Definitions matter – part one

Why is it that many of the key statistics used to measure progress in development have no agreed definition – ‘city population’, ‘water availability’, ‘slum’ and ‘poverty line’ to name but a few?

David Satterthwaite's picture
Blog by
10 November 2021

David Satterthwaite is senior fellow in IIED's Human Settlements research group

Informal housing and tall buildings behind.

Slums in Mumbai, India. Between 2000 and 2010, the world’s slum population fell by over 200 million according to the UN's definition (Photo: Christian Croft via FlickrCC BY-NC-ND 2.0)

This matters because progress towards many of the UN’s Sustainable Development Goals (SDGs) is strongly influenced by these definitions and related numbers. It can have a knock-on effect for any remedial actions, and can ultimately threaten progress itself.

Why have ‘slum’ population levels dropped dramatically globally, as well as levels of urban poverty? And how come 98% of the world’s urban population now has ‘improved’ water?  

Yet we know that hundreds of millions of urban dwellers suffer from very poor-quality provision for water. The scale and depth of urban poverty is greatly underestimated. So too are slum populations.

Here, it is the definition that is at fault. Set a poverty line or standards for water or ‘slum’ definitions far too low and these problems disappear. And the false figures are used because they seem to show success.

‘Improved’ – but what has actually improved?

Ninety-eight per cent of the world’s urban population has what the UN categorises as ‘improved’ water. Even low-income and the least developed nations record that 95% of their urban dwellers benefit from ‘improved’ provision.

But the definition of ‘improved’ sets the bar so low, ranging from having water from a tap in the dwelling, to water kiosks and tanker trucks from which water has to be fetched and carried. Improved provision does not necessarily mean that the water is free from faecal contamination, or that its provision is safe, accessible, regular or affordable.

The same is true of sanitation levels where ‘improved’ provision ranges from a flush toilet connected to a sewer, to a shared pit latrine with slab. Are pit latrines with slabs really improved? Or appropriate for multi-storey buildings or households living in tiny rented rooms?

Part two of this blog will focus on water and sanitation issues in more detail.

Defining poverty

How poverty is defined has huge implications for determining the number of people classified as poor. If the poverty definition of US$1/day (and more recent upward adjustments to allow for inflation) is applied to urban populations, then there is apparently very little or no urban poverty in most of the global South. Set the poverty line low enough and no one is poor.

But if poverty is defined by the actual income needed to afford food and non-food needs, adjusted for local costs, the numbers living in urban poverty increase dramatically.

International/global poverty lines make very little allowance for non-food needs. So they miss the scale and depth of poverty in cities where low-income groups pay rents that they struggle to afford. They don’t allow for a lack of public services and the high costs of purchasing alternatives – for instance health care and private school fees (as many unofficial schools in informal settlements are poor quality and suffer from overcrowding).

The urban poverty line definition used by the English social reformer Benjamin Seebohm Rowntree over 120 years ago when researching poverty in the city of York is actually far better than most current definitions: “A minimum weekly sum of money to enable families… to secure the necessaries of a healthy life". These included fuel and light, rent, food, clothing, and household and personal items. The amount was also adjusted for family size.

Rural or urban?

You would think that statistics on rural and urban populations are not controversial. But if settlements benefit from rural development programmes, their local governments may resist being reclassified as urban despite fitting the official definition for an urban centre. Or populations may be exaggerated by city or state government if this brings in more funds from higher levels of government.

Comparisons are often made between countries’ or regions’ urbanisation levels – the percentage of the population living in urban centres. Most nations define urban centres using a population threshold, or include this among other criteria. But this threshold can vary from a few hundred to 50,000 inhabitants. As a result, in most nations, a substantial proportion of their population live in what can be considered and defined as large villages or small urban centres.

India’s population is predominantly urban. This actually is not true. But it would be if India chose to use Sweden’s definition of urban areas as: “Built-up areas with 200 inhabitants or more and where houses are at most 200 metres apart.” If this was applied to India, most of its rural settlements would indeed be urban.

City limits

For many decades London was the world’s largest city. But by 2020, it had fallen out of the top 30. If London wanted to ‘boost’ its population, it could easily re-establish itself among the world’s largest by following the example of Chinese cities that have boundaries extending far beyond their built-up areas.

This could be achieved if the Greater London Authority was able to convince the UK government that a new London Metropolitan region be created, incorporating neighbouring counties such as Surrey, Kent, Essex, Hertfordshire, Buckinghamshire and Berkshire – although one suspects there would be a certain reluctance on the part of most people living in these counties to such a change.

We compare city populations or population growth rates without recognising the many different ways in which city boundaries are defined – which limits the validity of these comparisons. London, Los Angeles, Tokyo, Buenos Aires or Mexico City can be correctly stated as having populations that are declining and expanding in recent decades, depending on which city boundaries are used.

Slum dwellers or not?

How accurate are the statistics on slum populations? Previous blogs described the lack of data about what the UN defines as slum dwellers: “those lacking improved water sources, sanitation facilities, durable housing or sufficient living spaces”. According to the UN, between 2000 and 2010, the world’s slum population fell by over 200 million, as they benefited from improvements in one or more of the factors in the UN’s definition.

Many countries are reported to have experienced a large decline in the proportion of their urban population living in ‘slums’. For example, in India the proportion dropped from 54.9% in 1990 to 41.5% in 2000 and 29.4% in 2009. Bangladesh, Egypt, Mali, Indonesia, Ghana and Nigeria reported large declines too.

But this apparent progress is hard to reconcile with other sources that highlight the growth of informal settlements in these and many other nations. It seems it was simply a change in the definition of a ‘slum’ that produced the dramatic fall in slum populations.

What needs to be done?

Many other important definitions need careful scrutiny. For instance whether carbon dioxide emissions are measured by production-based or consumption-based definitions (the latter greatly increases emissions attributed to wealthy people and economies).

Will the SDGs continue to use definitions that only show great progress because they are based on standards that are far too low?

Was this page useful to you?