ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846

A+| A| A-

Income and Inequality across Rural–Urban, Occupational, and Caste Divides

The evolution of income in India from 2014–19 is analysed, and it is found that the lower end of the income distribution has experienced significant losses—the bottom ventile shows not only a decline in income share of ~41% but also a negative real average income growth of -5.5% per annum. Further investigating the composition of this part of the distribution using rural and urban splits, it is found that even as income shares at the bottom of the urban distribution have increased over time, those at the bottom of the rural distribution have decreased—income share of the bottom decile of the rural income distribution declined by ~43%, and the real average income growth of this decile was -5% per annum. The bottom ventile of the consolidated Indian income distribution is composed primarily of rural incomes and therefore the decline in real incomes is essentially a rural phenomenon.

The past 30 years have seen the inexorable rise of income inequality in India (Chancel and Piketty 2019). This has been caused through a combination of global technological and economic changes as well as structural conditions in the Indian economy since the 1980s (Banerjee and Piketty 2005; Chancel and Piketty 2019; Deaton and Dréze 2002; Kohli 2012; Dev and Ravi 2007; Milanovic 2016).

Immediately after independence in 1947, the state took ownership of the “commanding heights” of the economy and ensured a progressive taxation structure with high marginal rates for top incomes, with the explicit goal of curbing elite economic power and driving income convergence; indeed by the early 1970s, the top effective marginal tax rate had risen to 97.5% (Banerjee and Piketty 2005; Chancel and Piketty 2019; Acharya 2005). This period, until 1980, saw a sustained decline in income inequality, with the share of the top 10% of income earners reducing from 37% in 1951 to 31% in 1981 and that of the bottom 50% rising from 21% to 24% in the same period (Chancel and Piketty 2019). This decline in inequality should, however, be contextualised by the fact that the poverty rate in India remained practically unchanged in this time—from 56% in 1954 to 53% in 1978 (Dutt and Ravallion 2009). Essentially, low economic growth (3.4% per annum between 1951 and 1980) and high population growth (98% between 1951 and 1981) meant that even with the reallocation of income from the rich to the poor within the distribution, the poverty rate remained persistent at over 50% and the poverty head count doubled (Nagaraj 1990; Census of India 2011; Dutt and Ravallion 2009).

However, since 1980, there has been a progressive dismantling of the socialist architecture of the Indian economy, with enhanced private participation, deregulation of prices, and reduction in tax rates (the top marginal tax rate had declined to 30% in 1998), though still retaining a progressive taxation structure that has resulted both in increased economic growth and rising income inequality (Banerjee and Piketty 2005; Kohli 2012; Chancel and Piketty 2019; Rodrik and Subramanian 2004; Basole 2014). The share of the top 10% of income earners has sharply increased to 56% and that of the bottom 50% has declined to 15% as of 2015 (Chancel and Piketty 2019). However, increased economic growth in the period from 1980 to 2015 (average annual growth rate of 6.05%) has also resulted in a significant decline in the poverty rate to 21% in 2006 (Deaton and Dréze 2002; Panagariya and More 2014; Dhongde 2007; Panagariya and Mukim 2014; Dutt and Ravallion 2009; Bhagwati and Panagariya 2013).

It has been argued that although worsening inequality could be a consequence of economic growth (which yields poverty reduction), the benefits of this growth are spread across the distribution in India, leaving individuals, on average, better off than before (Bhagwati and Panagariya 2013). This interpretation is consistent with an income distribution where worsening income shares for those lower in the distribution occur on account of differential income growths at different points in the distribution, essentially income growth is higher, on average, higher in the distribution, meaning that the lower end of such an income distribution would see lesser than average growth over time and account for a progressively reducing share of the total income. In such a distribution, despite the rich benefiting more than the poor, everybody is better off than before. Recent work has revealed that while this was potentially the mechanism underlying the increasing income inequality from 1980 to 2000, it is likely that the nature of redistribution has entered a fundamentally new regime since the early 2000s where income growth at the bottom of the distribution has not only (on average) been lower but also negative, while incomes higher in the distribution experienced positive growth. This essentially implies that real incomes at the bottom of the distribution have actually been declining and that the income distribution is essentially diverging (Sahasranaman and Jensen 2021).

In this work, we study the dynamics of income distribution and inequality in India from 2014 to 2019, with a particular focus on the bottom of the distribution. Using panel data from the Consumer Pyramids Household Survey (CPHS) to construct the Indian income distribution, we explore the evolution of income inequality for the country as a whole, as well as for the rural and urban India separately, to generate an understanding of distributional features and inequality. We also empirically examine inequality through the lenses of occupation and caste to identify the typology of India’s most economically vulnerable populations. Finally, in order to explore the dynamics operating at the very bottom of the distribution, we use a stochastic model to quantify: (i) the extent and direction of redistribution occurring in the distribution, and (ii) the likelihood of incomes at the bottom transitioning higher in the distribution over time.

Data and Methods

We use the data from the CPHS published by the Centre for Monitoring Indian Economy. The CPHS is a pan-India panel household survey of roughly 1,70,000 households collecting monthly data on income, consumption, demographics, assets, and borrowing by households. The CPHS data set creates a geographically representative data set by sampling one or more homogeneous regions for each state from a set of neighbouring districts that have a similar agroclimatic condition, urbanisation levels, female literacy, and family size as per the 2011 Census. The CPHS visits each household in the panel thrice a year (each visit is known as a “wave”), and all household level is captured at monthly frequency. Using this data, we compute monthly per capita income by adjusting the total household income reported for each month, with the size of the households using a square root equivalence scale (Deaton 2003). We use household income as the basis to construct the Indian income distribution by adjusting each income by the appropriate weighting factor (provided by the CPHS) to ensure appropriate representation of all household types in the income distribution and then cumulate these adjusted incomes over each percentile to construct the income distribution. Annual income distributions are obtained by adding the corresponding percentiles in the 12 monthly income distributions.

An important consideration to keep in mind while analysing data from the CPHS is that the incomes at the top of the distribution are likely to be underestimated; this is a more general concern with income surveys where the highest incomes are unlikely to participate. More specifically in this context, we have the Indian income distribution from Chancel and Piketty (2019) who use tax data—a more reliable indicator of top incomes—to find that the income share of the top decile in 2015 was 56%, as against the 32.8% as per the CPHS data. Therefore, it is possible that the magnitude of income shares of the bottom half is overestimated in this analysis, which potentially makes the situation of the bottom decile of the distribution even more precarious than our findings indicate.

In past work, the income share of the top 10% of income earners has been used as a metric for inequality (Chancel and Piketty 2019), but given the concern expressed about underestimation of this component, we propose to use the ratio of income share of the bottom 50% (S50%) to that of the bottom decile (S10%) : S50% /S10%. This metric focuses attention on the lowest part of the distribution and, more specifically, provides us a sense of the relative income earned by the bottom decile of the income distribution. A perfectly equal distribution would yield a ratio of 5.

Additionally, we also construct ventiles and deciles of the income distribution to explore the extent of income earned in different parts of the distribution. For each ventile/decile, we compute the average real income for each year from 2014 to 2019, adjusting for annual inflation using data from the Reserve Bank of India’s (RBI) database on Indian economy (RBI 2021). We compute the compounded average growth rate over the period 2014–19 to quantify the magnitude of real income growth at all points in the distribution and construct Growth Incidence Curves (GICs) for India (overall) and also for rural and urban India. We also construct the occupational and caste composition for each ventile to enable an examination of these characteristics across the entire distribution.

Finally, in order to explore the nature of redistribution occurring in India for the period 2014–19, we use a simple stochastic model of geometric Brownian motion with reallocation (RGBM) to model income dynamics (Berman et al 2017). There is a body of empirical work which indicates that real world income, expenditure, and wealth distributions are reasonably approximated by lognormal distributions across many national contexts (Chatterjee et al 2016; Ghosh et al 2011; Banerjee et al 2006; Drăgulescu and Yakovenko 2001; Souma 2001). The RGBM models income growth as a multiplicative process described by the geometric Brownian motion, yielding a widening lognormal distribution over time. However, given the context of real economies where a number of mechanisms for redistribution are in place (such as taxes, transfers, and public spending), RGBM also incorporates a reallocation parameter (τ) to capture the extent and direction of transfer occurring within the income distribution. Income dynamics in RGBM are described using the following stochastic differential equation:

dxi = xi (µdt+σdWi) – τ(xi– áxñN), … (1)

where dxi is the change in income of i over time dtµ is the drift and σ the volatility of income, dWi is a Wiener process increment with mean 0 and variance dtτ is the reallocation parameter, and áxñN is the mean income: áxñN = ∑i=1xi. The first term of equation 1 is the income growth term encompassing growth due to both systemic (µdt) and idiosyncratic (σdWi ) components, and the second is the reallocation term where the reallocation parameter (τ) is applied to the net difference between individual i’s income and the average income of the society. If τ >0, it is indicative of progressive redistribution where resources are being reallocated from the top to the bottom of the distribution, which is the reality we would expect in most modern societies; and if τ <0, over a period of time, the income distribution is divergent and redistribution occurs from the bottom to the top of the distribution, indicative of a perverse state of economic inequity. τ is most appropriately understood as a cumulative measure of the overall redistribution occurring in an economy, implicit in the nation’s resultant income distribution.

In order to derive µ and σ for the Indian income distribution, we use the previous work of Sahasranaman and Jensen (2021) as the basis for obtaining the values of parameters µ and σ for the Indian income distribution. In that work, using time series data on the Indian income distribution (Chancel and Piketty 2019) and time series of wholesale prices for staple Indian crops and commodities (rich, wheat, and jaggery) as well as short-term time series of wages, it was estimated that µ = 0.0231 and σ = 0.15 for India’s income distribution. Given µ and σ, the RGBM algorithm is executed by propagating equation 1 over a set of N incomes over T = 5 time periods (corresponding to the period 2014–19) such that at each time period, the reallocation parameter τ(t) is obtained by minimising the distance between the income share of the bottom half of the simulated income distribution and that of the empirically observed Indian income distri­bution (S50%(t)). In summary, τ(t) is chosen to minimise at each time period t, resulting in a time series τ(t) that describes the temporal evolution of both the extent and direction of reallocation apparent in the income distribution. A detailed exposition on the RGBM algorithm is available in Berman et al (2017) and its application to the Indian income distribution for 1951–2015 in Sahasranaman and Jensen (2021).

Once the complete time series of income distributions is constructed using RGBM, we can study the transitions of incomes across the first decile so as to assess the persistence of low incomes over time. Specifically, we seek to understand how much a past in the lowest part of the distribution is a predictor of a present in this part of the distribution. Using methodology developed in Sahasranaman and Jensen (2021), we define the persistence probability pstic(t,tD1 ) as the probability of having spent at least tD1 consecutive time periods in the bottom decile at time t–1, given that the individual is in decile 1 at time t (equation 2):

 … (2)

where N1(t, tD1 ) is the number of individuals who have been in the first income decile for at least tD1 years at time t–1 and are still in the first decile at t; and N1(t) is the population of the first decile at t. This metric helps us understand how income dynamics impacts those who are deepest in poverty, and the extent to which escape from the lowest parts of the distribution is feasible.

Income Distribution and Inequality

We find that income inequality measured as S50%  /S10% increased from 13.4 in 2014 to 18.4 in 2019, but the rise is non-monotonic (Figure 1a, p 25). There is a rise and a dip in 2015 and 2016, respectively, and this is followed by a consecutive rise till 2019, with a particularly sharp increase between 2018 and 2019. Overall, this means that the relative income of the poorest decile has decreased between 2014 and 2019, implying that there is relative compensatory increase in other parts of the distribution.

In order to develop a more nuanced understanding of this phenomenon, we explore the evolution of incomes of all deciles of population and find that income shares of top 2 deciles declined from 51% to 49%, while the income share of each decile from the 3rd to the 9th decile showed an increase (Figure 1b, p 25). Given this generally progressive trend, it is the bottom decile that emerges as a concern because it has lost income share in this period from 1.6% to 1.2%, a decline of ~25% in this five-year period. This decline becomes even more pronounced the deeper we go into the income distribution, with the income share of the bottom 5% (first ventile) declining ~41%, from 0.27% in 2014 to 0.16% in 2019.

However, losses in income share over time are not necessarily representative of declines in real income levels; it is possible that certain parts of the income distribution gained share at the expense of others but that all parts of the distributions experienced absolute increases in real income.

To test this, we construct the GIC of average real incomes by ventile for 2014–19 and find that while the top 95% of the distribution saw positive compounded annual real income growth from 2014–19, the bottom 5% had a decline in real income of -5.5% per annum in this time period (Figure 1c, p 25). Our concerns about declining income shares at the bottom of the income distribution are exacerbated by this finding that a significant proportion of the lowest incomes experienced negative real income growth. The second ventile also shows an average real income growth of only 0.5%, suggesting that a portion of incomes in this ventile could also be subject to real income declines—indeed, when we analyse the percentiles comprising the second ventile (6th–10th percentiles), we find that both the 6th and 7th percentiles demonstrate real income declines as well at -7.3% and -3%, respectively.

We now attempt to dig deeper into the composition of the bottom of the distribution by splitting the Indian income distribution into its rural and urban income components.

Rural and Urban Dimensions of Income Inequality

As with the consolidated income distribution, we find that the k-index values for both the rural and urban distributions are between 0.64–0.65 for each year from 2014 to 2019, suggesting that the inequality in the distribution of income remained consistent in both distributions for this period. However, when we assess the change in income shares across parts of the distribution over time, we find that in the urban income distribution, the share of the bottom half increased from 22.8% to 24.4%, while the top half saw a decline in income share from 77.2% to 75.6% (Figure 2b). Even the bottom decline saw an increase in income share by ~7%, from 2.5% to 2.7%. In the rural distribution, on the contrary, while we saw a decline in income share for the top decile (from 34.1% to 33.7%), the middle part of the distribution—from the 3rd to the 8th decile—experienced an increase in income shares, and the bottom two deciles experienced significant declines (Figure 2a). Overall, the bottom half of the rural income distribution saw a marginal decline in income share, from 22.7% to 22.1%, but within this, the income share of the bottom decile declined sharply by ~43%, from 1.3% to 0.76%. Given the low levels of income and low income share in this part of the distribution, this represents a drastic reduction. In view of these distinct dynamics in urban and rural distributions, the decline in income share of the bottom decile in the consolidated Indian income distribution appears to be driven primarily by the declines apparent in the rural distribution.

We find evidence of this when we compute the average annual income per percentile for both the rural and urban distributions—the average annual income of the first percentile of the consolidated Indian income distribution corresponds to second percentile of the rural distribution and is much lower than the first percentile of the urban. Indeed, the first ventile of the consolidated Indian income distribution, which we observe to exhibit negative income growth (Figure 1c), is composed of the bottom 7 rural income percentiles and only part of the second urban income percentile. This confirms that negative income growth at the bottom of the Indian income distribution is essentially a rural phenomenon. The first decile of the consolidated distribution corresponds to the bottom 14 percentiles of the rural population and only the bottom 3 percentiles of the urban population. Indeed, average annual urban incomes are superior to rural incomes at all points in the distribution as evinced by the ratio of average annual income of each percentile in the rural distribution to the corresponding urban percentile (Figure 2c, p 26). This also cautions us against drawing too many inferences from relative performance of equivalent portions in the rural and urban distributions, given these vast discrepancies in income levels at corresponding points in the distribution.

We also examine the growth rates of real average income for each decile in the urban income distribution and find that income growth rate is positive (and progressively declining) across the deciles, pointing to a mildly convergent trend in the urban incomes and consistent with an increase in income share for the bottom half of the distribution (Figure 2d, p 26). However, while the top 9 deciles of the rural income distribution exhibit positive real growth, the bottom decile shows real income decline to the tune of -5% per annum (Figure 2d). This result is in agreement with the negative real growth exhibited by the bottom ventile of the consolidated Indian income distribution (Figure 1c), as the bottom ventile of the Indian distribution is comprised largely of incomes in the bottom 7 percentiles of the rural income distribution (Figure 2c).

Occupation and Caste at the Bottom

We look into the data to assess the profiles of households at the bottom of the Indian income distribution and find that the bottom decile is comprised primarily of small and marginal farmers, organised farmers, and wage labourers; on average, 80% of the bottom decile is composed of households with these primary occupations from 2014–19 (Figure 3a). Among these occupations, small and marginal farmers comprise the single largest occupational segment (~44%), followed by organised farmers (~23%), and then wage labourers (~13%) (Figure 3a). These occupations, which have experienced declines in real income, therefore currently comprise the most economically vulnerable workers in the Indian income distribution. These findings are consistent with previous studies showing that incomes from casual and wage labour have an inequality reducing effect on the income distribution (Ranganathan et al 2016; Shariff and Azam 2009). Studying rural income inequality between 1993 and 2005, it was found that income from casual labour represented a source of decreasing inequality, meaning that a rise in labour incomes acted as a countervailing force to inequality because it represents the income of those at the bottom of the distribution (Shariff and Azam 2009). Farm income and salaries, corresponding to higher parts of the rural income distribution, were found to be inequality enhancing sources of income. Analysis of the India Human Development Survey data also found inequality decreasing effects of income from casual labour and remittances in 2011–12 (Ranganathan et al 2016). The vulnerability of these occupations, especially of small and marginal farmers, has also been evinced in the growing incidence and extent of farmer indebtedness over time as well as the tide of farmer suicides since the 1990s (Vakulabharanam and Motiram 2011; Suri 2006; Narayanamoorthy and Kalamkar 2005; Vaidyanathan 2006; Kennedy and King 2014; Nagaraj et al 2014). Overall, these outcomes from the 1990s and 2000s combined with our findings for 2014–19 suggest that small and marginal farmers and wage labourers continue to comprise the most economically vulnerable populations in India, with their reducing share in the income distribution compounded by declining real incomes over time.

Caste provides yet another salient stratification of the income distribution. Prior work on a small rural sample of 8 Indian villages revealed that Scheduled Caste (SC) households were substantially worse off that other communities, being over-represented in the bottom income quintile and under-represented in the top quintile (Rawal and Swaminathan 2011). More worryingly, the study indicated that more prosperous villages were characterised by greater income inequality and caste segregation. Exploring the impact of caste on farm income per unit of cultivated land, it was found that SC- and Scheduled Tribe (ST)-owned land yielded significantly lower farm returns compared to other castes and that this difference was substantially explained by caste-based inequality (Singh 2011). The SC households are also more likely to be in the lowest consumption quintile than households of other castes (Borooah et al 2014). Given this context, it is not surprising that the SC and ST populations progressively decline as we go higher in the income distribution (Figure 3b, p 27)—for 2014–19, the SCs comprise 30% of the bottom income ventile and only 8% of the top ventile, while the STs comprise 16% of the bottom ventile and only 2% of the top ventile. In contrast, we see that the Upper Caste (UC) population fractions increase higher in the distribution from 17% of the bottom ventile; the UC population increases to 44% of the highest income ventile. The Other Backward Castes (OBCs) appear to be more evenly distributed across the income distribution, comprising 32%–40% of each ventile, except for the top ventile where they account for only 24% of the population.

This finding highlights the fact that the most oppressed castes (SC and ST) occupy the most vulnerable occupations (small farmers and labourers), yielding not only the lowest incomes in the distribution but also declines in real incomes for the period 2014–19. Of deeper concern is the fact that the trend of declining real incomes in 2014–19 is likely part of a longer-term trend of income erosion experienced by these vulnerable castes and occupations since the early 2000s (Sahasranaman and Jensen 2021).

Dynamics at the Bottom of Income Distribution

The low, and declining, income shares of the bottom decile and the bottom ventile as well as the decline in real income growth in this part of the distribution raise real concerns about the nature of reallocation occurring within the income distribution. Previous econometric modelling work indicates the possibility that the bottom of the income distribution in India has been witnessing negative growth since the early 2000s and that the overall reallocation within the distribution has turned regressive—meaning that instead of the expected progressive redistribution from rich to poor in a modern regulated market economy, there is a regressive transfer of resources from the bottom of the distribution to the top (Sahasranaman and Jensen 2021).

Using the RGBM, we model the evolution of the Indian income distribution from 2014–19 and find reasonable concurrence between the modelled results and empirical findings. Figure 4a describes the modelled incomes for the bottom five deciles (dotted lines) and the corresponding empirical observations (solid lines). Incidentally, the model is able to reasonably simulate the evolution of even the top decile of the distribution and given that the GBM produces a lognormal distribution, this concurrence supports the concern we had highlighted earlier that the CPHS data is not capturing the power law tail of the Indian income distribution. Given this correspondence between model and observation, we turn our attention to the time series of τ(t) that describes the time evolution of reallocation within the income distribution (Figure 4b). τ is declining from 2015 to 2018, indicating that the extent of
reallocation within the distribution is reducing over time and redistribution is becoming less progressive. Further, for the consolidated (and rural) income distribution, τ turns negative in 2018 (2017 and 2018 for rural), highlighting the risk that continued negative reallocation in the future could result in persistent divergence in the income distribution, yielding a perverse reallocation of resources from the bottom of the distribution to the top. Indeed, the empirical observation of negative real income growth at the bottom of the consolidated and rural income distributions (Figures 1c, 2d) is already in concurrence with the emergence of negative τ in these distributions (Figure 4b). While τ does emerge positive in 2019 across all distributions, tracking the trend in τ over a longer period of time will be required to ascertain whether the negative τ regime is short-lived or not. Longer term perpetuation of negative τ progressively increases the probability that income growth lower in the distribution turns persistently and deeply negative, calling into question extant policies of economic growth and redistribution in India.

Given this temporal pathway of τ, we now turn to understanding how the dynamics impact those at the very bottom of the distribution, in terms of the difficulty involves in escaping from the bottom decile to higher in the income distribution. Using equation 2, we study the persistence probabilities for tD1 = 1 and tD1 = 5 years and find that incomes in the bottom decile display a high level of stickiness to that part of the distribution, with pstic (t,1)=0.84 and pstic (t,5)=0.53 (Figure 5). This means that a past in deep poverty remains an effective predictor of future poverty with an 84% chance that an individual who is in the bottom decile currently has been in the bottom decile for at least the past year, and a 53% chance that an individual who is currently in the bottom decile has been in the bottom decile for at least the past five years. We do not see much of a difference when we compute these probabilities across rural and urban income distributions, though the bottom incomes in the urban distribution exhibit marginally lower stickiness than rural incomes (Figure 5). Effectively, we see that for individuals low in the income distribution, escapping seems an unlikely prospect. Our empirical analysis reveals that these are the incomes also declining over time, which makes their economic existence truly perilous.

Previous work has demonstrated that India was likely in a prolonged, decade-long period of negative reallocation beginning in 2002, with a significant proportion of the population at the bottom of the income distribution experiencing negative income growth (Sahasranaman and Jensen 2021). The growing informalisation of the formal workforce in manufacturing and services as well as the rising agrarian distress have meant that employment is increasingly characterised by greater insecurity and uncertainty (Vakulabharanam and Motiram 2011; Vaidyanathan 2006; Suri 2006; Mehrotra 2019). Our work here also uncovers the caste and occupational dimensions of those at the bottom of the income distribution, highlighting the concern that the SC and ST populations engaged in smallholder farming and wage labour have experienced a process of deepening povertisation over the last two decades. This process combined with dynamics, which reveal that the incomes of the poorest in the income distribution also exhibit very high likelihoods of remaining in the bottom decile over time, are the cause for alarm going into the future as well. Providing income support for the bottom deciles in the Indian income distribution is therefore a clear and present policy concern.


We study the income distribution for India from 2014 to 2019 and find that income shares at the very bottom of the distribution decline substantially. Income shares of the bottom decile and ventile decline by 25% (from 1.6% to 1.2%) and 41% (from 0.27% to 0.16%), respectively. Using GICs for 2014–19, we find that not only do the income shares at the bottom of the distribution decline but also that real income growth is negative—the bottom ventile has an annual growth rate of -5.5% and the subsequent percentiles (6th and 7th) also exhibit negative growth.

In order to understand the composition of the bottom of the income distribution, we explore rural–urban splits of income distribution and find that while the bottom decile of the urban distribution gains income share from 2.5% to 2.7%, the bottom two deciles of the rural distribution see significant declines, with the income share of the bottom decile declining sharply by 43%, from 1.3% to 0.76%. We also find that the GICs for the urban and rural distributions reveal that while all deciles in the urban distribution experienced positive real income growth from 2014–19, the bottom decile of the rural distribution experienced negative real income growth for this period at -5% per annum. The bottom ventile of the composite Indian distribution that experienced negative income growth is therefore composed largely of rural incomes. We validate this empirically and find that the bottom ventile of the composite distribution is composed of the first 7 rural percentiles of rural incomes and only part of the second percentile of urban incomes, thus confirming that negative income growth is essentially a rural phenomenon.

Using household data for each percentile, we find that the bottom decile of the Indian income distribution is comprised largely of small, marginal, and organised farmers as well as agricultural and wage labour and also that these occupations are performed predominantly by the SCs and STs whose participation in the workforce is maximised at the bottom of the income distribution. Taken together, we can confirm that the lowest incomes in India today are comprised predominantly of the SC and ST populations working in the most economically precarious occupations and suffering from the double whammy of declining income shares and real income declines. We also highlight the possibility that these groups have been in a negative growth spiral since the early 2000s.

In order to assess the redistribution occurring within the income distribution, we use the RGBM model to quantify the nature and extent of reallocation inherent in the income distribution. We find that reallocation rates are declining in all distributions (consolidated, rural, and urban) from 2015 to 2018 and are even negative for the consolidated (in 2018) and rural (2017 and 2018) distributions. This means that the extent of redistribution is decreasing continuously but the decline into negative τ indicates the potential risks of continued negative reallocation—regressive redistribution of resources from the bottom to the top of the distribution. We also find that a past in the bottom of the income distribution is a good predictor of a present in that part of the distribution. This dynamic combined with our evidence of negative real income growth at the bottom of the Indian income distribution makes for a worrying prognosis of the future.

This fragility of incomes lower in the distribution is reflective of broader economic trends including the informalisation of formal workforce and agrarian distress. The design of sustained income support policies for marginal farmers and wage labourers is therefore an area that requires immediate attention.


Acharya, S (2005): “Thirty Years of Tax Reform in India,” Economic & Political Weekly, Vol 40, No 20, pp 2061–70.

Banerjee, A and T Piketty (2005): “Top Indian Incomes, 1922–2000,” WBER, Vol 19, No 1, pp 1–20.

Banerjee, A, V Yakovenko and T Di Matteo (2006): “A Study of the Personal Income Distribution in Australia,” Physica A, Vol 370, pp 54–59.

Basole, A (2014): “Dynamics of Income Inequality: Insights from World Top Incomes Database,” Economic & Political Weekly, Vol 40, pp 14–17.

Berman, Y, O Peters and A Adamou (2017): “An Empirical Test of the Ergodic Hypothesis: Wealth Distributions in the United States,” SSRN.

Bhagwati, J and A Panagariya (2013): Why Growth Matters: How Economic Growth in India
Reduced Poverty and the Lessons for Other Developing Countries,
 New York: PublicAffairs.

Borooah, V, D Diwakar, V Mishra, A Naik and N Sabharwal (2014): “Caste, Inequality, and Poverty in India: A Re-assessment,” Development Studies Research, pp 279–94.

Census of India (2011): Census of India.

Chancel, L and T Piketty (2019): “Indian Income Inequality, 1922-2014: From British Raj to Billionaire Raj?” The Review of Income and Wealth, Series 65, No S1, pp S33–S62.

Chatterjee, A, A Chakrabarti, A Ghosh, A Chakraborti and T Nandi (2016): “Invariant Features of Spatial Inequality in Consumption: The Case of India,” Physica A, Vol 442, pp 169–81.

Deaton, A (2003): “Health, Inequality, and Economic Development,” Journal of Economic Literature, Vol 41, No 1, pp 113–58.

Deaton, A and J Dréze (2002): “Poverty and Inequality in India: A Re-examination,” Economic & Political Weekly, Vol 37, No 36, pp 3729–48.

Dev, S M and C Ravi (2007): “Poverty and Inequality: All-India and States, 1983–2005,” Economic & Political Weekly, Vol 42, No 6, pp 509–21.

Dhongde, S (2007): “Measuring the Impact of Growth and Income Distribution on Poverty in India,” Journal of Income Distribution, Vol 16, No 2, pp 25–48.

Drăgulescu, A and V Yakovenko (2001): “Exponential and Power-law Probability Distributions of Wealth and Income in the United Kingdom and the United States,” Physica A, Vol 299, pp 213–21.

Dutt, G and M Ravallion (2009): “Has India’s Economic Growth Become More Pro-Poor in the Wake of Economic Reforms?” World Bank Policy Research Working Paper, 5103.

Ghosh, A, K Gangopadhyay and B Basu (2011): “Consumer Expenditure Distribution in India, 1983–2007: Evidence of a Long Pareto Tail,” Physica A, Vol 390, pp 83–97.

Kennedy, J and L King (2014): “The Political Economy of Farmers’ Suicides in India: Indebted Cash Crop Farmers with Marginal Landholdings Explain State-level Variation in Suicide Rates,” Globalization and Health, Vol 10, No 16.

Kohli, A (2012): Poverty Amid Plenty in the New India, New York: Cambridge University Press.

Mehrotra, S (2019): Informal Employment Trends in the Indian Economy: Persistent Informality, but Growing Positive Development, Geneva: ILO.

Milanovic, B (2016): Global Inequality: A New Approach for the Age of Globalization, US: Harvard University Press.

Nagaraj, K, P Sainath, R Rukmani and R Gopinath (2014): “Farmers Suicides in India: Magnitudes, Trends, and Spatial Patterns 1997–2012,” Review of Agrarian Studies, Vol 4, No 2, pp 53–83.

Nagaraj, R (1990): “Growth Rate of India’s GDP, 1950-51 to 1987-88,” Economic & Political Weekly, Vol 25, No 26, pp 1396–403.

Narayanamoorthy, A and S Kalamkar (2005): “Indebtedness of Farmer Households Across States: Recent Trends, Status, and Determinants,” Indian Journal of Agricultural Economics, Vol 60, No 3, pp 290–301.

Panagariya, A and V More (2014): “Poverty by Social, Religious and Economic Groups in India and its Largest States,” Indian Growth and Development Review, Vol 7, No 2, pp 202–30.

Panagariya, A and M Mukim (2014): “A Comprehensive Analysis of Poverty in India,” Asian Development Review, Vol 31, No 1, pp 1–52.

Ranganathan, T, A Tripathi and B Rajoriya (2016): “Changing Sources of Income and Income Inequality among Indian Rural Households,” National Seminar on Dynamics of Rural Labour Relations, Institute of Economic Growth, New Delhi.

Rawal, V and M Swaminathan (2011): “Income Inequality and Caste in Village India,” Review of Agrarian Studies, Vol 1, No 2,

RBI (2021): Database on Indian Economy, retrieved from Reserve Bank of India,

Rodrik, D and A Subramanian (2004): “The Mystery of the Indian Growth Transition,” NBER Working Paper, 10376.

Sahasranaman, A and H Jensen (2021): “Dynamics of Reallocation within India’s Income Distribution,” Indian Economic Journal, Vol 56, pp 1–23.

Shariff, A and M Azam (2009): “Income Inequality in Rural India: Decomposing the Gini by Income Sources,” SSRN,

Singh, A (2011): “Farm Income Inequality and the Role of Caste: New Evidence from India,” Economics Bulletin, pp 2847–62.

Souma, W (2001): “Universal Structure of the Personal Income Distribution,” Fractals, Vol 9, No 4, pp 463–70.

Suri, K (2006): “Political Economy of Agricultural Distress,” Economic & Political Weekly, Vol 41, No 16, pp 1523–29.

Vaidyanathan, A (2006): “Farmers’ Suicides and the Agrarian Crisis,” Economic & Political Weekly, Vol 40, pp 4009–13.

Vakulabharanam, V and S Motiram (2011): “Political Economy of Agrarian Distress in India since the 1990s,” Understanding India’s New Political Economy: A Great Transformation?, Abingdon, Oxon: Routledge, pp 117–42.


Dear Reader,

To continue reading, become a subscriber.

Explore our attractive subscription offers.

Click here

Updated On : 11th Jul, 2022
Back to Top