Written by Tanushree Goyal and Sam van Noort.

Child marriage is a major human rights violation that disproportionately affects young women, and is a widespread problem in the developing world. In India, marriage is legally allowed at the age of 18 for females, except for Muslim women who wish to marry under Mohammedan Law or Sharia law. Child marriage casts a long shadow not only on the young woman’s life but also on that of her children. Worldwide, more than 700 million women were married before their 18th birthday. Moreover, the dismal rate of progress on reducing child marriages is outpaced by population growth. Despite legal requirements in most developing countries that fix the age at marriage to be 18 for women, one in nine young girls worldwide is married off before 15.

We find that the Indian police is severely understaffed as per global standards. India has on average 159 policemen per 100,000 inhabitants, and there is almost no change in staffing trend over time. These numbers are much below the global mean of 300 policemen per 100,000 inhabitants

This suggests weak policy and law enforcement by the state, an aspect that has surprisingly remained unexamined. While cultural beliefs and gender norms sustain the practice of child marriage, an emphasis on them limits the scope for policy intervention. On the other hand, by examining the role of the state, our research shifts the emphasis towards the state’s ability in implementing its own policy throughout its territory (state capacity). We argue that individual women who live in regions with higher state capacity have lesser chances of being a child bride.

We focus on the Indian context, a country with the highest absolute number of child brides in the world, our research shows. Despite India’s long-standing and continued vocal commitment to end child marriage, as well as setting legal amendments that have raised the marriage age for women over three times in a century, one in three global child brides are in India. Towards this end, we examine the relationship between state capacity and a woman’s propensity to marry before the nationally set legal age of 18 years, across Indian states.

Using recent micro-level survey from IHDS, we find that an Indian woman living in a region with higher state capacity has a significantly lower probability to become a child bride. We measure state capacity along two dimensions: coercion and care. Using hierarchical logistic regression models, we find that the relationship between coercive state capacity and child marriage is substantive, and alone explains 53 percent of the subnational variation, while, caring state capacity has no effect.

Examining millennial child brides in India

Our main sample consists of women, between 18-28 years of age as of 2012, that is, young millennials, born in the years between 1984 and 1994. These women grew up in the economically expanding and well-established democracy of India. Yet, we find that even in this recent sample, 22 percent of women are married before 18, with a mean age at marriage of 15.3 years.

Like previous studies, we find substantive within country subnational variations in India (see figure 1). This phenomenon is not unique to India and is shared by many other countries in West and Central Africa and Latin America. These within-country variations are substantive and continue to puzzle policymakers.

Figure 1. Child marriage rates across India

Regional variations in state capacity across Indian states

We focus on two core aspects of state capacity. The first is coercive state capacity, which we define as the capacity of the state to coerce citizens to abide by its laws and regulations. We measure this as the average number of police officers per 100,000 inhabitants from the years 2005 to 2012.  We find that the Indian police is severely understaffed as per global standards. India has on average 159 policemen per 100,000 inhabitants, and there is almost no change in staffing trend over time. These numbers are much below the global mean of 300 policemen per 100,000 inhabitants and below the United Nations recommended minimum police strength.

The subnational picture in India is even more striking, with a range from 64 in Bihar to 1044 in Mizoram (a border state with separatist/insurgency movements). This is in line with criticism of the view that India’s bureaucracy is bloated, instead the problem is too many officials in the wrong place and not enough in the right place.

The second dimension we examine is caring state capacity, that is, the capacity of the state to effectively deliver public services that empower and inform vulnerable societal groups, such as adolescent girls and pregnant women against child marriage. We measure this dimension of state capacity by two variables: (a) the average quality of Anganwadi centres (AWCs); and (b) average number of AWCs per 100,000 inhabitants from 2003 to 2011. AWCs are rural mother and child care centers in India which are additionally responsible for implementing programs targeted to raise awareness about child marriage among teenage girls and their parents, and for maintaining birth and death registration. Importantly, these centers also have the duty to inform the police about probable or planned incidences of child marriage.

The importance of effective states

We find three important results. Studying women with the equivalent individual and household characteristics, who live in different states, we find that about 10 percent of all differences in child marriage rates exist between states, and therefore, can only be explained by contextual factors that vary between them.

Second, an increase of 85 police officers per 100,000 (one standard deviation) leads to, on average, a 32% decrease in the probability of a woman being married before attaining 18 years of age. This is the equivalent of a 7.3% decrease in average state-level child marriage rates with every standard deviation increase in coercive State capacity. Coercive State capacity, even in several robustness checks, is always statistically significant, and explains more than 53% of all (macro) differences in child marriage rates across our sample of large Indian states. The decline is stronger for the most at risk girls as compared to the common Indian woman (see figure 2).

Figure 2. Decline in probability of child marriage for most common and at-risk woman with increase in coercive State capacity

Third, the caring dimension of state capacity is neither strongly related to child marriage nor explains many variations. This is a surprising finding. Nevertheless, this ties in with findings from qualitative research, that find that AWCs need reform, and in their current form are overburdened and inefficient. Anganwadi workers have been found to be too close to the communities they serve to remain impartial, or face social sanction when they go against their communities, while some may even reinforce norms that promote child marriage.

These findings are robust to many control variables such as direct-indirect colonial rule, ethnic party in legislature, corruption, the presence of natural resources, social capital, regional cultural norms – variables that are plausibly related to both coercive state capacity and child marriage. The results are robust to many alternative samples (including all 33 regions, except Andaman and Nicobar and Lakshadweep), excluding Muslim women, J&K, and so on, and estimation techniques (linear model, probit).

Lastly, we identify the causal effect of coercive state capacity on child marriage by exploiting the fact that some Indian woman are exposed to significantly higher levels of police officers per 100,000 inhabitants because they happen to live in a state with insurgency/ independence movements. Using instrumental variable probit models, we find that Hindu women that live in such states are significantly less likely to be married as a child, as compared to Hindu women in other regions, despite equivalent individual-level background characteristics. The identifying assumption is that insurgency movements do not have a direct effect on child marriage. Since we possibly cannot assign state capacity randomly, this is, in our view, the most convincing evidence possible for the hypothesised relationship.

Concluding remarks

Our findings are broadly applicable to other contexts and stand in contrast to the singular emphasis in policy research on the role of an individual level economic or cultural and gender norms in shaping individual incentives to marry their daughters in childhood. We do not find that these factors are unimportant, we simply suggest that reducing child marriage is much more of a law enforcement problem than is widely believed, and that governments are well-advised to pay more attention to law enforcement than they have done in the past.

Our findings also speak to the dominant institutional literature that examines the role of state capacity on a host of macro outcomes, such as economic growth, civil wars, human rights and so on. By contrast, we contribute by showing that state institutions have a deep impact on the very private behaviour of individuals.  To understand exactly how state capacity is related to child marriage, we need detailed research to illuminate the mechanisms and use surveys designed to elicit citizen’s interaction with state institutions. Our study strongly suggests that this exploration is necessary and worthwhile, given the gravity of the matter at hand.

Tanushree Goyal is a PhD candidate at the University of Oxford where her research focuses on what shapes political and bureaucratic effort towards access and quality of public goods provisions using newer data. Sam van Noort is a PhD candidate at the University of Oxford where his research focuses on the effect of state capacity on economic growth. Image Credit: CC by Wikipedia Commons.