The aim of the paper is to evaluate the impact of value-webs as an innovation in agricultural production on welfare of cassava smallholders in Nigeria. The estimation procedure involved the alternative process of multivalued treatment models when treatment units have multiple values. The study thus extends previous impact studies which focused on estimating causal effects from binary treatment units. The treatment units were determined from the extent of utilization of cassava which informed the classification of households into value-web groups. Value-web is defined here as a measure of joint linkages of product chains within the cassava system. The determinants of the choice of utilization were also estimated. Results show that value-web groups include non-cassava based households; low-level, middle-level and high-level value web groups at 32.4%, 34.1%, 24.4% and 9.1%, respectively. Resource allocation to cassava, farming experience, and access to improved cassava varieties increased probability of higher value-web activities. The ATE estimated from the model shows significant increases of up to N11, 560.14 (USD 37.9) and N11, 296.57(USD 37.04) in monthly farm income if non-cassava based and low-level web households became high-level web households. Keywords: Cassava, Value-webs, Causal effect, Smallholders, Multivalued treatment JEL: C31; D13; O31, Q12 Acknowledgement : The authors wish to acknowledge the BiomassWeb Project funded by the German Federal Ministry for Education and Research (BMBF) and the Deutsche Gesellschaft f r Internationale Zusammenarbeit (GIZ) GmbH. The Research Fellowship offered by the International Institute of Tropical Agriculture, Ibadan, Nigeria to the first author is gratefully acknowledged. The research is supported by CGIAR Program on Humidtropics and, the Roots, Tubers and Bananas (RTB).