Land Inequality in India: Nature, History, and Markets

 

WORLD INEQUALITY LAB  |  WORKING PAPER 2026/09

Land Inequality in India:

Nature, History, and Markets

Nitin Kumar Bharti  ·  David Blakeslee  ·  Samreen Malik

April 2026  ·  University of Western Australia  /  NYU Abu Dhabi

 

 

Key Statistics at a Glance

71.1 Mean village Gini all households

46% Landless households of rural India

45.9 Landowner Gini among owners only

270,000 Villages in sample covering 650M people

 

 

Introduction

Land is the primary productive asset in agrarian economies — yet in much of rural India, a small number of households control a vast share of agricultural land while a large population remains landless or operates marginal plots. Understanding why this is the case, and which forces sustain it, is one of the central questions of development economics.

A new working paper from the World Inequality Lab by Nitin Kumar Bharti, David Blakeslee, and Samreen Malik offers the most comprehensive village-level analysis of Indian land inequality ever conducted. Drawing on data from 270,000 villages and 650 million individuals — believed to be the first such census-level dataset for a large developing country — the authors disentangle the contributions of three broad forces: agricultural suitability, historical institutions, and market integration.

The diversity of land inequality levels across Indian states is almost as large as the variation between countries at the world level.

Their findings are striking. Historical factors and agricultural geography are roughly equally important, each explaining around 38–39% of what the model can account for. Market access plays a supporting role at around 20%. Crucially, while economic modernisation can erode the inequality driven by geography, it appears to leave the inequalities rooted in colonial history and caste almost entirely intact.

The Scale of the Problem

Indian villages are extraordinarily unequal by any measure. The mean village-level Gini coefficient — where 0 is perfect equality and 100 is total concentration — stands at 71.1 when all households are included. Nearly half of all rural households (46%) own no agricultural land whatsoever. Among those who do own land, the average holding is 6.2 hectares, but the distribution is deeply skewed: the wealthiest household in a typical village controls 12.4% of all local agricultural land, and in 3.8% of villages a single landlord owns more than half the land.

The picture varies sharply by state. Kerala records the highest Gini at 90, followed closely by Punjab and Bihar. At the lower end, Karnataka and Rajasthan sit below 65. Landlessness is highest in Punjab (73%) — despite its advanced commercial agriculture — and lowest in Rajasthan (34%). Bihar and Punjab also have the highest share of villages dominated by a single large landlord.

 

State	Gini (all HHs)	Landlessness	Top 10% land share
Kerala	90.0	65.4%	65.2%
Punjab	83.5	72.8%	40.5%
Bihar	82.8	59.4%	56.1%
Tamil Nadu	80.9	66.5%	43.6%
West Bengal	77.7	55.5%	48.0%
Madhya Pradesh	71.5	50.7%	42.7%
Maharashtra	70.9	48.1%	40.8%
Uttar Pradesh	67.5	39.4%	46.0%
Karnataka	64.6	38.6%	41.2%
Rajasthan	62.1	34.0%	41.9%

 

 

Agricultural Suitability: More Productive Land Is More Unequal

One of the paper's most novel findings is the relationship between agricultural productivity and land inequality. Existing literature has focused almost entirely on plantation agriculture as a driver of inequality — think sugar and cotton in colonial Latin America. But Bharti et al. show that general agricultural suitability, independent of crop type, also strongly predicts inequality in India.

Using satellite-derived vegetation indices to measure agricultural output potential, the authors find a strong positive relationship between productivity and inequality up to the 60th percentile of agricultural suitability, after which the relationship flattens. This pattern holds equally in British-ruled and princely-state areas, ruling out the possibility that it is simply a proxy for historical institutions.

Areas within government irrigation schemes have a Gini coefficient roughly 1 percentage point higher than comparable villages just outside the boundary — confirmed by a rigorous border discontinuity design.

The mechanism appears to work through the expansion of large farms at the expense of small ones. As agricultural potential rises, the share of land held by households with more than 10 hectares increases sharply, while small and semi-medium farmers lose ground and more households become landless. The share held by medium-sized farms (4–10 hectares) shows almost no change — it is the very large holdings that benefit disproportionately.

The good news is that this form of inequality appears susceptible to economic modernisation. In areas where the non-agricultural sector has grown significantly, the influence of agricultural suitability on inequality is reduced by 50–100%. Proximity to towns also attenuates the relationship by roughly a third. This suggests that as rural economies diversify, the returns to agricultural land ownership become less dominant.

The Colonial Legacy: Zamindars and Princely States

Historical institutions cast a long shadow over Indian land ownership. The paper examines two key legacies of British rule: the distinction between areas directly governed by the British and those that remained under indigenous royal households (the 'princely states'); and, within British-ruled areas, the distinction between zamindari and ryotwari systems of land revenue collection.

The zamindari system

The zamindari system, introduced by the British to increase tax revenue, vested erstwhile revenue collectors with formal ownership rights over large swathes of agricultural land, effectively creating a landlord class. The consequences are still visible today. The Gini coefficient in zamindari areas is 3–5 percentage points higher than in ryotwari areas, where cultivators retained ownership rights. This finding is robust across multiple specifications and is confirmed by a border discontinuity design comparing villages on either side of the historical zamindari boundary in Uttar Pradesh and Bihar.

The elevated inequality in zamindari areas is driven by multiple mechanisms: a substantial reduction in the share of land held by smallholders, a marked increase in the share held by very large landlords, and greater landlessness. In roughly two thirds of cases, the increase in large holdings can be traced directly to the share controlled by the single wealthiest household in the village.

Princely states

Villages within the former princely states — territories governed by Indian princes under British supervision — show around 2–3 percentage points lower Gini than those in directly ruled British areas. This difference is driven primarily by lower rates of landlessness rather than more equal distribution among owners. Interestingly, princely state areas also show a slightly higher incidence of dominant landlords, suggesting that the lower overall inequality does not simply reflect an absence of large holdings but rather a different pattern of land concentration.

Caste: Exclusion, Not Just Concentration

Perhaps the most striking and robust finding in the paper concerns caste. Villages with a higher share of Scheduled Caste (SC) population — those historically subjected to the most severe forms of social and economic marginalisation — have substantially higher land inequality. A one standard deviation increase in SC population share is associated with a 3–5 percentage point increase in the all-household Gini, significant at the 1% level across virtually every specification the authors try.

The caste-inequality relationship operates through exclusion, not concentration. SC presence raises landlessness — it does not make the distribution among landowners more unequal.

This is a crucial distinction. When the authors examine the Gini coefficient calculated only among landowning households, the SC population share has essentially no effect — or even a slightly negative one. The inequality is not about SC households owning smaller plots than upper-caste households who also own land. It is about SC households being excluded from land ownership almost entirely.

The relationship is also non-linear. It rises steeply as the SC share increases from 5% to around 80% of the village population, then turns negative above 80%. In predominantly SC villages, the excluded group is so dominant that the between-group inequality resolves into within-group variation among a more homogeneous population.

The limits of land reform

Post-independence land reforms attempted to address these historical inequities, but the consensus in the literature is that they largely failed the lowest castes. One cited study found that tenancy reforms actually benefited middle-caste tenants while reducing land access for poorer lower-caste tenants. The migration of SC agricultural workers may mechanically inflate landlessness figures in destination villages, but the authors control for this and find the result unchanged.

Kerala and West Bengal: the exceptions

Two states stand as significant exceptions to this pattern. In Kerala and West Bengal, the relationship between SC population share and land inequality is substantially smaller and statistically insignificant. Both states were governed for extended periods by left-wing parties and are widely considered to have implemented the most successful post-independence land reforms in India. This comparison provides important positive evidence: the caste-land link can be broken by deliberate political intervention, even if it has rarely been.

 

State	SC effect on Gini (pp per 1SD)	Significant?	Note
Bihar	+4.1	Yes	Strong link
Maharashtra	+4.7	Yes	Strong link
Punjab	+5.2	Yes	Strongest link
Tamil Nadu	+4.3	Yes	Strong link
Rajasthan	+4.0	Yes	Strong link
Uttar Pradesh	+3.5	Yes	Moderate link
Karnataka	+3.2	Yes	Moderate link
Madhya Pradesh	+3.8	Yes	Moderate link
West Bengal	+0.9	No	Land reform success
Kerala	+0.4	No	Land reform success

 

 

Market Integration and Inequality

A third driver of inequality is proximity to markets. Villages closer to towns, major highways, and railway stations are systematically more unequal. The effect of town proximity is the largest and most persistent, with elevated inequality detectable up to 10 kilometres from a town. The effects of roads and railways attenuate more quickly, within 2.5 kilometres.

The paper also examines the role of banks and government-sanctioned agricultural markets (mandis). Villages that host these facilities have higher inequality, but there is no spillover to nearby villages — the effect is localised to the facility itself, suggesting that the mechanism is not simply general economic development but something more specific about the institutions.

For highways, there is an important historical dimension. The long-established Golden Quadrilateral highway network shows elevated inequality up to 25 kilometres from the road, rising by around 2 percentage points in its immediate vicinity. The more recently built North-South-East-West corridor shows a much weaker effect, barely detectable beyond 5 kilometres. This contrast suggests that it is the durability of market integration — the accumulation of advantages over time — that drives inequality rather than the mere presence of a road.

What Can Market Forces Actually Fix?

One of the paper's most policy-relevant findings is the asymmetry in how economic modernisation affects different drivers of inequality. The authors interact indicators of structural transformation and market access with agricultural suitability and historical variables to test whether economic development moderates either.

The results are clear. In areas where the non-agricultural workforce share is more than one standard deviation above the mean, the influence of agricultural suitability on inequality is reduced by 50 to 100%. Town proximity reduces it by roughly a third. The presence of an agricultural market also attenuates the agriculture-inequality link.

Structural transformation and market integration have essentially no effect on the historical drivers of inequality — the caste and colonial tenure effects persist unchanged in the most economically advanced villages.

None of these moderators, however, make any significant dent in the inequality driven by historical institutions or caste. The interaction coefficients for zamindari tenure, princely state status, and SC population share are all statistically indistinguishable from zero, regardless of how economically developed or market-integrated the village. This implies that the inequities rooted in history and social structure are qualitatively different from those rooted in geography — they are not susceptible to being eroded by market forces alone.

Inequality and Public Goods: A Nuanced Relationship

The paper also examines how land inequality affects the provision of public goods — schools, paved roads, health clinics, and sanitation campaigns. The conventional view, supported by a large literature, is that inequality depresses public goods provision by empowering elites who can privatise services and block redistribution.

The Indian evidence, however, is more nuanced. Consistent with earlier work by Banerjee and Somanathan (2007), the authors find a positive and concave relationship between inequality and public goods. Some degree of inequality appears to improve outcomes, perhaps because wealthier landowners are better positioned to lobby the government for village-level resources. Landlessness itself is also associated with more public goods, possibly reflecting government targeting of the most deprived populations.

But this positive relationship has limits. At very high levels of inequality, the beneficial effect disappears or turns slightly negative. And most strikingly, villages dominated by a single large landlord — defined as one individual owning more than 30% of the land — have systematically worse public goods outcomes, even after controlling for overall inequality levels. The negative effect is largest for government primary schools, which are 10 percentage points less likely to be present in a landlord-dominated village. The authors suggest that individual dominant landlords may behave differently from a diffuse landed class — they may be absentee, or may have less interest in the collective welfare of the village.

Conclusions and Policy Implications

This paper is a landmark contribution to our understanding of land inequality. Its scale — 270,000 villages, 650 million people — sets it apart from any previous analysis, and its methodological rigour, including multiple regression discontinuity designs, gives confidence in its causal claims.

Three conclusions stand out for policymakers and researchers.

First, agricultural productivity is an underappreciated driver of land inequality. Higher agricultural potential expands large holdings at the expense of small farmers, increasing landlessness. This process can be attenuated by structural transformation and market integration, suggesting that policies that support rural non-farm employment and market development can help.

Second, colonial institutions have left durable and measurable imprints on Indian land ownership. The zamindari system created a landlord class whose influence is still visible nearly eight decades after independence. Princely states fared better. These differences are not simply absorbed by economic development.

Third — and perhaps most importantly — caste remains the most persistent and intractable driver of land inequality. The mechanism is exclusion: SC households are disproportionately landless, not merely land-poor. Neither economic modernisation nor market integration reduces this effect. The exceptions — Kerala and West Bengal — show that deliberate redistributive policy can make a difference, but such policies have been the exception, not the rule. In the absence of renewed political will to redistribute land or its economic equivalent, caste-based land inequality is likely to persist for generations.

 

 

Source

Bharti, N.K., Blakeslee, D., & Malik, S. (2026). Land Inequality in India: Nature, History, and Markets. World Inequality Lab Working Paper 2026/09.