In emerging economies like India, where income data are often unreliable or incomplete, consumption offers a more stable and accurate measure of household welfare and inequality than income-based metrics.
Image Source: Getty Images
With the World Bank reporting a decline in the consumption Gini for India from 0.288 in 2011 to 0.255 in 2022, indicating a decrease in inequality, a raging debate has arisen. This “declining inequality” thesis has been challenged on grounds of the irreconcilability between the World Bank's numbers and the World Inequality Database’s (WID) income Gini, which is high and rising over time, having increased from 0.59 in 2011 to 0.61 in 2022.
In this rigmarole of Indian performance and which measure is a better reflector of “(in)equality”, this article presents four propositions: a) economic well-being should be measured in terms of consumption, not income; b) given the same, consumption Gini is a better measure of inequality than income Gini; c) WID’s database fails to capture the true nature of inequality in emerging economies like India; and d) the World Bank must improve its inequality data to render comparability between nations.
While household income has traditionally been touted as a measure of purchasing power and economic well-being, an increasing body of research, especially in the context of developing economies like India, suggests that consumption is a more reliable, robust, and meaningful measure of household well-being. This preference for consumption becomes particularly important when measuring inequality and designing policies to address poverty and disparities. This delineation of consumption as a more reflective well-being parameter emerges organically and normatively.
Income represents the flow of earnings from wages, interest, profits, or transfers over a period of time. Consumption, by contrast, reflects the goods and services a household purchases, and is often seen as a more direct measure of the standard of living. Therefore, household consumption is the most important facet of quality of living. In India, household consumption data collected in various NSS rounds since 1950-51 have been extensively used for studies on standards of living, measurement of absolute poverty, and disparities across space and time.. As such, size distributions of consumption also reflect the measurement of inequality.
Since consumption reveals actual living conditions, it has emerged as a better indicator and a more effective tool for poverty measurement and inequality, which justifies its ubiquitous use in calculating poverty lines and food security indicators in the developing world.
Further, as long posited by Milton Friedman in his Permanent Income Hypothesis, consumption is typically less volatile than income, providing a more stable lens through which to view household welfare. Income can fluctuate significantly due to a host of factors. Factors like seasonal employment, informal sector volatility, or temporary unemployment are features prevalent in middle-income economies. In various economies, including the higher-income ones, households smooth their consumption in response to these shocks by drawing on savings, borrowing, or relying on community networks.
Therefore, there are four good reasons for choosing consumption over income as an indicator of welfare. First, consumption is clearly the “realised” material well-being, with people deriving utility from what they consume, not from the potential purchasing power alone. Second, in emerging economies like India, where a large share of employment and transactions occur informally or in kind, income is often underreported or unobservable, while consumption (especially of essentials) is easier to track. Third, consumption is associated with lower measurement error at the bottom, as compared to income. Studies show that poor households tend to underreport income more than consumption. The tangible expenditures are more easily recalled by respondents in a survey than fluctuating or undocumented earnings. Fourth, since consumption reveals actual living conditions, it has emerged as a better indicator and a more effective tool for poverty measurement and inequality, which justifies its ubiquitous use in calculating poverty lines and food security indicators in the developing world.
The above points indicate that the consumption-based inequality measure is a better indicator of the distributional impacts of economic growth and prosperity than the income-based Gini. Apart from being more stable over time due to the circumvented overreaction to short-term income shocks and being better aligned with permanent income and long-run welfare, consumption remains more grounded in reality, particularly when governments report official consumption data more consistently than income. This is the case with India.
However, there have been criticisms that consumption inequality may understate disparities at the top of the distribution. In contrast, income (particularly from tax records) can provide better insights into the concentration of economic power among the top 1 percent or 10 percent. While consumption may increase with income, the marginal propensity to consume (i.e., the increase in consumption resulting from a rise in income) declines as incomes rise. Therefore, after a certain level of income, consumption may not increase proportionally, as higher-income households tend to allocate a larger portion of their increased disposable income to savings compared to lower-income households. The latter tend to spend a larger component of their income on consumption than the former.
Apart from being more stable over time due to the circumvented overreaction to short-term income shocks and being better aligned with permanent income and long-run welfare, consumption remains more grounded in reality, particularly when governments report official consumption data more consistently than income.
However, that does not detract from the fact that India's consumption inequality has been declining over time. Rather, it is a reflection that more money is coming into the hands of the lower-income groups, which helps them satisfy their consumptive needs and improve their standards of living. More money in the hands of the richer household does not get into the consumption stream, but has the potential to get into the savings stream that, axiomatically, creates the potential investible pool for the economy. Therefore, a declining consumption Gini is a clear indication of not just reduced inequality, but also improved standards of living. The increase in consumption of the poorer households in India is evident from the decline in extreme poverty from 27.1 percent in 2011-12 to 5.3 percent in 2022-23, as reported by the World Bank, leading to approximately 269 million people being lifted out of extreme poverty.
Table 1 shows that the decline in consumption Gini has been associated with increasing per capita consumption at constant prices – at a much higher rate for the lower 10 percent and 50 percent, as compared to the upper 10 percent and 50 percent between 2011 and 2022 – evidencing the narrative of an “equalising India”.
Table 1: Consumption Gini and Per Capita Consumption in India (2011 and 2022)
| 2011 | 2022 | Change ( percent) | |
| Consumption Gini | 0.2878 | 0.2551 | -11.34 |
| Per capita Consumption (INR constant prices) | 40582 | 67518 | 66.37 |
| Per capita consumption of the bottom 50 percent (INR constant prices ) | 25126.99 | 44317.54 | 76.37 |
| Per capita consumption of the bottom 10 percent (INR constant prices) | 17246.13 | 30454.61 | 76.59 |
| Per capita consumption of the top 50 percent (INR constant prices) | 56037.42 | 90717.61 | 61.89 |
| Per capita consumption of the top 10 percent (INR constant prices) | 100653.35 | 149431.54 | 48.46 |
Source: Author’s estimates based on data from the World Bank; Ministry of Statistics and Programme Implementation (MoSPI); Census of India; and Lok Sabha Unstarred Question No. 890, Ministry of Home Affairs (MoHA), Government of India.
On the other hand, the WID’s fundamental contention on inequality is based on income tax records, which are fraught with fundamental limitations. With only about 5 percent of the population filing tax returns in a country like India, the entire inequality analysis is contingent on a narrow base, eliminating the vast informal sector and the majority of the lower and even middle-income population. This organically skews the distribution, making it top-heavy and biased towards the wealthy.
Further, in a scathing critique of the WID database, James Galbraith points out the inherent inconsistencies emerging from the WID’s reliance on tax records due to the unevenness of the quality and continuity of the tax data in most economies. In India’s case, historical income tax records stretch back to 1922, but estimation of the bottom 90 percent and middle 40 percent income shares involves heroic assumptions and interpolations due to many years of missing or sparse data. This challenges the validity of trends, particularly when dramatic policy shifts affect income distribution but may not be well-captured in tax filings. Additionally, tax-based inequality measures bear the risk of conflating changes in compliance or filing behaviour with actual shifts in economic inequality.
With only about 5 percent of the population filing tax returns in a country like India, the entire inequality analysis is contingent on a narrow base, eliminating the vast informal sector and the majority of the lower and even middle-income population. This organically skews the distribution, making it top-heavy and biased towards the wealthy.
Another concern with WID lies in the broader methodological claim of its superiority over household surveys in terms of revealing consumption. As previously noted, in developing economies like India, even lower-income groups tend to underrepresent their incomes, while providing more accurate accounts of their consumption. Further, the WID estimates are inconsistent with other alternative inequality measures, including those derived from payroll or industrial earnings data, such as the University of Texas Inequality Project's Estimated Household Income Inequality (UTIP-EHII series). So, any inference on inequality based on the WID data does not hold water.
That doesn’t mean that everything is hunky-dory with the World Bank’s consumption Gini. Presently, the consumption Gini can only serve as a metric for gauging an economy’s performance on inequality over time. The data remains incomparable across nations, as the World Bank reports consumption-based Gini for some nations and income-based Gini for others. Even the periods of reporting are not the same. There are only 25 other nations that have comparable data to India over the indicator and time, and adjustments need to be made there as well. Therefore, three clear suggestions emerge for the World Bank to improve the existing inequality database: a) there is every reason to move towards consumption-based Gini for all economies; b) the method of data collection needs to be uniform; and c) the reporting years need to be compatible across nations.
The author benefited from discussions with Gautam Chikermane.
Nilanjan Ghosh is Vice President - Development Studies at the Observer Research Foundation.
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.
Dr Nilanjan Ghosh heads Development Studies at the Observer Research Foundation (ORF) and serves as the operational and executive head of ORF’s Kolkata Centre. He ...
Read More +
PREV
