In patriarchal societies there are often expectations for a woman to live with her husband’s family after marriage. Kandiyoti (1988) identified the patrilocal extended household as a key element to the operation of ‘classic patriarchal systems’, which are most often found stretching from the Mediterranean to the Pacific, also termed the ‘mother-in-law belt’ ( Ember & Ember, 2003). This practice may be further perpetuated by modern labor migration where a woman’s husband leaves the home for extended periods of time and sends remittances to his family, thereby necessitating the woman to live with her in-laws for support during his absence ( Galam, 2017; Jama-Shai et al., 2017).

Women typically enter their new household with little status or power and their mother in-law, as the senior woman in the house, gains higher social status and more decision-making authority ( Kandiyoti, 1988). In Nepal, women reported they always were the last to eat in the household and sometimes went without food at the direction of their mother-in-law as punishment, even during pregnancy ( Pun et al., 2016). Women have also reported their mother-in-law is a primary decision maker in their healthcare decisions ( Acharya et al., 2010; Ganle et al., 2015), including needing their mother-in-law’s permission to leave the house and visit a health center ( Hyder et al., 2007).

A number of studies have identified the mother-in-law as a gatekeeper or barrier for accessing maternal health services for women ( Gupta et al., 2015; Pun et al., 2016). A study in Mali found that maternal health behaviors of women were strongly associated with the preferences of their mother-in-law ( White et al., 2013). For example, women whose mother-in-law reported negative attitudes towards delivering in a health facility were less likely to receive care in a facility regardless of the woman’s own beliefs towards traditional birthing practices ( White et al., 2013). Talbert et al. (2016) explained that women in Kenya felt pressure to adhere to the breastfeeding advice from their mother-in-law because they lived together and were expected to always obey.

A handful of studies in India have reported some positive effects for women living with their mother-in-law. Varghese & Roy (2019) found that women who lived with their mother-in-law during pregnancy were significantly less likely to experience severe maternal anemia and more likely to take regular iron supplements. Another study evaluated the impact of an intervention to improve the mother-daughter-in-law relationship and found a positive relationship with the mother-in-law protecting against violence from a woman’s husband ( Krishnan et al., 2012).

To date, most published literature stems from qualitative research ( Hyder et al., 2007; Jama-Shai et al., 2017; Krishnan et al., 2012; Nasrullah et al., 2015) with limited quantitative data. Studies which do include quantitative data have used a variety of methods for measuring co-residence status with the mother-in-law, from primary surveys which ask coresident status directly ( Falb et al., 2013; Falnes et al., 2011; Gupta et al., 2012; Jewkes et al., 2019; White et al., 2013), to estimates from national surveys ( Kiros & Kertzer, 2000) and census data ( Gibson & Mace, 2005; Huber et al., 2017). Previous research has taken advantage of household schedules to identify mother-in-law/daughter-in-law cohabitation ( Speizer et al., 2015; Varghese & Roy, 2019). It is also unknown how the globally increasing age of marriage ( Liang et al., 2019) has impacted trends in co-residency.

In order to gauge the extent to which married women around the world live with their mothers-in-lay, there remains a need for an established quantitative method for measuring coresident status across countries and time using publicly available data. Research focusing on the health and well-being of daughters-in-law could be greatly expanded if researchers were able to use the large datasets collected by nationally representative surveys. Research could span demographic, economic, health, and many other fields, documenting both how living with a mother-in-law impacts daughters-in-law, but also how this relationship has changed over time and varies across settings.

This paper measures co-resident status with mothers-in-law using widely available data to compare trends across time and region. The Demographic and Health Surveys (DHS) program collects nationally representative data on a range of health topics approximately every five years in most countries, making it a useful survey to examine the health status within and across countries and time. Using data available from the household schedules of 250 surveys from 75 countries we are able to estimate the share of married women who live with their mothers-in-law, compare prevalence between countries and regions, and examine changes over time. For select countries with large changes we decompose changes in co-residence into changes in co-residence by age and changes in the age structure of married women.

To identify relationships between household members in the DHS, we used household schedules. As part of the household questionnaire, an interviewer asks one member of the household to list all usual members and visitors and collects data on each person’s age, sex, relationship to household head, and other information ( DHS). While a household schedule was collected as part of Phase 1 of the DHS (1984–1989), relationship to head of household was introduced in Phase 2 (1988–1993) ( Institute for Resource Development/Macro International, Inc, 1990). 258 surveys have available data for Person Recode files which are constructed from the household questionnaire with each member of a household as his or her own row.

Using the Person Recode files, we identified the labels and values for possible responses to the question of relationship to the household head. These relationships were then cleaned and classified into one of the following categories: head, wife or husband, son or daughter, son-in-law or daughter-in-law, parent, parent-in-law, sibling, grandparent, grandchild, adopted child, other relative, not related, and other. During data cleaning, one survey was removed from the analysis because it did not contain labels for DHS variable “hv101,” the variable containing the relationship to head of household, (Jordan 1990) and seven surveys were excluded because they did not differentiate between parents and parents-in-law (Bolivia 2003, Dominican Republic 1996, Dominican Republic 2002, Dominican Republic 2007, Dominican Republic 2013, Nicaragua 2001, and Peru 1992). Because relationships are only recorded in reference to the household head, we created a relationship matrix to identify the potential relationships between two people based on their relationship to the head ( Table 1).

For each household member, we created a variable with the household identification information, line number, age, and sex of the household member. Transforming this data from long to wide created a dataset where each line represents a household, and each column is a single member. This dataset was then merged to the original Person Recode file, so that each member of a household also contained information on all the other household members. This file was reshaped from wide to long, resulting in dyads of two household members as observations. Finally, the relationship matrix was used to identify the relationship of the two members in the dyad.

To identify mother-in-law/daughter-in-law cohabitation we selected all female residents between the ages of 15 and 49, in order to match the age range of women eligible for the individual questionnaire. We classified a woman as living with their mother-in-law if they had a dyad relationship where they are the “son-in-law or daughter-in-law” of the other member and if the sex of the other member of the dyad is female. We merged this information with the Individual Recode file (women 15-49) and selected currently married women. It was necessary to include the Individual Recode file because not all household schedules recode marital status.

With the Individual Recode file we calculated the percent of married women who live with their mothers-in-law in total and by age group. For four example countries, we constructed a decomposition of two rates ( Kitagawa, 1955) to determine if changes over time are attributable to changes in the age structure of married women or changes in the rate schedule by age.

All code used to transform household rosters and individual recode files into mother-in-law analysis and figure is written in R using Rstudio version 1.2.5019 and is archived on Zenodo ( Bietsch, 2021). The code uses multiple packages which need to be installed in R; the full list of packages and links to their Cran documentation is available on the author’s GitHub.

While no overarching global pattern emerges, many interesting regional trends appear from this analysis. In Central and Southern Asia, which had some of the highest rates of co-residence in the 1990s, most countries have seen significant increases in co-residence. Nepal’s recent large increase in co-residence coincides with large scale migration of husbands to foreign countries for work. However, when looking at the remainder of Ember and Ember’s ‘mother-in-law belt,’ we see a decline in women living with their mothers-in-law. This is particularly true in Egypt and Turkey, which have had large declines, not caused by the age structure of the married population but by declines in daughters-in-law living with their mothers-in-law. Most countries in sub-Saharan Africa are also experiencing a large decline in co-residence, though co-residence was never as common as in other regions.

Given the global variation in the proportion of women living with their mother-in-law and the limited research which has accounted for this factor, there remain many future directions for research. Future studies may assess the timing of childbearing. Furthermore, does living with their mother-in-law influence the type of contraception a woman uses? As prior studies have reported some women need permission from their mother-in-law to leave the house ( Hyder et al., 2007; Rew et al., 2013), this could impact her ability to visit a family planning clinic or resupply short-term methods. Other studies have reported a woman’s mother-in-law has decision-making power over a daughter-in-law’s health ( Acharya et al., 2010; Ganle et al., 2015), therefore future studies may determine whether living with a mother-in-law influences who is the primary family planning decision-maker. Finally, how does living with the mother-in-law impact a woman’s perceived autonomy and empowerment within the household, the relationship with her husband, and in her healthcare decisions?

One limitation in this research is that we are unable to identify mother-in-law residence in some households that contain several branches of a family co-residing. Because we are identifying in-laws through the household head, if the household head is the not directly related to the mother-in-law or daughter-in-law, in some cases the relationship cannot be established. For example, looking at the relationship matrix in Table 1, if either member is identified as “other relative” then a mother-in-law relationship cannot be established. We believe that these cases will be infrequent, but may cause an undercount of co-residence.

By creating a methodology for studying household dyad relationships, we have established a means of identifying mother-in-law co-residence for women in Demographic and Health Surveys. We have found global change in the number of women living with their mothers-in-law; in some regions this trend is increasing, and in others it is decreasing. In countries with the largest change, the change is overwhelmingly caused by changes in the rates of women cohabiting, not in the age structure of married women. Our goal in this research was to shed light on how many women live with their mothers-in-law and produce replicable code to allow other research to continue to explore how co-residence impacts women’s lives in a myriad of ways.