In this paper we challenge the presumption that academic tenure is an outmoded institutional form for the small teaching university. Starting from the premise that tenure is granted on the basis of research (reflected in a minimum required number of publications), we argue that tenure has value for a university concerned solely with teaching (as opposed to research) because research enhances human capital and incentives for its accumulation are necessary to improve the quality of faculty teaching over the lifecycle. However, while human capital accumulation and research effort create future value, contracting on either basis is not feasible because neither can be measured objectively. Numbers of publications, the usual proxy for research, meter the desired activity only imperfectly due to randomness in the publication process. In these circumstances, an employment contract that offers tenure, compared with contracts that a) reward only teaching and b) supplement teaching payments with a direct reward for publications, can better generate the optimal level of human capital. The minimum publication requirement of the tenure contract induces the optimal level of research with less variation in expected income, avoiding inefficient behavioural responses to the greater riskiness of a contract rewarding only realised publications.