The global experience has been that as countries develop, rural-to-urban migration accelerates, and decelerates only when the urbanisation level is very high – usually well over 50%. In contrast, migration in India began decelerating when urbanisation was below 25%. In the article, Pronab Sen deconstructs this puzzle.
The global experience has been that as countries develop, rural-to-urban migration accelerates, and decelerates only when the urbanisation level is very high – usually well over 50%. In India, however, migration began decelerating when urbanisation was below 25%, and the trend continued over three Censuses – 1991, 2001, and 2011. This principal puzzle contains three subsidiary puzzles, which I shall discuss here.
A simple explanation of the principal puzzle is that India is actually far more urbanised than official measures suggest, and hence, not an outlier. As per recent studies, the Indian ‘urban’ definition is among the most stringent in the world, involving three criteria – population size, population density, and proportion of adult males employed in agricultural activities. Most countries have two criteria; some have just one. If we drop one of our criteria, the urban proportion is 40-70% versus the official estimate of about 32%.
But the migration rate began decelerating in the 1980s, when the urbanisation rate was low by any yardstick. Also, the “smell” test suggests that the Indian definition is likely closer to the mark than others. Then, the puzzle remains.
As the principal strategist in the Planning Commission for four five-year plans, my curiosity was roused by a compelling characteristic of Indian policymaking: Almost all government schemes for improving rural conditions are justified on the grounds that they will reduce migration to cities. Surely, improving village living conditions should be sufficient for obtaining political approval? Why tag on potential effect on migration? This was inexplicable given that growth is strongly linked to urbanisation.
The urbanisation-growth link: Theory and evidence
Global and Indian experiences (across states) show that productivity and growth are strongly correlated with urbanisation. Given that urban productivity is much higher than rural, a shift in labour force from rural to urban activities is a key source of growth. International evidence also suggests that allowing existing urban agglomerations to grow may be a more efficient strategy than creating new urban areas.
The theory in favour of urbanisation is also quite compelling. There are three economic explanations for the link between urbanisation and productivity. First, it deepens local product and labour markets, leading to greater competition and efficiency. Second, it permits more specialisation and division of labour, and hence, productivity improvements. Third, it enables greater learning at all levels of workers, and better technology adoption and innovation. Also, urbanisation lowers transaction and logistical costs, and permits greater economies of scale and scope, all of which contribute to growth.
Urbanisation is also strongly linked to rapid improvements in social indicators, such as health and education, as economies of scale and scope are more pronounced in the supply of these services than even in industry. Health and education indicators are persistently higher in urban areas relative to rural, across and within Indian states.
What explains political aversion to rural-urban migration?
This raises the first subsidiary puzzle: If the evidence is so compelling, why is our political system, across parties and ideologies, so averse to rural-urban migration?
The reason may lie in the imperatives of gaining and retaining political power. In a country where political success is driven by managing the 3 Cs of Indian society – caste, community, and class – no incumbent political leader would like to see any uncontrolled change in the social configuration of the constituency, and therefore of the winning coalition. Migration causes this both in the originating villages and destination towns. Initially these effects may be relatively small, but can snowball over time since much of migration is driven by social networks.
The political system seems to have succeeded: 80% of Indian urban growth is organic in that it arises from three predominant sources: (a) natural population growth; (b) absorption of neighbouring villages; and (c) designating existing villages as “census towns”. None of these involve spatial movement of people and hence, do not alter the social composition of constituencies. Migration accounts for the remaining 20%, most of which is for marriage. This too may not upset the political calculus.
What constrains voluntary migration?
However, our understanding of migration behaviour is based mainly on economic factors. Theory suggests that if investment in non-agricultural activities grows faster than in agriculture (which is true for India for four decades), there will be voluntary movement of people in response to higher urban wages. This raises the second subsidiary puzzle: Why don’t more Indians migrate voluntarily in response to the growing divergence in economic opportunities between rural and urban areas?
One reason may be that non-agricultural opportunities are coming up in rural areas. This looks plausible as over 50% of rural incomes now originate from non-agricultural activities. On the other hand, rural-urban wage differentials don’t support this proposition. But it’s not obvious what urban wage rates should be used to assess the wage differential. Given the large informal sector in urban areas, should it be the formal wage, informal, or an average? Each will give a different result.
Rationality doesn’t only mean economic rationality. We also need to consider rural-urban differences in living conditions. Although most social indicators are significantly better in urban areas than rural, comparing averages completely misses the point. As with urban jobs, new migrants don’t have access to many urban facilities. Hence, it is important to know the characteristics of the right category of the urban population. Assuming that most first-time migrants initially live in slums, we may use socioeconomic characteristics of the slum population as a fair representation of living conditions of new migrants. Existing data point to the vast difference between living standards of slum-dwellers and urban averages. Although there are no comparisons with village living standards, mainly due to lack of an appropriate yardstick, this may be a better explanation of low migration rates than wage differentials.
This comparison too may understate the true extent of hardship of first-time migrants. Having worked in some detail on the characterisation and identification of Indian slums, I am acutely aware of the fact that practically every village meets the formal characterisation of a slum. Hence, many villages that are absorbed into urban agglomerates automatically become part of its slum component. But there is significant difference between ‘urban villages’ and migrant-populated slums. First, most residents of urban villages have some form of property rights while migrant slum-dwellers have none. Second, urban villages have established social networks, enabling access to opportunities, benefits, and informal insurance. Migrant slums have weaker social networks, though these may strengthen over time as more migrants join. Third, residents of urban villages are far more attuned to the ways of urban life than migrants.
Thus, any analysis of rural-urban differences in living conditions for determining migration behaviour should be based on migrant slums only. This is challenging given the lack of regular processes for identifying slums.
What explains aversion to identifying slums?
The final subsidiary puzzle is: Why is the Indian political and administrative system averse to identifying slums?
This is important since slums are a natural outcome of rural-urban migration, and the mainstay of urban growth. Slums start reducing only when average urban incomes become fairly high. Slums may be the epicentre of urban angst, crime, violence, etc., but shutting one’s eyes to their reality won’t make the problems go away, and may indeed make them worse since corrective measures can’t be undertaken.
Political resistance to migration doesn’t entirely explain this deliberate blindness. After all, slums are a consequence of migration that has already happened, and politicians would want to attract votes of this new addition to the electorate. An interesting twist is that in India we classify slums into notified, recognised, and unrecognised. Municipalities are assigned responsibilities, ranging from granting property rights in notified slums to providing some minimum amenities and resettlement in the recognised to utter neglect of the unrecognised. It would be interesting to know the composition of each of these categories - I suspect that most notified slums are urban villages where property rights already exist, and true migrant slums are unrecognised.
Despite the benefits of granting in situ property rights to migrant slum-dwellers, such as reduced insecurity and greater sense of belonging (De Soto 1989) - are politicians entirely wrong in not doing so? If we accept that urbanisation levels are too low in India and migration will continue for a long while yet, we do need to consider where the future stream of migrants will settle if all prime public land is already occupied.
This analysis assumes away policy barriers to migration in India. By constitution, all citizens are free to move to wherever they choose - but are they really free? A barrier may be the ubiquitous ration card - both a source of subsidy and means of identification - that is location-specific and not easily portable. Aadhaar was meant to overcome this, but its progressive linkage with practically all public services and social benefits and procedural complexity of getting any change made, especially location, may end up making it the biggest barrier of all.
This article is based on the 8th B.G. Kumar Memorial Lecture presented at the Centre for Development Studies, Thiruvananthapuram on 20 February 2017. The full version is available here: http://cds.edu/wp-content/uploads/2017/04/BGKumarLecture_ProfPronabSen.pdf. The lecture is based on research carried out at the IGC. The author gratefully acknowledges the valuable contribution of Vikas Dimble.