Plaintiffs, a class composed of 77 municipalities which purchase water from the defendant City of Detroit, appeal from the trial judge’s December 2, 1981, order dismissing their action for breach of defendant’s contracts with them for the sale of water.
On May 1, 1976, defendant put new water rates into effect which increased the old rates by 39% for customers both inside and outside of Detroit. The rates were to remain in effect until January 1, 1981. Plaintiffs argue that the increase, as it relates to them, violates a provision in defendant’s contracts with 67 of the 77 plaintiff municipalities to the effect that the rates charged them shall always be reasonable in relation to the costs incurred by the City of Detroit for the supply of water.1 The trial consisted largely of expert testimony regarding the proper allocation between *158suburban use and intracity use of the costs of the system as a whole and the investment in the entire system; the amount of net revenue derived from the suburban users and the intracity users, respectively; and the reasonableness of the rate of return on the investment in the share of the system which serves the suburbs.
Plaintiffs argues at trial that, because the rate of return from intracity users was negative, the positive rate of return on the investment devoted to suburban use was unreasonable per se because it supported or subsidized the rates charged intracity users. The trial judge concluded in his written opinion that there is no theory justifying examination of the relative rates of return of the two classes of users and that therefore the only question was the reasonableness of the rate of return on the share of the system’s investment devoted to suburban use. The trial judge determined that the rate of return on suburban investment was between 8% and 9% and that he could not say that this rate was unreasonable where the evidence showed that regulated utilities in Michigan were allowed higher rates during the same period.
Plaintiffs’ first contention on appeal is that the trial judge failed to reach the question of whether defendant breached the "reasonable in relation to costs” provision in the contracts, but instead concluded erroneously that defendant’s power over the rates charged its suburban customers is legislative in nature. Plaintiffs’ argument is unfounded.
The trial judge did note in his opinion that the setting of rates is a legislative matter and that the courts will not interfere unless the rate is arbitrary, capricious, or unreasonable. See Detroit v Highland Park, 326 Mich 78, 92; 39 NW2d 325 (1949). The trial judge, however, cited Highland *159Park in support of his conclusion that a greater rate of return on suburban sales than on city sales is not improper.
The same reason lies behind the trial judge’s citation of Meridian Twp v East Lansing, 342 Mich 734; 71 NW2d 234 (1955). After concluding that these two cases demonstrate that there is no theory justifying a comparison of the rates of return of intercity and intracity rates, the trial judge then stated that the only question is whether defendant’s rate of return on the investment devoted to its suburban customers "is so unreasonable as to make the rate unreasonable”.
Although the parties disagreed over whether the suburban rate of return should be compared with the city rate of return, they did agree that the reasonableness of the rates in relation to costs depends upon the reasonableness of the rate of return. Thus, the trial judge specifically considered whether the rates were reasonable in relation to costs under the contracts in question.
In a related matter, defendant argues that any water rates charged by defendant to its suburban customers are reasonable as a matter of law if they fall within the minimum and maximum standards set by 1917 PA 34, MCL 123.141; MSA 5.2581. At all times pertinent to the instant case, this section provided as follows:
"Municipal corporations having authority by law to sell water outside their territorial limits, hereinafter referred to as corporations, may contract for such sale with cities, villages or townships having authority to provide a water supply for their inhabitants, but the price charged shall not be less than nor more than double that paid by consumers within their own territory. The price charged may be more than double that paid by consumers within their own territory if the *160water is delivered to a city, village or township lying outside the county within which the corporations are situated, and lying more than 10 miles beyond the territorial limits of the corporations. Any price charged that is more than double shall bear a reasonable relationship to the service rendered.”2
The uncontroverted evidence at trial showed that no plaintiff was charged as much as twice the price charged consumers within the city.
Defendant bases its contention that MCL 123.141; MSA 5.2581 mandates the conclusion that rates to suburbs which are less than double those charged customers inside the city are reasonable as a matter of law on dicta in Meridian Twp, supra. In that case, the plaintiff argued that the rate it was being charged violated a contractual provision which stated that the "rates shall always be reasonable in relation to the costs incurred by the city for the supply of water”. The majority of the Court stated:
"We note in passing, that the Michigan legislature has clearly, with respect to cities and villages, resolved the competing considerations of exhorbitant rates versus fair profits in CL 1948, § 123.141 (Stat Ann § 5.2581), by providing that the rate charged such outside municipalities shall not be more than double the rate paid by consumers within their own territory. This statute, even if applicable, as to which I express no opinion, has not been offended. The rate charged is not alleged to be greater than double that charged East Lansing consumers within their own territory. * * *
*161"The burden of proof was on the plaintiff to show that the rates charged were, in fact, unreasonable with relation to costs. It has not done so. I find nothing unreasonable in the charges which the city seeks to make, charges which are permissible under the contract, which are well within good accounting practice, and which are, in fact, less than the maximum authorized by the legislature for similar fringe areas.” 342 Mich 748, 753.
The Court did not hold that the statutory provision would have taken precedence over the contractual provision had the statute applied to townships. We conclude that the statute does not render reasonable as a matter of law rates within its maximum and minimum provisions in the face of a contractual provision which states that rates shall be reasonable in relation to costs. Regardless of how the statute reads, defendant has limited its discretion in setting rates by agreeing to the contractual provision.
Thus, the sole question is whether the rates charged were unreasonable in relation to costs. The answer to that question, as recognized by the trial judge, is controlled by the determination of the reasonableness of the rate of return on the investment in the share of the system devoted to suburban use.
Plaintiffs argue that the rates defendant charged them were unreasonable in relation to costs because they did not reflect the 3.3 million dollars to 3.6 million dollars of nonoperating income generated annually during the period in question. This income apparently consisted primarily of interest earned on idle bond funds and unused reserves for capital improvement. Plaintiffs argue that defendant breached its contracts with them to the extent of the amount of this interest and that there*162fore they are entitled to an award or credit for that amount.
At trial, the primary disagreement regarding this nonoperating income was whether it should be used in arriving at the total amount of revenue derived from the share of the investment in the system devoted to the city. Defendant’s expert argued, in essence, that the nonoperating income should be allocated to the city because such income is the property of the owner of the system. Plaintiffs’ experts, one of whom was an expert in ratemaking and the other an accounting expert, both testified that the nonoperating income should not be credited as revenue to either the city or the suburbs. Plaintiffs’ counsel argued for the first time at closing argument that, if the nonoperating income is going to be used to add to revenue, then the suburbs should get credit for the interest that was earned on the amounts contributed by defendant’s suburban customers.
The trial judge ruled that the nonoperating income should not be added to revenues from intracity sales. In addition, the trial judge, in his computations of the rate of return on the investment used to provide suburban service, adopted plaintiffs’ experts’ treatment of the nonoperating income by not adding net income to revenue at all.
Plaintiffs argue for the first time on appeal that they are entitled to an award of the amount of the nonoperating income for each year the rates in question were in effect because defendant breached the contracts’ cost standard by "requiring payments that have such a non-service related, built-in deprivation-of-earnings 'cost’ ”. By failing to raise this issue below, plaintiffs have waived it on appeal. Sowels v Laborers’ International Union of North America, 112 Mich App 616, 623; 317 NW2d *163195 (1981). In any event, the trial judge’s treatment of the nonoperating income was proper and, in fact, accorded with plaintiffs’ experts’ opinions on the issue.
Before proceeding to the question of the reasonableness of the rate of return on the share of the investment devoted to the suburbs, it is necessary to deal with plaintiffs’ primary argument at trial. Plaintiffs contended, and their ratemaking expert John B. Gillett testified, that the rates charged the suburbs were unreasonable per se because they were set sufficiently high to subsidize a negative return on the share of the investment in the system devoted to the city.3 The trial judge specifically found, and we agree, that the experts’ studies leave no doubt that the application of the challenged rates produced negative rates of return on intracity sales. The trial judge also concluded, however, that there is no theory justifying the comparison of the rates of return for the two classes of users and that the reasonableness of the rate of return on suburban investment would determine whether the rates charged were reasonable in relation to costs.
The Supreme Court has held, in a case dealing with the reasonableness of the rate charged Highland Park for sewage treatment and disposal, "The mere fact that there is a differential between the rate charged to the residents of Detroit and the rate charged to the adjacent municipalities does not make the latter rate unreasonable.” Detroit v Highland Park, supra, p 101. Similarly, in Meridian Twp, supra, the Supreme Court upheld the water rate charged the township under a contrac*164tual provision essentially identical to the provisions in the instant case, despite the fact that the rate was set at an effective level of 187.5% of the rate charged intracity users. See also Oakland County v Detroit, 81 Mich App 308; 265 NW2d 130 (1978).
Mr. Gillett acknowledged that Michigan case law sanctions differential rates of return, but distinguished these cases from the situation where the rates charged suburban users are set high enough to subsidize another class of users. Plaintiffs’ position seems to be supported by the following statement made by the Supreme Court in Highland Park, supra, p 101: "Plaintiff may not charge the adjacent municipalities a rate sufficient to pay part of the cost of furnishing service to its own residents * * The proposition that one class of users of a particular utility cannot be made to subsidize another class is also supported by case law of other jurisdictions. See Northern States Power Co v City of St Paul, 256 Minn 489; 99 NW2d 207 (1959); State of North Carolina ex rel Utilities Comm v City of Greensboro, 244 NC 247; 93 SE2d 151 (1956).
The rate of return on suburban investment, however, may be reasonable despite the presence of a negative rate of return on the city investment. This situation may result where the rates for the system as a whole were set too low to provide for a reasonable rate of return on the investment in the entire system.
Thus, the crucial question in determining whether the rates charged by defendant to its suburban users were reasonable in relation to their costs is whether, based on accepted methods of determining and evaluating rates of return, the rate of return on the share of the investment in *165the system devoted to suburban use was reasonable. Similarly, the primary issue on appeal is whether the trial judge erred in ruling that the rate of return on the share of the investment in the system devoted to suburban use was not unreasonable. The analysis of this issue depends in large part upon the testimony at trial and upon the method used to evaluate the rate of return.
Plaintiffs introduced an exhibit compiled by Mr. Gillett, their ratemaking expert, which represented his conclusion that, based on the revenue and expense projections compiled by defendant in 1975 and used in setting the water rates at issue in the instant case, the rate of return on the net capital investment related to suburban service for a 1977 test year was 9.8%. Plaintiffs also introduced exhibits which represented Mr. Gillett’s rate of return computations for fiscal years 1977, 1978, and 1979, based on actual revenues and expenses derived from the system for those years. Mr. Gillett felt that the rate of return on the investment related to suburban service was 9.76% in 1977, 9.04% in 1978, and 9.48% in 1979. 1979 was the last year for which actual figures were available at the time of trial.
During the testimony of Mr. Howard J. Lobb, defendant’s ratemaking expert, defendant introduced Mr. Lobb’s computations for fiscal year 1978 based on actual revenues and expenses and for a test year of 1979. Mr. Lobb concluded that the rate of return on suburban investment was 5.96% for the 1979 test year and 6.55% for fiscal year 1978.
Mr. Gillett was asked on cross-examination whether the 9.8% rate of return on the investment devoted to suburban use for the 1977 test year arid 9.48% based on actual revenue and expenses in 1979 were unreasonable when viewed without re*166gard to the negative intracity rates of return. Mr. Gillett, although he qualified his answers by adhering to his opinion that the suburban rates which make up for the intracity deficits are unreasonable per se, testified that the reasonableness of the rate of return depends on the capital structure of the investment devoted to suburban use and on what it would take for the utility to attract capital. Mr. Gillett then testified that he did not know whether 9.48% was a reasonable rate of return because he was not aware of the capital structure or of what it would take for the utility to attract capital.
Mr. Jan Chapski, an accounting expert called by plaintiffs, testified that, to determine a fair rate of return, one must ascertain the capital structure and then analyze the components of that capital structure, such as bonds, embedded costs, and the proper return on equity. Mr. Chapski testified that one way to ascertain the proper return on equity is to compare earnings of various organizations of comparable size. Mr. Chapski then testified that the overall rates of return (as opposed to return on equity) for various regulated utilities during 1977, 1978, and 1979 were 9% to 11%.
The testimony of defendant’s expert, Mr. Lobb, also indicates that the reasonableness of a particular rate of return depends on what is required for the city to attract capital and on the outstanding bond indebtedness. Mr. Lobb testified that 9% was reasonable for the test year because the municipal bond market was at a peak at the time defendant was determining the rates.
The trial judge found that the evidence showed that investment returns in excess of 10% were usual and that average interest rates on municipal bonds varied from 5.4% to 7.25% in the period *167from 1977 through 1979. The trial judge concluded that "the Court cannot say that the rate of return on extra-city sales shown herein, something between 8% and 9%, was unreasonable when the evidence showed rates in excess thereof being allowed to regulated utilities in Michigan during the same period”.
Plaintiffs argue on appeal that the trial judge erred in finding that the rate of return was not unreasonable. Plaintiffs base their argument on the contention that, as demonstrated by the cost of capital method of ascertaining rate of return, the trial judge allowed defendant a clearly impermissible 17.6% return on the equity in the suburban portion of the total investment in the system. Defendant counters this argument with the contention that plaintiffs failed to argue below that the trial judge allowed defendant an excessive rate of return on equity and that there is no evidence supporting the cost of capital method of computing the rate of return on the investment devoted to suburban customers.
The cost of capital method is a widely accepted method of determining the overall rate of return to be allowed a public utility. See Michigan Bell Telephone Co v Public Service Comm, 332 Mich 7, 39; 50 NW2d 826 (1952); Washington Gas Light Co v Public Service Comm of Dist of Columbia, 450 A2d 1187, 1209 (DC App, 1982).
To calculate the overall rate of return, the costs of each component of capital are weighted according to the ratio each bears to the total capital structure of the utility and the resultant figures are added together to yield a sum which represents the overall rate of return.4 See Washington *168Gas Light Co, supra, p 1209, fn 30; Application of Hawaii Electric Light Co, Inc, 594 P2d 612, 618 (Hawaii, 1979). In the public utility situation, capital structure generally consists of debt, which is often divided into short-term and long-term debt, preferred stock, if any, and common stock, which is referred to as equity or common equity. See Michigan Bell Telephone Co, supra, p 39; Washington Gas Light Co, supra, p 1209, fn 30.
Debt and preferred stock generally have a contractual cost and can consequently be determined with a high degree of precision and, therefore, do not generally supply a source of controversy in contested rate cases. Application of Hawaii Electric Light Co, Inc, supra, p 618; Sun City Water Co v Arizona Corp Comm, 26 Ariz App 304; 547 P2d 1104 (1976), vacated on other grounds 113 Ariz 464; 556 P2d 1126 (1976).
Determination of the rate of return on equity, however, is not capable of such mathematical precision and therefore provides the primary controversy in many utility rate cases. See Washington Gas Light Co, supra, p 1209; Application of Hawaii Electric Light Co, Inc, supra, p 619. In such cases, the reasonableness of the overall rate of return depends largely upon the reasonableness of the return on equity.
The return on equity must be sufficiently high to attract investors to the utility. See Bluefield Water *169Works & Improvement Co v Public Service Comm, 262 US 679, 692; 43 S Ct 675; 67 L Ed 1176 (1923); Application of Hawaii Electric Light Co, Inc, supra, p 619. Accordingly, a fair return on equity is generally determined by comparing what capital would earn in other enterprises of corresponding risks and hazards. Sun City Water Co v Arizona Corp Comm, supra.
So far as we can determine, the cost of capital method has never been used to determine the proper rate of return to be received by a municipally owned utility. Use of this method in this context would differ somewhat from the public utility situation, in that equity in the system is based on the going rate on municipal bonds rather than upon the rate of return on common stock necessary to attract investors, but the general application of the method should remain the same.
As noted previously, all the expert witnesses testified that the reasonableness of the rate of return on the suburban investment depended upon the capital structure. Although none of the experts put it into words, ostensibly because none of them were asked to, they were saying that the reasonableness of the rate of return could be evaluated by employing the cost of capital method. It was not until the rebuttal testimony of Mr. Gillett, plaintiffs’ ratemaking expert, that plaintiffs’ counsel sought to elicit evidence of the capital costs of the system. At that time, Mr. Gillett testified that the interest cost on the debt remaining on the bond issue floated in 1971 was 4.52% in 1977, 1978, and 1979. Mr. Gillett also testified that, in 1979, there was a spread of nearly 5% between the cost of the debt and the rate of return on suburban investment and that a reasonable rate has to be closer than 5% to the cost of attracting capital. *170Finally, Mr. Gillett testified that the 5% differential indicated that the suburbs might be subsidizing the intracity customers.
The trial judge did not consider the cost of the outstanding bonded indebtedness of the system in ruling that a rate of return on suburban investment of "somewhere between 8% and 9%” was not unreasonble. Instead, the trial judge seemed to base his opinion solely on the fact that regulated utilities in Michigan were allowed rates of return higher than 8% or 9% during the same period.
This was error. In comparing the overall rate of return in the instant case to the overall rates of return allowed regulated utilities without comparing their costs of outstanding debt and cost of attracting capital, the trial judge may well have sanctioned an unreasonably high rate of return on the equity on the suburban investment.
The trial judge’s errors warrant reversal of his decision. There are simply too many unanswered questions, however, to grant plaintiffs their requested relief of a refund or a credit for an amount representing the difference in the amounts paid by the suburbs under the alleged 17.6% rate of return on equity and an allegedly reasonable rate of return on equity of 9%.5 Accordingly, we must remand the case to the trial court to allow the court to make several determinations.
Normally, the reasonableness of the rate of return set by a regulatory body using the cost of capital method is determined by a court before the rates go into effect. In cases like the instant case, *171the cost of capital method will have to be used to work backwards from the actual overall rate of return on suburban investment to arrive at the actual rate of return on the equity in the suburban system. This is the case because this lawsuit involves not the reasonableness of rates which are not yet in effect, but instead the reasonableness of rates charged over a period of time that has already passed.
Thus, the trial judge must, on remand, first ascertain the actual overall rate of return for each of the fiscal years the contested rates were in effect. The trial judge must then determine, for each of the years in question, the debt-to-net-investment ratio of the share of the system devoted to suburban use6 and the actual rate of return, or cost, of the debt. The actual rate of return on equity can then be determined by working backwards from the actual overall rate of return.7
After ascertaining the actual rate of return on *172equity for each fiscal year in question, it will then be necessary to determine the reasonableness of that rate of return. This can be done by comparing the rate of return on equity in the instant case with the rates of return on equity allowed other utilities with similar credit ratings during the same periods, and by ascertaining for each year the going interest rates on municipal bonds issued by municipalities with credit and bond ratings similar to the ratings of the defendant during the years in question. Rates of return on equity in extra-city shares of other water systems owned by municipalities with credit and bond ratings similar to the ratings of the defendant during the period in question would be especially useful as a guide in evaluating the reasonableness of the rate of return in the instant case.
Finally, should the trial judge conclude that the rate of return on equity is unreasonable, the judge will have to ascertain what reasonable rates of return on equity would be, as it appears that plaintiffs’ remedy for unreasonable rates depends on the difference in cost between the rates as actually charged and reasonable rates. Although such a determination is akin to setting a rate, a function eschewed by the Supreme Court in Meridian Twp, supra, p 752, the determination is necessary if plaintiffs are to be provided with a remedy for breach of contract. Thus, this case differs from Meridian in that the Meridian Court found that the rates charged in that case did not violate the reasonable-in-relation-to-costs contractual provision.
We acknowledge that our resolution of this appeal will probably require an extensive evidentiary hearing and introduction of proofs not introduced at trial. Our disposition of this appeal, however, is necessary to achieve a reasoned result.
*173Reversed and remanded for proceedings in accordance with this opinion. We do not retain jurisdiction.
No costs.
There are several different contractual provisions among the 10 municipalities whose contracts do not contain the reasonable-in-relation-to-costs provision. Those plaintiffs with contracts that provide that their rates shall be comparable to or the same as other cities’ rates but not more than double than the rate paid by Detroit residents stand in the same shoes as the 67 whose contracts contain the reasonable-in-relation-to-costs provision. Those municipalities, however, with contracts that provide for rates as set by Detroit’s Board of Water Commissioners or that state that rates shall be no more than double the rate paid by Detroit residents may only recover if in fact their rates were more than double those paid by Detroit residents. See MCL 123.141: MSA 5.2581: see infra.
1981 PA 89, effective July 2, 1981, amended MCL 123.141; MSA 5.2581. The section as amended no longer contains the double limit but instead provides that the price a city charges "its customers shall be at a rate which is based on the actual cost of service as determined under the utility basis of rate-making”.
Plaintiffs take somewhat of a different approach on appeal, arguing that the parties did not intend that there be a negative rate of return on intracity investment and that therefore defendant breached the contracts to the extent of the negative intracity rate of return.
The following is an example taken from Washington Gas Light Co, supra, p 1209, fn 30, of the cost of capital method of ascertaining overall rate of return:
*168Type of Capital Proportion of Capital Cost Rate Weighted Cost
Debt 48.3% 7.04% 3.40%
Preferred Customer 12.5 6.59 .82
Deposits Equity 0.9 38.3 6.0 13.0 .05 4.98
Overall 100.0 9.25
Plaintiffs arrived at the 17.6% figure for return on equity by assuming a debt-to-net-investment ratio for the suburban portion of the system equal to the ratio for the system as a whole. Plaintiffs also assumed an actual 9% overall rate of return, 4.52% rate of return on bonded debt, and 3.55% rate of return on long-term debt. None of these figures have been found as facts by the trial judge.
The ratio for each year does not depend, as assumed by plaintiffs, solely upon the debt-to-net-investment ratio for the system as a whole. The reason for this is that the ratio of the share of the system devoted to suburban use may be higher than the ratio for the system as a whole due to the fact that the suburban system is generally newer than the part of the system that serves the city.
For example, assuming an actual overall rate of return of 9% and a 5% cost of a debt which represents 40% of the investment in the suburban portion of the system, the method of determining the actual rate on equity would be as follows:
Type of Capital Proportion of Capital Cost Rate Weighted Cost
Overall 100% 9%
Debt 40 5% 2%
Equity 60 x
The first step is to subtract the weighted cost of debt from the overall rate of return, yielding a weighted cost of equity of 7%. The actual cost of equity can then be ascertained by dividing the ratio of equity to net investment into the weighted cost of equity, yielding in this case a figure of .117, or 11.7% actual return on equity.