Working Paper: NBER ID: w26220
Authors: Steven Shavell
Abstract: This article studies a model of liability for automobile accidents in the coming world in which automobiles will be autonomous. In that world, travelers will not be drivers, rendering liability premised on driver fault irrelevant as a means of reducing accident dangers. Moreover, no other conventional principle of individual or of manufacturer liability would serve well to do so. Indeed, in the model considered, strict manufacturer liability, recommended by many commentators, would actually tend to leave accident risks unchanged from their levels in the absence of liability. However, a new form of strict liability––the hallmark of which is that damages would be paid to the state––would be superior to conventional rules of liability in alleviating accident risks and would be easy to administer.
Keywords: autonomous vehicles; accident liability; strict liability; tort law
JEL Codes: K13; K2; K32
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Conventional liability based on driver fault (K13) | Irrelevance in a world of autonomous vehicles (L92) |
| Strict manufacturer liability (K13) | No effective reduction in accident risks (R41) |
| New form of strict liability to the state (K13) | Better incentives for risk reduction (G52) |
| Individuals bear full social costs of accidents (J28) | Motivation to demand safer vehicles and moderate mileage (R48) |
| Strict liability to the state (K13) | Individuals incur total costs of 50,000 per accident (G52) |
| Strict liability framework (K13) | Improved manufacturers' incentives to produce safer vehicles (L62) |
| Manufacturers' liability for total harm (K13) | Optimal safety levels and mileage choices (R48) |