) with this outcome as the dependent variable are reported in Panel B of Table 6.3.6 Columns 1–3 report estimates based on the woman’s perception of the frequency of positive activities and Columns 4–6 report estimates based on the husband’s perception. Across all specifications, we find that a higher bride price value is associated with a greater frequency of (p.131) positive interactions. According to the estimates from Column 6, a one-standard-deviation increase in the bride price index is associated with an increase of 0.114×1.44=0.164, which is equal to approximately 0.164/0.816=0.20 standard deviations of the dependent variable. In addition, all estimates are highly significant, and the magnitude of the point estimates remain stable with the inclusion of additional covariates. Thus, the results suggest that a higher bride price is not associated with a lower-quality marriage, and in fact seems to be associated with better marriages, at least according to this particular measure of marriage quality.

4.5. Bride Price and Self-Reported Happiness

We now examine a related outcome, and one that directly measures women’s well-being: self-reported happiness. We measure happiness using the following survey question: ‘Taking all things together, would you say you are very unhappy, somewhat unhappy, neither happy nor unhappy, somewhat happy, or very happy? You can also look at these pictures to help you with your response.’ The respondents were presented with emoticons of frowning and smiling faces to visualize the scale. Based on the respondents’ answers, we create a variable that takes integer values between 1 and 5 and is increasing with their self-reported happiness.

We then estimate equation (1) with the wife’s happiness as the outcome variable. Estimates are reported in Panel C of Table 6.3 in Columns 1–3. We find a robust positive relationship between the amount of bride price paid and the wife’s self-reported happiness. The estimates are highly significant and stable across our three baseline specifications. The point estimate of 0.192 (from Column 3) suggests that a one-standard-deviation increase in the bride price index is associated with an increase of 0.192×1.44=0.28 points on the happiness scale. This is a sizeable effect given that the mean level of the happiness in the sample is 2.7 and that the standard deviation is 0.88.

The natural next question is what effect a higher bride price has on the husband’s happiness. It is possible that a higher bride price is also associated with greater happiness on the part of the husband. However, it is also possible that the wife’s happiness comes at the expense of the husband’s happiness and so a higher bride price is associated with a less happy husband. We examine this by estimating equation (1) with the husband’s happiness measure as the outcome of interest. Estimates are reported in Panel C of Table 6.3 in Columns 4–6. The estimated coefficients are very close to zero.

Overall, the estimates are consistent with a higher bride price resulting in greater happiness on the part of the wife, but having no effect on the happiness of the husband. These estimates are (perhaps) surprising given the (p.132) general concern that a high bride price leads men to treat women poorly, leaving women in unhappy marriages where they are worse-off. Our estimates provide no evidence to support the validity of this concern. If anything, our findings appear to suggest that a higher bride price is associated with a better marriage, where the wife is happier.

4.6. Repayment Practices and Happiness

Another concern with the bride price is that it is often the case that if the marriage dissolves then the bride price must be paid back, in whole or in part, by the wife’s family. The extent to which this convention is present, and is enforced, can vary depending on which party is perceived to be at fault, and whether or not the woman has had any children. The requirement of paying back bride price is believed to be a significant obstacle to women who would like to leave bad marriages, causing them to be stuck in these marriages.

In an attempt to provide evidence on this consequence of the bride price, we examine whether the presence of a requirement to pay back the bride price upon divorce is associated with less self-reported happiness by wives. If the practice of repayment causes women to be trapped in low-quality and unhappy marriages, then we would expect that married women who perceive there to be a requirement of repayment of the bride price upon divorce, on average, will be less happy.

To do this, we use information obtained from the following question: ‘In the event of a divorce, how much of the bride price must be repaid?’. Respondents choose one of the following responses: all of it; most of it; some of it; a little bit of it; none of it; or don’t know. Using their response, we then code a variable that equals 0 if none of it must be returned, 1 if some amount of it must be returned, and 2 if all of it must be returned. We add to our baseline estimating equation this measure, which is increasing in the extent to which the bride price must be repaid upon divorce.7

The estimates are reported in Columns 1–3 of Table 6.4. As shown, we find no evidence that the requirement to pay back the bride price upon divorce is associated with wives being less happy in their marriage. The coefficient on the repayment variable is not statistically different from zero. In addition, the coefficients are not negative, but positive. Though not significant, these estimates suggest that, if anything, a repayment requirement is positively associated with the wife’s happiness.

(p.133) It is possible that the requirement to repay the bride price upon divorce is more detrimental when the value of the bride price paid is higher. We test for this possibility by including an interaction between the repayment variable and the value of the bride price measure. This allows the effect of the repayment requirement on the wife’s happiness to vary depending on the value of the bride price that was paid at marriage.

The estimates are reported in Columns 4–6 of Table 6.4. Our estimates produce a coefficient for the interaction term that is negative and significant. According to the estimated magnitudes, the effect of the repayment index, which takes on the value of 0, 1, or 2 is given by: 0.926 – 0.146 × bride price amount. Thus, the estimate relationship between the repayment index and the wife’s happiness moves from being positive to negative when the value of the bride price index is greater than 0.926/0.146 = 6.34. Therefore, when the value of the bride price index is 7, 8, or 9, then a requirement to repay the bride price is associated with lower happiness of the wife. The relationship is not significant for a bride price value of 7 (p = 0.30), but is significant for a value of 8 (p = 0.05) and a value of 9 (p = 0.02).

Table 6.4. Bride price repayment customs and self-reported happiness

Dep. Var.: wife’s happiness

(1)

(2)

(3)

(4)

(5)

(6)

Bride price amount

0.186***

0.191***

0.190***

0.265***

0.276***

0.272***

(0.038)

(0.038)

(0.038)

(0.041)

(0.041)

(0.041)

Amount pay back

0.069

0.078

0.065

0.926***

0.992***

0.933***

(0.076)

(0.077)

(0.077)

(0.251)

(0.248)

(0.252)

Bride price amount*amount pay back

−0.146***

−0.156***

−0.148***

(0.043)

(0.043)

(0.043)

Wife-level controls, Xw

Y

Y

Y

Y

Y

Y

Marriage-level controls, Xi

Y

Y

Y

Y

Husband-level controls, Xh

Y

Y

Observations

317

317

317

317

317

317

Mean

2.681

2.681

2.681

2.681

2.681

2.681

Notes: Robust standard errors in parentheses. Amount pay back is a variable from 0 to 2, where (0) means none of the bride price has to be repaid upon divorce, (1) means some of the bride price must be repaid, and (2) means all of the bride price must be repaid. Wife’s Happiness asks respondents to rate how happy they are on a scale of 1 = very unhappy to 5 = very happy. *p<0.1; **p<0.05; ***p<0.01.

Source: Authors.

Overall, there is no evidence that the requirement to repay the bride price upon divorce is associated with wives being stuck in less happy marriages. However, this average effect does mask heterogeneity. When the value of the bride price that was paid at marriage is very high (above US$1,000), then there is a negative and significant relationship between the repayment requirement and the happiness of wives.

(p.134) 5. Taking Stock and Comparing Our Findings to Other Studies

To date, there is limited evidence about the correlates of bride price in the African context. One of the most robust findings, although about the causes of bride price and not its consequences, is that higher education is associated with higher bride price payments at marriage. The reason behind this association is explored in depth by Ashraf et al. (2016), who provide evidence that, within Zambia, the positive effect of the wife’s education on her bride price is causal and widely known. In addition, they also show that this effect of education is taken into account when parents make the decision of whether to send their daughters to school. In our setting, we have also found a strong positive relationship between education and the value of the bride price.

A number of previous studies have examined the consequences of bride price values. In a recent working paper, Mbaye and Wagner (2013) examine over 2,000 respondents from eight regions in Senegal and find a significant negative relationship between higher bride price and fertility. This contrasts with our finding of no robust relationship between the value of the bride price and fertility. Although we estimated a positive relationship, the point estimates were generally not statistically different from zero. Along somewhat similar lines, the estimates from Mbaye and Wagner (2013) are only marginally significant and they are very small in magnitude. According to their estimates, if bride price increases by 100 per cent—a very large increase—fertility falls by only 0.04 children. Thus, taken together, our findings and the findings from Mbaye and Wagner (2013) seem to indicate that the value of the bride price appears to have no sizeable or robust relationship with fertility.

Although not the focus of their analysis, the study by Mbaye and Wagner (2013) also provides estimates of the relationship between bride price and a measure of appreciation of the wife by the husband. They find that a higher bride price is associated with less appreciation, although the coefficients are generally not statistically different from zero. This can be contrast to the spirit of our findings, which show a positive association between the value of the bride price and our different measures of the quality of the marriage.

The previous findings that are most directly comparable to the findings in this chapter are from qualitative studies. Comparing our findings to these studies, we find significant differences. For example, our findings that higher bride price is correlated with less acceptance of domestic violence, higher marriage quality, and greater happiness for women stand in stark contrast to the conclusions from casual observation or qualitative studies. A number of focus group and survey-based studies have shown that men and women tend to believe that the bride price results in less empowerment of women, worse marriages, and lower overall well-being. Results of this nature have been found in Uganda (Hague et al. 2011; Kaye et al. 2013) and Ghana (p.135) (Dery 2015; Horne et al. 2013). For example, in Hague et al. (2011), 84 per cent of 151 respondents reported that they believed that there was a strong connection between the value of the bride price and domestic violence.

There are many explanations for the differences in the findings. First, Africa is not a homogeneous unit. Thus, there could be significant heterogeneity across the large continent, which may result in differences in our relationships of interest. Thus, the effects of bride price may be different in the DRC than in Uganda, Kenya, or Ghana. A second possibility is that the actual effects of the bride price custom may be different from the perceived effects. Individuals observe the practice of the bride price, high levels of domestic violence, and low levels of female empowerment and may draw a link between them. Whether there is a general relationship in the data when looking across a large number of individuals is an empirical question. A third possibility stems from the fact that in qualitative studies participants’ answers must be interpreted, and this is done through the lens of the researcher. Thus, there is concern that the researcher’s prior assumptions affect the mapping that is made from statements in the focus groups to conclusions. Further, the presence of a researcher in the focus groups may affect the statements made by participants, and one worries, in particular, about ‘demand effects’, where participants are more likely to inadvertently (and without conscious realization) say what they feel the researcher wants to hear. This can be contrasted with surveys, which, though less rich in some dimensions, have the advantage that researchers are not present when the questions are answered. Instead, local enumerators ask the questions.

6. Conclusion

Bride price, which is payment from the groom and/or groom’s family to the bride’s family at the time of marriage, is an important cultural practice of many African societies. In recent years, there have been widespread concerns that the practice may have negative effects for women. One concern is that the monetary payment received by the bride’s family at marriage may incentivize early marriage, leading to higher fertility. It is also believed that it may promote the view that husbands have ‘purchased’ their wives, resulting in worse treatment of wives. In many locations, the bride price must be paid back to the groom’s family upon divorce. This may cause an obstacle to divorce and result in women being trapped in unhappy marriages. Thus, in general, there has been widespread concern that the practice is detrimental to the well-being of women.

In this chapter, we have used data related to these issues in an attempt to provide a better understanding of the potential effects of bride price. We did (p.136) this by examining the empirical relationship between bride price payments and various outcomes of interest, using a sample of 317 married couples from Kananga, a city located in the DRC, a setting where almost everyone pays a bride price and marriages are not recognized as legitimate unless a bride price is paid.

We found no evidence that a larger bride price payment is associated with earlier marriage or with higher fertility. We also found that larger bride price payments are associated with better-quality marriages as measured by beliefs about the acceptability of domestic violence, the frequency of engaging in positive activities as a couple, and the self-reported happiness of the wife. We also examined the correlates of the requirement for the bride price to be paid back upon divorce. Contrary to general concerns about this aspect of the custom, we found no evidence that this requirement is associated with women being less happy in their marriages. In fact, we found a positive association, although the coefficient was statistically insignificant. However, we did find that if the value of the bride price paid was very high (over US$1,000), then the requirement is, in fact, negatively associated with the happiness of the wife.

Overall, we found that the evidence does not support the notion that the practice of bride price has detrimental effects on the well-being of married women. Perhaps surprisingly, in general, a higher bride price tends to be associated with good outcomes. The one exception is that the presence of a high bride price and the requirement for the bride price to be paid back upon divorce does appear to be associated with less happiness on the part of the wife.

We end by reminding the reader about an important caveat. Although informative and valuable, the relationships that we estimate cannot be taken as definitive evidence of the causal effect of high bride price and/or repayment requirements on women’s well-being. Despite our attempts to control for potentially omitted factors, it is very possible that they still influence the estimates and impeded our ability to interpret them as causal. However, we do feel that the estimates we report here lead one to pause and recognize the need for greater research to understand the causal effects of the custom, particularly given the calls to abolish the practice in many countries within Africa.

References

Bibliography references:

Adams, B.N., and E. Mburugu (1994). ‘Kikuyu Bridewealth and Polygyny Today’. Journal of Comparative Family Studies, 25(2): 159–66.

Anderson, S. (2007). ‘The Economics of Dowry and Brideprice’. Journal of Economic Perspectives, 21(4): 151–74.

(p.137) Ashraf, N., N. Bau, N. Nunn, and A. Voena (2016). ‘Bride Price and Female Education’. Working Paper. Cambridge, MA: Harvard University.

Biryabarema, E. (2015). ‘Uganda Court Says Divorced Husbands Cannot Demand “Bride Price” Refund’. Reuters Africa, 6 August.

Bishai, D., and S. Grossbard (2010). ‘Far Above Rubies: Bride Price and Extramarital Sexual Relations in Uganda’. Journal of Population Economics, 23(4): 1177–87.

Borgerhoff Mulder, M. (1995). ‘Bridewealth and Its Correlates: Quantifying Changes Over Time’. Current Anthropology, 36(4): 573–603.

Chondoka, Y.A. (1988). Traditional Marriages in Zambia: A Study in Cultural History. Cambridge: Cambridge University Press.

Corno, L., and A. Voena (2016). ‘Selling Daughters: Age of Marriage, Income Shocks and the Bride Price Tradition’. Working Paper. Chicago, IL: University of Chicago.

Corno, L., N. Hildebrandt, and A. Voena (2016). ‘Weather Shocks, Age of Marriage and the Direction of Marriage Payments’. Working Paper. Chicago, IL: University of Chicago.

Dery, I. (2015). ‘Bride-Price and Domestic Violence: Empirical Perspectives from Nandom District in the North Western Region of Ghana’. International Journal of Development Sustainability, 4(3): 258–71.

Dudley, O. (2014). ‘Highlights of the Marriage Act, 2014’. Available at: http://kenyalaw.org/kenyalawblog/highlights-of-the-marriage-act-2014/ (accessed 5 June 2017).

Eryenyu, J. (2014). ‘Payment of Bride Price Turns Women into Commodities’. Daily Monitor. Available at: www.monitor.co.ug/OpEd/Letters/Payment-of-bride-price-turns-women-into-commodities/806314-2447104-n41iwj/index.html (accessed 5 June 2017).

Goldschmidt, W. (1974). ‘The Economics of Brideprice Among the Sebei and in East Africa’. Ethnology, 13(4): 311–31.

Government of Uganda (2001). A Compendium of Laws Relating to Domestic Relations in Uganda. Kampala: LDC Publishers.

Hague, G., R.K. Thiara, and A. Turner (2011). ‘Bride-price and its Links to Domestic Violence and Poverty in Uganda: A Participatory Action Research Study’. Women’s Studies International Forum, 34: 550–61.

Horne, C., F.N.-A. Dodoo, and N.D. Dodoo (2013). ‘The Shadow of Indebtedness: Bridewealth and Norms Constraining Female Reproductive Autonomy’. American Sociological Review, 78(3): 503–20.

IRIN News (2006). ‘Study Links Payment of Bride Price to Abuse of Women’. Available at: www.irinnews.org/report/59032/tanzania-study-links-payment-bride-price-abuse-women (accessed 5 June 2017).

Kaye, D.K., F. Mirembe, A.M. Ekstrom, G.B. Kyomuhendo, and A. Johansson (2013). ‘Implications of Bride Price on Domestic Violence and Reproductive Health in Wakiso District, Uganda’. African Health Sciences, 5(4): 300–3.

Kelly, A. (2006). ‘Why Girls are Economically Challenged’. The Guardian. Available at: www.theguardian.com/katine/2009/aug/17/money-women (accessed 5 June 2017).

Lowes, S. (2017). ‘Matrilineal Kinship and Spousal Cooperation: Evidence from the Matrilineal Belt’. Working Paper. Cambridge, MA: Harvard University.

Lowes, S. and N. Nunn (2017). ‘Bride Price and the Wellbeing of Women’. WIDER Working Paper 2017/131. Helsinki: UNU-WIDER.

(p.138) Mbaye, L.M., and N. Wagner (2013). ‘Bride Price and Fertility Decisions: Evidence from Rural Senegal’. Discussion Paper 7770. Bonn: IZA.

Mwesigwa, A. (2015). ‘Uganda Court Rules Against Refund of “Bride Price” After Divorce’. The Guardian, 17 August.

Tembo, S. (2014). ‘Court Warns Against High Bride Prices’. Times of Zambia, 3 May. Available at: www.times.co.zm/?p=21383 (accessed 5 June 2017).

Vansina, J. (1966). Introduction a l’Ethnographic du Congo. Kinshasa: Centre de Recherche et d’Information Socio-Politiques.

Vroklage, B. (1952). ‘Bride Price or Dower’. Anthropos, 47(1/2): 133–46.

Notes:

(1) Cammie Curtin provided excellent research assistance. Sara Lowes acknowledges generous financial support from the National Science Foundation Graduate Research Fellowship Program, Harvard IQSS, and the Harvard Center for African Studies.

(2) The data were collected for a project on matrilineal kinship systems and intra-household bargaining (Lowes 2017).

(3) Village of origin is where an individual’s family originates and is not necessarily the same as village of birth.

(4) When wives are able to estimate the amount of bride price paid, the amounts of bride price paid reported by husbands and wives are highly correlated. However, women are more likely than men to report that they do not know the value of the bride price paid.

(5) This does not take into account stillbirths or children that have died subsequently. Almost 25% of the women in the sample have had at least one child die during its first year of life.

(6) The sample size falls from 317 to 315 due to non-response.

(7) An alternative strategy is to create indicator variables for the categories of the index. All conclusions we report here remain if we do this.