In Heimer v. Companion Life Insurance Co.,1×1. 879 F.3d 172 (6th Cir. 2018). the Sixth Circuit purported to rule on the by-now venerable question of “whether a contract should mean what it says.”2×2. Id. at 173. The panel majority answered this question in the affirmative by finding a disputed insurance policy provision unambiguous.3×3. Id. at 174–75. And yet, perhaps to assuage any possible doubt, the majority also deployed the canon of contra proferentem: even if the provision were ambiguous, the panel would still have been bound to construe it exactly as it did — against the drafter.4×4. Id. at 176. Heimer supports the position that contra proferentem is a penalty, rather than a majoritarian, default rule.5×5. A default rule is a rule that courts apply to “fill the gaps in incomplete contracts.” Ian Ayres & Robert Gertner, Filling Gaps in Incomplete Contracts: An Economic Theory of Default Rules, 99 Yale L.J. 87, 87 (1989). Some defaults are called “majoritarian” to indicate that they supply the term most parties “would have wanted.” Id. at 93. Others are called “penalty defaults” to indicate that they supply the term most parties would not have wanted. Id. at 91; see also Eric A. Posner, There Are No Penalty Default Rules in Contract Law, 33 Fla. St. U. L. Rev. 563, 564 (2006). Although penalty defaults usually address inefficient bargaining between parties, the Sixth Circuit deployed contra proferentem to shift the costs of ambiguity away from itself and onto least cost avoiders.6×6. The “least cost avoider” is the party that can most cheaply prevent a socially undesirable outcome. See Guido Calabresi, The Costs of Accidents 135 (1970).
On the night of his accident, plaintiff Beau Heimer, age twenty-two,7×7. Heimer v. Companion Life Ins. Co., No. 15-cv-338, 2016 WL 10932755, at *1 (W.D. Mich. Aug. 12, 2016). was riding a dirt bike with his friends “after nightfall in a farm field.”8×8. Heimer, 879 F.3d at 176 (McKeague, J., dissenting in part and concurring in the judgment). Heimer and his friends had been drinking that night — Heimer’s own blood alcohol content towered at 0.152, nearly twice the legal limit for the operation of a motor vehicle.9×9. Heimer, 2016 WL 10932755, at *1–2. The group had collectively courted this acute state of inebriation to the end of playing a dangerous “game” with simple rules: two players would “hurtle themselves toward each other at high speed . . . to see who might ‘chicken out’ at the last second.”10×10. Heimer, 879 F.3d at 176 (McKeague, J., dissenting in part and concurring in the judgment). The winner would be the rider who stayed his course, the loser the rider who swerved away. Unfortunately, when Heimer took his turn, neither he nor his opponent “chicken[ed] out.”11×11. Id. He suffered “catastrophic injuries” in the resulting collision, treatment for which generated medical bills in excess of $197,333.12×12. Heimer, 2016 WL 10932755, at *1.
Peter Heimer, plaintiff’s father, submitted a medical claim form to defendant Companion Life Insurance, requesting reimbursement for plaintiff’s medical treatment.13×13. Id. Defendant denied coverage “under the health benefit plan exclusion #43,” which barred any compensation for “treatment of any injury or [s]ickness which occurred as a result of a [c]overed [p]erson’s illegal use of alcohol.”14×14. Id. (internal quotation marks omitted). Administrative appeal of this denial failed; defendant had determined that plaintiff’s “use of alcohol directly contributed to his illegal act of operating an off road vehicle while impaired.”15×15. Id. at *1–2 (internal quotation marks omitted). Plaintiff appealed the administrative judgment under the Employee Retirement Income Security Act of 197416×16. Pub. L. No. 93-406, 88 Stat. 829 (codified as amended in scattered sections of 26 and 29 U.S.C.). (ERISA), arguing in the United States District Court for the Western District of Michigan that defendant had denied him coverage in breach of contract17×17. Heimer, 2016 WL 10932755, at *1. : Exclusion 43 did not apply because he had not illegally consumed alcohol on the night of the accident.18×18. Id. at *2.
The district court heard this case as a “narrow issue of policy construction” on the scope of Exclusion 43.19×19. Id. at *1. Judge Neff held for the plaintiff on the grounds that “constru[al of] the unambiguous terms of the policy . . . as they are commonly understood” did not yield a reading of Exclusion 43 that covered any part of plaintiff’s behavior on the night of the accident.20×20. Id. at *2–3. Plaintiff’s consumption of alcohol was not, in any usual sense, “illegal,” as he was drinking neither as a minor nor in violation of a court order.21×21. Id. at *3. Although plaintiff’s operation of the dirt bike was illegal, the unambiguous meaning of the phrase “illegal use of alcohol” did not include any “illegal post-consumption conduct.”22×22. Id. The court distinguished cases that defendant cited as evidence of alternative reasonable meanings by observing that such cases were based on different plan language and decided under deferential standards of review.23×23. Id. at *3–4. Whereas the court decided the interpretive issue in Heimer de novo, the cases defendant cited were decided under an “arbitrary and capricious” standard deferential to the administrative decisions of the insurer. Id. Finally, the court stated that even if the phrase were ambiguous, “any ambiguities in the language of the plan [must] be construed strictly against the drafter of the plan.”24×24. Id. at *4 (alteration in original) (quoting Regents of the Univ. of Mich. v. Emps. of Agency Rent-A-Car Hosp. Ass’n, 122 F.3d 336, 340 (6th Cir. 1997)).
The Sixth Circuit affirmed.25×25. Heimer, 879 F.3d at 176. Writing for the panel, Chief Judge Cole26×26. Chief Judge Cole was joined by Judge Stranch. held that the plain meaning of the phrase “use of alcohol” unambiguously referred to the “act of consuming alcohol,” not “post-consumption conduct.”27×27. Heimer, 879 F.3d at 174. The majority cited three dictionaries as evidence that “[i]n this context” the “ordinary meaning” of the word “use” is “[t]he action of consuming something as food, drink, a drug, etc.”28×28. Id. (second alteration in original) (quoting Use, Oxford English Dictionary (3d ed. 2011)). In rejecting “employ” and “apply” as reasonable alternative meanings of “use,” the court stated that because no “natural English speaker” would substitute either “employ” or “apply” for the word “use” in relation to alcohol, those more expansive definitions were inapposite.29×29. Id. at 175. Therefore, the majority reasoned, given that Heimer’s consumption of alcohol did not itself violate any law, the phrase “illegal use of alcohol” did not accurately describe the conduct that resulted in his injuries.30×30. Id. at 174. The majority rejected precedent that defendant had raised in support of alternative reasonable meanings of “use” on this same basis, arguing that these opinions “[gave] short shrift to the ordinary meaning” and “elided the distinction between ‘use of alcohol’ and post-use conduct.”31×31. Id. at 176.
The majority asserted that the policy language also supported its plain meaning analysis. Because the phrase “under the influence” appeared explicitly in other exclusion provisions, the defendant had demonstrated its capacity to express precisely that ground for exception.32×32. Id. at 175. The decision not to use the words “under the influence” in Exclusion 43, then, was best construed as a deliberate choice not to invoke the concept at all.33×33. Id. Moreover, there was no evidence that “illegal use of alcohol” was a “term of art” that extended to drunk driving.34×34. Id.
Finally, the majority held that, even if the contested phrase were ambiguous, “ordinary contract principle[s]”35×35. Id. at 176. would have required the court to construe any such ambiguity “strictly against the drafter.”36×36. Id. (internal quotation marks omitted) (quoting Regents of the Univ. of Mich. v. Emps. of Agency Rent-A-Car Hosp. Ass’n, 122 F.3d 336, 340 (6th Cir. 1997)). Application of this canon, therefore, would dictate the same result.
Dissenting in part but concurring in the judgment, Judge McKeague argued that the majority had mistakenly characterized Exclusion 43 as unambiguous.37×37. Id. (McKeague, J., dissenting in part and concurring in the judgment). The concurrence relied on different dictionaries and intuitions to find that “use” may include “[t]he application or employment of something . . . for [a] purpose.”38×38. Id. at 177 (alteration and emphasis in original) (internal quotation marks omitted) (quoting Use, Black’s Law Dictionary (10th ed. 2014)). To illustrate, Judge McKeague asked: “In one sense, can we really doubt that a person who has a blood-alcohol level of over twice the legal limit continues to ‘use’ the alcohol in his bloodstream after he ingests it?”39×39. Id. There would, accordingly, be no abuse of language in describing the plaintiff as having “used” alcohol “to play chicken in the same way that one ‘uses’ (abuses) cocaine after he inhales it to get high.”40×40. Id. The concurrence also cited to cases that interpreted the phrase “illegal use of alcohol” to mean engagement in any activity that is illegal when intoxicated.41×41. Id. Finding ambiguity as between this interpretation and the majority’s, Judge McKeague concurred with the majority’s final judgment — “any ambiguities . . . [must] be construed strictly against the drafter,”42×42. Id. (alteration in original) (quoting Regents of the Univ. of Mich. v. Emps. of Agency Rent-A-Car-Hosp. Ass’n, 122 F.3d 336, 340 (6th Cir. 1997)). because the drafter ought to “bear the consequences of its sloppy drafting.”43×43. Id.
Both the majority and the concurrence relied on the canon of contra proferentem,44×44. Graydon S. Staring & Hon. Dean Hansell, Law of Reinsurance § 13:2, Westlaw (database updated Mar. 2018). a doctrine whose role in contract law remains unsettled.45×45. See, e.g., Ian Ayres, Default Rules for Incomplete Contracts, in 1 The New Palgrave Dictionary of Economics and the Law 585, 587 (Peter Newman ed., 1998); Posner, supra note 5, at 578–80. The first question is whether, given that canons and default rules traditionally occupy “separate doctrinal categor[ies],”46×46. Posner, supra note 5, at 579. it is appropriate to analyze contra proferentem as a default rule.47×47. Professor Eric Posner has identified this problem as “probably unimportant,” id. at 579, and, indeed, this topic is not the subject of much academic discussion. Nevertheless, examining this question clarifies what the more active debates about contra proferentem may take for granted. If not, then courts could only weigh its implications against those of other interpretive canons, rather than according it a special, tie-breaking status. Second, there is the more actively contested question of whether the framework of majoritarian defaults or of penalty defaults best accounts for contra proferentem’s distinctive features.48×48. See Ayres, supra note 45, at 587; Posner, supra note 5, at 579–80. Professor Eric Posner argues that contra proferentem is a majoritarian default and that there are, in fact, no true penalty defaults.49×49. See Posner, supra note 5, at 565. Professors Ian Ayres and Robert Gertner, however, identify contra proferentem as a penalty default that “induces the drafter to educate the offeree ex ante about the contract terms or risk being penalized by an unfavourable reading ex post.”50×50. Ayres, supra note 45, at 587 (referring to the work of Ayres and Gertner). In the case of Heimer, contra proferentem functions as a default rule, but not as a majoritarian default. Instead, it is a penalty default that forces least cost avoiders to internalize the costs of ambiguous drafting rather than externalize them through the courts. The weight of the Heimer opinion, then, falls behind Ayres and Gertner in their dispute with Posner. This weight is significant: a case like Heimer that addresses the canon specifically in its area of greatest application — insurance — has particular salience for the doctrine as a whole. It also invites legal scholars and practitioners to consider the judiciary’s growing recognition of its own economic constraints and its consequent willingness to privilege efficiency in its own interactions with private parties over efficiency in the interactions of private parties with each other.
The Heimer court deployed contra proferentem as a default rule and not narrowly as an interpretive canon. Default rules are “off-the-rack”51×51. Ian Ayres, Preliminary Thoughts on Optimal Tailoring of Contractual Rules, 3 S. Cal. Interdisc. L.J. 1, 4 (1993) (internal quotation marks omitted). rules that “fill the gaps in incomplete contracts; they govern unless the parties contract around them,”52×52. Ayres & Gertner, supra note 5, at 87. whereas interpretive canons are “rules of interpretation” that clarify ambiguity in the communicative content of a legal text.53×53. Posner, supra note 5, at 565–66. Tellingly, each Heimer opinion considered contra proferentem only after exhausting all of the available interpretive tools54×54. See Heimer, 879 F.3d at 176; id. at 177 (McKeague, J., dissenting in part and concurring in the judgment). and understood it as the kind of canon that would permit the court to assign a meaning when the “tools of legal interpretation [ran] out,”55×55. William Baude & Stephen E. Sachs, The Law of Interpretation, 130 Harv. L. Rev. 1079, 1110–11 (2017) (identifying contra proferentem as a “closure rule”); see 5 Margaret N. Kniffin, Corbin on Contracts § 24.27 (Joseph M. Perillo ed., 1998). and thereby to stipulate the contract’s “legal content.”56×56. Lawrence B. Solum, Communicative Content and Legal Content, 89 Notre Dame L. Rev. 479, 507 (2013) (defining “legal content” as the “content of [the] legal norms” of a legal text, which norms include “standards, principles, obligations, [and] mandates”). The majority argued for the reasonableness of an understanding of “use,” limited to “consum[ption],”57×57. Heimer, 879 F.3d at 174. whereas the concurrence found a broader understanding reasonable as well.58×58. Id. at 177 (McKeague, J., dissenting in part and concurring in the judgment). The gap in Heimer was the ambiguity between these two definitions of the insurer’s obligations to cover Heimer’s medical costs. The court’s application of contra proferentem was therefore equivalent to the application of a default.
But exactly what kind of a default rule is contra proferentem as applied in Heimer? The domain of the default rule consists of two primary categories: majoritarian defaults and penalty defaults. Majoritarian default rules — rules that default to whatever arrangement the majority of similarly situated parties would prefer — are often most efficient because parties contracting to maximize mutual gains will incur the lowest transactional costs contracting around them.59×59. See Ayres & Gertner, supra note 5, at 93; Alan Schwartz & Robert E. Scott, Contract Theory and the Limits of Contract Law, 113 Yale L.J. 541, 569 (2003); see also Frank H. Easterbrook & Daniel R. Fischel, The Economic Structure of Corporate Law 15 (1991). Penalty default rules, however, are antimajoritarian: they set terms that the majority of parties possess strong incentives to override.60×60. See Ayres & Gertner, supra note 5, at 91. Heimer cuts against the view that contra proferentem functions as a majoritarian default61×61. There is, in fact, a general presumption that contractual default rules are majoritarian. See Omri Ben-Shahar, A Bargaining Power Theory of Default Rules, 109 Colum. L. Rev. 396, 396 (2009). — it is unlikely that a majority of parties, similarly situated to the parties in Heimer, would have agreed upon the terms that the Heimer court imposed.
To understand why the majoritarian default is a poor fit, consider a simplified model in which the population of insurers is homogeneous, but the population buying insurance splits into a high-risk category and a low-risk category. Insurers and buyers must negotiate about whether to adopt a term that specifies whether provisions concerning coverage of injuries suffered as a result of illegal, high-risk activities should be construed broadly or narrowly. The insurer’s optimal strategy is to offer only broad construal. Low-risk customers have no incentive to negotiate for narrow construal, because their preferences for insurance are satisfied by relatively cheap and limited coverage. Offering only broad construal also prevents high-risk customers from exploiting their private knowledge of their high-risk status. If the high-risk customer counteroffers for the narrow construal provision, he signals his high-risk status, collapsing the information asymmetry; the insurer has no incentive to sell narrow construal to a high-risk customer.62×62. See William Samuelson, Bargaining Under Asymmetric Information, 52 Econometrica 995, 1004 (1984). In fact, the insurer will not offer broad construal even at a higher price, because of Professor George Akerlof’s problem of adverse selection: offering coverage for illegal, high-risk activities attracts high-risk customers, but an insurance pool filling up with high-risk customers requires raising premiums, which forces relatively low-risk customers out, which requires raising premiums — a vicious cycle that will force the insurer out of business.63×63. See George A. Akerlof, The Market for “Lemons”: Quality Uncertainty and the Market Mechanism, 84 Q.J. Econ. 488, 492–94 (1970). Stated concretely, individuals attempting to negotiate coverage for drunk-driving injuries do not typically succeed in negotiating for any policy at all.
The penalty default framework better explains the Heimer court’s use of contra proferentem, albeit in a somewhat unusual way. For Ayres and Gertner, the typical penalty default encourages revelation of information that one party would otherwise “strategically withhold” to “increase the total gains from contracting . . . in order to increase [the withholding party’s] private share of the gains.”64×64. Ayres & Gertner, supra note 5, at 94; see also Posner, supra note 5, at 569. Others have linked this “information-forcing paradigm” to the idea of least cost avoidance,65×65. Robert R. Niccolini, The Voidability of Actions Taken in Violation of the Automatic Stay: Application of the Information-Forcing Paradigm, 45 Vand. L. Rev. 1663, 1676 (1992); see also Mark Cantora, The CISG After Medellin v. Texas: Do U.S. Businesses Have It? Do They Want It?, 8 J. Int’l Bus. & L. 111, 129 (2009). the notion that the party that can most cheaply prevent a negative outcome should bear responsibility for doing so.66×66. See Calabresi, supra note 6. If the negative outcome is a market inefficiency driven by information asymmetry, then the least cost avoider is the party that possesses the informational advantage. As Ayres argues, by forcing the repeat player — the insurer — to rectify this asymmetry during negotiations, contra proferentem efficiently discourages opportunistic behavior.67×67. Ian Ayres, Ya-Huh: There Are and Should Be Penalty Defaults, 33 Fla. St. U. L. Rev. 589, 597–98 (2006).
In Heimer, however, it is the insured, not the insurer, who primarily benefits from concealing information relevant to the transaction.68×68. David Miller worries that contra proferentem actually encourages inefficient behavior on the part of buyers. David S. Miller, Note, Insurance as Contract: The Argument for Abandoning the Ambiguity Doctrine, 88 Colum. L. Rev. 1849, 1863 (1988). Furthermore, because of the expectation that a driver’s insurance premiums will rise significantly as a result of accidents caused by drunk driving, it is implausible that any ambiguity in Exclusion 43 systematically profited the insurance company.69×69. See Frank A. Sloan et al., Effects of Tort Liability and Insurance on Heavy Drinking and Drinking and Driving, 38 J.L. & Econ. 49, 50 (1995). Reflection on these points suggests drafter “negligence” rather than “opportunism.”70×70. See George M. Cohen, The Negligence-Opportunism Tradeoff in Contract Law, 20 Hofstra L. Rev. 941, 969 (1992) (explaining that “negligence and opportunism can both be viewed as costly types of behavior that . . . should be deterred”). In this limited context, Posner is correct that “[t]here is a mismatch between the model [of information-forcing defaults] and the function of [contra proferentem].”71×71. Posner, supra note 5, at 580.
To clarify the way in which contra proferentem fits into the penalty default framework, one must introduce a third party into the least cost avoidance analysis — namely, the court itself. On this point, Ayres and Gertner relate private parties to the courts in the context of market-level efficiency, observing that “ex ante contracting can be cheaper than ex post litigation”72×72. Ayres & Gertner, supra note 5, at 97. when it is “systematically easier” for the parties to supply the missing term than for the courts to do so.73×73. Id. at 96. Drafters, however, can generally draw up ambiguous contracts at a lower private cost than unambiguous contracts, especially if they can pass some of those costs onto the courts through litigation.74×74. See Alan Schwartz & Joel Watson, The Law and Economics of Costly Contracting, 20 J.L. Econ. & Org. 2, 11 (2004) (“Contracting costs are positive and words are ambiguous, however, so that what the contract directs seldom can be made perfectly clear.”). Responding directly to this hazard, Judge McKeague wrote that insurers must “bear the consequences of [their] sloppy drafting.”75×75. Heimer, 879 F.3d at 177 (McKeague, J., dissenting in part and concurring in the judgment). The Heimer court used contra proferentem, therefore, to signal to insurers that the judiciary will push any attempt to externalize the costs of poorly drafted provisions back onto the drafters. The sloppiness of defendant’s drafting is aggravated in two ways: first, the exception defendants attempted to articulate through Exclusion 43 is a commonly included exception to insurance coverage;76×76. See ER Visits After Drinking May Not Be Covered, PBS: NewsHour (Apr. 30, 2012, 4:30 PM), https://www.pbs.org/newshour/health/er-visits-after-drinking-may-not-be-covered [https://perma.cc/2SJM-2HRD]. second, defendant’s insurance policy was a group medical insurance policy issued as part of an ERISA welfare benefit plan.77×77. Heimer v. Companion Life Ins. Co., No. 15-cv-338, 2016 WL 10932755, at *1 (W.D. Mich. Aug. 12, 2016). The former factor implies that the costs defendant would have incurred to write Exclusion 43 unambiguously were low. The latter factor implies that, because such group medical plans are designed to cover large numbers of people, the probability of important ambiguities being litigated is high. The Heimer holding conserved judicial resources by using a simple rule to shift the burden back onto the party better positioned to bear it. Rightly understood, Heimer represents a victory for the Ayres and Gertner side of the contra proferentem default debate, and indicates the judiciary’s increased willingness to seek efficiency, not between private parties, but between drafters and the courts.