How collective settlements camouflage the costs of shareholder lawsuits.
Corporations insure against liability in shareholder lawsuits by buying tiered coverage from multiple insurers who each cover a distinct segment of the potential damages range. Rather than negotiating to settle individually with the plaintiff, the insurers seek to reach a single, collectively binding settlement agreement. This combination of segmented coverage and collective settlements produces a conflict of interests: the corporation's managers and some insurers are better off if the case settles pre-trial for the expected damages, while other insurers are better off going to trial. To force reluctant insurers to settle, courts have created a duty that can require an insurer to pay its policy amount when the plaintiff makes a settlement demand that exceeds that amount and another insurer or the corporation is willing to pay the rest. This "duty to contribute" biases negotiations toward settlements that overcompensate plaintiffs, thereby encouraging lawsuits of doubtful merit. The conflict of interests in settlement negotiations could be eliminated by allowing defense-side parties (defendants and their liability insurers) to settle separately their respective segments of the damages range. But this "segmented" approach to settlements is contrary to the private interests of managers because it eliminates the justification for the duty to contribute. That duty forces insurers to pay for settlements that they think are excessive or contractually uninsurable, thereby shielding corporate earnings reports--and managers' incentive-based pay--from the costs of shareholder lawsuits resulting from the managers' conduct.
TABLE OF CONTENTS Introduction I. Structural Conflict When Settlements Are Collective A. Liability and Insurance in Shareholder Litigation B. The Settlement-Trial Liability Gap C. The Conflict-Control Devices Now in Use 1. Settlement Demands Within the Policy Limit: The Duty To Settle 2. Demands Above the Cap: The Duty To Contribute II. The Social Costs of Defense-Side Conflict Control A. Plaintiff Overcompensation Under the Duty To Contribute B. Insurer Underspecialization C. Overspending on Defense Lawyers III. Eliminating the Structural Conflict Through Segmented Settlements A. The Mechanics of Segmented Settlements B. Segmentation's Economic Benefits C. Other Reform Proposals: Strict Liability and Vertically Sliced Towers IV. Reform Obstacles: Old-Fashioned Judges and Profit-Smoothing Managers A. Judicial Conservatism and the Duty To Defend B. Collective Settlements as Reputation and Compensation Shields 1. Insurer Resistance Due to Plaintiff or Insurer Overconfidence 2. Insurer Resistance Due to Coverage Exclusions 3. Segmentation's Impact on Towers and Profit Reports C. Reform's First Step: Reversing the Bias of the Duty To Contribute Conclusion Appendix
Liability insurance makes defendants overeager to settle risky lawsuits. If a lawsuit goes to trial, the damages award could be greater than the coverage limit on the defendant's liability policy, forcing the defendant to pay the excess. But the plaintiff will usually be willing to settle before trial for a discounted amount that factors in the possibility of a verdict for the defendant. Settling pre-trial thus compresses the liability burden, increasing the proportion that falls within the insurance policy limit and hence is borne by the insurer. This opportunity to concentrate liability on the insurer can make the defendant better off settling before trial even when the plaintiff's settlement demand exceeds the expected (that is, risk-discounted) damages. And the insurer has the opposite bias: it often is better off going to trial, even when the plaintiff is willing to settle for less than the expected damages.
The same dynamic arises in the more complex situation in which the defendant has multiple liability insurers, as is typical in shareholder lawsuits against corporate managers for securities fraud and breaches of fiduciary duties. To cover the costs of such suits, most public corporations purchase not just one directors-and-officers (D&O) liability insurance policy, but rather a stack of them, forming a so-called insurance "tower." The tower's ground floor is occupied by a "primary" insurer that bears the initial liability in a lawsuit up to its policy limit. Upper floors are occupied by a series of "excess" insurers, each of whose liability begins at the limit of the policy immediately below it in the tower. As liability mounts, the policies are exhausted in succession. When a lawsuit's trial outcome is uncertain, the primary insurer is biased toward trial, the insured defendants are biased toward settling before trial, and the excess insurers divide in their biases based on where the expected damages fall within the tower.
To encourage settlements by insurers that are structurally biased toward trial, courts have read two duties into liability insurance contracts. The first duty requires an insurer to settle before trial when it can do so for a reasonable amount within its policy limit. This "duty to settle" has been thoroughly analyzed by academic commentators. The second duty requires an insurer to contribute its policy amount to a settlement when the plaintiff's settlement demand is above the insurer's policy limit but another defense-side party--that is, another insurer or the defendant--is willing to pay the above-limit portion. This second duty has been ignored by commentators, and indeed there is no conventional name for it: this Article calls it the "duty to contribute." This duty distorts negotiations by permitting "cramdown" settlements in which a party on an upper floor of an insurance tower settles with the plaintiff and then shifts most of the liability to the insurers below. Because cramdown settlements concentrate liability on non-consenting third parties, they will tend to overcompensate plaintiffs, the predictable consequence of a judgemade duty that favors the parties in settlement negotiations who are overeager to settle.
A more efficient way to resolve the conflict of interests in settlement negotiations would be to eliminate the conflict at its source. Courts and commentators have treated the conflict as an inevitable byproduct of insurance policy limits. But the conflict is not inevitable; rather, it arises only when a settlement takes the form of a collective resolution that binds all defense-side parties. The presumption that settlements must be collective is so widespread that it has gone essentially unnoticed. Yet the presumption is worth questioning, as it is the reason that settlement causes some defense-side parties to pay more, and others to pay less, than their expected liability at trial. This divergence between trial burdens and settlement burdens encourages strategic behavior, as it enables defense-side parties to shift liability onto each other when deciding whether to accept or reject a settlement offer.
This conflict of interests in settlement negotiations would be eliminated if each defense-side party were allowed to settle separately its respective slice of the damages range. Under this "segmented" approach to settlements, trial would occur unless all slices settled, and the plaintiff would collect at trial only those awarded damages (if any) that fell within the unsettled slices. Segmenting settlements in this way would eliminate the mismatch between trial liability and settlement liability that encourages strategic behavior. Unlike the judge-made duties to settle and contribute, this segmented approach would address both sides of the conflict: lower-level insurers would no longer be biased against settling, and upper-level insurers and defendants would no longer be biased in favor of settling. With these biases eliminated, the duties to settle and contribute could be abandoned as obsolete. And because the duty to contribute produces cramdown settlements, its elimination would remove the plaintiff-overcompensation hazard.
If segmented settlements really would be more efficient, why have they not been adopted already? Outside the context of shareholder litigation, the most likely obstacle to the segmented approach is liability insurers' traditional "duty to defend." For example, automobile and homeowners liability policies state that the insurer, in addition to covering the policyholder's liability in a lawsuit, will defend the policyholder through trial. Courts might deem an insurer to have breached its duty to defend if it settled out of a lawsuit without obtaining a complete release for the policyholder as well.
D&O policies, by contrast, cover corporate managers who typically disclaim the insurer's duty to defend and insist on controlling their own defenses. Therefore, in shareholder lawsuits the most likely explanation for the persistence of collective settlements is not the duty to defend, but rather the managers' private interests. Managers benefit from the collective approach because it provides a justification for the duty to contribute, which can be used to overcome insurer resistance to settlement. Yet structural bias caused by the collective settlements approach is not the only reason that insurers might resist settling. They also might resist because they think that the plaintiff is demanding more than the lawsuit is worth, or that coverage is likely to be excused at trial by a finding that the defendants deliberately engaged in misconduct. There is no public-policy reason to force insurers to settle when their resistance is based on these alternative grounds. But the duty to contribute does not discriminate, as it forces insurers to settle regardless of their reasons for rejecting a plaintiff's settlement demand.
Under the segmented approach to settlements, there would be no mechanism to force settlements by insurers that believe that the plaintiff is overreaching or that a coverage exclusion is likely. For this reason, the segmented approach would cause the proportion of settlements paid by D&O insurers to fall. The additional liability would be borne not by the managers who defend such suits, but rather by their corporations, which invariably agree to indemnify managers for liability that they incur on the job. But managers would still be averse to the change, because insurance coverage prevents a large settlement payment from causing a drop in the corporation's reported earnings that could draw unfavorable investor attention and reduce managers' performance-based pay. In this way, insurance shields managers who are sued by reducing the volatility of their corporation's reported earnings.
Although insurance coverage for settlements benefits corporate managers, it imposes a host of costs on shareholders. First, by camouflaging the costs of shareholder lawsuits, insurance reduces the usefulness of a firm's reported earnings as a measure of the contribution of that firm's managers to diversified shareholder wealth. Second, by encouraging plaintiff overcompensation, the duty to contribute gives plaintiffs' attorneys an incentive to file lawsuits of doubtful merit. Third, both the duty to contribute and the duty to settle require potentially expensive follow-up lawsuits over whether the duties have been breached. Fourth, cramdown settlements discourage insurer specialization by concentrating liability for both low-risk and high-risk lawsuits on primary insurers. And fifth, collective settlements lead to overspending on defense attorneys, as D&O policies combine coverage for litigation expenses with coverage for liability to discourage insurers from rejecting reasonable settlement offers. Each of the last four costs drives up D&O insurance premiums, which shareholders ultimately pay.
Courts seem unaware that reading a duty to contribute into D&O policies advances managers' interests at the expense of shareholders. Judicial opinions cite the bias of primary insurers against settling but do not mention the countervailing bias that makes defendants and some excess insurers overeager to settle. Courts could increase social wealth by being much leerier of the duty to contribute in shareholder lawsuits. Given that defendants in such cases are sophisticated enough to insist on running their own defenses and negotiating directly with plaintiffs to settle, the justification for the duty to contribute does not apply. And without this mechanism for camouflaging the costs of shareholder lawsuits, managers would be less reluctant to adopt an alternative settlements approach that would increase shareholder profits.
The rest of this Article has four parts. Part I explains the conflict of interests that occurs in settlement negotiations when settlements are collective, and it describes the various legal devices--including the duty to contribute--that have been developed to overcome it. Part II describes the social costs of these various conflict-control devices. Part III explains how the defense-side conflict of interests that courts and commentators attribute to policy limits would disappear if settlements were segmented rather than collective. Part IV describes why, despite the efficiencies of the segmented approach, corporate managers probably prefer the status quo. A brief conclusion and an appendix follow.
I. STRUCTURAL CONFLICT WHEN SETTLEMENTS ARE COLLECTIVE
The conflict of interests among liability insurers and defendants in settlement negotiations is caused by the interaction of two factors: policy limits, which serve valuable economic functions, and the collective approach to settlements, which often does not. Several legal devices have been developed to manage this conflict, but these presuppose--and indeed tend to reinforce--the collective approach, thereby failing to correct the problem at its source.
The conflict of interests is perhaps most conspicuous in the context of D&O insurance, which protects public corporations and their mangers against the most significant source of civil liability they face: shareholder litigation. For this reason, this Part begins by describing the structure of D&O coverage purchased by public firms. Many of the qualitative observations about D&O insurance are, however, also true of other important types of commercial liability coverage.
A. Liability and Insurance in Shareholder Litigation
As Professors Thomas Baker and Sean Griffith describe in an important recent book on shareholder litigation, the most important source of civil liability for American public corporations is the shareholder class action alleging fraud on the securities markets. (1) These lawsuits, the majority of which are brought under Rule 10b-5 (2) of the Securities Exchange Act of 1934, (3) can be highly lucrative for plaintiffs and their attorneys. (4) Between 2003 and 2008, the average settlement payout in 10b-5 class actions was $45 million. (5) And a handful of cases are worth much more: a case against McKesson HBOC settled in 2008 for $1.1 billion. (6) By comparison, the average profits of a Fortune 500 company between 2003 and 2008 were $256 million per quarter. (7) Thus, the average settlement in this period would have reduced the average Fortune 500 corporation's earnings by 18 percent in the quarter in which it was reported--that is, if the settlement were not covered by insurance. (8)
Because of the role that managers play in preparing and reviewing corporate financial reports, virtually all 10b-5 actions name at least one corporate manager as a defendant. (9) As a practical matter, however, the managers rarely bear personal liability in securities class actions. (10) One important reason for the managers' de facto immunity is that almost all public corporations agree to indemnify their managers for liability that they incur on the job. (11) To be sure, general incorporation statutes forbid indemnification when the corporate agent is shown to have acted in bad faith or with wrongful intent. (12) But these indemnification disqualifiers are almost never established, as essentially all 10b-5 claims are either dismissed or settled before trial, (13) and the plaintiffs' attorneys have no financial incentive to insist that the defendants admit wrongdoing in the settlement agreement. Indeed, the incentives go the other way: the corporation usually has much deeper pockets than its managers, giving the managers and the plaintiffs' attorneys a common interest in ensuring that indemnification by the corporation remains available. (14)
To cover the costs of securities actions, virtually all public corporations purchase D&O insurance. (15) Besides covering the managers for their personal liability, the typical D&O policy covers the corporation itself, both for its indemnification obligations to its agents and for any vicarious liability it incurs under the doctrine of respondeat superior. (16) And, in addition to providing liability coverage--that is, coverage for amounts paid to plaintiffs in settlements or judgments--D&O policies provide defense coverage-that is, coverage for defense attorneys and for other litigation expenses. (17)
D&O insurance is expensive in nominal terms. In 2008, public companies with market capitalizations of at least $10 billion paid an average of $2.2 million in D&O insurance premiums. (18) In relative terms, however, this was a small expense--less than 1 percent of these companies' average annual profits. (19) And the companies seem to have gotten a good deal of coverage for their money. A recent study of securities class-action settlements found that D&O insurers paid for the full settlement in 53 percent of cases, and they paid a portion--usually a large portion--of the settlement in 35 percent more. (20) Because directors and officers themselves rarely pay anything in such cases, the amounts not covered by the insurers were paid almost entirely by the corporations directly.
Besides covering liability in securities actions, D&O insurance covers the costs of a second type of shareholder lawsuit: the derivative suit. (21) Like securities class actions, most derivative suits are brought by shareholders against corporate managers. The main difference is that a derivative suit is brought on behalf of the corporation, meaning that the corporation rather than its shareholders recovers any judgment or settlement payment. (22) General incorporation statutes prohibit corporations from indemnifying managers for personal liability in derivative suits, which makes sense given that otherwise the money flow would be circular, with the corporation paying for its own recovery. (23) But corporations can--and almost always do--buy D&O insurance for their managers that covers their personal liability in derivative litigation. (24) In this way, corporations fund their own derivative-suit recoveries ex ante (through D&O insurance premiums) even though they are prohibited from doing so ex post (by reimbursing the defendants for the judgment or settlement). (25)
D&O policies always specify that coverage is unavailable if the defendants are found to have engaged in deliberate fraud or to have enriched themselves at the expense of the corporation. (26) Like the statutory indemnification disqualifiers, these coverage exclusions encourage managers to settle before trial to avoid an adverse finding by a judge or jury that could leave them responsible for their own legal bills.
Even though most D&O policies cover defense costs, the corporate managers rather than the insurers control the defenses of shareholder lawsuits. (27) In this way, D&O policies are different from, for example, automobile and homeowners liability policies, which assign the insurer both the right and the duty to run the defense. (28) One potential reason for this difference is that insurers may be more likely to assert coverage defenses for intentional misconduct in the D&O context, creating a conflict of interests if the insurers were also running the defenses of the lawsuits. Other reasons that corporate managers would rather run their own defenses in shareholder lawsuits are that the managers typically have their own preferred lawyers, and that a public trial could impose reputational costs that, unlike monetary liability, cannot be shifted to the insurer.
The managers' right to run their own defenses entails a right to negotiate directly with plaintiffs to settle. D&O policies provide, however, that the insurer must consent to a settlement agreement to be bound by it. (29) Such provisions are strictly enforced, with courts holding that policyholders forfeit coverage if they settle without first seeking the insurer's permission. (30) This does not mean, however, that the insurer has an absolute veto right, as the insurer can be held liable for rejecting a settlement offer that a court later decides was reasonable. (31)
As with other types of liability insurance, D&O policies always come with coverage caps, known as policy amounts or limits. (32) Policy limits make liability insurance marketable, as a policyholder with finite wealth will be unwilling or unable to buy infinite wealth protection. (33) And policy limits also protect insurers against losses they cannot bear in a cost-effective manner. Thus, insurance creates economic value by enabling the risk-averse policyholder to incur a certain cost (the insurance premium) in exchange for protection against the risk of a much larger, uncertain loss. The insurer, in turn, pools this risk with uncorrelated risks from other policies that it sells, thereby building a diversified portfolio of contingent liabilities whose overall performance is predictable. (34) But an insurer may be unable to diversify against the risk of an especially large loss on a particular policy and hence may use a policy limit to exclude the risk from its liability portfolio.
A characteristic feature of the D&O insurance market is that most public corporations do not buy all of their coverage from a single insurer. Rather, they buy tiered coverage from several insurers, constructing what is called an insurance "tower." (35) The tower's ground floor is occupied by the primary insurer, which bears the initial costs of a lawsuit up to the primary policy limit. Above the primary insurer is a first-layer excess insurer, which provides coverage for losses greater than a specified amount, known as the "attachment" point. (36) The attachment point typically equals the policy limit of the primary policy, making coverage continuous. (37) Additional layers of excess insurance can be stacked atop the first, creating a column of policies that are exhausted in succession as the costs of a lawsuit mount. Sitting atop the tower are the defendants themselves, who bear any residual liability after all policies have been exhausted. (38) Towers with numerous stories are the norm: in 2008, public companies with market capitalizations of at least $10 billion owned an average of seven full layers of D&O coverage. (39)
Why do corporations erect multi-insurer towers when they seemingly could save on transaction costs by consolidating coverage in a single policy? One commonly cited explanation for D&O towers is that corporate managers want more liability coverage than any individual insurer is willing to sell. (40) Thus, by covering only a portion of a company's potential liability, an insurer avoids the risk of a large loss that the insurer cannot easily hedge through diversification. In this way, towers are a substitute for reinsurance, an arrangement in which an insurer sells a policyholder the desired amount of coverage but then purchases from a "reinsurer" its own coverage for some or all of its liability on the primary policy. (41) The reinsurer is comparable to an excess insurer, as it accepts the risk of losses beyond those that the primary insurer is willing to bear. The converse explanation for insurance towers is that the policyholders want to diversify their coverage to protect themselves against the risk that an insurer will fail. Evidence for this explanation is the fact that corporate managers who are responsible for choosing their companies' D&O insurers rank "financial strength" among their most important selection criteria. (42)
Starting in Part II, this Article advances two additional explanations for D&O towers--one benign, the other worrisome. The benign explanation is that towers permit insurers to specialize by focusing their coverage on discrete aspects of litigation risk. Specialization in this form could create social wealth by enabling insurers to reduce their operating costs, a benefit that would translate into lower premiums.
The second, and more troublesome, explanation for insurance towers this Article proposes is that towers create settlement conflicts among insurers that serve the interests of corporate managers. Due to the settlement obligations that courts place on liability insurers, these conflicts increase the likelihood that the settlement of a shareholder lawsuit will be paid by the defendant corporation's D&O insurers rather than by the corporation itself. Shifting liability for settlements to insurers reduces the earnings volatility of those corporations whose managers are sued. In this way, insurance towers camouflage the costs of shareholder lawsuits resulting from managers' conduct in office, costs that shareholders ultimately bear. While this result is good for the corporation's managers, it harms diversified shareholders by diminishing the value of a firm's reported earnings as an indicator of the amount of shareholder wealth created by that firm's management team. The next section lays a foundation for the discussion of these additional functions of insurance towers by describing how conflicts of interests among defense-side parties arise when settlements are collective.
B. The Settlement-Trial Liability Gap
It is widely recognized among courts and commentators that policy limits introduce a conflict of interests in settlement negotiations. (43) What these observers have not recognized is that policy limits alone are not sufficient to produce this conflict. The other necessary element is a presumption that any settlement will be a collective resolution that binds all defense-side parties. When the trial outcome is uncertain, the collective approach to settlements drives a gap between the distribution of the settlement burden and the distribution of the expected trial liability, biasing some defense-side parties in favor of settling and others against it.
The collective approach to settlements creates a conflict of interests whenever the trial outcome is uncertain and the potential damages--that is, the damages the plaintiff will win if he prevails at trial--exceed the limit of the defendant's primary liability policy. As an illustration, consider a hypothetical lawsuit against an insured defendant with a single, $2M liability policy. (44) Assume that the plaintiff has a 50 percent chance of winning $3M at trial and a 50 percent chance of winning nothing. Assume further that the defendant has enough wealth to pay any above-limit damages. (45) On these assumptions, the actuarially fair settlement amount, meaning the amount of the expected damages, is $1.5M. (46) And settling for $1.5M rather than going to trial would minimize the combined costs to the defendant and the insurer, (47) as trial would entail $1.5M in expected damages plus additional litigation expenses that settlement avoids.
Consider what would happen, however, if the plaintiff offered to settle the case pre-trial for the actuarially fair amount.
The defendant would be happy if this case settled for $1.5M, which is within the policy limit and hence would be paid entirely by the insurer. If the case instead went to trial, there would be a 50 percent chance of a $3M verdict, $1M of which would exceed the policy limit and thus be the defendant's responsibility. The defendant's expected trial liability therefore is $0.5M--that is, the expected damages above the policy limit--plus the additional attorneys' fees and other litigation expenses associated with trial. By contrast, her liability is zero if the case settles for the actuarially fair amount. Thus, as Table 1 indicates, the benefit to the defendant of an actuarially fair settlement relative to trial, in terms of liability only (that is, excluding defense costs), is $0.5M.
To the insurer, on the other hand, trial is the cheaper option in expected value terms. Although settlement would cost the insurer $1.5M, trial presents expected liability of only $1M, equal to the policy amount of $2M multiplied by the 50 percent chance of a verdict for the plaintiff. Thus, the cost to the insurer of a fair settlement relative to trial is $0.5M. Put another way, settling pre-trial for the expected damages instead of going to trial shifts $0.5M in liability from the defendant to the insurer, a shift the insurer will naturally resist.
If the insurance policy covers not just liability to the plaintiff but also defense costs such as attorneys' fees, then the insurer will be more inclined to settle pre-trial, as it will bear the defense's trial expenses. (48) Unless, however, those expenses would be at least $1M, the insurer will still be better off vetoing the actuarially fair settlement offer. (49) In this way, bundling defense coverage with liability coverage only partly corrects the insurer's underincentive to settle.
The same structural conflict that makes the insurer too reluctant to settle makes the defendant overeager to do so. To see why, assume that the defense's trial costs in the same hypothetical lawsuit would be $0.4M and that the plaintiff demands a settlement payment of $2M rather than the actuarially fair $1.5M. Now it is in the combined interest of the defense-side parties to reject the settlement demand, which exceeds the expected damages plus the defense's trial expenses. Yet the defendant individually remains better off accepting the plaintiff's demand instead of going to trial, (50) as the $2M settlement would be fully covered by the insurer and would avoid the risk of an above-limit damages award. (51) If the defendant and plaintiff could agree to enter into a settlement that bound the insurer, they might do so for an amount that overcompensates the plaintiff relative to the expected damages.
If the defendant were a public corporation, it probably would have one or more excess insurance policies in addition to its $2M primary policy. (52) To reflect this possibility, the hypothetical could be changed to assume that the defendant has a $3M excess policy to supplement its $2M primary policy. The consequence of this change would be that the excess insurer would step into the shoes of the defendant as summarized in Table 1, facing the same potential liability and thus being similarly biased toward pre-trial settlement. (53) Stated in general terms, defense-side parties on upper floors of an insurance tower tend to be biased in favor of settling before trial, and parties on lower floors tend to be biased against it. (54)
C. The Conflict-Control Devices Now in Use
At least three legal devices have been developed to address the defense-side conflict of interests over settlements. One device has already been mentioned: the bundling of defense coverage with liability coverage. (55) That device is only partially effective because, as was illustrated by the hypothetical lawsuit summarized in Table 1, even when the insurer bears defense costs there will be cases in which the insurer's expected trial liability is less than its share of a pre-trial settlement for the expected damages.
The other two conflict-control devices are a pair of quasi-contractual duties that are placed on liability insurers. Both duties require insurers to accept certain settlement offers even when doing so is contrary to their private interests. One of these, known as the "duty to settle," arises when a plaintiff would be willing to settle within or at the insurer's policy limit. This duty is straightforward in its implications and has been well analyzed in the academic literature. (56) Much less famous--yet more problematic--is the second duty, which arises when the plaintiff makes a settlement demand above the insurer's policy limit and the defendant or an excess insurer is willing to pay the above-limit portion. The question in such cases is not, strictly speaking, whether the insurer has a duty to settle, but rather whether it must "tender"--that is, contribute--its policy amount in support of a settlement negotiated and partly funded by another defense-side party. Courts have held that the answer often is yes, thereby creating the duty that is termed here the "duty to contribute."
1. Settlement Demands Within the Policy Limit: The Duty To Settle. The standard statement of the duty to settle is that a liability insurer must agree to settle a case if it can do so for a reasonable amount within its policy limit. (57) The penalty for breach is forfeiture of the limit, leaving the insurer responsible for the full damages award at trial plus any other consequential damages. (58) The duty has a contractual basis: while liability policies typically require the policyholder to seek the insurer's consent before settling, they provide that consent will not be "unreasonably withheld." (59) Courts, however, impose the duty to settle even on those insurers that do not assume it explicitly, most often by holding that the duty is an aspect of the covenant of good faith and fair dealing implied in every contract. (60)
To determine whether an insurer breached the duty to settle, courts ask whether "a prudent insurer without policy limits would have accepted the settlement offer." (61) As a matter of theory this approach has some appeal. Recall the earlier hypothetical lawsuit (summarized in Table 1) in which the actuarially fair settlement amount was $1.5M. Settling for that figure was in the combined interests of the defendant and the insurer, as trial entailed that amount in expected damages plus additional litigation expenses. Yet if the insurer were to consider its financial interests alone, it might reject a $1.5M settlement demand, as its policy limit causes its expected trial liability to be only $1.0M. On the other hand, if its policy lacked a limit, the insurer's expected trial liability would be $1.5M, making the insurer better off accepting a settlement demand for that amount to avoid the expenses of trial. (62) Thus, assuming that a court would deem the $1.5M settlement demand reasonable, applying the duty to settle would overcome the insurer's structural disincentive to settle and produce what appears to be the socially preferable result.
Although courts consistently specify that the duty to settle only requires insurers to accept settlement demands that are reasonable, the duty would seem to serve its conflict-correction function even if this qualification were dropped and the duty were triggered by an): settlement demand within the insurer's policy limit. (63) Under this alternative, the prospect of uncapped trial liability would automatically align the insurer's interest with those of its policyholder. For example, if the insurer in the hypothetical case summarized in Table One faced unlimited trial liability, and the defense's trial expenses would be $0.4M, then the insurer would face total expected trial costs (expected damages plus expenses) of $1.9M. In that case the insurer would accept a pre-trial settlement demand of $1.5M but not $2.0M, because if the plaintiff insisted on $2.0M then the insurer would be better off going to trial. In this way, the duty to settle is self-regulating, as it encourages insurers to accept actuarially fair settlement demands but not those in which the plaintiff is overreaching.
Based on similar reasoning, several commentators have concluded that the duty to settle should be converted to a strict-liability rule under which an insurer automatically forfeits its policy limit whenever it rejects a settlement demand within that limit, regardless of whether the demand is reasonable. (64) This version of the rule has two apparent advantages. It removes the structural conflict of interests caused by the policy limit when the settlement demand is within that limit, thereby avoiding strategic bargaining by the insurer and the policyholder. And it further reduces litigation costs by eliminating the need for follow-up lawsuits between policyholders and insurers over the reasonableness of rejected settlement offers. (65)
Despite the apparent advantages of a strict-liability approach, courts continue to subject the duty to settle to a reasonableness standard. (66) In practice, however, many insurers treat any settlement demand within their policy limit as creating a high risk of liability. (67) These insurers probably fear hindsight bias: the duty to settle is litigated only after a plaintiff has won an above-limit damages award, (68) at which point the insurer may find it difficult to convince a judge or jury that the within-limit settlement demand that the insurer previously rejected was unreasonable. Indeed, the California Supreme Court has held that the mere fact of an above-limit damages award supports an inference that accepting a pre-trial demand within the limit would have been "the most reasonable method of dealing with the claim." (69)
2. Demands Above the Cap: The Duty To Contribute. Scholars who have analyzed the settlement conflict among liability insurers and policyholders have focused on the fact pattern that triggers the traditional duty to settle: a settlement offer within the policy limit. Much less has been written about the nature of the conflict when the plaintiff demands more. (70) And next to nothing has been written about how courts actually define insurer duties in above-limit demand cases. (71) One possible explanation for this lack of commentary is that the analysis of such cases seems, at least at first glance, straightforward: the policyholder has paid for only a limited amount of coverage, and so the insurer should be under no presumptive obligation to settle for amounts above that limit. (72) To hold otherwise would effectively read coverage limits out of policies, which ultimately would hurt policyholders by driving up insurance premiums. It is probably for this reason that most judicial statements of the duty to settle are careful to specify that the duty applies only to within-limit settlement offers. (73)
Matters become more complicated, however, when a settlement demand is above the insurer's policy limit but the defendant or an excess insurer is willing to pay the above-limit portion. If the insurer is asked to participate in such a settlement and refuses, what should its liability be, if any? Strictly speaking, this fact pattern is not encompassed by the traditional duty to settle, as the total settlement is greater than the policy amount. Moreover, the consequences of imposing a duty on the insurer in such a case are different in important ways from those of the traditional duty to settle. For these reasons, it is useful to have a separate term for the insurer's obligations in such a case, which is why this Article refers to a distinct "duty to contribute."
The duty to contribute almost never comes up in cases involving personal liability coverage such as automobile and homeowners insurance, (74) as plaintiffs in such cases rarely try to settle for more than the defendant's policy limit. For the typical holder of, for example, a personal automobile policy, the policy itself may be her most valuable recoverable asset: her home may be mortgaged to the bank, and her other personal assets may be trivial. A plaintiff suing such a policyholder for injuries covered by her policy will often be better off demanding no more than her policy amount, thereby permitting the plaintiff to negotiate exclusively with the liability insurer, an experienced litigant who has deep pockets and who is accustomed to reaching into them to end a lawsuit. (75)
Fact patterns that implicate the duty to contribute are more common in lawsuits against corporate defendants, as such defendants often buy tiered liability coverage from multiple insurers. Tiered coverage means that liability above the primary policy limit is borne not by the defendants but rather by another insurer. It is thus predictable that, in lawsuits against defendants protected by insurance towers, the plaintiff often makes a demand that exceeds the primary limit, thereby pulling at least one excess insurer into the settlement negotiation. (76) And the excess insurer and plaintiff may then reach a settlement conditioned on the willingness of the primary insurer, and any other subsituated excess insurers, to contribute their policy amounts. The question of interest then becomes, what happens if one (or more) of these lower-level insurers refuses to participate? One possible outcome is that the proposed settlement falls apart and the case goes to trial. Courts might then be tempted to hold that the logic underpinning the traditional duty to settle extends to this fact pattern as well, making the dissenting insurer liable for the full damages award--including any portion outside its policy limits--if the rejected settlement offer was reasonable. (77) And this is in fact how courts presented with this scenario have ruled. (78)
The second potential outcome of such a settlement negotiation--and the one that seems more common in practice--is that, after the lower-level insurer refuses to participate, the excess insurer that negotiated the settlement pays the full settlement out of its own pocket and then sues the dissenting insurer for the latter's policy amount. In such cases, must the dissenting insurer, which in most cases is the primary insurer, pay over its policy amount to defray the costs of a settlement to which it did not consent? In the decisions that have addressed this question directly, most courts have held that the answer is yes. (79) California courts in particular have consistently held that the dissenting insurer's duty to contribute in such cases is essentially automatic. (80) By contrast, opinions from other jurisdictions hold that the dissenting insurer's duty to contribute (again, not the courts' term) shares a common foundation with the duty to settle and hence arises only if the overall settlement amount is reasonable. (81)
Although none of the decisions recognizing a duty to contribute says so explicitly, one possible rationale for the duty is that it, like the duty to settle, counteracts the structural disincentive of lower-level insurers to accept actuarially fair settlement demands. As an illustration, consider again a hypothetical defendant who has a primary liability policy of $2M and an excess policy of $3M. As before, we will assume that the defendant faces a lawsuit in which the plaintiff has a 50 percent chance of prevailing, but now we will assume that the potential damages are $5M rather than $3M. In contrast with the hypothetical lawsuit summarized in Table 1, the expected damages in this case ($2.5M) exceed the primary policy limit. (82)
To avoid trial expenses, the defense-side parties are collectively better off settling pre-trial for the expected damages of $2.5M. And the plaintiff prefers this result as well, for the same reason. But the primary insurer will resist, as its liability would then be $2M (its policy amount), whereas its expected trial liability is only $1M. (83) To overcome this obstacle to an insured settlement, a court might create a rule under which the primary insurer forfeits its policy limit whenever it refuses to participate in a settlement that requires it to contribute no more than its policy amount. (84) Alternatively, the court could permit an excess insurer who settles the case for a "reasonable" amount to recover the primary policy amount from the primary insurer. In either of these forms, a duty to contribute placed on the primary insurer seemingly moves settlement negotiations toward the socially preferable result.
Professors Baker and Griffith have observed that excess D&O insurers often put settlement pressure on the insurers below them in an insurance tower, especially when the excess insurers are willing to pay the portions of the settlement demand that fall within their slices of the liability range. (85) Court decisions recognizing a duty to contribute explain the source of this pressure. Objecting to settlement is of little advantage to a lower-level insurer if the case will settle anyway and a court then will force the insurer to pay over its policy amount--plus, most likely, interest. And if its objection scuttles the settlement, the insurer may be forced to pay the entire subsequent damages award, including the portion above its policy limit.
Besides counteracting lower-level insurers' bias against settling when the trial outcome is uncertain, the duty to contribute may also ameliorate a problem with collective settlements that can arise even in cases in which the plaintiff has a 100 percent chance of prevailing at trial. Settling such a case pre-trial confers a benefit on the insurers in the tower that are responsible for liability in excess of the lawsuit's potential damages, as it is those insurers, rather than the insurers below them, whose overall liability would be increased by the litigation expenses that settlement avoids. Knowing that settlement reduces the liability of these upper-level insurers, lower-level insurers might strategically withhold their consent to settlement in order to negotiate for a reduced share of the overall settlement burden. This risk of holdouts, which constitutes a classic collective-action problem, increases with the number of insurers in the tower. Strategic negotiating of this type can introduce delay and cause negotiations to collapse. (86) The duty to contribute can be seen as a device for defeating such holdouts and forcing them to pay their contractually assigned share of the settlement burden.
The discussion to this point has considered only the apparent virtues of the duty to contribute, one of which--counteracting lower-level insurers' bias against settling--it shares with the duty to settle. The implication might seem to be that the duty to contribute should be governed by a strict-liability rule rather than a reasonableness standard, just as scholars have advocated for the duty to settle. (87) But drawing this inference would be a mistake. As will be discussed in the next Part, the duty to contribute lacks the duty to settle's self-regulating virtue. Rather than discouraging plaintiff overreach, the duty to contribute rewards it, producing a plaintiff-overcompensation hazard that does not arise under the duty to settle even when that duty is not restricted to "reasonable" settlement offers. This downside of the duty to contribute raises the question whether the better solution to the conflict that arises from above-limit settlement offers would be to eliminate the collective settlement process that is that conflict's source.
II. THE SOCIAL COSTS OF DEFENSE-SIDE CONFLICT CONTROL
The previous Part described how the bundling of liability coverage with defense coverage, the duty to settle, and the duty to contribute all act against an insurer's structural disincentive to settle when the trial outcome is uncertain and the potential damages exceed the insurer's policy limit. None of these conflict-control devices directly addresses the countervailing bias, which is the overeagerness of defendants and upper-level excess insurers to settle pre-trial when doing so would shift liability onto primary and lower-level excess insurers. (88) The implication is that the current set of conflict-control devices increases the proportion of lawsuit settlements paid by lower-level insurers rather than by upper-level insurers and defendants. To the extent that defendants are risk-averse, this shifting of their liability onto insurers may seem efficient. (89) But the conflict-control devices achieve this result by introducing several costly distortions, all of which are byproducts of the practice whereby defense-side parties settle collectively.
A. Plaintiff Overcompensation Under the Duty To Contribute
Although scholarly commentary on insurer-policyholder conflict in settlement negotiations has been extensive, it has overlooked how the duty to contribute systematically encourages settlements that overcompensate plaintiffs. (90) And courts that enforce that duty seem unaware of this hazard as well. Yet the hazard should be heeded, as systematic plaintiff overcompensation will tend to generate a variety of social costs including too many lawsuits, overspending on lawyers, and, ultimately, more expensive liability insurance.
To see how the duty to contribute distorts negotiations in favor of plaintiffs, consider again the hypothetical lawsuit summarized in Table 2, in which the defendant has a $2M primary policy and a $3M excess policy, and the plaintiff has a 50 percent chance of losing at a trial and a 50 percent chance of winning $5M. Rather than assuming, however, that the plaintiff makes an actuarially fair settlement demand of $2.5M, consider what could happen instead if the plaintiff made a more aggressive demand of $3M. In that case, it is cheaper in expected value terms for the defense-side parties to go to trial, at least as long as their trial expenses would be less than $0.5M. (91) But from the individual perspective of the excess insurer, accepting the $3M settlement demand is still better than trial as long as it can use the duty to contribute to force the primary insurer to tender its policy amount of $2M. The excess insurer's net settlement liability would then be $1M, whereas its expected trial liability, as noted in Table 2, is $1.5M (plus potential trial expenses). (92) In this way, the duty to contribute has what we might call a "cramdown" effect: it enables a plaintiff and excess insurer (or defendant) to improve their positions relative to trial by shifting liability down the tower onto the primary insurer. The result in this hypothetical lawsuit is a settlement that overcompensates the plaintiff relative to the expected damages.
A possible objection is that the plaintiff's settlement demand of $3M seems noncredible given that the expected damages are only $2.5M. The excess insurer seemingly could negotiate the plaintiff down to the actuarially fair amount, saving itself $0.5M. By the same logic, however, the excess insurer's rejection of the $3M settlement demand is noncredible given that settling for that amount would cost it only $1M, whereas its expected trial liability is at least $1.5M. The plaintiff therefore should be able to negotiate the total settlement amount up to $3.5M, thereby capturing the $2.5M in expected damages plus the $1M potential benefit to the excess insurer of avoiding trial. In fact, any settlement between $2.5M and $3.5M would be mutually beneficial to the excess insurer and the plaintiff relative to trial. (93) And only at the bottom end of this range is the plaintiff not overcompensated as measured by the expected damages.
What is happening here is that the plaintiff and excess insurer are bargaining to divide between themselves a $1M transfer away from the primary insurer. The transfer equals the difference between the primary insurer's policy amount ($2M) and its expected trial liability ($]M). (94) The transfer is available to the other parties to a collective settlement, which is why the range of settlements mutually beneficial to the plaintiff and excess insurer is $1M wide. Excluding trial expenses, the transfer away from a primary insurer produced by the duty to contribute, and hence the potential amount of plaintiff overcompensation, can be expressed as follows:
(1) If L < D, then T = L - pL
(2) If L [greater than or equal to] D, then T = 0
where T is the transfer, L is the primary limit, D is the potential damages (the damages award if the plaintiff wins at trial), and p is the probability (between 0 and 1) of a verdict for the plaintiff. Equation (1) states that, as long as the potential damages exceed the primary limit, the transfer equals the primary limit minus the primary insurer's expected trial liability. Equation (2) states that no transfer will occur if the potential damages do not exceed the primary limit. (95) In that case the excess insurer (or, in the absence of excess insurance, the defendant) has no financial incentive to participate in a settlement, and thus the duty to contribute will not normally be invoked. (96)
These equations have two worrisome implications from a social-welfare perspective. The first is that the overcompensation hazard is greatest in those cases that are the least meritorious. Our proxy for merit is p, the plaintiff's probability of winning at trial. If we hold L (the primary limit) constant in Equation (1), T (the value transfer) rises as p falls. (97)
The second worrisome implication is that, if the defendant has constructed an insurance tower, then the overcompensation hazard increases with the number of policies in the tower. To see why, we must first broaden the definition of L to signify not the primary policy limit per se but rather the limit of whichever policy in the tower is closest to the potential damages without exceeding them. The equations remain valid with this modification because once any excess insurer or the defendant agrees to a settlement that would fall within its contractually allocated slice of the liability range, all subsituated insurers (the primary insurer plus any lower-level excess insurers) are collectively subject to the duty to contribute, and hence are like a single primary insurer for purposes of calculating the value transfer captured by the settlement.
With this adjustment in mind, two observations lead to the conclusion that the overcompensation hazard rises with the number of policies in the tower. First, note that if the plaintiff's probability of success (p) and the potential damages (D) are held constant, the transfer (T) increases with the policy limit (L) as long as the limit does not reach the potential damages. Put another way, the transfer is largest when the distance between the potential damages and the nearest underlying policy limit is smallest. Second, note that if the defendant faces a set of possible lawsuits with potential damages that are randomly distributed across the liability range encompassed by the tower, then the expected distance between the potential damages in any given suit and the nearest subsituated policy limit decreases as the number of policies in the tower increases. In combination, these observations mean that, if the total damages range covered by the tower is held constant, then the potential for plaintiff overcompensation increases with the number of policies in the tower.
When an excess insurer negotiates with a plaintiff to divide the wealth transfer created by the duty to contribute, which of them will have the stronger bargaining position? At least in the context of shareholder litigation, there are reasons to think that both parties will be under pressure to settle. Because the excess insurer will have sold a diversified portfolio of policies, it may be more risk-neutral than the plaintiff. On the other hand, the plaintiff's settlement decision will likely be made by a plaintiffs' attorney, who may be part of a firm that itself has a diversified set of pending lawsuits. In addition, the excess insurer will probably be under strong settlement pressure from the defendant-managers, who will want the lawsuit settled quickly. Professors Baker and Griffith report that corporate managers desire quick settlements of shareholder lawsuits, and are willing to tolerate higher settlement amounts to attain them, in order to avoid ongoing negative publicity. (98) These observations suggest that plaintiffs and excess insurers typically negotiate on roughly equal footing. If this suggestion is correct, then settlement negotiations indeed are biased strongly toward plaintiff overcompensation, which is avoided only if the excess insurer has such a strong bargaining advantage that it captures the entire value transfer for itself.
A question to be addressed at this point is whether the insurers in a tower could prevent the plaintiff-overcompensation hazard by bargaining among themselves after a lawsuit is brought. For example, in the hypothetical lawsuit summarized in Table 2, could the excess insurer pay the primary insurer to accept responsibility for the excess policy? Doing so would effectively collapse the primary and excess policies into one, rendering the duty to contribute inapplicable and eliminating the cramdown dynamic that leads to plaintiff overcompensation. (99) To transfer its policy to the primary insurer, however, the excess insurer would need permission from the policyholder under the common law rule that forbids a promisor (here, the excess insurer) from delegating its obligations to a third party (the primary insurer) without a "novation" from the promisee (the policyholder). (100) And the policyholder in our hypothetical would have little reason to permit the transfer. Reassignment of policies among insurers would defeat one of the main functions of insurance towers, which is to reduce the policyholder's exposure to each insurer's individual bankruptcy risk. (101) Indeed, if the insurers in a tower could freely reassign liability among themselves, they could mutually profit by transferring all of their liability to the insurer who is least creditworthy, which is the type of result that the novation rule seems intended to avoid. (102)
Alternatively, the primary insurer might try to prevent a cramdown settlement by selling a reinsurance policy to the excess insurer for the full excess-policy amount. Like an outright transfer of the excess policy to the primary insurer, such a reinsurance policy would seemingly consolidate liability in the primary insurer, but without requiring a novation from the policyholder. Reinsurance by itself would, however, only exacerbate the plaintiff-overcompensation risk, as then the excess insurer would be fully insulated from liability and hence have no incentive to refuse any settlement demand at all. For this reason, a cramdown settlement could be avoided only if the excess insurer bought a reinsurance policy from the primary insurer and transferred to the primary insurer full authority to settle on behalf of both insurers. But such a delegation of authority would create a significant liability risk for the excess insurer, as courts have consistently held that an insurer cannot use reinsurance to relieve itself of the duty to settle it owes its policyholder. (103) This rule against delegation of settlement authority is consistent with the novation rule, as both protect the policyholder from actions by an insurer that might deprive the policyholder of the full benefits of the insurance contract.
Going a step further, the excess insurer could try to include a clause in its policy that expressly permitted it to delegate settlement authority to another insurer. But there is reason to doubt that corporate managers would be willing to buy a D&O policy that contained such a clause. As will be discussed in Part IV, managers have a strong incentive to preserve the duty to contribute's cramdown dynamic in order to encourage insurer-covered settlements even when the total settlement amount exceeds the expected damages. Therefore, ex post bargaining among insurers in a tower--that is, bargaining that occurs after a lawsuit is filed--does not seem to be a viable solution to the plaintiff-overcompensation hazard created by the duty to contribute.
Although they seem unable to bargain around cramdown settlements ex post, insurers can adjust to the risk of such settlements ex ante by charging higher premiums. The burden of the higher premiums is ultimately borne by the shareholders of the corporations that buy the insurance. For this reason, corporate shareholders, despite being the plaintiffs in shareholder lawsuits, derive no net benefit from systematic plaintiff overcompensation in such cases. Rather, the lawsuits benefit shareholders only if they deter managerial misconduct in a cost-effective manner, a result that excessive settlements will tend to undermine. (104) And if the damages set by the legal system already overstate the actual costs of the underlying managerial conduct--as is likely the case in securities class actions (105)--then excessive settlements will only exacerbate the problem.
While the actual plaintiffs in shareholder litigation do not generally benefit from excessive settlements, the same cannot be said for their attorneys, who are paid on contingency. (106) This boon for plaintiffs' attorneys comes at the expense of social wealth, as it encourages the filing of marginal lawsuits that the attorneys deem worthwhile only because of the possibility of an excessive settlement. The costs of these suits are like a tax on shareholder returns and hence discourage capital formation.
Importantly, the duty to settle does not produce a plaintiff-overcompensation hazard analogous to that caused by the duty to contribute. If an insurer believes that a within-limit settlement demand is greater than the expected damages plus any trial expenses the insurer would bear, then the insurer is better off rejecting the demand and going to trial, even if by doing so it breaches the duty to settle and hence forfeits its policy limit. (107) This is the duty to settle's self-regulating virtue mentioned earlier: the fact that the duty, even when not subject to a reasonableness standard, cannot be used to force an insurer to settle for more than the insurer's estimate of the lawsuit's value.
As the hypothetical lawsuit summarized in Table 2 illustrates, the duty to contribute does not share this virtue. Thus, whatever the merits of applying a strict-liability rule to the duty to settle, the analysis here shows that applying the same rule to the duty to contribute would be a mistake. Insurers who refuse to participate in above-limit settlements need some means for vindication, as otherwise excess insurers or policyholders will collude with plaintiffs to force the dissenting insurers to pay for settlements that overcompensate the plaintiffs and hence are socially inefficient. Fortunately, some duty-to-contribute precedent holds that the duty is breached only if the settlement amount was reasonable. (108) But this solution is hardly ideal: enforcing this standard requires follow-up litigation over the reasonableness of settlement amounts, which besides being costly may often be distorted by a pro-settlement bias among courts that makes overcompensation more likely.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Introduction to II. The Social Costs of Defense-Side Conflict Control A. Plaintiff Overcompensation Under the Duty to Contribute, p. 1-35|
|Publication:||Duke Law Journal|
|Date:||Oct 1, 2012|
|Previous Article:||Regulatory moratoria.|
|Next Article:||How collective settlements camouflage the costs of shareholder lawsuits.|