Accounting for Heterogeneity in Growth Incidence in Cameroon Using Recentered Influence Function Regression

This paper frames growth incidence analysis within the logic of social impact evaluation understood as an assessment of variations in individual and social outcomes attributable to shocks and policies. It uses recentered influence function (RIF) regression to link the growth incidence curve to household characteristics and to perform counterfactual decomposition à la Oaxaca–Blinder to identify sources of variation in the distribution of consumption expenditure in Cameroon in 2001–2007. We find that the sectors of employment and geography are the main drivers of the observed pattern of growth through the structural effect. The composition effect accounts for a greater proportion of the observed variation in the social impact of growth. In particular, that effect tends to reduce poverty while the structural effect tends to increase it. This conclusion is robust with respect to the choice of poverty measures and RIF regression models. An important methodological lesson emerging from this study is that linear and non-linear specifications of the RIF regression lead to qualitatively similar results. Copyright 2013 , Oxford University Press.