The Public Service Commission (PSC) found that National Fuel’s allocation of the insurance settlement proceeds based on the premiums paid by each of its subsidiaries “was unjust and unreasonable at the time it was made” and that the proceeds should have been apportioned, instead, based upon the potential liabilities each subsidiary faced for certain environmental remediation expenses. Since there is substantial evidence to support the PSC’s conclusion, I would reverse.

Among the general powers accorded to the PSC by statute is the authority to evaluate proposed utility rate changes and to set just and reasonable rates (see Public Service Law § 66 [12] [f]; § 72). “Indeed, it has been recognized that when it comes to setting rates for [gas and electric] service the Commission has been granted ‘the very broadest of powers,’ the Legislature mandating only that the rates fixed be ‘just and reasonable’ ” (Matter of Niagara Mohawk Power Corp. v Public Serv. Commn. of State of N.Y., 69 NY2d 365, 369 [1987] [citations omitted]). The burden of proving that any proposed rate change is just and reasonable rests with the utility (see Public Service Law § 66 [12] [i]).

Our review of PSC determinations involving ratesetting has always been deferential. We have emphasized that the standard of review is meant to be flexible and that the determinations *372“ ‘may not be set aside unless they are without rational basis or without reasonable support in the record’ ” (Matter of New York Tel. Co. v Public Serv. Commn. of State of N.Y., 95 NY2d 40, 48 [2000], quoting Matter of Rochester Tel. Corp. v Public Serv. Commn. of State of N.Y., 87 NY2d 17, 29 [1995]). It is well settled that “substantial evidence consists of proof within the whole record of such quality and quantity as to generate conviction in and persuade a fair and detached fact finder that, from that proof as a premise, a conclusion or ultimate fact may be extracted reasonably—probatively and logically” (300 Gramatan Ave. Assoc. v State Div. of Human Rights, 45 NY2d 176, 181 [1978]). Such deference is appropriate since ratesetting “presents ‘problems of a highly technical nature,’ ” which are well within the particular expertise of the PSC (see New York Tel. Co., 95 NY2d at 48, quoting Matter of Abrams v Public Serv. Commn. of State of N.Y., 67 NY2d 205, 211-212 [1986]).

The determination of whether a utility acted prudently is made by assessing whether its actions were reasonable under the circumstances existing at the time they were made, without the benefit of hindsight (see Matter of Long Is. Light Co. v Public Serv. Commn. of State of N.Y., 134 AD2d 135, 143-144 [3d Dept 1987]). Significantly, “the PSC’s broad rate-making powers . . . are sufficient to allow it generally to assess the prudence of a utility’s actions as those actions impact upon the ratepayers. Indeed, a specific function of the rate-making power is to protect the utility’s ratepayers” (Matter of Crescent Estates Water Co. v Public Serv. Commn. of State of N.Y., 77 NY2d 611, 617 [1991]). Through this power, the Commission ensures that unreasonable or excessive rates are not inflicted upon the utility’s customers (see e.g. Matter of General Tel. Co. of Upstate N.Y. v Lundy, 17 NY2d 373, 381 n 3 [1966]).

Even assuming that our traditional deference to the PSC is skewed toward the utility for the purpose of prudence determinations (see majority op at 368-369), the PSC has provided a rational basis for its determination that National Fuel acted imprudently.

The PSC order notes that, at the time the insurance proceeds were distributed, National Fuel was aware of the projected liability for environmental remediation faced by its subsidiaries, including that NFG Distribution’s potential liability accounted for 64% of the total estimated liability. Instead, only 46% of the proceeds were allocated to NFG Distribution based on its proportionate share of the premiums paid for the subject general liability insurance policies. By use of this allocation method, *373a portion of the settlement proceeds was distributed to National Fuel’s nonregulated affiliated companies. The PSC concluded that National Fuel should have taken into account the available estimates of potential environmental liability for each subsidiary that were available at the time the proceeds were allocated and that the appropriate distribution should have been made in proportion to that prospective liability. The PSC therefore determined that the 46% allocation to NFG Distribution was “unjust and unreasonable.”

A review of the record as a whole provides ample basis for the PSC’s conclusion. PSC’s expert testified that the allocation of proceeds on the basis of insurance premiums paid was “not accurate and [made] no sense.” He concluded that the settlements were reached on the basis of projected environmental liabilities and “had no relation to the amount of insurance premiums paid.” PSC’s expert relied, in part, on the answers to interrogatories that were provided by National Fuel. It was clear from this evidence that National Fuel sought coverage for its environmental liabilities under its general liability insurance policies. After those carriers initially denied the claims, National Fuel retained environmental consultants to estimate the potential environmental site investigation and remediation costs. The attorney who represented National Fuel in the settlement negotiations submitted an affidavit, affirming that the environmental report was prepared “explicitly for insurance settlement negotiation purposes only, in order to identify actual and potential environmental risks” and that counsel used the report “to present settlement demands” to the insurance companies. Although the insurers asserted that the policies did not apply to environmental liability, they ultimately settled.

The evidence in the record therefore establishes that the insurance proceeds were pursued and obtained for the specific purpose of dealing with potential environmental liability. Given that the policies at issue were general, and not environmental, liability policies, the proportion of premiums paid by the various entities was completely unrelated to the settlement proceeds. The IES report was at least related to the subject of the settlement—environmental liability. The methodology the majority embraces as beyond the reach of the PSC has no demonstrated relationship to that issue, the settlement negotiations or the settlement proceeds. We simply are not confronted with a choice of two prudent methodologies.

The idea of dividing insurance proceeds among insureds in proportion to premiums is strange at best. The reason for *374buying insurance is to protect against unforeseen losses, not to get a return on premiums paid. The “premiums paid” methodology, as applied in this case, gives “protection” against environmental liabilities to companies that suffered no environmental loss, including some that may never have had any environmental risk. If a utility had negotiated such an allocation at arm’s length with independent coinsureds, the agreement would raise a serious question of prudence. Here, where the allocation of insurance proceeds was among affiliated companies, whose management had an obvious interest in maximizing the burden on the ratepayers in order to minimize cost to the shareholders, the PSC was all the more justified in looking at that allocation with a skeptical eye.

In these circumstances, the PSC clearly had a rational basis to conclude that the allocation of proceeds to National Fuel’s subsidiaries based on premiums paid was imprudent. There is substantial evidence to support the PSC’s determination and therefore the Appellate Division judgment should be reversed.

Judges Ciparick, Read and Jones concur with Judge Graffeo; Chief Judge Lippman dissents and votes to reverse in a separate opinion in which Judges Smith and Pigott concur.

Judgment affirmed, with costs.

OPINION OF THE COURT

In this appeal, we consider the legal standards that apply when a utility company seeks permission from the Public Service Commission (PSC) to recoup from ratepayers certain environmental remediation costs it has incurred. We hold that when the PSC reviews a management decision of a utility to assess its prudence, the Department of Public Service (DPS) bears the initial burden of showing that the utility may have acted imprudently based on what was known at the time the challenged decision was made. Furthermore, there must be a rational basis in the record evidence to support the grounds cited in a PSC order for a finding of imprudence.

I

Petitioner National Fuel Gas Distribution Corp. (NFG Distribution) is a natural gas delivery utility that operates in western New York and is regulated by the PSC under article 4 of the Public Service Law. NFG Distribution is a subsidiary of National Fuel Gas Company (National Fuel) and has a number of corporate affiliates.

In the 1990s, National Fuel (the parent company) began pursuing insurance coverage for potential environmental cleanup costs at its former manufactured natural gas plants. National Fuel had commissioned an environmental report (the IES report), issued in 1996, which estimated that site investigation and remediation (SIR) expenses at the former plants would be approximately $300 million. The IES report further attributed 64% of the potential SIR liabilities to NFG Distribution. To determine the extent of its insurance coverage for these estimated remediation expenses, National Fuel filed notices of potential claims with its general liability insurance companies and provided copies of the IES report to its insurers. All of the insurance carriers initially denied coverage.

Eventually, the insurers and National Fuel reached two separate settlements in 1999, totaling approximately $37 million. All *365but one of the insurers settled with National Fuel for about $16 million. Additionally, AEGIS insurance agreed to issue a replacement policy providing $20.8 million in future SIR coverage. Since the estimated SIR liability of just one of National Fuel’s subsidiaries (such as NFG Supply) could have exceeded the total amount of the settlements, it was conceivable at the time the parties entered into the agreements that one environmental remediation claim by a single subsidiary could exhaust the entire settlement fund to the detriment of the other subsidiaries and their ratepayers.1

National Fuel therefore decided to allocate the proceeds of the settlements among its subsidiaries that had been covered by the insurance policies through the use of a “premiums paid” formula of allocation. Under this approach, each subsidiary received an amount from the settlements proportionate to its share of the insurance premiums paid and its contribution to the costs incurred in obtaining the settlements. As a result, NFG Distribution received almost 46% of the settlement proceeds—approximately $8.3 million (about 52% of the cash settlement) and approximately $8.5 million in future coverage under the AEGIS policy (about 41% of the new SIR insurance). A similar percentage of the total recovery was allotted to two other regulated National Fuel subsidiaries (NFG Supply and Pennsylvania NFG Distribution). The unregulated subsidiaries received approximately 7% of the total settlement, but their share apparently did not consist of cash proceeds, just future SIR coverage under the AEGIS policy.

Between 1998 and 2006, NFG Distribution incurred actual SIR expenses of almost $27 million—85% of National Fuel's aggregate environmental remediation costs during that period— which depleted the company’s proceeds of the monetary settlement and its share of coverage under the AEGIS policy. As a result, in 2007, NFG Distribution petitioned the PSC for tariff amendments to increase its rates in order to pass its uninsured SIR costs to its customer base. At that time, NFG Distribution was collecting $600,000 per year for SIR expenses from ratepayers. The tariff request sought to increase that amount to $1.7 million.

*366DPS challenged the requested increase, arguing that it had been unreasonable for National Fuel to use the premiums paid methodology to allocate the settlements because the percentage of premiums paid by each subsidiary bore no relation to the amount of the settlement funds. According to DPS, the settlements should have been distributed to the subsidiaries based on the actual SIR expenses incurred. DPS requested that the PSC impute approximately 85% of the total settlement to NFG Distribution, thereby reducing the proposed tariff request accordingly. NFG Distribution countered that National Fuel chose the premiums paid methodology in 1999 because the IES report provided only preliminary estimates of potential SIR expenses as of 1996; that not all of National Fuel’s former manufactured natural gas sites were included in the report; and it would have been too speculative to attempt to determine the actual SIR costs that each subsidiary would ultimately incur. Consequently, NFG Distribution urged that it was reasonable for its corporate parent to utilize the premiums paid formula at the time the settlements were disbursed in 1999.

The administrative law judge (ALJ) ruled in NFG Distribution’s favor, concluding that the premiums paid formula was “not unreasonable on its face” since DPS had failed to demonstrate that some other settlement distribution method would have been reasonable at the time the corporate decision was made.2 The recommended decision determined that no additional portion of the settlements beyond the 46% that NFG Distribution actually received should be imputed to the company. The ALJ recommended that NFG Distribution be permitted to increase its rates to collect $1.7 million annually for environmental remediation expenses, the amount requested by NFG Distribution.

Exceptions were filed to the ALJ’s decision and the matter was brought before the PSC. DPS continued to assert that it was unreasonable for National Fuel to have employed the premiums paid method and asked the PSC to impute 85% of the settlements to NFG Distribution. In the alternative, DPS argued that the settlements should have been distributed to the subsidiaries based on the percentage of SIR costs that were attributable to them in the IES report, which method would have *367resulted in 64% of the settlement proceeds directed to NFG Distribution. During the evidentiary hearing, a DPS employee testified that it had been unreasonable to allocate the settlements based on premiums paid because the claims that were presented to the insurance carriers were premised on costs associated with specific sites and the amount of settlement proceeds “was not related in anyway [sic] to the insurance premiums paid.” According to the DPS employee, the settlements were “based on the estimated remediation costs, and presumably other litigation factors, which had no relation to the amount of insurance premiums paid.”

The PSC concluded that National Fuel had acted imprudently, finding that “the 46% allocation of the insurance proceeds was unjust and unreasonable at the time it was made” because National Fuel “should have taken into account the estimates that were available at the time of the liabilities that each subsidiary company was facing.” The PSC determined that the proper allocation to NFG Distribution in 1999 should have been 64% of the settlements and ordered that the company be imputed with an additional 18% of the settlements in developing the proper rate structure.3

NFG Distribution then commenced this CPLR article 78 proceeding to contest the PSC’s determination. The case was transferred to the Appellate Division, which annulled the PSC’s imputation of additional settlement proceeds to NFG Distribution (71 AD3d 62 [3d Dept 2009]). The court held that it was reasonable for National Fuel to use the premiums paid formula in 1999, reasoning that the IES report contained only preliminary estimates of SIR costs, it did not address every potential claim of former manufactured gas sites and there was no evidence in the record to demonstrate that the report was intended to reflect all accurate SIR claims when it was drafted in 1996 or when the settlements were allocated in 1999. The Appellate Division also observed that the premiums paid formula allowed National Fuel to reasonably compute each subsidiary’s share of the insurance premiums and the methodology assured that each subsidiary and its ratepayers would receive the benefit of the settlements in proportion to what each subsidiary had paid for insurance coverage. Although the court agreed with the PSC that it would have been a reasonable alternative for National *368Fuel to have allocated the settlements in 1999 using the potential claims that were listed in the IES report, the Appellate Division determined that the record did not support the PSC’s finding that the premiums paid formula was unreasonable. The Appellate Division therefore concluded that the PSC erred by imputing additional settlement proceeds to NFG Distribution beyond the 46% that it actually received.

We granted leave to appeal (14 NY3d 709 [2010]) and now affirm.

II

Generally, a deferential standard of review applies to PSC orders because establishing utility rates is a “highly technical” matter (Matter of Abrams v Public Serv. Commn. of State of N.Y., 67 NY2d 205, 211-212 [1986]) and the PSC possesses specialized knowledge and expertise in ratesetting matters (see Matter of New York Tel. Co. v Public Serv. Commn. of State of N.Y., 95 NY2d 40, 48 [2000]). Judicial review is therefore limited to determining whether record evidence provides a rational basis for a PSC order (see Matter of Rochester Tel. Corp., 87 NY2d at 28-29).

But it is also a bedrock principle of administrative law that a “ ‘court, in dealing with a determination . . . which an administrative agency alone is authorized to make, must judge the propriety of such action solely by the grounds invoked by the agency’ ” (Matter of Scherbyn v Wayne-Finger Lakes Bd. of Coop. Educ. Servs., 77 NY2d 753, 758 [1991], quoting Matter of Montauk Improvement v Proccacino, 41 NY2d 913, 913 [1977]; see e.g. Matter of Consolidated Edison Co. of N.Y. v Public Serv. Commn., 63 NY2d 424, 441 [1984]). If the reasons an agency relies on do not reasonably support its determination, the administrative order must be overturned and it cannot be affirmed on an alternative ground that would have been adequate if cited by the agency (see Matter of Scherbyn, 77 NY2d at 758).

Although both of these principles are relevant here, this case is not the typical appeal of a PSC order. In assessing the requested tariff amendments to increase NFG Distribution’s rates, the PSC first had to consider whether National Fuel’s settlement distribution formula was “prudent” when it was devised in 1999 (see Matter of Crescent Estates Water Co. v Public Serv. Commn. of State of N.Y., 77 NY2d 611, 617 [1991]). A utility’s decision is prudent if it acted reasonably based on the information that it had and the circumstances that existed at *369the time (see Matter of Long Is. Light Co. v Public Serv. Commn. of State of N.Y., 134 AD2d 135, 143-144 [3d Dept 1987]). A decision may be viewed as prudent even though a different course of action would ultimately have been more advantageous to the utility or its ratepayers. In this regard, hindsight is irrelevant to a prudence analysis because the utility must make a determination that addresses its business prospectively. Thus, if more than one course of action was reasonable at the time of decision making, the utility may choose among them. The PSC cannot overturn a prudent decision by a utility because it believes that another course of action would have been preferable.

The usual burdens of proof are also slightly different in this case. A utility company seeking a rate change has the burden of proving that the requested regulatory action is “just and reasonable” (Public Service Law § 66 [12] [i]; see id. § 72; 16 NYCRR 61.1; Matter of St. Lawrence Gas Co. v Public Serv. Commn. of State of N.Y., 42 NY2d 461, 464 [1977]). However, a utility’s decision to expend monetary resources is presumed to have been made in the exercise of reasonable managerial judgment (see Matter of Long Is. Light Co., 134 AD2d at 144). DPS carries the initial burden of providing a rational basis to infer that the utility may have acted imprudently before the burden shifts to the utility to demonstrate that its decision was prudent when made (see Matter of New York Tel. Co. v Public Serv. Commn. of State of N.Y., 190 AD2d 217, 221 [3d Dept 1993]; Matter of Long Is. Light Co., 134 AD2d at 144).

Based on these principles, the issue before us distills to whether DPS adequately raised a reasonable inference of imprudence and, if so, whether there is a rational basis in the record to support the grounds cited by the PSC for its conclusion that National Fuel acted imprudently when it used the premiums paid formula for the distribution of the settlement proceeds in 1999.

We conclude that DPS failed to meet its initial burden of rebutting the presumption of prudence. The only DPS employee to testify—a public utility accountant—opined that the premiums paid formula was unreasonable because the settlements were not procured in relation to the amount of premiums paid. In our view, this testimony did not provide a rational basis to infer that National Fuel acted imprudently or sought to maximize future recovery from utility rates. The record does not reveal what factors prompted the insurers to settle, nor does it definitively exclude as a relevant factor the amount of *370premiums collected from each subsidiary. In deciding to settle, the insurers may have factored in the total amount of premiums that they had been paid in conjunction with their maximum financial exposure, along with the likelihood that a court might have concluded that the terms of the policies extended coverage to environmental remediation costs.4 The DPS employee also did not explain why the factors that led the insurers to settle should have dictated National Fuel’s allocation method. In light of the uncertainty of coverage under the policies and the lack of definitive information pertaining to all potential SIR claims, we view the testimony of the DPS employee as too conclusory to call into question the prudence of National Fuel’s allocation decision or to shift the burden of proof to the utility.

We also note that the only ground cited by the PSC was that the premiums paid methodology was unjust and unreasonable because National Fuel could have used the IES report’s estimate of SIR subsidiary liabilities. But there are often multiple ways to reasonably address an issue that arises in a business setting and the fact that it may have been prudent for National Fuel to use the estimated liabilities method did not, standing alone, make the use of the premiums paid approach imprudent.5

In addition, the 1996 IES report supplied only preliminary estimates of potential SIR liabilities and was prepared as part of the effort to convince the various insurers to settle and *371maximize National Fuel’s recovery. Simply put, its purpose was to persuade the insurers to avoid litigation, not to determine the extent of environmental contamination with scientific certitude. Nor does the record reveal that the IES report addressed every potential SIR site and, for the sites that were reviewed, the report was not designed to definitively assess the costs of individual SIR claims. Hence, the figures included in the IES report were estimations, whereas each subsidiary’s share of the premium payments could be accurately tallied. As a result, the IES report was not the sole rational means for reasonably allocating the proceeds of the settlements.

In sum, there was no evidentiary foundation to infer that National Fuel may have acted imprudently in 1999 when it decided to use the premiums paid formula. We therefore conclude that the PSC’s finding of imprudence was erroneous as a matter of law.

Accordingly, the judgment of the Appellate Division should be affirmed, with costs.