Beyond local climate: rainfall variability as a determinant of household nonfarm activities in contemporary rural Burkina Faso

Abstract

At the household level, nonfarm activities are thought to help rural poor households buffer against agricultural risks related to local climate variability by providing them with cash to buy food in the case of harvest shortfalls. Over the recent decades, households in rural Sub-Sahara have been found less dependent on land and subsistence agriculture and an increasing number of households here derive their income from nonfarm activities. This study tests the hypothesis that rural households in Burkina Faso have diversified to the extent that they no longer rely on nonfarm activities as a safety net against adverse local rainfall events. Results show that household decisions to participate in the nonfarm economy could not be directly linked with local rainfall events during the study period in the mid-2000s. However, household participation was determined by adverse rainfall conditions in the major staple food production zone of the country, presumably because these caused a rise in food prices. Results also suggested that Burkinabe households adopted a flexible approach to nonfarm participation in terms of locality and plurality, depending on short-term rainfall conditions.

Keywords:

• income diversification
• nonfarm activities
• rainfall variability
• panel analysis
• multilevel analysis
• West African Sahel
• Burkina Faso

Acknowledgements

This study was part of the African Monsoon Multidisciplinary Analysis (AMMA) project. Based on a French initiative, AMMA was created by an international scientific group and was subsequently funded by a large number of agencies, especially from France, the UK, the USA and Africa. It was the beneficiary of a major financial contribution from the European Community's Sixth Framework Research Programme. Detailed information on scientific coordination and funding is available on the AMMA International website: http://www.amma-international.org.

We are grateful to Thierry Lebel of the Laboratoire d'étude des Transferts en Hydrologie et Environnement in Grenoble for providing the longitudinal decadal rainfall data and to Michel Kone of the Institut National de la Statistique et de la Demographie in Burkina Faso for facilitating access to and analysis of the socio-economic survey data.

Notes

1. In this article, ‘nonfarm activities’ refers to all activities that are not agricultural or not local. Activities performed during migration can be agricultural in nature, but figure, for the purpose of this article, in the category ‘nonfarm activities’.

2. The poverty indicator was calculated based on the following assets: building material of the walls (earthen or solid), floors (sand/earth, concrete and tiles), source of drinking water (surface water, pit or tap water), number of rooms (<2, 3, 4 and >4), possession (yes/no) of a radio, bicycle or motorbike and the number of cattle. The final index aggregation function is determined through polychoric Principal Component Analysis (pPCA), which unlike ordinary Principal Component Analysis (PCA), developed for continuous variables, is especially designed to deal with continuous, binary and categorical variables at the same time. Measuring poverty in an African context is challenging due a general lack of data and due the intertwinement of the different causal factors and effects of poverty in general. One of the main additional challenges of deriving poverty indices from existing survey data is that the choice of components is limited. This is predominantly because the surveys from which data are obtained are originally not designed for the measurement of poverty. Another reason is that the index is used for comparing poverty over time. To make comparison more straightforward, only assets common to all years are included, thereby further reducing the choice of variables. Of the many dimensions of poverty suggested in the literature, the index in this paper only explores consumption poverty (asset poverty). The index, therefore, can only be seen as an approximation of wealth/poverty. Moreover, different areas and groups within Burkina Faso may value different kinds of assets; in the north, well-off households may have few consumer durables, as they tend to invest in livestock more than assets, while the opposite may be true in the south. The paper tries to accommodate for this by including both consumer items and livestock possession in the index.