• After focusing on expanding physical access for last few decades, India is confronting the challenge of 'small' schools and moving toward school closure or rationalisation. But potential impact of such a policy remains under-investigated. • 'Small' schools are situated in villages which are relatively backward compared to the villages in which non- 'small' schools are situated. Students attending 'small' schools have fewer alternatives in that village. • 'Small' schools have higher probability of multi-grade teaching and are less likely to have basic infrastructural facilities. • Learning outcomes of students in 'small' schools are not different than that of students in non- 'small' schools. • School rationalisation is likely to have adverse consequences on children's access to schooling without any gains as far as learning levels are concerned. Alarmed by declining enrolment in government schools and potentially adverse academic, administrative and fiscal consequences associated with it, policy makers in India have initiated experimenting with closure of government schools with low enrolments ('small' schools), an exercise commonly referred to as 'school rationalisation'. However, the impact of this policy on access to schooling and learning remains empirically unexplored. Utilising ASER 2014 data, this paper asks three key questions: (a) what are the characteristics of villages in which 'small' schools are located?, (b) what options would students have if 'small schools' were to be closed, and finally (c) what are the differences in characteristics of 'small' and non-'small' schools? Results indicate that the villages which have 'small' schools are more disadvantaged in terms of essential public services such as all-weather roads leading to village, availability of government health facilities or banks and post offices. Additionally, these villages are less likely to have an alternative to the 'small' school, either government or private. Results also show that 'small schools' are much more likely to have multi-grade teaching. They are less likely to have basic infrastructural facilities. Interestingly, learning levels are unlikely to be different in 'small' schools than non- 'small' schools even after controlling for child, household and village attributes. Thus, the analysis suggests that school rationalisation can potentially have severe consequences on children's access to schools without any meaningful impact on learning levels in a 'business as usual' scenario. [ABSTRACT FROM AUTHOR]