BEA ANN SMITH, Justice.
We withdraw our opinion issued April 5, 1995 and substitute the following in its place.
The Court on its own motion has considered this cause en banc. See Tex.RApp.P. 79(d), (e). El Paso Electric Company (“EPEC”), the City of El Paso, the Office of Public Utility Counsel (“OPUC”), the State of Texas, and the Public Utility Commission (“Commission”)1 appeal from a district-court judgment rendered in a suit for judicial review under the Public Utility Regulatory Act. Tex.Rev.Civ.StatAnn. art. 1446c (West Supp. 1995) (hereafter “PURA”). The district court’s judgment reverses and remands certain aspects of the Commission’s final order and affirms the remainder. We will affirm the district-court judgment in part and reverse and render in part. See Administrative Procedure Act, Tex.Gov’t Code Ann. § 2001.174 (West 1994) (hereafter “APA”).2
BACKGROUND
EPEC is an electric utility that provides retail electric service in Texas and New Mexico and wholesale service in Texas, New Mexico, California, and Mexico. In 1978 the Commission awarded EPEC a certificate of convenience and necessity to expand its power generation system by purchasing an interest in the Palo Verde Nuclear Generating Station. Palo Verde consists of three separate units; EPEC’s ownership interest3 entitles it to 200 megawatts of power from each unit, for a total of 600 megawatts.
This appeal arises from the Commission’s final order in Docket 9945 entered in response to EPEC’s request for a rate increase of $131 million. See Tex.Public Util.Comm’n, Application of El Paso Electric Co. for Auth. to Change Rates, Docket No. 9945, 18 Tex. P.U.C.Bull. 9 (Feb. 6, 1992). As part of its proposed rate increase, EPEC requested inclusion in rate base of the $300 million cost of constructing Unit 3. The Commission permanently disallowed $27.23 million of costs for imprudence resulting in construction delays. The Commission then found that the utility had imprudently failed to reduce its [854]*854massive investment in the Palo Verde project, yielding more capacity than its customers needed. Because the Commission concluded that this “excess capacity” in the form of Unit 3 is not currently “used and useful,” its costs were excluded from rate base at this time. See PURA § 41(a). However, rather than permanently disallow such a huge expenditure because of EPEC’s imprudent level of participation in the project, the Commission decided that inclusion of Unit 3 in rate base be postponed only until the capacity becomes “used and useful.” Anticipating that Unit 3’s capacity would become useful in increments, the Commission proposed to include increasing percentages of Unit 3 costs in rate base as the capacity becomes needed to serve customers. EPEC challenges the Commission’s findings that EPEC imprudently failed to reduce its level of participation in the Palo Verde project, argues that Unit 3 does not represent excess capacity, and complains of the Commission’s postponement of its recovery of Unit 3 construction costs. The other appellants raise additional points of error asserting the res judicata effect of decisional imprudence holdings in prior dockets, and challenging the treatment of income tax expenses, deferred accounting charges, the tariff classification of state agencies, and the award of attorney’s fees.
DISCUSSION
Finality of the Commission’s Order
Before proceeding to EPEC’s points of error, we must address an argument raised by the State challenging the finality of the Commission’s order. The State argues that the Commission’s order was not a final, appealable order because the Commission continued to approve a rate moderation plan that fails to assign revenue responsibility for deferred revenues.4 Cf. City of Corpus Christi v. Public Util. Comm’n, 572 S.W.2d 290, 299-300 (Tex.1978) (only final agency orders are appealable); State v. Public Util. Comm’n, 840 S.W.2d 650, 654 (Tex.App.—Austin 1992), rev’d in part on other grounds, 883 S.W.2d 190 (Tex.1994) (final order leaves nothing open for future disposition). The State contends that the Commission abdicated its responsibility to ensure just and reasonable rates under PURA because it failed to identify the class of consumers that will pay for these costs in the future.
As we have noted in the analogous situation of deferred accounting treatment for capital costs, “the presence of a condition in an order does not automatically destroy its finality.” State v. Public Util. Comm’n, 840 S.W.2d at 654-55. Courts will treat as final an order that is definitive, promulgated in a formal manner, and issued with an expectation of compliance. Office of Pub. Util. Counsel v. Public Util. Comm’n, 843 S.W.2d 718, 723 (Tex.App.—Austin 1992, writ denied) (citing Texas-New Mexico Power Co. v. Texas Indus. Energy Consumers, 806 S.W.2d 230, 232 (Tex.1991)). For example, the Commission may issue an order determining the prudence of a utility’s investment before including the costs in rate base; the order is final because it entitles the utility to recover the costs, even though ratepayers remain unaffected by the order until rates incorporating these costs are set. See id. at 723-24 (order need only establish legal relationship between parties to be final and need not resolve all potential related controversies). Similarly, the Commission’s order did not need to assign responsibility for revenue increases beyond those incorporated in the current rate design to be final. The State’s first point of error is overruled.
In a related point of error, the State argues that the Commission’s order is unreasonable and discriminatory to future ratepayers because the failure to allocate deferred revenues by customer class allows customers who leave the system to escape responsibility for their fair share of the cost of service provided for their benefit. The State’s argument is premature; there has been no show[855]*855ing that the deferred revenues are guaranteed or that the State has been discriminated against by the rate plan currently authorized. Without expressing an opinion on the merits of the State’s argument, we overrule the State’s second point of error as not ripe for our consideration.
Unit 3: Excess Capacity?
PURA section 41(a) provides: “Utility rates shall be based upon the original cost of property used by and useful to the public utility in providing service including construction work in progress at cost as recorded on the books of the utility.” In its first point of error, EPEC asserts that Palo Verde Unit 3 is currently “used and useful” in providing service to its customers and does not represent excess capacity; hence, the costs of this unit should have been included in rate base immediately. EPEC urges several ancillary arguments in support of this point of error: (1) when considered as part of EPEC’s total generating system, Unit 8 operates to capacity to serve EPEC’s retail customers and “off-system” wholesale customers; (2) “off-system” sales must be considered in determining whether Unit 3 is used and useful, because the off-system customers absorb a disproportionate share of EPEC’s total system costs, which in turn benefits Texas ratepayers; (3) the Commission allowed EPEC to recover Unit 3’s operating expenses and the lower capital costs of a hypothetical generating plant with the same capacity as Unit 3, proving that Unit 3 capacity is used and useful; and (4) the Commission’s decision is arbitrary and capricious because it failed to analyze all relevant factors before concluding that Unit 3 is not currently used and useful.
In order to properly evaluate EPEC’s arguments, we must first review the bases and effects of the Commission’s decision. The Commission argues that EPEC improperly focuses its argument on whether Unit 3 is used and useful and ignores the underlying foundation of the Commission’s actions: that EPEC was imprudent in failing to reduce its level of investment in the Palo Verde Nuclear Project after its initial participation in 1978. EPEC’s imprudent level of participation in the Palo Verde project resulted in unneeded capacity in the form of Unit 3, which precluded inclusion of any of Unit 3’s construction costs in rate base in this docket. However, the Commission recognized that the effects of this imprudence were temporary and determined that Unit 3 costs would be included in rate base as increments of the plant become used and useful.5
Much of EPEC’s argument under this point of error rests on the assumption that if a utility’s plant is “actually being used” to produce power for whatever purpose, the plant is per se “used and useful” and therefore includible in rate base. Under EPEC’s view, it is irrelevant whether the utility acquires capacity to serve its on-system customers or to sell power wholesale abroad. We cannot agree with this contention. PURA was enacted to “protect the public interest inherent in the rates and services of public utilities” and to protect consumers in the utilities’ service areas from the unfair or unreasonable rates that could result from a utility’s monopolistic dominance of its service area. PURA § 2. PURA seeks to accomplish this goal by “establish[ing] a comprehensive regulatory system which is adequate to the task of regulating public utilities as defined by this Act, to assure rates, operations, and services which are just and reasonable to the consumer and to the utilities.” Id. To monitor achievement of this goal, PURA directs electric utilities such as EPEC to prepare a detailed biannual report assessing load and resource forecasts for their service areas. See PURA § 16(c).
[856]*856Thus, PURA is concerned primarily with balancing the interests of utilities and their consumers in the area over which the utilities enjoy a monopolistic advantage. As part of this balance, a utility is entitled to have the prudently incurred costs of “used and useful” property included in its rate base when providing service to customers in its service area. See PURA § 41(a); see also Gulf States Utils. Co. v. Coalition of Cities for Affordable Util. Rates, 883 S.W.2d 739, 743 n. 3 (Tex.App.—Austin 1994, writ requested) (discussing the “prudent investment test,” which is not specifically set out in PURA). Even when property is used solely to provide service to on-system ratepayers, the Commission has significant discretion to determine what property is “useful.” When a utility does not use property exclusively to provide service to local ratepayers — and is therefore not operating under the strictly regulated monopoly balance governed by PURA — the Commission has yet more latitude to determine what portion of costs, if any, is “useful to” the utility in serving its “native-system” consumers.
The Commission’s construction of PURA is entitled to great weight. State v. Public Util. Comm’n, 883 S.W.2d 190, 196 (Tex.1994); Southwestern Bell Tel. Co. v. Public Util. Comm’n, 745 S.W.2d 918, 923-24 (Tex.App.—Austin 1988, writ denied) (court defers to agency’s interpretation if construction is reasonable and does not contradict plain language of statute); see also Tarrant Appraisal Dist. v. Moore, 845 S.W.2d 820, 823 (Tex.1993). Although PURA does not distinguish between property that the utility uses exclusively to provide service to its on-system ratepayers and property that benefits these ratepayers only indirectly through sales to off-system wholesale customers, we conclude that the purpose and overall statutory scheme of PURA support the Commission’s interpretation that capacity used for off-system sales is not per se useful to the utility in providing service to its local ratepayers.
Having concluded that EPEC is not entitled per se to inclusion of Palo Verde Unit 3 costs in rate base merely because that unit is in use, we now examine the agency record to determine whether substantial evidence supports the Commission’s decision that Unit 3 represents excess capacity that is not currently used and useful. In City of League City v. Texas Water Commission, 777 S.W.2d 802 (Tex.App.—Austin 1989, no writ), this Court summarized the standards governing appeal of an administrative order under the substantial-evidence rule: (1) The findings, inferences, conclusions, and decisions of an agency are presumed to be supported by substantial evidence, and the burden is on the party contesting the order to prove otherwise; (2) in applying the substantial evidence test, the reviewing court is prohibited from substituting its judgment for that of the agency on questions committed to agency discretion that involve weight of the evidence; (3) substantial evidence is more than a scintilla, but the evidence in the record may preponderate against the decision of the agency and nonetheless amount to substantial evidence; (4) the true test is not whether the agency reached the correct conclusion, but whether some reasonable basis exists in the record for the action taken by the agency; and (5) the agency’s action will be sustained if the evidence is such that reasonable minds could have reached the conclusion that the agency must have reached in order to justify its action. Id. at 805 (citing Texas Health Facilities Comm’n v. Charter Medical-Dallas, Inc., 665 S.W.2d 446, 452-53 (Tex.1984)). In short, the substantial evidence rule is a reasonableness or rational basis test. State v. Public Util. Comm’n, 883 S.W.2d at 203.
Citing extensively to the examiner’s report and the testimony of its own experts, EPEC argues that Unit 3 capacity was useful because it operated on a daily basis to serve EPEC’s base load requirements, contributed to fuel cost savings and fuel diversification, and provided flexibility and system reliability in maintenance scheduling. It is well established that the Commission is not bound by the findings or recommendations of its hearing examiners. See, e.g., Cities of Abilene v. Public Util. Comm’n, 854 S.W.2d 932, 946 (Tex.App.—Austin 1993, writ requested). Moreover, the agency is the arbiter of the credibility and weight to be [857]*857accorded to expert testimony. See Southern Union Gas Co. v. Railroad Comm’n, 692 S.W.2d 137, 141-42 (Tex.App.—Austin 1985, writ ref'd n.r.e.); see also APA § 2001.174. In response to EPEC’s arguments, the City introduced evidence through its expert witness, David Schlissel, that Unit 3 capacity was used to sell wholesale power to off-system purchasers in Mexico, California, and New Mexico, and that “no part of EPEC’s 200 MW [megawatt] share of the capacity represented by this Unit will be required to meet reasonable reserve margins until 1995 at the earliest.” Schlissel also opined that EPEC’s experts relied on unnecessarily high reserve margins in their calculations. In light of this conflicting evidence, we conclude that substantial evidence in the underlying record supports the Commission’s decision that Unit 3 capacity is excess capacity not currently used and useful in providing service to EPEC’s native-system customers.
EPEC further argues that the Commission’s refusal to include Unit 3 costs in rate base is arbitrary because the following facts contradict the Commission’s finding that Unit 3 capacity is not currently used and useful: (1) off-system customers absorb $44 million of system costs, which Texas ratepayers would bear if Unit 3 did not enable EPEC to make off-system sales; and (2) the Commission included an allowance for purchased power costs equal to Unit 3’s capacity of 200 megawatts and for operating and fuel costs of Unit 3.
As the Supreme Court has noted, “The economic judgments required in rate-making proceedings are often hopelessly complex and do not admit of a single correct result.” Duquesne Light Co. v. Barasch, 488 U.S. 299, 314, 109 S.Ct. 609, 619, 102 L.Ed.2d 646 (1989). The Commission discharges its statutory duty to balance the interests of utilities and their consumers by determining whether the utility has acted prudently, whether property is used and useful to the utility in providing service to the customers it is bound to serve, whether the utility’s operating expenses are reasonable and necessary, and whether rates allow the utility a reasonable return on its investment. See PURA §§ 2, 38-41. In the instant cause, the Commission recognized that although Unit 3 is not necessary to provide service to on-system customers, it nonetheless produces certain indirect benefits for these customers. Acknowledging EPEC’s poor financial condition, the Commission compensated EPEC for these temporary current benefits to Texas ratepayers through an allowance in cost of service, even though Unit 3’s capital costs could not yet be added to rate base. Specifically, the Commission: (1) included an allowance for 200 megawatts of purchased power costs at $17.3 million as a substitute for the reallocation of Unit 3 costs EPEC claimed would be needed if Unit 3 did not exist; (2) added Unit 3 operating and fuel costs to EPEC’s cost of service; and (3) allowed EPEC to retain for rate year 1991 all profits from its off-system sales to Mexico.6 In effect, the Commission “segregated” EPEC’s off-system business, allowing EPEC to enjoy the profits generated by the off-system sales, while protecting Texas ratepayers from the massive cost burden of unneeded capacity.
Rather than support EPEC’s position that Unit 3 capacity was actually used and useful, the Commission’s allowances for off-system sales resemble a capital transition allowance. City of El Paso v. Public Util. Comm’n, 609 S.W.2d 574, 577 (Tex.Civ.App.—Austin 1980, writ ref'd n.r.e.); cf. Southwestern Bell Tel. Co. v. Public Util. Comm’n, 571 S.W.2d 503, 516 (Tex.1978) (Commission should consider nature and present usefulness of land held for future use [858]*858when determining inclusion in rate base). In City of El Paso, we held that the Commission had authority to grant EPEC a capital transition allowance, even though such an allowance was not specifically mentioned in PURA or the Commission’s own substantive rules. We concluded that the transition allowance was merely “an adjusted rate of return.” 609 S.W.2d at 577. Similarly, we conclude that the Commission’s inherent authority to establish just and reasonable rates includes the power to grant EPEC allowances for 200 megawatts of purchased power costs and fuel and operating expenses to continue its off-system sales. See PURA § 39(a). These allowances constitute an interim adjustment to EPEC’s rate of return until the excess capacity caused by its deci-sional imprudence is transformed into capacity “used by and useful to” EPEC in providing service to local ratepayers.
Having addressed all of EPEC’s arguments on this issue, we overrule its first point of error.
Adequacy of Findings
1. Prudence Findings
In its fourth point of error, EPEC challenges the adequacy of the Commission’s findings underlying its conclusions that EPEC failed to sustain its burden of proof that (1) its participation in Unit 3 was prudent, and (2) Unit 3 is currently used and useful. An agency’s findings of fact that are expressed in statutory language or that “represent the criteria that the legislature has directed the agency to consider in performing its function” must be supported by findings of underlying fact. Charter Medicalr-Dallas, 665 S.W.2d at 451; APA § 2001.141. Although not statutorily defined, a prudence determination also requires findings of underlying fact. See City of El Paso, 839 S.W.2d at 908. EPEC claims that the findings in this cause are not sufficiently detailed to satisfy APA section 2001.141(d), which requires a “concise and explicit statement of underlying facts” for ultimate findings set forth in statutory language.7 See also Charter Medical-Dallas, 665 S.W.2d at 450-51.
With regard to the prudence issue, the Commission found that Unit 3 represented temporary excess capacity caused by EPEC’s decisional imprudence. The Commission’s ultimate finding that EPEC did not meet its burden of proof in establishing the prudence of its continued level of investment in Palo Verde is supported by findings of fact 49 to 55:
49. EPEC’s management was imprudent in failing to attempt to market an interest in Palo Verde during the 1978-1981 time frame following the entry of the Commission’s final Order in Docket No. 1981.
50. The Commission granted a Certificate for Palo Verde in Docket No. 1981 in part to facilitate EPEC’s ability to sell a portion of its Palo Verde entitlement on reasonable and favorable terms.
51. EPEC’s Board of Directors affirmatively decided not to attempt to market or otherwise dispose of any portion of its Palo Verde entitlement prior to 1981, notwithstanding the intent of its regulators that it do so.
52. The studies EPEC relied upon to justify its refusal to sell a portion of its Palo Verde entitlement are flawed, and EPEC’s reliance on them was unreasonable.
53. EPEC failed to demonstrate by preponderance of the credible evidence that EPEC could not have sold an interest in Palo Verde prior to 1981 had it tried.
54. Based on the evidence in the record, Palo Verde Unit 3 currently represents excess capacity.
55. Because the result of EPEC’s imprudence was excess capacity, it is appropriate to quantify EPEC’s impru[859]*859dence as the level of excess capacity on EPEC’s system.
EPEC claims that these findings fail to elucidate the connection between the flaws in EPEC’s decision-making and a level of investment in Palo Verde that represents excess capacity.
We have previously addressed a similar nexus argument in City of El Paso v. Public Utility Commission, 839 S.W.2d 895 (Tex.App.—Austin 1992), rov’d in part on other grounds, 883 S.W.2d 179 (Tex.1994). In that case we rejected the Office of Public Utility Counsel’s contention that findings of underlying fact must identify the processes and acts found to be imprudent, the nexus between those acts and the Commission’s remedy for the imprudence, or the evidentiary support for the Commission’s remedy or disallowance. Id. at 908. Here, the Commission’s findings taken together reveal EPEC made no attempt to sell a portion of its Palo Verde investment because it relied on studies it should have known were flawed and ignored regulators’ concerns that continuing its full participation in Palo Verde was misguided. These findings certainly support and concisely explain the Commission’s conclusion that EPEC’s continued investment in Palo Verde Unit 3 was imprudent. See APA § 2001.141(d); Charter Medical-Dallas, 665 S.W.2d at 453.
In a related point of error, the City urges that findings of EPEC’s decisional imprudence for its level of investment in Palo Verde in earlier dockets should have foreclosed the Commission’s reconsideration of prudence in the current docket. As authority for its position, the City relies primarily on the supreme court’s decision in Coalition of Cities for Affordable Utility Rates v. Public Utility Commission, 798 S.W.2d 560 (Tex.1990), which held that res judicata and collateral estoppel prohibit a utility’s relitigating historical investment facts such as the prudence of construction expenditures. Id. at 564-65. However, there are salient differences between that case and this one: (1) EPEC’s decisional imprudence was not an historical investment fact but a continuing error in judgment that resulted in temporary excess capacity; (2) the prior dockets’ findings reflected EPEC’s imprudence in “continued” participation in Palo Verde for discrete time frames; (3) the current docket was EPEC’s first application to include in rate base its $300 million investment in Unit 3, as distinct from its investment in Units 1 and 2 and common facilities; and (4) two of the earlier dockets relied on by the City considered the inclusion in rate base of construction-work-in-progress costs.
To invoke the doctrine of collateral estoppel, a party must establish: “(1) the facts sought to be litigated in the second action were fully and fairly litigated in the prior action; (2) those facts were essential to the judgment in the first action; and (3) the parties were cast as adversaries in the first action.” Bonniwell v. Beech Aircraft Corp., 663 S.W.2d 816, 818 (Tex.1984). Earlier Commission findings addressing EPEC’s continued level of involvement in Palo Verde or earlier phases of the project do not constitute actual litigation of EPEC’s decisional prudence regarding Unit 3, which EPEC is only now attempting to include in rate base. Furthermore, the prudence and financial integrity findings in the orders granting a construction work in progress allowance are ambiguous and do not preclude relitigation of these issues.8 The specific facts regarding [860]*860Unit 3’s capacity and the propriety of EPEC’s continued investment in Unit 3 in light of its most recent load projections were not essential to the Commission’s earlier orders and could not have been fully litigated. We overrule the City’s first point of error.
2. Used and Useful Findings
EPEC similarly charges that the Commission’s underlying findings are inadequate to support its conclusion that Unit 3 is currently not used by or useful to EPEC in providing service. While we agree that the Commission’s findings on the “used and useful” issue are not as extensive or conveniently grouped in its order as the findings regarding the prudence issue, we nonetheless conclude that the findings in their totality support the Commission’s ultimate finding that Unit 3 is not currently used by or useful to EPEC in providing service. The Commission made findings of fact discussing EPEC’s use of capacity for off-system sales and concluded that this capacity, as represented by Unit 3, was “excess” considering EPEC’s native-system load. In light of the Commission’s decision that capacity for off-system sales is not “used by and useful to” EPEC in providing service, a factor “clearly implicit in the Commission’s findings,” the findings as a whole support “in a general way the determinations reached by the agency.” See Public Util. Comm’n v. Houston Lighting & Power Co., 715 S.W.2d 98, 109 (Tex.App.—-Austin 1986), rev’d in part on other grounds, 748 S.W.2d 439 (Tex.1987) (quoting Amtel Communications v. Public Util. Comm’n, 687 S.W.2d 95, 107 (Tex.App.—-Austin 1985, no writ)). We are satisfied that the Commission’s findings adequately support its ultimate conclusion that EPEC’s imprudence in failing to reduce its participation in Palo Verde resulted in temporary excess capacity used only to make off-system sales and not used by and useful to EPEC serving its local customers. EPEC’s fourth point of error is overruled.
Certifícate of Convenience and Necessity
In its third point of error, EPEC contends that some of the Commission’s findings underlying its prudence determination are arbitrary because they presume that EPEC was granted a certificate of convenience and necessity in 1978 to facilitate a sale of its interest in Palo Verde. EPEC poses the apparent contradiction as follows: “It is impossible to reconcile the statutory requirement of necessity with the Commission’s conclusion that it certificated Palo Verde to facilitate a sale of unnecessary facilities.” This statement oversimplifies both the Commission’s findings and its statutory obligations. PURA specifically directs the Commission to consider inter alia “the effect of the granting of a certificate on the recipient of the certificate” when evaluating an application for a certificate of convenience and necessity. PURA § 54(c). This statutory obligation is consistent with and reflected in finding of fact 50, in which the Commission notes that EPEC’s certificate was granted “in part to facilitate EPEC’s ability to sell a portion of its Palo Verde entitlement on reasonable and favorable terms.”
The supreme court has acknowledged that a certificate of convenience and necessity “affords only a right to begin construction, not a guarantee that every inefficient or imprudent expenditure will be passed on to the consuming public.” Texas-New Mexico Power Co. v. Texas Indus. Energy Consumers, 806 S.W.2d 230, 233 (Tex.1991). The Commission’s order granting the certificate made clear that EPEC was to “evaluate its continued level of participation in the Palo Verde Nuclear Project.” Tex.Public Util.Comm’n, Application of El Paso Elec. Co. for a Rate Increase, Docket No. 1981, 4 P.U.C.Bull. 480, 481 (Nov. 9, 1978). Commission-approved settlement agreements in the wake of subsequent rate increase requests reinforced the directive that EPEC “present evidence at a future ... date regarding the continued need for [861]*861EPEC’s current participation in [Palo Verde].” Tex.Public Util.Comm’n, Application of El Paso Elec. Co. for a Rate Increase Within El Paso, Hudspeth, and Culberson Counties, Docket No. 2641, 5 P.U.C.Bull. 27 (Sept. 13, 1979). A witness for the City also testified that in 1976 the El Paso City Council investigated EPEC’s participation in Palo Verde and in 1980 adopted a report concluding that EPEC should reduce its interest in the project. As the Commission points out, EPEC waived any complaint about the use of such testimony by failing to object at the hearing.9 This record evidence amply supports the Commission’s findings that EPEC was on notice to reevaluate the level of its investment in Palo Verde. We overrule EPEC’s third point of error.
Regulatory Restraints and Other Considerations
In its second point of error, EPEC asserts as an alternative basis for finding Unit 3 capacity used and useful that the Commission failed to consider that federal and state regulations, combined with long-term oil and gas price and availability forecasts, required EPEC to replace its gas-fired generation facilities with nuclear-powered generating facilities. EPEC argues that the Commission’s failure to give dispositive weight to these federal and state mandates to “get off gas” renders its decision arbitrary and capricious because the Commission has not “taken a hard look at the salient problems and has not genuinely engaged in reasoned decision-making.” See Starr County v. Starr Indus. Servs., Inc., 584 S.W.2d 352, 356 (Tex.Civ.App.—Austin 1979, writ ref'd n.r.e.); see also Public Util. Comm’n v. Gulf States Utils. Co., 809 S.W.2d 201, 211-12 (Tex.1991). In essence, EPEC argues that regulatory constraints were a relevant factor which the Commission was required to explain in making its prudence and usefulness determinations. See Gulf States Utils. Co., 809 S.W.2d at 212.
We do not accept EPEC’s contention that federal and state regulations encouraging utilities to diversify and “get off gas” present “a factual theory that would be dispositive of the prudence issue without regard to the parties’ other contentions regarding prudence.” EPEC claims that the examiner’s report and testimony of Commission staff witnesses constitute unrefuted evidence that full participation in the Palo Verde project was needed to complete displacement of its gas-fired generation facilities, even if the capacity was not needed to serve an increased demand for electricity. However, the only finding regarding the issue in the examiner’s report stated that Unit 3 “contributes to EPEC’s fuel diversification.” Although regulatory constraints may have generally encouraged fuel diversification, they do not resolve the question whether EPEC’s participation in this project on this scale was prudent. Cf. Gulf States Utils. Co. v. Public Util. Comm’n, 784 S.W.2d 519, 533 (Tex.App.—Austin 1990), aff'd, 809 S.W.2d 201 (Tex.1991) (generalities are insufficient justification for specific actions in specific cases). Nor does PURA or the APA require the Commission to explain its rationale for rejecting a hearing examiner’s recommendations. E.g., Cities of Abilene, 854 S.W.2d at 946. The trial court did not err by refusing to consider regulatory restraints or other factors generally promoting the utility’s replacement of gas-fired facilities when reviewing EPEC’s particular level of participation in this particular plant. We overrule EPEC’s second point of error.
Rate Not Confiscatory
By its fifth point of error EPEC claims that the Commission’s failure to include Unit 3 costs in rate base, while expropriating the advantages of Unit 3 for the benefit of Texas ratepayers, resulted in a confiscatory taking of EPEC’s property in violation of the United States and Texas Constitutions. See U.S. Const, amend. V; Tex. Const, art. 1, § 17. EPEC suggests [862]*862that its filing for bankruptcy is clear proof that the Commission’s rates were confiscatory and insufficient to allow EPEC to attract further debt or equity financing. EPEC cites several state and federal cases which hold that the requirement of a reasonable return on investment is met only “if the return is sufficient to assure confidence in the financial integrity of the enterprise, so as to maintain its credit and to attract capital.” Suburban Util. Corp. v. Public Util. Comm’n, 652 S.W.2d 358, 362 (Tex.1983); see also Federal Power Comm’n v. Hope Natural Gas Co., 320 U.S. 591, 603, 64 S.Ct. 281, 288, 88 L.Ed. 333 (1944). However, as the United States Supreme Court has also noted, “[A]ll of the subsidiary aspects of [rate base] valuation [may] not properly be characterized as having constitutional dimension, despite the fact that they might affect property rights to some degree.... ‘If the total effect of the rate order cannot be said to be unreasonable, judicial inquiry ... is at an end.’” Duquesne, 488 U.S. 299, 310, 109 S.Ct. 609, 617, 102 L.Ed.2d 646 (1989) (quoting Hope Natural Gas, 320 U.S. at 602, 64 S.Ct. at 288). As previously discussed, the Commission compensated EPEC for benefits Texas ratepayers received from off-system sales by including Unit 3’s operating expenses, fuel costs, and an allowance for purchased power capacity in EPEC’s cost of service.
Rate-setting requires balancing investor and consumer interests; the result of this balance does not necessarily insure that the utility will produce net revenues. Hope Natural Gas, 320 U.S. at 603, 64 S.Ct. at 288. Constitutional protections merely reinforce the idea that agencies may not arbitrarily disallow expenditures; they do not entitle utilities to immediate return on investments that are imprudent or are not used and useful in providing service to customers. See Suburban Util. Co., 652 S.W.2d at 363. Furthermore, although EPEC complains of the “rate order,” it does not challenge the actual rate of return allowed by the Commission; EPEC merely couches in constitutional takings language its challenge to the Commission’s failure to immediately include EPEC’s Unit 3 investment in rate base. Having already addressed the Commission’s findings and record evidence supporting these decisions, we overrule EPEC’s fifth point of error.
Attorney’s Fees to City
Under PURA section 24(a), a municipality is entitled to reimbursement for the reasonable cost of experts it employs to participate in ratemaking proceedings:
The governing body of any municipality participating in or conducting ratemaking proceedings shall have the right to select and engage rate consultants, accountants, auditors, attorneys, engineers, or any combination thereof, to ... assist with litigation in public utility ratemaking proceedings .... The public utility engaged in such proceedings shall be required to reimburse the governing body for the reasonable cost of such service to the extent found reasonable by the applicable regulatory authority.
In its fourth point of error, the City argues that the Commission erred by failing to reimburse it (1) for the services of an assistant city attorney in the present docket at $75 per hour and (2) for costs it incurred in connection with its participation in earlier dockets. The City presented expert testimony that a reasonable fee for the services performed by the assistant city attorney was $75 per hour. However, the City also presented evidence of the actual cost of the attorney’s services, based on her salary, fringe benefits, one-half of her shared secretary’s salary and benefits, and an allowance for overhead. The Commission awarded the City only these actual costs, which the City calculated to equal an hourly rate of approximately $42.
The reasonableness of attorney’s fees is a question of fact. Tesoro Petroleum Corp. v. Coastal Ref. & Mktg., Inc., 754 S.W.2d 764, 766 (Tex.App.—Houston [1st Dist.] 1988, writ denied). Consequently, we are constrained by the well-known rules governing our review of agency fact findings: we will affirm the agency decision if substantial evidence, i.e., some reasonable basis, supporting the agency action exists in the record. Charter Medical-Dallas, Inc., 665 S.W.2d at 452. The deference we grant an [863]*863agency’s findings under substantial evidence review is similar to the deference afforded a trial court’s award of attorney’s fees. Cf. Mack v. Moore, 669 S.W.2d 415, 420 (Tex.App.—Houston [1st Dist.] 1984, no writ).
We conclude that substantial evidence supports the Commission’s award of attorney’s fees as calculated. Although the City argues that the law does not require it to present a detailed cost study to document every element of its costs, neither does the law require that the Commission accept the City’s conclusion of what is reasonable and award it amounts in excess of actual costs.
In addition to attorney’s fees incurred in Docket 9945, the City requested reimbursement for its participation in prior dockets. The Commission denied reimbursement because the earlier proceedings were not “ratemaking proceedings” under PURA section 24(a). Both the City and the Commission agree that this Court has defined a ratemaking proceeding as one in which a “Statement of Intent” for a rate increase is filed. See City of El Paso, 609 S.W.2d at 579 (holding that hearing on application for certificate of convenience and necessity is not “ratemaking proceeding” for which municipality may recover costs). Although no “Statement of Intent” for a rate increase was filed in the earlier dockets in the instant cause, the City insists that the proceedings were so closely connected to ratemaking that it had to participate to effectively discharge its statutory duties as a regulator under PURA sections 22, 23, and 24 and that recovery of expenses incident to this participation must therefore be allowed.10
This is precisely the issue we addressed in City of El Paso. In holding that PURA section 24 provides for reimbursement only in proceedings held pursuant to a formally filed “Statement of Intent” for a rate change, we expressly rejected the argument that “[u]tility rates seem to be dictated, in substantial part at least, by an accumulation of prior decisions often made in proceedings not exclusively devoted to ‘rates’ as such. Matters are frequently decided in these other proceedings that ‘make’ the rates that are ‘set’ later.” Id. (cited with approval in State v. Public Util. Comm’n, 883 S.W.2d 190, 198 (Tex.1994)). While we sympathize with the City’s desire and arguable obligation to conscientiously participate in other proceedings that may affect rates, we are unable to contravene the mandate of unambiguous statutory language or fill alleged gaps in legislation. See City of El Paso, 609 S.W.2d at 579; Brushy Creek Mun. Util. Dist. v. Texas Water Comm’n, 887 S.W.2d 68, 70-71 (TexApp.—Austin 1994, writ granted). The City’s fourth point of error is overruled.
Rate 41 Tariff Dispute
During the rate-design phase of Docket 9945, the State requested that the University of Texas at El Paso and other state agencies in EPEC’s service area be reclassified in rate class 41, the tariff class applicable to EPEC’s city and county customers.11 The State urges that the Commission’s continued refusal to reassign the State to rate class 41 is unreasonably discriminatory and thereby violates the Commission’s mandate to ensure just and reasonable rates under PURA section 38.12 The State also [864]*864contends that one of the Commission’s findings is arbitrary and capricious because it imposes a new requirement on customers attempting to show that a rate structure is unreasonably discriminatory. We will overrule the point of error.
The Commission exercises broad discretion in designing rates. Texas Alarm & Signal Ass’n v. Public Util. Comm’n, 603 S.W.2d 766, 772 (Tex.1980). Implicit in this authority to establish reasonable, not perfect, classifications is the acknowledgment that some unequal treatment may result. Amtel Communications v. Public Util. Comm’n, 687 S.W.2d 95, 101-02 (Tex.App.—Austin 1985, no writ). However, unequal treatment does not necessarily produce unlawful discrimination; as long as a substantial and reasonable basis exists for the distinction, unequal treatment is neither a violation of PURA nor a reason to invalidate agency decisions. Id.
Cost of service and quantity of service are proper factors for distinguishing between utility customers or classes of customers. Id. at 102; see also Texas Alarm, 603 S.W.2d at 772. Existing classification schemes previously approved by the Commission are prima facie reasonable; once classified, a customer seeking reassignment must show that its “conditions of service” are similar to other members of the proposed class.13 City of El Paso, 839 S.W.2d at 932. Similarity is a question of fact to be determined from the evidence in each case, and the party seeking reassignment bears the burden of proof. Amtel Communications, 687 S.W.2d at 102 (quoting Ford v. Rio Grande Valley Gas Co., 174 S.W.2d 479, 480 (Tex.1943)). Despite the State’s argument that its agencies, the University of Texas at El Paso, and current rate class 41 members perform similar governmental and educational functions, record evidence demonstrates that transferring the State to rate class 41 could significantly affect the volume and cost of service in different classes, which in turn would affect overall rate design. For example, the University of Texas at El Paso is a “large power” customer, with a monthly demand requirement more than 100 times that of the average rate class 41 customer. A Commission staff witness also testified that current load analysis data was insufficient to conclusively determine the load and cost characteristics of the State agencies, and that the impact of a reclassification could be significant.
“The absence of reliable load research data to determine the overall impact on other classes and customers” is a reasonable basis for the Commission’s decision to exclude the State from rate class 41. Furthermore, the Commission was not imposing a new requirement on the State. Its refusal to reclassify state agencies and the University of Texas at El Paso as rate class 41 customers merely indicated that the State failed to meet the burden of proving that its conditions of utility service are sufficiently similar to current rate class 41 members to justify reclassification. See Pedernales Elec. Coop. Inc. v. Public Util. Comm’n, 809 S.W.2d 332, 337 (Tex.App.—Austin 1991, no writ) (finding of fact demonstrating party’s failure to carry its burden of proof). Because the Commission failed to find the similarity necessary for reclassification, it was not required to support its conclusion with underlying findings stating the evidence it rejected or upon which it did not rely. See id. We overrule the State’s fifth point of error.
Deferred Accounting Treatment and Deferred Carrying Costs
In order to alleviate the harsh [865]*865effects of regulatory lag,14 the Commission had previously permitted EPEC to defer certain costs for Palo Verde Units 1, 2, and 3. As part of Docket 9945, the Commission placed $90.8 million of Palo Verde Unit 1 and 2 deferred costs in rate base and postponed consideration of $69.1 million in deferred costs related to Unit 3. The Commission also allowed EPEC to continue accruing a carrying charge of 9.43 percent on the Unit 3 deferred costs. Relying on our decision in City of El Paso, the district court reversed the Commission’s inclusion in rate base of deferred carrying charges for Units 1 and 2. Because the supreme court has reversed that holding and declared that PURA permits the inclusion in rate base of post-in-service carrying costs, we will sustain EPEC’s sixth and the Commission’s second points of error and reverse this portion of the district-court judgment.
As they have done in prior cases, the City, the State, and OPUC all contest the authority of the Commission to permit deferred accounting treatment of post-in-service costs and to ultimately include these costs in rate base. In two recent cases, the supreme court has confirmed that deferred accounting practices and the inclusion in rate base of deferred post-in-service carrying, operating, and maintenance costs do not violate the language or purpose of PURA section 41(a), the rule against retroactive ratemaking, the historical test year requirement, or the rule against single-issue ratemaking. See State of Texas v. Public Util. Comm’n, 883 S.W.2d 190, 198-201 (Tex.1994); City of El Paso v. Public Util. Comm’n, 883 S.W.2d 179, 187-88 (Tex.1994).
However, the State further argues that the Commission erred and abused its discretion by failing to determine what portion of deferred expenses for Units 1 and 2 were necessary to maintain EPEC’s financial integrity. The State’s complaint comes too late. The Commission previously approved inclusion in rate base of the Unit 1 and Unit 2 deferrals in an earlier docket. Tex.Public Util.Comm’n, Application of El Paso Electric Co. for Auth. to Change Bates, Docket No. 7460, 14 Tex.P.U.C.Bull. 932, 1079, 1282 (June 16, 1988). As the supreme court noted at the culmination of that docket’s administrative appeal, the failure of the parties to raise the issue of financial integrity at that time waived the argument. City of El Paso, 883 S.W.2d at 187 n. 17. The current docket merely adjusted the amount of deferrals based on actual costs. See Tex.Public Util.Comm’n, Application of El Paso Electric Co. for Auth. to Change Rates, Docket No. 9165, 16 Tex.P.U.C.Bull. 605, 683 (Aug. 22, 1990) (explaining history of Unit 1 and 2 deferrals). Just as parties may not perpetually challenge a prudence determination once made, neither may they revisit in successive dockets the question of financial integrity after the Commission approves deferred costs for rate base recovery. Accordingly, we overrule the State’s third point of error, the City’s second point of error, and OPUC’s first and second points of error.
The State also argues that the Commission erred in allowing deferrals for Unit 3 because EPEC failed to prove that these deferred costs were prudently incurred. The State has confused the standards for recording deferred charges and recovering deferred charges in rate base: the Commission may authorize deferred cost accounting when necessary to preserve the utility’s financial integrity; a prudence inquiry is required only before the deferred costs are placed in rate base. See id. at 198, 201-02. The Commission declined to consider Unit 3 deferrals for inclusion in rate base in this docket because Unit 3 investment costs were temporarily excluded after the Commission found that Unit 3 currently represents excess capacity. Consequently, the Commission’s failure to make prudence findings regarding Unit 3 deferral costs was not error. We overrule the State’s fourth point of error.
Federal Income Tax Expense
The Commission concluded that it was not constrained by the holding in Public Utility Commission v. Houston [866]*866Lighting & Power Co., 748 S.W.2d 489 (Tex.1987), to apply an “actual-taxes-paid” methodology in computing EPEC’s federal income tax expenses. Accordingly, the Commission refused to recognize subsequent net operating losses and expenses not integral to the production of utility service that diminished the utility’s actual tax liability or its deferred tax account; the district court affirmed this ruling. Likewise, the Commission refused to include disallowed expenses actually claimed as deductions in calculating the utility’s tax liability; the district court reversed this ruling. The supreme court’s recent opinion in Public Utility Commission v. GTE-Southwest, Inc., 901 S.W.2d 401 (1995), validates both aspects of the Commission’s treatment of the utility’s income tax expenses for rate-making purposes.
In GTE-Southwest, the supreme court flatly refused to interpret literally section 41(c)(2) of PURA or the “actual taxes incurred” language of Houston Lighting & Power as requiring the Commission to apply an actual-taxes-paid methodology in calculating income tax expenses in setting rates.15 Id. at 411. The court also held that the Commission was not required, in fact was not permitted, to consider the income tax deductions actually taken for disallowed expenses because such consideration would violate the mandate of section 41(c)(3) that the Commission “shall not consider for ratemak-ing purposes ... any expenditure found by the regulatory authority to be unreasonable, unnecessary, or not in the public interest.” Id. at 411. These two holdings govern all of the challenges to the Commission’s treatment of income tax expenses urged in this appeal.
The City and OPUC contest the district court’s judgment affirming the Commission’s exclusion of deductions for “expenses not integral to the production of utility service.” The City and OPUC also assert that a hypothetical tax expense of $38.5 million for 1987 should have been deleted because carry-back operating losses in 1988, 1989, and 1990 eliminated this tax liability.16 The Commission responds alternatively that (1) the actual-taxes-paid doctrine does not require a utility to pass on to ratepayers tax savings realized as the result of non-utility operations, and (2) the Commission’s decision not to adjust EPEC’s deferred tax account involves an issue of timing within the Commission’s discretion under PURA section 27(e). GTE-Southwest supports both of the Commission’s positions.
Because all of the City’s and OPUC’s challenges to the Commission’s treatment of income-tax expenses are grounded in the now abandoned position that all tax savings actually incurred by the utility must be passed on directly to ratepayers, we overrule all points of error challenging the Commission’s calculation of EPEC’s income tax expenses. Specifically, we overrule the City’s and OPUC’s third points of error regarding application of the actual-taxes-paid doctrine to interest expense, non-utility expenses, and net operating losses. We sustain the Commission’s first and EPEC’s seventh points of error. [867]*867The trial court correctly affirmed the Commission’s refusal to include as deductions those expenses not integral to the production of utility service but incorrectly reversed the Commission’s decision to disregard deductions flowing from disallowed expenses. We will affirm all aspects of the Commission’s decision regarding federal income tax expenses to be included in EPEC’s cost of service.
CONCLUSION
We reverse that portion of the district court’s judgment ordering the Commission to reconsider the deferred carrying costs and render judgment that the Commission’s order be in that aspect affirmed. We also reverse the district court’s judgment that finds error in the Commission’s decision to exclude income tax deductions flowing from disallowed expenses and render judgment that all aspects of the Commission’s decision regarding EPEC’s federal income tax expenses be affirmed. Having determined that the Commission’s order was in the foregoing respects correct, we reverse the district court’s remand of this cause to the Commission and affirm the remainder of its judgment upholding the Commission’s order.
Reversed and Rendered in Part; Affirmed in Part.