400 Million Dreams!

Movement of people is a fundamental economic activity having significant influence on public policy. Typically, studies on migration have relied either on the decennial Census or on irregular Surveys. By the time Census data is published, typically 6-7 years after collection, it is already dated. Moreover, Census data – though comprehensive and granular – captures a stock of migrants rather than dynamic, seasonal trends. This paper seeks to address this gap by harnessing three high frequency and granular datasets – Indian Railway Unreserved Ticketing System (UTS) data on passenger volumes; mobile telephone roaming data from the Telecom Regulatory Authority of India (TRAI); and district level banking data, to provide near-real time insights into migration flows. Our findings indicate that overall domestic migration in India has slowed. Since Census 2011, the number of migrants has reduced by about 11.78%, to ~40 Crore and the migration rate (percentage of migrants in the population) reduced to 28.88%. We hypothesize that this is on account of improved conditions and opportunities at the places of origin. Further, spatial trends in migration appear to be in line with gravity models of migration, with short distance migration accounting for the largest share. We go on to identify popular destinations, major origin districts, and build dyads of popular migration routes.