What’s behind pro-poor growth? The role of shocks and measurement error

Standard growth incidence curves describe how growth episodes impact on the overall income distribution. However, measuring the pro-poorness of the growth process is complex due to (i) measurement errors and (ii) effect shocks that may hit the percentiles of the income distribution in different ways. Therefore, standard growth incidence curves may misrepresent the true growth process and its distributive impact. Relying on a non-anonymous axiom, we compare actual growth episodes at each percentile of the initial personalized distribution with counterfactual mobility profiles which rule out the presence of shocks. We consider Indonesia in 2000–07 and 2007–14—two growth spells in which there was substantial, significant upward mobility among the initially poorer, a sizeable part of which cannot be explained by unobserved individual endowments or standard socioeconomic attributes. The difference between actual and expected growth can largely be attributed to individual recovery from previous negative losses, rather than resulting from purely exogenous positive shocks.