JUSTICE YOUNG delivered the opinion of the Court. As part of a larger settlement agreement, the parties in this case trusted Wesley Holmes, an attorney representing the owners of CTMI, to hold about $1 million in escrow. Holmes held those funds in an account under CTMI’s name—at least, he did so up until he drained the money from the account and took it for himself. The settlement agreement provided that Fischer, the party adverse to CTMI and its owners, would receive those funds if this Court ruled in his favor, as it eventually did. Only after that ruling did the parties learn of Holmes’s actions. The bottom-line question is whether CTMI was still required to pay Fischer or whether CTMI’s payment of the required amounts into the escrow account discharged its liability. In other words, the outcome of the parties’ dispute ultimately turns on who bore the risk of loss caused by the escrow’s failure. The court of appeals resolved the case by holding that there never really was an escrow agreement. The court believed that an escrow holder must be neutral. Holmes—a lawyer on one side of the dispute— was decidedly not neutral. Placing the funds with him, the court of appeals accordingly held, could not constitute an escrow. And if Holmes did not hold the money in escrow, he must have held it in the ordinary way that lawyers routinely hold client funds. A lawyer’s theft of his client’s funds, while outrageous and unfortunate, would not satisfy that client’s contractual obligations to a third party like Fischer. Presumably based on this logic, the court of appeals ruled for Fischer. We disagree with the premise of this rationale. Making one side’s lawyer an escrow holder may seem imprudent, especially in hindsight. But it is not unlawful. If parties find it sensible and efficient for a non-neutral party to be an escrow holder, the law does not forbid them from making that choice. The law only demands clarity. While a non-neutral individual like one party’s attorney is presumptively not an escrow holder, the parties’ unmistakable agreement to the contrary will rebut that presumption. We hold that the parties unambiguously created an escrow to be held by Holmes until the conclusion of their litigation, at which point the prevailing party should have received the escrowed funds. We nonetheless affirm the court of appeals’ judgment. The choice for parties in adversarial litigation to use one party’s attorney as an escrow holder and to keep the funds in an account opened under that party’s name will not defeat an escrow if all parties clearly consent, but that choice remains significant for the risk of loss. When one party agrees to retain title to property and to allow its attorney (or other fiduciary) to control the property, that party presumptively retains the risk of loss. More to the point, the party that lacks title to the escrowed property and whose lawyer is not the escrow holder cannot be regarded as having accepted the risk of loss without something more than the existence of the escrow itself. Express contractual language or necessary implication from its text is indispensable in this situation. Nothing in these parties’ contractual relationship, however, establishes any such agreement. We therefore hold that CTMI bore the risk of loss. I This is the second time this Court has taken up this case. Our prior decision described the facts in greater detail, see Fischer v. CTMI, L.L.C., 479 S.W.3d 231, 233–36 (Tex. 2016), and we provide only a condensed version relevant to the dispute as it returns to us. Ray Fischer agreed to sell his tax-consulting business to CTMI, a business that Mark Boozer and Jerrod Raymond created for the purpose of receiving the assets of, and then operating, Fischer’s old company. See id. at 233. Disputes arose; litigation commenced.[1] The parties settled in open court through a Rule 11 agreement. See Tex. R. Civ. P. 11. Their settlement agreement, however, left one claim unresolved: CTMI’s request for a declaratory judgment regarding Fischer’s entitlement to revenue for projects begun, but not finished, before he left the company. See Fischer, 479 S.W.3d at 234–36. This claim was severed from the remainder of the case, and Fischer ultimately prevailed on it in this Court. Id. at 244. The present litigation concerns the aspect of the settlement agreement that provided for what Fischer would receive if he won in the prior appeal and the mechanics of how those funds would be handled. At Fischer’s demand, the parties agreed that the disputed funds would not stay with CTMI but would “be[] held in escrow.” As revenue from specified projects came in, 15% of it would, “upon receipt, be paid into an interest bearing escrow account.” If Fischer won after all appeals were exhausted, the settlement agreement provided that Fischer “will receive payment of” the funds. But who would hold them in the meantime? One option was to make Fischer’s attorney the holder of the amounts to be held in escrow. Fischer’s counsel, however, made clear that he did not want “to be paying income tax on the interest that accrues on it.” The parties then agreed that CTMI would place the funds in an account to be controlled by Wesley Holmes, the attorney for Boozer and Raymond. With the tax question in mind, they agreed that the account would “be a CTMI account, but” that they would “have it be controlled by [Holmes].” As Holmes put it, and the parties agreed, CTMI “will pay the taxes on it, but I’m on the hook.” After the attorneys hammered out these details in open court, they emphasized that “it’s agreed by the parties that what is being entered onto the record in this Rule 11 setting, does not contemplate the parties sitting down and executing any further settlement agreement to be written or agreed upon. This is to be used as the settlement agreement, this transcript.” The parties each took the stand and confirmed that they understood and agreed to these arrangements. The judge then called the jury back in, gave them the news that the case had settled, and complimented the lawyers on “some real fine lawyering.” Little did the judge—or anyone else—know what would come to pass. At first, all went smoothly. Based on the agreement, Holmes opened the account in CTMI’s name, but he was the sole signatory and he alone had authority over the account. The address listed for CTMI was Holmes’s law office. Overtime, as the agreement required, CTMI transferred funds totaling nearly $1 million—$990,175.66, to be exact. After the court of appeals ruled for CTMI, this Court granted review and held oral argument. Only this Court’s decision and mandate were necessary to determine the final distribution of the money. If we had denied the petition or affirmed the judgment, Holmes would have been obligated to return control of the funds to CTMI. If, as in fact happened, we reversed the judgment, then Holmes’s duty was to pay the funds over to Fischer. After we ruled for Fischer, he demanded that CTMI pay him the amounts due. By then, however, Holmes had absconded with the funds, which came to light only after we announced our decision. In response, CTMI filed this lawsuit, seeking a declaration that it had fulfilled its obligations under the settlement agreement. CTMI reasoned that, because it had paid every dollar of the funds at issue as the escrow agreement required, it owed nothing further to Fischer. Fischer counterclaimed for breach of contract, arguing that the agreement also required CTMI to pay him the proceeds. In three partial summary judgments, the trial court ruled for CTMI, absolving it of further responsibility. The trial court first held that CTMI accurately calculated and properly deposited the funds into the escrow account; second, it concluded that the parties had created an escrow with Holmes as the sole escrow agent and that Holmes had converted the funds; and third, the trial court concluded that CTMI had no further liability to Fischer. Based on these rulings, it rendered a final judgment in favor of CTMI.[2] The court of appeals reversed. 665 S.W.3d 46 (Tex. App.—Fort Worth 2021). It held that the settlement agreement did not create an escrow and therefore that CTMI’s failure to pay Fischer the amounts due was a breach of the settlement agreement. See id. at 55–57. The court also awarded attorney’s fees to Fischer and remanded for the trial court to address pre- and post-judgment interest. See id. at 57–59. CTMI petitioned this Court for review of the court of appeals’ judgment. Like that court’s opinion, the parties’ briefing here focused primarily on whether the settlement agreement created an escrow.[3] II The first question presented by this appeal is whether parties may form an escrow agreement that selects one party’s attorney to serve as the escrow holder, even if those parties are actively engaged in litigation against each other. We conclude that they may. A Given how frequent the use of escrow is, and how familiar a concept it is even to nonlawyers, remarkably few discussions of escrow appear in this Court’s cases. And despite a myriad of statutory provisions that refer to escrow or govern aspects of its use in specific contexts,[4] the legislature has not adopted a baseline, generally applicable definition of the concept.[5] For the most part, this dearth of law about escrow’s core attributes likely reflects that we know what escrow is through frequent experience. But cases on the margins, like this one, also highlight the need for a clear, easily administrable definition. Such a definition is necessary in this case, where the parties have identified no statute, regulation, or precedent of this Court that directly governs whether the parties’ arrangement constituted an escrow. We have found none, either. We therefore articulate and apply a definition of escrow thatdrawsonavailableauthoritytoencapsulateourlongstanding,though often implicit, common-law understanding of the concept.[6] The concept of escrow originated in the context of land conveyances. See William A. Ingraham, Jr., Comment, Escrow Agreements,8 Mia. L.Q. 75, 75–76 (1953). As Blackstone explained, the key distinction between an immediately effective deed delivered to the party itself and one held in escrow is that the delivery of the deed held in escrow depends on the fulfillment of a condition: A delivery may be either absolute, that is, to the party or grantee himself; or to a third person, to hold till some conditions be performed on the part of the grantee: in which last case it is not delivered as a deed, but as an escrow; that is, as a scrowl or writing,[7] which is not to take effect as a deed till the conditions be performed; and then it is a deed to all intents and purposes. 2WilliamBlackstone,Commentaries*307. The applications of an escrow have since expanded to include other forms of property, such as money or indeed any type of real or personal property. Charles H. Walker & William D. Eshee, Jr., The Safeguards and Dilemmas of Escrows, 16Real Est. L.J. 45, 49 (1987). This Court has only infrequently addressed the fundamental features of an escrow. In an adopted Commission of Appeals opinion early in the twentieth century, however, we set out the basic nature of an escrow as a “writing[8] by the grantor, promisor or obligor” deposited “with a third person not a party thereto” that was “to be kept until the performance of a condition or the happening of a certain event, then to be delivered to take effect.” Green v. Priddy, 250 S.W. 656, 660 (Tex. [Comm'n Op.] 1923) (internal quotation omitted). We also glean insights from other cases from this Court that reference escrow. First, to create an escrow, the parties must agree to do so. See City of Fort Worth v. Pippen, 439 S.W.2d 660, 664 (Tex. 1969); Pickle v. Whitaker, 224 S.W.2d 741, 745 (Tex. Civ. App.—El Paso 1949, writ ref’d); Tyler Bldg. & Loan Ass’n v. Biard & Scales, 171 S.W. 1122, 1122–23, 1125 (Tex. 1914).[9] This is because of the fundamental principle that an escrow holder owes fiduciary obligations to both parties, not just one.[10] See Pippen, 439 S.W.2d at 664. Second, the escrowed property leaves the depositor’s control through the duration of the escrow. See Pickle, 224 S.W.2d at 745. These principles yield the following basic common-law elements of an escrow: a deposit of property (which could be a legal instrument such as a deed or contract) upon an agreement by the parties with a third party, who will owe fiduciary obligations to both parties for purposes of the property held in escrow who will hold that property outside of the depositor’s control and who will deliver that property to the other party upon the performance of a certain condition or the happening of a certain event, or otherwise will relinquish the property. Escrow arrangements that satisfy these requirements can come in all shapes and sizes.[11] Compare Harris v. Rowe, 593 S.W.2d 303, 304–05 (Tex. 1979) (two-page escrow agreement), with Martin Operating P’ship v. QEP Marine Fuel Inv., LLC, 525 S.W.3d 712, 726 (Tex. App.—Houston [14thDist.]2017,nopet.)(thirteen-page,single-spacedescrowagreement). Our point is not that slapdash or abbreviated escrow agreements— even those that satisfy the requirements listed above—are wise or desirable. Our concern today is only to identify the minimum common-law requirements for forming an escrow, not to advocate for any particular choice that parties may make. Written agreements that clearly articulate their terms and the escrow holder’s duties remain the gold standard, especially to avoid doubts about the inevitable contingencies that surround escrow. The clearer an agreement is about exactly what should happen under any given set of circumstances, the better. And as with any other contract, failure to include an essential term—a term “parties would reasonably regard as vitally important” to the bargain—renders the contract unenforceable. Fischer, 479 S.W.3d at 237 (internal quotation omitted). B To determine whether these parties created an escrow, we look to the parties’ intent as expressed by the terms of their agreement. See Green, 250 S.W. at 660; URI, Inc. v. Kleberg County, 543 S.W.3d 755, 763 (Tex. 2018). As usual when construing agreements, no magic words are necessary, but the words the parties used in their agreement—especially the word “escrow” itself—are the clearest indicators of that shared intent. While not dispositive or sufficient, “the use of the word ‘escrow’ . . . indicates more clearly than any other their actual intention.” Lechner v. Halling, 216 P.2d 179, 185 (Wash. 1950). This principle is merely a manifestation of our general approach to contracts. See, e.g., URI, 543 S.W.3d at 764 (noting that we “presume parties intend what the words of their contract say” (internal quotation omitted)). We also look to whether the agreement fits the basic elements of an escrow. As with other contracts, we may look to the objective circumstances of the escrow agreement’s execution and the property’s deposit as appropriate to help elucidate (but not to inject ambiguity into) the agreement’s text. See Green, 250 S.W. at 660; see also URI, 543 S.W.3d at 758 (noting that appropriate context can include “objectively determinable facts and circumstances that contextualize the parties’ transaction”); Harris, 593 S.W.2d at 306 (noting, in the context of construing an escrow agreement, that courts “consider the wording of the instrument, in the light of surrounding circumstances”). In this case, the words the parties chose and the structure of their transaction both demonstrate that they created an escrow. The agreement repeatedly used the word “escrow” to describe the arrangement. It stated that the funds would “be paid into an interest bearing escrow account” and “be[] held in escrow,” with the funds being placed in the “escrow account” of either Holmes or Fischer’s attorney. The parties expressly agreed that the transcription of the hearing would itself constitute the text of the agreement. In addition, the agreement’s substance satisfies the fundamental requisites of an escrow. The parties agreed that CTMI would place control of the funds with a third party—Holmes—who agreed to safeguard the funds as a fiduciary for both parties in accordance with the parties’ agreement. The agreement expressly stated that those funds were to be “controlled by” Holmes alone, not CTMI. And most importantly, the parties and Holmes agreed that Holmes would hold the money until the fulfillment of the escrow condition: the judicial resolution, upon the exhaustion of all appellate avenues, of the parties’ contractual dispute. Fischer would then “receive payment” if the courts resolved the dispute in his favor. The attorneys for all parties and the parties themselves agreed to this arrangement; Holmes accepted his duty, too. CTMI then deposited the funds into the account, completing the steps necessary to create an escrow. All the indicia of an escrow were present and the parties denominated it as such, so we cannot credit Fischer’s post hoc characterization of their use of that term as merely casual or colloquial. To the contrary, it cannot have been an offhand way of describing a mere trust account with IOLTA funds—the reason for Holmes to serve as the escrow agent was because the account would generate taxable interest and tedious filings, which Fischer’s counsel did not want to pay or manage. Fischer continued to describe the arrangement as an “escrow” long after the open-court agreement—including in formal briefs filed in this Court. Those briefs described the arrangement as an escrow; we accepted that description and used it in our opinion. See Fischer, 479 S.W.3d at 236. Only when it became inconvenient was the escrow label disclaimed. Our point is not that Fischer is “bound” by that usage in the briefs—our prior decision did not turn on whether it was an escrow. Rather, the repeated use of “escrow,” combined with the substantive characteristics of an escrow, illustrate that it was an escrow as a matter of law. C Fischer also argues, however, that the law does not authorize parties to choose an escrow holder who is not a neutral third party. In support of this proposition, Fischer cites several lower-court opinions stating that the escrow holder must be a “neutral third party” or “stranger” and cannot be the agent of one of the parties.[12] If Fischer is right, the arrangement in this case would not be a legally permissible escrow regardless of the parties’ intent because Holmes was neither neutral nor a stranger to the transaction. He was Boozer and Raymond’s attorney and was engaged in active litigation against Fischer. Fischer is correct up to a point, but it is an important point: when one party’s attorney holds funds, the law will presume that the parties have not created an escrow. As discussed above, one fundamental feature of an escrow is that the escrow holder owes fiduciary duties to both parties, not just one. Ordinarily, however, an attorney owes fiduciary duties to—and serves as an agent for—his own client, not an adverse party. Presumptively, therefore, when a party gives his attorney funds to hold, the attorney holds those funds as his client’s agent. See Pickle, 224 S.W.2d at 745. The cases Fischer cites reflect this understanding of an attorney’s usual role. None of them addresses the question of whether a party’s attorney may serve as the escrow holder when, up front, the parties clearly agree to that role. The cited cases do one of two things. Several look backward in time to retrospectively determine whether an arrangement that lacked a clearly expressed intent was in fact an escrow—and, of course, no escrow could be created under such circumstances.[13] Others simply use the “neutral third party” wording or similar language when accurately explaining the typical characteristics of an escrow—and we reaffirm today that having a “neutral” or a “stranger” is the norm.[14] These cases are not binding on us and we express no view on them other than to confirm this basic point, which we think emerges from them and many others: as a default, a lawyer or other agent for one side will be deemed not to have been an escrow holder without clear evidence that the parties (and the attorney)agreed upfront that the attorney would not merely hold funds but do so as the parties’ escrow holder. As with other default rules, parties are free to contract around this one, even if doing so creates odd or inadvisable results. See Rosetta Res. Operating, LP v. Martin, 645S.W.3d212, 220(Tex. 2022);Endeavor Energy Res., L.P. v. Energen Res. Corp., 615S.W.3d144,153 (Tex.2020). Thiscontractualfreedomisespeciallyapplicablewhenthepartiesinvolved are sophisticated, well-counseled, and not subject to domination by the other side. See Forest OilCorp.v.McAllen,268S.W.3d51,58(Tex.2008). Our conclusion is consistent with longstanding precedent from other states’ high courts, which recognize that one party’s attorney or agent may serve as the escrow holder as long as doing so presents no conflict with the client’s interests.[15] It is likewise consistent with the Restatement, which clearly contemplates that lawyers representing one side in a dispute may serve as an escrow holder for both sides: “the arrangement under which the lawyer receives property of a third person of adverse interest—for example, an escrow arrangement—can imply that the client and third person have agreed that the lawyer is to protect the third person’s interests.” Restatement (Third) of the Law Governing Lawyers § 44 cmt. h (emphases added). Neither the Restatement nor other states’ courts are binding on us, of course, but they confirm that our understanding is based on shared and noncontroversial common-law premises. Finally, and important to both the bar and the public, our recognition of the parties’ contractual authority to choose one side’s lawyer as an escrow holder is also consistent with our Rules of Professional Conduct. The rules state that an attorney may serve as an “escrow” holder in connection with a representation and must uphold the standards of a fiduciary. See Tex. Disciplinary R. Prof’l Conduct 1.14(a) & cmt. 1. Comment 4 further suggests that, when a lawyer serves as an escrow holder, it normally is as part of a representation. It notes that “even” when a lawyer happens to be an escrow holder aside from a representation, the normal fiduciary responsibilities of lawyers still fully apply: “For example, a lawyer who serves as an escrow agent is governed by the applicable law relating to fiduciaries even though the lawyer does not render legal services in the transaction.” Id. R. 1.14 cmt. 4 (emphasis added). Attorneys understand that when they agree to serve as an escrow holder, they take on fiduciary duties to both parties for the limited purpose of the escrow agreement. This Court has noted that fiduciaries who handle the funds of others owe duties of loyalty, of full disclosure, and “to exercise a high degree of care to conserve the money and pay it only to those persons entitled to receive it.” Pippen, 439 S.W.2d at 665. Our courts of appeals have properly understood an escrow agent to owe these duties to all parties to an escrow agreement. See, e.g., Williams v. Land Title Co. of Dall., No. 05-96-00039-CV, 1997 WL 196345, at *5 (Tex. App.—Dallas Apr. 23, 1997, no writ); Trahan v. Lone Star Title Co. of El Paso, Inc., 247 S.W.3d 269, 287 (Tex. App.—El Paso 2007, pet. denied).[16] An attorney should be well-equipped to comply with the high fiduciary duties of an escrow holder. Those duties do not generate a conflict in a context like this one, where an agreed-upon objective standard governs the distribution of funds and where the attorney acknowledges himself to be “on the hook,” rather than merely acting for his client.[17] D Finally, Fischer argues that the arrangement was not an escrow because CTMI was the owner of the escrow account and was responsible for the taxes due on the interest generated by the account. In contrast, CTMI points out that Holmes was the account’s sole signatory and that his law-office address, not CTMI’s address, was listed on the account. We agree with CTMI. As we described above, what matters for the creation of an escrow is not formal ownership but control. See, e.g., Pickle, 224 S.W.2d at 745 (“In an escrow the paper passes out of the control of the maker.”). Escrow agreements generally do not require or even contemplate the transfer of legal title from one of the parties until the occurrence of a triggering event, upon which the escrow holder must distribute the property according to the agreement’s terms. See Cowden v. Broderick & Calvert, 114 S.W.2d 1166, 1169 (Tex. 1938) (“[L]egal title did not pass to the lessee until the conditions of the escrow agreement were satisfied . . . .”); Bell v. Rudd, 191 S.W.2d 841, 843 (Tex. 1946) (“As [the escrow] agent[, the bank] was to deliver the instruments according to the terms of the escrow agreement and not otherwise.”). After all, parties usually need an escrow because one of them has property that is subject to dispute or that they wish to transfer under specified conditions. Exclusive control over the property by a third-party escrow holder (and no one else, regardless of formal title) is often essential for a transaction to proceed smoothly. Ownership will eventually change—but only if the stated contingency occurs. The fact that the account was under CTMI’s name is certainly not irrelevant to the creation of the escrow, however(or, as we discuss below, to the risk-of-loss analysis). To the contrary, if all we knew was that the funds were held in a CTMI account, we would agree that there was no escrow. But we would reach that conclusion because account owners ordinarily can control their own accounts. The control element remains central, so once the parties agreed to an arrangement in which Holmes would exclusively control the funds in the account—that is, an arrangement dividing control and ownership—the impediment to such an account being held in escrow evaporated. Once CTMI placed the funds into the escrow account, it could not direct Holmes regarding what to do with them.[18] Holmes alone controlled the funds, which is why he was able to abscond with them. Had this Court either denied the petition for review or affirmed the judgment in the parties’ first appeal, Holmes’s conduct would have been to CTMI’s (and his own clients’)detriment. Holmes’s conduct likewise harmed CTMI and his own clients even though the Court ruled for Fischer: by draining the account, Holmes depleted the funds that CTMI had intended to be set aside for compliance with the settlement agreement. That conduct caused CTMI serious legal problems, as CTMI sought to hold Holmes accountable and to defend against Fischer’s demand for payment. The court of appeals reasoned that understanding the parties’ agreement to constitute an escrow agreement would render three terms of the agreement “meaningless”: (1) the parties’ choice to place the funds in a CTMI-owned account; (2) the parties’ choice to have CTMI pay taxes on the account; and (3) the parties’ agreement that Fischer “will receive” payment of the funds. See 665 S.W.3d at 56–57. As discussed above, the first two reasons do not prevent the agreement from creating an escrow. And the fact that Fischer did not receive payment cannot make the “will receive payment” term meaningless—it simply means that the escrow failed, as could happen in any context. Accordingly, that the account was held in CTMI’s name did not defeat the parties’ effort to create an escrow. * * * The parties in this case referred to their arrangement as an escrow. They instituted a condition upon which Fischer would be entitled to the funds held by Holmes: a final judicial determination that the contractual provision at issue was enforceable. They agreed that Holmes would control the funds. They were sophisticated parties already engaged in litigation with each other. Their attorneys hammered out the details and CTMI deposited the funds in the designated account. We hold, therefore, that they created an escrow. III CTMI has sufficiently rebutted the presumption against the creation of an escrow when one party’s attorney holds funds and title remains formally vested in that party. The successful creation of an escrow, in turn, generates various consequences. For example, because Holmes was an escrow holder, he owed fiduciary duties to CTMI and Fischer, not just his own clients. His status as an escrow holder subjected him to personal liability from all sides; the account’s status as an escrow made it an escrow for all general purposes. None of this, however, is material to the resolution of this appeal. Rather, the question is whether the successful creation of an escrow in this unusual context affects CTMI’s obligation to Fischer with respect to the disputed amounts. We thus address whether, by entering into or fully complying with the escrow agreement, CTMI shifted the risk of loss to Fischer. The parties have not pointed us to—nor are we aware of—any statutory risk-of-loss rules that apply here, even though the legislature has adopted risk-of-loss rules in other contexts. Nor has this Court adopted a common-law risk-of-loss rule relating to the failure of escrow that would apply when the escrow agreement itself is silent on this issue. Fischer argues that we should do so now by adopting the “entitlement rule” for determining how and when risk of loss shifts in the escrow context. We decline to reach that question here, however, for several reasons. The court of appeals did not address it; the parties only lightly raise it in their briefing; the experience of many of our sister states reflects the host of complicated questions that we would have to answer before doing so (or refusing to do so); and, most importantly, we can resolve this dispute without proceeding to consideration of the entitlement rule, which we therefore may reserve for future cases. Today’s dispute, in other words, does not require us to develop any general principles that broadly apply to escrow as opposed to other kinds of arrangements. Instead, whether such general rules exist, and whatever they are, we conclude that the principle discussed in Part II also applies here: when parties involved in active litigation agree for one party’s attorney (or, as applicable here, an attorney on one side of the litigation) to serve as escrow holder, and when that party agrees to retain title to the property in escrow, that party must bear the risk of loss unless the parties clearly indicate a contrary contractual intent. The parties did not displace that presumption here. CTMI rather than Fischer therefore bears the risk of loss occasioned by Holmes’s misconduct. We begin by explaining why we decline to establish a general risk-of-loss rule that would apply to all escrow arrangements. We then turn to how we resolve the parties’ dispute on narrower grounds. A Texas has not adopted or rejected the “entitlement rule,”[19] which is a general escrow risk-of-loss rule that Fischer suggests. Jurisdictions that have adopted it in the escrow context describe it this way: “If property in the custody of an escrow holder is either embezzled or lost by it, then as between the seller and the buyer, the loss falls on the one who owns the said property at the time of the embezzlement or loss.” Schmidt v. Fitzsimmons, 226 P.2d 304, 306 (Or. 1951). Whether to join those jurisdictions should await a case in which that question is fully presented and aired. For one thing, criticism of that doctrine,[20] including in the Restatement,[21] warrants consideration not only from this Court but from the lower courts, parties to escrow-related litigation, and amici. Likewise, how far to take the entitlement rule is an important and nonbinary question, even in jurisdictions that have applied it. They often have done so in contexts highly distinct from this one, most commonly involving real-estate transactions or some other purchase of property. See, e.g., Foster v. Elswick, 4 S.W.2d 946, 946–47 (Ark. 1928); Schmidt, 226 P.2d at 305–07; Paul v. Kennedy, 102 A.2d 158, 158–59 (Pa. 1954); Stuart v. Clarke, 619 A.2d 1199, 1199–1200 (D.C. 1993). And even in the real-estate context, how the rule is applied can vary depending on the type of escrow. See Bixby Ranch Co. v. United States, 35 Fed. Cl. 674, 679–80 (1996) (distinguishing between deed-and-money escrows and “set-aside” escrows). Even when the rule applies in a given context, it has its exceptions. Some courts, for example, depart from the rule when one party is more at fault than the other (such as when one party unreasonably delays fulfilling the escrow condition), see Foster, 4 S.W.2d at 947, or when “the depositor would not be entitled to the return of the subject matter under any circumstances, irrespective of performance of the terms of the agreement,” Cradock v. Cooper, 123 So. 2d 256, 258 (Fla. Dist. Ct. App. 1960). In particular, some courts have departed from the rule when it was more equitable to place the risk of loss on the seller as opposed to the buyer, based at least in part on the principle that “where one of two innocent persons must suffer from the wrongful act of a third, the loss should be borne by him who put the wrongdoer in a position of trust and confidence[.]” Paul, 102 A.2d at 160 (internal quotation omitted) (placing the loss on the seller because the escrow holder was acting primarily as the seller’s agent and the seller had not performed his obligations before the embezzlement); see also, e.g., Jones v. Lally, 511 So. 2d 1014, 1016 (Fla. Dist. Ct. App. 1987) (placing the loss on the sellers because they agreed to receive an ordinary check from the escrow holder instead of the cash or equivalent to which they were entitled and failed to properly endorse the check, making them “best able to avert the loss and . . . the least innocent”). Some commentators have suggested that when neither party is more to blame, the logic of an escrow requires that they share the risk of loss—after all, an escrow holder is a joint fiduciary rather than a fiduciary for only one party.[22] We cannot adopt the entitlement rule, or refuse to adopt it, without taking these pivot points into account. In an escrow, are both parties equally defrauded by the escrow holder, such that they should share the loss? Should the structure of this escrow transaction be informative, as described above, such that the rule that might apply in a deed-and-money escrow would not necessarily apply in other types of escrow? Are there other helpful analogies from related areas of the law that may be important in understanding how to allocate escrow loss in various contexts? In short, when the parties’ agreement is otherwise silent on this issue, how does the existence of an escrow affect the question of who bears the risk of loss? We note these issues not to express any view on how courts should resolve them, nor to suggest that they are the correct (much less exclusive) questions to ask, but only to illustrate why we decline to adopt a general principle applicable to all escrows today. As the experience of other states demonstrates, risk of loss is not geared toward a one-size-fits-all solution. That principle is also reflected in the legislative fine-tuning of risk-of-loss rules in other contexts.[23] * * * Allocating the risk of loss in the escrow context appears to be a question of first impression in Texas. Given how widespread the use of escrow is—and how many different types of escrow agreements exist— it is to be expected that many in the academy, bar, and industry will have helpful views if and when Texas courts must address this question outside the narrow context presented today. B We finally turn to the resolution of this case. The court of appeals’ judgment was premised in part on its mistaken conclusion that no escrow had been created. CTMI therefore requests that we remand to the court of appeals to address the risk-of-loss issue in the first instance. CTMI is correct that we could exercise our discretion, as we often do, to remand for further proceedings in light of our holding on escrow.[24] We would do so if we believed that addressing the entitlement rule was necessary, for example. “As a court of last resort, it is not our ordinary practice to be the first forum to resolve novel questions, particularly ones of widespread import. . . . Rather, this Court’s preferred process is to decline to address and defer such questions until after complete vetting of the parties’ potential arguments in the lower courts.” In re Troy S. Poe Tr., 646 S.W.3d 771, 780 (Tex. 2022) (internal quotation omitted). The terms of the parties’ settlement and this case’s specific circumstances, however, allow us to narrowly resolve this dispute while leaving consideration of the “entitlement rule” for another day. A remand is therefore unnecessary. We begin with how the parties’ relationship should have unfolded. As we concluded in Part II, the parties’ agreement established an escrow in which Holmes agreed to serve as the escrow holder.[25] The parties agreed that CTMI was to pay Fischer the disputed funds if he won on appeal, and they set up a specific mechanism for how the payment process would work. This required CTMI to deposit the funds into the account controlled by Holmes, which it indisputably did. Holmes was then to distribute those funds once the escrow condition was fulfilled. Of course, things did not work out that way. As escrow holder, Holmes became liable for those funds. See Norman v. Wilson, 41 S.W.2d 331, 332 (Tex. Civ. App.—Austin 1931, writ ref’d). The question here, however, is not whether Holmes was, as he himself put it, “on the hook.” He was. The question is whether CTMI was also liable for those funds when they went missing, even though CTMI had completed the steps for fulfilling its part of the payment process.[26] To answer that question, we turn to the parties’ agreement in this specific set of circumstances. Risk of loss is generally a matter better allocated by contract than left for courts to determine from common-law principles. Cf. LAN/STV v. Martin K. Eby Constr. Co., 435 S.W.3d 234, 240 (Tex. 2014) (“Risks of economic loss tend to be especially well suited to allocation by contract.”(quoting Restatement(Third)of Torts: Liability for Economic Harm § 1 cmt. c (Tentative Draft No. 1))). This is particularly true when, as in this case, the parties are sophisticated and well-counseled, having had “a full chance to consider how to manage the risks involved,” such as by “making a contract that assigns the risk of loss to someone else.” Id. at 240–41 (quoting Restatement (Third) of Torts: Liability for Economic Harm § 1 cmt. c (Tentative Draft No. 1)). One would especially expect clarity when—as here—sophisticated parties already in adversarial litigation with each other create a settlement agreement for the purpose of facilitating the resolution of that litigation. The parties illustrated their ability to speak with clarity even in the odd circumstance of a settlement agreement read into the record in open court. To create the escrow in the first place, they sufficiently displaced the presumptions that one side’s attorney will not be an escrow holder and that a party controls the funds held in its account. Given the importance of clarity in escrows managed by one side’s attorney over property held in that party’s (or side’s) name when the parties are in the midst of adversarial litigation, however, we hold that the creation of an escrow alone is insufficient to also shift the risk-of-loss principles underlying the parties’ joint contractual obligations to each other. This holding governs the risk of loss here. Because an attorney representing CTMI’s owners was the escrow holder and CTMI continued to own the funds, the risk of loss presumptively lies with CTMI. The purpose of default rules like this one is “to effectuate [parties'] agreements” and “to enforce them consistently and predictably so that parties may write their agreements knowing how courts will interpret them.” Richards v. State Farm Lloyds, 597 S.W.3d 492, 500 (Tex. 2020). Parties may freely depart from applicable default rules by including contractual terms in their agreements that clearly rebut the presumption, either expressly or by necessary implication. See, e.g., Perthuis v. Baylor Miraca Genetics Lab’ys, LLC, 645 S.W.3d 228, 237 (Tex. 2022). Nothing in this agreement, however, rebuts the presumption. None of its express language gives Fischer recourse exclusively to whatever is held in escrow. Nor does the settlement agreement explicitly state that Fischer bears the risk of loss. And although CTMI joined in the creation of an escrow, which ceded control to Holmes, it did not, for example, transfer the funds to Fischer or to an account owned by Holmes. It agreed to (and perhaps found it in its interest to) keep the funds in its own name. The agreement’s terms and structure are in no way inconsistent with the default rule. To the contrary, the objective indicia that might be relevant to a risk-of-loss analysis in this case corroborate that CTMI continued to bear the risk of loss: the funds were controlled by Holmes, not Fischer’s attorney; CTMI agreed for the funds to be held in an account under its name, not Fischer’s or even Holmes’s; CTMI acknowledged its ownership of and responsibility for the account by agreeing to have it held in its name and to pay the taxes on the interest it generated; and, at the time the funds went missing, CTMI still had a legal (though contingent) right to those funds, which would remain CTMI’s unless and until this Court reversed the prior judgment in CTMI’s favor. It is true that CTMI properly divested itself of control of the account; otherwise no escrow could exist. But CTMI’s nominal ownership of the account, backed by the fact that CTMI had prevailed in the court of appeals when Holmes absconded with the funds, and CTMI’s special relationship with, access to, and confidence in Holmes (an attorney for CTMI’s owners) are circumstances that gave CTMI, as compared to Fischer, a far greater ability and incentive to monitor the escrow. We acknowledge CTMI’s emphasis that the agreement’s language stated that Holmes would be “on the hook.” True enough—that placed the risk of loss on Holmes. Of course it did: an escrow holder’s liability to both parties is an essential feature of an escrow, as we have discussed at some length above. Holmes’s statement was a truism; it confirmed his acceptance of his own role, which was necessary for the escrow’s very creation. But at least in the context of this case, such a truism, by itself, does not also clearly take CTMI off the hook. Said differently, Holmes’s liability under the escrow agreement does not terminate CTMI’s liability under the settlement agreement. We express no opinion about whether a different outcome might follow if the escrow holder had been a true neutral, or if the escrowed property had not been held in CTMI’s name, or both. But because both features were present, and because nothing rebutted the resulting presumption that CTMI also retained the risk of loss, we therefore hold that CTMI remained liable to Fischer. IV For the foregoing reasons, we disagree with the court of appeals’ rationale but agree with its disposition. We affirm its judgment and we remand the case to the trial court to address pre- and post-judgment interest. Evan A. Young Justice OPINION DELIVERED: June 30, 2023
