This paper addresses the role of product liability for the emergence and development

of smart products such as autonomous vehicles (AVs). We analyze how the

liability regime affects innovative activities, as well as the timing of market introduction

and market penetration of such smart products. We develop a dynamic

model in which at each point in time, a potential (monopolistic) innovator decides

on how much to invest in the safety stock of the smart product and on the product

price, once it has been launched. Calibrating the model to the U.S. car market, our

analysis reveals policy-relevant trade-offs when shifting more liability on the producers

of AVs. First, while this improves the safety of AVs in the long run, the safety

stock is accumulated more slowly. Second, it delays the market introduction of AVs,

and also slows down market penetration, which hampers the innovator’s incentives

for safety investments in the short- and intermediate term. As a result, the safety

level of AVs at a given point in time decreases as the liability regime becomes more

stringent. Furthermore, there is a threshold for the innovator’s burden of liability

beyond which she forgoes to develop the AV altogether. Finally, we find that direct

AV safety regulation is welfare-superior compared to a stringent liability regime, as

it induces higher levels of AV safety in the short and intermediate term.