Group causation theories and deterrence of tortious acts

Schweizer, Urs Article — Published Version Group causation theories and deterrence of tortious acts European Journal of Law and Economics Provided in Cooperation with: Springer Nature Suggested Citation: Schweizer, Urs (2025) : Group causation theories and deterrence of tortious acts, European Journal of Law and Economics, ISSN 1572-9990, Springer US, New York, NY, Vol. 59, Iss. 3, pp. 555-570, https://doi.org/10.1007/s10657-025-09838-y This Version is available at: https://hdl.handle.net/10419/330777 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ Vol.:(0123456789) European Journal of Law and Economics (2025) 59:555–570 https://doi.org/10.1007/s10657-025-09838-y Group causation theories anddeterrence oftortious acts UrsSchweizer1 Accepted: 31 March 2025 / Published online: 22 April 2025 © The Author(s) 2025 Abstract This paper examines a model with multiple actors, each of whom faces a binary action choice. The action choice imposes harm on a victim that depends on the number of actors who enter or, alternatively, who deviate from a due care standard. The paper examines incentive and welfare effects from per-capita liability under simultaneous as well as sequential action choice. The challenge is to cope with due care standards that, for whatever reasons, need not be efficient. As it turns out, welfare is enhanced in any equilibrium as compared to the situation where none of the actors enter or, alternatively, where all of them meet the due care standard. The range of parameter values, however, where a welfare maximizing outcome in equilibrium is induced by per-capita liability turns out to be rather limited. As an alternative rule, efficient per-capita liability is proposed that would lead to an efficient outcome quite generally for simultaneous as well as sequential action choice. Keywords Group causation· Per-capita liability· Efficient deterrence· Inefficient due care standards· Sequential action choice Mathematics Subject Classification K13 1 Introduction 1.1 Scope andfindings To hold an actor liable, there should be a causal link between his tortious act and the victim’s harm. In some cases, the but-for test establishes such a link. With multiple tortfeasors, however, this test turns out being too restrictive. To ensure The author has benefitted from extensive discussions with J. Shahar Dillbary on issues related to the present paper and he is grateful to Murat Mungan as well as to the editor and two reviewers for very helpful comments on an earlier version of this paper. * Urs Schweizer schw[email protected] 1 University ofBonn, Bonn, Germany 556 European Journal of Law and Economics (2025) 59:555–570 compensation of the victim and to deter tortious acts nonetheless, courts take resort to group causation theories such as concerted action, concurrent causes, substantial factors, alternative liability, overdetermined harm and multiple sufficient causes. In many such cases, the resulting damages rule is per-capita liability where negligent actors expect to share liability for the victim’s harm in equal parts. Under joint and several liability, the victim can request the full harm from any one of the tortfeasors. It then becomes the responsibility of the tortfeasors to sort out their respective proportions of liability and payment. For such cases, the German Civil Code, Section426, explicitly states that joint and several debtors are obliged in equal proportion in relation to one another. Thus, in Germany, percapita liability follows directly from the law. In a private conversation, Shahar Dillbary has told me that, under concerted action and alternative liability, there is no real way to apportion and, for that reason, per-capita is the way out. Moreover, when liability is joint and several, the defendant could recover from any one of multiple tortfeasors and, in contrast to German law, the paying party had no right to go after the others. Nonetheless, if the defendant picks a tortfeasor randomly, in expected terms, each defendant would perceive per-capita liability in this case as well. From the incentive perspective, expected damages only matter. In any case, Dillbary (2016) questions the justification of group causation theories and per-capita liability in terms of deterrence. In fact, dilution of liability may encourage group wrongdoing. Rather than deterrence, Dillbary offers a justification in terms of welfare. He views per-capita liability as a sorting device. Actors are deterred from engaging in welfare decreasing accidents but they are encouraged to engage in welfare increasing ones. To support his claim, Dillbary offers the following two numerical examples. The first one refers to the benefits from driving tortuously. A number N of actors consider whether to engage in a drag race. Each expects a benefit b=40 from participation. Independent of the number of tortfeasors, the harm to the victim is assumed to be H=90 . Since negligent actors share liability equally, each of n tortfeasors is liable for 90/n and, hence, the net benefit of each tortfeasor amounts to 𝜋n=40 −90∕n . There is dilution of liability because individual liability decreases with the number n of tortfeasors in the range n>0 . Moreover, since 𝜋1=−50 <𝜋 2=−5<0<𝜋 3=10 , it follows that, for N≤2 , no actor would participate. For N>2 , however, it is a Nash equilibrium when not a single actor is deterred from participation in the drag race. From a welfare perspective, this is not worrisome as, in fact, welfare is at its maximum when all N actors participate. Note, however, it would also be a Nash equilibrium when all actors are deterred. In fact, for a given actor, if no other actor participates, it is a best response for him not to participate either because, otherwise, his payoff would be negative, 𝜋1<0 . Therefore, in this example, while the efficient outcome is always a Nash equilibrium, a second equilibrium coexists that features over-deterrence. The second example has three tortfeasors who lean on a car causing the car to fall over the edge of a cliff. The car is destroyed, its value being H=90 . It is assumed 557 European Journal of Law and Economics (2025) 59:555–570 that any tortfeasor alone would not have exerted enough force to push the car over the edge but any two tortfeasors would have. Based on the but-for test, there would be no causal link between any single actor’s wrongdoing and the victim’s harm as the same level of harm would have occurred even if, ceteris paribus, this actor had not been leaning on the car. In the car example, the but-for causation requirement would prevent compensation of the victim entirely,1 As a substitute, the suitable group causation theory would offer the required causal link. It seems to be consensus that leaning on the car is a tortious act and that each of the three tortfeasors is liable for H∕3=30 . Nevertheless, actors may be tempted to lean negligently on the car as it comes with a private benefit b=40 for each tortfeasor. When all three are leaning on the car, each tortfeasor ends up with payoff 𝜋3=b−H∕3=10 . This constitutes a Nash equilibrium as no single actor could gain from not leaning on the car. In fact, not leaning on the car would deprive her of the benefit b and, while she would escape liability, her net payoff would be zero. Hence, all three leaning on the car is a Nash equilibrium and so per-capita liability would deter not a single actor from the wrongdoing. From a welfare perspective, though, the lack of deterrence would be less worrisome. Welfare in this Nash equilibrium amounts to w3=3b−H=30 . When none of them had leaned on the car, harm would have been avoided but there also would be no private benefits and, hence, welfare would be w0=0 . Therefore, while the damages rule does not deter tortious acts, it increases welfare. Note, however, the outcome in this Nash equilibrium falls short of maximizing welfare, i.e., it would not be efficient. Welfare would rather be at its maximum, w1=b=40 , when a single tortfeasor were leaning on the car. This actor would benefit without causing any harm to the victim. Since w1>w3 , the Nash equilibrium with all three leaning on the car is inefficient as it comes with under-deterrence. Note further that the Nash equilibrium in this example too is not unique. The efficient outcome itself coexists as a second Nash equilibrium. The single actor leaning on the car ends up with net payoff b=40 whereas the other two obtain zero. But none of them could unilaterally improve her position because, by a unilateral switch, the payoff would amount to b−H∕2=−5 and would be even less. Therefore a single actor leaning on the car is not only efficient. It even constitutes a second Nash equilibrium. To sum up, in both examples, the efficient outcome is a Nash equilibrium but inefficient equilibria coexist, in one case, with over-deterrence and, in the other case, with under-deterrence. These examples stimulated my curiosity and so I wanted to find out how robust the message drawn from them can be. To find out, the present paper offers a setting which goes beyond numerical examples but remains sufficiently simple to allow dealing with the issues explicitly. There is a finite set of actors and a victim who is affected by the actors’ action choice but remains passive otherwise. Action choice is binary, zero/one. We have 1 To confirm the view that the but-for standard for overdetermined harm is inadequate, see Rethinking Actual Causation in Tort Law Harvard Law Review, June 2017. 558 European Journal of Law and Economics (2025) 59:555–570 two interpretations in mind. First, the decision is whether to participate or not so that zero/one means staying out or entering. Second, zero means being in while meeting the due care standard or being in while deviating from it. For the second interpretation, there is a vast literature on negligence rules that provide efficient precaution incentives. The corresponding contributions have in common that due care standards are specified at efficient due care levels. Per-capita liability as laid out above, however, departs from due care levels that need not be efficient. Nonetheless, in the numerical examples at least, there exists an efficient Nash equilibrium. The compensation principle as established and extended by the present author shows that Nash equilibria under per-capita liability, where deviations from due care standards occur, come with a welfare level not lower as compared with the situation where all had met these inefficient standards. The principle remains silent, though, on the amount, by which welfare is enhanced in equilibrium. The above numerical examples suggest that welfare may be enhanced up to maximal welfare so that the equilibrium outcome would be efficient. Unfortunately, the efficiency claim has limits even in the proposed setting of binary action choice. To allow for determining equilibria and doing welfare comparisons explicitly, it is further assumed that benefits from negligent behavior as well as harm to the victim depend on the number of entrants/tortfeasors only but not their identity. All results apply equally well to the first interpretation where the binary decision is in/out. For notational simplicity, we refer in the text mainly to the first interpretation. In any case, the damages rule is per-capita liability among entrants or among negligent actors. In spite of its simplicity, the setting exhibits a sufficiently rich analytic structure to uncover the limits of the efficiency claim. Within this setting, the main findings are as follows. First, quite generally, the number of entrants in Nash equilibrium with simultaneous action choice is not unique. The lowest and highest numbers of entrants are derived explicitly. Second, under sequential action choice, the number of entrants in subgame perfect equilibrium as obtained by backward induction is (generically) unique. More surprisingly, the number of entrants in the subgame perfect outcome is equal to the highest number of entrants in Nash equilibrium.2 While backward induction leads to a unique outcome, it need not be the efficient one though (if an efficient Nash equilibrium exists at all). Third, except for borderline cases, entry occurs in equilibrium and welfare is strictly enhanced but need not be maximized. Fourth, only if welfare attains its maximum when all actors enter, then the efficient outcome is equal to the Nash equilibrium with the highest number of entrants for sure. While the above findings are of positive nature as per-capita liability, a rule used in legal practice, is examined, the fifth contribution of the paper is of normative nature. A modified version of per-capita liability is proposed that would provide 2 Thus, intuition gained from Cournot versus Stackelberg is thoroughly misleading. The Stackelberg equilibrium under sequential quantity choice, while also being calculated by backward induction, would never be a Cournot (Nash) equilibrium under simultaneous quantity choice. 559 European Journal of Law and Economics (2025) 59:555–570 efficient incentives quite generally. The rule, referred to as efficient per-capita liability, denies liability of entrants (or of negligent actors) as long as harm falls short of a suitably chosen limit. This limit is equal to the harm from the efficient number of entrants. Beyond this limit, some suitably chosen deviators will have to share harm beyond efficient harm in equal parts. Efficient per-capita liability may be at conflict with discrimination issues, at least when actors move simultaneously. When they move sequentially, however, it may be justifiable to discriminate among them according to the stage when it is their turn to decide. In any case, efficient per-capita liability would provide efficient incentives. Inefficient outcomes would not occur, no matter whether action choice is simultaneous or sequential. 1.2 Relevant literature There is a vast literature on damages rules for cases of multiple tortfeasors. A prominent source of a formal analysis would be Kornhauser and Revesz (1989). But there also exist prominent textbook treatments of tort law such as William and Posner (1987) and Shavell (1987, 2004). This literature mainly deals with rules that provide efficient incentives. For negligence rules, this means that due care standards are specified at their efficient level. Miceli and Segerson (1991) examine a setting with two actors who, at the first stage, simultaneously take an entry decision. At the second stage, the actors who have entered simultaneously decide on their activity levels. These authors do backward induction as the present paper does. By insisting on a socially best response in every subgame, however, they find out that the subgame perfect outcome fails to maximize welfare, at least for the damages rule proposed by them. The present paper, in contrast, deals with the subgame perfect outcome, not insisting on efficient replies off the equilibrium path. The main references of the present paper, however, are Dillbary (2013, 2016) where due care standards may be inefficient and where per-capita liability leads to second best problems that, as such, are notoriously more demanding to handle. The compensation principle as developed over time by the present author, however, provides guidance. A first version of the compensation principle was introduced in Schweizer (2005a) for a situation with two actors and in Schweizer (2005b) for multiple actors. These early versions were still confined to efficient reference profiles. The most general version so far can be found in Schweizer (2024). It allows for multiple actors as well as for inefficient reference profiles and covers the simultaneous move game of the present setting. The extension to sequential choice with multiple actor is provided by the present paper. 560 European Journal of Law and Economics (2025) 59:555–570 1.3 Organizational structure The present paper is organized as follows. In Sect. 2, the basic model of binary action choice under per-capita liability is introduced. In Sect.3, the Nash equilibria for simultaneous action choice are identified as functions of the parameter configuration. In Sect.4, backward induction for sequential action choice is carried out. Section5 summarizes the welfare statements and provides numerical examples. In Sect.6, efficient per-capita liability is specified as a rule that would provide efficient incentives quite generally. Section7 concludes. In the appendix, the extension of the compensation principle is established. 2 The model Each actor i out of a finite set I={1, ..., N} chooses an action xi from the set {0, 1} of available alternatives. If actor i decides xi=0 we say that i stays out. If he decides xi=1 we say that i enters. An alternative interpretation would be that all actors are in. In this case, i=0 means that i meets the due care standard while xi=1 means that he deviates. The results hold for both interpretations. For simplicity, the text mainly refers to the first one. An action profile x=(x1, ..., xN) lists exactly one action of each actor. At action profile x, the number of entrants is t(x)=∑i∈Ixi . By assumption, harm H0=0 ≤ Hn to the victim as well as the private benefit bn>0 before damages of each entrant depend on the number 0<n≤N of entrants but not their identity i. The drag race example from the introduction has H0=0<Hn=H=90 for 0<n≤N and the car example has N=3 and H0=H1=0<H2=H3=H=90 . Both examples have b0=0<bn=40 for all 0<n≤N . Under per-capita liability, actors who stay out (or meet the standard) are not liable whereas entrants (or negligent actors) are liable for total harm in equal shares. The victim’s claim is harm Hn when n>0 actors have entered, in which case each entrant owes damages hn=Hn∕n to the victim. With n entrants, welfare amounts to w(0)=0 and, for all 0<n≤N , where 𝜋n=bn−hn denotes the benefit net of damages of each entrant. Under per-capita liability, the payoff of actor i net of damages amounts to as a function of the chosen action profile x. The payoff functions (2) define a game. Rational actors end up, when they choose their actions simultaneously, with a Nash equilibrium of this game or, when actions are chosen sequentially, with a subgame perfect outcome. An action profile is a Nash equilibrium when it consists of mutually best responses. A subgame perfect equilibrium is obtained by backward induction. Formal details are provided in the next sections. (1) w (n)=n⋅b n −H n =n⋅ ( b n −h n) =n⋅𝜋 n (2) 𝜙i(x)=𝜋t(x) ⋅ xi 561 European Journal of Law and Economics (2025) 59:555–570 The victim is not an active player but, as she is fully compensated, her payoff including damages amounts to 𝜙v(x)=0 , no matter, what actions x are chosen. Relative to a given reference profile xo , a damages rule is called compensatory when no actor who keeps to the reference profile suffers from deviations by others, i.e., 𝜙i(xo i,x−i)≥𝜙i(xo) holds for all i and all deviations x −i by the other actors and if the victim is compensated for any deviations, i.e., 𝜙v(x)≥𝜙v(xo) holds for all action profiles. In this sense, per-capita liability is compensatory relative to the reference profile xo=(0, …,0) . By (1), this reference profile is efficient (i.e., welfare maximizing) if and only if max[𝜋1,…,𝜋N]≤0 . In this case, it follows from the compensation principle (see the appendix) that any Nash equilibrium under simultaneous action choice as well as any subgame perfect outcome under sequential action choice must be efficient. In all other cases, i.e., if max[𝜋1,…,𝜋N]>0 , it follows from the compensation principle that any Nash equilibrium and any subgame perfect outcome comes with a higher welfare level than under the reference profile, in our case w(0)=0 . The compensation principle does not tell us by what amount welfare is increased though. For the setting of binary action choice, this issue will be fully settled in the next two sections. 3 Simultaneous action choice A Nash equilibrium in pure strategies consists of an action profile x=(x1,…,xN) so that no actor can improve his payoff by unilaterally deviating from it, i.e., holds for all actors i and for all deviations xi≠xi by actor i. Let x=1 n =(1, …, 1, 0, …,0) denote the action profile where actors i≤n enter but the remaining ones do not. Since net benefits of an entrant depend only on the total number of entrants, the following must be true: if x=1 n is a Nash equilibrium then any action profile x=(x1,…,xN) with the same number t(x)=n of entrants is also a Nash equilibrium. This allows us to say the number n of entrants is Nash when 1 n is a Nash equilibrium. Finite games do not necessarily have a Nash equilibrium in pure strategies. In our model, however, there always exist such equilibria. The next two propositions show how the lowest number nL and the highest number nH that are Nash can be found in a constructive way. Proposition 1 The lowest number nL that is Nash is determined as follows: (i) If 𝜋1≤0 then nL=0 . (ii) If 𝜋n+1≤0<min[𝜋1,…,𝜋n] then nL=n . (iii) If 0<min[𝜋1,…,𝜋N] then nL=N . 𝜙i(xi,x−i)≥𝜙i(xi,x−i) 562 European Journal of Law and Economics (2025) 59:555–570 Proof To prove claim (ii), suppose 𝜋n+1≤0<min[𝜋1,…,𝜋n] and, in particular, 𝜋0=0<𝜋 1 . Then n=0 cannot be Nash. In fact, given profile (0, …,0) where each actor receives zero payoff, any one of them, say, i=1 could deviate ending up with a positive payoff 𝜋1>0 . Therefore (0, …,0) cannot be a Nash equilibrium. If 0<n<n then 𝜋n+1>0 . Such a number n can neither be Nash. In fact, consider the action profile 1n=(1, …, 1, 0, …,0) where actors i≤n deviate but actors j>n do not. Since n+1≤n , it follows that 𝜋n+1>0 and, hence, actor, say j=n+1 , could deviate, ending up with a positive payoff 𝜋n+1 compared with payoff zero under 1n . Therefore no 0≤n<n can be Nash. The profile 1 n , however, is Nash. In fact, any actor i≤n makes positive payoff 𝜋n>0 and would loose from deviating with xi=0 . Actor j>n makes zero payoff. Yet, by deviating with xj=1 , his payoff 𝜋n+1≤0 would not be higher and, hence, 1 n is a Nash equilibrium. Since n<n is not Nash, it follows that n=nL is the lowest number of entrants (deviators) that is Nash. Claim (ii) is established. Claims (i) and (iii) can be established similarly. ◻ This proposition allows for the following welfare comparisons. Under configuration (i), nL=0≤n∗ is obviously true and, hence, this Nash equilibrium need not maximize welfare. Under configuration (ii), welfare is enhanced as w(n L )=n L⋅ 𝜋 n L >0 holds for sure but it need not maximize welfare. Moreover, cases with over-deterrence, nL<n∗ as well as with under-deterrence, n∗<nL , can occur. Under configuration (iii), finally, n∗≤nL=N is obviously true. In this case, welfare is enhanced, w(n L )=n L⋅ 𝜋 n L >0 , but nL need not maximize welfare. Numerical examples supporting these claims are presented in section5 below. For the next result, to avoid case distinctions of little interest, we assume that parameter configurations are generic in the sense that 𝜋n≠0 holds for all 0<n≤N .3 Proposition 2 (generic parameter configurations) The highest number nH that is Nash is determined as follows: (i) If 0<𝜋 N then nH=N . (ii) If max[𝜋n+1,…,𝜋N]<0<𝜋 n then nH=n . (iii) If max[𝜋1,…,𝜋N]<0 then nH=0 . To economize on space, the proof is omitted. The argument is similar to the one in the Proof of Proposition1. As a first corollary of the above proposition, it is easily seen that, in Nash equilibria with the highest number of deviators, there is under-deterrence as, in fact, n∗≤nH holds in all cases (i)–(iii). Under-deterrence may be strict. Moreover, the proposition allows for the following welfare comparison. In case (i), while welfare is enhanced, w(nH)=N ⋅ 𝜋N>w(0) , it need not maximize welfare, i.e., n∗<nH cannot be ruled out. The same holds true for case (ii). In case (iii), 3 The assumption holds generically in the following sense: when parameter values are drawn from a distribution that has a density function, then the assumption is met with probability one. 569 European Journal of Law and Economics (2025) 59:555–570 holds as follows from the compensation requirement (3) with respect to player i=1 . Given this subgame perfect equilibrium, the subgame perfect outcome xsp =(x sp 1 ,a1[x sp 1]) will evolve. It follows that, in this subgame perfect outcome, 𝜙i(xsp)≥𝜙i(xo) holds for all actors i∈I . By adding up these inequalities, it follows that holds for any subgame perfect outcome xsp and so claim (iii) is established. Claim (iv) then follows from claim (iii) as claim (ii) follows from claim (i). The general version of the compensation principle is fully established. It remains to construct the functions ai[x1,…,xi−1,xi] explicitly by backward induction. At the final stage N, there are no subsequent stages left and, hence, aN[x1,…,xN] consists of the empty list whereas aN− 1[x1,…,x N− 1]=x sp N (x1,…,x N− 1 ) . Suppose ai[x1,…,xi] has already been constructed down to stage i. Take xsp i {x1,…,x i− 1 } as defined above. Then we specify at stage i. By repeating this process backwards, the functions ai[x1,…,xi] are constructed for i=N,N−1, …,1 and we are done. ◻ Funding Open Access funding enabled and organized by Projekt DEAL. The author did not receive any funding for this paper and the research behind it. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/ licenses/by/4.0/. References Dillbary, J. Shahar. (2013). Tortfest. The University of Chicago Law Review, 80(3), 953–1005. Dillbary, J. Shahar. (2016). Causation actually. Georgia Law Review, 51(1), 1–69. Kornhauser, Lewis A., & Revesz, Richard L. (1989). Sharing damages among multiple tortfeasors. The Yale Law Journal, 98(5), 831–884. William, Landes, & Posner, Richard A. (1987). The Economic Structure of Tort Law. Harvard University Press. Miceli, Thomas J., & Kathleen, Segerson. (1991). Joint liability in torts: Marginal and infra-marginal efficiency. International Review of Law and Economics, 11, 235–249. 𝜙1 (x sp 1 ,a 1 [x sp 1 ]) ≥𝜙 1 (x o 1 ,a 1 [x o 1 ]) ≥𝜙 1 (x o) w (xsp)= ∑ i∈I 𝜙i(xsp)≥ ∑ i 𝜙i(xo)=w(xo ) ai−1 [x 1 ,…,x i−1 ]=(x 1 ,…,x i−1 ,x sp i {x 1 ,…,x i−1 },a i [x 1 ,…,x sp i {x 1 ,…,x i−1}]) 570 European Journal of Law and Economics (2025) 59:555–570 Schweizer, Urs. (2005). Law and economics of obligations. International Review of Law and Economics, 25, 209–228. Schweizer, Urs. (2005). The pure theory of multilateral obligations. Journal of Institutional and Theoretical Economics, 161, 239–254. Schweizer, Urs. (2024). Liability for Accidents between Road Users whose Activity Levels are Verifiable. Journal of Institutional and Theoretical Economics: https:// doi. org/ 10. 1628/ jite20240022. Shavell, Steven. (1987). Economic Analysis of Accident Law. Harvard University Press. Shavell, Steven. (2004). Economic Analysis of Law. Harvard University Press. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.