The main issue raised by this cause is the effect to be given to two insurance policies covering the same risk and providing that their liability with regard to that risk shall be excess insurance over other valid, collectible insurance.
Buckeye proposes two reasons why State Auto should be primarily liable for the damages arising out of the Harrell accident. The first is that, “as a practical matter, the State Auto insured was responsible for the accident.” The second is that neither policy qualifies “by its terms” as a policy of primary insurance, and, therefore, “the policy which specifically covers a vehicle and its driver” provides “the primary coverage.” Buckeye contends, in the alternative, that “the Court of Appeals did not err in the manner in which it prorated the loss between the parties.”
Buckeye’s first argument is without merit. In Maryland Cas. Co. v. Banker’s Indemnity Co. (1935), 51 Ohio *215App. 323, the Court of Appeals for Cuyahoga County found it reasonable; between two insurers, to hold thé insurer of the party involved in the accident primarily liable even though both policies attempted to avoid liability. However, we see no'reason to adopt that court’s reasoning. The liability-limiting clauses in Maryland were not the same type as’ those in the instant cause. One policy contained an excess cláüse, and the other contained an escape clause. Moreover, .the primary tortfeasor theory espoused in Maryland has been criticized by later courts and commentators because-it “ * * * represents an attempt to arbitrarily impose primary liability on one of the insurers” (Note, Concurrent Coverage in Automobile Liability Insurance [1965], 65 Colum. L. Rev. 319, 322). Even more important, the theory fails to reflect “* * * the intent of each insurer to cover the tortfeasor whether he was expressly named in its policy or was an additional insured under its omnibus clause” (Cosmopolitan Mut. Ins. Co. v. Continental Cas. Co. [1959], 28 N. J. 554, 560, 147 A. 2d 529). This last criticism is especially compelling because an analysis based on the intent of the parties had they anticipated the occurrence of. conflicting excess clauses is the best approach to such problems. (Concurrent Coverage, supra, at page 321.) We thus reject Buckeye’s first argument.
Buckeye, also contends that, where neither policy is primary “by its terms,” because each includes coverage and excess clauses for the same risk, “the policy which specifically covers a vehicle and its driver.” provides “primary coverage.” Buckeye correctly asserts that; by definition, neither ' of two policies containing mutually repugnant excess clauses is primary. Indeed, it is because neither policy is primary where two policies contain. conflicting excess clauses that courts have,had to resort to judicial formulae, including the “specific versus general” theory espoused by Buckeye, to apportion liability.1 We are *216not persuaded, however, by Buckeye’s suggestion that we hold .State Auto primarily liable for the risk covered in the excess.clauses'on the ground that the rest of State Auto’s policy is more specific than its own. Buckeye’s argument is based on “circular reasoning.”2 In State Farm Ins. Co. v. Home Indemnity Ins. Co. (1970), 23 Ohio St. 2d. 45, 47, this court pointed out. that before one policy “can ride as excess insurance, the mother policy must be made to’ walk as primary insurance.” Since there can be no primary insurance of the risk where there are conflicting excess clauses, -the .excess clauses a fortiori cannot be a valid means of establishing only “secondary” liability. (See, also, Continental Cas. Co. v. Buckeye Union Cas. Co. [1957], 75 Ohio Law Abs. 79, 91; Cosmopolitan, supra, at-page 562.) For us to look to other provisions within the insurance policies in order to find one insurer primarily liable and give effect to an excess clause which has been neutralized, as a matter of logic, by the conflicting excess clause of the other policy, is to apportion liability on the basis of criteria unrelated to the excess clauses themselves and to run the same hazards of arbitrariness and disregard for the insurer’s intent that lead us to reject the primary tortfeasor doctrine. Therefore, we decline to accept Buckeye’s second contention.3 The Court of Appeals for Franklin *217County correetly concluded that “it was error” for the trial court to find that one party had “primar^” and the other “secondary, coverage in this case * *
Having rejected the methods of determining liability suggested by Buckeye, we must now consider what method of apportioning coverage to adopt. The Court of Appeals ruled that the “respective coverages” of the policies at issue in the present cause be “prorated.” We agree that some method of prorating coverage between Buckeye and State Auto is necessary. Proration has been adopted by a majority of jurisdictions dealing with the problem of conflicting excess clauses. Annotation, 69 A. L. R. 2d 1122, 1123-1124. Furthermore, it is consistent with the concept of primary and excess liability expressed in State Farm, supra, and Continental Cas. Co., supra, and does not conflict with any previous treatment of insurance policy clauses by this court.4 Finally, by prorating the coverage of insurers such as Buckeye and'State Auto, we can honor the expectations of policyholders that they receive protection, and we can still give at least partial effect to the insurer’s intent to escape primary liability.
The Court of Appeals held, and Buckeye asserts, that “the most equitable method of prorating is to divide the coverages equally to the limits of the lower limit policy.” We disagree. Although the Court of Appeals may have been accurate in stating that “insureds do not pay proportionate premiums for upper limits,” proration based on the cost to the policyholder has been criticized by commentators (Note, 1 Willamette L. J. 485, 490 [1961]; Case *218Notes, 11 Hastings L. J. 223, 226 [1959]) and seems irrelevant since the policyholder will receive protection under any method of proration. .......
The majority rule is to prorate liability according to the amount of coverage provided by each- insurer.5 This method of proration “assures indemnification for. the insured: up to the maximum amount of coverage afforded ¡by each policy,” and it takes into consideration the- respective! liabilities that the two insurers would have incurred had there,not been other insurance. Concurrent Coverage, supra (65 Colum. L. Rev. 319), at page 330. Wé hold, therefore, that where two insurance policies cover the same risk and both provide that their liability with regard to that risk shall be exeess insurance over other valid, collectible insurance, the two insurers become liable-in. proportion to the . amount of insurance provided by their respective policies.
The judgment of the Court of Appeals is modified, and, as modified, is affirmed.
■ Judgment affirmed.
• 0Neill,' C. J., Herbert, Celebrezze, P. Brown, Sweeney and‘LocHER, JJ., concur.The judicial formulae include the “prior in-time” theory, the “primary tortfeasor” theory, the “specific versus general” theory and thé proration theory. Concurrent Coverage, supra (65 Colum. L. Rev. 319), at pages 321-322.
Annotation, 69 A. L. R. 2d 1122, 1123; Oregon Auto Ins. Co. v. United States Fidelity & Guar. Co. (C. A. 9, 1952), 195 F. 2d 958, 960.
We choose instead to adopt the reasoning espoused by Judge Leach in Continental Cas. Co., supra (75 Ohio Law Abs. 79), at page 91, the only 'previous Ohio case dealing with two conflicting excess clauses. Starting with “the fact * * * that there can be no ‘excess' insurance in the absence of a ‘primary’ insurance,’’ the Continental court rejected “beginning” its “mental approach with one policy and then going in circles” and chose instead to reason as’ follows:
“* * * As to- Continental, its policy cannot be ‘excess’ because there is,no.t.in- effect, any other policy of ‘primary’ insurance. The ‘excess’ provision being, .inoperative and the general coverage language applicable, it is potentially liable to the full amount of its coverage to the same extent .as . if its policy contained no excess feature. Having so concluded, we do.not. say that because of such conclusion we may now apply the! ‘excess’, features of Buckeye’s policy. To do so would make *217the result depend upon which policy we considered first. Instead, and ignoring our conclusion as to the Continental policy, we consider the policy of Buckeye in the same way.”
In Trinity Universal Ins. Co. v. General Accident, Fire & Life Assur. Corp. (1941), 138 Ohio St. 488, this court aifirmed the rule that the specific insurer is primarily liable. The Trinity case is inapposite, however, because it involved a contest between an excess clause and a prorata clause, and that distinction was equally as important to the court as was the general specific coverage theory. Trinity, supra, at page 491.
See 69 A. L. R. 2d 1122, supra, at page 1124.