The efficiency of comparative causation

Francesco Parisi, Ram Singh

Research output: Contribution to journalArticlepeer-review

12 Scopus citations


Under a rule of comparative causation, an accident loss is shared between a victim and a tortfeasor when the parties are either both negligent (as with comparative negligence) or when they are both non-negligent. When one or the other party is found solely negligent, the negligent party bears the entire accident loss. Comparative causation is the only tort regime that allows sharing of an accident loss between a non-negligent injurer and his non-negligent victim. By allowing the sharing of an accident loss in equilibrium, the comparative causation rule has efficiency and risk-allocation properties that are never observed under conventional negligence or strict liability regimes. With respect to efficiency, the comparative causation rule spreads activity level and R&D incentives between the parties, rather than concentrating them on one or the other party (the residual bearer). With respect to the allocation of risk, comparative causation avoids the all-or-nothing allocations of the residual loss that we observe under other legal rules. In doing so, comparative causation provides a mechanism of mutual insurance of victim and tortfeasor with a more equitable apportionment of the residual loss. In spite of these interesting attributes, the existing literature still falls short of establishing the compatibility of comparative causation with efficiency. In this paper, we show that loss sharing can be achieved while preserving optimal care incentives for both parties.

Original languageEnglish (US)
Pages (from-to)219-245
Number of pages27
JournalReview of Law and Economics
Issue number2
StatePublished - 2010

Bibliographical note

Copyright 2010 Elsevier B.V., All rights reserved.


  • comparative causation
  • loss-sharing
  • negligence
  • strict liability
  • torts

Fingerprint Dive into the research topics of 'The efficiency of comparative causation'. Together they form a unique fingerprint.

Cite this