Beyond the Six Veils: Conceptualizing and Measuring Poverty.

• Author: Streeten, Paul
• Position: Review

In this age of progress and economic transformation, a disturbingly vast number of the world's population remain steeped in poverty: The World Bank estimates that the number of the poor, defined as those who live on less than U.S.$1 a day, will continue to rise from a level of 1.1 billion in 1990 to 1.3 billion by the year 2000.(1)

It may be asked why it is important to know the facts about poverty The reason is that what we know, or think we know, enters into our theories, models and policies. Ill-informed or ignorant action can be ineffective or counterproductive and can contribute to the aggravation of poverty Moreover, in an attempt to help one group we may hurt another, or even harm our intended beneficiaries.

One must concede that firm knowledge about poverty is very hard to gain. The fate of the English poor during the Industrial Revolution is still an unsettled issue--even after two hundred years. But action surely cannot wait for the full results of research, not least because proposals to study the issue sometimes serve as excuses for inaction. Clearly, we must act as best we can on whatever inadequate knowledge we possess.

This paper reviews a number of concepts that economists employ to identify the poor, as well as the associated tensions between the use of absolute versus relative measures of poverty and the various statistical indicators, including the construction of a poverty line. The paper will also weigh the pros and cons of several alternative approaches to measuring poverty, including the United Nations' Human Development Index--a single poverty index with considerable political appeal--and efforts to develop strategies that approach the issue of poverty from the perspective of self-assessment, life expectancy and social exclusion.

I will argue that the process of measuring poverty hinges on the removal of six veils, the lifting of which reveals different, yet important, aspects of the task. This process allows us to focus on three sets of indicators that hold the most promise for identifying, understanding and assisting the world's poor: real income, expenditure or consumption per adult-equivalent; calories per adult-equivalent evaluated against an impartial international standard; and the proportion of income or expenditure spent on food.

REASONS FOR UPROOTING POVERTY

The elimination of poverty and the full human development of every person is a worthy end in itself that needs no further justification. We can nonetheless identify four practical reasons why we should wish to eradicate poverty.

First, the elimination of poverty is a means to achieving higher productivity. A well-nourished, healthy, educated, skilled and alert labor force is the most important productive asset in an economy. This has been widely recognized, although it is odd that Hondas, beer and television are often unquestioningly accepted as final consumption goods, while nutrition, education and health services have to be justified on grounds of their productivity.

Second, the eradication of poverty would lower the desired family size, which is generally regarded as desirable. It is paradoxical that a policy that reduces infant mortality and raises health standards should generally lead to lower population growth--one might think that more survivors mean more mouths to feed. But evidence shows that poor people try to over-insure themselves against infant deaths, and that fewer child deaths lead to a lower desired family size. It is true that there is a time lag of about two decades between falling child mortality and lower fertility rates. However, other components of a human development strategy, such as better and longer education of girls, pay off sooner in smaller families.

Third, poverty reduction--and the attendant improvements in human development--are good for the physical environment. The poor are both a cause (though not as large a cause as the rich) and the main victims of environmental degradation. The rich are mainly responsible for global damage to the environment: global warming, the depletion of the ozone layer, etc. The poor tend to destroy their local environment through deforestation, desertification and soil erosion. These latter ills can be curtailed with poverty reduction.

The impact of population growth and population density on the environment is more controversial. The conventional view is that both are detrimental since they increase the demand for land for uses other than crop growing. Overcultivation of land, overgrazing by cattle and deforestation--the results of population growth--lead to desertification. In addition, bad cultivation practices strip vegetation from the topsoil and deprive it of nutrients, thereby exposing it to erosion from sun and wind. Fewer trees also adversely affect the environment. On the other hand, evidence is accumulating that in some environments more people can be a condition for less degradation and more sustainable agriculture. In the Machakos District in Kenya, a fivefold increase in population was associated with a shift from highly degrading to much more sustainable agriculture. In Guinea, more people have created more forests. In Nepal, increased erosion near the forest margins was found to be the result of depopulation, and the collapse of terraces was due to the lack of people to maintain them. In the Kakamega District of Kenya the density of trees varied with the density of population and the smallness of landholdings.(2)

Finally, reduced poverty contributes to a healthy civil society, democracy and greater social stability--at least in the long run. Although the very poor, being too weak, are not a source of rebellion, gross inequalities can lead to social conflict and possibly to civil war.(3)

IDENTIFYING THE POOR

The first question to be asked in a discussion of poverty is how the poor should be identified. The common practice of using fractiles of income recipients--deciles or quintiles or quartiles--has its uses, but also has serious defects.

First, these figures have often been applied to households rather than to persons, or adult-equivalents, and therefore make no allowance for the fact that some households are large while others are small, or that some consist of children and others of older people.(4) There is considerable disparity in ranking incomes, including imputed incomes from home-produced goods and services per household and incomes per head.(5) Large households tend to have low incomes per person but higher income per household. However, ranking by household income approximates ranking by income per adult-equivalent less poorly than by income per person, because larger households tend to have a higher ratio of children to adults. But income per head is a much better measure of poverty than income per household.

Michael Lipton notes that only under two circumstances might income per household be a better measure. First, if indivisible assets (such as a car or furniture), other economies of scale in consumption or forms of joint consumption are important, the division of the flow of services by the number of family members would understate the level of welfare enjoyed by each. However, this is not likely to be important in very poor families. In addition, as already noted, larger households tend to have a higher ratio of children to adults, and therefore their income approximates income per consumer unit more closely than does income per person.(6)

Figures for the ideal measure, income per adult-equivalent, are often not available. Nor is it possible to give precise weights to males and females at different ages for total requirements to attain a given level of welfare.

A second approach looks beyond mere income statistics to consider the causes of poverty, a method which also highlights the deficiency of using fractiles of income recipients. Using this method, the poor may be identified by:

* social and economic class--landless workers, the proletariat, the small peasantry and the former untouchables;

* residence--the rural poor and the urban poor;

* lack of human capital--people with low educational attainments who are stuck in low-paying, insecure jobs and lack access to retraining;

* ethnicity--the tribes in India or the Muslims of Malaysia;

* the region in which they live--frequently the south of the country, or the mountains or areas distant from the capital city;

* the stage they have reached in the age cycle--young families with children, or the elderly;

* vulnerability to employment and capital market discrimination on the basis of race or sex;

* gender--as a result of patriarchal systems of private and public power;

* household structure and position within the household--households with many children and other dependents or single-parent or female-headed families or widows;

* vulnerability to climatic conditions--especially the rainy season, when poverty rates increase; or

* physical or mental limitations--for example, the temporarily or chronically disabled are particularly prone to poverty in many developing countries.

As demonstrated by this long list, the multiple dimensions of poverty, some of which are mutually reinforcing, are not fully captured in income deciles--even when adjusted for adult-equivalents, the size and composition of households, relative price changes, post-tax incomes and free social services.

A third critique of the use of income deciles stems from the fact that knowledge of a household's or individual's share in total income does not tell us how long they have been in that fraction. Perhaps we need not be unduly worried if a student is poor (as long as poverty does not damage his or her health) if we know that he or she will be much better-off later. For example, Stephen Jenkins of Essex University has produced evidence of considerable income mobility in England--only 7 percent of the population remain in the bottom 20 percent of incomes for four consecutive years.(7)

Furthermore...