The dynamics of farmer's choice for rice irrigation technology in the Kilombero Valley, Tanzania: evidence from reduced farm household model

Smallholder farming households in developing economies make simultaneous decisions about irrigation technology, land allocation, labor deployment, and consumption, decisions that binary adoption models cannot adequately capture. This study addresses this gap by developing and estimating a Reduced Farm Household Model (RFHM) grounded in classical profit maximization, random utility, and technology adoption theories to identify the factors that jointly determine irrigation technology choice between the System of Rice Intensification (SRI) and Traditional Flooding Irrigation Technology (TFIT) among rice farmers in the Kilombero Valley, Tanzania. Using Seemingly Unrelated Regression (SUR) applied to data from 100 purposively and randomly selected households, we simultaneously estimate six reduced-form equations capturing land allocation, labor demand, rice supply, rice demand, demand for market goods, and external income under each technology system. Key findings indicate that household head education and gender are the strongest predictors of SRI adoption, with education increasing land allocated to SRI and male headship facilitating access to irrigable land. Rising land rent discourages SRI adoption more than TFIT, while remittances increase land allocation under both systems. The RFHM framework reveals important behavioral mechanisms including income-mediated dietary shifts and labor-intensive technology trade-offs that discrete choice models overlook. Findings support targeted policies including indicative land rent regulation, gender-inclusive extension services, carbon market for compensating higher SRI labor requirement and income diversification programs to accelerate the transition from TFIT to SRI in water-constrained rice systems.