This paper estimates differentiated effects of household asset ownership on educational outcomes of children ages 6 and above in Tanzania. The paper contributes to the literature by providing a theoretical framework that portrays a mechanism for different assets to have differential effects on child education. We use data from Living Standard Measurement Study – Integrated Surveys on Agriculture (LSMS-ISA) in Tanzania which provides panel data on both household wellbeing and agricultural practices and resources. Use of the LSMS-ISA data allows us to disentangle the complicated relationship between child education and agricultural assets in ways which would not be possible using traditional cross-sectional surveys of either household wellbeing or farm practices. We use the Hausman-Taylor instrumental variable (HTIV) panel-data estimator to efficiently control for time-invariant variables omitted from our specifications while allowing us to identify the effects of fixed controls while correcting for the endogeneity of assets. We find that, controlling for household income, different asset types have opposing effects on child educational outcomes. Household durables and housing quality characteristics have positive effects but agricultural assets have adverse effects on highest grade completed and test scores. We demonstrate that the negative effect of agricultural assets emerges from higher opportunity cost of schooling and that the effect is more pronounced among boys and children from poor households, grain crop farmers, and rural residents.