FAQ: How can I measure household income? (Part 2)

Simone Lombardini ICT4D, Methodology, Real Geek

In the second of a two-part blog on measuring household income, Oxfam GB’s Impact Evaluation Advisers explain their approaches to measuring household wealth.

Increasing household income is the final outcome indicator for many development projects. Given this, as Impact Evaluation Advisers, it is unsurprising that we are often asked how to measure it for monitoring or impact evaluation assessments.

Last week, we explained how we are approaching the challenges of measuring household income in Oxfam’s Effectiveness Reviews (our impact evaluations of livelihoods projects) by considering household consumption and household assets and wealth, in order to assess whether our projects had an impact on
household welfare.

We think it is useful to measure both consumption and wealth, keeping in mind the different underlying concepts they represent. While household consumption is considered as a flow variable, more volatile and subject to seasonal fluctuations and shocks, the indicator of household wealth is considered to provide a more long-term measure of household welfare.

In Oxfam’s Effectiveness Reviews, we measure household wealth using a household asset wealth index. As with our work to measure household consumption, this approach is relatively common practice, but we thought it might be helpful to share how the index is constructed:

A questionnaire is given to respondents, in which they are asked to provide information about their household’s ownership of various assets (including livestock and productive equipment), as well as about the conditions of the family’s house, both at baseline and at the time of the survey. This information on asset ownership and housing conditions is used to generate an index of overall household wealth.

The index is generated under the assumption that, if each of the assets and housing characteristics constitute suitable indicators of household wealth, they should be positively correlated with each other. That is, a household that scores favourably on one particular wealth indicator should be more likely to do so for other wealth indicators. A small number of items that are found to have low or negative correlations with the others are therefore not considered to be good wealth indicators, and so are excluded from the index.

A data reduction technique called principal component analysis (PCA) is used to assign weights to the different assets, to capture as much information as possible from the data. In particular, our wealth index is taken directly from the first principal component. Broadly speaking, PCA assigns more weight to those assets that are less correlated with all the other assets, as they carry more information. By contrast, items with more intra-correlation are given less weight.

In order to ensure the same weights are applied to assets for both the recalled wealth index and the wealth index at the time of the survey, the two waves of data are first pooled, before undertaking the PCA procedure.

After the weighting process, data are then reshaped to form two indexes of overall wealth; one based on the recalled data from the baseline, and one based on the household’s situation at the time of the survey. This procedure enables us to assign the same weighting system during the PCA procedure, and allows us to compare changes in the wealth index over time.

Our last step is to normalise the wealth index, by subtracting the mean and dividing it by the standard deviations. This means that the impact of the project can be directly understood as the number of standard deviations by which the project improved wealth. This helps us to compare the results of a specific Effectiveness Review to other similar evaluations.

Please download an example of this Stata Do-file for creating the index. You can also download more examples of Effectiveness Reviews’ questionnaires, along with the anonymised data, by following the link to the UK Data Service from the report’s page below, listed by theme and geographic focus.

We’d be very interested to hear your comments on our approach, and your own approaches to measuring household income. Please let us know in the Comments section below.

 Resilience    Women’s empowerment  Livelihoods
 • Resilience in Mali: Evaluation of increasing food security • Effectiveness Review: Supporting Rural Community Banks, Honduras  • Livelihoods in Colombia: Evaluation of market access and food security in the central region
 • Resilience in Nepal: Evaluation of mainstreaming disaster risk reduction and enhancing response capability  • Women’s Empowerment in Rwanda: Evaluation of women’s economic leadership through horticulture planting material business  • Livelihoods in Ethiopia: Impact evaluation of linking smallholder coffee producers to sustainable markets
 • Resilience in Nicaragua: Impact evaluation of climate change adaptation among small scale producers  • Women’s Empowerment in Pakistan: Impact evaluation of the empowering small scale producers in the dairy sector project  • Livelihoods in Honduras: Evaluation of strengthening small-scale farmers’ agribusiness capabilities
 • Resilience in Niger: Evaluation of improving livelihoods through integrated water resource management  • Women’s Empowerment in Rwanda: Evaluation of women’s economic leadership through horticulture planting material business  • Livelihoods in the Philippines: Impact evaluation of the project ‘scaling up sustainable livelihoods in Mindanao’
 • Resilience in Pakistan: Evaluation of enhancing food security and resilience of small-scale farmers  • Women’s Empowerment in Uganda: Impact evaluation of the project ‘Piloting gender sensitive livelihoods in Karamoja’  • Livelihoods in Somalia: Impact evaluation of community driven livelihood and food security initiatives in Lower and Middle Juba Region
 • Resilience in Thailand: Impact evaluation of the climate change community-based adaptation model for food security project

Read more:


Annie Kelly