Before you can start to measure poverty, you first have to decide what you actually want to measure. What is poverty? That’s not just a philosophical problem because depending on the definition of poverty you use, your measurements will be radically different (even with an identical definition, measurements will be different because of different measurement methods).
Among people who measure poverty, roughly 6 different definitions of poverty are used:
- insufficient income
- insufficient consumption spending
- insufficient caloric intake
- food consumption spending above a certain share of total spending
- certain health indicators such as stunting, malnutrition, infant mortality rates or life expectancy
- certain education indicators such as illiteracy.
None of these definitions is ideal, although the first and second on the list are the most widely used. A few words about the advantages and disadvantages of each.
In developed countries, income is a common definition because it’s easy to measure. Most people in developed countries earn a salary or get their income from sources that are easy to estimate (interest payments, the value of houses, stock market returns etc.). They don’t depend for their income on the climate, crop yields etc. Moreover, developed countries have good tax data which can be used to calculate incomes.
In developing countries, however, income data tend to be underestimated because it’s difficult to value the income of farmers and shepherds. Farmers’ incomes fluctuate heavily with climate conditions, crop yields etc. If you ask them one day what their income is, there’s no guarantee that this is a good estimate of their yearly income.
Another disadvantage is that people are generally reluctant to disclose their full income. Some income may have been hidden from the tax administration or may have been earned from illegal activity such as corruption, smuggling, drug trade, prostitution, theft etc. For this reason, using income to estimate poverty means overestimating it.
And, finally, some income may be difficult to calculate (e.g. rising value of livestock).
The main advantage of using consumption rather than income to measure poverty is that consumption is much more stable over the year and over a lifetime (see above). Hence, if you ask people about the level of their consumption, they can just tell you about their current situation, without having to go back in time or to predict the future – which they would have to do if you asked them about income. Their current consumption is likely to be representative of their long term consumption, which isn’t the case for income. This is even more true in the case of farmers who depend on the weather for their income and hence have a more volatile income. If you know that farmers are often relatively poor, then this issue is all the more salient for poverty measurement.
Another advantage of using consumption is that people aren’t as reticent to talk about it as they are about certain parts of their income. It’s also appears that people tend to remember their spending better than their income.
If you want to measure how much people consume, you have to include durable goods and housing. And consumption of those goods is difficult to measure because it’s difficult to value them. For example, if a household owns a house, you have to estimate what it would cost to rent that particular house and add this to the total consumption of that household, at least if you want to compare their consumption to the consumption of the household next door who has to rent its house. And you can’t make poverty statistics if you don’t make such comparisons. Then you have to do the same for cars etc.
Another difficulty in measuring consumption, is that in developing countries households consume a lot of what they themselves produce on the family farm. This as well is often difficult to value correctly.
And finally, different people have different consumption needs, depending of their age, health, work etc. It’s not clear to me how these different needs are taken into account when consumption is measured and used as an indicator of poverty.
Calorie intake: the problem with this is that different people need different amounts of calories (depending on their type of work, their age, health etc.), and that it isn’t very easy to measure how many calories people actually consume.
Food spending as a fraction of total spending: if you say people who spend more than x % of their total spending on food are considered poor, you still have to factor in relative food prices.
Stunting as an indicator of malnutrition and hence of poverty: stunting (height for age) is a notoriously difficult thing to measure.
Some aspects of life tend to be excluded from poverty measurement, even though they have a huge impact on people’s wellbeing. The amount of leisure time people have is perhaps a good indicator of poverty, in certain circumstances (excluding CEOs and US Presidents), but it’s hardly ever counted in poverty measurements.
Another thing: people may have comparable incomes or even consumption patterns, but they may face very different social or environmental conditions: an annual income of $500 may be adequate for people living in a rural environment with a temperate climate where housing is cheap, heating isn’t necessary and subsistence farming is relatively easy. But the same income can mean deep poverty for a family living in a crowded city on the edge of a desert. The presence or absence of public goods such as quality schools, roads, running water and electricity also makes a lot of difference, but poverty measurement usually doesn’t take these goods into account.