OPINION
PHILLIPS, Chief Justice.
This case is an administrative appeal from a final order of the Texas Public Utility Commission. Gulf States Utilities Company (GSU) sought the Commission’s approval of the sale of two of GSU’s generating units and of GSU’s proposed rate treatment of certain revenues and expenses associated with that transaction. GSU proposed to sell the two plants to a joint venture comprised of GSU and three of its Louisiana industrial customers; GSU then planned to purchase the entire electrical output from the plants for distribution to its customers. GSU requested the Commission to rule that, in future rate proceedings, it be allowed to recover from its ratepayers its total purchase price for the electricity and furthermore that it be allowed to allocate the proceeds from the sale of the depreciated plants to its shareholders rather than to its ratepayers.
The Commission concluded that the proposed sale was generally in the public interest but imposed two conditions on GSU’s treatment of the transaction in future rate proceedings. First, the Commission determined that it was not in the public interest for GSU’s ratepayers “to pay in excess of GSU’s avoided cost” for power purchased from the joint venture. Therefore, the Commission held that it would not allow GSU to recover from its Texas ratepayers payments for electricity from the joint venture in excess of GSU’s avoided cost. Second, the Commission determined that GSU must divide the proceeds from the sale of the plants between the ratepayers and its shareholders in proportion to the amount each group contributed to the cost of the plants.
The district court affirmed the Commission’s order. The court of appeals then reversed the judgment of the district court and remanded the case to the Commission. 784 S.W.2d 519. The Commission and the Office of Public Utility Counsel now appeal to this court. We must first determine whether the applicable state and federal regulations permit a utility purchasing power from a cogenerator to recover from its ratepayers payments in excess of its avoided cost. Second, we must determine whether the Commission’s allocation of the utility’s proceeds from the sale of its plants was proper. For the reasons that follow, we affirm the judgment of the court of appeals.
I
A
In 1978, in response to rising energy costs and recent fuel shortages, Congress sought ways to conserve energy to reduce [203]*203the nation’s dependence on foreign oil and on natural gas. See Federal Energy Regulatory Comm’n v. Mississippi, 456 U.S. 742, 745-46, 102 S.Ct. 2126, 2130, 72 L.Ed.2d 532, 538 (1982). One possible remedy was to increase the use of cogeneration. Cogeneration involves the simultaneous production of electrical power and thermal energy, such as heat or steam, usually at an industrial site. Id. at 750 & n. 11, 102 S.Ct. at 2132 & n. 11, 72 L.Ed.2d at 541 & n. 11. By making productive use of the “waste” heat produced in the generation of electricity, cogeneration can reduce fuel consumption by as much as one half. See C. Phillips, The Regulation of Public Utilities: Theory and Practice 452 n. 92 (1988); see also 45 Fed.Reg. 12,215 (1980).
To encourage cogeneration, Congress enacted section 210 of the Public Utility Regulatory Policies Act of 1978 (PURPA). Pub.L. No. 95-617, 92 Stat. 3117 (1978) (codified as amended at 16 U.S.C. § 824a-3 (1988)). An obstacle to the development of cogeneration facilities in the past had been the reluctance of the traditional utilities to purchase excess power from, or sell backup power to, these nontraditional facilities. FERC v. Mississippi, 456 U.S. at 750-51, 102 S.Ct. at 2132-33, 72 L.Ed.2d at 541. Section 210 of PURPA attempts to remove this obstacle by requiring that electric utilities purchase electrical energy from qualifying cogenerating or small power production facilities1 (QFs) and provide back-up power to QFs on a nondiscriminatory basis.2
In section 210, Congress directed the Federal Energy Regulatory Commission (FERC) to prescribe, within one year of the statute’s enactment, rules requiring electric utilities to purchase power from and sell power to QFs. PURPA § 210(a), 16 U.S.C. § 824-3(a) (1988). Congress further provided that the rates established by FERC for the purchases of electricity by a utility (1) shall be just and reasonable to the utility’s customers, and (2) shall not discriminate against QFs. PURPA § 210(b), 16 U.S.C. § 824-3(b) (1988). Section 210(b) also provides that “[n]o ... rule prescribed under subsection (a) of this section shall provide for a rate which exceeds the incremental cost to the electric utility of alternative electric energy.” Id. The “incremental cost of alternative electric energy” is defined as the cost the utility would have incurred if it had generated itself or purchased from another utility the same amount of power it purchased from the QF. PURPA § 210(d), 16 U.S.C. § 824a-3(d) (1988). By setting a ceiling of incremental cost on the amount a utility could be forced to pay for a QF’s power, Congress intended to encourage cogeneration without requiring a utility’s ratepayers to subsidize cogenerators. H.R.Conf.Rep. No. 1750, 95th Cong., 2d Sess. 98, reprinted in 1978 U.S.Code Cong. & Admin.News 7659, 7797, 7832.
Pursuant to this statutory authority, FERC has adopted regulations governing an electric utility’s purchases from QFs. See 18 C.F.R. pt. 292 (1990). The regulations require that each electric utility “purchase, in accordance with § 292.304, any energy and capacity which is made available from a qualifying facility.” Id. § 292.303. Section 292.304 sets the rate of payment for such purchases equal to the utility’s full avoided cost unless the state regulatory authority (or an unregulated utility) determines that a lower rate is in the public interest and is sufficient to encourage cogeneration. Id. § 292.304(b)(2). (The term “full avoided cost” is equivalent to PURPA’s “incremental cost.”) Finally, the regulations provide that “[njothing in this subpart ... [ljimits the authority of any electric utility or any [QF] to agree to [204]*204a rate for any purchase ... which differ[s] from the rate ... which would otherwise be required by this subpart.” Id. § 292.301(b)(1).
Because Congress intended that state regulatory authorities be the primary enforcers of PURPA, section 210(f) of PURPA orders each state regulatory authority to implement FERC’s rules.3 16 U.S.C. § 824-3(f) (1988). In 1981, the Texas legislature added a provision to the Texas Public Utility Regulatory Act (Texas PURA) to the effect that the Commission “shall make and enforce rules reasonably required to implement the rules and regulations of the Federal Energy Regulatory Commission pertaining to the production of electric energy by [QFs].” Act of Apr. 1, 1981, 67th Leg., R.S., ch. 31, § 2, 1981 Tex.Gen.Laws 70, 71. This provision was later amended to order the Commission to “make and enforce rules to encourage the economical production of electrical energy by [QFs].” Act of May 26, 1983, 68th Leg., R.S., ch. 274, § 1, 1983 Tex.Gen.Laws 1258, 1280 (codified at Tex.Rev.Civ.Stat.Ann. art. 1446c, § 16(g) (Vernon Supp.1991)). Pursuant to these directives, the Commission has enacted its own rules governing a utility’s purchases of power from a QF. Tex.Pub. Util.Comm’n, 16 Tex.Admin.Code § 23.66 (West Sept. 1, 1988) (“Rule 23.66”).
The Texas rules apply to FERC-certified QFs, id. § 23.66(a)(15), and are very similar to FERC’s rules. Rule 23.66(d) of the Texas rules provides that “[i]n accordance with subsections (e) — (h) of this section, each electric utility shall purchase any energy and capacity that is made available from a qualifying facility....” Id. § 23.66(d)(1)(A). Rule 23.66(e) requires that rates for such purchases “shall be just and reasonable to the consumers of the electric utility and in the public interest, and shall not discriminate against” QFs. Id. § 23.66(e)(1). The rates “shall not exceed avoided cost.” Id. § 23.66(e)(2). “Avoided cost” has the same meaning as in FERC’s regulations. Id. § 23.66(a)(2). Rates which equal avoided cost are deemed to be just and reasonable, id. § 23.66(e)(3), and such payments are considered “reasonable and necessary operating expenses of [the] utility.” Id. § 23.66(e)(5).4 Finally, as in FERC’s rules, the Texas rules provide that nothing in the rules “shall limit the authority of any electric utility or any qualifying facility to agree to a rate for any [205]*205purchase, or terms or conditions relating to any purchase, which differ from the rate or terms or conditions that would otherwise be required by this subsection.” Id. § 23.66(b)(2)(A).
The first issue we must address in this case concerns the proper interpretation of the above federal and state regulations governing a utility’s purchases of electricity from a QF. The questions presented are (1) whether the regulations prohibit a utility from contracting for and paying a rate in excess of its avoided cost and (2) whether, if a utility may pay more than avoided cost, the regulations impose a limitation on how much of these payments the utility can recover from its ratepayers. The controversy arises because both sets of regulations expressly allow a utility and a QF to agree to any rate for purchased power, see 18 C.F.R. § 292.301(b)(1) (1990); 16 Tex.Admin.Code § 23.66(b)(2)(A) (West Sept. 1, 1988), but in subsequent provisions limit purchased power payments to avoided cost. See 18 C.F.R. § 292.304(a)(2) (1990); 16 Tex.Admin.Code § 23.66(e)(2) (West Sept. 1, 1988).
B
We consider these questions in light of the circumstances presented by this case. GSU is a public electric utility located in Texas near the Louisiana border. In 1984, two of GSU’s largest industrial customers, both situated in Louisiana, notified GSU that they were planning to build a cogener-ation facility, which would enable them eventually to withdraw from GSU’s system.5 Normally, the loss of such major customers from GSU’s system would result in increases in the rates paid by the remaining customers. These increases would occur because the utility’s fixed costs would have to be spread over a smaller number of ratepayers. As a way to keep the industrial customers in its system and thereby to prevent such rate increases, GSU proposed the formation of a joint cogeneration venture with the customers.
GSU and three of its Louisiana-based industrial customers eventually entered into an agreement to form the Nelson Industrial Steam Company Project (the Venture). Under the terms of this agreement, GSU agreed to sell two of its electrical generating plants to the Venture and to run the plants on behalf of the Venture.6 GSU further agreed to purchase the entire electric output from the plants for its general power supply at a price calculated by a contractual formula.7 In return for GSU’s undertakings, the three industrial members of the Venture agreed to continue as GSU’s customers. FERC later certified the Venture as a QF, contingent on the Venture’s reaching certain milestones in converting the plants from natural gas to petroleum coke operation.
Completion of the agreement to form the Venture was made contingent on the Commission’s approval of GSU’s proposed rate treatment of certain revenues and expenses arising out of the sale. GSU requested such approval under section 63 of the Texas PURA. This section requires public utilities to report to the Commission any sale of a plant when the total consideration exceeds $100,000. The Commission then must determine if the transaction “is consistent with the public interest.” Tex.Rev. Civ.Stat.Ann. art. 1446c, § 63 (Vernon Supp.1991). If the Commission finds that the transaction is not in the public interest, the Commission “shall take the effect of the transaction into consideration in the rate-making proceedings” and “shall ... disallow the effect of such transaction if it will unreasonably affect rates or service.” Id.
[206]*206Section 63 does not require a utility to obtain the Commission’s approval prior to such sales. Furthermore, the Commission does not set electric utility rates in a section 63 proceeding. Nevertheless, the Commission must determine the effect of a transaction on future ratemaking proceedings in order to determine whether the transaction as proposed will be in the public interest. In their section 63 application, therefore, the parties to the Venture requested a prospective ruling on the future rate treatment that would be granted the transaction. Specifically, GSU requested a ruling from the Commission that GSU be allowed to pass on the full amount of its payments to the Venture for electricity (the purchased power payments) to its ratepayers and that it be allowed to allocate the entire proceeds from the sale of the plant to its shareholders.
After a hearing, the Hearing Examiner determined that the sale of the plants to the Venture was generally in the public interest. The Examiner then found that GSU’s purchased power payments might at times exceed GSU’s avoided cost. The Examiner determined that avoided cost was the maximum rate that would be deemed just and reasonable and in the public interest under Rule 23.66(e). However, she concluded that GSU should be allowed to seek recovery of any purchased power payments in excess of its avoided cost in future rate or fuel-reconciliation proceedings.8 In order to recover the payments, GSU would have to show that the
costs were necessary as a means to keep its industrial load on the system, that the ratepayers benefited by the retention of the industrial load on the system, that the costs in excess of avoided costs were at a minimum, and that given these circumstances, its fuel costs were at their lowest reasonable level.
Tex.Pub.Util.Comm’n, Application of Gulf States Utils. Co. for Approval of a Joint Venture Cogeneration Project and Treatment of Revenues, Docket No. 7147, 14 Tex.P.U.C.Bull. 49, 66-67 (Mar. 21, 1988) [hereinafter Docket No. 7147], The Examiner further determined that it would be unreasonable for the ratepayers to absorb all the costs associated with the purchased power payments. As to the allocation of the sale proceeds, the Examiner concluded that the ratepayers should be credited with 83 percent of the proceeds because they paid for 83 percent of the cost of the plants. The ratepayers’ 83 percent share of the proceeds would be deemed “other electric utility income” and would be used to offset the amount of revenue otherwise required to be generated by the rates.
The Commission's final order deleted the Hearing Examiner’s findings that GSU could seek recovery of the full amount of its purchased power payments and substituted findings that GSU could not recover anything above avoided cost. The final order thus stated that “[i]t is not in the public interest for GSU’s Texas ratepayers to pay in excess of GSU’s avoided cost for purchased power from a qualifying cogen-eration project” and that “[i]n future rate proceedings, [GSU] is limited to recovering those purchased power payments to the Venture that do not exceed GSU’s avoided costs.”9 Docket No. 7147, supra p. 10, at 98. The parties to the Venture later went ahead with the transaction despite the Commission’s disallowance of the requested rate treatment.
GSU appealed to the district court, which affirmed the Commission’s order. GSU then appealed to the court of appeals, which reversed the judgment of the district court and remanded to the Commission. The court of appeals held that the Commission’s interpretation of Rule 23.66 as imposing a ceiling on negotiated purchased power payments was unreasonable under the text of the provision and contrary to the provisions and purposes of the Texas [207]*207PURA. 784 S.W.2d at 528. The court of appeals also held that the Commission’s allocation of the sale proceeds was not supported by substantial evidence in the record. 784 S.W.2d at 533.
II
We first address that part of the Commission’s order holding that GSU cannot recover from its ratepayers purchased power payments in excess of avoided cost. The Commission’s ruling is based on its interpretation of the state regulations, in particular Rule 23.66. As described above, Rule 23.66(b)(2)(A) allows utilities and QFs to agree to any rate, while Rule 23.66(e) states that purchases of energy from QFs shall not exceed avoided cost. Thus, the two rules appear to conflict. The Commission attempts to harmonize the two rules by interpreting them to permit a utility and a QF to contract only for rates below avoided cost. The Commission also argues that, even if Rule 23.66(b) allows a utility to pay a rate above avoided cost, Rule 23.66(e) prohibits a utility from recovering from its ratepayers any payments, whether negotiated or compelled, in excess of avoided cost. According to the Commission, Rule 23.66(e) provides that only those payments that are equal to avoided cost are just, reasonable, and in the public interest. Therefore, it is not just, reasonable, or in the public interest for the ratepayers to pay more than avoided cost under any circumstances. Finally, the Commission argues that FERC’s regulations also place a limit of avoided cost on purchased power payments.
The Commission’s interpretation of its own regulations is entitled to deference by the courts. See Udall v. Tallman, 380 U.S. 1, 16, 85 S.Ct. 792, 801, 13 L.Ed.2d 616, 625 (1965); Lloyd A. Fry Roofing Co. v. State, 541 S.W.2d 639, 644 (Tex.Civ.App.—Dallas 1976, writ ref’d n.r.e.). Our review is limited to determining whether the administrative interpretation “is plainly er roneous or inconsistent with the regulation.” United States v. Larionoff, 431 U.S. 864, 872, 97 S.Ct. 2150, 2155, 53 L.Ed.2d 48, 56 (1977) (quoting Bowles v. Seminole Rock Co., 325 U.S. 410, 414, 65 S.Ct. 1215, 1217, 89 L.Ed. 1700, 1702 (1945)). However, if the Commission has failed to follow the clear, unambiguous language of its own regulation, we must reverse its action as arbitrary and capricious. See Sam Houston Elec. Coop., Inc. v. Public Util. Comm’n, 733 S.W.2d 905, 913 (Tex.App.—Austin 1987, writ den’d); see also Lewis v. Jacksonville Bldg. & Loan Ass’n, 540 S.W.2d 307, 310 (Tex.1976), and Ex parte Roloff, 510 S.W.2d 913, 915 (Tex.1974).10
We hold that the Commission acted arbitrarily in adopting an interpretation contrary to the plain language of its regulation. Rules 23.66(b)(2)(A) and 23.66(e) can be harmonized, but not in the manner suggested by the Commission. We read Rule 23.66(e) as operating solely to set the rates that the Commission can compel a utility to pay for a QF’s power if the utility and the QF are unable to reach a voluntary agreement. Rule 23.66(e) does not impose a ceiling on the amount a utility can contract to pay for a QF’s power, nor does it limit the amount a utility can recover from its ratepayers under such voluntary arrangements.
Our interpretation is based on the interaction between Rule 23.66(e) and Rules 23.-66(d) and 23.66(b). Rule 23.66(d) obligates a utility to purchase power from a QF if the QF makes such power available. Subsection (d) also provides that the rates for purchases made under that subsection are governed by Rule 23.66(e). See 16 Tex.Admin.Code § 23.66(d)(1)(A) (West Sept. 1, 1988) (“In accordance with subsection[ ] (e) ... of this section, each electric utility shall purchase any energy and capacity that is made available from a [QF]_”). Rule 23.66(e)’s avoided-cost rules therefore do [208]*208not apply to voluntary contracts arranged outside the requirements of Rule 23.66(d). This interpretation of subsections (d) and (e) is in harmony with the plain language of Rule 23.66(b), which permits a utility and a QF “to agree to a rate which differ[s] from the rate that would otherwise be required” whether above, below, or equal to avoided cost. We find no justification in the language of Rule 23.66(b) to support the Commission’s interpretation that subsection (b) allows a utility to contract for any price below avoided cost but not for a price above avoided cost. Finally, because Rule 23.66(e) does not apply to contractual arrangements between utilities and QFs, it does not impose any limitation on the amount a utility can recover from its customers for such contractual payments. The provision that payments equal to avoided cost “shall be considered reasonable and necessary operating expenses of that utility” merely ensures that a utility can recover the full amount of the payments it is compelled to make by the Commission. Subsection (e)(5) does not provide that contractual payments above avoided cost can never be considered reasonable and necessary operating expenses of the utility.
We must also consider the application of PURPA and the federal regulations because the transaction between GSU and the Venture is governed by federal law as well as by Texas law. The interstate activities of GSU clearly bring it within the reach of the federal regulations. FERC v. Mississippi, 456 U.S. 742, 757, 102 S.Ct. 2126, 2136, 72 L.Ed.2d 532, 545 (1982). In addition, to the extent that the state regulations implement PURPA, the federal regulations serve as a guideline to and may control the scope of the state regulations.
We hold that the federal regulations do not grant the Commission any more authority to regulate negotiated purchases than do the state regulations.11 Like the Texas regulations, the federal regulations do not authorize the Commission to alter the terms of a purchased power contract between a utility and a QF. Neither do they set a limit of avoided cost on the amount of such payments a utility can recover from its ratepayers. Thus, section 292.303(a) of the federal regulations provides that a utility is obligated to purchase any energy and capacity made available from a QF. Such purchases must be made in accordance with section 292.304, which sets the rate for purchases at avoided cost. Because section 292.304 is applied through section 292.303, the avoided-cost limit applies only to compelled purchases. This interpretation of sections 292.303 and 292.-304 is in harmony with the language of section 292.301(b)(1), which provides that nothing is to limit the authority of a utility or QF “to agree to a rate for any purchase ... which differ[s] from the rate ... which would otherwise be required by this sub-[209]*209part.” 12 18 C.F.R. § 292.301(b)(1) (1990). See American Paper Inst., Inc. v. American Elec. Power Serv. Corp., 461 U.S. 402, 416, 103 S.Ct. 1921, 1930, 76 L.Ed.2d 22, 35 (1983) (“The Commission’s [full-avoided-cost] rule simply establishes the rate that applies in the absence of a waiver or a specific contractual agreement.”);13 see also In re Vicon Recovery Sys., 153 Vt. 539, 572 A.2d 1355, 1358 (1990) (“The rate provisions of § 292 apply only ... in a situation where the electric utility is forced to purchase power from the small producer. The regulations make clear that utilities and [QFs] can agree to a rate different than would otherwise be mandated.”); Barasch v. Public Util. Comm’n, 119 Pa.Cmwlth. 81, 546 A.2d 1296, 1300 (“privately negotiated contracts setting rates for QF power are essentially outside the federal and state rules”), modified, 119 Pa.Cmwlth. 81, 550 A.2d 257 (1988) (holding that previous opinion was to have prospective effect only); Bates Fabrics Inc. v. Public Utils. Comm’n, 447 A.2d 1211, 1214 (Me.1982) (“We conclude that the federal scheme expressly excludes from its reach all otherwise binding contracts between utilities and” QFs.). Finally, the regulations say nothing about how much of its purchased power payments a utility can recover from its ratepayers.
Our holding that the state and federal regulations governing a utility’s purchases of power from a QF do not apply to voluntary arrangements between a utility and a QF does not deprive the Commission of its regulatory authority over the amount of such contractual payments that a utility may recover from its customers. GSU may contract for any purchase price it wishes; however, whether such cost will be fully recoverable from the ratepayers will be subject to the Commission’s ordinary ratemaking powers. GSU’s purchase of electricity from the Venture is a fuel cost, see 16 Tex.Admin.Code § 23.23(b)(2)(B) (West Sept. 1, 1988), which, like any other expense, is subject to disallowance by the Commission upon a finding that the expense is unreasonable, unnecessary, or not in the public interest.14 See Tex.Rev.Civ. StatAnn. art. 1446c, §§ 38, 39(a), 41(c)(3)(D) (Vernon Supp.1991);15 16 Tex. [210]*210Admin.Code § 23.23(b) (West Sept. 1, 1988) (utility must show that contract negotiations for purchased power have produced the lowest reasonable cost of fuel to ratepayers); see also Suburban Util. Corp. v. Public Util. Comm’n, 652 S.W.2d 358, 362 (Tex.1983) (“The PUC’s ratemaking power includes the discretion to disallow improper expenses.”). In fact, section 63, under which this proceeding was brought, expressly orders the Commission to disallow any effect of the transaction that will unreasonably affect rates.
The Commission suggests that it would never be fair to the ratepayers for a utility to pay in excess of its avoided costs for power since by definition the utility could obtain the same amount of power elsewhere by paying avoided cost. But using this argument to set an absolute ceiling on purchased power payments ignores the plain language of the Commission’s own rules, which do not authorize the Commission to impose a ceiling of avoided cost on such payments. Under the rules, the Commission must conduct an independent, factual inquiry in each case to determine whether payments in excess of avoided cost are reasonable and necessary, considering the effects of the transaction as a whole.16
Because the Commission was not called upon to set rates in the section 63 proceeding from which this case arose, we do not remand this part of the case to the Commission. Instead, we order the Commission to allow GSU to recover purchased power payments in excess of its avoided cost in future rate proceedings if GSU establishes to the Commission’s satisfaction that the payments are reasonable and necessary expenses. The Commission contends that, since GSU and the industrial customers have already gone ahead with the Venture, GSU can no longer justify the payments on the grounds that they are necessary to retain the customers in the system. However, we believe that it may still be possible for GSU to establish the necessity of the Venture and the contractual rates. GSU should be allowed to show that, absent the Venture, the industrial customers would have left its system because independent cogeneration was economically more attractive than remaining in the system, that the contractual rates are necessary to make the Venture more attractive than independent cogeneration, and that such rates are at the minimum level. If GSU is able to satisfy the Commission that payments above avoided cost are justified, then the Commission should determine what portion of the costs of the Venture it is reasonable and necessary for the ratepayers to bear, given the distribution of benefits from the Venture to the ratepayers and to the shareholders.17
Ill
The second issue we must resolve is whether the Commission properly allocated the gains from the sale of the plants between GSU’s ratepayers and its shareholders. Under the agreement, the Venture is to pay for the plants in twenty annual installments of $6.35 million each (the fixed asset payments), for a total income stream of $127 million. GSU requested that it be allowed to allocate the entire amount of each fixed asset payment, i.e., the entire proceeds of the sale, to its shareholders. The Commission instead ordered that GSU allocate 83 percent of each fixed asset payment, or 83 percent of the total sale proceeds, to its ratepayers, thereby reducing the amount the utility was entitled to recover through its rates by that amount.
We must reverse the Commission’s allocation of the profits if it is not [211]*211reasonably supported by substantial evidence in the record or if the Commission’s action was arbitrary, capricious, or an abuse of discretion. Administrative Procedure and Texas Register Act, Tex.Rev.Civ. Stat.Ann. art. 6252-13a, § 19(e)(5), (6) (Vernon Supp.1991); Railroad Comm’n v. Continental Bus Sys., 616 S.W.2d 179, 181 (Tex.1981). Our inquiry is whether the Commission’s decision is reasonably supported by substantial evidence in view of the reliable and probative evidence in the record as a whole. Texas Health Facilities Comm ’n v. Charter Medical-Dallas, Inc., 665 S.W.2d 446, 452 (Tex.1984). Agency decisions that are not supported by substantial evidence are deemed arbitrary and capricious. Id. at 454. In applying this test, we may not, however, substitute our judgment as to the weight of the evidence for that of the agency. Id. at 452.
The Commission’s decision to allocate 83 percent of the fixed asset payments to the ratepayers is based on the testimony of the Office of Public Utility Counsel’s witness Dr. Steven Anderson and of the Commission’s staff accountant Paul Bellon. GSU has indirectly recovered 83 percent of the cost of the plants from the ratepayers through depreciation expense allowed for the plants and figured into the rate base. Both Anderson and Bellon testified that “because the Company and the shareholders have recovered 83 percent of the costs related to [the generating facilities], the ratepayers should receive 83 percent of the fixed asset payment GSU receives from the sale of the units.” Docket No. 7147, supra p. 10, at 56.
We hold that the Commission’s decision is not reasonably supported by substantial evidence in the record. The Commission based its decision solely on conclu-sory testimony that cites only the evidence of the ratepayers’ contribution to depreciation. The Commission did not consider any other factors that would be relevant to allocation of a utility’s proceeds from the sale of assets. The issue of the proper allocation of such proceeds is a complicated one that cannot be resolved simply by reference to who paid for the property. See Washington Pub. Interest Org. v. Public Serv. Comm’n, 393 A.2d 71, 72 (D.C.1978); see also Democratic Cent. Comm. v. Washington Metro. Area Transit Comm’n, 485 F.2d 786 (D.C.Cir.1973); Kansas Power & Light Co. v. State Corp. Comm’n, 5 Kan.App.2d 514, 620 P.2d 329 (1980). A proper allocation of proceeds can be determined only by an analysis of all the equities involved. See, e.g., Democratic Cent. Comm., 485 F.2d at 821; Kansas Power & Light Co., 620 P.2d at 340.
The equitable principles commonly used to resolve allocation problems are that “benefits should follow burdens” and that “gain should follow risk of loss.” Democratic Cent. Comm., 485 F.2d at 806; Washington Pub. Interest Org., 393 A.2d at 92. In other words, in the general case, the gain should be allocated to that group (as between shareholders and ratepayers) that has borne the financial burdens (e.g., depreciation, maintenance, taxes) and risks of the asset sold. In addition to these two general equitable factors, courts have also considered numerous other factors, including whether the asset sold had been included in the rate base over the years, whether the asset was depreciable property, nonde-preciable property, or a combination of the two types, the impact of the proposed allocation on the financial strength of the utility, the reason for the asset’s appreciation (e.g., inflation, a general increase in property values in the area), any advantages enjoyed by the shareholders because of favored treatment accorded the asset, the dividends paid out to the shareholders over the years, and any extraordinary burdens borne by the ratepayers in connection with that asset. See Democratic Cent. Comm., 485 F.2d at 791-92, 821-22; id. at 831-32, 841-44 (MacKinnon, J., concurring in part and dissenting in part); Washington Pub. Interest Org., 393 A.2d at 89, 91-92; Kansas Power & Light Co., 620 P.2d at 341.
In the present case, the Commission allocated the sale proceeds in direct proportion to the amount the shareholders and the ratepayers contributed to the cost of the plants. This formula is based on only one of the above principles, that benefits should follow burdens. In addition to ignoring the [212]*212other considerations outlined above,18 the Commission did not articulate its reasons for concluding that the ratepayers are entitled to a percentage of the proceeds exactly equal to the percentage of the cost they have paid. The Commission cited only the testimony of witnesses Anderson and Bel-lon but there is no evidence in the record to explain how they reached their conclusions. We therefore hold that the Commission’s allocation is not supported by substantial evidence because we find that the evidence the Commission cited does not support its decision and because the Commission failed to consider evidence relating to other relevant factors. We remand the ease to the Commission for it to recalculate the allocation of the fixed asset payments along the lines of the analysis suggested above, although we do not require the Commission to consider all of the above factors nor forbid it from considering others. However, the Commission should set forth the factors it considers relevant and should explain how these factors are evaluated in the present case.
IV
For the reasons stated above, we affirm the judgment of the court of appeals. This case is therefore remanded to the Commission for it to determine the proper allocation between the shareholders and ratepayers of the fixed asset payments after consideration of the factors the Commission determines are appropriate. The Commission is further ordered to allow GSU to recover its purchased price payments in future rate proceedings if GSU shows that the payments are reasonable and necessary expenses.