Opinion
PETERSON, J.
In these consolidated appeals, we affirm orders of the former Insurance Commissioner (Commissioner) and trial court which (1) [1413]*1413required the appellant insurer to accept assignment of automobile insurance risks on a statewide basis; and (2) created, and thereafter suspended prospectively, a special Urban Credit Program for risk assignments in certain urban areas. These actions of the Commissioner were consistent with her statutory duties, were neither arbitrary nor capricious, and were supported by substantial evidence in the administrative record.
I. Facts and Procedural History
Before discussing the merits, we will first briefly summarize (a) the most relevant statutes and regulations governing the assignment of automobile risks, in order to provide necessary background; (b) the actions of the Commissioner which are in issue here; and (c) the procedural background of these appeals.
A. The Assigned Risk Plan
The California Automobile Assigned Risk Plan (CAARP) was originally created by the Legislature to provide insurance to “those marginal motorists who, because they were considered ‘bad risks,’ were otherwise unable to secure and maintain such insurance.” These persons included “ ‘violators of traffic laws, . . . persons with minor physical disabilities, the young and the old drivers, and, of course, those who had bad accident records.’ (Cal. State Auto. [etc. Bureau v. Downey (1950) 96 Cal.App.2d 876,] 881-882 [216 P.2d 882].) To this extent the plan complements the state’s financial responsibility laws by providing a limited fund of insurance to compensate persons injured by drivers who otherwise would be uninsurable.” (Nipper v. California Auto. Assigned Risk Plan (1977) 19 Cal.3d 35, 41 [136 Cal.Rptr. 854, 560 P.2d 743].)
Insurance Code1 section 11620 gives the Commissioner a mandate to design and implement a “reasonable plan for the equitable apportionment” of assigned risks among insurers. In general, insurers receive their “equitable” assignments of drivers based upon a ratio or quota, theoretically derived from the percentage of voluntary liability policies they write in the state. (Cal. Code Regs., tit. 10, § 2445.) After a particular driver is assigned to an insurer pursuant to a “reasonable” plan, tibe insurer is required to issue a policy to the driver. (Cal. State Auto. etc. Bureau v. Downey (1950) 96 Cal.App.2d 876, 884-885 [216 P.2d 882], affirmed sub nom. California Auto. Assn. v. Maloney (1951) 341 U.S. 105 [95 L.Ed. 788, 71 S.Ct. 601].)
Unfortunately, there are obvious inherent conflicts in this statutory scheme, which reflect the underlying problems which the Legislature sought [1414]*1414to remedy by its creation. The reason insurers would not cover certain drivers in the first place was because those drivers were not profitable to insure at any price the drivers could pay, and any insurer which voluntarily accepted these drivers in large numbers at lower rates could be driven out of business. Certain legal issues related to this conundrum are now pending before our Supreme Court (cf., e.g., California Auto. Assigned Risk Plan v. Gillespie (1991) 235 Cal.App.3d 1825 [280 Cal.Rptr. 217], review granted July 24, 1991 (S016766)); those issues are irrelevant here except to the extent that they reflect the pressures faced by the Commissioner in attempting to fashion a reasonable plan for the equitable assignment of risks.
The Legislature directs the Commissioner to mediate this inherent conflict, by forcing insurers to share among themselves, in an “equitable” and “reasonable” way, the burdens of covering these drivers at low rates. One unintended result of this statutory scheme may be that every insurer has an increased incentive to avoid covering assigned risks, and seeks to shift those assigned risks to other insurers while covering only voluntary, profitable risks; it would be the job of the Commissioner, as defined by the Legislature, to try to defeat such a strategy and devise a reasonable method of equitably distributing to insurers certain lines of business which they might prefer not to handle.
B. The Commissioner’s Orders
1. Statewide Assignment
In 1987, problems in the assignment of risks—specifically, the continuing increase in the number of assigned risks in Southern California—led the Commissioner to institute a new method of statewide, random risk assignment which is in issue here.
Prior to this 1987 action, appellant California State Automobile Association Inter-Insurance Bureau (Bureau) had not been accepting, for about five years, its allotted portion of the state’s assigned risks. The Bureau had traditionally accepted only assignments of risks from Northern and Central California, its preferred area of operations; most of the assigned risks awaiting distribution were from Southern California. As a result, the Bureau was apparently not doing its allotted part in meeting this statewide need.
Further, for unrelated actuarial reasons, insurers were losing more money on their Southern California assigned risk business than in other areas. It was concluded that the Bureau was not taking its share of this undesirable business—a situation which caused disproportionate losses to other insurers.
[1415]*1415In order to remedy these problems, in 1987, the Commissioner approved a new approach, the statewide assignment of risks. The new rules provided that the Bureau could no longer turn down Southern California risks; all the state’s assigned risk drivers would be randomly assigned to insurers in proportion to their business conducted in the state, on a statewide basis.
2. The Urban Credit Program
The goal of equitable assignment of risks also led to the instigation—and later, the abandonment—of another new approach approved by the Commissioner, the Urban Credit Program.
Certain urban areas of the state—primarily those located in the heavily urbanized areas of Southern California, though also including parts of Oakland and Berkeley—appear on an actuarial basis to be exceptionally undesirable to automobile insurers as a result of the nature of the risks involved and the alleged inadequacy of the rates which insurers are allowed to charge for coverage in those areas.
In order to deal with the problem of the equitable assignment of this undesirable urban business, the Commissioner álso authorized in 1987 the new Urban Credit Program. The gist of the program was the idea that insurers who voluntarily wrote automobile insurance business involving risks located in these specified urban areas would receive credit against their quota of involuntary, assigned risk business in the same areas.
The result of application of this new program proved, whether by intention or not, to redound strongly against the interests of appellant Bureau. Since the Bureau did most of its business in Northern California, it was unable to receive credits for doing business in the Southern California urban credit areas, although the Bureau did do some urban area business in Northern California.
Further, other insurers whose operations in Southern California were more extensive found that their quota of assigned risks could be effectively reduced to zero, while the quota of the Bureau had to be increased correspondingly. One apparently anomalous feature of the Urban Credit Program was the fact that it treated the renewal of a previously existing policy, covering a risk located in the designated area, as a new voluntary policy written in that area. Thus, the program could not effectively serve to increase the net coverage by insurers on a voluntary basis. Since most insurers could reduce their assigned risk quota to zero simply by renewing old policies, they had little incentive to write additional voluntary policies in these areas [1416]*1416and thereby reduce the need for compulsory assigned risk coverage, which had been one of the original rationales of the Urban Credit Program.
The Urban Credit Program also interacted with and exacerbated the underlying problems posed by the Bureau’s limited chosen area of operations. Even if the Bureau received all the assigned risk coverage business in the northern two-thirds of the state, it could not effectively assume its allotted quota of risks without an expansion of its traditional area of operations. In order to deal with these problems, in September 1987, the Governing Committee of CAARP approved “Motion #1” which required the Bureau to accept statewide assignments, i.e., random assignments from Southern California which were outside the Bureau’s chosen territory; and “Motion #2” which caused the Bureau’s assigned risk quota to increase by three percentage points per month in order to reduce the backlog of underassignments to the Bureau. The Commissioner subsequently approved and ratified these actions.
The interaction of the Urban Credit Program and the statewide random assignment of risks approved by the Commissioner, thus, caused the Bureau, which had previously limited its business to the northerly portions of the state, to suddenly begin to receive large numbers of assignments from actuarially undesirable urban portions of Southern California. The Bureau contends that it thereby suffered an actuarial loss in the tens of millions of dollars.
C. Procedural History of This Litigation
Although the procedural history of this litigation is complex and confusing, it may be briefly summarized as follows.
1. Gillespie I
In November 1987, the Bureau filed its first mandamus action (Gillespie I) in the San Francisco Superior Court. After certain proceedings which are not directly relevant here, in March 1988, the superior court issued a peremptory writ, directing the Commissioner to hold public hearings on “Motion #1,” on “Motion #2,” and on the Urban Credit Program. The implementation of these administrative actions was also stayed by the superior court, pending the Commissioner’s holding of public hearings directed to those issues.
In accordance with the order of the superior court, in June 1988, the Commissioner held public hearings directed to the issues specified by the superior court; in September 1988, the Commissioner issued a final decision, in light of the evidence presented at those hearings.
[1417]*1417The Commissioner’s decision indicated that the objections of the Bureau and others to the implementation of the Urban Credit Program had sufficient merit to require that this program should be immediately suspended, pending further study. However, the Commissioner also ruled that it was necessary to implement statewide assignment of risks, the subject of “Motion #1,” as soon as possible.
2. Gillespie II
In response to this decision, the Bureau filed another mandamus action (Gillespie II) in the San Francisco Superior Court. The gist of this action was (1) the Bureau’s claim that there was insufficient evidence in the record at tiie June 1988 administrative hearing to show the Commissioner’s order requiring statewide assignment of risks was in compliance with section 11621, which required that risk assignments be “consistent” with an insurer’s scope of territorial operations, “[ijnsofar as possible”; and (2) the Bureau’s contention that, even though the Commissioner had prospectively suspended the Urban Credit Program, the prior assignment of risks based upon that defunct program was improper.
In February 1989, after further proceedings, Judge Poliak ruled that there was insufficient evidence in the record to support the Commissioner’s ruling allowing statewide assignment of risks. The superior court remanded to the Commissioner for further proceedings on that point, without ruling on the other issue pertaining to the Urban Credit Program.
In March 1989, the Commissioner held yet another hearing on statewide assignment of risks; and in November 1989, the Commissioner issued yet another ruling approving this method of risk assignment.
3. Gillespie III
The Bureau then filed its ultimate mandamus action in the superior court (Gillespie III), challenging the statewide assignment of risks. The Bureau also sought a ruling on the issue remaining in dispute and not yet ruled upon in Gillespie II, relating to the prior assignment under the suspended Urban Credit Program.
After further proceedings, in April and May of 1990, Judge Brown ruled against the Bureau on these issues, entering judgments against it and dissolving the stay which had remained in effect during the pendency of litigation.
The Bureau timely appealed from the judgments against it; we subsequently ordered the appeals from Gillespie II and Gillespie III consolidated in [1418]*1418these proceedings, and denied the Bureau’s request for a stay pending appeal.
II. Discussion
We affirm in all respects. The statewide, random assignment of automobile insurance risks was consistent with the Commissioner’s statutory mandate, not arbitrary or capricious, and supported by substantial evidence. (Cf., Bixby v. Pierno (1971) 4 Cal.3d 130, 148-149 [93 Cal.Rptr. 234, 481 P.2d 242].) “We conclude that the trial court in the present case properly reviewed the commissioner’s approval of the . . . plan to determine whether substantial evidence supported the commissioner’s determination, [f] The trial court reviewed the entire administrative record and found substantial evidence to support the commissioner’s lawful determination that the [plan in question] was ‘fair, just and equitable.’ Our scope of review on appeal from such a judgment is identical to that of the trial court [Citations.] Our review of the record discloses ample evidence to support the commissioner’s decision and the trial court’s judgment.” (Ibid.., fn. omitted; accord, County of Alameda v. Board of Retirement (1988) 46 Cal.3d 902, 906-910 [251 Cal.Rptr. 267, 760 P.2d 464].)
Moreover, although the Urban Credit Program—which has since been suspended indefinitely—presents a closer question, we ultimately determine under the deferential standard applicable here that this program too was within the ambit of the Commissioner’s powers, and was not invalid from its inception. (See Bixby v. Pierno, supra, 4 Cal.3d at pp. 148-149.) Hence, although the Urban Credit Program has been suspended, the previous assignments made to the Bureau under this scheme remain valid.
A. Statewide Assignment
Appellant Bureau contends the Commissioner’s order requiring statewide random assignment of risks violates the principle stated in section 11621, which provides, in relevant part: “Insofar as possible, assignments under the plan shall be consistent with the scope of territorial operations, and underwriting policies of each [insurer].” (Italics added.) We disagree with the Bureau’s contention for two reasons. First, the principle stated in section 11621 is explicitly made subject and dependent to considerations of practicality by an introductory disclaimer (“[i]nsofar as possible”); here, the Commissioner’s conclusion that statewide assignment was necessary is supported by substantial evidence, and neither arbitrary nor capricious.
Second, this result is also supported by the recent amendment of the statutory scheme by the addition of new provisions to section 1861.02 (Stats. [1419]*14191989, ch. 1099, § 2), and by one such provision which requires that the Bureau do business on a statewide basis in any event (§ 1861.02, subd. (b)(3)(C)(i)), as we discuss below.
Moreover, as a result of an evidentiary record compiled at public hearings mandated by the trial court, the Commissioner concluded that “No reasonable alternative had been proposed that equitably apportions risks among CAARP participants while still allowing for consideration of operating territory [as limited by section 11621 and the Bureau’s desires].” The Commissioner, thus, properly satisfied the requirements of section 11620 and section 11621, by attempting to find a method of risk assignment which will meet the two mandatory considerations—reasonableness and equity— which are specified by section 11620, while also attempting to satisfy the Bureau’s preference as to its operating area, which was not a mandatory consideration but a desideratum specified in section 11621. (Cf., Nipper v. California Auto. Assigned Risk Plan, supra, 19 Cal.3d at pp. 42-43 [“It is apparent that the Legislature prefaced consideration of [certain matters] with discretionary and permissive language while retaining . . . clear mandatory language relating to [other] elements which a plan must contain. . . . Had the Legislature intended otherwise, there was readily available language to accomplish such a purpose. It was not used.”].)
The Commissioner’s conclusion that the Bureau’s wishes could not be allowed to govern to the exclusion of reasonableness and equity in risk assignment was also supported by the evidence. There was ample evidence presented that the Bureau could not be assigned its required quota if its assignments were limited, as it wished, only to Northern California. The Bureau’s underassignment meant that a part of its allotted risks had to be parcelled out to other insurers as additions to their quota—a result which is not reasonable or equitable. Moreover, the Bureau did not present any evidence that it would be unduly burdensome to it to service assigned risks in Southern California, either through independent adjustors or the Bureau’s own long-standing arrangement for this purpose with the Southern California Auto Club.
Further, the 1989 amendment of section 1861.02 provides in pertinent part that “membership in a motor club [i.e., the precondition to coverage by the Bureau] may not be based on residence in any area within the state.” (§ 1861.02, subd. (b)(3)(C)(i).) Thus, months before statewide assignment finally went into effect, the Bureau was forbidden from limiting its operations so as to exclude Southern California drivers as it had wished. This specific statutory language would obviously control over the prior hortatory language of section 11621 which the Bureau relies upon.
[1420]*1420We find no infirmity in the Commissioner’s decision which required the Bureau to accept statewide random risk assignments.
B. The Urban Credit Program
The Urban Credit Program presents a closer question. This initiative was problematical from its inception and has since been abandoned by the Commissioner. However, the assignments made during the window period in which the program was in effect remain with the Bureau. The Bureau contends the program was legally void, and the assignments given the Bureau should be abrogated as well. Although the issue presented is in some ways a troubling one, we ultimately conclude the Commissioner acted properly.
First, it is clear that the Urban Credit Program was conceived as a response to a valid subject of concern—the situation faced by insurers operating in primarily urban areas of Southern California, where the costs of operations were causing insurers to suffer large losses. As originally conceived, the Urban Credit Program might have caused the number of assigned risk drivers in these areas to decrease, by giving insurers an added incentive to cover these drivers voluntarily. For each driver from the affected area which it covered voluntarily, the insurer could reduce its quota of assigned risks by one driver as well. In theory, this would achieve the goal of the assigned risk laws by covering drivers in high risk areas, and would also tend to reduce the number of drivers who had to be assigned as risks to other insurers.
However, in the initial implementation of the Urban Credit Program certain decisions were made, perhaps by default, which caused the idea to backfire. The major problem was that, by treating each periodic renewal of an existing policy as a new policy, the Urban Credit Program essentially allowed certain insurers to exempt themselves from further risk assignments without writing any more policies in the affected areas than they had done previously. The Commissioner wisely reconsidered the program as this defect became apparent, and suspended the program indefinitely.
With the benefit of hindsight we can now say that the Urban Credit Program, as implemented, was misconceived. However, that is a far cry from saying it was an abuse of the Commissioner’s discretion or in violation of applicable law. It was not. Rather, in this statutory context, the Legislature has obviously given the Commissioner great discretion in fashioning a specific response to the problems encountered in fulfilling the overall goals of the assigned risk laws. “Much is necessarily left to the Insurance [1421]*1421Commissioner, who has broad discretion to adopt rules and regulations as necessary to promote the public welfare.” (Calfarm Ins. Co. v. Deukmejian (1989) 48 Cal.3d 805, 824 [258 Cal.Rptr. 161, 771 P.2d 1247].) The Commissioner’s powers “are not limited to those expressly conferred by statute; ‘rather, “[i]t is well settled in this state that [administrative] officials may exercise such additional powers as are necessary for the due and efficient administration of powers expressly granted by statute, or as may fairly be implied from the statute granting the powers.” ’ ” (Id. at pp. 824-825, italics in original.) In light of the broad mandate given by the Legislature to the Commissioner, the courts must necessarily recognize her “broad discretion” to enact specific measures which are consistent with the goals of the statutory scheme.
We also understand the Bureau’s contention that the existing problems of risk assignment, which led to both statewide assignment of risks and to the establishment of the Urban Credit Program, might perhaps be traced to the regulatory process itself—and to the provision of arguably inadequate rates for assigned risk automobile insurance. However, the question of the adequacy of statewide insurance rates is not the proper subject of this litigation, and is pending in another forum where all the parties affected may present their positions. We will not reach out to address that issue here, even though the Bureau contends we could. It is clear the Commissioner could properly determine that the Urban Credit Program, if properly implemented, might be a useful method of dealing with the actual problems of risk assignment, all other things being equal. When evidence and experience disproved this theory, the Commissioner properly cancelled the program. These actions were within the scope of the broad powers delegated by the Legislature to the Commissioner, and the program was not void.
However, we have certain additional concerns in this case. Here, the Commissioner’s ultimate ruling on the Urban Credit Program was to suspend it prospectively, but not to cancel the previous assignments made under the program. We requested additional briefing of this issue from the parties, in light of authorities on the question of the ability of an administrative agency to take corrective action which is purely prospective, without cancelling the negative effects on a party “previously affected by” the newly abrogated ruling. (Cf., e.g., Securities Comm’n v. Chenery Corp. (1947) 332 U.S. 194, 202-203 [91 L.Ed. 1995, 2002, 67 S.Ct 1575]; N.L.R.B. v. Majestic Weaving Co. (2d Cir. 1966) 355 F.2d 854, 860-861; Laidlaw Corporation v. N.L.R.B. (7th Cir. 1969) 414 F.2d 99,107; McDonald v. Watt (5th Cir. 1981) 653 F.2d 1035, 1042-1044.)
Upon close review of the issue in light of the additional briefing submitted, we ultimately conclude the Commissioner also had the authority here to [1422]*1422fashion prospective relief from the unintended consequences of the Urban Credit Program, without relieving the Bureau of its already accrued obligations under that program.
A number of considerations support this conclusion. First, it is desirable to allow the Commissioner to rethink and amend programs in light of experience, without requiring the Commissioner to also undertake the monumental task of revising and recalculating all liabilities previously accrued under the prior administrative system.
We must recognize that, when the Commissioner has a legislative grant of power entailing “broad discretion to adopt rules and regulations as necessary” (Calfarm Ins. Co. v. Deukmejian, supra, 48 Cal.3d at p. 824), this also necessarily grants power to change existing rules and regulations in light of experience. A legislative grant of power to develop new rules is by definition the power to create rules having prospective application. While we might have preferred in this case that the Commissioner not only suspend the Urban Credit Program but also cancel its untoward effects on the Bureau, this is not an issue for a court to decide in the first instance.
Moreover, the Commissioner’s policy decisions here—after a full public hearing, to first enact and later suspend the further operation of a new and innovative program in pursuit of her legislative mandate—would certainly implicate many of the interests sought to be protected by the doctrine of separation of powers. Thus, “the courts exercise limited review out of deference to separation of powers between the Legislature and the judiciary, and to the presumed expertise of the agency within the scope of its authority. A reviewing court will determine whether the agency acted within the scope of its delegated authority, whether it employed fair procedures, and whether its action is arbitrary, capricious, or lacking in evidentiary support. [Citations.] A reviewing court will not substitute its policy judgment for the agency’s in the absence of an arbitrary decision . . . .” (Western Oil & Gas Assn. v. Air Resources Board (1984) 37 Cal.3d 502, 509 [208 Cal.Rptr. 850, 691 P.2d 606].)
Although the Commissioner obviously had the discretion to grant the relief which the Bureau requests, we may not control administrative discretion as a matter of course in mandate proceedings such as these. It is not to the courts that the Legislature granted this discretion and power. Rather, where as here the discretion in issue resides in the administrative agency, “mandate will not lie to compel the exercise of such discretion in a particular manner.” (Hunt v. Board of Chiropractic Examiners (1948) 87 Cal.App.2d [1423]*142398, 101 [196 P.2d 77].) A writ of mandate “is not a writ of right to be freely issued whenever a court disagrees with the policy of the administrative action.” (Ibid.; accord, Barrett v. Stanislaus County Employees Retirement Assn. (1987) 189 Cal.App.3d 1593, 1613-1614 [234 Cal.Rptr. 900] [Under this legal principle governing mandate proceedings, the trial court erred when it directed an agency to exercise its discretion in such a way as to grant retroactive relief to petitioners.].) Rather, as our Supreme Court has previously noted in summarizing the relevant law, “ ‘The courts have nothing to do with the wisdom or expediency of the measures adopted by an administrative agency to which the formulation and execution of state policy have been entrusted, and will not substitute their judgment or notions of expediency, reasonableness, or wisdom for those which have guided the agency.’ ” (Faulkner v. Cal. Toll Bridge Authority (1953) 40 Cal.2d 317, 329 [253 P.2d 659].)
In short, though we may disagree with the Commissioner’s decision not to grant the Bureau relief from the after-effects of the abortive Urban Credit Program, we cannot dispute her discretionary power to so act—and therefore, we must respectfully decline the invitation to second guess or abrogate her decision. The proper implementation of an agency’s rule-making or adjudicative power “ ‘lies primarily in the informed discretion of the administrative agency.’ ” (Agricultural Labor Relations Bd. v. Superior Court (1976) 16 Cal.3d 392, 413 [128 Cal.Rptr. 183, 546 P.2d 687], quoting Securities Comm’n v. Chenery Corp., supra, 332 U.S. at p. 203 [91 L.Ed. at p. 2002].)
In the present case, the provision of retroactive relief to the Bureau would also simply shift the liability for covering the risks in question to other state insurers, who would have their share of assigned risks retroactively increased. We are loath to create such a retroactive liability on the part of third parties in order to provide retroactive relief to a litigant. This creation of a new retroactive liability would obviously defeat the very purpose of providing retroactive relief. (See McDonald v. Watt, supra, 653 F.2d at p. 1044.)
Finally, we believe that where, as here, the Commissioner has the greater power, i.e., to effect changes to an administrative program in a highly regulated industry, that power also necessarily includes the lesser power to make certain changes take effect prospectively only. In the present case, the suspension of the Urban Credit Program did not require the Commissioner to recalculate all previous risk assignments between the Bureau and third parties as if the program had never existed. Since the Urban Credit Program was not void, the Bureau is not entitled to relief.
[1424]*1424HI. Disposition
The judgments are affirmed.
Smith, Acting P. J., and Benson, J., concurred.
A petition for a rehearing was denied June 12, 1992, and appellant’s petition for review by the Suprme Court was denied August 20, 1992.