«JOURNAL OF LAW, ECONOMICS & POLICY VOLUME 10 SPRING 2014 NUMBER 2 EDITORIAL BOARD 2013-2014 Steve Dunn Editor-in-Chief Crystal Yi Meagan Dziura Sarah ...»
The conclusions of the analysis contained herein nevertheless have relevance for evaluating policies aimed at reducing the level of frivolous litigation. To the extent that these policies succeed, they may have the unintended consequence of mitigating the deterrence benefits of the litigation process.
APPENDIX This appendix lays out the details of the theoretical model described in
the text. The following notation will be used:
z = level of the risky activity in which the injurer (defendant) engages;
V(z) = gross value of the activity, where V(0)=0, V’0, and V”0;
x = dollar spending on care by the injurer per unit of the activity;
p(x) = probability of an accident per unit of the activity, where p’0 and p”0;
q = probability of a “piggyback suit” being filed per unit of the activity;
L = harm suffered by a genuine victim in the event of an accident;
k = victim’s cost of filing suit;
C = cost of a trial for victims (plaintiffs);
C = cost of a trial for the defendant;
S = settlement amount.
We first examine the outcome of the model when the defendant can perfectly distinguish genuine and piggyback plaintiffs (the perfect information model). We then turn to the model where the defendant cannot distinguish between the two types of plaintiffs (the imperfect information model).
The Perfect Information Model We examine the players’ decisions in reverse sequence of time. Thus, consider first the settlement–trial decision, assuming that both types of plaintiffs have filed suit. Under a rule of strict liability, genuine plaintiffs will win with certainty and be awarded compensation of L, but since they have to pay their own trial costs, the minimum amount they will accept to settle is L–C 0. We therefore assume that the defendant will make a take-it-or-leave-it settlement offer of S=L–C to all genuine plaintiffs, and they will accept this offer. As for piggyback plaintiffs, they will lose at trial with certainty, and so the defendant, who by assumption can perfectly distinguish them from genuine plaintiffs, will offer S=0 and they will drop their suits rather than go to trial.
Now move back to the filing stage. Since all genuine plaintiffs expect to settle for L–C, they will only file suit if
which we assume is true. Thus, all genuine plaintiffs will file. In contrast, no piggyback plaintiffs will file suit since they do not expect to receive a positive settlement offer.
Finally, consider the optimal care and activity choices of the injurer– defendant. Since the injurer anticipates that only genuinely injured plaintiffs will file suit and all will settle for S=L–C, the expected value of activity is given by
Note that xc* is independent of his level of activity because accident costs are assumed to be proportional to z. Given x, the injurer’s optimal activity level, denoted zc*(x), solves
The Imperfect Information Model We now turn to the outcome of the model when the defendant cannot distinguish between genuine and piggyback plaintiffs. As above, we begin at the settlement–trial stage, where the defendant again chooses between two offers: S=L–C and S=0. Note that the first is a “pooling” offer because it will induce both types of defendants to behave the same way— namely, to accept and settle. In contrast, the second is a “separating” offer because it will induce genuine plaintiffs to opt for trial while piggyback plaintiffs will drop their suits.146 In choosing between these two offers, the defendant faces the following trade-off. On one hand, if she offers the higher amount, both genuine and piggyback plaintiffs will accept, so she avoids trial costs, but she ends up paying a positive amount to piggyback plaintiffs. On the other hand, if she offers zero, any piggyback plaintiffs who filed will drop their suits, but genuine plaintiffs will go to trial, costing the defendant L+C L–C.
In order to derive the equilibrium in this case,147 we need to define two additional variables. Let = probability that the defendant offers a settlement of S=L–C rather than zero, = the probability that a piggyback plaintiff files suit.
(The probability that a genuine plaintiff files suit is one, given (A1).) Note that in the perfect information model, =1 and =0, but that outcome is not possible under imperfect information.
Consider first the defendant’s settlement strategy after a suit is filed by a plaintiff of unknown type. Using Bayes’ rule, she first calculates the conditional probability that the plaintiff is genuine to be, (A5) which depends on his prior choice of care. Note that this expression ranges from p(x)/(p(x)+q)1 when =1 (i.e., all piggyback plaintiffs file with certainty) to 1 when =0 (i.e., no piggyback plaintiffs file). Given (A5), if the defendant offers S=0, his expected cost per suit will be (because piggyback plaintiffs will drop), whereas if he offers 146 Note that it would never make sense for the defendant to offer more than L–C (since both types would settle for the lesser amount), nor would it make sense to offer an amount between 0 and L–C (since genuine plaintiffs would reject it and go to trial, while piggyback plaintiffs would “accept” an offer of 0).
147 The derivation of the equilibrium follows Katz, supra note 20.
Note that the first two lines represent pure strategies, while the third line constitutes a mixed strategy under which the defendant offers L–C with probability and zero with probability 1–.
Now consider piggyback plaintiffs, who must decide between filing and not filing. Prior to filing, their expected return is (L–C )–k, which is strictly positive if =1 (by (A1)), and negative if =0. Their decision rule is therefore
where the first two lines are pure strategies and the third is a mixed strategy.
It turns out that there are two types of equilibria of the settlement–trial subgame. The first (Type 1), occurs when
In this case, = =1 is an equilibrium; that is, all piggyback plaintiffs file suit and the defendant settles all cases for S=L–C. This pure strategy equilibrium occurs when q, the probability of a piggyback suit, is small.
Alternatively, suppose that
In this case, if =1, the defendant’s optimal strategy would be to set =0 by the first line of (A6); that is, offer S=0. But then the optimal strategy of piggyback plaintiffs would be to set =0 (i.e., not file), in which case =1 would be optimal for the defendant. Clearly, no pure strategy equilibrium exists in this case. There is, however, a mixed strategy equilibrium in which piggyback plaintiffs are indifferent between filing and not filing, and 2014] THE IMPACT OF FRIVOLOUS LAWSUITS ON DETERRENCE 341 defendants are indifferent between offering S=0 and S=L–C. From the third lines of (A6) and (A7), this implies that
where the latter condition also makes use of (A5). This mixed strategy (Type 2) equilibrium occurs when q is relatively large.
Care and Activity Choices The injurer’s choice of care (x) and activity (z) will depend on which type of equilibrium he expects to arise in the settlement–trial subgame. If it is a Type 1 equilibrium in which all piggyback plaintiffs file suit and all
cases settle, the injurer’s problem is to choose x and z to maximize the following expected value of engaging in the activity:
In contrast, if the expected equilibrium is of Type 2, all genuine plaintiffs and a fraction * of piggyback plaintiffs will file suit. Of these suits, the defendant offers S=L–C to a fraction *, all of which settle, and S=0 to the remainder, of which only the genuine plaintiffs opt for trial. After making the appropriate calculations, it turns out that the defendant’s expected costs in this case are equivalent to the cost he would incur if only genuine plaintiffs filed suit and all went to trial. Thus, his problem under a Type 2 equilibrium is to choose x and z to maximize the following expected value
Comparison of (A13) and (A16) shows that x1*x2*, while comparison of (A14) and (A17) shows that z1*(x) z2*(x) for any x.
Given these results, we first ask how the defendant’s equilibrium care and activity choices compare to those in the certainty model above. For care, comparison of (A3), (A13), and (A16) shows that xc*=x1*x2*. Thus, the possibility of piggyback suits induces the defendant to take either the same or more care as compared to a world without such suits. For the activity level, comparison of (A4), (A14), and (A17) shows that for any x, zc*(x) is larger than either z1*(x) or z2*(x). Thus, for any level of care, the possibility of piggyback suits reduces the defendant’s activity level compared to a world without such suits.
This section compares the defendant’s equilibrium care and activity choices to the socially optimal choices—that is, the choices that a social planner would choose, assuming it could perfectly distinguish between genuine and piggyback plaintiffs. Since the planner would settle with all genuine plaintiffs, and no piggyback plaintiffs would file suit, the planner’s objective function is
Consider first the choice of care. Comparing (A19) to the conditions for equilibrium care under the perfect information and imperfect information models implies that xs*xc=x1*, but xs* x2*. Thus, the defendant takes less than the socially optimal level of care under the perfect information and Type 1 imperfect information models, but she may take too much or too little care under the Type 2 imperfect information model. (The comparison depends on the relative magnitudes of k and C.) As for the defendant’s activity level, comparison of (A20) to the conditions for equilibrium activity under both models implies that zs*(x)zc*(x), but zs*(x) may be larger or smaller than z1*(x) and z2*(x). Thus, the defendant overengages in the activity in the perfect information model compared to the social optimum, but she may overengage or underengage in the activity in the imperfect information model. These conclusions show that from a 2014] THE IMPACT OF FRIVOLOUS LAWSUITS ON DETERRENCE 343 pure deterrence perspective, the existence of piggyback suits is not necessarily socially undesirable.
Fee-Shifting Rules A switch to the English rule for allocating legal costs, or the imposition of sanctions on frivolous suits that shifts the defendant’s legal fees to the plaintiff, are often proposed as responses to the problem of frivolous litigation. In the context of the current model, these two responses have identical effects and therefore can be examined together under the heading of fee-shifting rules.
The first effect of such a rule on the model is to change the minimum amount that a genuine plaintiff will accept to settle to S=L+k. This is true because if the plaintiff wins at trial (which we assume will happen with certainty for a genuine plaintiff), the defendant will be responsible for both the plaintiff’s trial costs and her (sunk) filing costs. As for piggyback plaintiffs, they will lose at trial and will therefore be responsible for the defendant’s trial costs, which only reinforces their decision to drop their cases if presented with a settlement offer of zero. Thus, in the imperfect information model, the defendant’s choice in the settlement–trial subgame is between offering S=L+k and S=0, and the same two types of equilibria (pure and mixed strategies) exist. After working through the details, we calculate that the pure strategy (Type 1) equilibrium arises if the following condition holds
as the equilibrium probabilities that the defendant offers a positive settlement amount, and that piggyback plaintiffs file suit, respectively. The resulting expected values of the activity to the defendant are
under the mixed strategy (Type 2) equilibrium. Comparing these expressions to those for the imperfect information model above (expressions (A12) and (A13)) shows that the injurer chooses more care and a lower activity level in both types of equilibria compared to the situation without fee shifting. Finally, comparing these expressions to the expression for social welfare in (A18), we further find that injurers invest in efficient care under the Type 1 equilibrium, and too much care under the Type 2 equilibrium, but they engage in too little activity under both equilibria. Generally, therefore, fee shifting results in overdeterrence compared to the social optimum.
NATURE ABHORS A VACUUM AND SO DO LOCAL GOVERNMENTS: BUT
VACANT PROPERTY ORDINANCES GO TOO FAR
Abstract This article evaluates whether vacant property ordinances are a justifiable exercise of a municipality’s police power. Most commentaries about vacant property ordinances reviewed by this author apparently assume the regulations are valid. This article does not make such an assumption, but rather takes a contrary view. The question of validity needs to be raised because hundreds of municipalities have enacted these types of ordinances, which, in this writer’s opinion, will have a negative cumulative effect upon lenders’ costs of doing business and borrowers’ costs for new loans. Regulations are not a panacea for the country’s mortgage and housing crises.
The ordinances have yet to be tested in court, though they should be scrutinized before the government makes the economic situation worse. Enacted in response to the housing mortgage financial crisis, the regulations may not be valid if they were promulgated without substantial evidence that blight—a common justification—actually existed in the municipalities’ respective jurisdictions. The type of ordinance discussed in this article rewrites the terms of mortgage loan agreements such that a borrower’s contractual duties to maintain and keep secure the property are shifted to his mortgage lender and impose strict liability when the lender fails to comply.