VoxEU Column Development Poverty and Income Inequality

No-one left behind?

The proportion of the population in the developing world that lives below the poverty line has been decreasing. This column argues, however, that such a counting approach is not the only way to assess whether the poorest have been reached. An alternative is to use the success in raising the consumption floor as an indicator of social progress. When applying such an approach, the author finds little progress for the poorest.    

At the launch of the Millennium Goals Report, UN Secretary-General Ban Ki-moon said that:

“The poorest of the world are being left behind. We need to reach out and lift them into our lifeboat.”

This view is heard quite often. Yet, other observers tell a very different story. They use aphorisms such as “a rising tide lifts all boats” or they point to evidence that the poor are “breaking through from the bottom” (Radelet 2015).     

Can we make sense of these seemingly conflicting views on this important question? The preferred approach of economists and statisticians has been to count the poor in some way. One might track the proportion of the population living below some deliberately low poverty line or use a more sophisticated measure giving higher weight to poorer people. I dub this the counting approach. The theoretical foundations of the approach are found in the literature on poverty measurement, in which various axioms have been proposed.

This approach does not seem to offer much support for Ban Ki-moon’s view. The evidence using the counting approach indicates falling incidence and depth of absolute poverty in the developing world over recent decades, when judged by poverty lines typical of developing countries (Chen and Ravallion 2010, Ravallion and Chen 2013). Ban Ki-moon’s view could still be right if past applications of the counting approach have used too high a poverty cut-off, thus missing the poorest. The poorest subset of the poor can be called the ‘ultra-poor’. In principle, progress against poverty can be achieved in large part by lifting people near the poverty line out of poverty, with little gain to the ultra-poor.

The consumption floor

That is not what I find in a new paper (Ravallion 2014a). Drawing on the results from household surveys for developing countries spanning 1981-2011, I find considerable progress against ultra-poverty using the counting approach; indeed, the bulk of either the inter-temporal or the cross-country variance in rates of poverty reduction for either $1.25 or $2.00 a day is accountable to progress for those living under $0.87 or even $0.77 a day. There is first-order dominance over the 30 years, implying an unambiguous reduction in absolute poverty by the counting approach over all lines and all additive measures. By this version of the counting approach the poorest as a whole have not been left behind.

The counting approach is not the only way of assessing whether the poorest have been reached, although it has been (by far) the main approach used by economists. If overall economic progress is not to ‘leave the poorest behind’ then it must, in due course, raise the lower bound of the distribution of permanent consumption levels in society. That lower bound can be called the consumption floor. Human physiology makes the existence of a positive floor plausible, given the nutritional requirements for basal metabolism. This can be called the ‘the biological floor’. In practice, however, the actual consumption floor can rise above the biological floor. Private social interactions can provide some degree of protection. A consumption floor above the biological minimum may also stem from a policy-supported minimum income.

Figure 1. Same reduction in the poverty count but different implications for the poorest

Figure 1 illustrates the difference between this perspective and the counting approach. Each panel gives two cumulative distribution functions (CDFs). In each case, the upper CDF is the initial one and the lower CDF is for some later date. The drop in the incidence of poverty (which could be for an ultra-poverty line) is similar in panels (a) and (b).

  • In panel (a), the counting approach might reasonably claim that many of the poorest have been reached even though the floor has not risen, so some people still remain living at the same very low level.
  • In panel (b), the same reduction in the poverty rate has come with a rising floor—implying that none of the poorest are left behind.

The idea that we should judge progress in part at least by success in raising the floor is missing from all standard poverty measures, as used in the counting approach. The concept of the consumption floor is conceptually distinct from existing poverty lines. Naturally, any poverty line aims to reflect what ‘poverty’ means in a specific society, on the understanding that (potentially many) people live below that level. The poverty line is a normative concept, while the consumption floor is a positive one.

Nor do standard poverty measures necessarily reflect any progress in raising the floor. Although a higher consumption floor (ceteris paribus) automatically reduces any measure of consumption poverty satisfying the monotonicity axiom, none of the standard axioms of poverty measurement attach any explicit weight to the level of the floor. This appears to be due at least in part to the difficulties in identifying the floor. This is surely a gap in the current ‘dashboard’ of development indicators.

Focusing on the floor can also draw support from the literature, though largely outside economics. An important school of moral philosophy has argued that we should judge a society’s progress by its ability to enhance the welfare of the least advantaged, following the principles of justice proposed by John Rawls (1971). By this view, a higher floor (as in Figure 1b) is not only preferred, but is the main criterion of distributive justice (subject to other criteria of liberty, as identified by Rawls). While Rawls’s ‘difference principle’ is often interpreted as ‘maximin’ (to maximise the minimum level of welfare), Rawls insisted that some degree of averaging was required in defining the ‘least advantaged’. I call the idea of focusing on the expected welfare of the poorest stratum in assessing social progress the Rawlsian approach.

While the Rawlsian approach of using success in raising the consumption floor as an indicator of social progress has not been favoured by economists, it has deep roots in development and social policy thinking. Versions of the approach thrive today in policy discussions. In a famous example in 1948 (shortly before his assassination), Mahatma Gandhi was asked “How can I know that the decisions I am making are the best I can make”? He answered:

“I will give you a talisman. Whenever you are in doubt, or when the self becomes too much with you, apply the following test. Recall the face of the poorest and the weakest man whom you may have seen, and ask yourself if the step you contemplate is going to be of any use to him. Will he gain anything by it?”

The spirit of Gandhi’s talisman was echoed (in somewhat dryer terms) 65 years later in a report initiated by the UN on setting new development goals, which argued that:

“The indicators that track them should be disaggregated to ensure no one is left behind and targets should only be considered ‘achieved’ if they are met for all relevant income and social groups.” (UN 2013, Executive Summary; my emphasis)

Endorsing this view in the UN report, Watkins (2013) refers explicitly to Gandhi’s talisman, and argues that “As a guide to international cooperation on development, that’s tough to top.”

Assessing the progress for the poorest

Estimating the level of the consumption floor is a difficult task. I have proposed two methods of estimating the consumption floor, both of which can be implemented with readily available data sources (Ravallion, 2014a). Both methods use data from national poverty lines in poor countries, but otherwise they make very different assumptions.

  • The first method aims to identify an expected consumption floor amongst those who are identified as poor in absolute terms by the standards of poor countries. It would not be sensible to simply measure the lowest observed consumption level in the available data. This could well be subject to sizeable measurement errors, and also idiosyncratic transient effects in consumption or income. Echoing Rawls, it makes more sense to form an expectation over a stratum of people with low observed consumption levels. The lowest observed consumption is assumed to have the highest probability of being at the floor, but that probability is less than one. The probability declines linearly as consumption rises above the lowest observed value up to some critical point, above which there is zero probability of being the poorest.
  • The second method assumes that national poverty lines comprise the expected value of the consumption floor plus a relative component proportional to actual mean consumption in that country.

Both methods of quantifying the level of the floor suggest a consumption floor today that is about half of the international poverty line of $1.25 a day. This is probably close to the consumption of essential foods for those living around $1.25 a day.

My principal empirical finding is that, while the counting approach shows huge progress for the poorest, the Rawlsian approach does not. The distribution of the gains amongst the poor has meant that the expected value of the consumption floor has risen little over the last 30 years.  Figure 2 gives the estimated level over time for the developing world as a whole. It is evident that there has been very little progress in raising the floor despite the progress (accelerating since 2000) is raising the overall mean consumption. 

Figure 2. Mean consumptions for the developing world

Figure 3 gives another perspective on the issue. Here, I give the absolute gain in consumption in the developing world over 1981-2011 by percentile. Consistently with the lack of progress in raising the floor, we see that the gains are close to zero for the poorest, but rising to quite high levels. This is also consistent with what we know about rising absolute inequality in the developing world; on this see Ravallion (2014).

  • While there has been marked progress in reducing the numbers of the ultra-poor, and the poorest have not been left behind, the expected value of the lowest level of living amongst those who are considered poor by developing country standards has advanced rather little. 

Figure 3. Absolute gains by percentile 1981-2011

Concluding remarks

To anticipate one response, it might be argued that progress in lifting the floor is a second-order issue, as long as fewer people live near the floor. That is implicit in the traditional counting methods used to assess progress against poverty. However, proponents of this view must surely take pause when one notes that for a long time, and across countries at very different levels of development, social policies have often claimed that they aim to ensure a minimum level of living above any biological consumption floor required for mere survival. Negative income tax schemes and (formally-equivalent) basic-income guarantee schemes financed by progressive income taxes aim to raise society’s consumption floor above the biological minimum. And such efforts are not confined to rich countries; indeed, the two largest anti-poverty programmes in the world today (in China and India) aim to raise the floor.

While it may well be ill-advised to look solely at the level of the floor in a given society, it can be acknowledged that this has normative significance independently of attainments in reducing the number of people living near that floor. The argument here is not that progress against poverty should be judged solely by the level of the consumption floor, but only that the latter should not be ignored.

In demonstrating the difference between the counting and Rawlsian approaches, this new research also throws light on how the idea of the consumption floor can be interpreted as transformations of standard poverty measures. This permits straightforward applications to the task of monitoring progress in raising the level of living of the world’s poorest. Our success in assuring that no-one is left behind is thus readily monitored from existing data sources.


Chen, S, and M Ravallion (2010), “The Developing World is Poorer than we Thought, but no Less Successful in the Fight against Poverty,” Quarterly Journal of Economics 125(4): 1577-1625.

Gandhi, M (1958), The Last Phase, Volume II, Ahmedabad: Navajivan Publishing House.

Radelet, S (2015), The Great Surge: The Unprecedented Economic and Political Transformation of Developing Countries Around the World, Simon & Schuster.

Ravallion, M (2014a), “Are the World’s Poorest being Left Behind?” NBER Working Paper, 20791.

Ravallion, M (2014b), “Income Inequality in the Developing World,” Science 344, 851-5.

Ravallion, M, and S Chen (2013), “More Relatively Poor People in a Less Absolutely Poor World,” Review of Income and Wealth 59(1): 1-28.

Rawls, J (1971), A Theory of Justice, Cambridge MA: Harvard University Press.

UN (2013), A New Global Partnership: Eradicate Poverty and Transform Economies Through Sustainable Development. The Report of the High-Level Panel of Eminent Persons on the Post-2015 Development Agenda, United Nations.

Watkins, K (2013), “Leaving No-one Behind: An Equity Agenda for the Post-2015 Goals,” Overseas Development Institute, London.

4,304 Reads