Inequality benchmark incomes: A neglected tool for analysing income distributions

Laurence Roope 18 March 2021



How best to grow the incomes of the world’s poorest has long been a focus of development economics and policy (Ravallion 2020). In recent years, there has also been much discussion of the role of income growth at the very top of the distribution in exacerbating inequality (Piketty et al. 2017). 

It is no surprise that inequality increases when the incomes of the very richest go up and decreases when the incomes of the very poorest go up, but what effect do income rises for people on middle incomes have on inequality? This is an important question for policymakers to consider. It could illuminate the likely impact on inequality both of income subsidies and of policies likely to grow incomes mainly in particular sections of the income distribution. 

Income inequality is typically measured using a variety of indices, the most popular of which is the Gini coefficient. Theoretical studies have shown that nearly all income inequality measures are associated with a benchmark income or position, above which income gains increase inequality, and below which income gains decrease inequality (Hoffmann 2001, Lambert and Lanza 2006, Roope 2019).

The benchmark income is a distinguishing feature of inequality measures that sets them apart from poverty measures. While poverty is unaffected when people above the poverty line get richer, inequality goes up when people above the benchmark income become richer. Part of the essence of inequality measures is that there is something socially undesirable about these increases. 

Beyond this common characteristic, different inequality measures can have very different properties. In particular, some pay more attention to the relative gaps between different incomes, while others pay more attention to the absolute gaps between incomes. These differences reflect the fact that there are fundamental differences in how people conceive of inequality (Ravallion 2015, Niño-Zarazúa et al. 2016). Moreover, trends in relative and absolute inequality often move in opposite directions, leading to considerable debate and controversy (Ravallion 2015, Niño-Zarazúa et al. 2016).   

Advantages of benchmark incomes

Poverty lines are often criticised for being arbitrary, with no firm basis for choosing a particular poverty line. An attractive feature of the benchmark income approach is that it is not arbitrary. Once the choice has been made to adopt a particular inequality measure, the benchmark income arises naturally and is determined by the inequality measure itself. If one has agreed to use a particular inequality measure, unlike poverty lines, there can be no disagreement over the level of the benchmark income.

Knowledge of the benchmark income could be used, for example, to predict the impact on inequality of subsidies to incomes in particular parts of the distribution. The benchmark income can be used to identify the richest person for whom it might be deemed fair to subsidise income financed by taxation (Corvalan 2014). Equivalently, the benchmark income can signify the poorest person whom it is just and fair not to subsidise their income (Roope 2021).

Despite the intuitive appeal of benchmark incomes, they are hardly ever used. In one of the first studies to illustrate where benchmark incomes lie in practice (Roope 2021), I find that, across a selection of ten contrasting countries, benchmark incomes associated with the Gini coefficient ranged from the 62nd percentile to the 85th percentile (Figure 1). 

Figure 1 Benchmark income and poverty line percentiles in 2010

The figure shows that all benchmark incomes for all countries lay far above official poverty lines. (The only exception is South Africa, where the national poverty line lay marginally above the 50th percentile, which is where the benchmark income implied by the absolute Gini measure lies). Across the ten countries studied, on average, half of the income distribution lay above the official poverty line but below the benchmark income implied by the Gini coefficient.

It might seem surprising that inequality – even of the absolute kind – can be reduced by income rises so far up the income distribution. This helps to illuminate just how distinct the concepts of inequality and poverty are, and underscores the importance of not conflating the two. 

There is, rightly, considerable interest in how to sustain economic growth while keeping inequality in check (Jalles and de Mello 2020) – and doing so in an environmentally sustainable way (Ravallion 2020). However, an important implication of my study is that economic growth alongside falling inequality need not necessarily be poverty-reducing. If, due to a particular pattern of economic growth, gains are made mainly to incomes above the poverty line but below the benchmark income, inequality will fall but poverty will not.

Policymakers should therefore be careful not to assume that an increase in average incomes, together with falling inequality, will necessarily result in people being lifted out of poverty. The economy can grow, and become more equal, while the incomes of the poor – and even those on middle incomes – are left behind. It will always be important to examine the overall pattern of growth, and in which percentiles of the economy it occurs.


Corvalan, A (2014), “The impact of a marginal subsidy on Gini indices”, Review of Income and Wealth 60(3): 596–603. 

Hoffmann, R (2001), “Effect of the rise of a person’s income on inequality”, Brazilian Review of Econometrics 21(2): 237-262. 

Jalles, J T and L de Mello (2020), “Pursuing inclusive growth”,, 22 October.

Lambert, P and G Lanza (2006), “The effect on inequality of changing one or two incomes”, Journal of Economic Inequality 4(3): 253–277.

Niño-Zarazúa, M, L S J Roope and F Tarp (2016), “Income inequality in a globalising world”,, 20 September.

Piketty, T, E Saez and G Zucman (2017), “Economic growth in the US: A tale of two countries”,, 29 March.              

Ravallion, M (2015), “Challenges in maintaining progress against global poverty”,, 23 December.

Ravallion, M (2020), “A (very) short history of the idea of ending poverty”,, 26 October.

Roope, L S J (2019), “Characterizing inequality benchmark incomes”, Economic Theory Bulletin 7(1): 131-145.

Roope, L S J (2021), “First estimates of inequality benchmark incomes for a range of countries”, PLOS One 16(3): e0248178.



Topics:  Poverty and income inequality

Tags:  income inequality, benchmark incomes, inequality measures, Gini coefficient

Senior Researcher, Health Economics Research Centre, University of Oxford


CEPR Policy Research