The measurement of dynamic poverty with geographical and intertemporal price variability: evidence from Rwanda

Little attention has been devoted to the effects of absolute and relative prices variability at local and seasonal levels, for the measurement of living standards in LDCs. In particular, it is not known if a substantial share of welfare or poverty indicators may be the consequence of price differences rather than of differences in living standards across households and seasons. With exogenous poverty lines, we show how the directions of effects of accounting for price variability can be theoretically established for popular poverty indices. With endogenous poverty lines, using data from Rwanda, we show that the composition of the population of the poor can be notably modified by accounting for price variability. The change in aggregate living standards due to price correction is moderate although significant in every quarter, in contrast with the change in poverty which can be considerable. The correction yields generally a larger transient seasonal share of annual poverty. In terms of impact of the price correction on the assessment of poverty, the poverty line or the quarter are generally more influential than the formula of the poverty indicator. Though, poverty indicators giving a high importance to the severity of poverty are more likely to lead to a strong effect of prices. However, the sign of change in poverty indicator is not systematically related with parameters of poverty indices or with poverty lines. These results support the necessity in poverty measurement, of an accurate correction for geographical and seasonal price effects, whose directions cannot be easily predicted. Nonetheless with exogenous poverty lines, for an important class of axiomatically valid poverty indices, the change in poverty with compensation for keeping the harmonic aggregate mean of price indices constant, is always positive.