Click here for search results

Measuring Vulnerability

Insecurity is an important component of welfare and can be understood as vulnerability to a decline in well-being. The shock triggering the decline can occur at the micro (household) level, (e.g. illness, death); at the meso or community level (pollution, riots); and/or at the national or international level (national calamities, macroeconomic shocks).

Vulnerability is defined here as the probability or risk today of being in poverty or to fall into deeper poverty in the future. It is a key dimension of welfare since a risk of large changes in income may constrain households to lower investments in productive assets--when households need to hold some reserves in liquid assets--and in human capital. High risk can also force households to diversify their income sources, perhaps at the cost of lower returns. Vulnerability may influence household behavior and coping strategies and is thus an important consideration for poverty reduction policies. The fear of bad weather conditions or the fear of being expelled from the land they cultivate can deter households from investing in more risky but higher productivity crops and affect their capacity to generate income.

Vulnerability is difficult to measure: anticipated income or consumption changes are important to individuals and households before they occur—and even regardless of whether they occur at all—as well as after they have occurred. The probability of falling into poverty tomorrow is impossible to measure, but one can analyze income and consumption dynamics and variability as proxies for vulnerability. Such analysis could be replicated for specific non-monetary variables likely to fluctuate, e.g. health status, weight, asset ownership, etc.

Measuring income and consumption dynamics and variability requires specific types of data (see Data for Measurement for a summary of what can be measured with different sources of data).

  • In countries where only one cross-sectional survey is available, quasi-panel data can sometimes be derived if income and consumption are recorded at different points in time. Even when no quasi-panel components are available, it may be possible to build measures of household vulnerability that rely on the variation within communities or other subgroups; or on external information on the seasonality of prices and production.
  • When two or more cross-section surveys are available, changes and trends in levels and patterns of poverty over time can be analyzed. Repeated cross-sections reveal trends for population groups but do not allow tracking of individuals or households within groups over time. They reveal only net aggregate changes; they would not capture large movements into or out of poverty.
  • Panel data follow the same households over time and relate their patterns of consumption and income to changes in other characteristics. The welfare and income variability of households can be followed only when panel data are available. Panel data allow the analyst to determine factors that underlie mobility; estimate changes at the individual level.
  • Alternatively, qualitative information can complement the picture by allowing the analysis of important aspects of vulnerability, such as households’ participation in informal networks; variation patterns in household income and consumption (e.g. seasonal variations), people’s perceptions of their vulnerability and its determinants, and/or various strategies households put in place to reduce their vulnerability (depletive strategies, diversification, consumption reduction, etc.).

Some measures that can be used as proxies for vulnerability are discussed below: movements in and out of poverty, length of poverty spells, and income variability and mobility.

  • Movements in and out of poverty, entry and exit probability: When two observations in time are available (in a panel or in a cross-section which contains a quasi-panel component), transition matrices can be used to map changes-- improvement or decline--in household welfare. The table below presents an example of a transition matrix depicting the movements in and out of poverty for households in rural Ethiopia between 1989 and 1995. The headcount index of poverty declined from 61 percent to 46 percent. This type of information would be revealed by an analysis based on two cross sections of data. The use of panel data provides a more revealing picture. Despite poverty reduction between the two years, half of those that were poor in 1989 remained poor in 1995 (31 out of 61). The other half of the population which was poor in 1989 had emerged from poverty by 1995, but more than a third of the non-poor in 1989 had fallen in poverty by 1995 (15 out of 39). The data still suggests significant flows in and out of poverty, a sign of vulnerability.

Movements In and Out of Poverty in Rural Ethiopia

When data are available for several periods within the same year, the analysis can also distinguish between seasonal and non-seasonal poverty. Another way to look at flows into and out of poverty is to compute poverty entry and exit rates—the probability that a household enters in, or emerges from, poverty.

  • Length and frequency of poverty spells: When several years of panel data are available, it becomes possible to distinguish households according to the time they spend in poverty and the frequency of their poverty spells. There are many different ways of naming these groups, and we only present one of them. Some households will have a very low probability of falling below the poverty line (some time referred to as the transiently poor) – they are not very vulnerable, even if they do experience poverty every now and then. Others will have a higher probability of falling into poverty (some time referred to as the chronic poor) – they are vulnerable. Some households will typically spend most of their time in poverty and have a very high probability of falling into poverty (can be called the persistently poor) – they are very vulnerable.

In the example from Rural China presented in the table below, households have been classified as “very vulnerable” or “persistently poor” when their income is always below the poverty line; as “vulnerable” or “chronically poor” when their income is on average below the poverty line but sometimes above it; and as “not very vulnerable” or “transiently poor” when their income is on average above the poverty line but sometimes below the line. Table 18 shows that, over the period 1985-1990, 33 percent of households were not very vulnerable, 14 percent vulnerable, and 6 percent very vulnerable. Analysis of the characteristics of these groups would inform on the determinants and correlates of vulnerability and on the policy options. 

Classification of Households in Rural China over 1985-1990 (Percentage):

  • Income variability and mobility: Some households may be on average slightly below the poverty line and experience low income variability—an unskilled wage worker in urban areas, for example. Other households may be on average slightly above the poverty line but experience higher income variability—a rural agricultural household, for example. Standard static poverty analysis might classify the first type of household as poor, and the second as non-poor. However, both types experience some form of poverty, and if the second type of household does not have access to instruments to smooth its consumption, it may need some form of temporary support from the state. By contrast, the first type of household may need a very different type of support on a more regular basis. The first group could be called “non-vulnerable” while the second group is “vulnerable”. The analysis of income variability thus reveals alternative policy options for alternative groups of households.

Related Sites

For information on the measurement of vulnerability, explore the site on Social Risk Management.

Back to Measuring Poverty


Permanent URL for this page: