We use panel data of smallholder farm households from Tanzania to empirically assess a largescale
fertilizer subsidy program in Tanzania with respect to its ability to meet its stated
targeting criteria and the effect of subsidy receipt on both smallholder commercial fertilizer
demand and total fertilizer use. We find that the majority of subsidy recipients met the
targeting criteria in practice in regards to area cultivated to maize and that few of them had
used inorganic fertilizer on maize or rice in the previous five years. However, we also find that
depending on the year, between 25 to 37% of households receiving a fertilizer voucher did not
use it, implying that these households did not gain the experience using fertilizer on maize or
rice as envisioned by NAIVS. We find that receipt of one kilogram of subsidized fertilizer has a
small (0.11 kg) but significant positive effect on smallholder commercial fertilizer demand. This
implies that NAIVS is the only large-scale fertilizer subsidy program in Sub-Saharan African
during the 2008-2014 period that managed to avoid ‘crowding-out’ of smallholder commercial
fertilizer demand at a national level, on average. When we adjust the effect of household
receipt of subsidized fertilizer on total smallholder fertilizer use (given that 25 to 37% of
subsidized fertilizer was not actually used by intended recipients) the adjusted effect of an extra
kilogram of subsidized fertilizer on total fertilizer use in 2012/13 (2008/09) was 0.827 kg (0.697
kg).