Street safety and girls’ secondary schooling in India

Street harassment in India is still a constraint on women's education, dimming their future. Analysing the case of ‘SHE teams’ in the state of Telangana, this article shows that the introduction of special police units dedicated to women's safety on streets led to a significant increase in girls finishing grade 12. Further, it establishes that the impact is not driven by non-street safety determinants of schooling.

A Government of India-endorsed survey across the states of Assam, Delhi-NCR (National Capital Region), Madhya Pradesh, Maharashtra, Telangana and West Bengal, reveals a harrowing reality about street safety perception among girls (Save the Children, 2018). Over a third of teenage girls report facing distressing situations in public areas, such as unwelcome comments, uncomfortable stares, stalking and harassment by men. Alarmingly, only a third of teenage girls and their parents trust local police to act effectively against cases of street harassment. As a mark of desperation, 20% of parents consider early marriage to shield their daughters from such dangers in public spaces. These distressing experiences can push many girls to leave school early, a trend echoed in studies by Hebert et al. (2019) and Hardt et al. (2022). A World Bank Report estimates a loss of US$15 trillion to US$30 trillion in lifetime productivity and earnings due to failing to provide 12 years of schooling to girls in countries like India (Wodon et al. 2018).

Research shows that when choosing colleges, girls often prioritise proximity to home over the quality of education (Borker 2021). In Bihar, providing bicycles made commutes safer and increased girls’ secondary school enrolment (Muralidharan and Prakash 2017). Similarly, placing civilian guards along school routes in Chicago reduced crime in school areas and improved student attendance (McMillen et al. 2019). In recent research (Ashutosh 2023), I highlight how improving street safety for women can have a powerful impact, enabling more girls to complete their schooling till grade 12.

Special police units to prevent street harassment

The deployment of special police units to prevent street harassment has the power to transform perceptions of safety among girls and their parents. In theory, these units can serve dual purposes: first, by apprehending offenders and deterring potential harassers through visible patrolling; and second, by fostering trust through empathetic support for victims. This newfound sense of security can significantly reduce school dropouts, empowering more girls to complete grade 12.

However, when harassment cases are handled insensitively by these specialised units, they risk drawing unwanted attention to the victim or her family, potentially discouraging girls from attending school. This adverse outcome may be exacerbated by the presence of these units, as their proactive monitoring and capture of such cases could unintentionally increase public visibility of the incidents. In contrast, the absence of such units might result in cases remaining unreported, thereby shielding victims from unwanted scrutiny, albeit at the cost of justice and accountability. Hence, the impact of such an intervention is a priori unclear and warrants an empirical evaluation.

I examine whether introduction of special police units in the state of Telangana led to rise in the rate of girls completing their schooling till grade 12 due to an enhanced sense of street safety.

‘SHE teams’ in Telangana

On 24 October 2014, the Telangana state government launched 'SHE Teams' in the capital city of Hyderabad – a specialised police unit focused on addressing violence against women and girls on the streets. Police officers in SHE Teams receive orientation and capacity-building training designed to enhance their gender sensitivity when addressing street harassment cases. They are also trained to utilise SHE software and dashboards effectively, ensuring optimal implementation of women’s safety measures on streets.

This initiative seeks to enhance the safety of women and girls in public spaces by deploying officers, both female and male, in uniform and plain clothes to patrol harassment-prone areas such as outside schools and colleges, transportation hubs, markets, and parks. By 1 April 2015, the intervention had expanded to all districts, providing victims with additional reporting options through QR codes, WhatsApp, and email. While crime reported for child-related offenses (for both girls and boys) between 2010 and 2014 was nearly identical in what is now Telangana and Andhra Pradesh¹, the introduction of SHE Teams in 2014 was followed by a significant increase in reported cases in Telangana, highlighting a shift in reporting practices (Figure 1).²

Figure 1. Crime reported against children before and after introduction of SHE Teams

Source: Author’s illustration using data from National Crime Records Bureau (Government of India).

Data and methodology

I use data from two rounds of the National Family Health Survey (NFHS) (2015-2016 and 2019-2021) to estimate the impact of SHE Teams on grade 12 completion among girls in Telangana. My sample includes girls and boys aged 18 to 21 from both rounds of the survey who attended school in any grade between 6 to 12 in Telangana and its neighbouring states (Andhra Pradesh, Karnataka and Chhattisgarh).³ Those in the age cohort between 18 to 21 years during 2019-2021 would have been of age between 15 to 18 years during 2015-2016, a period when they should have been in grades 9 to 12. This means girls of Telangana in those age cohorts were exposed to SHE Teams before finishing school. In contrast, those aged 18 to 21 in 2015-2016 completed grade 12 without the intervention. Comparing grade 12 completion rates for this age cohort (between 18 to 21 years) between the two rounds of NFHS, that is, between 2015-2016 to 2019-21, allows the estimation of the effect of SHE Teams on schooling completion.

I employ 'triple difference-in-difference' methodology to estimate the causal impact of the SHE intervention on grade 12 completion rates among school-going girls in Telangana, compared to their counterparts in neighbouring states. This approach incorporates three layers of comparison: gender (girl or boy), state domicile (residing in Telangana or in one of the neighbouring states), and time period (pre- and post-SHE intervention).

The primary outcome variable is whether an individual who attended any grade between 6 and 12 successfully completed grade 12. Interpreting the coefficient of the ‘interaction terms’ representing these three layers of comparison, is crucial. This coefficient estimates the relative change in the percentage of school-going girls in Telangana completing grade 12, vis-à-vis girls in neighbouring states, attributable to the SHE intervention. The term 'relative' here signifies that the analysis measures changes in the grade 12 completion rate for girls relative to boys within both the intervention region (Telangana) and the comparison region (neighbouring states), before and after the intervention. This approach highlights the differential impact of the intervention on girls' schooling outcome. Incorporating boys into the analysis strengthens its robustness by accounting for non-gendered factors influencing schooling, such as changes in overall school budget allocations at both the central and state levels of administration.

Similarly, I also compare the rate of dropout from school between Telangana and the neighbouring states among girls due to a list of reasons unrelated to street safety, to rule out the gains in schooling completion after SHE intervention from other sources. These reasons include work responsibilities (“household work”, “work on farm / family business”, “outside work for payment in cash or kind” and “care of siblings”), child marriage, accessibility to schooling (“distance from home”, “schooling fees”), and facilities at school (“hygiene areas for girls”, and “inadequate women teachers at school”). Finally, I compare the dropout rate from school among girls due to feeling unsafe while commuting to school before and after SHE intervention between Telangana and neighbouring states.

Further, I also compare the rate of women of Telangana being employed as compared to women in the neighbouring states, to show that gains in schooling outcome were unrelated to labour market conditions for women.

Findings

In Figure 2 below, I present the findings. Following the intervention, the percent of school-going girls in Telangana who completed schooling up to grade 12, relative to school-going boys in the state, was 13.5 percentage points higher compared to the corresponding percent for girls relative to boys in neighbouring states. Since before the intervention, only about 36% of girls in Telangana completed grade 12 (NFHS, 2015-2016), this means that the SHE intervention caused an increase of 37.3% in school completion for girls in Telangana.

Figure 2. Changes in outcomes after SHE Teams interventions

Notes: (i) School completion refers to finishing grade 12. (ii) No change in school completion rates for women aged 25 to 30 years (who completed school long before the SHE Teams intervention) provides evidence that the dependence of schooling gains observed for the younger group (18 to 21 years) on the intervention are genuine. (iii) The intervals around the average values in the figure are calculated at 95% confidence interval (CI). A 95% CI means that, if you were to repeat the experiment over and over with new samples, 95% of the time the calculated CI would contain the true effect.

My methodology ensures that this change remains unaffected by time-varying non-street safety determinants of girls’ schooling at the district level such as construction of roads, schools, expansion of electricity access to households, etc. Also, the methodology ensures that the estimated gain in schooling completion remains unaffected by differences in time-invariant determinants of schooling between Telangana and neighbouring states such as cultural preference of households for girls’ schooling.

Examining the drivers behind this progress in schooling completion, I find that almost 34 percentage points fewer girls in Telangana dropped out after the SHE intervention compared to girls in neighbouring states, because they no longer felt unsafe travelling to school (as shown in Figure 2).⁴This highlights a crucial impact: school-going girls in Telangana felt significantly safer on their commutes post-intervention than their peers in neighbouring regions. Furthermore, the same figure shows that difference women's chances of employment in the labour force remained unchanged between Telangana and neighbouring states after the intervention. This supports the argument that the rise in school completion rates is not driven by improved labour market conditions for women, but rather by enhanced safety for girls while going to school.

Moreover, the gain in schooling completion after SHE intervention appears unlikely to be driven by non-street safety channels (see Figure 2). The rate of dropping out of school among girls relative to boys remained similar after the intervention, between Telangana and its neighbouring states due to non-street safety reasons including work responsibilities, facilities at school, physical and financial accessibility to school, and marriage.

Conclusion

The SHE intervention in Telangana illustrates the transformative impact of ensuring street safety for girls. By fostering a secure environment, the intervention empowered more girls to complete their education, demonstrating the crucial role of safety in shaping educational outcomes. Policymakers should consider scaling such initiatives nationally, providing dedicated and empathetic police forces to protect girls and build community trust. However, it is vital to continuously monitor and evaluate these interventions to prevent potential misuse (such as excessive moral policing of girls by special police units) or unintended consequences, ensuring that safety measures remain effective and supportive for young women's education.

Notes:

In 2014, due to public demand, the state of Andhra Pradesh was split into Telangana and the residuary Andhra Pradesh by an Act of Parliament.
National Crime Records Bureau (NCRB) collected gender-disaggregated data on crimes against children, including stalking cases against girls, only for the years 2014 and 2015. Due to this data limitation, this study examines all forms of crimes against children in Telangana and Andhra Pradesh from 2010 to 2020 to assess shifts in crime reporting after the introduction of SHE Teams in Telangana. Given that girls are more susceptible to violence and crime than boys in India, the study assumes that any observed changes in the reporting of crimes against children will also reflect changes in the reporting of crimes against girls.
These states share their geographical boundaries with Telangana. Moreover, during 2015-2016 to 2019-2021, these neighbouring states did not have similar women safety centric police units like SHE units in Telangana.
NFHS collects information on school dropouts due to safety concerns exclusively for girls. As a result, this analysis compares the dropout rates of girls in Telangana with those of neighbouring states, omitting boys from this specific comparison.