concurring in part and concurring in the judgment.
I fully join the decision to certify the duty-to-defend question to the New Hampshire Supreme Court. I do not agree, however, that the Stratford policy, as originally issued, provided no coverage for the *87Ryder tractor. In my view, the original policy did provide such coverage, but the retroactive endorsement then made such coverage excess instead of primary. And thus I end up where the majority does, but by a different route.
As the majority notes, Stratford’s Senior Litigation Specialist, on reviewing the language of Stratford’s policy, concluded “there was an exposure out there that wasn’t intended to be covered by the policy.” Maj. Op. at 81. But Stratford’s response was not to argue the parties’ intent trumped the policy’s text nor to suggest the text was less than clear, Instead, Stratford reached an agreement with its insured to change the policy’s text via a retroactive endorsement that expressly limited that otherwise concerning exposure.
Thus, it is not surprising that in the District Court, even Stratford — following its Senior Litigation Specialist’s lead — did not dispute that the original policy covered the Ryder tractor. Of course, now that the District Court has bypassed the parties’ arguments about how best to deal with the exposure and, of its own accord, ruled such exposure never existed, Stratford agrees. But I believe Stratford had it right the first time. It did provide such coverage, even if it did so due to poor drafting.
No matter, though. Stratford worked out a deal. That deal limited the exposure by making coverage of tractors hired from Ryder excess rather than primary. And thus, the parties to the Stratford policy reached a sensible and practicable result. I think we should bless it rather than read the policy in a way that, I worry, may suggest to some that insureds should have less confidence in the text of their policies than I read precedent to show they should.
I.
The majority accepts, as it must, that the most directly pertinent portion of the policy — the definition of “hired ‘auto’ ”— indisputably includes the Ryder tractors. Maj. Op. at 80-81. That definition includes any “land motor vehicle, ‘trailer,’ or semitrailer designed for travel on public roads” that DAM “lease[s], hire[s], rent[s], or borrow[s]” from third parties. And the policy then states that it broadly covers “all sums an ‘insured’ legally must pay as damages because of ‘bodily injury’ or ‘property damage’ to which this insurance applies, caused by an ‘accident’ and resulting from the ownership, maintenance or use of a covered ‘auto’ ” — including, for this policy, a “hired ‘auto.’ ” The language could not be clearer.
Nonetheless, I can understand the temptation to look beyond the policy’s plain text. It does seem, from what the law calls extrinsic evidence (which is to say, evidence not found within the language of the policy itself), that neither party thought, at the time of drafting, about the kind of tractor at issue in this case. Stratford apparently did not know about this part of the insured’s business, and the insured apparently did not think Stratford was the source of coverage for these tractors.
But the majority acknowledges — as it must — that extrinsic evidence becomes relevant to limit exposure only if the text is actually ambiguous. White v. Vt. Mut. Ins. Co., 106 A.3d 1159, -, 2014 WL 6533298, at *3 (N.H.2014) (“[Ajbsent ambiguity, our search for the parties’ intent is limited to the words of the policy.” (quoting Bates v. Phenix Mut. Fire Ins. Co., 156 N.H. 719, 943 A.2d 750, 753 (2008))). And that rule reflects the fact that while the intent of the parties controls, that intent is found first and foremost in the words of the policy — words that, when *88clear, are determinative. See id. I do not read the majority’s key case, Tech-Built, to say otherwise. See Tech-Built 153, Inc. v. Va. Sur. Co., 153 N.H. 371, 898 A.2d 1007 (2006). There, the court first found the text of the policy at issue ambiguous and only then turned to extrinsic evidence to limit the reach of the coverage. Id. at 1009-10.
Nor does Tech-Built, as I read it, support finding ambiguity here. In that case, Tech-Built, a construction company, was a client of an employee leasing company called Surge. Id. at 1008. Tech-Built claimed the insurance policy that covered Surge actually also covered Surge’s clients. Id. at 1009. Tech-Built based that surprising contention on the following policy language: Surge’s policy listed the “insured” as Surge “etal [sic].” Id. (alteration in original). The policy then listed both Surge’s address and its “other workplaces,” which Surge described by listing the various companies (Tech-Built included among them) to which it had leased employees. Id. From this thin textual basis — an “interplay,” the court charitably called it — Tech-Built claimed Surge’s carrier had plainly agreed to supply insurance coverage to 150 or so of Surge’s clients, id., a most implausible claim. The New Hampshire Supreme Court had no trouble rejecting it. It simply noted that “[o]ther language” in the policy would be rendered “nonsensical” if Tech-Built’s interpretation was right. Id. at 1009-10.
That case is thus very far from this one. The definition of “hired ‘auto,’ ” unlike the definition of “insured” in Tech-Built, is crystal clear. And while the majority is quite right that the policy text must be considered as a whole, giving this “hired ‘auto’ ” language its plain meaning does not make nonsense of other parts of the policy. In fact, a plain reading of “hired ‘autos’ ” does not even lead to an “internally inconsistent policy.” Maj. Op. at 80 (citing Tech-Built, 898 A.2d at 1010). A review of the three other aspects of the policy the majority relies on in finding ambiguity reveals why.
First, the majority notes the policy describes DAM’s business as consisting of the delivery of office supplies and appliances, and makes no reference to any business involving the transport of “palletized freight.” Maj. Op. at 81. If the majority’s point is that the description of DAM’s business is not consistent with a policy term that plainly covers leased tractors, I cannot agree. After all, such tractors could readily be used to deliver office supplies and appliances.
More fundamentally, Stratford’s policy does not contain explicit language tying coverage to the description of DAM’s business. And the general rule (which, so far as I can tell, New Hampshire does not reject) is that “business descriptions” do not limit coverage to the precise type of business described. See, e.g., Mount Vernon Fire Ins. Co. v. Belize NY, Inc., 277 F.3d 232, 239 (2d Cir.2002) (rejecting the argument that a policy’s description of the insured’s business as “Carpentry” served to limit the coverage “to carpentry operations,” because “[t]he Policy simply fails to provide that the classifications define the covered risks”); GRE Ins. Grp. v. Metro. Bos. Hous. P’ship, 61 F.3d 79, 82 (1st Cir.1995) (rejecting the argument that a “business description” of “office” meant “only liability arising from [the insured]’s office operations was covered” because while the fact that the type of business was “office” rather than “skating rink” was “obviously relevant to coverage,” the description “d[id] not show a clear understanding to restrict coverage to liability arising out of [insured]’s office only”).
That rule makes sense. An insurer may require the description to inform the pre*89mium the insurer sets. But if the insurer wants to strictly limit coverage to activities within that description, it should explicitly say so. Mount Vernon Fire Ins., 277 F.3d at 239; see also, e.g., Wickramasekra v. Associated Int’l Ins. Co., 890 So.2d 569, 574 (La.Ct.App.2003) (construing a policy with an endorsement that explicitly restricted coverage to the type of business shown on the declaration); Sun Indem. Co. v. Lovell, 6 Conn.Supp. 337 (Conn.C.P.1938) (same).
Second, the majority notes the policy covers “specifically described ‘autos’ ” and that the two “specifically described ‘autos’ ” listed in the policy are smaller vans and not Ryder tractors. Maj. Op. at 81. But the policy provided coverage for two separate categories of “autos”: “hired ‘autos’ ” that, by terms, DAM does not own, and “specifically described ‘autos’” that, by terms, must be described on a separate schedule. That schedule in turn refers to the listed vehicles as “covered autos you own.” The leased Ryder tractors, therefore, were not listed as “specifically described ‘autos’ ” for the simple reason that DAM did not own them. As vehicles DAM only leased, these tractors would be covered as “hired ‘autos.’ ” If the majority’s point is that an objective reader would have to assume that the rented vehicles' necessarily would be similar in type to the owned vehicles specially listed on the schedule, I do not see why. Companies might rent vehicles of a different type precisely because the ones they own are not adequate to every task.
Finally, the majority notes the policy lists an “estimated cost of hire” for all hired autos at $5,000 per year. Maj. Op. at 81. That number is small. It does seem fit for a modest business using vans rather than for a large one using tractors. But the text of the policy contemplates insureds may lowball their estimates — presumably to reduce their premiums. That is why — as the majority acknowledges, Maj. Op. at 81 the policy expressly provides that estimates supplied by the insured on this portion of the policy are not binding. In fact, the policy further provides that Stratford may audit the insured’s actual hired-auto expenditures and retroactively increase the premium should the estimate prove too low. In other words, even though no intentional low-balling occurred here, Stratford’s policy made clear that an insured’s low estimate is not binding in the insured’s favor. I thus do not believe we should rely on the estimate to create an ambiguity in what is otherwise clear and plainly binding policy language— the “hired ‘auto’ ” definition.
For these reasons, I would hold the policy language is clear. That being the case, it seems to me that New Hampshire law requires us to give effect to the plain meaning of the “hired ‘autos’ ” definition. White, 2014 WL 6533298, at *3; Bates, 943 A.2d at 753. And thus we may not limit the coverage by looking about for outside evidence of the parties’ understandings (here, the contract between DAM and Ryder). But that does not mean Stratford is without recourse, as I will now explain.
II.
Faced with an exposure it wished to limit, Stratford reached out to DAM to provide excess coverage, rather than primary coverage, for the leased Ryder tractors. I would give the deal Stratford and DAM struck full effect.
True, as Old Republic points out, a party’s affirmation of a preexisting duty is not generally adequate consideration for a new contract. Melotte v. Tucci, 319 Mass. 490, 66 N.E.2d 357, 358 (1946). But relinquishing a contestable claim is. Pitkin v. Noyes, 48 N.H. 294, 304 (1869); see also Mathewson Corp. v. Allied Marine Indus., *90Inc., 827 F.2d 850, 856 (1st Cir.1987) (Massachusetts law). I conclude Stratford’s pre-existing duty to provide 'coverage was clear. But I do not believe a contrary claim would be frivolous. And, of course, neither would the majority. Thus, Stratford gave up something real. It foreclosed its right to argue it could deny coverage altogether. See Chisholm v. Ultima Nashua Indus. Corp., 150 N.H. 141, 834 A.2d 221, 225 (2003) (“Consideration is present if there is either a benefit to the promisor or a detriment to the promisee.”); see also Pitkin, 48 N.H. at 304 (“[Compromise of doubtful claims” is consideration unless the claims are “utterly without foundation and known to be so”); Mathewson Corp., 827 F.2d at 856 (finding consideration if the surrendered claim is not “‘vexatious or frivolous’”) (quoting Blount v. Dillaway, 199 Mass. 330, 85 N.E. 477, 479 (1908)).
Old Republic also argues the endorsement prejudiced Old Republic without Old Republic’s consent. But no term in the Stratford policy protected Old Republic from such modification. See Restatement (Second) of Contracts § 311(2) (absent a term in the contract forbidding modification of a duty to a third party “the promisor and promisee retain power to discharge or modify the duty by subsequent agreement”); see also Brooks v. Trs. of Dartmouth Coll., 161 N.H. 685, 20 A.3d 890, 900 (2011) (following the Second Restatement’s description of third-party beneficiary rules — though not mentioning this particular one — as a matter of New Hampshire law).
Nor does New Hampshire law make Old Republic a vested third-party beneficiary entitled to such protection in the absence of such a policy term. The original Stratford policy did not “satisfy some obligation owed by the promisee [DAM] to the third party [Old Republic],” nor was that policy “so expressed as to give the promisor [Stratford] reason to know that a benefit to a third party [Old Republic] is contemplated by the promisee [DAM] as one of the motivating causes of his making the contract.” Brooks, 20 A.3d at 900. In fact, Stratford did not have reason to know about the Old Republic policy, much less to know DAM contemplated Old Republic would benefit from the Stratford policy.
That leaves only Old Republic’s argument that the endorsement violates a public policy against “post-claim underwriting.” The out-of-New Hampshire cases on which Old Republic exclusively relies (from Louisiana and Mississippi, respectively) are readily distinguished. They involved retroactive policy alterations that limited the coverage for an injured party (in two instances, for an injured party who was also a party to the policy). See Mattox v. W. Fid. Ins. Co., 694 F.Supp. 210, 216 (N.D.Miss.1988); Washington v. Savoie, 634 So.2d 1176, 1180 (La.1994); Lewis v. Equity Nat’l Life Ins. Co., 637 So.2d 183, 188-89 (Miss.1994). And those alterations were made without the insurer first obtaining the injured party’s consent. See Mattox, 694 F.Supp. at 216; Washington, 634 So.2d at 1180; Lewis, 637 So.2d at 188-89.
Nothing like that happened in this case. Here, the insured consented to the change at the time it was made. And, further, the party injured by the insured will not suffer any reduction in total coverage in consequence of the change. Thus, the only party “harmed” by this change is an insurer, Old Republic, who is not a party to the policy changed and whose harm is hard to divine. Old Republic simply must provide coverage it thought it was on the hook for all along — coverage Old Republic also must provide under the majority’s approach.
*91III.
An insured should be able to rely on what the policy says. New Hampshire agrees. Like other states, it provides that even ambiguous policies “will be construed against the insurer,” Catholic Med. Ctr. v. Exec. Risk Indemn., Inc., 151 N.H. 699, 867 A.2d 453, 456 (2005), at least absent sufficient extrinsic evidence to show the parties intended otherwise. All the more reason, therefore, to be wary of resorting to extrinsic evidence too easily arid then relying on it to defeat coverage for the insured. As this case and others I refer to show, the non-binding aspects of a policy may not reflect the full extent of the coverage contained in a policy’s binding passages. But it is those binding passages that should control when clear. And as I find them clear here, I also find them controlling.