This will decide whether recent changes to the City of Fort Worth’s (the “City”) pension plan violate the law. Under certain circumstances, the Texas Constitution forbids reduction of public pension “benefits accrued by a person.” Tex. Const, art. XVI, § 66(d). Primarily, the plaintiffs argue that the City’s pension reforms violate this “Section 66.” Two district courts have already ruled in favor of the City, sparking two separate appeals. We hereby consolidate the plaintiffs’ appeals for decision and affirm.
BACKGROUND
Fort Worth operates a defined benefits pension plan for the benefit of its employees.1 All of the plaintiffs are vested members of the plan. At the time each of the *533plaintiffs vested, the City employed a “High 3” formula. Under this formula, the three highest annual salaries received by the retiring employee were averaged to reach a base amount, which was then multiplied by the employee’s years of service and then subjected to a 3% multiplier. The plaintiffs also had the right to a cost-of-living adjustment, or “COLA.” Each of the plaintiffs in these cases had the option of choosing a 2% simple COLA or an “ad hoc COLA,” which allowed for a variable, compounded rate between 0% and 4% depending on the financial strength of the pension plan. Each of the plaintiffs chose the ad hoc COLA, which was described as an “irrevocable election.”
Like most public pension-plans in Texas, Fort Worth’s is underfunded. Over the years, Fort Worth has sought to improve the financial condition of its pension plan. In 2012, with the passage of Ordinance No. 20471-10-2012, the City made two primary changes. For new employees, it replaced the High 3 formula with a High 5 formula. Not only does this High 5 formula average the five highest paid years, it also uses a 2.5% multiplier instead of a 3% multiplier. In light of Section 66, the City sought to ensure that the reform would affect vested members of the pension plan only prospectively by adopting a bifurcated “High 3/High 5” formula. As Judge John McBryde’s opinion explains:
The calculation of benefits for employees who work both before and after the amendment is a combination of the two calculations. The part accrued before the amendment stays the same. It is only future benefits that are calculated under the new formula.
Tate v. City of Fort Worth, Tex., No. 4:15-CV-115-A, 2015 WL 4486793, at *2 (N.D. Tex. July 22, 2015).
The second noteworthy change concerned the COLA. The City eliminated cost-of-living adjustments for future employees, provided that current employees would henceforth receive a simple 2% COLA, and allowed current employees who had previously taken the ad hoc COLA “to revert to 2% simple.”
Due to a collective bargaining agreement, City firefighters were not affected by Ordinance No. 20471-10-2012. Shortly after that agreement expired, however, the City imposed essentially the same reform on its firefighters with Ordinance No. 201510-10-2014. The two lawsuits before us challenge those ordinances. One is brought by a pair of police officers, the other by a trio of firefighters. We refer to the challenged ordinances collectively as the “Pension Reform.”
Ultimately, both cases were resolved at the summary judgment stage. On April 7, 2015 Judge Terry Means rendered judgment in favor of the City and against the police officers in Case No. 15-10416, finding that the Pension Reform complied with Section 66 and that all other claims were contingent on a threshold finding of incompatibility with Section 66. Judge McBryde issued a similar ruling against the firefighters in Case No. 15-10796 on June 22, 2015. Both sets of plaintiffs timely appealed.
STANDARD OF REVIEW
The district courts’ orders granting summary judgment are subject to de novo review. Time Warner Cable, Inc. v. Hudson, 667 F.3d.630, 638 (5th Cir. 2012). Likewise, the constitutionality of the Pension Reform is a question of law subject to de novo review. See Nat’l Fed’n of the Blind of Tex., Inc. v. Abbott, 647 F.3d 202, 208 (5th Cir. 2011).
Despite this standard of review, the plaintiffs argue that we should defer to a relevant opinion of the Texas Attorney *534General. See Tex. Att’y Gen. Op. GA-0615, 2008 WL 982266 (2008) (hereinafter, the “AG Opinion”). We consider that opinion, of course, but it does not change our task. Because the Texas Supreme Court has not yet interpreted Section 66, our Erie function is to predict how it would rule. McCaig v. Wells Fargo Bank (Texas), N.A., 788 F.3d 463, 472 (5th Cir. 2015). The Texas Supreme Court would consider the AG Opinion “persuasive” but “not controlling”, Holmes v. Morales, 924 S.W.2d 920, 924 (Tex. 1996). We accord it the same stature. Any other approach would put us out of step with the Texas Supreme Court and impair our ability to accurately prognosticate that court’s ruling. See Batts v. Tow-Motor Forklift Co., 66 F.3d 743, 750 (5th Cir. 1995) (explaining our duty to do “no more” than “predict how the state court will decide a question”).
ANALYSIS
I.
A.
Section 66(d) provides:
On or after the effective date of this section, a change in service or disability retirement benefits or death benefits of a retirement system may not reduce or otherwise impair benefits accrued by a person if the person:
(1) could have terniinated employment or has terminated employment before the effective date of the change; and
(2) would have been eligible for those benefits, without accumulating additional service under the retirement system, on any date on or after the effective date of the change had the change not occurred.
Tex. Const, art. XVI, § 66(d).
We must decide whether Section 66 prohibits pension reform that would decrease expected but as-yet unearned benefits. For the reasons that follow, we conclude that such an interpretation is inconsistent with Section 66’s text, which prohibits only the reduction or impairment of “benefits accrued.”
We interpret the Texas Constitution as would Texas courts. See Cerda v. 2004-EQR1 L.L.C., 612 F.3d 781, 786 (5th Cir. 2010). Texas courts “presume the language of the Constitution was carefully selected, interpret words as they are generally understood, and rely heavily on the literal text.” In re Allcat Claims Serv., L.P., 356 S.W.3d 455, 466 (Tex. 2011).
This case comes down to the meaning of the word accrued — or whether it means anything at all. Section 66(d) has a two-part structure. The introduction of subsection (d) provides which benefits are covered — “benefits ¿ccrued by a person”— while subparagraphs (d)(1) and (d)(2) provide who is covered — vested present and past employees. In short, Section 66(d) prohibits the impairment of accrued benefits for vested employees. This understanding essentially resolves the case.
There is an understood difference between the concepts of benefit accrual and vesting. See, e.g., Shanks v. Treadway, 110 S.W.3d 444, 445 n. 2 (Tex. 2003) (“Pension plan benefits become vested when the employee has an unconditional ownership interest in them; that is, the employee has the right to receive the accrued benefits upon retirement whether or not he is working for the same employer.”); Dewey v. Dewey, 745 S.W.2d 514, 518 (Tex. App. 1988) (recognizing that “accrued benefits” begin to accumulate prior to vesting). Benefits accrue on an ongoing basis as service is performed, and accrued benefits are those benefits that have been earned to date. Meanwhile, vesting is a one-time event giving rise to a right to the accrued *535benefits.2 “In summary, the notion of benefit accrual quantifies actual benefit accumulations. The concept of vesting determines the nature of an employee’s legal right in the accrued benefits.” Brown, 37 Baylor L. Rev. at 123.3 By its terms, Section 66 prohibits only the reduction or impairment of accrued benefits, and the plaintiffs cannot complain about the reduction of benefits that have not yet accrued.
The plaintiffs agree that Section 66 has a two-part structure and that sub-paragraphs (d)(1) and (d)(2) represent a vesting requirement. But they don’t think the term “accrued” means anything. As the plaintiffs see it, this “dispute ... boils down to whether ‘benefits accrued’ means merely money, or instead the right to the method for determining the compensation base for calculating retirement benefits.” “If ‘benefits’ means the right to the formula, then the right becomes ‘accrued’ under the Constitution once the employee vests in the plan.” Under their proposed interpretation, the term “accrued” repeats the vesting requirement and has no operative effect.4 In short, according to the plaintiffs, a benefit formula accrues at the moment of vesting and becomes constitutionally protected, meaning that even wholly prospective formula adjustments are foreclosed by Section 66.
The simple observation that accrual and vesting are distinct and vital concepts in the pension plan lexicon renders plaintiffs’ interpretation unsustainable. The rule that Texas courts “refuse, whenever possible, to construe constitutional language in a way that renders it idle or inoperative” likewise precludes such an interpretation. See Spradlin v. Jim Walter Homes, Inc., 34 S.W.3d 578, 580 (Tex. 2000) (rejecting an interpretation that would yield “an immediate redundancy”). Examine the plaintiffs’ key predicate claim: Does the term “benefits” include a “right to the formula”?
Traditional tools of statutory interpretation yield a clear answer. The term “benefits” refers to payments and does not encompass the formula by which those payments are calculated. There are numerous indications that the term “benefits” refers only to payments. We start with Section 66 itself. When addressing other benefits, Section 66 expressly refers to “disability benefits” that “are no longer payable.” Section 66(c) (emphasis added). This makes sense only if benefits are generally something to be paid.
*536Further, Section 66 must be read in pari materia with Section 67, entitled “State and local retirement systems.” Duncan v. Gabler, 147 Tex. 229, 215 S.W.2d 155, 159 (1948) (“An important established rule for construing the Constitution is that all of its provisions affecting the same thing must be construed together.... ”). In Section 67, the term “benefits” represents a payment. Section 67 covers limitations on the right to “receive benefits.” Section 67(a)(2). It references “benefits payable.” Section 67(d)(1). It ensures that local retirement systems hold assets “for the exclusive purposes of providing benefits.” Section 67(f)(2). The constitutional requirement that “[f]inancing of benefits must be based on sound actuarial principles” makes sense only if benefits refers to payments of money. Section 67(a)(1). So, too, the phrase “fractional benefit.” Section 67(a)(2). Similarly, the requirement that “[benefits under these systems must be reasonably related to participant tenure and contributions” essentially requires that the measure of benefits be the product of a lawful formula. Section 67(c)(2). Finally, Section 67 refers to “the benefit formula used,” a statement that would be inscrutable if the term “benefits” encompassed the benefit formula.5 Section 67(a)(2).
The plaintiffs direct us to Black’s Law Dictionary and argue that the word “benefits” means “privilege,” as in “[t]he right to participate in a plan under a particular formula.” But that same dictionary provides an alternative definition: “Financial assistance that is received from an employer, insurance, or public program (such as social security) in time of sickness, disability, or unemployment.” Blaok’s Law Dictionary 188 (10th ed. 2014). Between the two possibilities — one generic, the other contextual — it is clear which is more apt. Likewise, the Glossary of Insurance Terms defines the term “benefit” as “The amount to be paid to a participant of a retirement plan or to the participant’s beneficiary at retirement, at death, or at termination of services.” GLOSSARY OF INSURANCE TERMS 27 (6th ed. 1996).
The plaintiffs also rely on the AG Opinion, which found that the term “benefits” to encompass the benefit formula. We have carefully considered the AG Opinion. For the following reasons, we do not believe that it accurately predicts how the Texas Supreme Court will interpret Section 66.
The AG Opinion begins with Section 66’s text but finds it ambiguous and then considers legislative history, which it also finds unhelpful. See Tex. Att’y Gen. Op. No. GA-0615, at 2-6. By the end of the analysis, the text of Section 66 has been left behind and the opinion is instead based on holdings from other state supreme courts, particularly those of New York, Illinois, and Alaska:
The New York, Illinois, and Alaska court decisions suggest that the authorized method for determining the base compensation of vested employees is a *537constitutionally protected “right” that “accrues” upon vesting.... In the absence of any Texas judicial authority, we believe this case law is persuasive authority, even recognizing that the Texas constitutional language is similar, but not identical to these states’ constitutional language.
Id. at 7.
Already, we have found that Section 66’s text is not ambiguous. It is certainly not such a muddle that it must be thrown out and replaced with case law from other states — states with different policy objectives, value judgments, and histories. And yet that is what the AG Opinion does, concluding that Section 66 — a lengthy and detailed multi-part provision — is to be interpreted identically as, for example, New York’s sweeping constitutional decree that “membership in any pension or retirement system of the state or of a civil division thereof shall be a contractual relationship, the benefits of which shall not be diminished or impaired.” N.Y. Const, art. V, § 7.
Just recently, the Texas Supreme Court issued a reminder that when interpreting the Texas Constitution, courts must interpret the Texas Constitution, and not resort reflexively to the constitutions of other states. Patel v. Texas Dep’t of Licensing & Regulation, 469 S.W.3d 69, 91 (Tex. 2015) (“[WJhether [a particular licensing requirement] violates the Texas Constitution is not determined by the relationship between other states’ statutes and regulations and their respective constitutions.”); see also id. at 98 (Willett, J., concurring) (“[W]hat happens in the Aloha State makes not the slightest constitutional difference in the Lone Star State.”).
In the context of public pension plans, it is particularly problematic to assume that Texans suddenly decided (with the enactment of a constitutional provision that looks nothing like the supposedly similar constitutional provisions of other states) that Texas would henceforth copy states like New York and Illinois with respect to protecting public employees’ retirement plans. When it comes to public pension protection, Texas is known to be an outlier.6 In 1937, the Texas Supreme Court decided City of Dallas v. Trammell and held that pensioners’ rights to accrued benefits were subject to the legislative power of the state “to amend, modify, or repeal the law upon which the pension system is erected.” 129 Tex. 150, 101 S.W.2d 1009, 1014 (1937). The ruling meant that C.W. Trammell, a retired Dallas police officer whose monthly pension was cut from $183.33 to $72.16, had no recourse. While other states enacted laws to protect public pensions from similar cuts, Texas held its course — until the enactment of Section 66. As one Texas appellate court put it, Section 66 “was proposed and adopted specifically to change the result of the Trammell decision, albeit 70 years later.” Davidson v. McLennan Cty. Appraisal Dist., No. 10-11-00061-CV, 2012 WL 3799149, at *5 (Tex. App. Aug. 30, 2012) (mem. op.).
As we have interpreted it, Section 66 reverses the core unfairness of the Tram-mell decision by ensuring that earned ben-*538efíts cannot be reduced. By going no further, our interpretation of Section 66 stays true to Texas’ long-held flexible approach permitting municipalities to revise their pension plans in light of changing economic conditions. The AG Opinion not only discards Section 66’s text and replaces it with the text of other states’ constitutions, it also replaces Texas’ history and policy objectives with the history and policy objectives of disparate states. In doing so, the AG Opinion takes a path we do not expect the Texas Supreme Court to follow.7
B.
We have concluded that Section 66 permits prospective changes to the pension plans of the public employees within its reach. If the changes to the pension plan impact only benefits that have not yet accrued, amendment is permissible. We now must determine whether the Pension Reform complies with this restraint.
The plaintiffs do not dispute that, under the interpretation adopted by the district courts and now by this court, the bifurcated High 3/High 5 approach is permissible. The reform has been designed to protect all accrued benefits while impacting only the rate at which future benefits accrue. This aspect of the Pension Reform therefore passes constitutional muster.
The plaintiffs view the COLA reform differently. According to them, even *539under the district courts’ understanding of Section 66, the changes to the COLA are unconstitutional.8 According to the plaintiffs’ evidence, if the election is not reopened, their ad hoc COLA election can be expected to carry a 1% effective rate. If the election is re-opened and pension plan members flock to the 2% fixed option, the value of the ad hoc COLA election will drop to 0%. This predicted drop is the reduction or impairment of which they complain.9
The plaintiffs chose a variable rate, guaranteed only to be within 0% and 4%. They were not guaranteed a 1% rate of return. The enactment of Section 66 coupled with their evidence of a presently expected 1% return does not provide them constitutional protection against the risk of downward fluctuations inherent in variable rates. Any number of changes to the pension plan made by the City between now and the various dates upon which the various plaintiffs will retire could cause the variable-rate ad hoc COLA to decrease in value. Under the plaintiffs’ reasoning, all such changes violate Section 66. We reject this argument. Section 66 did not turn the plaintiffs’ variable-rate COLA into a one-way ratchet capable only of upward movement. In all challenged respects, the Pension Reform complies with Section 66.
II.
The plaintiffs attack the Pension Reform on additional grounds, which we now consider. According to Richard Van Houten, Jr. and Stephen Hall, the plaintiffs of Case No. 15-10416, only the Texas state legislature has the Trammell “reserved power” to amend pension plans and thus abrogate contractual rights, meaning the Ordinances represent a breach of contract and violation of Texas constitution’s contracts clause. This argument is foreclosed. Klumb v. Houston Mun. Employees Pension Sys., 458 S.W.3d 1, 16 (Tex. 2015) (“[N]o vested property right exists when a pension fund can be amended or abolished by the governing authority; it makes no difference whether the authority with the power to abolish the pension system is the Legislature or some other entity.”). Here, the “governing authority” is the City. See Tex. Rev. Civ. Stat. Ann. art. 6243i §§ 1.02(3), 4.03(a).
All of the plaintiffs argue that the Pension Reform violates the United States Constitution’s contracts clause and takings clause. Neither the contracts clause nor the takings clause create property rights. Rather, they protect property rights. Accordingly, the existence of the right depends on state rather than federal law. As the Supreme Court has observed, with reference to the takings clause, “[b]ecause the Constitution protects rather than creates property interests, the existence of a property interest is determined by reference to ‘existing rules or understandings that stem from an independent source such as state law.’ ” Phillips v. Washington Legal Found., 524 U.S. 156, 164, 118 S.Ct. 1925, 1930, 141 L.Ed.2d 174 (1998) (quoting Bd. of Regents of State Colleges v. Roth, 408 U.S. 564, 577, 92 S.Ct. 2701, 2709, 33 L.Ed.2d 548 (1972)).
Under Texas law, to the extent there is any sort of contractual .right to pension plan benefits, it is a right expressly “made subject to the reserved power of the Legislature to amend, modify, or repeal the law upon which the pension system is erected, *540and this necessarily constitutes a qualification upon the anticipated pension and a reserved right to terminate or dimmish it.” Trammell, 101 S.W.2d at 1014. As we have already seen, the Texas Supreme Court reaffirmed this view in Klumb. See 458 S.W.3d at 16. Thus, Texas law remains “clear that a person’s property right in a public pension is subordinate to the state’s power to determine to whom benefits are to be paid, to set conditions for receiving such benefits, to modify benefits paid, or to abolish the pension and accrued benefits altogether.” Kunin v. Feofanov, 69 F.3d 59, 63 (5th Cir. 1995). As a matter of state law, the right to public pension benefits in Texas is subject to legislative power. Legislative reduction of such benefits therefore cannot be the basis of a U.S. Constitution contracts clause or takings clause challenge.
CONCLUSION
For the foregoing reasons, the judgments in favor of the City in the cases numbered 15-10416 and 15-10796 are AFFIRMED.10
. "Generally, an employee participating in a defined benefit plan will receive a future benefit based on a specified formula that often takes into account earnings, length of service, or both.” Shanks v. Treadway, 110 S.W.3d 444, 445 (Tex. 2003) (citing Steven R. Brown, Comment, An Interdisciplinary Analysis of the Division of Pension Benefits in Divorce and Post-Judgment Partition Actions: Cures for the Inequities in Berry v. Berry, 37 Baylor L. Rev. 107, 115-16 (1985)).
.For another explanation, see John F. Buckley IV, ERISA Law Answer Book § 11-2 (8th ed. 2014).
The concepts of "accrued” and ‘Vested” are related but not the same. "Accrual” refers to the amount of benefits a plan has earned to date and is usually expressed in the form of an annual benefit commencing at normal retirement age. “Vesting” refers to the point in time at which accrued benefits become nonforfeitable. In other words, a plan’s accrual provisions provide a formula for calculating the amount of a normal retirement benefit that a participant has earned at any given time-, vesting provisions do not affect the amount of the accrued benefit but rather govern whether all or a portion of the accrued benefit is non-forfeitable.
Id. (emphases added) (citations omitted).
. Mr. Brown’s comment has been cited several times by Texas courts, including by the Texas Supreme Court in describing basic features of pension plans as we do here. See Shanks, 110 S.W.3d at 445 n. 1 & 2. Mr. Brown also noted that it "is essential to recognize and maintain the distinctions between ‘vesting’ and 'accrual.' ” 37 Baylor L. Rev. at 117.
. Plaintiffs describe the term “accrued” as just “a single verb” worthy of little attention, contend that term "[ajccrued does nothing to indicate when the benefits become protected, and does not act as a limitation of any kind,” and claim "the use of the term ‘accrued’ was not intended to create any additional limitation.”
. The Government Code’s treatment of "benefits” further supports our conclusion. For example, the phrase "Public retirement bill or resolution” is defined to mean "a bill or resolution that proposes to change the amount or number of benefits or participation in benefits of a state-financed public retirement system. ...” Tex. Gov. Code Ann. § 802.305(1X1) (emphasis added). Other examples abound. See Tex. Gov. Code Ann. § 802.203(a)(1)(A) (“providing benefits”); Tex. Gov. Code Ann. § 802.207(b) ("paying benefits”); Tex. Gov. Code Ann. § 802.1024(a-2), (b) ("overpayment of benefits”); Tex. Gov. Code Ann. § 803.302(b) ("amount of a benefit payable,” located in section entitled “Computation of Certain Benefits”); Tex. Gov. Code Ann. § 803.401(b)(2) ("benefits payable”); Tex. Gov. Code Ann. § 804.001(1), (4) (same); Tex. Gov. Code Ann. § 804.003(c) (same) Tex. Gov. Code Ann. § 805.007(a) (same).
. See, e.g., Anna K. Selby, Note, Pensions in A Pinch: Why Texas Should Reconsider Its Policies on Public Retirement Benefit Protection, 43 Tex. Tech L. Rev. 1211, 1230 (2011) (identifying Texas as one of only two states that takes a "gratuity approach” to public pensions, meaning pension benefits are viewed as gratuity rather than a contractual or statutory right); Amy Monahan, Public Pension Plan Reform: The Legal Framework, 5 Educ. Fin. & Pol’y 617, 621 (2010) (same); T. Leigh Anen-son, Alex Slabaugh, Karen Eilers Lahey, Reforming Public Pensions, 33 Yale L. & Pol'y Rev. 1, 15 (2014) (noting Texas' retention of the gratuity approach while "[a]n overwhelming majority of states ... have transformed tradition and retreated from the notion of pensions as unprotected gratuities”).
. The Dissent finds our interpretation "not unreasonable” but perceives a flaw in our analysis. Respectfully, the Dissent’s reasoning is perplexing. The Dissent focuses on our consideration of the term "benefits” and claims that we “ignore instances” where the term appears to refer to a formula rather than to a payment. (Dissent at 540.) One of these instances is found "in section 66(d) itself,” no less. {Id.) We did not overlook that use of the word; our very enterprise was to ascertain the meaning of the term "benefits" within Section 66(d). Surely the Dissent does not believe that the word, used four times in Section 66(d), might bear different meanings with each use. According to the Dissent, as used in Section 66(d), "[t]he phrase 'a change in ... benefits’ plainly encompasses a change in the formula." (Id.) True, a change to the benefits (i.e., the payment) can be effected via a change to the benefit formula. That does not mean, however, that the term "benefits” is any broader here than it is in other parts of Section 66, Section 67, or the Government Code. We interpret the phrase "a change in ... benefits” to include any change to the pensioner’s bottom-line — the actual payments. Thus, Section 66 is not a narrow reform myopically concerned with changes to benefit formulas. Read this way, Section 66(d) operates efficaciously no matter the particular design of the particular pension plan, an important feature because not all plans are so formula-dependent as the defined benefit plans at issue today. Thus, our reading treats Section 66 as an objective-oriented reform that contemplates any "change in ... benefits,” whether that change is effected via revision of the benefit formula or some other way. To be protected, however, the benefits must be "accrued.” It says so in the text.
The next use of the word "benefits" that we have purportedly ignored appears in Section 66(e). The Dissent’s reasoning has the virtue of simplicity, but not soundness. Section 66(e) provides: "Benefits granted to a retiree or other annuitant before the effective date of this section and in effect on that date may not be reduced or otherwise impaired.” According to the Dissent, ”[o]ne would not describe a monetary payment as being 'in effect' on a fixed date, but one would certainly describe a formula for calculating monetary payments that way.” (Dissent at 541.) This is facially uncompelling because of course one might describe monthly pension payments of a set amount as being "in effect” as of the date Section 66 was enacted. Further, if as the Dissent believes, Section 66(e) protects a benefit formula and the term "accrued” is used by Section 66(d) in an idiosyncratic way to denote the moment in time at which a right to a benefit formula becomes inviolable, then Section 66(e) should apply to "benefits accrued” as Section 66(d) does. It does not. The absence of the term "accrued” in Section 66(e) confirms that the words of Section 66 were chosen deliberately.
. We share the parties' mutual assumption that, under Texas law, the ad hoc COLA represents an “accrued benefit.”.
. Strangely, the plaintiffs find no comfort in the fact that switching to the 2% fixed COLA would, according to their own evidence, double the value of their COLA.
. The Van Houten plaintiffs’ motion to certify questions to the Texas Supreme Court is DENIED. Finding Section 66 to be ambiguous, the Dissent would grant the motion. That is not the standard, however. “We do not lightly abdicate our mandate to decide issues of state law” and are " 'slow to honor a request for certification from a party who chose to invoke federal jurisdiction.’ ” Jefferson v. Lead Indus. Ass’n, Inc., 106 F.3d 1245, 1248 (5th Cir. 1997) (quoting 17A Charles A. Wright, Arthur R. Miller, Edward H. Cooper, Federal Practice and Procedure § 4248, at 176 (1988)). We will not certify even “important and complex” questions where the answer is “sufficiently clear.” Patterson v. Mobil Oil Corp., 335 F.3d 476, 487 (5th Cir. 2003); see also Cerda v. 2004-EQR1 L.L.C., 612 F.3d 781, 785 n. 8 (5th Cir. 2010) (applying "sufficiently clear” standard); Moore v. State Farm Fire & Cos. Co., 556 F.3d 264, 269 (5th Cir. 2009) (same). At the appellate level, we expect even the losing party to present plausible arguments. Likewise, we deal routinely with ambiguous statutes. Were the arguments implausible and the statutes unambiguous, our role would be ornamental. We do not decline our obligation to resolve cases simply because the answer is not facially evident to all jurists. See Transcon. Gas Pipeline Corp. v. Transp. Ins. Co., 958 F.2d 622, 623 (5th Cir. 1992). The question of law presented by this case is undoubtedly an important one, but it is one we can answer "with confidence.” Hughes v. Tobacco Inst., Inc., 278 F.3d 417, 426 (5th Cir. 2001). Our duty, therefore, is to decide.