In 1999, the Legislature substantially revised the Public Utility Regulatory Act (PURA) to bring about a major restructuring of the electric power industry in Texas to allow retail electric rates to be determined by competition. As part of that restructuring, the PURA permits existing utilities to recover “regulatory assets” and “stranded costs” through securitization financing. Securitization is accomplished through a financing order issued by the Public Utility Commission that authorizes a utility to issue a transition bond. The transition bond is repaid or secured by transition charges to electric power consumers in a utility’s service area. Central Power and Light Company, an existing utility, applied for and the Commission approved a financing order that assures that CPL will recover certain of its regulatory assets through securitization. Two separate proceedings were brought in a Travis County district court seeking review of that order on different grounds. Final judgments were rendered in both proceedings affirming the Commission’s order. We granted direct appeals from those judgments pursuant to section 39.303(f) of the PURA and consolidated the two proceedings.

Power Choice, Inc., the appellant in one of the appeals, contends that the securiti-zation provisions of the PURA are facially unconstitutional under the Texas Constitu*236tion because they impose a tax that is not for a public purpose, constitute a taking without adequate compensation, or are an appropriation or grant for private purposes. We affirm the trial court’s judgment that the securitization provisions are not unconstitutional on any of those grounds.

In the appeal by numerous cities including the City of Corpus Christi, and by Texas Industrial Energy Consumers and the Office of Public Utility Counsel, we also affirm the trial court’s judgment. We hold that: 1) regulatory assets known as “SFAS 109 assets” may be securitized even though they currently earn no return and have no carrying costs; 2) the Commission properly treated investment tax credits; 3) the Commission did not err in securitizing regulatory assets reflected in CPL’s SEC Form 10-K rather than the balance of those assets as of December 31, 2001; 4) the PURA authorizes the Commission to prescribe what it calls a “nonstandard true-up”; 5) the Commission did not err in declining to adjust the allocation factors for industrial customer classes to reflect load loss; 6) the Commission did not err in its allocation of transition charges to non-firm industrial customer classes; and 7) the Cities were not denied due process in the proceedings before the Commission. Accordingly, we affirm the trial court’s judgment. Justice Owen’s concurring opinion is the opinion of the Court with respect to the issues that it addresses, and Justice Hecht’s concurring opinion is the opinion of the Court with respect to the issue that it addresses.

Justice HECHT filed a concurring opinion, in which Chief Justice PHILLIPS, Justice ABBOTT, Justice HANKINSON, and Justice JEFFERSON joined.

Justice OWEN filed a dissenting opinion, in which Justice ENOCH and Justice BAKER joined.

joined by Chief Justice PHILLIPS, Justice HECHT, Justice ENOCH, Justice BAKER, Justice ABBOTT, Justice HANKINSON, and Justice JEFFERSON, concurring.

I

The Public Utility Regulatory Act (PURA) first came into being in 1975. At that time, the Legislature established a comprehensive regulatory system for electric utilities.1 The Legislature had concluded that these utilities were “by definition monopolies in the areas they serve,” and that as a result “normal forces of competition which regulate prices in a free enterprise society do not operate.”2 Regulation was intended to be a substitute for competition.3 Although there were changes over the years in the manner in which the PURA regulated the electric power industry, and although by 1997 the Legislature had recognized that the wholesale electric industry was becoming more *237competitive,4 regulation under the PURA remained “comprehensive.”5

In 1999, the Legislature decided to chart a new course for the provision of electric service in Texas. In the years intervening since 1975, partial deregulation at the federal level and deregulation in other states had wrought significant changes in the electric industry on a national level. The Legislature concluded that it was in the public interest to establish a “fully competitive electric power industry” in Texas.6 It enacted chapter 39 of the PURA and other amendments to accomplish that goal and “to protect the public interest during the transition.”7

In order to achieve competition in the retail market for electricity, the amendments to the PURA require existing utilities to “unbundle” the services that they provide.8 Each electric utility must separate its business activities into distinct units: 1) a power generation company; 2) a retail electric provider; and 3) a transmission and distribution utility.9 This is to be accomplished through the creation of separate nonaffiliated companies, the creation of separate affiliated companies owned by a common holding company, or the sale of assets to a third party.10

Underpinning the Legislature’s decision to restructure the electric power industry was its finding that regulation was no longer warranted, except for regulation of transmission and distribution services and regulation of the recovery of stranded costs:

The legislature finds that the production and sale of electricity is not a monopoly warranting regulation of rates, operations, and services and that the public interest in competitive electric markets requires that, except for transmission and distribution services and for the recovery of stranded costs, electric services and their prices should be determined by customer choices and the normal forces of competition.11 Stranded costs have a precise, technical definition under chapter 39 of the PURA:

“Stranded cost” means the positive excess of the net book value of generation assets over the market value of the assets, taking into account all of the electric utility’s generation assets, any above market purchased power costs, and any deferred debit related to a utility’s discontinuance of the application of Statement of Financial Accounting Standards No. 71 (“Accounting for the Effects of Certain Types of Regulation”) for generation-related assets if required by the provisions of this chapter. For purposes of Section 39.262 [true-up proceeding], book value shall be established as of December 31, 2001, or the date a market value is established through a market valuation method under Section 39.262(h), whichever is earlier, and shall include stranded costs incurred under Section 39.263 [stranded cost recovery of environmental cleanup costs].12

Stranded costs can more generally be described as the portion of the book value of a utility’s generation assets that is project*238ed to be unrecovered through rates that are based on market prices. Under the regulatory scheme that existed prior to 1999, a utility would have had an opportunity to recover prudent capital investments in its rates through depreciation. The Legislature concluded, after intensive study by the Public Utility Commission and others, that those investments are unlikely to be recovered once a competitive retail market based on the market price of electricity is established. That is because for years to come, existing utilities would have costs associated -with historical costs of service or investments in facilities while new sellers of electricity would not. Accordingly, new marketers should be able to sell electricity at prices that are lower than prices that would permit incumbent utilities to recover all their existing, embedded costs and capital investments. In order to compete, existing utilities, or more precisely their shareholders, would have to absorb those costs and losses on investments because market prices would not provide a sufficient return. The largest part of stranded costs for utilities in Texas, including CPL, is attributable to investments in nuclear power plants that the Commission previously found in rate proceedings were prudently incurred.

The Legislature determined that it is in the public interest for existing utilities to recover certain stranded costs in charges that are “nonbypassable.” 13 That means that with limited exceptions, all retail electric customers in an existing utility’s service area will pay charges to allow that utility to recover stranded costs regardless of whether those customers purchase their electricity from that utility, switch to one of its competitors, or generate their own electricity.14

The Legislature similarly concluded that “regulatory assets” should be recoverable through nonbypassable charges.15 The definition of “regulatory assets” is, like the definition of stranded costs, technical:

“Regulatory assets” means the generation-related portion of the Texas jurisdictional portion of the amount reported by the electric utility in its 1998 annual report on Securities and Exchange Commission Form 10-K as regulatory assets and liabilities, offset by the applicable portion of generation-related investment tax credits permitted under the Internal Revenue Code of 1986.16

Regulatory assets are a subset of generation-related costs reasonably incurred by a utility that the Commission has determined in prior rate cases could be included in rates and recovered over a period of years instead of at the time the expenditure was made. They differ from “stranded costs,” as defined above, because stranded costs are investments in or the cost of tangible assets. Regulatory assets are essentially bookkeeping entries that reflect a charge to be included in a utility’s future rates. In the Commission’s reports to the Legislature about major issues that would arise if the electric power industry were deregulated, the Commission projected that all of a utility’s regulatory assets would be stranded if competition in retail sales occurred and no provision was included to address them. That is because regulatory assets have no market value absent a regulatory scheme that assures their recovery.

In deciding to proceed with deregulation of retail rates, the Legislature concluded that incumbent utilities should be allowed *239to recover stranded costs and regulatory assets through a nonbypassable “competition transition charge” imposed under sub-chapters E and F of chapter 39 of the PURA.17 That charge would ultimately be paid by retail electric customers in an incumbent utility’s geographic service area as part of the rates they pay for electric service regardless of whether they receive service from the incumbent utility or choose a new provider.18 The only exemption from competition transition charges is for users who receive power from certain qualifying cogeneration or on-site facilities.19

The Legislature also established an alternative method by which a utility could recover certain stranded costs and regulatory assets, which is through securitization financing under subchapter G of chapter 39 of the PURA. The Legislature authorized the Commission to adopt financing orders that permit an electric utility to issue transition bonds or other evidences of indebtedness.20 The proceeds of the bonds must be used to reduce the amount of recoverable regulatory assets and stranded costs through refinancing or by retiring a utility’s debt or equity.21

The Commission is directed by section 39.301 of the PURA to ensure that securi-tization of costs by a utility results in benefits to consumers. Section 39.301 provides:

The commission shall ensure that securi-tization provides tangible and quantifiable benefits to ratepayers, greater than would have been achieved absent the issuance of transition bonds.... The amount securitized may not exceed the present value of the revenue requirement over the life of the proposed transition bond associated with the regulatory assets or stranded costs sought to be securitized.22

A securitization financing order establishes “transition charges” to be paid by retail customers in a utility’s service area that allow recovery of “qualified costs.”23 Qualified costs include all of an electric utility’s regulatory assets and seventy-five percent of certain stranded costs, plus the cost of issuing, supporting and servicing transition bonds, and costs of retiring or refunding existing debt and equity securities in connection with the issuance of transition bonds.24 Transition bonds are to be secured by or payable from “transition property,” which includes the right to impose and collect transition charges.25 Accordingly, transition charges to retail customers will retire the transition bonds by paying all principal and interest. The transition charges, like competition transition charges, are nonbypassable and are allocated among electric consumers in an incumbent utility’s service area in the same manner as competition transition charges.26

The PURA sets forth a “true-up” mechanism that is intended to ensure that no utility overrecovers stranded costs through competition transition charges.27 There is *240also a true-up mechanism for transition charges imposed as part of the securitization procedure to ensure that those charges will be sufficient to meet transition bond obligations but will not generate any revenue beyond those requirements.28

The financing order at issue in this case authorizes CPL to obtain securitization financing for the net amount of $763,734,489 in regulatory assets. The financing order establishes transition charges to ensure payment of the transition bonds. None of CPL’s stranded generating plant costs are included in the financing order. Stranded generating plant costs are to be estimated in a proceeding that was pending before the Commission at the time this case was submitted. That proceeding is not part of these appeals. Nevertheless, the appeal by Power Choice broadly challenges the constitutionality of transition charges, irrespective of whether the underlying transition bonds securitize regulatory assets or stranded costs. Because Power Choice’s principal challenges would moot the appeal by the City of Corpus Christi and others, we turn first to Power Choice’s appeal.

II

Power Choice hopes to compete with CPL in the geographic area in which CPL is currently authorized by the Commission to provide service. The financing order under consideration establishes transition charges and directs all retail electric suppliers in CPL’s service area, which would include Power Choice, to collect those transition charges and pass them on to CPL.

Power Choice contends that the entire securitization scheme under the PURA is unconstitutional on several grounds. Power Choice argues that the nonbypassable charges are not rates because, it contends, they are unrelated to the cost of electricity that is actually used by retail customers. Power Choice also contends that transition charges are not rates because they are unrelated to any service provided to customers if they choose to purchase electricity from retail providers other than an incumbent utility. Power Choice further asserts that transition charges have no underlying public purpose because they are only a transfer of money from one private party to another. The result, Power Choice contends, is that the securitization provisions violate “the protections of the tax, taking, and appropriation and grant clauses of the Texas Constitution.” Power Choice relies on Article VIII, Section 3; Article I, Section 17; Article XVI, Section 6, and Article III, Section 51 of the Texas Constitution.

In approving CPL’s financing order, the Commission did not specifically address Power Choice’s constitutional arguments, apparently concluding that it lacked the power to rule on them. Power Choice appealed to a Travis County district court in accordance with section 39.303(f) of the PURA. That court rejected all of Power Choice’s arguments, holding that the secu-ritization provisions of the PURA did not violate any of the sections of the Texas Constitution on which Power Choice relied.

In this Court, as in the district court, Power Choice primarily mounts a facial challenge to the constitutionality of the securitization provisions of the PURA. In Texas Workers’ Compensation Commission v. Garcia we explained the difference between a facial challenge and an “as applied” challenge.29 In the latter a party concedes that a statute is generally constitutional but contends that the statute is unconstitutional when applied to a particular person or set of facts.30 To sustain a *241facial challenge, the complaining party “must establish that the statute, by its terms, always operates unconstitutionally.”31 We consider each of Power Choice’s constitutional challenges in turn.

Ill

Power Choice asserts that transition charges constitute a taking of money from consumers of electricity that is uncompensated because the charges are unrelated to the current provision of electric service in a competitive market. Power Choice contends that for customers of an incumbent utility, the transition charges bear no relation to the current market value of the electricity received, and that customers of a new provider will receive nothing at all in return for their payment of transition charges. We are unpersuaded that transition charges are a taking without adequate compensation in violation of Article I, Section 17 of our Texas Constitution.32

A

The first question that we must resolve is what analytical framework applies in determining whether charges by a utility amount to an unconstitutional taking. Historically, states have regulated utilities and utility rates under the police power. The United States Supreme Court has said that “the regulation of utilities is one of the most important of the functions traditionally associated with the police power of the States.”33 States have the authority to regulate public utility rates as long as the rates do not result in a “deprivation qf property without due process of law or the taking of private property for public use without just compensation.”34 Within those confines, a state has broad discretion.35 The United States Supreme Court has also said that in reviewing rates set under a state’s authority, courts “do not sit as a board of revision, but to enforce constitutional rights.”36 The question is “whether the rates as fixed are confiscatory,” and a “[cjourt may not interfere with the exercise of the State’s authority unless confiscation is clearly established.” 37

*242This Court has not previously considered in any depth when utility rates constitute a taking within the meaning of the Texas or United States Constitutions. We have noted the similarity between the federal takings clause and the prohibitions of Article I, Section 17 of the Texas Constitution.38 And in Railroad Commission v. Houston Natural Gas Corp., we looked to United States Supreme Court decisions for general guidance on when rates would constitute a taking of a utility’s property.39 In this case, a utility is not contending that rates have been set so low that there has been a taking of its property. Instead, Power Choice contends that rates in the form of transition charges will constitute a taking from consumers regardless of the amount of those rates and regardless of which consumers pay them.

The reasoning in decisions of the United States Supreme Court that have arisen in the context of complaints by consumers that a utility’s rates are unjust and unreasonable is instructive.

In Federal Power Commission v. Hope Natural Gas Co.,40 consumers and regulatory agencies contended that rates charged by a utility were excessive and unreasonable within the meaning of the Natural Gas Act. The United States Supreme Court reaffirmed in that case that Congress had the constitutional authority to regulate the rates at issue and that Congress’s authority was “at least as great under the Fifth Amendment as is that of the States under the Fourteenth [Amendment] to regulate the prices of commodities in intrastate commerce.”41 Therefore, when a consumer filed a complaint, the Federal Power Commission was free to establish rates as long as those rates were just and reasonable.42 The United States Supreme Court later explained in Permian Basin Area Rate Cases, that “the just and reasonable standard of the Natural Gas Act ‘coincides’ with the applicable constitutional standards,” and that “any rate selected by the [Federal Power] Commission from the broad zone of reasonableness permitted by the Act cannot properly be attacked as confiscatory.”43 The Supreme Court had also explained in Hope Natural Gas that the “rate-making process under the [Natural Gas] Act, i.e., the fixing of ‘just and reasonable’ rates, involves a balancing of the investor and the consumer interests.”44 The Court elaborated in Permian Basin that as long as a regulatory authority has balanced the interests of consumers and investors, rates established within a zone of reasonableness are not subject to attack as confiscatory:

Accordingly, there can be no constitutional objection if the [Federal Power] Commission, in its calculation of rates, takes fully into account the various interests which Congress has required it to reconcile. We do not suggest that maximum rates computed for a group or geographical area can never be confiscatory; we hold only that any such rates, *243determined in conformity with the Natural Gas Act, and intended to “balanc[e] ... the investor and the consumer interests,” are constitutionally permissible.45

Rates may be substantially higher than the lowest reasonable rate that is not confiscatory to a utility and yet not be excessive when charged to a consumer.46 The United States Supreme Court has said that a regulatory authority “is not required to prescribe rates so low as to be barely sufficient to withstand attack on the ground of confiscation, but is at liberty within limits that [it] may find to be just and reasonable to establish higher rates.” 47 Accordingly, the zone of reasonableness for takings clause purposes lies between rates that would constitute a confiscation of a utility’s property and rates that would be excessive or unjust or unreasonable if a consumer were required to pay them.

Recently, the Supreme Court of New Hampshire utilized the “zone of reasonableness” principles when it was called upon to review legislation that deregulated certain aspects of that state’s electric power industry.48 In considering the contention that charges for stranded costs are an unconstitutional taking, the court’s analysis turned on whether the rate allowing recovery of stranded costs was a just and reasonable one “ ‘falling] within the zone of reasonableness between confiscation of utility property or investment interests and ratepayer exploitation.’ ”49

Armed with these basic principles, we consider Power Choice’s arguments.

B

Power Choice contends that consumers should not be burdened with obligations that an incumbent utility incurred in the past to provide service or build facilities because those costs are unrelated to the current provision of electric sendee. The first question that we must decide, then, is whether, “during the transition to and in the establishment of a fully competitive electric power industry,”50 it is unjust or unreasonable for the Legislature to establish rates that will “allow utilities with uneconomic generation-related assets and purchased power contracts to recover the reasonable excess costs over market of those assets and purchased power contracts.” 51 Stated another way, is it unjust or unreasonable, and therefore confiscatory as to consumers, for the Legislature to allow an incumbent utility to recover through transition charges costs incurred for construction of a power generation infrastructure?

We have held that it is not unjust or unreasonable to permit a utility to recover past costs over a future period through rates.52 Indeed, we previously held in State v. Public Utility Commission that *244the largest part of regulatory assets addressed by the financing order at issue in this case, which is CPL’s Deferred Accounting asset, may be recovered over a period of years in rates.53 CPL participated in the construction of the South Texas nuclear plant. When that plant first went into operation, there were costs incurred between the date of commercial operation and the effective date of new rates set by the Commission. The Commission created deferred accounts for those costs and found that all amounts in certain of CPL’s deferred accounts relating to the South Texas Nuclear Project were prudent and reasonable and should be recovered in CPL’s rates over the forty-year life of the South Texas project.54 This Court approved of the creation of deferred accounting assets for CPL’s investment in the South Texas Nuclear Project.55 Under the regulatory scheme that existed until the 1999 amendments to the PURA that are at issue in the case, existing and future customers served by CPL’s system would pay those costs over time, even though the actual expenditures or investments were made a number of years ago.

The transition charges that are established under the 1999 amendments to the PURA are essentially a conversion from one form of rate recovery to another. Through transition bonds, the costs that a utility would otherwise have had the opportunity to recover in rates are to be refinanced. At least one court has characterized transition charges like those at issue here as “nothing more than a different manifestation of the previously regulated rates.”56 That court went on to explain that transition charges are simply a novel means of continuing regulation:

Had Pennsylvania continued the system in effect prior to the Competition Act, PUC would have allowed PECO to recover these costs through the rates it charged for its “bundled” services without labeling them transition charges. The stranded-cost provisions are simply a novel way to accommodate the need for recovery of these costs consistent with traditional state regulation.57

Spreading the recovery of costs incurred in the past over a period of years is not a concept unique to regulation of utilities in Texas. Recently, the United States Court of Appeals for the District of Columbia Circuit had occasion to review orders of the Federal Energy Regulatory Commission that partially deregulated the interstate electric industry and provided for recovery of stranded costs, including regulatory assets.58 The D.C. Circuit observed that regulatory assets are “nonrecurring costs approved by regulators that, in order to avoid rate increases, were recovered over a period of years instead of at the *245time the expenditures were made.”59 Similarly, the Supreme Court of Connecticut has observed that “[i]n order to avoid rate shock, commissions often will permit utility companies to recover their expenses from ratepayers on a deferred basis, listing the ratepayers’ debt as a ‘regulatory asset.’ ”60 The fact that rate recovery of these costs was and is to be spread over time does not render the rates unjust or unreasonable. The Supreme Court of New Hampshire recently rejected the argument that stranded cost recovery was unconstitutional because it allowed recovery of past investment in generation assets.61 That court confirmed that “ ‘current rates often include past costs that utilities deferred in order to avoid rate increases.’ ”62

Implicit, if not explicit, in these fundamental ratemaking principles is the recognition that even though a particular consumer does not derive a direct benefit from the use of particular assets, that does not render rates that include costs associated with those assets unjust or unreasonable. We considered a similar proposition in Cities for Fair Utility Rates v. Public Utility Commission,63 The issue was whether existing consumers can be required to bear costs that would benefit future customers, rather than whether consumers can be required to bear historical costs. But the argument made in Cities was essentially the same argument that Power Choice makes. In Cities, existing consumers complained that there would be an unfair and unlawful allocation of costs as between them and future ratepayers if a utility were allowed to recover costs incurred in connection with facilities that were not planned to go online for several years. We held that expenditures for “plant held for future use, or PHFU” could be included in a utility’s rate base before the plant went into service,64 We recognized that a regulated utility must make long-term plans and investments to meet the future needs of those in its service area. We said that in balancing the differing interests between present and future ratepayers, it was permissible for the Commission to require existing customers to pay costs associated with plans for future service:

While it is not fair to charge present ratepayers with the cost of future service, neither is it fair to burden future ratepayers with unnecessarily high acquisition costs because a utility was discouraged from making prudent long-term plans. PHFU expenses are used and useful because they are a necessary part of planned investments. Most states include PHFU in rate base in some circumstances.65

We accordingly held that a utility could include approximately $93 million of costs incurred in connection with a generating plant that it planned to build in ten years.66 Even though the utility ultimately can-celled completion of the plant, there was evidence that the utility had plans to bring *246the plant online at some point in the future when it included costs associated with the plant in its rate base.67 We had previously held in Southwestern Bell Telephone Co. v. Public Utility Commission that a utility could include in rate base the cost of land acquired for future use.68

Under the PURA, a utility’s recovery of regulatory assets through rates has, by definition, been determined in previous rate cases to be within the zone of reasonableness as between a utility and the customers who would pay its rates. And by definition, stranded costs were incurred in providing electric power service. Under the 1999 amendments to the PURA, a utility may recover only “net, verifiable, nonmitigable stranded costs incurred in purchasing power and providing electric generation service.”69 Rates that permit recovery of regulatory assets or net, verifiable, nonmitigable costs incurred in connection with the provision of electric generation service are not confiscatory, even if the time period over which those costs are recovered does not coincide with the time that the costs were actually incurred or the time that the electric generation service was provided.

Were it otherwise, no residential rates could pass constitutional muster. Residential customers continually move in and out of a utility’s service area. Rates are not adjusted so that each consumer pays only for costs incurred by the utility when that consumer actually lived in the area and was served by the utility. It would be utterly impracticable to set and then continually adjust rates on such a basis.

Another ground on which Power Choice bases it takings argument is that unless a utility has the constitutional right to recover stranded costs and regulatory assets, it would be unconstitutional for the Legislature to allow recovery of those costs and assets. We need not and do not decide whether an incumbent utility’s constitutional rights would be abridged if it were not permitted to recover stranded costs. As we have seen from some of the decisions discussed above, a regulator may lawfully allow a utility to recover certain costs even though a denial of recovery of those costs would not amount to confiscation from a utility.70 In balancing the competing interests of consumers and a utility’s investors, there is a zone of reasonableness within which a regulatory authority may set rates.71 The Constitution does not prohibit the regulator from permitting recovery, as long as the rates consumers are required to pay are not excessive, unjust, or unreasonable.72 In balancing the interests of consumers and public utilities, the Legislature can constitutionally conclude that a utility is entitled to recover reasonable costs associated with the generation of electric power.

In sum, a regulatory authority, and certainly the Legislature, may conclude that it is appropriate to spread recovery of a utility’s costs over time. The fact that a particular consumer does not derive a direct benefit from the past use of particular *247assets does not render rates that include costs associated with those assets confiscatory.

C

Power Choice argues that the PURA’s securitization provisions result in a taking as to new customers moving into a utility’s service area and a utility’s existing customers, at least if these new and existing customers choose to purchase power from a new supplier. Those consumers, Power Choice contends, receive no benefit at all from a utility’s past investments or costs. This is an as-applied constitutional challenge, rather than a facial challenge.

For many of the reasons that we discussed in section IIIB above, we reject Power Choice’s argument that transition charges are confiscatory as to new and existing consumers who will buy electricity from suppliers other than CPL. If the Legislature had not enacted the 1999 amendments to the PURA, many new and existing consumers who will now want to buy power from a new supplier would have been served by an incumbent utility. As explained above, many customers would not have had the option under the previous regulatory regime to seek another supplier because there would have been no other supplier. Those consumers would have been required to pay rates that allowed the incumbent utility to recover costs associated with obligations incurred in the past, even though those particular consumers may not have been customers of the utility when it incurred the costs or obligations.

It is not unjust or unreasonable to require consumers for whose benefit an electric power generation infrastructure was constructed to share in some of the costs of that infrastructure, even though they may not directly benefit from it under the new regulatory scheme. Under the regulatory scheme in which regulatory assets were created and stranded costs were incurred, an electric utility had a public-service obligation to render service to all those in its service area at regulated rates.73 As one court has explained, stranded costs “are the costs prudently incurred by the local utilities that will not be recoverable through market-determined prices, and that result from the utilities’ reliance on the previous regulatory structure.” 74 Those costs will be stranded because market rates are likely to be set and driven by new suppliers who are unlikely to have historical generation costs. The new suppliers therefore can sell electricity at a lower price than most incumbent utilities. When the 1999 amendments to the PURA were enacted, it was projected that market prices would not, in all probability, generate enough revenue for existing utilities to cover their historical, embedded costs.

Regulatory assets were created and stranded costs were incurred as part of prudent planning not only for customers the utility was then serving but for anticipated future customers as well. It is not unjust or unreasonable, and therefore it is not confiscatory, to charge rates to present and future retail consumers in a utility’s service area that will allow a utility to recoup regulatory assets and stranded costs associated with these outlays of capital. This is so even if the consumers do *248not buy power from an incumbent utility such as CPL but obtain service from a new provider in CPL’s service area. Were it not for the historical expenditures made by incumbent utilities over the years to construct generating capacity, the extensive electric power grid that will now allow new suppliers to enter the market would not exist in its present form. There would be far fewer transmission and distribution lines in existence today if fewer generating facilities had been built. Competition, much less statewide service, would not be as economically feasible without generation facilities and the infrastructure that was connected to them.

Although it is not critical to our takings analysis, we note that retail customers who otherwise would have been served by an existing utility but who will choose another provider in the new era of competition may have their electricity generated by facilities built by the existing utility in their service area. Under the PURA, an existing utility must sell at auction at least fifteen percent of its generation capacity.75 That obligation continues until the earlier of five years from the date customer choice is introduced or the date the Commission determines that forty percent or more of the electric power that was consumed by residential and small commercial customers in an incumbent utility’s service area before customer choice began is provided by someone other than the incumbent utility or an affiliate.76 Accordingly, consumers in an incumbent utility’s service area or their new suppliers will have the opportunity to purchase at least fifteen percent of the power generated by the incumbent utility’s facilities, including nuclear plants that account for a large portion of stranded costs.

In the interim between September 1, 1999 and January 1, 2002, an electric utility’s rates are frozen at the September 1, 1999 level, with certain exceptions.77 Existing customers and new customers entering an incumbent utility’s service area are thus assured of rate certainty during this interim.

Residential and small commercial customers will also have the right to continue to purchase electricity from an incumbent utility or its affiliated retail electric provider at a frozen “price to beat” for a period of time up to January 1, 2007.78 There are detailed requirements in the PURA of how the “price to beat” is determined, but it is in essence six percent less than the rates in effect on January 1, 1999.79 Accordingly, new residential and small commercial customers will have the choice of purchasing power from an incumbent utility at frozen rates or purchasing power from another provider.

The foregoing scheme indicates that the Legislature intended that the overall structure and impact of the securitization provisions and the move to partial deregulation would benefit most consumers. There are tangible services that are likely to be provided to many existing customers as well as many new consumers who enter an incumbent utility’s service area after its regulatory assets were created and its stranded costs were incurred, particularly if those consumers would have been served by the incumbent utility had the former scheme of regulation remained in place. The Legislature’s scheme does not result in an unconstitutional taking as to these *249new or existing consumers. It is within the Legislature’s province to decide that all segments of the power industry, including new consumers, should bear the costs of partial deregulation through rates.

We do not decide whether the securitization provisions of the PURA would result in an unconstitutional taking as applied to a new consumer if that consumer could demonstrate that its electric service would not have been provided by an incumbent utility if regulation had been continued and that no part of the electric service that the consumer has or will receive is generated by an incumbent, its affiliate or a purchaser of its assets, or is transmitted or distributed over or through facilities that were owned or constructed by an incumbent utility or its affiliate. A challenge mounted on that basis would be a particularized “as applied” challenge that Power Choice does not make in this case. For the reasons considered above, we conclude that the district court did not err in rejecting Power Choice’s constitutional challenges to transition charges.

IV

Power Choice contends that the nonbypassable transition charges contemplated by the PURA are a tax, not a rate or a regulatory fee. With limited exceptions, transition charges will be paid by all electric power consumers in CPL’s geographic service area regardless of whether those consumers buy electricity from CPL or some other supplier.80 Power Choice asserts that the PURA will allow Texas utilities to pass through to retail consumers approximately $8.4 billion for generation capacity that is of no benefit to those consumers. The effect of these nonby-passable charges, Power Choice argues, is to shift risk and cost from a utility’s shareholders to consumers. This, Power Choice says, is an exercise of sovereign power that is not for a public purpose and therefore amounts to a tax in violation of Article VIII, Section 3 of the Texas Constitution.81 Alternatively, Power Choice contends that if transition charges are for a public purpose, then they should be collected from all the people of Texas as a general tax, not levied as a utility rate. Power Choice’s arguments are unpersuasive.

A

In determining whether transition charges are a tax rather than a utility rate, it is important to consider who is to pay those charges under the PURA’s regulatory scheme and what the charges represent. When the Legislature implemented the securitization provisions of the PURA, it made a conscious decision about who is to bear certain costs associated with the transition from regulated electric retail rates to market-based retail rates. The Legislature determined that with limited exceptions, certain costs of the transition should be borne by all consumers of electricity in an incumbent utility’s service area rather than by that utility’s shareholders. The Legislature’s decision was consistent with how those same costs would have been allocated if the former regulatory scheme had been left in place. As we discussed above, incumbent utilities had an obligation to prudently plan to serve future consumers. Consumers of electricity for the most part would have purchased power from an incumbent utility and would have paid through rates, not taxes, costs that the incumbent utility incurred in the past *250in preparation to render service in the future. It is only because of partial deregulation that other suppliers, such as Power Choice, will be able to compete with incumbent utilities, such as CPL, and consumers will have a choice of suppliers.

Many of the costs that are to be securi-tized and recovered through transition charges are costs that the Commission determined in prior rate cases should be recovered directly by a utility through its rates because those costs were prudently incurred in connection with providing electric service.82 If the PURA had not been amended to restructure the electric industry, CPL would have had the opportunity to recover stranded costs and regulatory assets through its rates. Existing and future customers served by CPL would have paid these costs over time, even though the actual expenditures or investments were made by CPL a number of years ago. The fact that rate recovery of these same costs will now be through transition charges does not convert the nature of these charges from utility rates to taxes. As we explained in more detail above, transition charges under the PURA are essentially a conversion from one form of recovering the same costs in rates to another.

B

In arguing that transition charges cannot be lawful rates and therefore must be considered a tax, Power Choice asserts that transition charges are not “related to any service, product, or commodity” and thus do not satisfy the “used and useful” test. Power Choice also asserts that transition charges amount to retroactive rate-making in violation of the filed rate doctrine.

In support of its “used and useful” argument, Power Choice cites this Court’s decision in Cities for Fair Utility Rates v. Public Utility Commission.83 In that case, we quoted former PURA section 2.203(a), which used the phrase “used and useful,” and we said that under that code provision, rates must include a component that allows a reasonable return on invested capital that is used and useful in rendering service.84

Stranded costs include so-called bricks and mortar capital expenditures for generating facilities.85 Those facilities were used or held in the past to enable the utility to provide service. Power Choice complains only that these investments will not be “used” or “useful” with regard to future service. Even were that true, this and other courts have recognized that a regulatory authority may legitimately conclude that costs incurred in the past should be spread out in rates in the future in order to avoid sharp increases in rates or “rate shock.”86 With regard to regulatory assets, by definition, the Commission has previously found that these costs are recoverable through rates because they were prudently incurred and used and useful in connection with the generation of power.87

*251The Supreme Court of New Hampshire confronted the contention that recovery of stranded costs was unconstitutional because it violated the used and useful principle of ratemaking.88 That court held that even if it were to agree that stranded cost recovery charges were associated with property that was no longer used and useful, that “principle is not constitutionally mandated.”89 Similarly, in Rural Telephone Coalition v. Federal Communications Commission,90 the United States Court of Appeals for the District of Columbia Circuit rejected an argument indistinguishable from Power Choice’s “used and useful” argument. In that case, the FCC allowed a gradual phase-out of terminal equipment costs in telephone carriers’ accounts even after those carriers no longer furnished terminal equipment because of deregulation. The reviewing court affirmed this rate treatment, holding that the FCC had the authority to conclude that immediate removal of embedded costs from rates would be unacceptably disruptive and that the doctrine of “used and useful” was one of “limited weight” and had “ ‘ceased to have any constitutional significance.’ ”91

Power Choice relies on our decision in State v. Public Utility Commission92 and the United States Supreme Court’s decision in Arkansas Louisiana Gas Co. v. Hall,93 contending that transition charges violate the filed rate doctrine. Power Choice asserts that under that doctrine, rates can only have prospective effect and that rates cannot allow a utility to recoup past losses. We explained in State v. Public Utility Commission that the “rule against retroactive ratemaking is often misunderstood and misapplied.”94 We held that utilities could defer costs associated with construction and start-up of new power generation facilities and include those costs in future rates. In that case the Commission’s orders setting rates did “not allow the utilities to recoup losses resulting from previously set rates which were insufficient.”95 Similarly, in enacting the 1999 amendments to the PURA , the Legislature did not conclude that previously set rates were insufficient. It determined that it was in the public interest to “allow utilities with uneconomic generation-related assets and purchased power contracts to recover the reasonable excess costs over market of those assets and purchased power contracts” during “the transition to and in the establishment of a fully competitive electric power industry.”96

Moreover, the filed rate doctrine does not prohibit a regulatory authority from finding that an existing rate is unreasonable and determining the just and reasonable rate to be charged thereafter.97 The filed rate doctrine prohibits a regulatory authority “from imposing a rate increase for [a commodity] already sold.”98 Transition charges do not increase rates for power already sold. They are part of *252the price for power that will be sold in the future. They allow the continued recoupment of costs that would have been recovered by utilities through rates under a prior regulatory scheme.

C

Power Choice cites this Court’s decision in Conlen Grain,99 arguing that because the Commission has assessed transition charges under a financing order, those charges purport to raise revenue for a public purpose and therefore constitute a tax. It does not matter, Power Choice says, that the money is not paid into state coffers. The state is using its power to order collection of transition charges.

In Conlen Grain, agricultural producers voted under the Texas Commodity Referendum Act100 to create a board that levied assessments to be used for developing research programs, disease control, education, and marketing. Under the Act, the board was expressly an agency of the state. This Court held that the assessment was an occupation tax that violated Article VIII, Section 1 of the Texas Constitution, which prohibited an occupation tax on an agricultural pursuit.101 During the course of our opinion in Conlen Grain, we said that one reason the assessments constituted a tax was because their primary purpose was to generate revenue to be used by an agency of the state and that the agency could use that revenue as it considered proper for public purposes:

[The assessments on producers] are levied periodically to provide a fairly constant source of revenue that is expended by an agency of the state as it considers proper for the support of programs calculated to increase the production and use of particular agricultural commodities. These programs doubtless promote the economic welfare of many who are engaged in producing the commodities, but the assessment paid by any particular person is not necessarily related to the benefits that will be received by that person through the Board’s expenditure of the money he paid. The levy is not a special assessment.102

We then said that because the power of the state was used to deprive commodity producers of money or at least the use of money until the assessment was refunded, and because the primary purpose of the assessment was to raise revenues, it was a tax.103

The transition charges imposed by the PURA are different from the assessments in Conlen Grain in at least one important respect. Transition charges are not used by an agency of the state. Transition charges are used by an electric utility to retire debt or equity associated with its stranded costs and regulatory assets related to the generation of electric power. Transition charges are analogous to an allocation of costs between intrastate and interstate telephone service providers that was held not to be a tax by the United States Court of Appeals for the District of Columbia Circuit in Rural Telephone Coalition.104 In that case, the FCC deter*253mined that interstate carriers should shoulder twenty-five percent of local phone exchanges’ “non-traffic sensitive” costs even though those costs did not increase as a result of increased interstate use.105 In holding that this allocation did not amount to a tax, the court observed that “a regulation is a tax only when its primary purpose judged in legal context is raising revenue,” and that “[tjhere is no reasonable way to construe the [non-traffic sensitive] cost allocation as having the primary purpose of raising federal revenue.”106

It is beyond dispute that in the prior, regulated environment, the Legislature had the authority to require retail electric customers to pay the costs represented by a utility’s regulatory assets through utility rates. That authority still exists even though the Legislature has decided to partially deregulate the electric power industry. The fact that the Legislature has chosen to continue to require consumers to pay the costs represented by a utility’s regulatory assets under a different mechanism than it did under the prior regulated regime does not transform what were undeniably utility rates into taxes. The secu-ritization provisions of the PURA do not constitute a tax and do not violate Article VIII, Section 3 of the Texas Constitution.

Power Choice’s alternative argument is that if transition charges are truly in the public interest, then the general public, not just energy consumers in particular incumbent utilities’ service areas, should pay those costs in the form of a tax. This would, of course, breathe life into another of Power Choice’s constitutional challenges, which is that tax dollars cannot be paid to private corporations because to do so would violate Article III, Section 51 of the Texas Constitution. (We consider that constitutional challenge in the next section.) But the Legislature has the unquestioned police power, as we have seen, to regulate utility rates.107 It acted well within that power when it decided that transition charges are to be collected as a utility rate for power consumed in the future rather than a general tax.

V

Power Choice contends that transition charges are a grant of public money to private corporations in violation of Article III, Section 51 of the Texas Constitution.108 We disagree. Transition charges are not state expenditures. They are an allocation of a utility’s costs between electric power consumers and a utility’s shareholders.

This Court held in State v. City of Austin that the purpose of Article III, Section 51 and of Article XVI, Section 6 of the Texas Constitution “is to prevent the application of public funds to private purposes; in other words, to prevent the gratuitous grant of such funds to any individual or corporation whatsoever.”109 In *254City of Austin, the Legislature enacted a statute that called for the state to pay for the relocation of utility lines due to highway improvements if the relocation was eligible for federal participation.110 This Court held that although a utility could be required to move facilities at its own expense, the Legislature had the authority to pay these costs. Power Choice relies on an observation in City of Austin that concludes that the state could not reimburse a utility for any expense it incurred in moving facilities prior to the enactment of the law allowing the state to pay for the relocation. But that reasoning is in-apposite. The PURA does not contemplate that any state or public funds will be used to pay transition charges. The securitization provisions of the PURA authorize rates that are to be paid by consumers for the power they consume.

For the same reason, this Court’s decision in Road District No. ⅛, Shelby County v. Allred is inapposite.111 In that case, a county road district issued bonds, the proceeds of which were to be used to construct roads. An agent absconded with $260,000 of the bonds, which he sold to the public for his own account. As a result, none of the $260,000 was spent by the road district. Bondholders obtained judgments against the road district when it did not pay in accordance with the terms of the bonds and the Legislature passed a statute directing that state funds were to be paid to the road district to cover these losses. This Court held that the road district was a corporation within the meaning of Article III, Section 51, and that the payment of funds to it was not for a public purpose since the funds were to be used only to reimburse the road district for its losses, not to build roads.112 Here, no public funds are being expended to pay transition charges. They are to be paid by electric power consumers.

For the reasons considered above, we conclude that Power Choice’s appeal is without merit. We now turn to the issues raised in the appeal by numerous cities served by CPL, the Office of Public Utility Counsel, Texas Industrial Energy Consumers, and the Texas Retailers Association.

VI

The financing order at issue allows CPL to securitize regulatory assets that include what are known as “SFAS 109 assets.” “SFAS 109” refers to Statement of Financial Accounting Standard 109. An SFAS 109 regulatory asset is essentially a receivable from a utility’s customers for the future payment of federal income taxes. Of the total amount of $763,734,489 of regulatory assets that the financing order at issue in this case allows to be securi-tized, SFAS 109 assets account for $139,182,000.

The Cities contend that it was improper for the Commission to include SFAS 109 assets in the amount to be securitized because these regulatory assets earn no return and have no carrying costs. The Cities argue that the PURA requires that each regulatory asset be analyzed on a stand-alone basis to determine whether securitization will lower that asset’s carrying costs “relative to the costs that would be incurred using conventional utility financing methods.”113 Securitization of SFAS 109 assets does not benefit consumers, the Cities contend, because the carrying costs of those assets cannot be decreased since they are already at zero, and securitization may increase the carrying *255costs to as high as 8.75 percent, which is the highest interest rate authorized in the financing order for the transition bonds.

Resolution of the Cities’ challenge to this aspect of the financing order lies in the provisions of the PURA. Those provisions make it clear that, contrary to the Cities’ position, the Commission is not permitted to decide what types or categories of regulatory assets may be securitized. The PURA says that all regulatory assets are to be securitized on application of a utility, subject to the requirement that “the total amount of revenues to be collected under the financing order” meets the requirements of sections 39.301 and 39.303(a).114

The PURA provides that an electric utility may “securitize 100 percent of its regulatory assets as defined by section 39.302.”115 Section 39.302 defines “regulatory assets” with some specificity:

“Regulatory assets” means the generation-related portion of the Texas jurisdictional portion of the amount reported by the electric utility in its 1998 annual report on Securities and Exchange Commission Form 10-K as regulatory assets and liabilities, offset by the applicable portion of generation-related investment tax credits permitted under the Internal Revenue Code of 1986.116

Each utility’s “regulatory assets” were fixed and ascertainable from 1998 Form 10-K’s at the time that the securitization provisions were enacted in the 1999 amendments to the PURA.117 The amounts reported in each utility’s 1998 10-K are not subject to revision by the utility or the Commission, except under section 39.262(f).118 SFAS 109 assets are among the regulatory assets listed in CPL’s and other utilities’ 1998 10-K’s. The Legislature did not exclude SFAS 109 assets from its definition of “regulatory assets.”119

Sections 39.302(4), 39.302(7), and 39.303(a) of the PURA reiterate that “100 percent of an electric utility’s regulatory assets” are “qualified costs” that can be recovered through transition charges imposed under a financing order upon application by a utility, subject to the requirements of sections 39.301 and 39.303(a). The PURA does not give the Commission discretion to single out certain types of regulatory assets from “100 percent of an electric utility’s regulatory assets” and declare that those particular types of regulatory assets are not “qualified costs.”120 Section 39.302(4) necessarily forecloses the Commission from making such a determination. The limitation on a utility’s right to recover all regulatory assets is determined by looking at the amount securitized on an aggregate basis. The Commission is expressly directed to look to “the total amount of revenues to be collected under the financing order” to determine whether that amount is “less than the revenue requirement that would be recovered over the remaining life of the stranded costs using conventional financing methods” and whether “the financing order is consistent with the standards in Section 39.301.” 121

*256The standards under section 39.301 include a present value test. The parties in this case, as well as in TXU Electric Co v. Public Utility Commission, 51 S.W.3d 275,122 which we also decide today, disagree on how the Commission is to calculate present value within the meaning of section 39.301.123 The Commission takes the position that because section 39.301 says that it must “ensure that securitization provides tangible and quantifiable benefits to ratepayers, greater than would have been achieved absent the issuance of transition bonds,” it must also apply another present value test, in addition to that required by the final two sentences of section 39.301. But it is undisputed, even by the Cities, that under both present value tests savings result to consumers when those tests are applied to the total amount of assets that the financing order allows CPL to securitize. The Commission found that present value benefits to ratepayers would be at least $90 million.

The Commission did not err in allowing CPL to securitize SFAS 109 assets.

VII

The Cities contend that the Commission erred by fading to offset the regulatory assets that CPL sought to securitize with all of CPL’s investment tax credits. The PURA’s definition of “regulatory assets” that may be securitized provides that regulatory assets are to be “offset by the applicable portion of generation-related investment tax credits permitted under the Internal Revenue Code of 1986.”124

The operative qualifier in section 39.302(5)’s offset provision is “applicable.” Section 39.302(5) does not require that CPL offset all its tax credits against its regulatory assets. It only requires an offset of applicable credits. Witnesses for CPL and the Commission testified that none of the investment tax credits that the Cities identified were applicable to the regulatory assets in question. Moreover, the Cities’ expert witness conceded that if the offset the Cities sought were required by the Commission, a normalization violation might occur, which would obligate CPL to pay to the IRS all its current tax credits, not just those associated with its investments in transmission and distribution facilities.

The financing order’s treatment of investment tax credits “conforms to the [PURA]” and is “within the authority of the commission” under chapter 39 of the PURA.125 The trial court did not err in its disposition of this issue.

VIII

The Cities assert that because the Commission used CPL’s 1998 SEC Form 10-K to identify the amount of certain regulatory assets to be securitized, rather than the balance of those assets as of December 31, 2001, CPL will over-recover. This occurs, the Cities contend, because CPL’s regulatory assets are in its existing rate base, and its rates provide for recovery of and a return on those assets. The actual balance of the regulatory assets will decline, the Cities contend, from the balance shown in CPL’s 1998 10-K because rate recovery will reduce that balance *257from December 31,1998 until the date that transition bonds are issued.

Once again, the express provisions of the PURA are dispositive. Section 39.302(5) defines “regulatory assets” as the amount reported in CPL’s 1998 SEC Form 10 K.126 The PURA provides that all “regulatory assets” are to be securitized on application of a utility, subject to the requirement that “the total amount of revenues to be collected under the financing order” meets the requirements of sections 39.301 and 39.303(a).127The PURA does not provide for an adjustment to the amount shown in the 1998 SEC Form 10-K to reflect subsequent rate recovery.

We note that the Commission and CPL contend that certain sections of the financing order and provisions of the PURA should mitigate or eliminate any over-recovery by CPL. They point to the final true-up provision of the financing order and sections 39.254 and 39.257 through 39.261 of the PURA. We need not decide whether any of those provisions will actually eliminate or ameliorate any over-recovery. That question may be raised in future proceedings. The financing order at issue in this case complied with section 39.302(5) in using the 1998 Form 10-K amount of regulatory assets. The Cities do not attack the validity of section 39.302(5). They say only that its express directive should not have been applied. Because the PURA does not authorize the Commission to use the balance as of December 31, 2001, in determining the amount of regulatory assets that may be securitized, our inquiry in this case is at an end.

IX

Another point of dispute is how section 39.253 of the PURA, which allocates transition charges among classes of customers, should be interpreted.128 CPL has eight classes of customers. They include the residential class, commercial classes, firm industrial customers, non-firm industrial customers, and others.

The allocation of stranded costs, including regulatory assets, has two basic components. One is determined by applying the same methodology used to allocate costs of the underlying assets in the electric utility’s most recent commission order addressing rate design.129 The other is the energy consumption of the respective classes130 “based on the relevant class characteristics as of May 1, 1999, adjusted for normal weather conditions.”131

CPL’s last rate design case was in 1997. In allocating transition charges among classes of CPL’s customers, the Commission applied the methodology used in that rate case to the power consumption data developed in that same case. This resulted in allocation factors, expressed as percentages, for each class of customers. The Commission used those allocation factors to allocate transition costs.

Texas Industrial Energy Consumers is a voluntary association of companies that operates small and large industrial facilities in CPL’s service area. It contends that in initially determining the allocation factor for each customer class, the Commission should have adjusted the historical data that reflected usage by the industrial classes to account for industrial customers that have since left or will leave CPL’s *258system to obtain service from exempt facilities.

To the extent a customer’s load is served by a qualifying cogeneration facility before a certain date, or by an on-site power facility that has a capacity of ten megawatts or less, that customer is exempt from paying transition charges under section 39.262(k).132 It is undisputed that a number of CPL’s industrial customers have switched to exempt power sources and thus will not be required to pay transition charges. TIEC requests this Court to remand the financing order with instructions that the allocation factors be recalculated and then adjusted at the end of each year during the life of the transition bonds to remove the load of industrial customers who switch to other service from exempt facilities and therefore become exempt from paying transition charges. We conclude that the PURA does not provide for the method advocated by TIEC to establish allocation factors.

TIEC first asserts that this is required by section 39.253(f), which says that “[e]xcept as provided by Section 39.262(k), no customer or customer class may avoid the obligation to pay the amount of stranded costs allocated to that customer class.” 133 TIEC argues that when a customer becomes exempt from paying transition charges, that customer’s class is also exempted from paying what would have been that customer’s share of transition charges. We disagree. The exemptions in section 39.262(k) apply to specific customers. Section 39.262(k) does not give a corresponding exemption to a class when one of its customers obtains power from a qualifying facility. Section 39.253(i) reaffirms that all customer classes must pay their share of transition charges but recognizes that an entire class of customers may switch to exempt power facilities.

Moreover, to give effect to TIEC’s position, we would have to construe section 39.253 as directing the Commission to establish allocation factors based on data obtained at the end of each year over the life of the transition bonds rather than the data used in the utility’s most recent rate design case. As noted above, section 39.253 says that one component of how transition costs are allocated among all other customer classes is “the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design.” 134 We conclude, as we similarly conclude today in TXU ,135 that this phrase *259is not entirely clear.136 It could be construed to permit the Commission to apply the methodology used in a utility’s last rate case to data in that rate case or to more current data to arrive at allocation factors for each class.137 However, the Commission has construed section 39.253 to mean that the methodology is to be applied to the data in the most recent rate case, which is historical data. That interpretation is reasonable and does not contradict any plain language in section 39.253. It is also consistent with the Legislature’s directive that historical data is to be used in applying the other component for allocating transition charges. That component is the energy consumption of the customer classes as of May 1, 1999, adjusted for normal weather conditions.138 We accordingly accept the Commission’s construction of section 39.253 in this regard.139

In light of the Commission’s and our interpretation of section 39.253(c)-(i), the Commission did not err in declining to use customer load data other than the data in CPL’s last rate case and the energy consumption data as of May 1, 1999, adjusted for normal weather conditions, in determining how transition charges are to be allocated among CPL’s customer classes.

X

Another disagreement is how section 39.253 allocates transition costs to non-firm industrial customers. A utility may interrupt service to non-firm customers for specified reasons, typically during periods of high demand from other customers on that utility’s system. Nucor Steel, one of CPL’s non-firm industrial customers, and TIEC insist that the literal terms of the PURA require that after the amount to be allocated to residential customers has been determined, section 39.253 directs that that amount is to be subtracted from the total amount of transition charges before an allocation is made to non-firm industrial customers. The Commission did not make that subtraction. It based the percentage of transition costs to be borne by non-firm industrial customers on the entire amount of the transition charges.

Section 39.253 is complex and, we conclude, unclear in this regard. It says:

(c) The allocation to the residential class shall be determined by allocating to all customer classes 50 percent of the stranded costs in accordance with the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design and allocating the remainder of the stranded costs on the basis of the energy consumption of the classes.

(d) After the allocation to the residential class required by Subsection (c) has been calculated, the remaining stranded costs shall be allocated to the remaining customer classes in accordance with the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design. Non-firm industrial customers shall be allocated stranded costs equal to 150 percent of the amount allocated to that class.

(e) After the allocation to the residential class required by Subsection (c) and the allocation to the nonfirm industrial class required by Subsection (d) have been calculated, the remaining stranded *260costs shall be allocated to the remaining customer classes in accordance with the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design.140

There is tension between paragraph (c) on the one hand, and paragraphs (d) and (e) on the other, partly because of the use of the word “remaining” in the latter two paragraphs. There is also ambiguity because paragraph (c) seems to contemplate an allocation of all costs to all classes based fifty percent on demand allocation factors and fifty percent on energy consumption. . Paragraph (c) says that fifty percent of stranded costs (which also means transition charges for regulatory assets141) are allocated to all customer classes in accordance with the methodology used in the utility’s most recent rate case, and that the remaining fifty percent is allocated on the basis of the energy consumption of the classes. Then, paragraph (d) says that after the residential allocation is made under (c), “the remaining” stranded costs are allocated to the customer classes based on the methodology used in the utility’s last rate case. Paragraph (d) similarly says that after the residential and nonfirm industrial allocations are made, “the remaining” costs are allocated among the other classes (i.e. firm non-residential classes) based on the methodology used in the utility’s last rate case.

TIEC offers an example to demonstrate the difference between its application of section 39.253 and the Commission’s. In that example, the transition charges total $100. Residential customers are allocated forty percent of those costs. The example further assumes that based on the last rate case, the non-firm industrial customers’ demand allocation factor is ten percent and the demand allocation factor for all other classes totals fifty percent. Because paragraph (d) requires non-firm industrial customers to bear 150 percent of the amount allocated to that class, the demand allocation factor for that class becomes fifteen percent. Both TIEC’s and the Commission’s application of section 39.253 to this example allocates $40 of the $100 of transition charges to the residential class. TIEC and the Commission diverge on what happens next.

TIEC would subtract the residential class’s $40 share of costs from the $100 total, and then apply the fifteen percent demand allocation for non-firm industrial customers to the remaining $60 of costs. That results in an allocation of $9 to that class. TIEC would then spread the remaining costs of $51 among the firm nonresidential classes based on their proportionate demand allocation factors, rather than applying the demand allocation factors which total fifty percent for those classes. TIEC refers to this latter step as “grossing up” the allocation factors for the firm non-residential classes, which is not in conformity with the literal terms of paragraph (e). TIEC argues that “grossing up” the firm non-residential class’s demand allocation factor is nonetheless necessary under its interpretation of section 39.253 because otherwise, there would be an undercollection of transition charges.

Using the Commission’s application of section 39.253, the fifteen percent demand allocation factor for the non-firm industrial customers is applied to the total $100 in transition costs. The non-firm industrial customers would pay $15, rather than $9 in TIEC’s example. The Commission’s application of section 39.253 then allocates the remaining $45 of costs among the firm non-residential classes. As Nucor Steel *261points out in its amicus brief, the Commission’s methodology would result in an ov-ercollection of $5 if the fifty percent demand allocation factor for those classes were applied to the $100 total rather than simply spreading “the remaining” $45 of stranded costs proportionately among the firm non-residential classes. The Commission counters that it is applying the words “the remaining stranded costs” used in paragraph (e) in its application of section 39.253.

TIEC’s construction of section 39.253 results in an undercollection of transition charges. The Commission’s construction, if carried to its logical conclusion, would result in an overcollection. No one suggests a construction that allows recovery of 100 percent of the transition charges, but no more. As we explained above and as we explain today in TXU,142 when faced with an ambiguous code provision, we give some deference to the Commission’s interpretation when it is reasonable and does not conflict with the code’s clear language.143 Under the circumstances presented here, we cannot say that the Commission’s interpretation of section 39.253 is an unreasonable one or that it conflicts with that section’s plain language. Accordingly, we hold that the Commission, did not err in the manner in which it allocated transition charges to non-firm industrial customer classes.

XI

Finally, the Cities argue that the Commission denied them due process by: (1) not allowing adequate time to prepare for the contested case hearing on the financing order; (2) unduly restricting the Cities’ opportunity to cross-examine witnesses; and (3) issuing a financing order that adopts a non-unanimous stipulation without first holding a separate hearing on that stipulation or making findings that it is reasonable, supported by the evidence, and in compliance with the PURA. We conclude that the Cities were not denied due process in this case.

A

The PURA contemplates that adoption of a financing order will be accomplished on an accelerated procedural schedule. The Legislature has directed that a financing order shall be issued within ninety days after a utility files a request for a financing order.144 A financing order is not subject to rehearing by the Commission.145 The time for filing an appeal in a Travis County district court is fifteen days after the financing order is signed by the Commission.146 There is a direct appeal from the district court to this Court with no intermediate review by a court of ap*262peals.147

CPL filed its request for a financing order with the Commission on October 18, 1999. The Cities were among eighteen parties that intervened. Consistent with the PURA’s expedited procedures, the Commission established a briefing and hearing schedule. The Cities argue that this schedule allowed insufficient time for discovery. They were particularly aggrieved, they say, by the hearing schedule because they were afforded only “39 minutes of cross examination [of 36 witnesses] reduced by any time presenting an opening statement.” Accordingly, the Cities argue, their due process rights were violated in the first instance because they did not have a reasonable amount of time to prepare, citing Lowe v. City of Arlington,148 or to cross examine witnesses, citing Rector v. Texas Alcoholic Beverage Commission,149

This Court has held that in administrative proceedings, due process requires that parties be accorded a full and fair hearing on disputed fact issues.150 This requirement includes the right to cross-examine adverse witnesses and to present and rebut evidence.151 But due process does not require that administrative proceedings have the full procedural framework of a civil trial.152 “ ‘[D]ue process is not so rigid as to require that the significant interests in informality, flexibility, and economy must always be sacrificed.’ ”153

While we recognize that an administrative agency is entitled to considerable procedural flexibility, the Commission has failed to offer an adequate explanation of why, in a case of such public importance, it so severely restricted the time for cross-examination of witnesses. Even within the confines of the ninety-day statutory deadline for concluding proceedings when a request for a financing order has been filed, the Commission could have devoted more time in the schedule to the actual hearing. Furthermore, CPL filed an amended application after the evidentiary hearing had concluded that gave the Commission an additional seventy-one days beyond the original ninety days it would otherwise have had to issue an order. Yet the Commission did not allow additional hearing time. We do not condone the truncated hearing schedule that was established in this case. Nevertheless, based on the record before us, the Cities have failed to demonstrate any harm flowing from the Commission’s actions.

Before the hearing on the merits, the parties joined issue by submitting written discovery and pre-filing direct testimony. CPL filed its direct testimony with its request for a financing order, and all of the intervenors were able to respond to this testimony. At the hearing, witnesses were grouped into panels by topic, and counsel shared or traded time allotted to allied counsel for the examination and cross-examination of witnesses. After the hearing, all parties submitted an initial brief and a reply brief. With only one exception, the Cities do not identify any instance in which, the Commission’s procedural schedule pre*263vented them from making arguments or offering proof on a fact or policy issue material to the Commission’s decision. As to the one issue that the Cities did identify — whether the assumed interest rate of 8.5 percent on the transition bonds benefits consumers — the Cities in fact made their argument and offered proof. The Commission made findings as required by section 39.301 of the PURA. The merits of that issue were determined on an adequate record.

In view of the entire record in this case, we are persuaded that the Cities were accorded due process, particularly since the dispositive issues in these appeals are legal ones involving constitutional provisions and statutory construction. The Cities have failed to identify any material disputed fact issue had they been given more time to prepare that might have been resolved differentlyfor the hearing, present evidence, or cross-examine witnesses. In spite of the shortness of the hearing and the limited time for cross-examination, the Cities have not shown that substantial rights were violated by the procedures afforded.154

B

The Cities additionally contend that their due process rights were violated when the Commission ignored its own rule by adopting the non-unanimous stipulation without conducting an additional hearing on that stipulation. The Cities point to section 22.206 of the Commission’s rules.155 They urge that the failure to follow section 22.206 was arbitrary and capricious and requires that the financing order be vacated. The Cities cite Public Utility Commission v. Gulf States Utilities Co.,156 and Sam Houston Electric Cooperative, Inc. v. Public Utility Commission157 in support of their position.

We have recognized the importance of requiring an administrative agency to consider a non-unanimous stipulation on its merits.158 We have also said that due process requirements are satisfied “[i]f [the agency] makes an independent finding supported by ‘substantial evidence on the record as a whole’” and if the agency “provide[s] all parties, including non-signatories, the opportunity to be heard on the merits of the stipulation.”159 Indeed, the Commission’s own regulation provides that “[w]here some of the parties have reached a settlement on some or all of the issues, each party in the proceeding shall have the right to have a full hearing ... on issues that remain in dispute.”160

Although the Commission did not provide the non-settling parties an evidentiary hearing following the non-unanimous stipulation, we conclude that the Cities have failed to show that the Commission’s procedure denied them due process. The Commission adopted the financing order incorporating in part the non-unanimous stipulation only after extensive briefing, open meetings, and comments on the proposed order. The parties continued to *264comment after the stipulation was filed. The Commission urges that nothing in the financing order shows that it blindly adopted the stipulation, and we agree.

Moreover, we have previously held that failure to follow procedural requirements of statutes or rules is not reversible error without a showing of harm.161 The Cities have not demonstrated that they were harmed by the procedure followed by the Commission. The parameters of partial deregulation under the 1999 amendments to the PURA were the subject of much attention while under consideration by the Legislature, and the parties to this case were provided the opportunity to address the issues raised here within the time constraints imposed by the PURA. We are troubled by the appearance of haste in proceedings as significant as these, but we cannot say, on balance, that the Cities were denied constitutional due process.

joined by Chief Justice PHILLIPS, Justice ABBOTT, Justice HANKINSON, and Justice JEFFERSON, concurring.

We join fully in the Court’s judgment affirming the district court and in Justice Owen’s concurring opinion. This is the opinion of the Court regarding the validity of the “non-standard true-up” included in the Public Utility Commission’s financing order for Central Power and Light Company.

Under chapter 39 of the Public Utility Regulatory Act,1 the Public Utility Commission may issue a financing order authorizing an electric utility to use securitization financing by issuing transition bonds secured by or paid from transition charges.2 Transition charges are allocated among and collected from the retail electricity customers in the utility’s geographical certificated service area as it existed on May 1,1999.3 The allocation is by customer class (e.g., residential, commercial, industrial, etc.), and the rate per unit of service is affected by the energy consumption of the class.4 Because consumption varies over time, the unit rate must be adjusted periodically so that the total transition charge revenue is neither more nor less than the amount necessary to discharge the transition bond obligations and related financing costs. Section 39.307 refers to this adjustment as a “true-up” and provides:

A financing order shall include a mechanism requiring that transition charges be reviewed and adjusted at least annually, within 45 days of the anniversary date of the issuance of the transition bonds, to correct any overcollections or undercollections of the preceding 12 months and to ensure the expected recovery of amounts sufficient to timely provide all payments of debt service and other required amounts and charges in connection with the transition bonds.5

The Commission determined that the true-up for CPL should not only adjust the transition charge rate for each class based on changes in consumption within the class — what the Commission called a “standard true-up” — but should also adjust the allocation of transition charges among the classes if any class’s consumption is *265forecast to drop more than ten percent below its consumption for the year ending April 30,1999 — termed by the Commission a “non-standard true-up”. The Office of Public Utility Counsel and others contend that this non-standard true-up is not authorized by section 39.307, is contrary to PURA’s purposes, and is not supported by the evidence. Before we consider OPC’s arguments, we must explain the allocation of transition charges in more detail.

Section 39.303(c) states that “[transition charges shall be collected and allocated among customers in the same manner as competition transition charges under Section 39.201.”6 Section 39.201® states that “[a]ny competition transition charge shall be allocated among retail customer classes according to Section 39.253.”7 Section 39.253 prescribes the allocation of stranded costs, including regulatory assets, among a utility’s classes of customers. The parties agree that transition charges must also be allocated in the manner section 39.253 prescribes. Section 39.253 provides in pertinent part:

(c) The allocation to the residential class shall be determined by allocating to all customer classes 50 percent of the stranded costs in accordance with the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design and allocating the remainder of the stranded costs on the basis of the energy consumption of the classes.

(d) After the allocation to the residential class required by Subsection (c) has been calculated, the remaining stranded costs shall be allocated to the remaining customer classes in accordance with the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design. Non-firm industrial customers shall be allocated stranded costs equal to 150 percent of the amount allocated to that class.

(e) After the allocation to the residential class required by Subsection (c) and the allocation to the nonfirm industrial class required by Subsection (d) have been calculated, the remaining stranded costs shall be allocated to the remaining customer classes in accordance with the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design.

(f) Notwithstanding any other provision of this section, to the extent that the total retail stranded costs, including regulatory assets, of investor-owned utilities exceed $5 billion on a statewide basis, any stranded costs in excess of $5 billion shall be allocated among retail customer classes in accordance with the methodology used to allocate the costs of the underlying assets in the electric utility’s most recent commission order addressing rate design.

(g) The energy consumption of the customer classes used in Subsections (a)(2) and (c) shall be based on the relevant class characteristics as of May 1, 1999, adjusted for normal weather conditions.

(h) For purposes of this section, “stranded costs” includes regulatory assets.

(i) Except as provided by Section 39.262(k), no customer or customer class may avoid the obligation to pay the amount of stranded costs allocated to that customer class.

*266Thus, transition charges are to be allocated among a utility’s classes of customers based in part on the utility’s most recent Commission order addressing rate design and in part on the classes’ characteristic, weather-adjusted energy consumption as of May 1, 1999. CPL has eight customer classes. The Commission calculated each class’s transition charge allocation factor as follows:

Residential 37.0664%

Commercial & Small Ind. — Energy 21.5756%

Commercial & Small Ind. — Demand 26.9570%

Large Industrial — Firm 4.4891%

Large Industrial — Non-firm 5.5190%

Standby — Firm 1.4227%

Standby — Non-firm 0.3844%

Municipal & Cotton Gin 2.5858%

Total 100.0000%

To derive each class’s transition charge rate per unit of service, CPL’s financing order requires the transition bonds servi-cer to determine the total revenue needed to meet obligations for the upcoming year, multiply that amount by the percentage factors above to get each class’s dollar allocation, and then divide that allocation by the class’s usage forecast for the year in terms of billing units (e.g., kilowatt-hours for residential customers and kilowatts for demand customers), to arrive at a per-unit transition charge rate to bill the customer. Thus, usage and rate are inversely related. If usage increases, the unit rate decreases, and vice versa. If actual usage during the year varies from the forecast so that a class is charged more or less than its allocation, the class rate for the following year is adjusted up or down to compensate for the overpayment or underpayment the prior year. This adjustment is the standard true-up, and it is not challenged in this case.

Transition charges are nonbypassable; that is, they must be paid by every consumer of electricity in the utility’s service area whether the consumer buys electricity from the utility or not.8 Two exceptions are for customers using co-generation facilities operational by September 1, 2001, and customers using on-site generators with a rated capacity not more than ten megawatts.9 Typically, these are industrial customers. If such customers avoid transition charges by one of these exceptions, or simply by relocating outside the utility’s service area, the decrease in consumption in their class will result in an increase in the class rate, thereby prompting other customers in the class to look for a way out. Each departure of a customer from the class increases the transition charge burden on those remaining, encouraging further departures. In the Commission’s words, this “cascading load loss” could reach a “death spiral” so that transition charge rates become prohibitive, driving all customers from the class. If the class’s transition charge allocation is not shifted to the remaining classes or funded by some other means, the total transition charge revenue will be insufficient to meet transition bond obligations. No one disputes that the possibility of such an eventuality is very real and would adversely affect transition bond ratings by financial markets and impair their marketability. In this case, the Commission determined that a non-standard true-up should be used to reallocate transition charges among CPL’s customer classes whenever a class’s annual consumption is forecast to be less than ninety percent of its consumption for the year ending April 30, 1999.

It is undisputed that consumption by CPL’s industrial customers has already dropped more than ten percent since April 30, 1999. Texas Industrial Energy Consumers, a party to this case, states that *267CPL has fewer than twenty-five customers in each of its two industrial classes, so that a loss of only a few customers could significantly increase the transition charge rate paid by the others. By contrast, residential consumption in CPL’s service area is increasing. A witness for the Commission staff testified that most other states had used a procedure like the non-standard true-up to avoid cascading load loss in a class of customers. Analyzing the revenues projected to be needed to service CPL’s transition bonds, he concluded that “if any class experiences a decrease in [usage] in excess of 4%-7% (assuming no over- or under-collection) that class will see a higher transition charge in the prospective period than in the previous period, and thus may be at risk for a cascading load loss scenario.” He recommended that a non-standard true-up be used and that it be triggered when a class’s projected consumption decreased more than ten percent below consumption for the year ended April 30, 1999. A witness for CPL testified in favor of a more fluid reallocation each year, but after prompting by the Commission for the parties to resolve as many differences as possible, CPL agreed to the non-standard true-up proposed by the Commission staff.

To illustrate the operation of the nonstandard true-up, OPC offers the following example. Suppose a utility must allocate $100 million in annual transition charges among four classes of customers. Assuming allocation percentages and class usage, the rates for each class are calculated as follows:

_Residential Commercial Industrial Other

(1) Allocation of total annual transí- 40% 20% 10% 30%

tion charges (assumed)_

(2) Annual dollar allocation $40 MM $20 MM $10 MM $30 MM

(multiply $100 MM by line 1)_

(3) Annual unit usage (assumed)_700 MM_260 MM_23 MM 500 MM

(4) Rate per unit (divide line 2 by 5.7c 7.7c 43.5<t 6.0c line 3)

If forecast industrial usage drops to 16 MM units, and consumption by all the other classes remains the same, industrial customers will provide only $7 million in transition charges at a 43.5c rate, leaving a $3 million deficit. If that $3 million is reallocated to the four classes using the non-standard true-up prescribed in CPL’s financing order, the rates for each class would be calculated as follows:

_Residential Commercial Industrial Other

(1) Allocation of total annual transí- 40% 20% 10% 30%

tion charges (assumed)_

(2) Dollar allocation of deficit $1.2 MM $600 K $300 K $900 K

(multiply $3 MM by line 1)_

(3) Dollar allocation of balance_$40 MM_$20 MM_$7 MM $30 MM

(4) Total after reallocation_$41.2 MM $20.6 MM $7.3 MM $30.9 MM

(5) Annual unit usage_700 MM_260 MM_16 MM 500 MM

(6) Rate per unit (divide line 4 by 5.9c 7.9c 45.6c 6.2c line 5)

If industrial class customers had borne the entire rate increase due to their reduced consumption, their rate would have increased to 62.5c and the other three rates would have stayed the same. One will notice in this illustration, although OPC *268does not point it out, that the non-standard true-up increases the rate for industrial customers by 4.8%, while increasing the rates for the other three classes only about 2.5% to 3.5%. Without the non-standard true-up, rates for these three classes would remain the same while the industrial class rate increased 43.7%. Moreover, as the Commission points out, there is no basis for an assumption that consumption in the non-industrial classes will not increase. An increase in their total consumption by slightly over three percent would make up the $3 million deficit in annual transition charges caused by the reduction in industrial class consumption, leaving rates unchanged after a non-standard true-up. A greater increase in consumption in those classes would result in rate reductions, although those reductions would be larger without a non-standard true-up.

With this understanding of the non-standard true-up, we return to OPC’s arguments. First, OPC argues that however reasonable and beneficial the non-standard true-up might appear, it simply is not authorized by the PURA. Section 39.253 requires a fixed allocation based on historical data and does not contemplate realloca-tions among classes based on future changes in consumption of electricity. To read section 39.307 to allow for adjustments in section 39.253 allocations among classes, OPC contends, would violate the clear provision of section 39.253(f) that with the exceptions we have already noted, “no customer or customer class may avoid the obligation to pay the amount of stranded costs allocated to that customer class.” Rather, OPC argues, section 39.307 permits only intra-class rate adjustments required by inaccurate forecasts of usage and does not allow for adjustments in the allocations made under section 39.253. OPC points to the language in section 39.307 that true-up adjustments are “to correct any overcollections or undercollec-tions of the preceding 12 months”.

We agree with the Commission and CPL, however, that OPC’s argument ignores broader language in section 39.307 requiring adjustments “to ensure the expected recovery of amounts sufficient to timely provide all payments of debt service and other required amounts and charges in connection with the transition bonds.” A significant shrinkage of CPL’s industrial classes would unquestionably threaten the collection of “amounts sufficient to timely provide all payments of debt service”. OPC does not dispute this fact, nor does it argue that we should simply invalidate the non-standard true-up procedure in the financing order and leave no means to address the problem. Rather, OPC urges us to remand the case to the Commission to find a different solution. But the requirement of section 39.307 that adjustments be made to protect the discharge of transition bond obligations does not limit the means available to the Commission. The nonstandard true-up does not permit classes to avoid allocated charges, as prohibited by section 39.253(i); the procedure merely allows minimal adjustments to ensure the payment of transition bonds — which benefits all of the classes.

OPC argues that reading section 39.307 to allow a non-standard true-up makes section 39.253 a dead letter, allowing the Commission to make whatever allocations it may choose. We disagree. Section 39.307 does not convey such broad discretion, and the Commission makes no claim for such authority. Rather, as the Commission itself recognized in CPL’s financing order, transition charge allocations among customer classes must start with section 39.253, and that allocation must remain undisturbed unless and until its structure threatens the recovery of sufficient revenue to pay the transition bonds *269and other costs. Any load loss for a class will not trigger a non-standard true-up; the loss must be at least ten percent of the consumption in the year ending April 30, 1999. Section 39.307 allows only minor and essential adjustments in class allocations.

OPC argues that the non-standard true-up will raise residential rates, making it more difficult for non-incumbent retailers to compete and thereby defeating the purpose of chapter 39 deregulation. But as its own illustration shows, the impact of a non-standard true-up on residential rates may well be minuscule. Somewhat inconsistently, OPC argues that a non-standard true-up is unnecessary because the threat of industrial load loss to CPL is not great. Of course, if that turns out to be correct, the non-standard true-up may not be much used or have much effect. In any event, we are not persuaded that the Commission’s order jeopardizes retail competition in residential electricity markets simply because it adopts a procedure used successfully in other states to protect transition bonds.

OPC argues that even if the PURA authorized the Commission to adopt some type of non-standard true-up procedure, the evidence in this proceeding did not support the Commission’s decision in CPL’s financing order to trigger the nonstandard true-up procedure whenever forecast consumption in one class falls below ninety percent of consumption for the year ending April 30, 1999. OPC argues that the ten percent figure was arbitrarily chosen, but as we have noted, the record contains testimony by a Commission staff witness that a class with a decrease in usage of four to seven percent “may be at risk for a cascading load loss scenario.” We have not been cited to any evidence that the ten percent figure was too low. OPC’s argument that the record does not support measuring the ten percent reduction against the year ending April 30,1999, instead of the year immediately preceding the true-up, has more force. OPC contends that the Commission did not fully consider that because CPL’s industrial load loss since April 30, 1999, already exceeds ten percent, the non-standard true-up will be triggered to determine the first rates, and if industrial usage does not increase, may be triggered every year afterward, thereby becoming the standard true-up. As explained more fully in Part XI of, JustiCe Owen’s concurring opinion, we are troubled that the Commission’s haste in this proceeding may have resulted in an incomplete consideration of the complex and critical issues presented. With respect to the ten percent trigger, however, OPC argued its position to the Commission fully, explaining in detail the possible effects of the trigger, just as it has here. While the evidence supporting the Commission’s decision is slight, we cannot say in this case that the decision was arbitrary.

Finally, OPC argues that the Commission did not fully consider alternatives to the non-standard true-up. With respect to one alternative advanced by OPC’s witness — a non-standard true-up among three “super-classes”, combining CPL’s eight customer classes — was fully presented and discussed in the record. The only other alternative that OPC has advanced here is that customers in a class likely to have decreased usage should be required to pay increased transition charges in advance to provide a fund for bond payments later. But if increased rates due to decreased usage will drive customers from a class, we fail to see why increased rates due to anticipated decreased usage will not have the same effect. We find no support for this alternative in the record. OPC has not suggested an alternative solution that *270is consistent with its construction of section 39.307.

Justice Owen’s dissenting opinion makes three additional arguments which we address briefly. First, she notes that there can be no default on transition bond payments until a class has been completely vacated, leaving no one to pay its share of the transition charges. If the Commission’s concern were really “to ensure the expected recovery of amounts sufficient to timely provide all payments of debt service and other required amounts and charges in connection with the transition bonds” as authorized by section 39.307, the Commission need not have provided for any adjustments to class allocations until one class was vacant. However, we do not agree that the Commission was required to do nothing until the predicted exodus from the industrial class was complete, thereby impairing, at present, the marketability of the transition bonds. Moreover, if the nonstandard true-up prevents the complete vacancy of the industrial class, the other classes will benefit because industrial customers will remain to pay a portion of the transition charges.

Second, Justice Owen argues that section 39.307 is a general provision and therefore cannot be used to “nullify” section 39.253. We disagree that the nonstandard true-up “nullifies” the allocations prescribed by section 39.253. At most, the adjustments are slight, and some adjustments are almost certainly unavoidable if bond obligations are to be met. In our view, Justice Owen would deprive section 39.307 of its express purpose of protecting the means of satisfying bond obligations.

Lastly, unlike OPC, who argues for a remand so that the Commission can devise an alternate solution, Justice Owen argues that the dilemma facing the Commission and parties in this case is simply inescapable without legislative solution. We recognize that we cannot alter a statute’s plain meaning merely to make it more workable, but as we have explained, we believe our reading of section 39.307 is faithful both to its language and to PURA’s purposes.

Accordingly, we conclude that the nonstandard true-up procedure in CPL’s financing order does not violate the PURA.

dissenting.

I agree with the Court’s disposition of the direct appeals in these cases in all respects but one, which is the Court’s conclusion that the Commission was authorized by section 39.307 of the PURA to adopt a non-standard trueup. Accordingly, I respectfully dissent in part from the Court’s judgment.

I recognize that a mechanism for dealing with what the parties describe as “cascading load loss” may be critical to the marketability of transition bonds and therefore to the viability of securitization financing. However, I am constrained to conclude that the PURA ties the Commission’s hands. Neither section 39.253 nor section 39.307 gives the Commission discretion to reallocate transition charges among customer classes in a manner that is different from the allocation required by section 39.253.

In somewhat simplified terms, the nonstandard trueup provides that if in a given year, the predicted load within a class is projected to decrease by more than ten percent of what the load for that class was for the twelve months ended April 30, 1999, then the amount of transition charges attributable to the projected lost load are reallocated among all customer classes. In short, customers in some classes pay more transition charges than the allocation mechanism prescribed in *271section 39.253 permits, while others pay less than that allocation mechanism would require. To see why that is so, the mechanics of section 39.253 must be understood.

Section 39.303(c) of the PURA says that transition charges “shall be collected and allocated among customers” in the manner prescribed by section 39.201.1 Section 39.201(j) in turn says that section 39.253 governs.2 The allocation of stranded costs under section 39.253, which expressly includes regulatory assets,3 has two basic components. One is determined by applying the same methodology used to allocate costs of the underlying assets in the electric utility’s most recent commission order addressing rate design.4 The other is the energy consumption of the respective classes5 “based on the relevant class characteristics as of May 1, 1999, adjusted for normal weather conditions.”6

CPL has eight classes of customers among which transition charges are allocated. They include the residential class, commercial classes, firm industrial customers, non-firm industrial customers, and others. Section 39.253 allocates transition charges to the residential class differently than to the other classes.7 Section 39.253 also requires that non-firm industrial customers are to be treated differently from other classes of customers.8 Non-firm industrial customers are allocated a larger share of transition charges. Their allocation is increased by fifty percent of what it would otherwise be applying allocation demand factors, so that their allocation is 150 percent.9

Although the Court correctly concludes that there is some ambiguity in section 39.253 about how allocations are to be made among classes of customers, it is nevertheless clear that section 39.253 requires a financing order to establish a fixed percentage for each class that determines how much of the transition charges are to be allocated to each class based on historical data. Section 39.253 does not allow subsequent adjustment of the allocation to take into account growth or, conversely, load loss within each class during the life of the transition bonds, much less projected load loss.

The Commission applied the methodology used in CPL’s last rate design case to the consumption data in that case to arrive at allocation factors, that is percentages, for each class of customers. The Court correctly says that the Commission could have chosen to apply the same rate design methodology to more recent, but nonetheless historical, data.10 Section 39.253 is not crystal clear in that regard. But the Commission has now construed section 39.253 to require application of the latest rate design methodology to the consumption data that was part of the same rate case in which the rate design methodology was established. The Court agrees with that construction.11 Application of the rate *272design methodology to historical consumption data results in a fixed percentage, also known as a demand allocation factor. That percentage is applied to the transition charges and results in an allocation to each class of a dollar amount of transition charges for which it is liable. There is no mechanism in section 39.253 for reallocating some or all of one class’s responsibility to another class. To the contrary, section 39.253(i) says that “no customer or customer class may avoid the obligation to pay the amount of stranded costs allocated to that customer class,” with certain exceptions not relevant here.12 When transition charges that would otherwise be allocated to a customer class under section 39.253 are reallocated to another class, the class whose transition charges are reallocated has “avoid[ed] the obligation to pay the amount of stranded costs allocated to that customer class.”13

This can be seen from the examples given by the Court. The Court assumes that $100 million in annual transition charges must be allocated among four classes of customers based on each class’s historical usage. Applying the allocation factors assumed by the Court, each class is assigned the obligation to pay a dollar amount:

_Residential Commercial Industrial Other

(1) Allocation of total annual transí- 40% 20% 10% 30% tion charges (assumed, based on historical usage)_

(2) Annual dollar allocation $40 MM $20 MM $10 MM $30 MM (multiply $100 MM by line 1)

In the Court’s example, a load loss of 16 MM units is forecasted in the industrial class, which is more than a ten percent decrease in that class’s load as of the year ended April 30, 1999. That load loss would trigger the non-standard trueup, and each class’s responsibility to pay its assigned dollar amount is reallocated. Under the Court’s example, $3 million in transition charges that would otherwise be borne by the industrial class is reallocated across all classes:

_Residential Commercial Industrial Other

(1) Allocation of total annual transí- 40% 20% 10% 30% tion charges (assumed, based on historical usage)_

(2) Dollar allocation of deficit $1.2 MM $600 K $300 K $900 K (multiply $3 MM by line 1)_

(3) Dollar allocation of balance_$40 MM $20 MM $7 MM $30 MM

(4) Total after reallocation_$41.2 MM $20,6 MM $7.3 MM $30.9 MM

(5) Resulting allocation factor after 41.2% 20.6% 7.3% 30.9% reallocation of annual transition charges

It is readily apparent that allocation factors for each class have in reality been changed. The residential class no longer bears the percentage of responsibility assigned to it by the methodology set forth in section 39.253. Its obligation is a higher percentage. Likewise, the industrial class is no longer required to shoulder 150 percent of the transition charges allocated to *273it. But its obligation is something less than section 39.253(d) requires. And more importantly, under the Court’s decision, there is nothing to prevent the Commission from more drastically altering the allocation factors in a non-standard trueup.

I agree with the Office of Public Utility Counsel and the Texas Retailers Association that the non-standard trueup contravenes requirements in section 39.253 that are not ambiguous. I agree with those parties that the Commission may not take actions that are in excess of or inconsistent with express statutory provisions. The Third Court of Appeals “restated the familiar principles” in Southwestern Bell Telephone Co. v. Public Utility Commission:

“[A]n agency can adopt only such rules as are authorized by and consistent with its statutory authority.” Railroad Comm’n v. Lone Star Gas Co., 844 S.W.2d 679, 685 (Tex. 1992) (quoting State Bd. of Ins. v. Deffebach, 631 S.W.2d 794, 798 (Tex.App. — Austin 1982, writ ref d n.r.e.)). In this connection, it is well settled that an agency rule may not impose additional burdens, conditions, or restrictions in excess of or inconsistent with the relevant statutory provisions.14

The Commission and those who side with it rely on section 39.307 for authority to adopt non-standard trueups. That provision of the PURA says that a financing order must include an adjustment mechanism “to correct any overcollections or un-dercollections of the preceding 12 months and to ensure the expected recovery of amounts sufficient ■ to timely provide all payments of debt service and other required amounts and charges in connection with the transition bonds.”15 But the only circumstance under which the standard trueup would result in insufficient collections to cover debt service and other costs in connection with transition bonds would be if all customers in a class were lost. Theoretically, at least, as long as one customer remained in a given class, that customer would be obligated to pay all transition charges allocated to that class.16 The non-standard trueup is not designed to remedy the default of a class of customers in paying its allocated share of transition charges. The non-standard trueup is designed to more equitably spread transition charges when projected load loss within a class is just ten percent.

The Commission knows how to design a provision that would protect transition bondholders from the complete loss of customers in a class. It did so in the financing order at issue in TXU Electric Co. v. Public Utility Commission.17 TXU’s financing order says: “Should any of the Regulatory Asset Recovery Classes cease to have any customers, the [allocation factors] will be adjusted proportionately such that the sum of the [allocation factors] equals 100.0000%.” The Commission nevertheless included a non-standard true-up provision in TXU’s financing order identical to the one in CPL’s financing order. This underscores that non-standard true-up provisions do not and are not designed to “ensure the expected recovery of amounts sufficient to timely provide all payments of debt service and other re*274quired amounts and charges in connection with the transition bonds,”18 which is the Court’s touchstone for sanctioning nonstandard trueups. A non-standard trueup is designed to try to forestall a death spiral, but it is not until the death spiral has occurred and no customers are left in a class that debt service is impaired. A non-standard trueup does not cure that impairment. Under a non-standard true-up, transition charges are still allocated to all classes, even if there are no customers in one or more of those classes to pay the charges.

My fundamental problem with the Court’s holding is that it reads section 39.253 out of the PURA without express language in section 39.307 that can be used to do that. The Court has construed section 39.307 to allow the Commission to allocate transition charges in any manner that it chooses, as long as the Commission deems that necessary to ensure payment of the transition bonds. To give but one example, the Commission could decide not to afford residential customers the protections that section 39.253 provides by choosing to allocate transition costs in a manner entirely different from the one set forth by the Legislature. I recognize that the Commission did not take such a drastic step in CPL’s financing order. But once the moorings of section 39.253 are cut by giving the Commission authority under section 39.307 to allocate transition costs in any manner that it deems necessary, section 39.253 becomes a dead letter.

The general directive that section 39.307 gives to the Commission to “ensure the expected recovery of amounts” to retire transition bonds cannot override the more specific directives in other sections of the PURA about how transition charges are to be allocated among classes of customers. This Court’s decision in State v. Jackson19 is instructive. In that case, a statute authorized the Game and Fish Commission to close certain waters from all forms of netting and seining, except for minnow seines, whenever the Game and Fish Commission deemed that was best for protection of fish life. In another statute, the Legislature expressly said that it was lawful to use nets of a certain size in Galveston and Trinity Bays. Thereafter, the Game and Fish Commission issued a proclamation prohibiting all seines or nets for fishing in Galveston and Trinity Bays. This Court held that when the Legislature acts with specificity, an administrative agency cannot nullify that action under a more general grant of regulatory authority:

When the Legislature acts with respect to a particular matter, the administrative agency may not so act with respect to the matter as to nullify the Legislature’s action even though the matter be within the agency’s general regulatory field.

There is little case law announcing the rule last stated, no doubt because it is self-evident.20

In the case before the Court today, section 39.307 is a general grant of regulatory authority. It cannot nullify the specific directive in section 39.253 about how transition charges are to be allocated among customer classes.

I recognize that the Commission found itself in a dilemma. Section 39.253 allocates transition charges in such a way that load loss leading to a death spiral is not unlikely. In Jackson, this Court recognized a similar dilemma: “The State puts its position in these words: ‘The need for administrative closing of the bays increases when and as the Legislature increases *275the area of legal netting.’ ”21 We nevertheless were required to conclude, “Let it be so; the problem is one for legislative, not judicial solution.”22 This Court recognized that we must give effect to the plain meaning of a statute, even if to do so may effectuate a plan that is impracticable:

“The problem of statutory construction is to ascertain the intent of the Legislature. When we abandon the plain meaning of words, statutory construction rests upon insecure and obscure foundations at best. It should perhaps be reiterated that Courts have no concern with the wisdom of legislative acts, but it is our plain duty to give effect to the stated purpose or plan of the Legislature, although to us it may seem ill advised or impracticable.”23

I would hold that the Commission did not have the statutory authority to include a non-standard trueup in CPL’s financing order. Accordingly, I respectfully dissent.