EXECUTIVE SUMMARYRaising agricultural productivity to meet growing food demands while increasing the resilience of rain-fed farm systems to climate variability is one of the most pressing contemporary development challenges in Sub-Saharan Africa (SSA). Anchored on the three core principles of minimum tillage (MT), crop residue retention, and crop rotation; conservation agriculture (CA) technologies have been actively promoted over nearly the last two decades as potential solutions to raise farm productivity in the context of increased climate variability. Despite the long CA promotion histories in the region, there is a dearth of evidence of its yield impacts on smallholder farmers' plots and under typical smallholder management practices and conditions. In this paper, we examine the yield effects of CA under smallholder systems in Zambia. In particular, we assess the effects of MT on maize yields under smallholder conditions in Zambia. Maize is the most widely grown smallholder crop in Zambia, while MT is the most prevalent component of CA practiced by smallholders.Whether CA does indeed raise farm productivity under real smallholder farm conditions in SSA has been the subject of intense debates over the recent past. The intensity of these debates is, in large measure, the result of limited and at times contradictory empirical evidence. Indeed, the bulk of the evidence on CA is derived from experimental plots, based on small samples and relies on bivariate mean comparisons. This paper uses nationallyrepresentative survey data from nearly 48,000 smallholder maize plots from 2008-2011 to estimate the ceteris paribus effects of planting basins and ripping on maize yields in Zambia. These data are drawn from the Ministry of Agriculture and Livestock and the Central Statistical Office's crops forecast surveys and are panel at the standard enumeration area (SEA) level for the period under consideration. The large sample and meso-panel structure of the data 1 , respectively, allow the current analysis to benefit from asymptotic properties and thereby obviate the small sample biases, and to exploit panel data methods to control for higher order unobservables that may confound the results.After applying a pooled ordinary least squares-correlated random effects approach to control for time invariant unobserved heterogeneity at the enumeration area level, we find positive maize yield gains from minimum tillage over conventional tillage methods when tillage is done before the onset of the rains (holding other factors constant). When tillage is done before the rains, rip tillage confers average maize yield gains over conventional plow tillage of 577 kg/ha nation-wide, and 821 kg/ha in agro-ecological zones 1 and 2a (the two zones most suitable for CA). Planting basins also increase maize yields relative to conventional hand-hoe tillage when tillage is done before the onset of the rains, but the yield gains are smaller at an average of 191 kg/ha nation-wide, and 194 kg/ha in agro-ecological zones 1 and 2a. A caveat in interpr...