1. Policyholders: Policyholders purchase insurance to reduce financial hardships from losses. Policyholders also have contractual expectations for insurance, which may be defeated by courts deciding that public policy bars coverage for an otherwise covered claim, particularly if the policyholder and insurer expressly negotiated for molestation coverage.
2. Insurers: Insurers weigh the additional premiums they receive for providing molestation coverage against the cost of defending and indemnifying those claims. To avoid incurring these costs, insurers are well positioned to negotiate endorsements excluding molestation coverage. Some insurers, for example, add the Insurance Services Office’s “Abuse or Molestation Exclusion,” which many courts have interpreted expansively, in an insurer-friendly manner.
3. Victim: Many victims who bring civil suits against their molester depend on financial recovery through insurance; the molester has insufficient resources to pay an adverse judgment by himself or herself. To secure those proceeds, victims even join coverage suits to secure insurance recovery. By contrast, there is less need for insurance when the molester has the money to pay a civil judgment.
4. Society: Considering the interest of society has two salient goals. The first goal is to reduce the expected harm from molestation. That is the probability of molestation occurring times the harm (or “cost”) of any particular molestation. With that in mind, an important theoretical observation is that insurance for molestation creates the same moral hazard problem seen elsewhere in the insurance industry: Policyholders are more likely to engage in a covered act (here, molestation) if the cost of that act is borne, even in part, by someone else. Thus, insurance against molestation gives policyholders an incentive to molest by reducing one potential, large cost: a suit brought by the victim. Second, again as a theoretical matter, the cost of insurance (i.e., premiums) will increase if insurers are required to provide more coverage for the same insurance terms. Simply put, if insurers provide more coverage, then they will demand more money.
What Courts Are Doing
A survey of cases nationwide reveals that courts frequently engage in an incomplete or truncated public policy analysis to the detriment of all stakeholders in the public good: policyholders, insurers, victims, and society. There is no decision-making model in place, let alone a comprehensive one. By way of a particularly egregious example, in State Farm General Insurance Co. v. White, the Texas Appellate Court held that Texas public policy did not negate the homeowner’s insurer’s duty to defend because the insurer “has the ability to draft policies which expressly excludes [molestation] coverage.” That logic, without more, is a non sequitur. The Texas court ignored two of the stakeholders (the victim and society), opting instead to treat the public policy inquiry like a bilateral, contractual deal. The public policy argument is designed to prohibit insurance for certain conduct despite an otherwise valid contract between the insurer and policyholder. Contracts can create incentives (or remove the disincentives) that result in conduct that imposes terrible harm on parties not privy to the contract.
Failing to acknowledge all constituents is not the extent of the dysfunction. Courts often make empirical conclusions without empirical support. In cases where the molester seeks coverage under a homeowner’s policy, for example, courts summarily assert that permitting insurance for molestation claims would cause premiums to skyrocket, rendering the insurance “too expensive for a majority of the public.”
But where are the data? Insurers, presumably, are well positioned to introduce evidence of their premium rates with or without a molestation exclusion. Insurers also could compare rates in states where public policy bars and does not bar molestation coverage. If the data revealed that coverage for molestation would result in too high premiums, then policyholders could negotiate a lower premium in exchange for the insurer adding a molestation exclusion to the policy.
Despite the limitations, the case law reveals three trends, tacitly elucidating how courts would weigh and balance the interests of the four constituents under a comprehensive public policy framework. First, courts place prominent weight on society’s interest in deterring acts of molestation. With rare exception, courts bar insurance for insured molesters for fear that the insurance was an incentive for their conduct. Second, the type of insurance involved decides whether the policyholder has reasonable contractual expectations for molestation coverage. Third, society has a greater interest in accused molesters having insurance pay their criminal defense fees to maintain their innocence than convicted molesters having insurance pay their civil liability judgments.
Who is the molester? Courts are far more likely to allow insurance where the policyholder is not the molester, but rather, for example, an employer or witness to the abuse. The primary difference between the two categories of cases is consideration of the moral hazard problem.
In cases where the molester seeks insurance, courts—in the vast majority of cases—conclude that insurance increases the probability of molestation occurring by reducing its cost. While perhaps correct in theory, at least one court has questioned this logic in practice, concluding that “[d]enying coverage would not deter any crime.” In Allstate Insurance Co. v. Patterson, the Utah federal court placed the burden on the insurer to establish a link between the homeowner’s insurance and the molestation. The insurer failed to establish that the homeowner’s insurance was procured in contemplation of the molestation by the children of the named insured or that the children even knew about the insurance. Relatedly, note the obvious: Would-be molesters, if acting rationally, already have substantial incentives not to molest; namely, the specter of harsh criminal penalties. For that reason, insurance may have, at best, a marginal effect on the incentives at play, leading once again to the question: Where are the data?
By comparison, when a non-molester seeks insurance, such as the molester’s purportedly negligent employer, courts diminish the moral hazard problem or find it inapplicable. In American Justice Insurance Reciprocal v. Cates, the South Dakota federal court differentiated between the public policy implications of insurance for a molester and the implications of insurance for its negligent employer. The court held that South Dakota public policy permits insurance for the molester’s employer but not the molester himself. Insurance for the employer “would not serve a deterrent effect nor would it punish [the molester]”; thus, the moral hazard problem is “inapplicable” to insurance for the employer, unlike insurance for the molester. Four years later, the Ohio Supreme Court tracked the logic of Cates, holding that Ohio public policy does not prohibit the Roman Catholic Church from recovering insurance for its employee’s molestation.
To say that the moral hazard problem is inapplicable, however, is an overstatement. It appears that courts have not yet considered that the existence of insurance, by reducing the cost of liability, also gives employers an incentive to reduce training, monitoring, and other expenses aimed at reducing incidents of molestation by their employees. Even a Kentucky federal case that left open the prospect that Kentucky public policy may deny coverage to policyholders who negligently fail to prevent molestation did not raise this moral hazard issue. An empirical analysis of this issue is beyond the scope of this article, yet it is worth noting that there may be a tipping point dollar amount of coverage for each company when the higher expected harm to society from more incidents of molestation caused by reduced training and monitoring outweighs the benefits of insurance to the company.
What type of insurance? The type of insurance naturally determines the reasonable expectations of policyholders. The case law demonstrates (albeit with a small sample size) that policyholders have a reasonable expectation for molestation coverage under general liability and other non-homeowners’ forms of insurance. By comparison, policyholders have minimal expectations for molestation coverage under homeowners’ policies; indeed, many “cringe” at the very suggestion. In contrast to other types of insurance, homeowners’ policies are not designed to insure intentional acts. Courts also highlight that policyholders and insurers “certainly [cannot] have reasonably contemplated the risk of a sexual assault suit when entering into the contract for homeowners’ insurance.”
These homeowners’ insurance decisions reveal that courts emphasize the policyholder’s expectations (or lack thereof) for coverage as part of the public policy analysis, yet perhaps only to support denying coverage. These decisions further reveal that courts place relatively little weight on the victim’s interest in financial recovery as compared those of the other constituents. The Kentucky Court of Appeals expressed this sentiment well: “While the Court is sympathetic to the injuries suffered by the [victims], these sympathies do not justify holding the insurance carrier liable for the sexual abuse[.]”
What duty does the insurer owe? A recent high-profile case involving Gerald Sandusky, a former assistant coach at Pennsylvania State University, highlights an important difference between duty-to-defend and duty-to-indemnify cases. In Federal Insurance Co. v. Sandusky, the Pennsylvania federal court held that Pennsylvania public policy does not permit enforcement of directors and officers (D&O) and employment practices liability insurance policies with Sandusky’s nonprofit organization (under which Sandusky was an additional insured) to the extent they indemnify Sandusky for civil liability arising from his alleged sexual molestation of children. Yet, the court—at a time before Sandusky was convicted—did not also determine whether Pennsylvania public policy relieved the insurer of its obligation to pay Sandusky’s defense costs for his criminal prosecution and civil claims. This aspect of the court’s opinion flowed from Sandusky’s brief, in which he emphasized that D&O policies cover defense costs pending final resolution of guilt (unlike homeowners’ policies). Accordingly, the presumption of innocence is a goal of society that may outweigh the dangers from molestation coverage.
The article proposes a new framework for courts to apply when determining whether public policy allows insurance coverage for molestation claims. Unlike current court practice, this framework considers all constituents affected by the availability (or not) of coverage. Finally, the framework highlights the importance of data to support empirical claims—and, as a corollary, the pitfalls of theory alone.
Keywords: litigation, insurance, coverage, sexual molestation insurance, public policy, empirical analysis
Ravi S. Shankar is with Jenner & Block LLP, Chicago.
 Ravi S. Shankar is an associate in the Insurance Recovery and Counseling Group at Jenner & Block LLP.
 See State Farm Fire & Cas. Co. v. Gregory, 2012 N.J. Super. Unpub. LEXIS 1301, at *4 (N.J. Super. Ct. App. Div. June 8, 2012) (New Jersey public policy barred coverage based in part on the policyholder’s lack of reasonable expectations for molestation coverage, “at least without language far more inclusive than that involved here.”).
 See Allstate Ins. Co. v. Patterson, 904 F. Supp. 1270, 1287 (D. Utah 1995) (“Allstate is always free to write its homeowners policies to exclude coverage for all sexual contacts initiated by an insured. . . .”); Grinnell Mut. Reinsurance Co. v. Jungling, 654 N.W.2d 530, 539 (Iowa 2002) (“[T]he insurer was in a position to include in the policy a specific exclusion. . . .”) (internal brackets omitted); State Farm Gen. Ins. Co. v. White, 955 S.W.2d 474, 477 (Tex. App. 1997) (“State Farm has the ability to draft policies which expressly exclude [molestation] coverage.”).
 See Ravi S. Shankar, “Taking Care and Control to Ensure Abuse and Molestation Coverage,” Coverage, Vol. 23, No. 5 (Sept.–Oct., 2013).
 E.g., Atl. Emp’rs Ins. Co. v. Tots & Toddlers Pre-Sch. Day Care Ctr., Inc., 239 N.J. Super. 276, 278–79, 571 A.2d 300, 301 (N.J. Super. Ct. App. Div. 1990) (victims moved to intervene in the coverage litigation).
 See St. Paul Fire & Marine Ins. Co. v. Shernow, 222 Conn. 823, 835, 610 A.2d 1281, 1287 (1992) (dissenting opinion considers the insured molester’s ability to pay as part of the Connecticut public policy analysis).
 Am. Justice Ins. Reciprocal v. Cates, No. 95-5038, 1996 U.S. Dist. LEXIS 22834, at *16–17 (D.S.D. Feb. 13, 1996).
 State Farm Fire & Cas. Co. v. Brooks, 43 F. Supp. 2d 695, 704 (E.D. Tex. 1998).
 State Farm General Insurance Co., 955 S.W. 2d at 477.
 Brooks, 43 F. Supp. 2d at 704; see also Wilson v. Horace Mann Ins. Co., 2003 Ky. App. LEXIS 61, at *7–8 (Ky. Ct. App. Mar. 21, 2003) (“To find liability on the insurance carrier would subsidize sexual abuse of schoolchildren at the ultimate expense of other insureds to whom the added costs of indemnifying sex offenders will be passed. . . .”).
 E.g., Cates, 1996 U.S. Dist. LEXIS 22834, at *16–17 (“Public policy prohibits individuals from insuring against their own intentional acts because then this punishing and deterrent effect is lost.”).
 Allstate Ins. Co. v. Patterson, 904 F. Supp. 1270, 1287 (D. Utah 1995).
 Patterson, 904 F. Supp. at 1287.
 Patterson, 904 F. Supp. at 1287; see also Doe v. Shaffer, 90 Ohio St. 388, 391–92, 2000-Ohio-186, 738 N.E.2d 1243, 1246 (2000) (prohibiting insurance on Ohio public policy grounds “only for those intentional torts where the fact of insurance coverage can be related in some substantial way to the commission of wrongful acts of that character”) (internal quotations omitted).
 See Patterson, 904 F. Supp. at 1287.
 1996 U.S. Dist. LEXIS 22834, at *16–17 (D.S.D. Feb. 13, 1996).
 American Justice Insurance Reciprocal, 1996 U.S. Dist. 22834, at *17.
 American Justice Insurance Reciprocal, 1996 U.S. Dist. 22834, at *16–17.
 Shaffer, 90 Ohio St. at 395, 2000-Ohio-186, 738 N.E.2d at 1248–49; see also GNFH, Inc. v. W. Am. Ins. Co., 172 Ohio App. 3d 127, 2007-Ohio-2722, 873 N.E.2d 345, at ¶ 56 (Ohio public policy does not bar coverage for negligence claims against a corporation.); Roman Catholic Diocese of Dallas v. Interstate Fire & Cas. Co., 133 S.W.3d 887, 896 (Tex. App. 2004) (same under Texas public policy); Tots & Toddlers, 239 N.J. Super. at 284–85, 571 A.2d at 304 (same under New Jersey public policy).
 W. Am. Ins. Co. v. Embry, No. 04-0047, 2005 U.S. Dist. LEXIS 9387, at *11 (W.D. Ky. Apr. 25, 2005).
 See, e.g., U.S. Fid. & Guar. Co. v. Open Sesame Child Care Ctr., 819 F. Supp. 756, 757 (N.D. Ill. 1993) (multi-peril insurance); Roman Catholic Diocese of Dallas, 133 S.W.3d at 889 (general liability insurance); Shaffer, 90 Ohio St. at 390, 2000-Ohio-186, 738 N.E.2d at 1245 (same).
 Teti v. Huron Ins. Co., 914 F. Supp. 1132, 1142 n.11 (E.D. Pa. 1996); accord Troy v. Allstate Ins. Co., 789 F. Supp. 1134, 1136 (D. Kan. 1992); Hanover Ins. Co. v. Crocker, 1997 ME 19, 688 A.2d 928, 931–32 (Me. 1997).
 Brooks, 43 F. Supp. at 704.
 Wilson v. Horace Mann Ins. Co., 2003 Ky. App. LEXIS 61, at *8 (Ky. Ct. App. Mar. 21, 2003).
 Fed. Ins. Co. v. Sandusky, No. 11-2375, 2012 U.S. Dist. LEXIS 76880, at *11–12 (M.D. Pa. June 4, 2012).
 Sandusky, 2012 U.S. Dist. LEXIS 76880, at *14–15.
 Brief in Opposition at 7, Fed. Ins. Co. v. Sandusky, No. 11 CV-2375, ECF No. 30.