Health insurance has been accepted as one of the channels that can serve the purpose of universalisation of health care and especially useful in the period of health crisis. This is the emerging and challenging sector in India. The present study explores the health insurance penetration in India using the latest two rounds of nationally representative datasets of the National Family Health Survey (2005-06 and 2015-16). This study identifies the covariates of households' participation and their choice for different health insurance schemes, using the average marginal effects of binary and multinomial logit regression models with conditional categories after checking their Kernel density function. The results suggest that health insurance in India is more skewed towards the households belonging to higher wealth quintile and in the front runner states. Age, occupation and education level are also positively associated with health insurance enrolment. The variable 'weighted information index', which is estimated through Principal Component Analysis, is a significant determinant of private health insurance and community-based health insurance, and this also caters to the richer households. The study also identifies several shortcomings of not accessing health insurance, their choices and suggests reforms with the goal of universalization of healthcare. The robustness check has been performed using 20 percent, 50 percent and 80 percent sub-sampling of the models.
