Tradition and Change in Marriage Payments in Vietnam, 1963–2000

Abstract

Trends and determinants of marriage payments have rarely been examined at the population level despite their plausible implications for the welfare of family and the distribution of wealth across families and generations. In this study, we analyze population-based data from the Vietnam Study of Family Change to document prevalence and directions of marriage payments in Vietnam from 1963 to 2000. We investigate the extent to which structural and policy transformations (particularly market reform and the socialist policy that banned brideprice) influenced the practice of marriage payments as well as estimate how societal changes indirectly impacted payments via their effects on population characteristics. Results indicate that marriage payments surged following market reform but also reveal more nuanced trends and regional differences during earlier years. While the socialist attempts to eradicate brideprice appear to have been moderately successful in the North prior to economic renovation the evidence suggests they were largely unsuccessful in the South. Results suggest that structural and policy change explained most of the observed variations in marriage payments and that changing characteristics of the individuals who married mattered relatively little. We interpret the reemergence of marriage payments as attesting to resilience of traditional values and the unraveling of the socialist agenda, especially in the North, but also as a reflection of economic prosperity associated with market reform.

INTRODUCTION

Marriages are closely tied to the distribution of property within and between families. Trends in brideprice and dowry can be complex and multifaceted, so much so that prevalence, direction, magnitude, and property rights of marriage payments tend to vary considerably over time and across societies. While marriage payments tend to decline and eventually disappear in modern societies, it is not uncommon to observe practice of payments waning only to re-emerge later; they may even shift in direction in terms of whether the groom’s or bride’s side benefits or whether the newlyweds or parental generations are the primary beneficiaries (Anderson, 2007; Tambiah, 1973). Until recently, existing research on marriage payments relied primarily on anthropological and historical accounts and has rarely been based on systematic population-based data. Further, knowledge about contemporary marriage payments has drawn heavily of studies of dowry inflation in South Asia, even though the practice of payments is common in other contexts (Brown, 2009; Zhang and Chan, 1999).

This paper examines prevalence, directions, and determinants of marriage payments in Vietnam from 1963 to 2000 based on survey data from the Vietnam Study of Family Change. As opposed to results from fieldwork in one village or from convenience sample, our study describes population-based (i.e., regionally representative) trends, a contribution that is almost non-existent in prior research due to a lack of appropriate data.¹ Further, Vietnam provides an especially interesting setting that can advance our understanding about mechanisms underlying cohort changes in the payments. During the 20^th century, the country underwent major social upheavals including continuous wars, socialist collectivization, severe economic stagnation, and rapid economic growth following market liberalization. Yet, northern and southern Vietnam differed significantly in their cultural orientation and political, social, and economic trajectories (Rambo, 1973).

Arguably, the North has historically been more oriented than the South towards the East Asian model of patriarchal kinship and family structure, whereas the bilateral kinship system common in Southeast Asia is more common among the southern populations. Recent separation under different governments (1954–1997) further underlay the regional differences. Beginning in the 1950s, the North was exposed to socialist policies which promoted gender equality in education and workforce and prohibited Confucian-based family practices, including brideprice transfers. The government viewed the institution of family as key to the reproduction of social stratification and sought to devalue traditional customs as an essential step towards the elimination of private ownership and inequality. In contrast to the North, the South was ruled by a government heavily subsidized and influenced by the US within the context of “cold war” politics whose philosophy and policies were diametrically opposite to those of socialism. Only after Vietnam’s reunification in 1975 did the South experience the socialist policies. Past research indicates that these regional differences had widespread implications for marriage and family behaviors in recent decades (Goodkind, 1996; Knodel et al., 2000; Teerawichitchainan et al., 2010).

The present study builds on and extends current theoretical debates on marital transfers. First, we describe the overall trends in marriage payments in Vietnam. Second, we explore determinants of the trends by assessing the direct roles of structural transformation and state policies as well as by addressing their indirect influences through individuals’ socioeconomic and marriage characteristics. By grounding our interpretations in the specifics of the Vietnamese setting, we hope to advance understanding of marriage payment systems and their functions and contribute to the development of a cross-cultural theory on interfamilial and intergenerational wealth distribution at the time of marriage.

BACKGROUND

How and why prevalence, directions, magnitude, and property rights of marriage payments vary over time and across societies have been addressed by researchers from diverse disciplines ranging from history and anthropology to sociology and economics. Unlike other marriage behaviors such as marital timing and mate selection, theoretical perspectives regarding changes in payments are less unified and remain largely contentious. In this section we review recent developments in the literature on marriage payments and address how Vietnam can extend current theoretical perspectives on trends and determinants.

Brideprice was found mainly in primitive societies where polygyny was uncommon and women were active in agricultural production (Boserup, 1970). It generally functioned as a repayment from the groom’s side to the bride’s for her labor and reproductive capacity. Who held the property right of brideprice was not uniform. In the traditional Islamic marriage regime, brideprice served as a woman’s insurance against marital dissolution so the key beneficiary was the bride herself. In sub-Saharan Africa, however, brideprice was a collective property to be used for establishing marriages for the bride’s male siblings. Anthropologists argue that when societies became more stratified, the practice of brideprice typically waned and gave way to dowry (Harrell and Dicky, 1985). Dowry was traditionally considered a pre-mortem inheritance and transferred from the bride’s parents directly to the bride (Botticini and Siow, 2003). When societies became more commercialized, dowry served as a means to ensure an endogamous matching and became a direct transfer to the groom and his family (i.e., groomprice), as evidenced in South Asia (Billig, 1992; Caldwell et al., 1983).

In pre-modern Vietnam, brideprice was considered the most important transaction between two families when a marriage took place. Marriage payments were typically intensively negotiated among parental generations (Malarney, 2002). The actual magnitude of brideprice depended largely on the economic status of the groom’s family and could account for a substantial proportion of his household income (Gourou, 1955; Hickey, 1964). In addition to the traditional betel leaves and areca nuts, brideprice items could include edible goods such as tea, rice, and pork and/or valuable transfers such as money, gold, and land. Brideprice was used primarily to fund wedding feasts and sometimes, if in the form of cash, to buy gold or items for the bride’s dowry (Pham, 1999). While brideprice was compulsory, it was voluntary for the bride’s parents to reciprocate with dowry (Malarney, 2002). Dowry was usually transferred directly to the bride in the forms of gold, jewelry, cloth, or other household items.

At least in the view of the economist Anderson (2007), anthropologists whose work is based largely on anecdotal evidence argue that modernization is the key structural force underlying changes in marriage payments. In European and tribal African societies, urbanization, shifts from customary to civil marriage practices, and transformation from household production to wage economy were associated with declines in dowry and brideprice (Goody, 1973). The modernization perspective, however, is less explicit about the mechanisms that caused change and fails to address why modernization leads to a decline in payments in some societies but an increase in others.

More recently, economists have extended the modernization perspective by providing testable hypotheses and economic explanations for the multi-faceted trends in marriage payments. Their hypotheses focus on human capital and demographic factors, and economic functionality of the exchange. Becker (1991), for instance, argues that when there is a more efficient way for parents to invest in their daughter’s future welfare, such as giving her education, dowry ceases to function and disappear. In contrast Anderson (2007: 169) argues that it is not the acquisition of human capital by daughters but a relative increase in the heterogeneity of women’s earning opportunities that cause disappearance of dowries/groomprices. Further, Rao (1993) hypothesizes that a change in population composition (i.e., excess supply of brides over grooms in the respective marriageable age) is a major cause of dowry inflation and brideprice decline. The perspectives put forth by economists have resulted in rigorous empirical tests; yet, evidence remains inconclusive (Edlund, 2000; Rao, 2000). One common limitation in economic studies of marriage payments is a lack of baseline data that permits an evaluation of change over time (Anderson, 2007).

With the exception of surveys in two provinces in northern and southern Vietnam by Goodkind (1996), studies of marriage payments in Vietnam are dominated by historical and ethnographic research that rely largely on anecdotal evidence thus leaving their generalizability in question. They attribute changes in payments to key institutional transformations, including the socialist revolution. In 1960, the socialist government in northern Vietnam implemented the Marriage and Family Law aiming to eradicate Confucian-based family practices including arranged marriage, polygamy, and conspicuous consumption in wedding celebrations, to name but a few. The law forbade brideprice but did not explicitly ban dowry (Luong, 1993; Malarney, 2002). Village cadres reportedly portrayed a brideprice demand as an attempt to sell daughters to the highest bidder. Research claimed a moderate success of the socialist agenda in the North. The northern Vietnamese community that Goodkind (1996) studied witnessed a decline in marital transfer from over 30 percent among couples married in the 1950s to 10–20 percent among those married a decade later. The transfers were also limited to non-monetary and simple household items (Kleinen, 1999). Unlike the North, southern Vietnam was not exposed to the law until 1975. Limited evidence suggested limited influence of socialist policies on marriage practices in this region (Belanger, 2000; Goodkind, 1996).

Vietnam’s shift from a centrally-planned to a market economy in the late 1980s was also hypothesized to have affected marital exchanges, thanks to the government’s lessened interests in promoting Marxist orthodoxy and reduced control over household production which led to rapid economic growth and improved living standards. Pioneering research speculated the revival of pre-socialist marriage practices, including marriage payments. Village studies conducted in the early 1990s revealed that although families routinely recounted the socialist rhetoric against marriage payments, brideprice transfers were openly practiced in the North (Malarney, 2002) and increasingly involved a variety of gift items, particularly cash and gold (Kleinen, 1999). Evidence has thus far been limited to the periods immediately after the reform. Full impacts of market transformation have not been investigated.

By examining trends in marriage payments in Vietnam, we aim to extend the current literature in three ways. First, we empirically assess prevalence and directions of marriage payments using population-based survey data. Our findings document cohort changes in payments over the last four decades and thus add important evidence to existing studies based on ethnographic research and convenience samples in Vietnam. Second, we address the role of state policies in determining practices of interfamilial exchanges at the time of marriage. Current theoretical debates tend to place an emphasis on individuals’ or household’s rational actions and pay inadequate attention to institutional constraints that may shape their decisions about interfamilial and intergenerational transfers of wealth. Very little has been investigated beyond the context of India and China about the extent to which state interventions (or lack thereof) influence the practice of marriage payments (Billig, 1992; Gates, 1997; Sheel, 1997; Siu, 1993; Whyte, 1993; Yan, 2003). Provided that the two key regions of northern and southern Vietnam were exposed to socialist policies at different timing, the country provides a convenient setting to address this research agenda.

In addition to the influences of policies, our third objective is to examine the determinants of brideprice and dowry and how the importance of these determinants may have changed over time. The modernization perspective posits that structural changes are likely to have direct or indirect impacts on marriage payments. While ethnographers documented recent changes in payments in Vietnam, they have not yet teased out specific factors behind these changes. According to the human capital perspective, the Vietnamese government’s determination to accelerate modernization and industrialization process, first through collectivization and subsequently through market measures, may have affected marriage payments through their effects on intra- and inter-familial relationships. These include the extent to which family invests in son’s and daughter’s education, how children contribute to the household economy, the social position of daughters relative to son’s, and the degree children could exert their independence from parents’ influence (Pettus, 2003; Barbieri and Belanger, 2009).

DATA AND METHODS

Our data come from two regional surveys conducted by Vietnam’s Institute of Sociology that comprise the Vietnam Study of Family Change. The first survey was carried out in 2003 in seven provinces including Hanoi in the northern region of the Red River Delta. Using an identical questionnaire, the second survey was conducted exactly one year later in Ho Chi Minh City and six southern provinces covering substantial parts of the Mekong River Delta and the southeastern region. The survey sites are typical of key areas in northern and southern Vietnam. For convenience, we refer to the Red River Delta sample as the North and the sample of the Ho Chi Minh City and environs as the South. Each regional survey was administered to 1,296 currently married individuals. Three marriage cohorts were targeted for interviews. For the first cohort, the range of marriage dates (1963–1971) ends a few years prior to North Vietnam’s victory in 1975, thus representing northern and southern respondents who experienced their early marital years during the US-Vietnam War and in the southern cases, those who were exposed to the pre-socialist social structure and family system. We are limited in our assessment of pre-socialist marriage behaviors among northerners because the oldest cohort interviewed in the surveys was married in 1963 when the socialist revolution had already taken place. Further, the middle cohort (1977–1985) represents new marriages during the post-reunification period when Vietnam underwent a centrally planned economy and widespread economic recession. The most recent cohort (1992–2000) was selected to capture marital experience during the period when the 1986 market reform became widely effective and the country experienced economic revitalization. We refer to these three cohorts as the wartime, the post-reunification, and the renovation cohorts respectively.

Each regional sample was divided equally into 12 categories among male and female respondents, urban and rural settings, and the three marriage cohorts. Since only one respondent was interviewed in each surveyed household, the husbands and wives interviewed were not spouses of each other. The sample was designed to be self-weighting and representative within these 12 categories². In the following analyses, we include only couples whose current marriages are their first and who had at least one parent on both sides alive at the time of marriage because marriage payments were significantly less common among those who experienced marital dissolution and whose parents were already deceased. The restriction yields a sample size of 2,283 married couples, of which 52 percent are northerners and 48 percent southerners. Thirty-two percent of the sample is marriages contracted between 1963 and 1971, 33 percent between 1977 and 1985, and 35 percent between 1992 and 2000.

Measurement of marriage payments

The surveys asked each respondent if there were any brideprice or dowry transfers at the time of marriage. The Vietnamese language prominently distinguishes these two types of transfers. While recall errors especially among the older cohorts cannot be ruled out, we doubt that underreporting of brideprice transfers would be caused by stigma from the former brideprice ban since it essentially disappeared or by respondents’ tendency to provide normative answers.³ Village studies conducted in the 1990s and 2000s concurred the villagers’ willingness to discuss marital transfers, sometimes even in the presence of local officials (Goodkind, 1996; Teerawichitchainan et al., 2007).

Preliminary diagnostic analysis indicates that female respondents tended to report dowry more frequently than male respondents, while there was no significant gender difference in brideprice reporting. Some male respondents may not be fully aware of dowry overlooking items such as personal clothing since dowry transfers are less ritualized compared to brideprice handover (Kleinen, 1999). Given that by the survey design, essentially equal numbers of men and women were interviewed for each marriage cohort and region, it is unlikely that any biases associated with reporting distort observed cohort or regional differences in marriage payments – the focus of our study. This would only be the case if the extent or the nature of the bias differed systematically by region and cohort.

To measure prevalence and direction of marriage payments, our analyses use dichotomous variables indicating whether each married couple observed any brideprice, any dowry, or both types of payments (i.e., bidirectional transfers) at the time of marriage. Approximately 43 percent of couples reported brideprice, 39 percent dowry, and 28 percent bidirectional payments. To measure other aspects of marriage payments, we include additional variables to identify marriages without any payments, with only brideprice, and with only dowry. About 46 percent of couples observed no payments, 15 percent only brideprice, and 11 percent only dowry.

Measurement of independent variables

Key independent variables are marriage cohort and region. Since a marriage cohort represents an aggregate of married couples who had a common exposure to Vietnam’s historical events and social change, we use marriage cohort to measure the compounding effects of major structural and policy transformation that took place during each historical period, namely the US-Vietnam War, post-reunification, and market reform. Additionally, region is incorporated to account for potentially different cultural orientations between the North and South and, because of different North-South timing of exposure to socialism, to gauge the roles of socialist policies in determining marriage payments. Within a bivariate framework, we describe for each marriage cohort and region how prevalent marriage payments was, to what extent the payments were one-directional or bidirectional, and which direction the payments originated from. Within a multivariate framework, we use binary logistic regressions to examine the net effects of cohort and region and to address other determinants of brideprice, dowry, and bidirectional transfers.

While the modernization perspective attributes macro-structural transformation such as urbanization as a major reason for decline in marriage payments, the human capital hypotheses pinpoint the cause of payment change to changes in individual’s socioeconomic characteristics, age at marriage, manner of mate selection and parent-children relations as a result of the diversification of the labor market and other structural shifts. To test these hypotheses, our analyses incorporate spousal location of childhood residence, bride’s and groom’s educational attainment and marital age, and parental involvement in mate selection as additional independent variables. As evident in Table 1, overall Vietnamese brides and grooms in recent cohorts are more urbanized, better educated, and more likely to marry a spouse of choice at a later age compared to older cohorts. Certain regional and cohort patterns are also noteworthy.

Table 1.

Descriptive statistics, Charateristics of predictor variables by marriage cohort and region.

Covariates	Year of marriage
	1963–1971			1977–1985			1992–2000
	North	South	Sig.	North	South	Sig.	North	South	Sig.
	(N =366)	(N =353)		(N =403)	(N =359)		(N =418)	(N =384)
	Spousal place of growing up
Both rural	80	63	***	71	52	***	64	52	***
One or both urban	20	37		29	48		36	48
Total	100%	100%		100%	100%		100%	100%
Bride's education
Less than 12 years	76	91	***	68	81	***	58	78	***
High education	24	9		32	19		42	22
Total	100%	100%		100%	100%		100%	100%
Mean years of schooling	9.0	5.6	***	10.0	7.1	***	10.0	7.9	***
Groom 's education
Less than 12 years	58	75	***	64	69	n.s.	51	70	***
High education	42	25		36	31		49	30
Total	100%	100%		100%	100%		100%	100%
Mean years of schooling	10.1	7.7	***	10.1	8.4	***	10.3	8.7	***
Mean bride's marital age	21.4	19.9	***	22.1	21.4	**	21.7	22.9	***
Mean groom 's marital age	24.7	23.6	***	25.2	23.6	***	25.9	25.8	n.s.
Manner of mate selection
Chosen by children	87	57	***	96	72	***	97	82	***
Arranged by parents	13	43		4	28		3	18
Total	100%	100%		100%	100%		100%	100%

^***Based on ANOVA, difference between North and South is significant at p≤0.001;

^**Significant at p≤0.01;

^*Significant at p≤0.05;

n.s= Not significant p-value.

Source: Vietnam Study of Family Change, 2003–4

Spousal place of childhood residence indicates whether both husband and wife grew up in rural areas or at least one of them was raised in urban location. For every cohort, there were significantly higher proportions of southerners than northerners with urban childhood residence. Nonetheless, proportions of northern couples with urban upbringing increased steadily among the post-reunification and renovation cohorts, possibly as a result of the socialist government’s attempts to swiftly industrialize the country. While the regional difference appeared to narrow over time, it remained significant among the most recent cohort. Given that modernization perspective posits that the custom of marriage payments tend to decline first among urban populations followed by their rural counterparts, we anticipate that the odds of each type of marriage payment to be lower among married couples with at least one spouse being raised in urban areas.

Further, significant regional differences in educational attainment and marital timing are evident. Northern brides and grooms were generally better educated than their southern counterparts, reflecting earlier and longer exposure to socialist policies aimed at eradicating illiteracy and expanding educational opportunities. Gender gap in schooling was more salient in the South than in the North, especially among the older marriage cohorts. A tangible improvement was nonetheless observed among the recent marriage cohorts of southerners thus reducing the gender gap. Likewise, average age at first marriage was higher among the war and post-reunification cohorts of northerners, compared to their southern counterparts. This possibly echoes, among other factors, northerners’ higher educational attainment and exposure to the socialist family law, which mandated a minimum age of 18 for brides and 20 for grooms (Nguyen, 1998). Consistent with prior studies, regional differences in marital timing among grooms from the renovation cohort, became negligible, whereas average northern brides married at significantly younger age than their southern counterparts. Our analyses incorporate bride’s and groom’s education as dichotomous variables indicating whether they had at least 12 years of schooling and include their marital age as continuous variables.

Education and marital age are common indicators of human capital and socioeconomic status. As societies become more developed and labor markets more diversified, the human capital perspective argues that returns to education would increase and that it would be more rational for parents to invest in their children’s education, which likely causes marital delays, rather than to invest in marriage payments (Becker, 1991; Botticini and Siow, 2003). This implies that high educational attainment and delay in marital timing will reduce the likelihood of marriage payments. Nevertheless, high education, as a proxy for higher socioeconomic status, could increase the odds of marriage payments once the brideprice ban is lifted, since well-off parents are likely better able financially to provide resources to their children at the time of marriage.

Lastly, manner of mate selection is incorporated as a dichotomous variable indicating whether the marriage was arranged by parents (regardless of children’s approval). While there was a continual decline in dominant parental influence in mate selection among successive cohorts, it was consistently more common for southern marriages. This could be due to earlier marital timing among southern couples (especially the first two cohorts); some cultural preferences unique to the South; or the prohibition of arranged marriage in the North. Since marriage payments traditionally went with arranged marriage, we expect parental involvement in mate selection to increase the likelihood of brideprice, dowry, and bidirectional transfers.

A key contribution of our study is to portray population-level trends in marriage payments in Vietnam. There are a few caveats, nevertheless. First, we rely on cross-sectional surveys with retrospective questions about payments opening the possibility for recall errors. Given that for each married couple only one spouse was interviewed, we cannot check consistency between husband’s and wife’s reporting of payments. Further, information on magnitude, providers, and beneficiaries of marriage payments is unavailable in the surveys. Our choice of independent variables is also restricted. While the surveys contain information about each couple’s socioeconomic characteristics, we lack detailed economic characteristics of the respondents and their families such as household income, individual earnings, and type of pre-marital occupation. In addition, some respondents and/or their spouses migrated since they married and lived outside the sample areas covered by the two surveys. However, diagnostic analysis indicates such mismatches are very modest (Teerawichitchainan et al., 2010, footnote 7). Nevertheless, the regional comparisons presented are to some modest extent affected by post-marital migration. Despite these limitations, the survey data documents overall trends and information on the general context in which marriage payments take place – the first step to understand their impact on family welfare and the distribution of wealth within the society.

RESULTS

Trends in marriage payments

Table 2 describes prevalence and direction of marriage payments across successive marriage cohorts in the North and South. Several findings stand out. First, during war years, marriage payments were not common in the North, with less than 30 percent of the cohort observing the custom. In sharp contrast, some forms of payment occurred in nearly 6 out of 10 southern marriages contracted during the corresponding period. Following reunification, the North witnessed a moderate increase in marriage payments, whereas proportions of southern marriages that witnessed any payments declined slightly. More striking was the surge in prevalence of marriage payments during the 1990s in both regions but particularly in the North. This resulted in the North-South convergence, with approximately 70 percent of marriages in the two regions practicing some forms of marital exchange.

Table 2.

Descriptive statistics, Cohort and regional trends in prevalence and direction of marriage payments.

Percentage of marriages	Year of marriage
	1963–1971			1977–1985			1992–2000
	North	South	Sig.	North	South	Sig.	North	South	Sig.
	(N =366)	(N =353)		(N =403)	(N =359)		(N =418)	(N =384)
	% Any m arriage paym ents	29	58	***	39	55	***	70	71	n.s.
% Any brideprice	21	50	***	25	49	***	51	61	**
% Any dowry	20	36	***	29	30	n.s	63	52	**
% Bidirectional payments	11	28	***	16	24	**	44	42	n.s.
% One-directional payments	17	31	***	23	31	**	27	29	n.s.
(% B rideprice only)	(9)	(22)	***	(9)	(25)	***	(8)	(19)	***
(% Dow ry only)	(8)	(9)	n.s.	(14)	(6)	***	(19)	(10)	***

^***Based on ANOVA, difference between North and South is significant at p≤0.001;

^**Significant at p≤0.01;

^*Significant at p≤0.05;

n.s= Not significant p-value.

Source: Vietnam Study of Family Change, 2003–4

When brideprice and dowry are considered separately, it is evident that both increased substantially following market reform but that in both regions dowry rose more than brideprice. Moreover, for all three marriage cohorts, brideprice occurred more frequently in the South than in the North, even though the regional gap narrowed considerably by the time of market reform. In contrast, while dowry was more prevalent during war years in the South compared to the North, the difference reversed among the renovation cohort with dowry becoming more typical in the North than the South. Our analysis shows that marital exchanges were not necessarily reciprocal. During the war and post-reunification periods, one-directional payments accounted for a larger proportion of northern and southern marriages than did bidirectional transfers. It was not until the 1990s that bidirectional payments became the key pattern of marital exchanges in both regions. Evidence further refutes the claim that brideprice was compulsory and that dowry was voluntary in Vietnam. For example, when one-directional payments occurred, it could be either brideprice-only or dowry-only transfers. Although brideprice-only transfers were more common than dowry-only payments for all southern marriages, it was the opposite for northern couples married after reunification and economic reform.

To show further nuances, Figure 1 depicts North-South differences in prevalence of brideprice, dowry, and bidirectional transfers with each of the three cohorts subdivided into two groupings of 5-year and 4-year intervals. The main regional difference is with respect to the prevalence of brideprice (Figure 1a). Among northern marriages, it started already to increase during the early 1980s, whereas in the South brideprice payments continued to decrease moderately during the corresponding period compared to prior to the socialist collectivization. Only after market reform is an increase apparent with the prevalence of brideprice in the South surpassing even the levels observed during pre-socialist years. Also noteworthy is that following reunification the southern levels of dowry and bidirectional transfers dropped more significantly than that of brideprice, as evidenced in Figures 1b and 1c, before rebounding in the late 1980s and 1990s. On the contrary, the North witnessed a consistent rise in the practice of dowry and bidirectional payments across successive marriage cohorts, although only to a modest extent prior to economic reform. Another notable trend is that throughout the 1990s the frequency of all three patterns of marriage payments continued to increase and that by the late 1990s half of all marriages in both regions observed bidirectional transfers. Providing tighter time segments, Figure 1 confirms results presented in Table 2 and our choice of three historical periods to frame the study.

Figure 1.

Cohort and regional trends in brideprice, dowry, and bidirectional transfersa.

***Based on ANOVA, difference between North and South is significant at p≤0.001; **Significant at p≤0.01; *Significant at p≤0.05

^a Due to the exclusion of the bridge years between each marriage cohort (1972–77 and 1986–91), the curves appear more compressed than the actual historical timeline.

Although we lack information on the magnitude of marital transfers, the survey provides information of how brideprice and dowry were expressed. Figure 2 shows the percentages that involved money or gold and house or land. Considerable differences are evident between the North and South as well as between brideprice and dowry. Regional differences are particularly pronounced with respect to brideprice. Across all three cohorts brideprices involving money or gold were substantially less common for northern than southern marriages. In the case of property or land, the newlyweds, rather than the bride’s family, typically hold property rights of such transfers (Personal communication with Sociologist Vu Manh Loi). Dowries consisting of house or land are less common in both regions than are brideprices of this nature. This undoubtedly reflects the predominance of patrilocal living arrangements as the house or property provided is likely associated with and in close proximity to the groom’s parents. Particularly noteworthy in terms of trends is that in both regions and with respect to both brideprice and dowry, providing money/gold became far more common among the renovation cohort.

Figure 2.

Cohort and regional trends in content of marriage payments

Determinants of marriage payments

To move beyond the descriptive picture of trends in marriage payments, we turn to multivariate analyses to explore mechanisms underlying the observed temporal change and regional differences. We use binary logistic regressions to address determinants of three patterns of marriage payments: 1) brideprice, 2) dowry, and 3) bidirectional transfers. In the following analyses, each dependent variable is coded 1 if a particular payment (e.g., any brideprice) took place and 0 if not (e.g., no brideprice). For each analysis, we incorporate two models. The baseline model treats trends in payments (e.g., brideprice) as a function of marriage cohort and region. The saturated model adds other independent variables, including measures of bride’s and groom’s location of childhood residence, educational attainment, marital age, and manner of mate selection. First, we address the extent to which marriage cohort and region explain observed trends in payments, after considering changes in population characteristics. Second, we probe the extent to which variations in payments is a result of differences in couples’ characteristics, net of cohort and regional influences.

Our multivariate analyses, presented in Table 3, consist of two procedures. First, we analyze all marriages, regardless of their regions, in the same regression models. We refer to this as analyses of the combined sample. Coefficients are expressed as the odds ratios of a specific type of marriage payments to non-payments for each category relative to the comparable odds of the reference category for each variable. Odds ratios above 1 indicate that the particular category is associated with higher chances than the reference category that the particular marriage payment took place, whereas values below 1 indicate the opposite. The reference categories are as follows: the 1963–1971 marriage cohort, being northerners, both spouses growing up in rural areas, the bride and the groom having below 12 years of schooling, and the couple making their own decision about mate selection. The bride’s and the groom’s marital age is incorporated as continuous variables. Results from the combined sample are presented in the top panel of Table 3. To consider possible interaction effects between region and each of the control variables, the second procedure features separate regression analyses for each regional sample. While using a similar analytical approach as in the combined sample, only coefficients for marriage cohort are reported in the lower panel of Table 3 to make the discussion of our findings more focused and because analyses based on either combined or regional samples yield roughly similar results with respect to direction and size of additional coefficients.

Table 3.

Binary logistic regression analysis, The determinants of marriage payments in Vietnam, 1963–2000.

Covariates	Brideprice						Dowry						Bidirectional marital transfers
	Baseline			Saturated			Baseline			Saturated			Baseline			Saturated
	Odds ratios		Std. Error	Odds ratios		Std. Error	Odds ratios		Std. Error	Odds ratios		Std. Error	Odds ratios		Std. Error	Odds ratios		Std. Error
COMBINED SAMPLE
Cohort 1977–85	1.09		0.11	1.27		0.12	1.10		0.12	1.12		0.12	1.01		0.13	1.04		0.14
Cohort 1992–00	2.53	***	0.11	3.25	***	0.12	3.51	***	0.11	3.96	***	0.12	3.17	***	0.12	3.48	***	0.13
South	2.46	***	0.09	2.02	***	0.11	1.08		0.09	1.19		0.11	1.48	***	0.10	1.51	***	0.12
Growing up in urban areas				0.86		0.10				1.00		0.10				1.01		0.11
Bride having high education				0.71	**	0.12				1.35	**	0.12				1.06		0.13
Groom having high education				1.12		0.12				1.23	†	0.12				1.21		0.12
Bride's marital age				0.98		0.02				0.96	*	0.02				0.98		0.02
Groom's marital age				0.96	**	0.01				0.98		0.01				0.98		0.02
Arranged by parents				1.47	**	0.13				1.42	**	0.13				1.26	†	0.14
NORTH SAMPLE
Cohort 1977–85	1.26		0.17	1.35	†	0.18	1.69	**	0.17	1.65	**	0.18	1.43	†	0.21	1.46	†	0.22
Cohort 1992–00	4.04	***	0.16	4.22	***	0.17	6.86	***	0.17	6.78	***	0.17	6.01	***	0.19	5.81	***	0.20
SOUTH SAMPLE
Cohort 1977–85	0.99		0.15	1.14		0.16	0.76	†	0.16	0.73	†	0.17	0.82		0.17	0.79		0.18
Cohort 1992–00	1.62	***	0.15	2.15	***	0.17	1.89	***	0.15	2.03	***	0.16	1.90	***	0.16	1.94	***	0.17

Significance level:

^***p≤0.001;

^**p≤0.01;

^*p≤0.05;

^†p≤0.1

Source: Vietnam Study of Family Change, 2003–4

Marriage cohort –our measure of compounding effects of structural and policy transformations—is the most robust predictor of marriage payments. Marriages contracted in the 1990s were several times more likely than the war cohort to observe payments – be it brideprice, dowry, or bidirectional transfers. A comparison between baseline and saturated models indicate that these net cohort effects remained largely unchanged after additional covariates were considered. Not only were the coefficients for marriage cohort unaffected by other independent variables but they also increased modestly, emphasizing importance of cohort in predicting marriage payments. Given other characteristics being equal, marriages contracted after economic reform had 3–4 times greater chance than those married during wartime to experience brideprice, dowry, and bidirectional payments. Further, when each regional sample is analyzed separately, the independent positive effects of the 1990s marriage cohort remained strong. The North observed a more dramatic upward trend in dowry and bidirectional transfers, following economic reform than did the South. Saturated models suggest that northern couples married after reform were about 4–7 times more likely than their wartime counterparts to practice brideprice, dowry, and bidirectional transfers possibly because payments were suppressed to a greater extent during war years. Compared to the North, the South observed strongly positive, yet more modest, effects of economic reform on all three types of payments.

Change in payments following reunification was relatively minimal. According to the analyses of the combined sample, the difference in the likelihood of marriage payments between the wartime and post-reunification cohorts was negligible and not statistically significant. However, when each region was examined separately, certain nuances emerged. In the South, couples observed roughly no change in brideprice and bidirectional transfers between war and post-reunification years. The exception was dowry among southerners which showed a modest but significant decline following reunification. Among northern couples, moderate upward trends in all types of payments were evident soon after reunification. The evidence calls into question the extent of the impact of socialist policies in suppressing marriage exchanges, especially brideprice.

Although the North and South were potentially different in their cultural orientations as well as historical, social, and political trajectories, the effects of region on marriage payments were more modest than those of cohort. Results indicate that the South was more likely than the North to observe brideprice and bidirectional payments and that these net influences changed slightly after the introduction of additional variables. There was no significant difference between the two regions in dowry payments. It was only when analyzing each regional sample separately that regional divergence in dowry patterns emerged. While the North witnessed a steadily upward trend in dowry across successive cohorts, dowry transfers dipped among southerners married after reunification only to rise two-fold during the 1990s.

There were notable changes over the last four decades regarding pre-marital characteristics of brides and grooms and patterns of mate selection (Table 1). Such transformations are hypothesized to lead to a decline in marriage payments; yet, results provide only modest support to these hypotheses. Net of other effects, groom’s marital age significantly lowered the odds of brideprice but not other forms of payments. Likewise, bride’s marital age was inversely associated with only dowry, not brideprice or bidirectional transfers. Further, the independent effects of education are inconsistent. High education among brides significantly decreased the likelihood of brideprice and was positively associated with the odds of dowry. Meanwhile, the likelihood of dowry is greater among highly educated grooms but groom’s education did not explain variations in brideprice and bidirectional transfers. Manner of mate selection demonstrated consistent, independent effects on all types of marriage payments. The odds of payments increased when parents heavily influenced the spousal selection. Nevertheless, its net effects were less robust when compared to those of cohort and region. More importantly, the introduction of these additional variables into multivariate analyses hardly changed the extent marriage cohort and region determined brideprice, dowry, and bidirectional transfers.

DISCUSSION AND CONCLUSIONS

The present study provides population-based trends and regional differences in marriage payments and their direction for Vietnam during almost four decades of major political and socioeconomic change. Although the results are based on two regional surveys, they cover arguably the two most important areas in the country, namely the Red River Delta including Hanoi in the North and Ho Chi Minh City and surrounding provinces in the South. Not only is this the first study to provide such relatively broad population-based evidence for Vietnam, but comparable data are rare elsewhere in Asia as well. In addition, we explore potential determinants of these trends guided by several theoretical perspectives in the literature. In this final section we attempt interpretation with reference to the specific context of Vietnam. While we believe our interpretations are reasonable, limitations of the surveys and study design described in the data and methods section mean they are not definitive. Still, we believe our findings advance understanding of the situation within Vietnam and contribute to the more general debate concerning marriage payments.

Our most striking finding is the surge of marriage payments in both North and South following economic renovation. Results also reveal that during 1963–1971 (the earliest marriage cohort covered), payments were considerably more common in the South than the North but that these regional differences have substantially contracted and in the case of dowry payments even reversed. In addition more nuanced trends and regional differences are also apparent.

Given strong net effects of cohort and region on marriage payments, it is likely that the socialist agenda impacted behavior although considerably more so for the North than the South. While it seems reasonable to assume that marriage payments declined following the establishment of the socialist government in the North, population-based data for the pre-socialist period is lacking. Thus our supporting evidence is limited to the much lower level for the war cohort in the North compared to the South and ethnographic evidence suggesting that payments were pervasive much earlier. Moreover, the presumed suppressing effects in the North apparently weakened relatively soon. Percentage of dowry began to increase gradually following the early 1970s and its upward trend increased very substantially among post-reunification marriages. Even brideprice increased in the early 1980s. The earlier return of dowry could be because the regime did not attack dowry directly and because dowry was perceived as enhancing the status of newlywed daughters, which echoed the socialist goal in promoting women’s status. Consistent with past studies, we find little evidence that the regime’s attempts to restrict marriage payments in the South following reunification substantially reduced marriage payments.

Since marriage cohort as well as region both compound effects of socio-economic changes and policy transformation related to marriage payment practices, it is difficult to disentangle their respective impacts from the existing data thus complicating interpretation of our findings. For example, the observed North-South differences during war years could reflect the impact of difficult economic conditions severely limiting marital exchanges in the North when basic necessities were lacking and private property confiscated. Likewise the sharp increases in marriage payments following market reform coincides with both a reduced interest in the government in enforcing the suppression of marriage payments and with substantial improvement in the economic conditions enabling much wider affordability for families to practice payments.

Contrary to the North where payments increased modestly following reunification, the South observed a slight decrease in brideprice and bidirectional payments. Such decline could be interpreted either as a result of the socialist agenda or Vietnam’s severe economic stagnation at the time. Under collectivization, southerners may have been less able to afford brideprice. It is interesting, though, that the decline of brideprice in the South was less salient than that of dowry, perhaps suggesting that brideprice was deemed of higher priority.

The government’s greater success in suppressing marriage payments in the North is arguably attributable to several factors, including northern Vietnam’s stronger historical patterns of communality, longer exposure to socialism, more rigorous organization efforts by the Communist Party during the late 1950s, and the strengthening of local cadres during the US-Vietnam War (Goodkind, 1996; Malarney, 2002). Under collectivization, the state replaced the family to a certain extent by providing housing, jobs, and benefits to the population. In such a political climate, northern family strategy for survival likely required knowledge of the government policies, including complying to laws such as brideprice bans, and ideally vertical connections with local authorities as opposed to strong horizontal family networks (Luong, 1993; Kleinen, 1999). Nonetheless, northerners did not passively accept whatever was imposed on them. Malarney (2002) provides evidence suggesting that northern villagers actively negotiated their way and reworked the socialist agenda to fit their traditional beliefs. For example, in his study site villagers were asked to organize weddings at a local government’s building. Yet many resisted because it was deemed more important for the newlyweds to conduct an ancestral worship at home than to follow the Communist Party’s mandate.

During the 1970s and 1980s, young Vietnamese women encountered a serious deficit of male partners because of population growth, war, and excess male migration (Goodkind, 1997). Such a surplus of brides in Africa led to a decline in brideprice (Anderson, 2007) whereas in India it led to dowry inflation (Rao, 1993). Such demographic phenomenon, however, did not produce similar trends in marriage payments in northern Vietnam. Compared to the 1960s, for northerners there were almost no change in the prevalence of brideprice during the 1970s and rising trends in the 1980s. In the South, although brideprice declined slightly during the 1980s, data limitations prevent ruling out effects of population composition. Given that in Vietnam the beneficiary of dowry was the bride and not her spouse or in-laws, the observed increase in dowry during the 1970s among northerners was unlikely caused by a surplus of brides, as evidenced in India.

The reemergence of brideprice predated the market reform by a few years. Nonetheless, the impact of Vietnam’s market transition on marriage payments was likely paramount. During the 1990s, practices of brideprice and dowry surged in both regions, with a majority of marriages practicing some form of payment. The increase was particularly pronounced in the North, given the considerably low level of marriage payments in earlier decades. The increase for brideprice and dowry continued throughout the 1990s suggesting the increase did not peak by the time of the surveys. Also noteworthy is that bidirectional transfers reflecting reciprocity of marital exchanges became the dominant form of payments in recent years. Changes in marriage payments after market reform were plausibly facilitated by both decreased governmental control over household decisions concerning family life and growing household disposable income as a result of market liberalization. Also striking is that the increase in marital exchanges in the South during the 1990s led to levels that far preceded those found in pre-socialist years, thus calling for explanations beyond the unraveling of the socialist agenda.

According to the literature, market reform is anticipated to have profoundly affected economic opportunities and prospective earnings of young people and to have increased the returns to investment in human capital. Indeed, the actual and potential economic value of young women and the earning capabilities of young men are considered instrumental in shaping marriage transaction patterns and in determining the shift from brideprice to dowry in various societies (Anderson, 2007).

Our evidence indicates that marriage cohort was the most robust indicator of recent trends in marriage payments in Vietnam. While we are limited in teasing apart the effects of economic hardship or prosperity from the direct impacts of policies, the cohort effects rarely altered when temporal changes in brides’ and grooms’ socioeconomic characteristics were taken into consideration. At best, our findings appear to lend only modest support to the modernization and human capital perspectives. This implies that the Vietnamese marriage transaction patterns, at least after economic reform, may have become disconnected with socioeconomic potentials of married couples. In other words, structural and policy transformations in the 1990s in combination facilitated the reemergence of marriage payments via more permissive policy and conducive economic conditions, even though the payments might not have served the usual economic functions. Urban or rural upbringing did not differentiate the likelihood of marriage payments. Highly educated grooms were as likely as those with lower education to observe brideprice. The net effects of bride’s education on payments were inconsistent.

What then is the key function of marriage payments among recent marriage cohorts? The reemergence of marriage payments might reflect not only the persistence of traditional values but also a combination of the weakening of the socialist agenda, especially in the North, and the economic prosperity associated with market reform. As economic development does not necessarily bring pervasive cultural change, cultural values could be “an enduring autonomous influence on society” (Inglehart and Baker, 2000, p.19), As such, traditions such as marriage payments might well flourish when economic conditions are favorable. In recent years, brideprice, dowry, and bidirectional transfers may have increasingly become a status symbol for bride’s and groom’s families, as evidenced in higher prevalence of brideprice in the South preceding its pre-socialist level. Given the increasing prevalence of marriage payments, especially in the form of money or gold as either brideprice or dowry, this could potentially affect the wealth distribution across generations and families, even though marriage payments might have been valued for their symbolic meanings rather than their practical functions.

What do our findings suggest about trends in marriage payments in the near future? Younger generations will increasingly come from small families but not necessarily marry at much later ages given that age at marriage has been rising relatively slowly compared to other Southeast Asian nations (Jones and Gubhaju, 2009). Perhaps most importantly, wealth display as status symbol will likely continue to grow in importance as Vietnam’s economy expands. Within this evolving context, it likely that families will be better financially equipped to invest in both their children’s human capital (e.g., schooling) as well as in traditions closely tied to social prestige such as marriage payments. Thus, in years to come, brideprice, dowry, and bidirectional transfers will likely continue to be prevalent and culturally functional, regardless of their economic functionality.