Themes > Current Issues
25.07.2012
Inequality in South Asia*
Jayati
Ghosh
The
South Asian region used to be described as one of the less unequal
in the developing world, particularly in contrast with Latin America,
and more recently even in contrast with China, which has experienced
rapid increases in income inequality during its period of rapid growth.
This perception was always contested by those who felt that inequality
particularly in India has been underestimated. But in any case recent
increases even in recorded inequality suggest that the earlier complacency
is unwarranted.
South Asia is clearly a region in which the period of globalization has been associated with greater income and consumption inequalities. This is generally described as the result of the combination of the forces of globalisation (more open trade and capital flows, greater economic integration at different levels) with the forces of technological progress that reduce demand for labour per unit of output. Both of these forces are described as reducing the bargaining power of workers relative to capital, and in addition the process of economic growth has itself been described as one that requires higher levels of inequality in the high accumulation phase.
However, when the region is compared to other parts of the developing world, it is evident that this explanation is simplistic. Rather, the recent more unequal outcomes in South Asia appear to be also very much determined by policies and broader choices of economic growth strategy.
All this is particularly evident in the case of the largest economy in the region, India. The only available large-scale survey data in India relate to consumption expenditure, which tend to understate the extent of inequality by underestimating the tails of the distribution (excluding the very rich and the very poor) and because the poor are more likely to consume as much or even more than their income while the rich are more able to save.
Indeed, the first detailed income distribution estimates for India (described in Sonalde B. Desai, Amaresh Dubey, Brij Lal Joshi and Mitali Sen, ''Human Development in India'' Oxford University Press, 2010) reveal quite high income inequality, with a Gini coefficient of 0.54 – or around the same as Brazil. Estimates based on village surveys derive even higher Gini coefficients: on average 0.645 across households and 0.595 across persons even within villages (Madhura Swaminathan and Vikas Rawal ''Is India Really a Country of Low Income-Inequality? Observations from Eight Villages'', Review of Agrarian Studies 2011).
Even consumption data suggest increasing inequality of consumption, in both vertical and horizontal terms. The national Gini coefficient for consumption increased from 0.31 in 1993-94 to 0.36 in 2009-10, while the ratio of urban to rural consumption went up from 1.62 to 1.96. The largest increases in consumption expenditure were concentrated in the top decile of the urban population: between 1993-94 and 2009-10, the income of the top urban decile went from 7.14 times to 10.33 that of the bottom urban decile and from 10.48 to 14.32 times that of the bottom rural decile.
The movement of factor incomes corroborates the tendency towards greater inequality: the wage share of national income fell from 40 per cent at the start of the 1990s to only 34 per cent by 2009-10, while in the organised sector the wage share fell from 69 per cent to 51 per cent in the same period. Meanwhile, even though the unorganised sector continues to account for the overwhelming majority of workers in the country, including the self-employed, its share of national income fell from 64 per cent to 57 per cent.
So the gains from Indian growth were concentrated among surplus-takers, including profits, rents and financial incomes. A major reason for this is that the growth has not been suffficiently employment generating, and therefore around half of the work force continues to languish in low-productivity agriculture (even though that sector now accounts for around 15 per cent of GDP) and in low remuneration services. The growth was driven by internal and external liberalisation measures that attracted global financial investors. Capital inflows sparked a domestic retail credit boom, which combined with fiscal concessions to spur consumption of the better-off sections. This led to rapid increases in aggregate GDP growth, even though compressed public spending on basic needs, poor employment generation and persistent agrarian crisis reduced wage shares in national income and kept mass consumption demand low.
Similarly, Bangladesh moved from being a developing country with relatively low inequality in the early 1990s to one with moderately high inequality by the middle of the first decade of the 21st century, as the Gini coefficient for income (based on extrapolating from consumer expenditure survey data) increased from 0.276 in 1991-92 to 0.404 in 2005. This increase was ''steady, uninterrupted and pervasive.'' (Khan 2008) Farm incomes dwindled as proportion of total income over time. The growth of tenant farming increased access to land, but the increasingly concentrated nature of landownership made the distribution of rents from land very unequal. Increasing wage differentials in non-agricultural activities (between relatively less skilled wage workers and relatively more skilled salaried workers) added to the inequality. Growth in remittance incomes, especially those from workers abroad, also had a strongly inequalizing effect in Bangladesh.
Measurements of income inequality in Pakistan can vary widely. Most estimates suggest relatively stable Gini coefficients just below 0.4 for the 1980s and 1990s and slight increases thereafter. Consumer surveys indicate that consumption inequality decreased in the first half of the 1990s and then increased for the next decade. In the first period, the poorest quintile gained significantly in their consumption share while that of middle 60 percent and the richest quintile reduced. From 1996-97 to 2004-05 the consumption share of the poorest 20 quintile and the richest quintile increased, while the share of middle 60 percent decreased.
Both trade openness and financial instability have been linked to this pattern of greater economic inequality. The relatively low employment generation continues to trap around two-thirds of the work force in agriculture, while inadequate public investment in infrastructure and essential social services has added to the problem. It has been argued that these reflect the combination of globalisation with ''an institutional structure that excludes a large proportion of the population from the process of economic growth as well as governance.'' (Husain, Akmal, Haris Gazdar and Asad Sayeed ''Power dynamics, institutional instability and economic growth: The case of Pakistan'', The Asia Foundation, Islamabad 2008)
Sri Lanka was the first country in the region to engage systematically in greater global integration. The government liberalized the economy and adopted market-oriented reforms in 1978, but initially, in the 1980s inequality remained relatively low. But by the mid 1990s Sri Lanka’s inequality was higher than its neighbours, and the IMF even estimated that it recorded the highest increase in the Gini index among selected Asian countries between 1990 and 2004. By the mid 2000s, the Gini for consumption was 0.46 and that for income was 0.5. Inequalities in Sri Lanka have strong regional, sectoral and ethnic dimensions. Rising inequality reflects two components: first, growing vertical inequality within the fast growing modern industrial sector and region, driven by asset and differences in skill levels; and second, between the modern industrial fast-growing sectors and regions and the traditional agricultural lagging sectors and regions.
Just because this is common across these four countries in the region does not mean that it is inevitable. Other highly globally integrated regions and countries, such as those in Latin America, have actually shown that it is possible to reduce inequalities, albeit from high pre-existing levels, through active policies designed for redistribution, and that this can be done at more than respectable rates of income growth. Clearly, the region has much to learn from proactive policies for equity elsewhere in the world.
* This article was originally published in the Frontline, Vol. 29: No. 15, Jul 28-Aug 10, 2012.
South Asia is clearly a region in which the period of globalization has been associated with greater income and consumption inequalities. This is generally described as the result of the combination of the forces of globalisation (more open trade and capital flows, greater economic integration at different levels) with the forces of technological progress that reduce demand for labour per unit of output. Both of these forces are described as reducing the bargaining power of workers relative to capital, and in addition the process of economic growth has itself been described as one that requires higher levels of inequality in the high accumulation phase.
However, when the region is compared to other parts of the developing world, it is evident that this explanation is simplistic. Rather, the recent more unequal outcomes in South Asia appear to be also very much determined by policies and broader choices of economic growth strategy.
All this is particularly evident in the case of the largest economy in the region, India. The only available large-scale survey data in India relate to consumption expenditure, which tend to understate the extent of inequality by underestimating the tails of the distribution (excluding the very rich and the very poor) and because the poor are more likely to consume as much or even more than their income while the rich are more able to save.
Indeed, the first detailed income distribution estimates for India (described in Sonalde B. Desai, Amaresh Dubey, Brij Lal Joshi and Mitali Sen, ''Human Development in India'' Oxford University Press, 2010) reveal quite high income inequality, with a Gini coefficient of 0.54 – or around the same as Brazil. Estimates based on village surveys derive even higher Gini coefficients: on average 0.645 across households and 0.595 across persons even within villages (Madhura Swaminathan and Vikas Rawal ''Is India Really a Country of Low Income-Inequality? Observations from Eight Villages'', Review of Agrarian Studies 2011).
Even consumption data suggest increasing inequality of consumption, in both vertical and horizontal terms. The national Gini coefficient for consumption increased from 0.31 in 1993-94 to 0.36 in 2009-10, while the ratio of urban to rural consumption went up from 1.62 to 1.96. The largest increases in consumption expenditure were concentrated in the top decile of the urban population: between 1993-94 and 2009-10, the income of the top urban decile went from 7.14 times to 10.33 that of the bottom urban decile and from 10.48 to 14.32 times that of the bottom rural decile.
The movement of factor incomes corroborates the tendency towards greater inequality: the wage share of national income fell from 40 per cent at the start of the 1990s to only 34 per cent by 2009-10, while in the organised sector the wage share fell from 69 per cent to 51 per cent in the same period. Meanwhile, even though the unorganised sector continues to account for the overwhelming majority of workers in the country, including the self-employed, its share of national income fell from 64 per cent to 57 per cent.
So the gains from Indian growth were concentrated among surplus-takers, including profits, rents and financial incomes. A major reason for this is that the growth has not been suffficiently employment generating, and therefore around half of the work force continues to languish in low-productivity agriculture (even though that sector now accounts for around 15 per cent of GDP) and in low remuneration services. The growth was driven by internal and external liberalisation measures that attracted global financial investors. Capital inflows sparked a domestic retail credit boom, which combined with fiscal concessions to spur consumption of the better-off sections. This led to rapid increases in aggregate GDP growth, even though compressed public spending on basic needs, poor employment generation and persistent agrarian crisis reduced wage shares in national income and kept mass consumption demand low.
Similarly, Bangladesh moved from being a developing country with relatively low inequality in the early 1990s to one with moderately high inequality by the middle of the first decade of the 21st century, as the Gini coefficient for income (based on extrapolating from consumer expenditure survey data) increased from 0.276 in 1991-92 to 0.404 in 2005. This increase was ''steady, uninterrupted and pervasive.'' (Khan 2008) Farm incomes dwindled as proportion of total income over time. The growth of tenant farming increased access to land, but the increasingly concentrated nature of landownership made the distribution of rents from land very unequal. Increasing wage differentials in non-agricultural activities (between relatively less skilled wage workers and relatively more skilled salaried workers) added to the inequality. Growth in remittance incomes, especially those from workers abroad, also had a strongly inequalizing effect in Bangladesh.
Measurements of income inequality in Pakistan can vary widely. Most estimates suggest relatively stable Gini coefficients just below 0.4 for the 1980s and 1990s and slight increases thereafter. Consumer surveys indicate that consumption inequality decreased in the first half of the 1990s and then increased for the next decade. In the first period, the poorest quintile gained significantly in their consumption share while that of middle 60 percent and the richest quintile reduced. From 1996-97 to 2004-05 the consumption share of the poorest 20 quintile and the richest quintile increased, while the share of middle 60 percent decreased.
Both trade openness and financial instability have been linked to this pattern of greater economic inequality. The relatively low employment generation continues to trap around two-thirds of the work force in agriculture, while inadequate public investment in infrastructure and essential social services has added to the problem. It has been argued that these reflect the combination of globalisation with ''an institutional structure that excludes a large proportion of the population from the process of economic growth as well as governance.'' (Husain, Akmal, Haris Gazdar and Asad Sayeed ''Power dynamics, institutional instability and economic growth: The case of Pakistan'', The Asia Foundation, Islamabad 2008)
Sri Lanka was the first country in the region to engage systematically in greater global integration. The government liberalized the economy and adopted market-oriented reforms in 1978, but initially, in the 1980s inequality remained relatively low. But by the mid 1990s Sri Lanka’s inequality was higher than its neighbours, and the IMF even estimated that it recorded the highest increase in the Gini index among selected Asian countries between 1990 and 2004. By the mid 2000s, the Gini for consumption was 0.46 and that for income was 0.5. Inequalities in Sri Lanka have strong regional, sectoral and ethnic dimensions. Rising inequality reflects two components: first, growing vertical inequality within the fast growing modern industrial sector and region, driven by asset and differences in skill levels; and second, between the modern industrial fast-growing sectors and regions and the traditional agricultural lagging sectors and regions.
Just because this is common across these four countries in the region does not mean that it is inevitable. Other highly globally integrated regions and countries, such as those in Latin America, have actually shown that it is possible to reduce inequalities, albeit from high pre-existing levels, through active policies designed for redistribution, and that this can be done at more than respectable rates of income growth. Clearly, the region has much to learn from proactive policies for equity elsewhere in the world.
* This article was originally published in the Frontline, Vol. 29: No. 15, Jul 28-Aug 10, 2012.
© MACROSCAN 2012