Income inequality is one of the most significant socio-economic challenges confronting India, with potentially long-lasting implications for the future of its democracy and society. In this work, we explore income inequality in India, without assumptions of equilibrium, and illustrate the nature and direction of re-distribution within the income distribution in a dynamic sense. Given that both mean income and income inequality show a rising trend post the Industrial Revolution, we argue that such a process is appropriately modeled using Geometric Brownian Motion (GBM). Specifically, we use the mechanism of GBM with a reallocation parameter (which indicates the nature of re-distribution occurring in the income distribution) proposed by Berman et al. We find that since the mid-1990s, reallocation is negative, meaning that incomes are exponentially diverging, indicating that there is a perverse re-distribution of resources from the poor to the rich. It has been well known that static inequality is rising in India, but the assumption has been that while the rich may be benefiting more than proportionally from economic growth, the poor are also better off than before. The surprising finding from our work is that the nature of income inequality is such that we have moved from a regime of progressive to regressive re-distribution. Essentially, continued impoverishment of the poor is directly spurring multiplicative income growth of the rich. We characterize these findings in the context of increasing informality of the workforce in the formal manufacturing and service sectors, as well as the possible evolution of negative net incomes of the agriculture workforce in India. Significant structural changes may be required to address this phenomenon.