Anecdotal accounts suggest that farmers are turning to older, pre-software-era tractors as a workaround for repair restrictions on newer equipment. To examine whether this preference is real and what drives it, I estimate a random coefficients discrete choice demand model (BLP) using over 97,000 used tractor auction listings. Defining products by brand, horsepower, and era, I estimate farmer preferences for tractor characteristics and price sensitivities as functions of farm income and exposure to repair cost shocks. I find that while older 1970–1999 era tractors command higher market shares, farmers on average prefer newer, higher-horsepower John Deere tractors when controlling for price — suggesting that observed old-tractor market shares reflect price differences rather than an underlying taste for repairability. However, higher-income farmers show a significantly greater willingness to pay for older era tractors than their lower-income counterparts, consistent with older tractors serving a secondary or discretionary equipment role on larger operations. Exposure to repair cost shocks has no statistically significant effect on preferences for older tractors, though this null result may partly reflect limitations of the quinquennial Census data used to construct the shock measure. Own-price elasticity estimates from the preferred LRR market definition are consistent with prior literature, and cross-price elasticities reveal greater substitution toward John Deere across most categories. I discuss how market definition sensitivity and the omission of engine hours and new tractors constrain current counterfactual applications, and outline refinements needed to evaluate right-to-repair policy impacts directly. Download paper here