concurring and dissenting:
I concur in the majority’s opinion with the exception of section II.D. As to this subject — the meaning of “substantially equal” — I respectfully dissent from the decision not to certify this question to the Texas Supreme Court. Under Texas law, “[t]he Supreme Court of Texas may answer questions of law certified to it by any federal appellate court if the certifying court is presented with determinative questions of Texas law having no controlling Supreme Court precedent.” Tex. R.App. P. 58.1; see also Tex. Const. art. V, § 3-c(a). Our court should certify the question of whether sections 50(a)(6)(L) and 50(a)(6)(O) allow as great a variation in monthly payments as the Cerdas’ loan provides.
Great care is required when we interpret article XVI, section 50 of the Texas Constitution because of the harsh consequences that follow from its violation, see Tex. Const. art. XVI, § 50(a)(6)(Q)(x), and the broad reach of the section, see, e.g., James W. Paulsen, Acquiring Separate Property on Credit: A Review and Proposed Revision of Texas Marital Property Doctrine, 37 St. Mary’s L.J. 675, 679-80 & nn. 12-13 (2006) (reviewing government estimates of scope of home equity lending following the section 50 amendment and concluding that as many as fifteen percent of Texas homeowners have loans governed by section 50(a)). In the past, we have not hesitated to certify questions concerning the proper interpretation of this provision. For example, in 1999, in Stringer v. Cen *798dant Mortgage Corp., 199 F.3d 190, 191-92 (5th Cir.1999), we certified the question of whether, where the notice provision of the amendment conflicted with the substantive rights provision on the question, a lender could permissibly require a borrower to pay off third-party debt with the proceeds of the loan. A few months later, the Texas Supreme Court explained that the substantive provisions control over the notice provisions. Stringer v. Cendant Mortgage Corp., 23 S.W.3d 353, 357-58 (Tex.2000). In 2001, in Doody v. Ameriquest Mortgage Co., 242 F.3d 286, 290 (5th Cir.2001), we certified the question of whether a lender may bring itself into compliance with the three percent fee cap of section 50(a)(6)(E) by subsequently refunding the excess portion of the fee to the borrower. Later that year, the Texas Supreme Court answered that a timely refund cures the violation. Doody v. Ameriquest Mortgage Co., 49 S.W.3d 342, 345-47 (Tex.2001). Most recently, in 2005, in Norris v. Thomas (In re Norris), 413 F.3d 526, 530 (5th Cir.2005), we certified the question of whether a houseboat is a homestead under sections 50 and 51 of article XVI. The Texas Supreme Court responded that a boat could not be a homestead. Norris v. Thomas, 215 S.W.3d 851, 852 (Tex.2007). Indeed, certification from this court has generated much of the Texas Supreme Court’s jurisprudence concerning section 50.
The majority’s holding on this issue in this case has the potential for a much greater impact than those issues presented in Stringer, Doody, or Norris. In contrast to those cases, this case effectively adjudicates the validity of thousands of adjustable-rate home equity loans originated between at least the 1997 enactment of the amendment and the 2004 effective date of the regulations that now provide a safe harbor.
Certification is, of course, discretionary both by our court in certifying and the Texas Supreme Court in accepting the question. Patterson v. Mobil Oil Corp., 335 F.3d 476, 487 (5th Cir.2003); Tex. R.App. P. 58.1. Undoubtedly, construction of a state constitution in a case of first impression is the type of matter appropriate for certification. In such an area, we have generally declined to exercise that discretion only where the controlling state law issues presented by the case are, though “important and complex, ... sufficiently clear.” Id. That is patently not the case here. Rather, we are faced with two provisions of the Texas Constitution that four Texas administrative agencies acknowledged are “ambiguous” when read together. See Regulatory Commentary 8.
The majority resolves this ambiguity by relying upon prior non-binding agency determinations. Admittedly, however, we have no case law whatsoever construing these sections or resolving this ambiguity. Were the provisions in question clear, I would have no hesitation in following the agency determination which, prior to 2004, is entitled to some, though not controlling, deference. See Tex. Bankers Ass’n v. Ass’n of Cmty. Orgs. for Reform Now (ACORN), 303 S.W.3d 404, 408-09 (Tex. App. — Austin 2010, pet. filed) (holding that the constitutionally-authorized regulatory interpretations of article XVI, section 50 are “ ‘entitled to serious consideration, so long as the construction is reasonable and does not contradict the plain language of the statute’ ”) (quoting Tarrant Appraisal Dist. v. Moore, 845 S.W.2d 820, 823 (Tex. 1993)); Stringer, 23 S.W.3d at 357 (characterizing the Regulatory Commentary as “advisory and not authoritative” but looking to it as “support[ing]” the court’s conclusion). However, here, the construction of “substantially equal” payments to mean only “all payments that are not balloon payments” is arguably contrary to the plain language of the section. See, e.g., Moore, 845 S.W.2d at 823; see also La*799Salle Bank Nat’l Ass’n v. White, 246 S.W.3d 616, 619 (Tex.2007) (“When interpreting the Texas Constitution, we ‘rely heavily on its literal text and must give effect to its plain language.’ ” (quoting Stringer, 23 S.W.3d at 355)).
The majority’s approach is certainly plausible. But section 50(a)(6)(L) can easily be read to bar more than simply balloon payments: indeed, the language is much broader than the straightforward 1994 federal statutory language that does expressly prohibit only balloon payments. Compare Tex. Const. art. XVI, § 50(a)(6)(L) with 15 U.S.C. § 1639(e) (“No balloon payments. [Certain] mortgage[s] ... having a term of less than 5 years may not include terms under which the aggregate amount of the regular periodic payments would not fully amortize the outstanding principal balance.”). The Cerdas’ loan, if enforced as written, would allow payments ranging from $2,950 to almost $5,000 per month over the course of the loan. It is difficult to reconcile the plain English-language meaning of “substantially equal payments” with such huge variations. It is hard to imagine how, if the Texas legislature were to disagree with the majority’s reading, the provision even could be rewritten to, in fact, require substantially equal payments.
The danger of such wide fluctuations is real, as evidenced by the recent “subprime mortgage meltdown” during which the so-called “payment shock” from sharp increases in variable interest rates did result in many homeowners across the country losing their homes. See, e.g., Andrew J. Ceresney, Gordon Eng & Sean R. Nuttall, Regulatory Investigations and the Credit Crisis, 46 Am.Crim. L.Rev. 225, 261 (2009) (“One of the more common complaints ... concerns homeowners confronted with ‘payment shock’ regarding adjustable rate mortgages---- Often these loans were originated with very low ‘teaser’ rates ... during the first few months of the mortgage; at the end of the low interest-paying period, however, the rates reset at substantially higher rates, leaving homeowners unable to afford monthly payments and placing them at risk of foreclosure.”); Bernhard Grossfeld & Hansjoerg Heppe, The 2008 Bankruptcy of Literacy: A Legal Analysis of the Subprime Mortgage Fiasco, 15 Law & Bus. Rev. Am. 713, 722-29 (2009).1
When the plain language of the section is viewed through the prism of Texas’s conservatism on home equity lending in light of the state’s long history of “carefully protecting] the family homestead from foreclosure,” see LaSalle Bank, 246 S.W.3d at 618, the Cerdas’ argument that “substantially equal” means only small variations are permitted in the monthly payments seems at least equally plausible to the meaning advanced by the lender and adopted by the majority, perhaps more so. Indeed, the relationship between “payment shock” and Texas’s historical desire to protect the homestead is especially (though not exclusively) relevant to home equity loans, which are, by their nature, more likely than purchase money loans to be made to borrowers on fixed incomes for whom increases in monthly payments are particularly problematic. Cf. Kristine M. Young, Note, The Aging Population and Maturing Mortgage Loans: Ensuring a Secure Financial Lifeline for the Elderly Through Mortgage Lending, 16 Elder L.J. 477, 487-91 (2009) (expressing concern that variable-rate mortgages are likely inappropriate for older borrowers with sig*800nificant home equity but living on fixed incomes).
This question thus is one of the few examples of “genuinely unsettled matters of state law” where certification is most appropriate. See Jefferson v. Lead Indus. Ass’n, 106 F.3d 1245, 1247 (5th Cir.1997); see also De Checa v. Diagnostic Ctr. Hosp., Inc., 967 F.2d 126, 129 (5th Cir. 1992) (certifying a question of Texas law where there was “an apparent conflict between” multiple provisions of the same statute). The majority’s decision — as would be any decision on this point by a federal court at this juncture — is an “Erie guess” that is just that: a guess. It is, moreover, a guess as to a proposition of state constitutional law that potentially affects the rights of thousands of Texas homeowners and their lenders and note-holders.
Accordingly, I respectfully submit that the appropriate course of action on this issue would be to certify the question to the Texas Supreme Court, and I dissent from the majority’s decision not to do so.
. Although the change from one month to the next in the Cerdas' loan is more gradual than may occur with the so-called “teaser rates”— where originally very low rales are suddenly raised precipitously — the contemplated payments are not “substantially equal” as that term would be understood by a layperson.