Why is labor mobility in India so low?

From Kaivan Munshi and Mark Rosenzweig, here is the puzzle:

It follows that the real wage gap [rural to urban] in India is at least 16 percentage points larger than it is in China and Indonesia. There is evidently some friction that prevents rural Indian workers from taking advantage of more remunerative job opportunities in the city.

Indian migration to the cities is much lower than for China or Indonesia.  Here is part of the answer:

The explanation that we propose for India’s low mobility is based on a combination of well-functioning rural insurance networks and the absence of formal insurance, which includes government safety nets and private credit.

…In rural India, informal insurance networks are organized along caste lines. The basic marriage rule in India (which recent genetic evidence indicates has been binding for nearly two thousand years) is that no individual is permitted to marry outside the sub-caste or jati (for expositional convenience, we use the term caste interchangeably with sub-caste). Frequent social interactions and close ties within the caste, which consists of thousands of households clustered in widely dispersed villages, support very connected and exceptionally extensive insurance networks.

Households with members who have migrated to the city will have reduced access to rural caste networks…

I believe “deficient infrastructure” and “lack of good manufacturing jobs” should be a bigger part of that answer, but nonetheless an interesting question and discussion.  Might the multiplicity of languages in India be a factor too?  Here is my earlier argument that Indian cities may be undercrowded.

For the pointer I thank the excellent Samir Varma.


Gregory Clark's book on social mobility using surname analysis, The Son Also Rises, revealed a low level of social mobility all around the world. The chief exception was India, where social mobility instead was extremely low.

"To Western observers, Indian cities appear impossibly overcrowded. But in income-adjusted terms, are Indian cities overcrowded at all? Might they be undercrowded and still capable of reaping additional economies of scale?" [SNIP]


May I suggest you visit Pune on a weekend late afternoon, via taxi http://www.livemint.com/rf/Image-621x414/LiveMint/Period1/2015/08/11/Photos/[email protected]

Please let us know how you found the experience.

I've seen the point made that Indian cities are particularly restrictive of urban development; could that potentially play a role as well? Are urban rents high relative to wages?

Family, caste, marriage. Language seems a likely factor too, as you suggest, but usually the nearest major city will speak the same/similar language as the countryside (right?). Infrastructure is clearly deficient for many purposes, but there is cheap bus and rail transport to get just about anywhere, so I don't see how this would be a particularly major barrier - if you're going to start a new life in the city, is it really that much of a larger barrier if the bus/train takes twice as many hours - we're talking hours here, not weeks and months difference like in olden times?

Also, potential migrants have probably heard too many stories about arriving in cities and getting fleeced in various manners. India has lots of mafias, and if you don't have connections in a city there's a rather high risk that you'll run into problems with one of the mafias, whether it be related to work or accommodation.

Language doesn't seem like it should be much of a factor. Even many poor Indians are somewhat multilingual. Hindi and English (or reasonable facsimiles thereof) are lingua francas.

I think here "language" is a stand-in for the fact that India is a whole civilisation with all the variety that implies. I would be interested in comparing the labour mobility in India with that of the EU.

I've never been to India so the next questions are genuine, not rhetorical: Is Hindi widely spoken in the South? And are there any states at all where a majority speak English well enough to do more than buy cups of tea etc.

Hindi is not widely spoken in the south. In my experience, exceedingly few people outside of the highly educated (a small minority) and people in the tourism industry can speak English for much of any practical purposes. I've spent days at a time in India where landmark names were the only way to get any sort of directions whatsoever (most often completely wrong directions at that), and where I did not meet anyone who spoke more than just a handful of words in English. Maybe the experience is very different if you also speak Hindi.

India seems to lack a true address system, so landmarks are the only way to give and get directions. I lived in Singapore where every high rise had it's own zip code. Here in Bangalore there are maybe 3 zip codes for the whole city.


TO India??


A lot of outside observers accept it as an inevitability that India will follow in the footsteps of China's economic success. Why shouldn't it be the next Asian tiger? But this view really glosses over just how backwards Hindu culture is in many ways. It's easily on par with Islamic culture in terms of serious social and ideological pathologies. The latter's flaws are simply more visible. While Islam is the most evangelical of the major global religions, Hinduism is the most insular (even more so than Judaism, where at least some process exists to allow conversion for someone not born into the faith). Hindustan's warts tend not to be on display to the broader world, because its adherents have no reason to share. But they are ugly and numerous.

The East Asian economic miracle was made possible because the institutions of Western capitalism and the rule of law were easily bootstrapped on top of Confucian-Buddhist cultures. Outside Europe and East Asia, very few nations have managed to move out of the middle-income trap (excluding countries that stumbled upon stupid amounts of natural resources). Hindu culture offers no compelling case for why India should be an exception to this general pattern.

+ 1

India will never achieve what the Han Chinese did over the last three decades-in part, by recognising that the rising tide must lift all boats [other Asian populations in the CHI sphere]

Different parts of India do rather differently, even though they are all Hindu. East Asian "tigers" have a long history of political unity and working states. India as a whole has rarely had that, though regions (often corresponding to states) within the country have.

The political culture here is also part of the problem. Taxes are high just to support the political class.

Culture does matter but these sorts of explanations can easily devolve into elaborate question-begging. Islam, for all its "pathologies," hasn't stopped Turkey, Malaysia and others from pulling far ahead of India. Max Weber originally proposed the idea of the Protestant Work Ethic as an explanation for why Protestant countries were wealthy and others were not. Now we seem to have the Protestant-Jewish-Catholic-Confucian-Buddhist-moderate-Muslim work ethic. Some humility is in order for those claiming to know exactly how culture impacts economic growth.

For that matter Hinduism hasn't stopped e.g. Gujarat pulling ahead of Bengal. And however badly they may creak along India has managed to keep Democratic and Free Market institutions alive despite the tide of postcolonialist fervor that did nothing except create failed states throughout the Third World. For example Indira Gandhis attempt at dictatorship was resisted by all society not just westernized elites. Dougs argument is dumb.

IMO HBD offers a more plausible explanation than culture (except insofar as culture is affected by genetic factors.) It may just be that compared to other countries India has larger numbers of people who are simply incapable of taking advantage of newly available opportunities. But in that case progress would require more casteism not less. Education, institutions etc. would help too but who would provide that education? Who would man those institutions?

HBD is not a substitute to cultural backwardness. The level of intelligence in a region isn't randomly exogenous, but a by-product of the history of its social institutions. Numerous anthropologists have shown how Christianity was an essential component of the gene-culture co-evolution that led to high European IQs.

Much of Gujarat's success can be contributed to the high concentration of Parsis and Jains. The fact that India's economy is largely dominated by religious minorities is evidence of the economic handicaps that accompany Hinduism. You do bring up a good point. Hindustan has demonstrated much less of a political proclivity towards autocracy and dictatorship than Islamic, Sub-Saharran, Latin and even East Asian countries. It's of course tough to tell how fundamental this relationship is, because we only have one nation's history to look at. That being said the stability of democracy in India, especially given its low levels of income and education, is remarkable. It does seem like Hindu culture is very good at democracy, but very bad at capitalism. That seems like a fascinating topic for exploration.

Parsis and Jains don't help your thesis. They are if anything even more insular than typical Hindu castes. Well, Jain Vaishyas will marry Vaishnava Vaishyas providing they are of the same caste. Whatever the theological differences, socioculturally they are completely mainstream. Parsis on the other hand are on the brink of extinction due to their inbreeding but that hasn't stopped those that remain from excelling in capitalistic endeavours.

And what about the ubiquitous Patels, a 100% Hindu Gujarati caste? Of particular relevance to the question of social mobility, the Patels seem to have been rather low SES until the late 18th/early 19th centuries or so. (Patel comes from the Mughal/Mahratta adminstrative term Patidar which means a farmer who held the deed to his own land as opposed to e.g. a sharecropper.) There does not seem to be anything particular in their origins to suggest that they had more aptitude for capitalistic commerce than their neighbours yet today they are found across the globe.

Obviously some Hindu cultures are quite capable of participating in the modern world. Others are not. But isn't that true everywhere including Christendom?

Doesn't Gujarat have some rather special diaspora dynamics dating back some hundreds of years? However, in which case, my understanding is that caste and family ties may in fact contribute positively to this.

That's a fair point. The problem with assessing something like how does culture X affect economic development is that you really only have one data point. Maybe a few at most if you're clever about dividing regions. It's not like we can re-run history. So I'll fully concede that these types of assertions have to be made with conservative p-values.

That being said, I don't think your counterpoints really prove anything. Malaysia was successful almost certainly because of its Chinese minority. Ethnic Chinese make up 80% of Malaysia's CEOs and tycoons. It'd be very difficult to convince me that a Muslim-only Malaysia could achieve anything like its post-independence economic miracle. As for Turkey, its post-war history can be succinctly summed up as "continual military suppression of any religious culture." Its relative success doesn't exactly prove that Islam is conducive to advanced economic development.

I think Chinese culture is relevant, but it could easily be overstated for the simple fact that they share language with another very large economy, much larger than Indonesia (essentially similar language and culture to Malays), for example.

I personally think the problem with Islam at present is the fact of essentially theocratic rule in many Muslim majority countries, rather more so than the religious practice at the individual level. For example, education revolves around the teaching of Qur'anic law, which is really good for regime stability, but disastrous for anything linked to economic innovations. (An article linked to on this blog a couple/few months ago has some discussion of historical evidence of this, suggesting that technological advance was quite good under less theocratic conditions, but then eventually the theocracy basically took over - in analyzing the subject matter of Islamic libraries through history, the author was able to present some reasonably convincing evidence of this. I forget the title, so I'm not sure how to find the link to it.)

Related to the network is the (relative) sense of well-being: someone who is poor doesn't feel poor if she has more than her poor neighbors, which one recent study found is why many in very poor countries are content with their lives. In the city, indications of one's inferior status are everywhere. Which relates to my theory about Trump: vicarious experience. In poor black (mostly rural) communities, the minister often displays signs of relative affluence, in the clothes he wears, the car he drives, the house he resides in. Why doesn't his poor flock resent it? Because they live vicariously through the minister. And so do Trump's supporters.

Of course, the Trump/minister analogy extends beyond the vicarious experience. Like the minister, Trump preaches salvation, to be found in his (or is it His) books and university and message: the kingdom of Trump is there for those who are willing to reach out for it. Trump's message and the prosperity gospel overlap, which explains why so many evangelicals are drawn to him. Like Jesus, Trump is a rabble-rouser, conveying to his flock a message of salvation for those willing to receive it.

This still sounds like a good libertarian system must have strong puritanical religion enforce society values. Without religion, libertarian society can not enforce its norms on poor people.

The impact of social insurance networks goes beyond the financial safety net. It also serves as a social coccoon tethering the person to his village and community. In India we have a saying that translates to, "Every 10 miles, the language changes, every 100 miles the complexion". The outside can be very different and challenging to someone who can instead be comfortable in his shell.

I'm Indian and I've never heard that saying. What language is it in?

Bhojpuri/Hindi I think: "Ek kos mein badle paani, das kos mein vaani"

Ankur, that does not sound like the correct translation.

My understanding is that India is a very feudal country, with extremely high rents in most urbanized areas. A higher wage doesn't mean much if most of the higher wage is used to pay housing rent. Back at the village my (extended) family likely owns land that may significantly reduce living costs, as well as provide other economic and social support mentioned in the article.

That's not feudalism at work, that's just about everywhere in the world. Cities almost invariably have rents that are very high compared to median wages in the absence of public or subsidized housing. People cope with this problem by living in low-rent districts (including illegal slum developments as in many Indian cities), living with relatives or finding roommates or a shared housing arrangement of some sort. The point about India seems to be that this sacrifice doesn't appeal to as many rural Indians as it would in the absence of very strong social networks and safety nets at home.

@jvm, robert,

Your answers are the closest to reality. Urban rents in india are high. Urban housing prices are positively obscene. That explains most of the difference. Politicians and bureaucrats prefer to store their black money in urban real estate. This combined with rent control in many of the older cities leads to a great housing shortage.
I'm not denying the results, but would be interested in the results of a study that looked at disposable income after rent.

Munshi highlights one variable which has policy implications but, of course, understands very well that flawed Economic policies from the Sixties onward were the main culprit.
In China, big employers set up dormitories and thus could easily scale up. In India, property rights in jobs, security of tenure for worker-tenants in 'chawls', reservation of certain labor intensive industries for the small/medium sector (e.g. power-looms being favored over big Mills) and, to a lesser extent, politically instrumentalised labour militancy- all militated against rational relocation of Labor. This affected the 'pull' factors.
On the 'push' side, local politicians had an incentive to tie poor people to the land. Initially this was done through 'Bhoodan' (voluntary land redistribution) and improved tenancy rights- there was also some land redistribution favoring richer peasants. The problem was that no fungible assets were created more especially because of restrictions on sale to non-agriculturist castes. Thus, most people could not sell up in the village and use the money to set up elsewhere.
The experience of India and China shows that illiterate villagers are willing to move to completely different language zones or, indeed, continents and, within a generation or two, they can be doing very well though retaining some traditional customs and endogamous traditions.
In the case of India, land is hugely overvalued because of a heritage of stupid economic policies. Furthermore, it is unlikely that Labor and Tenancy law will be adequately reformed. Thus we can't get back to the situation in the Fifties when Industrialists were willing to set up Townships or at least provide 'chawl' dormitories.
Prof Munshi's solution- viz. proper universal insurance as opposed to informal sub-caste based support- does have some problems, most notably corruption and administrative paralysis where suspicion of corruption exists. On the other hand, it may be that India will soon begin to see a big change in this area because farmers are more and more pessimistic about the future of their industry as well as of the possibility that one brother will be able to earn a surplus in the City to help the other brother back home. Thus, somewhat belatedly, even 'son preference' is falling and psychologically demographic change has or is occurring for those with even a small amount of capital or property.

Comments for this post are closed