Early Cohabitation and Intimate Partner Violence: Evidence from Cambodia

Intimate partner violence (IPV) disproportionately affects women in many countries. The prevalence in low- and middle-income countries is particularly high for various reasons, ranging from severer gender-discriminatory cultures to wider gender gaps in education. The Covid-19 pandemic has worsened the pre-pandemic situation due to increased economic stress and frequent lockdowns. Many studies have documented that IPV has huge impacts on women’s health and children’s health. Given its prevalence and consequences, IPV remains an important issue.In developing countries, child marriage is often considered a leading cause for IPV.Policymakers assert that women who get married before age 18 are less likely to work, obtain lower educational attainments, and have limited decision-making power in their households, leading to their higher vulnerability to IPV. Supporting this claim, several empirical studies from different contexts have found that there is a negative correlation between age at marriage and probability of experiencing IPV and concluded that the prevention of child marriage will play a vital role in lowering the prevalence of IPV. In Cambodia as well, we have found negative correlations between these two variables. However, does this correlation indicate the causal effect of child marriage on IPV?Despite the importance of this question, few studies have examined the causal relationship. If there is a causal relationship between age at marriage and women’s vulnerability to IPV, policymakers can implement women-targeted policies to delay the timing of marriage (e.g., conditional cash transfers). However, if there is a confounder in this relationship, previous findings on the correlation do not imply causality, and thus, interventions based on this correlation may turn out to be ineffective. One potential confounder is a gender-discriminatory culture.Since the prevalence of the gender-discriminatory culture lowers women’s age at marriage and increases their vulnerability to IPV, the omission of this variable negatively biases the estimated coefficient. If this is the case, the correlational findings might have overestimated the potential effects of raising women’s age at marriage on their vulnerability to IPV, suggesting limited effectiveness of proposed policies. If the downward bias is so severe, simply raising women’s age at marriage may have unintended consequences. Thus, it is important to identify the causal effects of women’s age at marriage on their vulnerability to IPV instead of relying on correlational findings.Using the individual-level data from Cambodia, we show how early cohabitation affects the attitudes of ever-married women toward IPV. Due to the data availability, we use cohabitation instead of marriage. To isolate exogenous variation in the timing of cohabitation, we use age at menarche as an instrumental variable for age at first cohabitation, and thus, our estimates are relevant to women whose timing of cohabitation is related to the timing of menarche.We find that later cohabitation makes women more tolerant of IPV among women whose timing of cohabitation changes due to random variation in the timing of their menarche. Our data suggest that the correlation between cohabitation timing and timing of menarche is stronger in the northeastern provinces, where early cohabitation is more common. Exploring the mechanisms, we find that later-married women are less likely to work after marriage, their partners are less educated, and their household wealth is smaller. These findings seem to suggest a premium of younger brides in the traditional marriage market, which may potentially offset the negative vi Early Cohabitation and Intimate Partner Violence: Evidence from Cambodia effects of early cohabitation on IPV. In line with these findings, our suggestive evidence shows that the estimated effects of cohabitation timing on the probability of experiencing different forms of IPV are statistically indistinguishable from zero although this finding should be taken with caution as the sample size halves in this part due to the questionnaire design.Our findings imply that simply incentivizing women to delay their first cohabitation may have unintended consequences due to a potential premium of younger brides in the traditional marriage market. If brides are younger, spousal quality seems higher, suggesting that men’s preference for younger brides plays a role in determining women’s intra-household status and their vulnerability to IPV. Therefore, instead of solely focusing on women to prevent early cohabitation and IPV, comprehensive interventions that affect both men and women will be necessary.