OPINION
Bob Pemberton, JusticeState Farm Lloyds appeals a district court judgment affirming a final order of the Commissioner of Insurance determining that it had charged consumers homeowners insurance premiums that were excessive for several years during the 2000s and ordering refunds with interest. We will affirm the Commissioner’s order with respect to the first year at issue, but must reverse as to the remaining years and remand for further proceedings. We also hold that there was reversible error in the Commissioner’s award of interest on refunds he determined to be due.
BACKGROUND
This appeal arises from the proceedings on remand this Court required in Geeslin v. State Farm Lloyds (State Farm Lloyds I),2 and we will refer the reader to that opinion for a comprehensive explanation of statutory context and procedural prologue. On remand, the Texas Department of Insurance (TDI) noticed a “re-hearing” to determine whether the Commissioner should affirm the reduction TDI had ordered in the “initial rate” State Farm Lloyds had filed under former art. 5.26-1 of the Insurance Code,3 or require a greater or lesser reduction instead.4 The Office of Public Insurance Counsel (OPIC) intervened. The hearing was ultimately conducted over several days in 2009, and the record was closed in November of that year.
The State Farm Lloyds “rate” in dispute at the hearing referred to “the cost of insurance per exposure unit ... with an *92adjustment to account for the treatment of expenses, profit, and individual insurer variation in loss experience, and before any application of individual risk variations.” 5 Former art. 5.26-1 required that the “initial rate” filed by State Farm Lloyds and ultimately'approved by TDI be “just, reasonable, adequate, not excessive, and not unfairly discriminatory for the risks to which it applies.”6 These requirements, simply put, mean that an insurer charges consumers a price sufficient to recover both its projected expenses of asT suming risks under its policy and a profit yielding a “reasonable” rate of return on its capital, but not a profit “unreasonably” higher than this.7 The requirements also operate against a constitutional backdrop. This Court has held — most recently in State Farm Lloyds I — that government-set rates are deemed to effect an unconstitutional taking of an insurer’s property if the insurer cannot recover both its projected “operating expenses” and a “reasonable rate of return” on its capital.8
In advocating their views of a reasonable rate that would comply with these statutory and constitutional requirements, the parties, generally speaking, followed a common methodology in which they calculated an “indicated” rate comprised of the sum of several specified categories of projected expenses per exposure unit (e.g., estimated payments or losses related to hurricanes), plus an additional “underwriting profit” provision calculated so as to ensure that State Farm Lloyds obtained an overall profit (including net income from both premiums and investments) suf*93ficient to provide a rate of return on its capital equivalent to that which it could obtain in alternative investments of equivalent risk (i.e., a return compensating it for the “opportunity cost” of its capital).9 The indicated rate would then be compared to the premium State Farm Lloyds was projected to earn per exposure unit under the filed initial rate it had charged its customers. If the indicated rate was less than State Farm Lloyds’s projected premiums, its initial rate would be deemed excessive. Conversely, if the indicated rate equaled or exceeded the projected premiums, a rate reduction would be considered confiscatory. Although generally following the same method for calculating the indicated rate and any required reduction, the parties differed with respect to the cost or expense elements that should be included in the indicated rate and their amount.
The issues were further framed by an unusual procedural posture on remand. As explained in State Farm Lloyds I, former art. 5.26-1 (a component of the 78th Legislature’s watershed S.B. 14) had governed the first of three phases through which the Legislature imposed rate regulation on what had become a largely unregulated Texas homeowners insurance market.10 Under former art. 5.26-1, effective June 11, 2003 through August 31, 2004, insurers were required to file their initial regulated rates within 20 days of the statute’s effective date and implement them.11 After this initial filing, former art. 5.142, effective June 11, 2003, through November 30, 2004, provided temporary rate-regulation procedures.12 Under former art. 5.142, insurers were required to file their rates with TDI and await the Commissioner’s approval before implementing them.13 Then, beginning on December 1, 2004, a permanent file-and-use regime, governed by former art. 5.13-2 of the Insurance Code (now codified as chapter 2251 of that code)14 took effect whereby insurers file their rates with TDI and implement them, subject to the Commissioner’s power to disapprove the rates before they go into effect or disapprove further use of the rates after they go into effect.15 In State Farm Lloyds’s case, TDI had ordered, and the Commissioner had affirmed, a 12% reduction in the insurer’s initial filed rate. This reduction, all other things being equal, would have taken effect on September 7, 2003. However, in former art. 5.26-I, the Legislature had allowed an insurer who sought judicial review of a rate-reduction order, as State Farm Lloyds had, the *94option of charging its filed rate while litigation was pending,16 subject to mandatory refunds of “the difference in overcharged premium to each policyholder, plus interest,” if “on final appeal the court upholds the commissioner’s determination that the insurer’s rates are excessive.”17 State Farm had availed itself of this option and continued charging its initial filed rate (which we will term its “implemented rate”), notwithstanding the Commissioner’s opposition to it, pending litigation over its validity.
As events turned out, the first round of litigation over State Farm Lloyds’s implemented rate had not concluded until 2008, when this Court issued its State Farm Lloyds I decision. In remanding the case to the Commissioner for further proceedings despite former art. 5.26-1’s expiration on September 1, 2004, this Court observed that the Legislature had provided that the expiration ‘“does not affect an action or proceeding against an insurer subject to that law for failure to comply with that law before its expiration, regardless of when the action or proceeding was commenced, and that law is continued in effect for that purpose.’”18 In the meantime, however, State Farm Lloyds’s implemented rate has remained in effect not only as its operative initial rate under former art. 5.26-1, but also as the insurer’s operative rate under former art. 5.142, which had contained parallel provisions authorizing the insurer to charge its desired rate • pending appeal, subject to refunds with interest if the rate was later found excessive.19 State Farm Lloyds had similarly continued charging the implemented rate as the file-and-use regime began on December 1, 2004, and would ultimately do so, despite TDI’s attempts to restrict the insurer from continuing to charge the rate and the insurer’s efforts to obtain rate increases,20 until mid-2008, when the Commissioner finally approved increased rates effective on June 1 for new business and August 1 for renewal business. Consequently, the dispute on remand concerned whether State Farm Lloyds, through its implemented rate, had overcharged its policyholders not only between September 7, 2003, and August 31, 2004, the period in which former art. 5.26-1 had been in effect, but whether it had continued to do so for several years thereafter. The statutory standards governing the “excessiveness” of State Farm’s “rate” after September 1, 2004, were essentially identical to those applicable under former art. 5.26-1.21
But while the hearing on remand thus concerned whether State Farm Lloyds had overcharged its customers in the past, the Commissioner emphasized that “ratemak-ing is a prospective endeavor,” alluding to the longstanding principles barring retroactive ratemaking, the making of “a retrospective inquiry to determine whether a prior rate was reasonable and imposing a surcharge when rates were too low or a refund when rates are too high.”22 He *95likewise observed that both the relevant Insurance Code provisions23 and actuarial principles24 contemplated prospective ratemaking based on estimates of future costs, and he additionally deemed it “unfair to judge the reasonableness of any of the parties’ estimates of future costs based on information that was unknowable” at the time the estimates were made.” (Emphases in original.) Thus, the proper rate-making inquiry on remand, as the Commissioner reasoned, centered not on real-life events that had occurred while the implemented rate was in effect, viewed in hindsight — as he put it, “retrospective evidence ... has no business in a rate hearing” — but on the rate or rates the Commissioner should have set prospectively from the perspective of a time preceding the period or periods in which the rate was used. To that end, the Commissioner focused the inquiry on information known or “knowable” as of early September 2003, the date of the original hearing.25 However, the Commissioner did conclude that evidence of State Farm Lloyds’s financial condition as of the time of the rehearing would be relevant to assessing the impact of any refunds he might order.
A related question concerned the time intervals for the rate or rates the Commissioner would determine — not only the period in which former art. 5.26-1 had been in effect, and for which the cost estimates underlying State Farm Lloyds’s implemented rate had been made, but almost four years thereafter, far longer than the typical one-year term of a homeowner’s insurance policy. The Commissioner ultimately focused the inquiry primarily on determining the rate that should have applied during an “initial term” between September 7, 2003 and August 31, 2004, corresponding roughly to the period in which former art. 5.26-1 had been in effect and for which the implemented rate had been designed. Then, utilizing his initial-term indicated rate as a starting point, the Commissioner considered the extent to which adjustments or modifications to the indicated rate should be made during the “subsequent period” between September 1, 2004 and July 31, 2008.
A further development impacting the hearing’s procedural posture had been the 2004 discovery of an error in State Farm Lloyds’s original premium projections under the implemented rate that had caused it to significantly overstate the figure. TDI had also relied on the same premium projections, and the error had correspondingly skewed upward the amount of rate reduction it determined necessary to conform State Farm’s Lloyds’s implemented rate with the agency’s view of the proper *96indicated rate. The Commissioner deemed this error to be “knowable” at the time of the first hearing, and TDI agreed that its calculations should be revised and abandoned its advocacy of the original 12% rate reduction. Accordingly, the Commissioner found that TDI’s original 12% reduction would “produce a confiscatory rate” and, as such, that State Farm Lloyds had met its initial burden under former art. 5.26-1 (as modified by State Farm Lloyds I) to show by clear and convincing evidence that TDI’s rate would not be “just, reasonable, [and] adequate ... for the risks to which it applies.”26 And with that burden being met, the Commissioner reasoned, the relevant “issue for determination” became simply “the development of rates chargeable by SFL [State Farm Lloyds] which are just and reasonable and neither confiscatory nor excessive for the risks to which they apply.” In the Commissioner’s view, “[n]either [TDI] Staff nor SFL bore the burden of production or persuasion on this issue.”
After correcting for the erroneous premium projections, TDI determined that State Farm Lloyds’s implemented rate had been excessive by over 9% through June 2004. OPIC, on the other hand, did not correct for the error and urged rate reductions in line with TDI’s original position. In contrast, State Farm Lloyds insisted that the Commissioner adopt its implemented rate in lieu of any reduction because its costs had actually justified a 12% rate increase over its implemented rate, had it opted to seek one.27 Although there were other areas of divergence in the parties’ respective calculations of the indicated rate, this appeal would ultimately center on two.
The first, presenting what the Commissioner would term the most “vexing and difficult” issue in the proceeding, concerned a provision State Farm Lloyds had included in its rate to aid its recovery from large financial losses it had incurred a few years earlier. Between 2000 and 2002, State Farm Lloyds had incurred almost $2 million in underwriting losses from unanticipated mold claims and a multitude of non-hurricane catastrophes (e.g., tornadoes, hailstorms, and ice storms).28 The aggregate losses had dwarfed the preexisting amount of State Farm Lloyds’s surplus, the cash reserves that insurers are required to maintain beyond the amounts reserved for incurred and expected claims, which serve as a “cushion” against losses not anticipated in the existing rate. By statute, State Farm Lloyds was required to maintain a minimum surplus equaling one-third of its total premiums — 33$ for every $1 of premium, or a 3:1 premium-to-surplus ratio — or else be deemed “statutorily insolvent” and potentially subject to supervision or conservatorship.29 Amid its losses, there is no dispute that State Farm Lloyds would have been rendered statutorily insolvent absent large cash infusions, and thus it had turned to an affiliate, State Farm Mutual Automobile Insurance Com*97pany (State Farm Mutual), and obtained three advances — in November 2001, February 2002, and September 2002 — in amounts totaling over one billion dollars. When State Farm Lloyds filed its initial rate in June 2003, its surplus was just below $467 million, giving it a premium-to-surplus ratio of roughly 2.8:1.
The advances from State Farm Mutual had formed the consideration for a $1.05 billion “consolidated surplus debenture” that State Farm Lloyds had issued to State Farm Mutual. The parties agree that the terms of this instrument distinguish it as a “surplus note,” a type of financial instrument that evidently has been used in the insurance industry for decades30 and has specifically been addressed by TDI rule.31 As the term suggests, a “surplus note” is in the nature of a debt instrument, but the interests of the creditor are subordinated in a manner permitting the proceeds to be initially treated for accounting purposes as unencumbered surplus, similar to equity capital. In the case of State Farm Lloyds’s surplus note, it reflects that the insurer owes State Farm Mutual a principal amount of $1.05 billion, plus interest accruing at 7% per annum, compounded and payable semiannually, with a maturity date of December 31, 2016. However, tracking requirements from TDI’s rules,32 the instrument conditions State Farm Lloyds’s payment obligation on the existence of surplus exceeding specified threshold amounts — principal is to be repaid only out of “excess surplus funds” that exceed $900 million, while semi-annual interest payments are to be made only from surplus, if any, that exceeds $700 million. Similarly, only if and to the extent these payment obligations actually arise is any offsetting liability recognized.33 Of final note, TDI’s rules require that the Commissioner give prior approval to any surplus notes before issuance,34 and State Farm Lloyds obtained such approval here.
State Farm Lloyds had included in its rate indication a “surplus note” provision in the amount of 9% of premium. This amount, as explained by the insurer’s witnesses during the hearing, was intended to (1) capture the annual costs of repaying the principal on the surplus note, plus interest, amortized over a 15-year term,35 plus (2) provide additional funds to increase State Farm Lloyds’s surplus (net of the funds going toward the surplus-note payments) at a rate so as to equal the amount of the company’s projected annual premiums (i.e., 1:1 premium-to-surplus ratio) within nine years. Both TDI and OPIC opposed the inclusion of the surplus-note provision in the rate, urging that State Farm Lloyds was entitled to recover only the opportunity cost of the surplus-note proceeds, not additional sums to recapitalize the company, and that the compa*98ny’s rate would already ensure it this recovery through the rate of return secured by the underwriting-profit provision. In addition to its surplus-note provision, State Farm Lloyds had included an underwriting-profit provision in its rate, equal to 5% of premiums, determining that this amount would provide it a rate of return on its capital equivalent to alternative investments of similar risk. TDI had agreed that this underwriting-profit figure was appropriate for recovering State Farm Lloyds’s costs of capital — including any recoverable costs related to the surplus note. In contrast, State Farm Lloyds insisted that the costs of amortizing the surplus-note principal and “rebuilding” the company’s surplus, as it sought to recover through the surplus-note provision, were instead in the nature of expenses of providing homeowner’s insurance in the volatile Texas market, akin to projected payments on claims, and were appropriately compensated in the rate separately and apart from its costs of capital recovered through underwriting profit and rate of return.
The second material disputed rate component was a “contingencies” provision, equal to 2% of premium, that State Farm Lloyds had included in its rate. The Commissioner heard evidence, and the parties appear generally to agree, that a “contingencies” provision, like underwriting profit, is an accepted means by which ratemak-ers account for the cost of capital in light of risk.36 Underwriting profit is said to account for the undifferentiated risk that an insurer’s actual expenses will turn out to vary from the estimates that the rate is designed to recover within any given period.37 But where deviations between an insurer’s estimated and actual expenses can be shown to persist systematically over time and be incapable of being eliminated by changing other components of the ratemaking process, it is considered appropriate to include in the rate an additional contingencies provision to compensate for the expected range of unanticipated losses.38 Thus, to summarize the difference between the two types of provisions, a contingency provision seeks to recover an expected range of losses and is thus not expected to be realized as profit, while an underwriting-profit provision yields an expected profit that accounts for variability in results.39
In attempting to establish the requisite systematic deviations between estimated and actual expenses, State Farm Lloyds relied in part on evidence of past court decisions that had effectively broadened coverage beyond what was contemplated in its rate structure. Initially, both TDI and OPIC opposed inclusion of a contingencies provision. Although TDI eventually acceded to the provision, OPIC maintained that State Farm Lloyds had failed to demonstrate any pattern of deviations that had not been addressed by intervening regulatory changes. It emphasized, for example, that the Commissioner in 2002 had approved new homeowners-policy forms that unequivocally excluded mold coverage.
The Commissioner determined that neither State Farm Lloyds’s 9% surplus-note provision nor its 2% contingencies provision should be included in the insurer’s *99rate. Regarding the former, the Commissioner agreed with TDI and OPIC that State Farm Lloyds was entitled to recover only the cost of the surplus-note proceeds that would already be compensated through a reasonable return on capital ensured through the underwriting-profit provision. Consequently, he reasoned that “[a]llowing a separate provision for a write-down or amortization of SFL’s surplus note and related interest would be tantamount to allowing two returns on capital in the ratemaking formula.” The Commissioner similarly concluded that “[i]t is unreasonable to include principal and interest payments oh the surplus note as an expense in SFL’s rates” and that doing so “in addition to rate provisions which already contemplate SFL’s expected future costs, including its cost of capital, will produce excessive rates.” As for the contingencies provision, the Commissioner found that “[t]he evidence does not support a conclusion that there is a systematic variation between expected costs and actual costs.” Accordingly, he concluded that “[i]nelusion of a contingency provision over and above a reasonable underwriting profit provision which reflects the risks of SFL writing homeowners insurance in Texas in 2008 will produce excessive rates.”
But while deciding that these specific rate provisions were unwarranted and inappropriate, the Commissioner determined that the underlying evidence was nonetheless probative of risk that had not yet been sufficiently accounted for in State Farm Lloyds’s calculation of underwriting profit. Although stopping short of finding that a “systematic variation between expected costs and actual costs” actually existed so as to warrant a contingencies provision, the Commissioner found that there “may be ” such a variation (emphasis in original), that this “conclusion ... implies risk,” and that “this risk can be considered and addressed in deriving an appropriate underwriting profit ... provision.” Similarly, with respect to the surplus-note obligation, the Commissioner found that:
While the inclusion of the separate 9.0% surplus note provision in the ratemaking formula would be unreasonable, it is reasonable to consider the existence of the surplus note, and the need for both the principal and related interest to be paid by December 31, 2016, in considering the reasonableness of the underwriting profit provision. Specifically, the Commissioner finds it is reasonable to consider the existence of the surplus note, and SFL’s obligation to timely repay it when determining: (1) the level of risk faced by SFL in writing homeowners insurance in 2008; (2) an appropriate premium to surplus ratio [as explained below, a method of weighing risk] or cost of capital; and (3) an appropriate underwriting profit provision.
The Commissioner additionally found that the analysis should also consider another “indication of risk” that had been “brought forth in the record”: “[t]he fact that SFL is a single-state insurer deriving almost its entire premium from a single line of insurance which includes catastrophe coverage in a catastrophe-prone state.” And while the Commissioner acknowledged evidence that “SFL is making significant provision for catastrophe reinsurance” (i.e., shifting the risk of such losses to the reinsurer)— in fact, the Commissioner ultimately approved including a rate component to fund such coverage — he also found that “some catastrophe exposure” faced by State Farm Lloyds “is not fully covered by reinsurance.” 40
*100In light of these indicators of additional risk he had identified, the Commissioner, to summarize his analysis, recalculated underwriting profit using the same basic methodology that State Farm Lloyds had utilized, but with greater weighting of risk in the equation. Based on this analysis, the Commissioner concluded that an appropriate underwriting profit for State Farm Lloyds during the initial period was 8.5% of premium — in other words, an underwriting profit 3.5% higher than the company itself had sought, and 1.5% higher than the combined total of the company’s underwriting-profit and contingencies provisions.
Despite this increase in the underwriting-profit component of the Commissioner’s rate indication, State Farm Lloyds’s implemented rate remained excessive by comparison. In accordance with the difference, the Commissioner concluded that a “rate reduction of -6.2% ... produces rates that are both just and reasonable and neither excessive nor confiscatory for the period September 7, 2003, through August 31, 2004.” He then turned to the “subsequent period” between September 1, 2004 and July 31, 2008.
The Commissioner rejected arguments by OPIC that he should merely extend the initial-period rate indication unchanged through the subsequent period, finding that “[tjhere is no evidence in the record to suggest that SFL’s rates remained at the same level of excessiveness for the entire period” and that, to the contrary, the “evidence ... suggests] the level of excessiveness of SFL’s rates declined over time.” To account for these changes, the Commissioner, relying on the same data set of September 2003 information from which the initial rate had been derived, projected trends of increases in several cost components of the rate (which would tend to justify a higher rate) into the subsequent period to determine a single rate for that entire period. But while determining these cost elements prospectively from the perspective of September 2003, the Commissioner also reduced the underwriting-profit component of the rate based on events that had occurred much later.
In 2006 and 2007, State Farm Lloyds’s non-hurricane catastrophe losses had turned out to be significantly lower than the projections built into its implemented rate, causing the insurer’s surplus to increase to an extent that it had been able to pay both accrued interest on the surplus note and large amounts of principal. Finding that “[djuring the subsequent period SFL [had] repaid approximately $400 million, or more than one-third, of the surplus note principal,” the Commissioner deemed it “reasonable to consider the repayment of $400 million of surplus note principal during the subsequent period when evaluating the risk associated with the repayment of the $1.05 billion surplus note.” On that basis, he determined that a “reasonable underwriting profit provision used to determine the rate in the subsequent period is 8.0%,” a reduction of one-half of one percent from the underwriting profit in the ordered rate for the initial period. Incorporating the lower underwriting profit into the flat rate the Commissioner had calculated for the subsequent period yielded an indicated rate that was 3.4% lower than State Farm Lloyds’s implemented rate throughout the entire subsequent period.
*101In light of his determinations that State Farm Lloyds’s implemented rate had been excessive during both the initial and subsequent periods, the Commissioner ordered that the insurer refund the difference to its policyholders. He estimated the gross amount of these refunds to be $256.7 million, plus accrued interest of approximately $53.0 million. The Commissioner found, however, that the insurer could provide these refunds in the form of credits to policyholders during the initial or subsequent period who renewed their coverage.
State Farm Lloyds timely sought judicial review of the Commissioner’s final order in the district court below.41 On evidence consisting solely of the administrative record, the district court affirmed the Commissioner’s order in full. State Farm Lloyds then perfected this appeal.
ANALYSIS
State Farm Lloyds brings six points of error on appeal. Collectively, the points challenge, based on an assortment of evi-dentiary and constitutional grounds, (1) the Commissioner’s refusal to include the surplus-note provision in State Farm Lloyds’s rate; (2) the Commissioner’s refusal to include the 2% contingencies provision in the rate; (3) the Commissioner’s determination of the underwriting-profit component of the rate for the initial period; (4) the rate the Commissioner imposed for the “subsequent period;” (5) the refunds ordered by the Commissioner; and (6) interest awarded by the Commissioner on amounts he held to be due.
Standard of review
The parties agree that judicial review is governed by the familiar standard set forth in section 2001.174 of the Administrative Procedure Act (APA):
[A]court may not substitute its judgment for that of the agency as to the weight of the evidence on questions committed to agency discretion but:
(1) may affirm the decision of the agency in whole or in part; and
(2) shall reverse or remand the case for further proceedings if substantial rights of the appellant have been prejudiced because the administrative findings, inferences, conclusions, or decisions are:
(A) in violation of a constitutional or statutory provision;
(B) in excess of the agency’s statutory authority;
(C) made through unlawful procedure;
(D) affected by other error of law;
(E) not reasonably supported by substantial evidence considering the reliable and probative evidence in the record as a whole; or
(F) arbitrary or capricious or characterized by abuse of discretion or clearly unwarranted exercise of discretion.42
Essentially, this is a rational-basis test to determine, as a matter of law, whether an agency’s order finds reasonable support in the record.43 “The test is not whether the agency made the correct conclusion in our *102view, but whether some reasonable basis exists in the record for the agency’s action.”44 We apply this analysis without deference to the district court’s judgment.45 We presume that the agency’s findings, inferences, conclusions, and decisions are supported by substantial evidence, and the burden is on the contestant to demonstrate otherwise.46
Substantial-evidence analysis entails two component inquiries: (1) whether the agency made findings of underlying facts that logically support the ultimate facts and legal conclusions establishing the legal authority for the agency’s decision or action and, in turn, (2) whether the findings of underlying fact are reasonably supported by evidence.47 The second inquiry, which has been termed the “crux” of substantial-evidence review,48 is highly deferential to the agency’s determination: “substantial evidence” in this sense “does not mean a large or considerable amount of evidence” — in fact, the evidence may even preponderate against the agency’s finding — but requires only “such relevant evidence as a reasonable mind might accept as adequate to support a [finding] of fact.”49 Likewise, we “may not substitute [our] judgment for the judgment of the state agency on the weight of the evidence on questions committed to agency discretion.” 50 In contrast, the first inquiry, concerning the extent to which the underlying facts found by the agency logically support its ultimate decision or action, may entail questions of law that we review de novo.51
State Farm Lloyds’s points of error implicate both deferential aspects of substantial-evidence review and issues that we must examine de novo. Foremost among the latter are State Farm Lloyds’s arguments grounded in constitutional takings principles, which substantially overlap and interrelate with their non-constitutional grounds.52 Although these principles derive from both the Texas and federal constitutions,53 the parties have not suggested *103any material substantive difference between these provisions as they impact this case, so we will assume there are none.54 Similarly, in invoking competing views of these principles and their application here, both sides have relied primarily on jurisprudence from the United States Supreme Court applying the Takings Clause of the federal constitution, with much emphasis on its seminal Hope Natural Gas decision.55
Under these authorities, the United States Supreme Court has explained, “[t]he guiding principle has been that the Constitution protects [regulated entities] from being limited to a charge for their property serving the public which is so ‘unjust’ as to be confiscatory,” a charge that is “ ‘so unjust as to destroy [the] value of the property for all the purposes for which it was acquired,’ and in so doing ‘practically deprive[s] the owner of property without due process of law.’”56 But above this constitutional floor, a rate “may reduce the value of the property which is being regulated” or “limit stringently the return recovered on investment.”57 “All that is protected against, in a constitutional sense, is that the rates fixed by the [government] be higher than a confiscatory level.”58
Starting with Hope, the Supreme Court began focusing this inquiry not on the precise methodology through which a rate is formulated, but on a rate order’s “end result”:
It is not theory but the impact of the rate order that counts. If the total effect of the rate order cannot be said to be unjust and unreasonable, judicial inquiry ... is at an end. The fact that the method employed to reach that result *104may contain infirmities is not then important.59
Thus, as far as constitutional takings principles are concerned, a governmental rate-maker “[i]s not bound to the use of any single formula or combination of formulae in determining rates.”60
Hope further explained that evaluating the justness and reasonableness of a rate’s “end result” entails “a balancing of the investor and the consumer interests.” 61 In that analysis:
[T]he investor interest has a legitimate concern with the financial integrity of the company whose rates are being regulated. From the investor or company point of view it is important that there be enough revenue not only for operating expenses but also for the capital costs of the business. These include service on the debt and dividends on the stock. By that standard, the return to the equity owner should be commensurate with returns on investments having corresponding risks. That return, moreover, should be sufficient to assure confidence in the financial integrity of the enterprise so as to maintain its credit and to attract capital.62
In short, courts consider “whether the order may reasonably be expected to maintain financial integrity, attract necessary capital, and fairly compensate investors for the risks they have assumed, and yet provide appropriate protection for the relevant public interests.”63 A rate whose end result effects a reasonable balancing of insurer and investor interests relative to ratepayer or broader public interests “cannot properly be attacked as COnfiSCatO-py ” 64
In reviewing whether the Commissioner’s order effected a reasonable balancing of these interests, we give deference under the substantial-evidence standard to his findings regarding “initial questions of historical fact,” but review de novo his “second-order analysis [of] applying] th[e] historical facts to the legal standards.”65
*105Surplus note
We will begin with the single most pivotal question in this rate appeal, and the one on which State Farm Lloyds places the greatest emphasis — the validity of the Commissioner’s decision not to include the 9% “surplus note” provision in the ordered rate. State Farm Lloyds challenges this determination principally through its first two points of error.
Interest
In its first point of error, State Farm Lloyds insists that without the surplus-note provision, the ordered rate, in both the initial and subsequent periods, is confiscatory and unconstitutional because it fails to account for the interest it owes on the surplus note. This is so, State Farm Lloyds reasons, because interest on the surplus note is properly considered an “operating expense” distinct from the costs of capital compensated through rate of return on capital and underwriting profit, and must be compensated as such in the rate.66 To support this premise, State Farm Lloyds points to TDI’s rule governing surplus notes, which requires accounting recognition of interest payments as “liabilities” when they come due67 and accounting principles, generally incorporated into TDI’s rules,68 requiring that interest “shall be expensed in the statements of operation when approved for payment.”69 State Farm Lloyds also emphasizes Texas Supreme Court decisions involving utility ratemaking that, in its view, approve or contemplate recovery of interest or other “carrying costs” on debt as specific components of the rate separate from a return on capital.70 In response, TDI and OPIC continue to *106maintain that, as the Commissioner found, State Farm Lloyds is seeking to double-recover a cost of capital for which it is already being compensated in the rate.
To address State Farm Lloyds’s assertions, we begin by noting that the Hope court characterized “service on the debt” as a component of the “the capital costs of the business,” as opposed to “operating expenses,”71 and that interest on debt is commonly accounted for through the rate of return on capital in other ratemaking contexts.72 The Commissioner heard evidence that the same is true in insurance ratemaking — interest on an insurer’s debt is considered to be a cost of capital that is properly and customarily compensated through the rate of return secured through underwriting profit, not as a type of operating expense. We cannot conclude the Commissioner’s decision to address State Farm Lloyds’s surplus-note interest the same way is without reasonable support in the evidence.
But the more critical feature of Hope, as it relates to State Farm Lloyds’s ultimate constitutional complaint, is the Supreme Court’s emphasis on the “total effect of the rate order,” not the “theory” or “method employed,” in determining a rate’s constitutionality.73 Under Hope, whether the ordered rate addressed interest on the surplus note as an “operating expense,” a “cost of capital” addressed through underwriting profit and rate of return, or through some other means, is singularly of little constitutional significance. What matters is whether the rate, in its ultimate effect, sufficiently accounted for the cost in some fashion.
To the extent State Farm Lloyds contends that the ordered rate fails altogether to account for interest -on the surplus note, it overlooks that the reasonable cost of capital contemplated by the insurer’s own underwriting-profit model, later utilized by the Commissioner, would include the cost of the surplus-note proceeds. That is, the ordered rate provided State Farm Lloyds a return on its capital, including the surplus-note proceeds, calculated so as to approximate the returns the funds would have garnered in alternative investments of similar risk. TDI’s witnesses testified that these earnings are the cost or value of State Farm Lloyds’s capital in economic effect, and for ratemaking purposes, inasmuch as they represent the rate of return that must be paid to investors to risk their funds in the enterprise.
While the Commissioner did not attempt to distinguish or specifically account for the actual amount of State Farm Lloyds’s interest obligations under the surplus note in determining the rate of return on the proceeds74 — in effect, he “costed” all of *107State Farm Lloyds’s capital as if it were equity75 — TDI presented expert testimony that this was the norm in insurance rate-making. Further support for the Commissioner’s approach is found in the evidence that the surplus note, while a debt instrument, also had attributes of equity — the debt obligations are subordinated so as to ensure that the proceeds can be used to supply minimum levels of surplus first. Yet even if there were any error in failing to differentiate surplus-note interest from its other costs of capital, State Farm Lloyds has not demonstrated any harm from the Commissioner’s approach.
To the contrary, according to projections prepared by one of State Farm Lloyds’s own experts, Dr. David Appel, the 5% underwriting profit originally advocated by the insurer and TDI would generate earnings sufficient to cover the same amortized costs of repaying the surplus note that the insurer had built into its surplus-note provision — including both interest and principal — -plus grow additional surplus at a rate of at least 4% annually through 2008. Ap-pel also prepared a similar set of projections that assumed a 7.8% underwriting profit, which generated surplus growth of at least 9% annually above the amortized costs of repaying the surplus note through the same period. The rate ultimately ordered by the Commissioner, as noted, provided State Farm Lloyds an even higher underwriting profit in both the initial and subsequent periods.
This evidence reasonably supports the Commissioner’s findings that the ordered rate fully accounts for interest on the surplus note (and, indeed, also principal) through the return on capital secured through underwriting profit. It would follow that allowing a separate provision to recover interest would, as the Commissioner found, amount to a double-recovery and render the rate excessive. In light of these findings and their evidentiary support, we cannot conclude that the Commissioner’s refusal to separately provide for recovery of surplus-note interest renders the ordered rate unreasonable or confiscatory. Accordingly, we overrule State Farm Lloyds’s first point of error.
Recapitalization
In its second point of error, State Farm Lloyds asserts that the ordered rate, in both the initial and subsequent periods, “failed to adequately and measurably provide for [its] principal payments on the surplus note.” By this, significantly, State Farm Lloyds does not mean that the ordered rate failed to provide it sufficient earnings to repay principal owed on the surplus note — again, Appel’s projections tended to show that the rate would cover the annualized amortized principal and interest on the note even with a lower underwriting profit than the Commissioner ultimately ordered. Instead, the insurer’s contention is that the rate must also enable it to build additional surplus at a rate that would increase its total surplus (net of amounts reserved for surplus-note payments) to a 1:1 premium-to-surplus ratio within nine years. Unless the rate enables it to do both, State Farm Lloyds insists, it would be forced to operate with “depleted” levels of surplus for an inordinately prolonged period of time, and run the risk in the meantime that further catastrophes would derail its repayment of the surplus note. Such developments, it adds,- would undermine confidence in its financial integrity and prevent it from attracting future investment to an unreasonable and unconstitutional degree.76
*108This was the real point of Appel’s projections, as State Farm Lloyds emphasizes. According to Appel’s calculations, a rate with a 5% underwriting profit and no surplus-note provision would require 21 years (2024) to grow State Farm Lloyds’s surplus, net of surplus-note payments, to reach a 1:1 premium-to-surplus ratio. With a 7.8% underwriting profit, according to Appel, such “recapitalization” would take 17 years (2020).
The Commissioner was unpersuaded that State Farm Lloyds was entitled to recover any costs related to the surplus note beyond the aforementioned return on its capital secured through underwriting profit. Assuming the rate provided State Farm Lloyds returns on its capital commensurate with those to be earned in alternative investments of similar risk, as the Commissioner observed in his order, there would be incentives for investors to provide capital to State Farm Lloyds going forward.77 The Commissioner was similarly skeptical that there was any immediate financial compulsion to provide State Farm Lloyds anything more. Although aeknowl-edging that having a 1:1 premium-to-surplus ratio would be “ideal,”78 the Commissioner cited Appel’s projections indicating that the ordered rate (and even a rate providing a lower underwriting profit) would enable the insurer to gradually rebuild its surplus to that 1:1 level while amortizing the surplus note, albeit at a slower pace than would be possible with the surplus-note provision. In light of this evidence, the upshot of State Farm Lloyds’s arguments, as the Commissioner saw it, was that he should “establish a rate so [the] insurer can expect to grow its surplus to a pre-determined level over a specified period of time over and above its cost of capital.” He rejected that notion, emphasizing the Supreme Court’s holdings that “[a]ll that is protected against, in a constitutional sense, is that the rates ... be higher than a confiscatory level.”79 He similarly viewed the surplus-note provision as an unjustified attempt by State Farm Lloyds not only to recoup its legitimate costs of providing insurance to its then-current policyholders, but also to effectively shift its past losses to them as well.
*109On the other hand, the Commissioner was given some pause by evidence that State Farm Lloyds’s premium-to-surplus ratio in 2003 was relatively low, noting a concession by TDI’s expert that the insurer’s premium-to-surplus ratio was only 2.35:1 around the time of the original hearing. Until State Farm Lloyds regained more solid financial footing, the Commissioner acknowledged, further unanticipated catastrophes would have had potentially “severe” repercussions for the insurer’s ability to fund repayment of the surplus note, obtain additional capital, and ultimately maintain financial viability. But the Commissioner remained unconvinced that he should approve the surplus-note provision for these reasons, a provision that, in his view, purported to provide State Farm Lloyds extraordinary returns above its cost of capital. Rather, the Commissioner viewed these considerations as going to the cost of capital itself, more specifically risk — the potential that State Farm Lloyds’s actual costs would vary from the expected costs anticipated in its rate — that is factored into determination of the reasonable return on its capital and underwriting profit. Additionally, the Commissioner determined that the parties themselves had not sufficiently accounted for these and other considerations bearing on risk when determining cost of capital and underwriting profit. To determine the appropriate rate, then, the Commissioner recalculated underwriting profit with greater weighting of risk and determined that State Farm Lloyds should receive a substantially higher underwriting profit of 8.5% of premium during the initial period and 8.0% in the subsequent period. Thus, the net effect of these rulings was to provide State Farm Lloyds an additional 3.0% to 3.5% of premium on top of the 5% it had sought as underwriting profit during the initial period, but not the full 9% of premium it had sought in the form of the surplus-note provision..
As State Farm Lloyds recognizes, our ultimate consideration in addressing its constitutional challenge to the ordered rate is whether the “end result” represents a reasonable balancing between the interests of the insurer and its investors (chiefly State Farm Mutual) on one hand, and those of its policyholders and the broader public interest on the other. With respect to State Farm Lloyds’s second issue, that inquiry focuses on whether the Commissioner’s refusal to permit State Farm Lloyds recovery of the full 9% of premium it sought through the surplus-note provision caused the ordered rate to unreasonably disfavor its financial and investor interests relative to those of ratepayers and the public. On the company and investor side of the equation, a rate that provided State Farm Lloyds returns reflecting the opportunity cost of its capital would, as the Commissioner observed, tend to attract funds from economically rational investors, at least in theory. Moreover, as TDI and OPIC emphasize on appeal, the Commissioner heard evidence that if the expense and earnings estimates anticipated in the rate held true, State Farm Lloyds would be able to cover the annualized amortized costs of the surplus note and gradually grow its surplus. And while the Commissioner acknowledged the potentially troubling consequences if actual events did not turn out that way, especially during the initial term, he took account of this risk through a substantial increase in the underwriting profit provided in the rate.
On the other side of the equation, furthermore, were concerns that a higher rate would effectively force State Farm Lloyds’s then-current policyholders to bear the costs of the company’s past losses. The Supreme Court has long held that a “company may not insist as a matter of *110constitutional right that past losses may be made up by rates to be applied in the present and future.”80 On the other hand, State Farm Lloyds insists that the surplus-note provision would actually benefit, not harm, its ratepayers, and should accordingly be weighed favorably in the constitutional balance. Because accelerated rebuilding of its surplus would tend to reduce the risk of future financial perils, State Farm Lloyds urges, its current ratepayers should appropriately bear these along with other costs of providing them insurance. More broadly, State Farm Lloyds insists that if its rate fails to include the surplus-note provision, it and other insurers will be discouraged from seeking this sort of extraordinary financing to survive future perilous times, opting instead to simply withdraw from the market,81 and that future investment to aid financially strapped insurers in such circumstances will likewise be chilled. These outcomes, State Farm Lloyds argues, would in turn not only harm Texas consumers (especially if a large homeowners insurer like itself were to leave the market), but thwart the regulated private-market approach to homeowners insurance that the Legislature has heretofore generally preferred over more government-intensive policy alternatives.
We echo the Commissioner’s assessment that these are “vexing” and “difficult” questions of insurance regulation, economic theory, and public policy. But resolution of such issues within our governmental framework is vested first in the Commissioner or the Legislature, and it is only at the margins of the other Branches’ lawful powers that the Judiciary properly has a say. We are unpersuaded that the Commissioner, in refusing to afford State Farm Lloyds recovery of the full 9% of premium it sought through the surplus-note provision, has effected such an unreasonable balancing among the interests of State Farm Lloyds, its investors, ratepayers, and the public as to violate the constitutional norms.82 Consequently, we cannot conclude this decision renders the rate confiscatory. Accordingly, we overrule State Farm Lloyd’s second point of error.
Contingencies
In its third point of error, State Farm Lloyds challenges the Commissioner’s refusal to include the 2% contingencies provision in the rates for the initial and subsequent periods. Alternatively, it urges that to the extent the Commissioner effectively included a contingencies provision in the increased underwriting profit he awarded, he deprived it of a reasonable *111rate of return on its capital because contingencies are in the nature of anticipated expenses rather than anticipated profit. The predicate for both arguments is that the Commissioner’s failure to find the required systematic variation between State Farm Lloyds’s projected and actual expenses was arbitrary and not supported by substantial evidence. This is so, according to the insurer, because it presented “un-contradicted” and “unimpeached” evidence establishing the required systematic shortfalls.
Even accepting State Farm Lloyds’s view of the record, the Commissioner was not obligated to accept or credit the insurer’s evidence regarding the contingencies provision as long as he provided an explanation or findings that would establish a reasonable basis for his doing so.83 And, although State Farm Lloyds asserts otherwise, the Commissioner made underlying findings demonstrating a reasonable basis for his rejection of the carrier’s evidence of systemic variations. Among other facts, the Commissioner pointed to a significant intervening change in the ratemaking process — in 2002, he had approved the use of policy forms that unequivocally excluded most mold losses from basic coverage, thereby addressing a key driver of the adverse court decisions that State Farm Lloyds had cited as proof of a systemic pattern of unforeseen losses. In the face of such facts, State Farm Lloyds’s argument goes ultimately to the weight the Commissioner ascribed to its evidence, and we cannot conclude he acted unreasonably or exceeded his discretion in that regard.84 We overrule State Farm Lloyds’s third point of error.
Underwriting profit
We next consider State Farm Lloyds’s challenge to the Commissioner’s determination of underwriting profit in the initial period, a component of its fifth point of error. In addition to being an independent ground for reversal, our preceding analysis of State Farm Lloyds’s arguments regarding the surplus note rested, in part, on the assumption that the underwriting-profit provision in the ordered rate had sufficiently secured it a rate of return consistent with its cost of capital.
State Farm Lloyds argues that the Commissioner’s determination of underwriting profit is not supported by substantial evidence because “he used a method that no witness proposed and no record evidence supports as an acceptable method for determining the profit provision.” It thus invokes the principle that substantial-evidence review, however deferential its factual aspects may be, still requires that an agency decision find ultimate support in evidence or facts officially noticed, and mere “agency expertise cannot be a substitute for proof.”85 It follows from this *112principle, as State Farm Lloyds emphasizes, that an agency’s methodology for calculating a rate or a rate component must find ultimate support in the evidence, facts judicially noticed, or an agency rule.86 But this does not necessarily mean that there must be explicit testimony or evidence of a particular methodology or calculation — in a given case, for example, the evidence might well enable the agency to reasonably infer an appropriate method that combines elements presented by different witnesses but no single witness.87 Similarly, even where no single witness testifies that the calculation should yield a specific figure, competing testimony may present a range of reasonable options from which the agency has discretion to choose.88
The primary focus of State Farm Lloyds’s argument is that the Commissioner took account of the surplus-note obligation and the possibility of contingencies as factors probative of risk that should bear upon underwriting profit rather than viewing them as justifications for the specific rate provisions it had advocated. In State Farm Lloyds’s view, the evidence afforded the Commissioner only one alternative for addressing these issues — adopt the solution advocated by its experts and include the 9% surplus-note provision and 2% contingencies provision in the rate. State Farm Lloyds overlooks that TDI and OPIC’s experts advocated an alternative method for addressing the surplus-note obligation and the possibility of contingencies, albeit not the one the insurer preferred — regard any such costs as costs of capital and relegate State Farm Lloyds to recovering these costs through the underwriting-profit provision of the rate. As previously indicated, there was ample evidence to support the Commissioner’s rejection of State Farm Lloyds’s proposed *113alternative. In this respect, State Farm Lloyds’s complaint about underwriting profit is merely a corollary to its other arguments challenging the Commissioner’s refusal to include the surplus-note and contingencies provisions in the rate.
State Farm Lloyds is correct, however, that no witness explicitly advocated taking account of the surplus-note obligation and the possibility of contingency as additional elements of risk to be factored into underwriting profit above the risk that would already be reflected in the 5% underwriting profit State Farm Lloyds had advocated. But to the extent State Farm Lloyds is arguing that the Commissioner should have refrained from making his upward adjustments to underwriting profit after having rejected the two rate provisions, any error would only have been to the insurer’s benefit.89 Regardless, the evidence reasonably supported the Commissioner’s decision to make an upward adjustment to underwriting profit. A key underpinning of TDI and OPIC’s opposition to the surplus-note provision (and, ultimately, of the Commissioner’s rationale in rejecting it) was that the rate’s underwriting-profit provision already compensated State Farm Lloyds for its cost of capital. This reasoning logically presumes a reasonably accurate calculation of State Farm Lloyds’s cost of capital and underwriting profit. A key determinant of the cost of capital (and, in turn, underwriting profit), everyone agrees, is the level of risk the insurer faces (i.e., the potential for future random deviations between the estimated costs accounted for in the insurer’s rate and what its actual costs turn out to be). Hence, if the Commissioner found the existence of additional risk that had not been factored into the determination of underwriting profit and was material to it, it would follow that an adjustment to that provision would be necessary to ensure that the insurer was reasonably compensated for its cost of capital. The Commissioner found such an adjustment to be warranted here, and this finding is reasonably supported by at least two evidentiary bases. First, and most obviously, State Farm Lloyds itself had viewed the surplus-note obligation and contingencies issue as justifying separate rate provisions and presumably would not have factored them into the risk bearing on its underwriting-profit calculation. Second, the Commissioner heard evidence to the effect that the 5% underwriting-profit figure advocated by State Farm Lloyds was to some extent a standard or industry figure commonly used when fashioning property-insurance rates. This case was different, the Commissioner reasoned, finding that “[t]he element of risk associated with repaying a surplus note, plus interest, would not be present in a typical insurer, which has no surplus notes, and would not be prominent in an aggregation of all insurers, as most insurers do not have surplus notes.” State Farm Lloyds has not challenged that finding. In sum, while not advocated specifically by any one witness, the Commissioner’s decision to consider the surplus-note obligation and possibility of contingencies as indicia of additional risk and accordingly revisit underwriting profit is nonetheless grounded in reasonable inferences from the evidence the parties did present.90
State Farm Lloyds further contends that the specific methodology through which the Commissioner adjusted underwriting profit lacks support by substantial evidence. To the extent State Farm *114Lloyds is urging that the Commissioner should have adjusted underwriting profit to some amount higher than it did, we will' proceed to address its contention.91 To determine the necessary adjustments to underwriting profit, the Commissioner utilized the same basic formula that State Farm Lloyds had itself employed to calculate underwriting profit, a method that the evidence further indicates is widely accepted among actuaries.92 Under this method, simply described, State Farm Lloyds solved for the amount of underwriting profit the carrier would have to obtain to yield, when combined with estimated investment income and other income, and after taxes, what would be a reasonable average rate of return on a hypothetical risk-adjusted level of capital.93 According to State Farm Lloyds’s witnesses, the hypothetical level of capital — customarily expressed in terms of a premium-to-surplus ratio, with a lower ratio signifying higher risk and vice versa — represents a level at which the risk associated with the amount of capital allocated to each exposure unit is of a targeted average intensity. In determining its underwriting profit, State Farm Lloyds had chosen a hypothetical 1:1 premium-to-surplus ratio. With that input, State Farm Lloyds had determined that 5% underwriting profit, when combined with its estimated investment and other income, and after taxes, would yield an average rate of return on surplus of 12.4%, which it had adjusted to an estimated return on equity calculated under GAAP of just under 10%. Although OPIC advocated a lower underwriting profit of 2.5%, TDI staff ultimately concurred with State Farm Lloyds that a 5% underwriting profit was appropriate and that the rate of return it had achieved was reasonable.
The Commissioner found that “[i]t is reasonable to use SFL’s profit model in deriving an underwriting profit provision for use in this case,” with “further consideration of the premium to surplus ratio used in the profit model” in light of various risk factors he had identified. The Commissioner then made the following specific findings:
136. SFL’s theoretical [1:1] premium to surplus ratio was described by SFL as being “minimally prudent” [for purposes of determining underwriting profit].
137. One SFL witness testified that analyses of premium to surplus ratios for professional reinsurers and Bermuda property reinsurers, combined with “the fact that SFL is not charging anything in the rate for catastrophe reinsurance” could justify a premium to surplus ratio as low as 0.5 to 1.0. Elsewhere in his testimony, the same witness acknowledges that SFL has made some provision for catastrophe reinsurance, but that “since many exposures are not covered by the catastrophic reinsurance coverage and since in Texas, many catastrophes are too small to exceed the reinsurance deductible (or ‘attachment point’), the coverage is incomplete.”
138. Other evidence in the record demonstrates that SFL is making significant provision for catastrophe reinsurance.
139. The fact that SFL is making significant provision for catastrophe reinsurance greatly diminishes *115SFL’s assertion that a premium to surplus ratio as low as 0.5:1.0 could be justified; but the fact that some catastrophe exposure is not fully covered by reinsurance offers some support for using a ■ premium to surplus ratio of modestly less than 1.0:1.0.
140. The justification for using a premium to surplus ratio of less than 1.0:1.0 is strengthened significantly when other elements of risk are considered.
141. The fact that SFL is a single-state insurer deriving almost its entire premium from a single line of insurance which includes catastrophe coverage in a catastrophe-prone state was brought forth in the record as an indication of risk.
142. As noted earlier ... the existence of a surplus note, and the obligation to repay such note by a date certain, along with related interest, is an element of risk inherent in SFL’s writing of homeowners insurance in 2003.
143. The element of risk associated with repaying a surplus note, plus interest, would not be present in a typical insurer, which has no surplus notes, and would not be prominent in an aggregation of all insurers, as most insurers do not have surplus notes.
144. SFL’s surplus note, and the obligation to repay it, along with related interest, suggests that an otherwise reasonable premium to surplus ratio, or what might be construed as a “minimally prudent” premium to surplus ratio, may not be reasonable to determine an appropriate underwriting profit provision with respect to SFL’s writing of homeowners insurance in 2003.
145. SFL’s surplus note, and the obligation to repay it, along with related interest, by December 31, 2016, suggests that the use of a lower premium to surplus ratio to determine the underwriting profit provision may be, in consideration of SFL’s writing of homeowner’s insurance in 2003, more reasonable than the 1.0:1.0 ratio used in SFL’s analysis.
146. As noted earlier [in regard to the contingencies provision], the conclusion that there may be a systematic variation between expected costs and actual costs implies risk.
147. To the extent such risk is not contemplated in the l.Orl.O premium to surplus ratio used in the SFL profit model, there is a basis for using a premium to surplus ratio lower than 1.0:1.0.
With that prologue, the Commissioner found that:
148. ... (1) SFL is a single-state writer, writing almost exclusively homeowners insurance, which includes catastrophic coverage, in a catastrophe-prone state; (2) SFL faces an element of risk associated with the obligation to repay the $1.05 billion surplus note, along with related interest by a date certain; and (3) there may be systemic variation between expected costs and actual costs which has not been addressed in the rate-making process. For these reasons, it is reasonable to accord to SFL’s writing of homeowners insurance in 2003 [emphasis in the original] a greater measure of risk than is contemplated in the 1.0:1.0 *116premium to surplus ratio contained in SFL’s profit model.
As for what that specific ratio should be, the Commissioner further found that a “reasonable premium to surplus ratio for use in this proceeding would be somewhere below 1.0:1.0, but well above 0.5:1.0.” This range of hypothetical premiums-to-surplus ratios, the Commissioner observed, translated into “profit provisions ... in the range of 7.0% to 10.0%.” The Commissioner then selected 8.5% as “a reasonable profit ... provision, including a provision for hurricane risk, for SFL in conjunction with its writing of homeowners insurance in Texas in 2003.”
Although State Farm Lloyds is correct that no witness explicitly testified in support of the upward adjustment to underwriting profit the Commissioner ultimately made, this was hardly a case where the Commissioner “embarked on [his] own method that goes beyond anything in the record.”94 To the contrary, the Commissioner utilized a formula that had been supplied by the insurer’s own witnesses and adjusted but a single input — the appropriate hypothetical amount of risk-adjusted capital, expressed as a premium-to-surplus ratio — which all agree is the means by which risk is accounted for in that formula. While State Farm Lloyds had assumed a 1:1 premium-to-surplus ratio, the Commissioner heard evidence that a ratio as low as 0.5:1 could be appropriate depending on the extent to which the insurer was able to shift its risks of catastrophe losses through reinsurance. The Commissioner chose a ratio (and a corresponding underwriting profit) in the middle of this range, reasoning that while a 0.5:1 ratio was inappropriate in light of State Farm Lloyds’s provision for catastrophe reinsurance, there remained risks of losses that had not been fully addressed by that reinsurance, along with the additional risks presented by the surplus note and possible contingency.95 While State Farm Lloyds would insist on greater exactitude in correlating specific adjustments to the premium-to-surplus ratio to particular risk factors, we cannot conclude that the Commissioner “selected a figure that is outside the range of the applicable evidence, or that is otherwise unsupported by any evidence.”96 Accordingly, we overrule State Farm Lloyds’s challenge to the 8.5% underwriting profit the Commissioner awarded in the ordered rate for the initial period.
Subsequent period
The foregoing holdings resolve each of State Farm Lloyds’s challenges to the ordered rate for the initial period. We now turn to State Farm Lloyds’s arguments regarding the rate the Commissioner determined for the “subsequent period” between September 1, 2004 and July 31, 2008, which are further components of its fifth point of error.
As a threshold matter, State Farm Lloyds questions (although does not extensively brief or argue) whether the Commissioner had statutory authority to reduce its implemented rate with respect to *117any period after August 31, 2004, the last day former art. 5.26-1 remained in effect. While State Farm Lloyds is correct that the legality of its implemented rate, as applicable to periods after August 31, 2004, is necessarily governed by statutes other than former art. 5.26-1, it overlooks that each of the successor statutory regimes also empowered the Commissioner to order refunds based on a final determination that State Farm Lloyds’s implemented rate had been excessive. Under former art. 5.142, as previously noted, the Legislature provided for refunds with interest in a manner identical to former art. 5.142 in the event the Commissioner’s rate determination was upheld on appeal.97 Similarly, under the file-and-use regime, former art. 5.144 of the Insurance Code (now codified as chapter 2254 of that Code)98 has authorized the Commissioner to order refunds of any filed rate he determines to be excessive.99 State Farm Lloyds suggests that the latter provision is inapplicable here because it contemplates a right to a rate hearing before the State Office of Administrative Hearings, as opposed to the Commissioner alone,100 the procedure followed here. This assertion is in the nature of an unpreserved complaint of procedural error rather than a basis for holding former art. 5.144 inapplicable in the first instance. In short, the Commissioner has had statutory authority to determine whether and the extent to which State Farm Lloyds had charged its policyholders excessive rates, and to order refunds accordingly, not only with respect to the initial period in which former art. 5.26-1 was in effect, but also in the subsequent period that followed.
Assuming the Commissioner’s statutory authority to determine the appropriate rate for the subsequent period, State Farm Lloyds’s core complaint is that the Commissioner, in setting a rate that was flat or uniform throughout the entire subsequent period, failed to take account of significant intervening increases in its costs of providing insurance. State Farm Lloyds points to evidence that, beginning in 2006, its reinsurance costs had significantly increased as world markets responded to Hurricanes Katrina and Rita in late 2005 and eight other hurricanes that had hit Florida in 2004 and 2005. In fact, one of TDI’s experts, Dr. Mark Crawshaw, maintained during the hearing that State Farm Lloyds’s initial rate had ceased to be excessive from July 2006 onward and thereafter became inadequate. Based in part on these cost increases, State Farm Lloyds would ultimately obtain Commissioner approval for the rate increase that took effect in mid-2008.
State Farm Lloyds attacks the ordered rate through two primary legal theories. First, similar to its arguments regarding underwriting profit in the initial period, it asserts that the Commissioner’s formulation of a flat rate for the entire subsequent period was founded on a methodology that no witness had proposed and no evidence *118supported. Second, State Farm Lloyds insists that the Commissioner acted arbitrarily and capriciously in imposing a flat rate that excluded consideration of its significant cost increases during the subsequent term. Relatedly, it contrasts the Commissioner’s failure or refusal to account for these cost increases with his reduction of underwriting profit for the entire subsequent period — i.e., beginning September 2004 — based on surplus-loan repayments that did not occur until 2006.101 Such “arbitrary hindsight,” State Farm Lloyds urges, represented a selective departure from the prospective rate-making on which the Commissioner had otherwise insisted, forcing it to absorb rate reductions based on actual post-2003 events while simultaneously depriving it of the benefit of actual post-2003 events that would have increased its rates.
We are compelled to agree that the Commissioner’s reduction of underwriting profit based on post-2003 events, in the context of imposing a single flat rate for the entire subsequent period, was arbitrary and capricious. This becomes apparent when considering the methodology the Commissioner used to set the rate. Having found that the evidence “suggests the level of excessiveness in SFL’s rates declined over time,” the Commissioner considered two alternative methodologies for accounting for the changes that had been proposed by State Farm Lloyds’s actuary, Rob Kelley. Both methodologies were prospective in nature, although their points of reference differed. In the first, which the Commissioner would term the “Trend Projection Method,” Kelley had utilized the data known and knowable as of September 2003 (i.e., the same information from which the initial-period rate had been derived) and extended projections originally used to develop certain cost components of the initial-period rate. Applying these projections to State Farm Lloyds’s rate indication for the initial period, Kelley devised rates for each of three successive annual periods beginning on September 1, 2004, 2005, and 2006, and a final period between September 1, 2007, and the end of the subsequent period, July 31, 2008. In Kelley’s second method, termed the “Augmented Data Method,” he determined these annual rates based on data that would have been available at each interval.
The Commissioner found that the Trend Projection Method — the one relying solely on September 2003 information — “is most suitable for determining a rate for SFL for the subsequent period.” Taking the ordered rate for the initial period as his starting point, the Commissioner, similar to Kelley, “trended” increases in State Farm Lloyds’s estimated earned premiums and four cost components in the rate — (1) non-catastrophe loss and loss adjustment expense (LAE); (2) hurricane catastrophe loss and LAE; (3) non-hurricane catastrophe loss and LAE; and (4) fixed expenses — and made specific findings as to the annual percentage increases that should be applied to each.102 However, rather than determining a rate for each of the four annual segments within the subsequent period, as had Kelley, the Commissioner determined a single rate for the entire subsequent period by trending the aforementioned rate components up to February 15, 2007.
*119The Commissioner offered two justifications for this departure from Kelley’s approach. First, he explained that February 15, 2007, represented the “average date of loss” under the State Farm Lloyds policies that would have been sold during the subsequent period, represented by the midpoint between the first day such a policy could have been effective (September 1, 2004) and the date the last policy could have expired (July 31, 2009).103 According to the Commissioner, it was “common actuarial practice” to trend cost projections underlying a rate to the average date of loss under the policy, and he pointed out all parties had accordingly trended their projections underlying the initial period rate until September 1, 2004, the midpoint between the first day the initial-period policies could have been effective and the last day they could have expired. On the other hand, the Commissioner also acknowledged that “formulating a homeowner’s rate for such a long duration,” as he was doing in the subsequent period, “is simply unheard of.”
The Commissioner’s second justification was that “there is no substantive difference,” in terms of the amounts State Farm Lloyds would ultimately owe as refunds, between his trending of the rate components for the entire subsequent period as a whole and Kelley’s delineation of annual segments. “This is especially true,” the Commissioner added, in light of evidence that State Farm Lloyds “has ‘extremely high’ renewal rates.” In essence, the Commissioner reasoned that while the ex-cessiveness of State Farm Lloyds’s initial rate would differ at any given time within the subsequent period, the differences would average out at 3.4% over the. entire period, at least when viewed from the perspective of September 2003 looking forward. While urging that no evidence supports the Commissioner’s departure from Kelley’s year-by-year approach, State Farm Lloyds does not appear to dispute the Commissioner’s math.
In sum, the Commissioner set a flat rate for the entire subsequent period founded on cost projections made from the perspective of September 2003, approximately a year before the period to which the rate applied began. This model foreclosed consideration of actual cost increases that might occur during the subsequent period. From that standpoint, State Farm Lloyds’s evidence of significant increases in reinsurance costs beginning in 2006 would have had no bearing on the appropriate rate for the subsequent period. For that matter, these subsequent developments would have had no bearing upon rates determined in annual segments under Kelley’s original Trend Projection Method, either, because that method likewise relied on September 2003 data.
But having deprived State Farm Lloyds of the benefit of any intervening developments that could favorably impact its indicated rate, the Commissioner also forced it to bear the unfavorable consequences to underwriting profit that he derived from the 2006 surplus-note repayments. TDI and OPIC insist that both the Commissioner’s method in setting a single flat rate for the entire subsequent period and his downward adjustment of underwriting profit were supported by substantial evidence, and that it likewise supported the “amalgam” of rating methodologies the Commissioner devised.104 *120Even if so, the Commissioner’s actions are independently subject to reversal if arbitrary and capricious.105 An agency’s decision may be held arbitrary and capricious if, among other defects, “it is based on legally irrelevant factors or if legally relevant factors were not considered or if the agency reached an unreasonable result.”106 In the context of the single rate he established for the entire subsequent period, the Commissioner’s reduction of underwriting profit was based on a factor that he had acknowledged to be legally irrelevant and inappropriate — the use of 2006 and 2007 data to determine a rate that applied prior to that time. In effect, the Commissioner engaged in prohibited retroactive ratemaking with respect to those earlier periods. He also did so selectively and unreasonably, in essence cherry-picking post-September 2003 data where it served to reduce State Farm Lloyds’s indicated rate while excluding consideration of post-September 2003 data that could have increased it, with no evident justification for the differential treatment. Even while recognizing the constitutional leeway afforded to an agency’s choice of ratemaking methods, the Supreme Court has cautioned nonetheless that a “decision to arbitrarily switch back and forth between methodologies in a way which requires investors to bear the risk of bad investments at some times while denying them the benefit of good investments at others would raise serious constitutional questions.” 107
The Commissioner’s determination of the rate for the subsequent period was arbitrary and capricious, and must be reversed.108 To that extent, we sustain State Farm Lloyds’s fifth point of error.
Interest
Having determined that State Farm Lloyds’s initial rate was excessive during both the initial and subsequent periods, the Commissioner ordered refunds of the “overcharged premiums,” plus interest, citing the refund provisions of both former art. 5.26-1 and chapter 2254 (the former art. 5.144) without distinguishing between overcharges for the initial versus subsequent periods or, with respect to the latter, those occurring while former art. 5.142 was in effect (i.e., before December 1, 2004) as opposed to thereafter. The Commissioner similarly conflated the statutory provisions in awarding interest on the refunds he held to be due, determining that interest on all refunds should be awarded in accordance with former art. 5.26-1 for each year during both the initial and subsequent periods, but only for the interest that accrued up • to the date of his order. Interest accruing from the date of the Commissioner’s order forward, he reasoned, would accrue under chapter 2254, regardless of when the overcharge occurred. The statutory distinction is not merely formal or technical. Under former art. 5.26-1, the Legislature had specified that “[t]he interest rate is the prime rate plus one percent as published in the Wall Street Journal on the first day of each calendar year that is not a Saturday, Sun*121day, or legal holiday.”109 An identical interest provision appeared in former art. 5.142.110 In contrast, beginning in 2005, chapter 2254 (or its predecessor, former art. 5.144) have specified that an “interest penalty” accrues on refund amounts at a rate that is the lesser of 18 percent or the sum of six percent and the prime rate for the calendar year in which the Commissioner’s order finding the rate excessive is issued.111
Apart from State Farm Lloyds’s complaints regarding the propriety of the refunds themselves and the Commissioner’s underlying rate determinations, none of the parties have challenged the Commissioner’s award of interest on all refunds at the rate of one percent over the prime rate for each year at issue, consistent with the rate prescribed by former art. 5.26-1, up to the date of his order (a/k/a “pre-order” interest). However, in its sixth issue, State Farm Lloyds urges that the Commissioner reversibly erred in awarding post-order interest accruing on all refunds at the higher rate specified in chapter 2254. In State Farm Lloyds’s view, interest on any refunds, both before and after the date of the Commissioner’s order, is governed exclusively by former art. 5.26-1 regardless of when the overcharge occurred. With respect to the overcharges the Commissioner found to have occurred during the initial period, we agree — as previously explained, former art. 5.26-1, which continues in effect for these purposes, governs the validity of State Farm Lloyds’s implemented rate during the initial period, its policyholders’ right to refunds of “overcharged premium,” and interest on those amounts. To this extent, we sustain State Farm Lloyds’s sixth issue.
As for interest on any refunds for overcharges during the subsequent period, our definitive resolution of that issue must await the proceedings on remand to determine the appropriate underlying rate or rates during that time and whether or for what times any refunds are owed." We note, however, that we have held above that the Commissioner’s authority to order refunds during the subsequent period is governed by former art. 5.142 with respect to any overcharges between September 1, 2004 and November 30, 2004, and by former art. 5.144 and chapter 2254 for periods *122thereafter. On the other hand, appellees have not preserved any complaint in this appeal regarding the propriety of the Commissioner’s awarding of pre-order interest on all refunds held to be due in accordance with former art. 5.26-1. We likewise observe that the Legislature did not authorize interest awards under the former art. 5.144, at the higher prime-plus- ■ six-percent rate or otherwise, until 2005.112 On remand, if the Commissioner again finds that refunds are owing for the subsequent period or a portion of it, the parties will have the opportunity to address whether or how these observations bear upon the appropriate interest rate or rates.
Refunds
In its sole remaining point of error, its fourth, State Farm Lloyds argues that the refunds with interest ordered by the Commissioner were confiscatory in light of their impact on the insurer’s financial condition at the time of the order. Because we have reversed and remanded the Commissioner’s determination of an appropriate rate for the subsequent period, any constitutional challenge based on the financial impact of any refunds and interest ultimately awarded by the Commissioner must await those further proceedings.
CONCLUSION
We affirm the portion of the district court’s judgment affirming the Commissioner’s order determining a rate for the “initial period” and that State Farm Lloyds’s implemented rate was excessive by 6.2% during that time frame. However, we reverse the portion of the judgment affirming the Commissioner’s order setting a rate for the subsequent period and holding that State Farm Lloyds’s implemented rate was 3.4% excessive then. We remand the question of the appropriate rate and reductions for the “subsequent period,” as well as the ultimate determination of refunds and interest awards, to the Commissioner for further proceedings consistent with this opinion. We further hold that interest awarded on refunds for the initial period will be governed exclusively by the rate specified in former art. 5.26-1, both before and after the date of the Commissioner’s order.
Justice Henson not participating
. 255 S.W.3d 786 (Tex.App.-Austin 2008, no pet.).
. Act of June 2, 2003, 78th Leg., R.S., ch. 206, § 4.01, 2003 Tex. Gen. Laws 907, 921 ("former art. 5.26-1”).
.See former art. 5.26-1, § 4; State Farm Lloyds I, 255 S.W.3d at 800-01.
. See Act of June 2, 2003, 78th Leg., R.S., ch. 206, §§ 1.01, 2(7), 2003 Tex. Gen. Laws 907-8 ("former art. 5.142”) (defining "rate”); former art. 5.26-1, § 1(b) (incorporating definitions from former art. 5.142).
. Former art. 5.26-1, § 2(b) ("Initial Rate Filing”); see also former art. 5.142, § 3 (requirements applicable to initial filed rates calculated by the insurer). Specific considerations in setting this "rate” were to include "past and present loss experience”; "the insurer's historical premium, exposure, loss, and expense experience”; "catastrophe hazards within this state"; "operating income”; "investment income”; and "a reasonable margin for profit.” Id. § 3(b).
. See former art. 5.142, § 2(b) (defining "excessive,” "inadequate,” and "unfairly discriminatory”); State Farm Lloyds I, 255 S.W.3d at 801 (construing "just, reasonable, adequate, not excessive, and not unfairly discriminatory for the risks to which it applies” to mean that "the rate must allow for a 'reasonable profit’ but not one that' is ‘unreasonably high in relationship to the insurance coverage provided’") (quoting former art. 5.142, §§ 2(b)( 1 — 3), 3(d); former art. 5.26-1, § 2(b)); see also American Alliance Ins. Co. v. Board. of Ins. Comm'rs, 126 S.W.2d 741, 744 (Tex.Civ.App.-Austin 1939, writ ref’d) (reasoning that “unreasonable, unjust, excessive, or inadequate” in context of insurance rate-making implies “a duty both to the insuring public and the insurance carriers" to promulgate rates “which shall be as low to the insured as is consistent with a reasonable return to the insurer”).
.State Farm Lloyds I, 255 S.W.3d at 795 ("A government-set rate must allow a regulated company to not only recover its operating expenses, but also to realize reasonable returns on its investments sufficient to assure confidence in the continued financial integrity of the enterprise. A rate that does not allow for a reasonable rate of return is confiscatory and unconstitutional.” (citing Duquesne Light Co. v. Barasch, 488 U.S. 299, 307, 109 S.Ct. 609, 102 L.Ed.2d 646 (1989); Jersey Cent. Power & Light Co. v. Federal Energy Regulatory Comm’n, 810 F.2d 1168, 1181 (D.C.Cir.1987); Railroad Comm'n v. Houston Natural Gas Corp., 155 Tex. 502, 289 S.W.2d 559, 572 (1956)); see also American Alliance, 126 S.W.2d at 742 (holding that fire insurance rates "must be reasonable and nonconfiscato-ry ... [tjhat is, the rate prescribed, whether maximum or fixed, must be sufficient to yield a reasonable net capital return, áfter deducting all necessary and proper expenses.... Otherwise, the rate is confiscatory and viola-tive of the stated constitutional inhibitions.”).
. See Actuarial Standards Board, Actuarial Standard of Practice: Documentation and Disclosure in Property and Casualty Insurance Ratemaking and Loss Reserving, at 9 (1989) (superseded 1991) (incorporating "Statement of Principles Regarding Property and Casualty Insurance Ratemaking” adopted by the Board of Directors of the Casualty Actuarial Society in 1988) (currently available at http:// www.actuarialstandardsboard.org/pdf/ superseded/asop9_l 989.PDF) (hereinafter
“ASOP 9”); Actuarial Standards Board, Actuarial Standard of Practice No. 30: Treatment of Profit and Contingency Provisions and the Cost of Capital in Property/Casualty Insurance Ratemdking, at §§ 1.1, 2, 3.1 (1997) (currently available at http://www.actuarialstandards board.org/pdf/asops/asop030_057.pdf) (hereinafter "ASOP 30”).
. See State Farm Lloyds I, 255 S.W.3d at 792.
. See id. (citing former art. 5.26-1, § 2(a)).
. See id. (citing former art. 5.142).
. See id. (citing former art. 5.142, § 5).
. See Act of June 2, 2003, 78th Leg., R.S., ch. 206, § 6.04, sec. 3(a)(5), 2003 Tex. Gen. Laws 907, 926 (currently codified at Tex. Ins. Code §§ 2251.00-252) ("former Art. 5.13-2”).
. See State Farm Lloyds I, 255 S.W.3d at 792 (citing former art. 5.13-2, §§ 5, 7).
. See former art. 5.26-1, § 5(b).
. Id. § 6.
. See State Farm Lloyds I, 255 S.W.3d at 805 (quoting former art. 5.26-1, § 7).
. See former art. 5.142, §§ 13(b), 14; State Farm Lloyds v. Geeslin, 267 S.W.3d 438, 443-46 (Tex.App.-Austin 2008, no pet.) (State Farm Lloyds III).
. See Texas Dep't of Ins. v. State Farm Lloyds, 260 S.W.3d 233, 238-49 (Tex.App.Austin 2008, no pet.) (State Farm Lloyds II); State Farm Lloyds III, 267 S.W.3d at 441-47.
. See former art. 5.142, §§ 2, 3(d); former art. 5.13-2, §§ 1, 3.
. State v. Public Util. Comm'n, 883 S.W.2d 190, 199 (Tex.1994). These principles derive in part from the view that ratemaking is in the nature of a legislative act, having only *95prospective effect, as opposed to being an adjudication of rights in a controversy that has previously arisen. See Railroad Comm'n v. Houston Natural Gas Corp., 155 Tex. 502, 289 S.W.2d 559, 562-63 (1956); see also Central Power & Light Co. v. Public Util. Comm’n, 36 S.W.3d 547, 554 (Tex.App.-Austin 2000, pet. denied) (citing City of Alvin v. Public Util. Comm’n, 876 S.W.2d 346, 362 (Tex.App.-Austin 1993), judgment vacated w.r.m. sub nom. Public Util. Comm’n v. Texas-N.M. Elec. Co., 893 S.W.2d 450 (Tex.1994)).
.See former art. 5.142, § 2(b) (defining an “excessive” rate as one "likely to produce a long-term profit that is unreasonably high” and an "inadequate” rate as one "insufficient to sustain projected losses and expenses”); accord former art. 5.13-2, § 3(b)(a)-(b) (same).
. See ASOP 9, at 6 ("A rate is an estimate of the expected value of future costs.”).
. See, e.g., Southwestern Bell Tel. Co. v. Public Util. Comm’n, 615 S.W.2d 947, 954-55 (Tex.Civ.App.-Austin 1981, no writ) (distinguishing retrospective ratemaking from administrative proceedings on remand to set rates prospectively from date of the original agency order that had been reversed on appeal).
. Former art. 5.26-1, § 4; see State Farm Lloyds I, 255 S.W.3d at 800-01.
. To be clear, State Farm Lloyds was not seeking an increase from its implemented rate (nor could it do so retroactively), but only to oppose any reduction in its implemented rate. See State Farm Lloyds I, 255 S.W.3d at 792 (noting that the insurer had filed its then-current rate as its "initial rate" under former art. 5.26-1).
. "Catastrophe” in this context refers, according to the Commissioner, to "large and fortuitous occurrences that cause widespread property damage,” such as a hurricane or "large hailstorm in a major metropolitan area.”
. See Tex. Ins. Code §§ 441.051(1) (circumstances constituting insolvency), 822.205 (requiring surplus); see generally id. §§ 441.001-.351 (“Supervision and Conservatorship”).
. See, e.g., Harlan v. United States, 409 F.2d 904, 906-07 (5th Cir.1969).
. See 28 Tex. Admin. Code § 7.7 (2014) (TDI, Subordinated Indebtedness, Surplus Debentures, Surplus Notes, Premium Income Notes, Bonds, or Debentures, and Other Contingent Evidences of Indebtedness).
. See id. § 7.7(c), (e), (f); see also id. § 7.7(a)(3) (defining "surplus note").
. See id. § 7.7(f)(1).
. See id. § 7.7(b)(1).
.However, because the liability for and timing of the payments themselves were tied to minimum surplus levels, State Farm Lloyds was not projected to actually make any payments of interest or principal during the initial period, nor for some time thereafter. Rather, the schedule reflected the amortized cost of paying the required principal and interest over the course of the 15-year term as surplus levels increased past the threshold mínimums.
. See ASOP 30, at § 1.1 (observing that both underwriting provisions and contingency provisions "provide for the cost of capital”).
. See id. at § 3.7 (recommending inclusion of contingency provision), App. 1 at 6-7 (discussing role of contingency provision).
. See id. at § 3.7.
. See id.
. Elaborating, the Commissioner credited testimony that "since many exposures are not covered by the catastrophic insurance coverage and since in Texas, many catastrophes are *100too small to exceed the reinsurance deductible (or ‘attachment point’), the coverage is incomplete.”
.See former art. 5.26-1, § 5(a) ("Not later than the 10th day after the date of receipt of the commissioner’s order under Section 4 of this article, an insurer may file a petition for judicial review in a district court of Travis County.”).
. Tex. Gov’t Code § 2001.174; former art 5.16-1, § 5(a); see also former art. 5.142, § 13; Tex. Ins. Code § 2254.004(d).
. See Texas Health Facilities Comm'n v. Charter Med.-Dallas, Inc., 665 S.W.2d 446, 452-53 (Tex.1984).
. Slay v. Texas Comm'n on Envtl. Quality, 351 S.W.3d 532, 549 (Tex.App.-Austin 2011, pet. denied).
. See Texas Dep’t of Pub. Safety v. Alford, 209 S.W.3d 101, 103 (Tex.2006) (per curiam).
. See Slay, 351 S.W.3d at 549.
. See Vista Med. Ctr. Hosp. v. Texas Mut. Ins. Co, 416 S.W.3d 11, 26-27 (Tex.App.-Austin 2013, no pet.) (citing Charter Med.-Dallas, 665 S.W.2d at 453).
. See Granek v. Texas State Bd. of Med. Exam’rs, 172 S.W.3d 761, 778 (Tex.App.-Austin 2005, no pet.) (citing John E. Powers, Agency Adjudications 163 (1990)).
. Slay, 351 S.W.3d at 549.
. Tex. Gov’t Code § 2001.174(1).
. See Railroad Comm’n v. Texas Citizens for a Safe Future & Clean Water, 336 S.W.3d 619, 624 (Tex.2011); Montgomery Indep. Sch. Dist. v. Davis, 34 S.W.3d 559, 565 (Tex.2000) (citing Charter Med.-Dallas, 665 S.W.2d at 453); City of El Paso v. Public Util. Comm'n, 344 S.W.3d 609, 619 (Tex.App.-Austin 2011, no pet.); Buddy Gregg Motor Homes, Inc. v. Motor Vehicle Bd., 156 S.W.3d 91, 99 (Tex.App.Austin 2004, pet. denied).
. To this extent, we make an exception to the general rule that we reach constitutional issues only after exhausting non-constitutional grounds. See AEP Tex. Commercial & Indus. Retail Ltd. P'ship v. Public Util. Comm'n, 436 S.W.3d 890, 907 (Tex.App.-Austin 2014, no pet.) (citing In re B.L.D., 113 S.W.3d 340, 349 (Tex.2003) (“As a rule, we only decide constitutional questions when we cannot resolve issues on nonconstitutional grounds.’’)).
. The prohibitions against confiscatory government-imposed rates derive from Article I, Section 17 of the Texas Constitution, see Tex. Const. Art. I, § 17(a) ("No person’s property shall be taken ... for or applied to public use without adequate compensation being made .... ”), and the Fifth Amendment to the United States Constitution, see U.S. Const. Amend. V ("private property [shall not] be taken for *103public use, without just compensation"), as applied to the states through the Fourteenth Amendment.
. See Sheffield Dev. Co. v. City of Glenn Heights, 140 S.W.3d 660, 669 (Tex.2004) (characterizing the counterpart Texas and federal constitutional takings protections as "comparable"); City of Corpus Christi v. Public Util. Comm’n, 51 S.W.3d 231, 242 (Tex.2001) (observing that federal case law is instructive on whether rates constitute an unconstitutional taking of regulated entity’s property under the Texas Constitution); see also Bentley v. Bunton, 94 S.W.3d 561, 577-78 (Tex.2002) (noting that where parties do ■ not argue that differences in state and federal constitutional guarantees are material to a case, and none is apparent, "we will limit our analysis to the First Amendment and simply assume that its concerns are congruent with those of article I, section 8").
. See Federal Power Comm’n v. Hope Natural Gas Co., 320 U.S. 591, 64 S.Ct. 281, 88 L.Ed. 333 (1944). Although relying on these principles in State Farm Lloyds I to partly invalidate former art. 5.26-1, this Court did not have occasion to explore their parameters in detail. Instead, as State Farm Lloyds observes, the State Farm Lloyds I analysis "was based on a straightforward syllogism” that followed from the requirement that a government-set rate afford a regulated entity a reasonable rate of return on its capital — “rates can be confiscatory without necessarily leading to insolvency.” See State Farm Lloyds I, 255 S.W.3d at 794-95. Beyond its recognition of the basic constitutional concepts and that they apply to insurance rate regulation, State Farm Lloyds I provides us little guidance here.
. Duquesne, 488 U.S. at 307-08, 109 S.Ct. 609 (quoting Federal Power Comm’n v. Natural Gas Pipeline Co. of Am., 315 U.S. 575, 585, 62 S.Ct. 736, 86 L.Ed. 1037 (1942); Covington & Lexington Tpk. Rd. Co. v. Sandford, 164 U.S. 578, 597, 17 S.Ct. 198, 41 L.Ed. 560 (1896)).
. Federal Power Comm’n v. Texaco, Inc., 417 U.S. 380, 391-92, 94 S.Ct. 2315, 41 L.Ed.2d 141 (1974) (citing In re Permian Basin Area Rate Cases, 390 U.S. 747, 769, 88 S.Ct. 1344, 20 L.Ed.2d 312 (1968); Hope, 320 U.S. at 601, 64 S.Ct. 281).
. Id. (citing Natural Gas Pipeline, 315 U.S. at 585, 62 S.Ct. 736).
. Hope, 320 U.S. at 602, 64 S.Ct. 281 (internal citations omitted).
. Id. In the same vein, the Supreme Court would later observe:
The economic judgments required in rate proceedings are often hopelessly complex and do not admit of a single correct result. The Constitution is not designed to arbitrate these economic niceties. Errors to the detriment of one party may well be cancelled out by countervailing errors or allowances in another part of the rate proceeding. The Constitution protects the [regulated entity] from the net effect of the rate order on its property.
Duquesne, 488 U.S. at 307, 109 S.Ct. 609.
. 320 U.S. at 603, 64 S.Ct. 281.
. Id. (internal citations omitted).
. Permian Basin, 390 U.S. at 792, 88 S.Ct. 1344.
. Id. at 770, 88 S.Ct. 1344 (citing Hope, 320 U.S. at 603, 64 S.Ct. 281); see also Jersey Cent., 810 F.2d at 1189 (Starr, J., concurring) (suggesting that "confiscatory” "is a shorthand way of saying that an unreasonable balance has been struck in the regulation process so as unreasonably to favor ratepayer interests at the substantial expense of investor interests”).
.See City of Dallas v. Stewart, 361 S.W.3d 562, 568 (Tex.2012). Stewart concerned an administrative determination that private property was a nuisance, a determination tantamount to holding that constitutional takings protections were inapplicable. See id. at 569. While granting that judicial scrutiny of the agency's findings of "historical facts” (e.g., "whether or not the structure had foundation damage”) could be limited to assessing whether there was reasonable support by substantial evidence, the Texas Supreme Court held that due process required de novo review of the agency’s construction and application of the property owner's substantive rights under the takings clause (e.g., "Did the damage to the structure make it a threat to public *105health or safety such that the government may deprive a citizen of her ownership in the structure?”). Id. at 578-79 & n. 24. The court reasoned that such de novo review was required because agencies lacked delegated power to construe constitutional provisions and any legislative attempt to delegate them such power “ 'would raise serious and grave issues of a separation of powers violation.’ " Id. at 568 (quoting Ronald L. Beal, Texas Administrative Procedure & Practice § 9.3.1 [c] (2011)); see also id. at 577 (further emphasizing the "subordinate status” of administrative agencies “in our system of government” and "the ‘legitimacy deficit’ inherent in administrative adjudication”) (quoting Henry P. Monaghan, Constitutional Fact Review, 85 Colum. L.Rev. 229, 239 (1985)). As State Farm Lloyds emphasizes, its constitutional challenges are grounded in the same substantive constitutional rights as in Stewart, and that case, consequently, guides our review of those aspects of the Commissioner's order. Cf. Permian Basin, 390 U.S. at 792, 88 S.Ct. 1344 (indicating that reviewing court should "not ... supplant the [ratemakerj's balance of these interests with one more nearly to its liking, but instead [merely] assure itself that the [ratemaker] has given reasoned consideration to each of the pertinent factors”).
. See State Farm Lloyds I, 255 S.W.3d at 795 ("A government-set rate must allow a regulated company to not only recover its operating expenses, but also to realize reasonable returns on its investments sufficient to assure confidence in the continued financial integrity of the enterprise.”).
. See 28 Tex. Admin. Code § 7.7(f)(1) ("[T]he terms for payment of ... interest shall result in the reflection of a financial statement liability.”).
. See id. § 7.18 (2014) (TDI, National Association of Insurance Commissioners Accounting Practices and Procedures Manual) (adopting all Statements of Statutory Accounting Principles, except to the extent they contradict the Insurance Code or TDI rules).
. See National Association of Insurance Commissioners, Accounting Practices <& Procedures Manual of2005, Statements of Statutory Accounting Principles No. 41, par. 5, at 41-3.
. See Texas Indus. Energy Consumers v. CenterPoint Energy Houston Elec., LLC, 324 S.W.3d 95, 97 (Tex.2010); CenterPoint Energy, Inc. v. Public Util. Comm’n, 143 S.W.3d 81, 83 (Tex.2004).
. See Hope, 320 U.S. at 603, 64 S.Ct. 281.
. See, e.g., Central Power & Light, 36 S.W.3d at 553 (describing how costs of gas utility's capital, including debt capital, are determined in calculating the rate of return) (citing Southern Union Gas Co. v. Railroad Comm'n, 692 S.W.2d 137, 141 (Tex.App.-Austin 1985, writ ref'd n.r.e.)); Texas Water Comm’n v. Lakeshore Util. Co., Inc., 877 S.W.2d 814, 819, n. 6 (Tex.App.-Austin 1994, writ denied) (noting that “utility’s interest payments on long-term debt are used to compute the utility's rate of return”) (citing Southern Union, 692 S.W.2d at 141).
. See Hope, 320 U.S. at 602, 64 S.Ct. 281.
. Cf., e.g., Southern Union Gas, 692 S.W.2d at 141 (distinguishing different methods for "costing" long-term debt capital, common stock capital, and preferred stock capital within determination of gas utility's rate of return, and explaining that the cost of long-term debt capital is simply "the interest rate contractually agreed upon by the investor and the [regulated entity]”).
. See id.
. State Farm Lloyds summarizes its contentions as follows:
*108Regardless of how "capital costs” might ordinarily be treated for accounting or ratemaking purposes, and regardless of whether an insurer’s replenishing surplus depleted by catastrophes can properly be considered an "investment in capital” in the traditional sense (a company investing in an asset to turn a profit), the fundamental point is this: unless the rates set for the insurer allow it simultaneously to (1) rebuild and maintain its surplus at prudent levels, and (2) timely pay off both interest and principal on its surplus note, the financial integrity of the insurer will be crippled. Without rates allowing for both (1) and (2), the insurer will not be able to "operate successfully” (because it will have surplus below prudent levels) and it will not be able to "maintain its credit” and attract capital” (because lenders don't renew loans to deadbeat clients).
Reply br. at 20-21 (citing Hope, 320 U.S. at 603, 64 S.Ct. 281).
. See Hope, 320 U.S. at 603, 64 S.Ct. 281 ("[T]he return to the equity owner should be commensurate with returns on investments having corresponding risks.”).
. State Farm Lloyds suggests that the Commissioner conceded in his order that nothing short of a 1:1 premium-to-surplus ratio would be "prudent” in its insurance operations. This is not entirely accurate. The Commissioner did accept a characterization by Appel that a 1:1 premium-to-surplus ratio was "minimally prudent,” but this was only with respect to the hypothetical, risk-adjusted level of capital that was utilized in calculating underwriting profit. See infra at 40-41. The Commissioner explicitly rejected the 1:1 ratio as a requirement or minimum benchmark in assessing State Farm Lloyds's financial condition. Order at 101 & n.276.
. Texaco, 417 U.S. at 391-92, 94 S.Ct. 2315.
.Bluefield Waterworks & Improvement Co. v. Public Serv. Comm’n, 262 U.S. 679, 694, 43 S.Ct. 675, 67 L.Ed. 1176 (1923); see also Market St. Ry. Co. v. Railroad Comm’n of State of Cal., 324 U.S. 548, 567, 65 S.Ct. 770, 89 L.Ed. 1171 (1945) ("The due process clause has been applied to prevent governmental destruction of existing economic values. It has not and cannot be applied to ... restore values that have been lost by the operation of economic forces."); Galveston Elec. Co. v. City of Galveston, 258 U.S. 388, 395, 42 S.Ct. 351, 66 L.Ed. 678 (1922) (rate cannot be held confiscatory merely on the basis that it is insufficient to enable the company to recoup prior losses).
. State Farm Lloyds offers the cautionary example of an affiliate, then the largest homeowners insurer in Florida, who elected to withdraw from that market after it "was denied adequate rates after unanticipated surplus-draining hurricane losses [reduced] its surplus ... to dangerously low levels despite a multi-million dollar surplus-note loan from State Farm Mutual.”
. See, e.g., Permian Basin, 390 U.S. at 770, 88 S.Ct. 1344 ("... any rate selected ... from the broad zone of reasonableness permitted by the Act cannot properly be attacked as confiscatory”).
.See CenterPoint Energy Entex v. Railroad Comm’n, 213 S.W.3d 364, 373 (Tex.App.-Aus-tin 2006, no pet.) ("An agency may accept or reject the testimony of witnesses. An agency is not obliged to accept opinion testimony from an expert- — even if it is the sole evidence on the issue. But an agency must provide a basis for its rejection of uncontradicted, un-impeached testimony that is neither inherently improbable or conclusory. The [agency] can reject such uncontradicted evidence if it explains or makes findings that permit courts to review the reasonableness of that rejection, but a failure to explain can result in reversal.’’) (internal citations omitted).
. See, e.g., Pioneer Natural Res. USA, Inc. v. Public Util. Comm’n, 303 S.W.3d 363, 367-68 (Tex.App.-Austin 2009, no pet.) (agency " 'is the sole judge of the weight to be accorded the testimony of each witness ...’”) (quoting Central Power & Light, 36 S.W.3d at 561).
. Railroad Comm'n v. Lone Star Gas Co., 618 S.W.2d 121, 124-25 (Tex.Civ.App.-Austin 1981, no writ).
. See Railroad Comm'n v. Moran Util. Co., 728 S.W.2d 764, 766-67 (Tex. 1987); Central Power & Light, 36 S.W.3d at 557-58; Lone Star Gas, 618 S.W.2d at 124-25.
. See Pioneer, 303 S.W.3d at 367-70 (although no witness testified specifically that 35% of utility's computer system should be considered used and useful and included in its rate base, agency could reasonably infer that calculation from evidence that (1) 70% of the system’s cost was allocable to its regulated operations and (2) 50% of the system had been used and useful during the test year; observing that "even where no evidence suggests a specific figure explicitly, the [agency] may infer that figure if it is supported by the body of evidence on that issue”); see also Central Power & Light, 36 S.W.3d at 547 (observing that because an agency "may accept or reject in whole or in part the testimony of the various witnesses who testify ... [i]t follows ... that substantial evidence will support a method of calculating [a rate component] that is an amalgam of methods proposed by different witnesses”).
. See City of El Paso v. Public Util. Comm'n, 883 S.W.2d 179, 185-86 (Tex.1994) (where evidence "ranged from expert testimony that no imprudence disallowance should be imposed, to testimony that a 50% imprudence disallowance should be imposed, and finally to testimony that there is no known theory to quantify the flaws in [the utility’s] decision making process,” and "because of the admitted complexity in valuing decisional imprudence in this case," holding that "there is a reasonable basis for the Commission to, in its discretion, select an amount within the range of figures provided by expert testimony of the parties”); Pioneer, 303 S.W.3d at 373-74 (where parties presented competing evidence as to whether utility’s capital structure for purposes of ratemaking should be its actual 83/17% debt-to-equity ratio versus a hypothetical 60/40% ratio, holding that Commission's selection of a hypothetical 75/25 figure was supported by substantial evidence, as “the evidence presented in favor of any against each endpoint was such as would allow for a hypothetical capital structure to be selected between those endpoints”); cf. id. at 370 (also acknowledging that, in a particular case, the evidence might conceivably permit only a binary choice between two end points).
. See Tex. Gov't Code § 2001.174(2) (reversal required only "if substantial rights of the appellant have been prejudiced”).
. See Pioneer, 303 S.W.3d at 367-70.
. See Tex. Gov’t Code § 2001.174(2) (reversal required only "if substantial rights of the appellant have been prejudiced”).
. See ASOP 30, at 7.
. See id.
. See Central Power & Light, 36 S.W.3d at 557-58.
. See City of El Paso, 883 S.W.2d at 185-86.
. Pioneer, 303 S.W.3d at 370; see also id. at 373 & n. 1 (rejecting complaints that no evidence established that a 75/25 capital structure "was more correct than a 77/23 structure, a 79/21 structure, or a 81/19 structure,” as "our inquiry is not whether the evidence establishes that the capital structure selected was the best option available,” only whether there was a reasonable basis for it, and in that regard "the record evidence may preponderate against the agency’s decision and yet still ... satisfy the substantial evidence standard) (emphasis in original).
.See former art. 5.142, § 14. In fact, State Farm Lloyds acknowledged as much in State Farm Lloyds III, observing that "[i]f [it] loses, and the Commissioner’s rate reduction order is reinstated on appeal, then Articles 5.26-1 and 5.142 provide their own remedy for any overcharges during that period ... [it] will be required to refund the 'overcharged premium’ to each policyholder, plus interest, in accordance with the terms of those statutes.” 267 S.W.3d at 446.
. See Act of June 2, 2003, 78th Leg., R.S., ch. 206, § 1.01, 2003 Tex. Gen. Laws 907, 912-13 (currently codified at Tex. Ins. Code §§ 2254.001-.004) (“former art. 5.144”).
. See former art. 5.144(b).
. See former art. 5.144(c).
. State Farm Lloyds also argues that the specific 0.5% downward adjustment the Commissioner made was not supported by substantial evidence. We need not reach this contention.
. The Commissioner found that both earned premiums and non-catastrophe loss and LAE would increase by 1.3% each year, hurricane catastrophe loss by 1.8%, non-hurricane catastrophe loss and LAE by 4.85%, and fixed expenses by 3%.
. I.e., one year after the last day in which State Farm Lloyds’s implemented rate had been in effect, July 31, 2008.
. See Central Power & Light, 36 S.W.3d at 547.
. See, e.g., State Farm Lloyds II, 260 S.W.3d at 245 ("Even if supported by substantial evidence, however, an agency’s order may be arbitrary and capricious...
. Id. at 245 (citing City of El Paso, 883 S.W.2d at 184; Dunn v. Public Util. Comm'n, 246 S.W.3d 788, 791 (Tex.App.-Austin 2008, no pet.)).
. Duquesne, 488 U.S. at 315, 109 S.Ct. 609.
. See Tex. Gov't Code § 2001.174(2)(f) (requiring reversal of agency "findings, inferences, conclusions, or decisions” that are” "arbitrary or capricious”); Granek, 172 S.W.3d at 782 (reversing agency decision that was arbitrary and capricious).
. Former art. 5.26-1, § 6.
. Former art. 5.142, § 14.
. As originally enacted in S.B. 14, former art. 5.144 contained no requirement that interest be awarded on refunds held to be due. See former art. 5.144. During the Legislature's following regular session, it twice amended former art. 5.144 to require interest on refunds at the rate described above. See Act of May 29, 2005, 79th Leg., R.S., ch. 291, § 1, 2005 Tex. Gen. Laws 869; Act of May 29, 2005, 79th Leg., R.S., ch. 899, § 16.01, 2005 Tex. Gen. Laws 3098, 3112. However, during the same session, the Legislature recodi-fied the underlying provisions of former art. 5.144 — minus the amendments adding the interest requirement — into chapter 2254 of the Insurance Code. See Act of May 24, 2005, 79th Leg., R.S., ch. 727, § 2254.002, 2005 Tex. Gen. Laws 1752, 2145. Although omitted from the recodification, the new interest requirement was nevertheless preserved and effective by operation of law. See Tex. Gov’t Code § 311.031(c) (“The repeal of a statute by a code does not affect an amendment, revision, or reenactment of the statute by the same legislature that enacted the code. The amendment, revision, or reenactment is preserved and given effect as part of the code provision that revised the statute so amended, revised, or reenacted.”). The Legislature corrected the omission from the recodification during its next regular session with amendments to section 2254.003 of the Insurance Code, see Act of May 23, 2007, 80th Leg., R.S., ch. 484, § 3, sec. 2254.003, 2007 Tex. Gen. Laws 850, 851, and the provision has remained in substantially this form through the present time, see Tex. Ins. Code § 2254.003.
. See supra note 111.