MONTANA-DAKOTA UTILITIES CO., a DIVISION OF MDU RESOURCES GROUP, INC., Appellant,
v.
PUBLIC SERVICE COMMISSION OF the STATE OF NORTH DAKOTA, Appellee.
Civ. No. 870015.
Supreme Court of North Dakota.
September 29, 1987.*309 Pearce & Durick, and Steven G. Gerhart and Douglas W. Schulz, Montana-Dakota Utilities Co., Bismarck, for appellant. Argued by William P. Pearce. Appearance by Douglas W. Schulz.
Lynn L. Schloesser, Asst. Atty. Gen., North Dakota Public Service Com'n, Bismarck, for appellee.
Harold L. Anderson, Bismarck, and Finch & Viken, Rapid City, S.D., for amicus curiae U-13, IBEW.
ERICKSTAD, Chief Justice.
Montana Dakota Utilities Co. [MDU], a division of MDU Resources Group, Inc., appeals from a district court judgment affirming a Public Service Commission [PSC] order establishing rates for MDU's natural gas service to North Dakota customers. We affirm in part, reverse in part, and remand.
MDU is a public utility which provides natural gas and electric service to customers in a four-state area including North Dakota. On June 28, 1985, MDU applied for a rate increase for natural gas service to its North Dakota customers, requesting additional annual revenue of $3,435,324. Pursuant to N.D.C.C. § 49-05-06, the PSC suspended MDU's proposed rate increase pending final determination. After staff investigation and a formal hearing, the PSC issued an order authorizing a rate increase yielding additional annual revenue of $2,363,000.
Before addressing the specific issues raised by MDU, we will briefly outline *310 our standard of review on appeals from decisions of an administrative agency. When an administrative agency decision is appealed to the district court and then to this court, we review the decision of the administrative agency and not the decision of the district court. Skjefte v. Job Service of North Dakota, 392 N.W.2d 815 (N.D. 1986). Accordingly, we review the record compiled before the administrative agency, rather than the findings of the district court. Application of Zimbelman, 356 N.W.2d 99 (N.D.1984).
Our review is governed by N.D.C.C. § 28-32-19, which provides that we shall affirm the decision of the administrative agency unless we find that:
"1. The decision or determination is not in accordance with the law.
"2. The decision is in violation of the constitutional rights of the appellant.
"3. Provisions of this chapter have not been complied with in the proceedings before the agency.
"4. The rules or procedure of the agency have not afforded the appellant a fair hearing.
"5. The findings of fact made by the agency are not supported by a preponderance of the evidence.
"6. The conclusions and decision of the agency are not supported by its findings of fact."
Pursuant to that statute, our scope of review requires us to determine: (1) if the findings of fact are supported by a preponderance of the evidence; (2) if the conclusions of law are sustained by the findings of fact; and (3) if the agency decision is supported by the conclusions of law. Application of Zimbelman, supra. We have summarized our standards in making those determinations:
"1. We do not make independent findings of fact or substitute our judgment for that of the agency, but determine only whether a reasoning mind could have reasonably determined that the factual conclusions were supported by the weight of the evidence.
"2. We exercise restraint when we review administrative agency findings.
"3. It is not the function of the judiciary to act as a super board when reviewing administrative agency determinations.
"4. We will not substitute our judgment for that of the qualified experts in the administrative agencies." Skjefte v. Job Service of North Dakota, supra, 392 N.W.2d at 817-818.
With those standards of review in mind, we turn to the specific issues raised by MDU. MDU first argues that the part of the PSC order projecting residential gas sales of 9,196,010 Mcf. for 1986 is contrary to the evidence. Our assessment of MDU's argument requires a description of the interrelationship of the variables involved in rate setting, including the projection of residential gas sales.
An investor-owned public utility, such as MDU, is entitled to earn revenue that will allow it to meet its expenses and earn a fair and reasonable rate of return for its investors. Public Service Comm'n v. Montana-Dakota Utilities Co., 100 N.W.2d 140, 144 (N.D.1959); N.D.C.C. § 49-02-03; see Federal Power Commission v. Hope Natural Gas Co., 320 U.S. 591, 64 S. Ct. 281, 88 L. Ed. 333 (1944). MDU's revenue for natural gas sales is determined by multiplying the units of gas sold in thousand cubic feet (Mcf.) times the rate per Mcf. That revenue must be sufficient to meet MDU's expenses which include operating expenses, depreciation, taxes, and a fair rate of return for its investors. If MDU's revenues do not meet its expenses, a rate increase is necessary. In order to determine the amount of rate increase necessary to meet the revenue deficiency, the gas sales in a test year must be determined. The PSC permits the use of forecast test years to analyze both historical and projected expenses and gas sales. In this case MDU disputes the PSC's projected residential gas sales for the 1986 test year.
At the formal hearing evidence was presented to establish that average residential gas use per customer decreased *311 from 1977 to 1981, increased from 1981 to 1983, and decreased in 1984. In its application, MDU projected residential gas sales of 8,570,147 Mcf. to 66, 189 residential customers in 1986 in contrast to 1984 when it sold 8,664,440 Mcf. to 63,525 residential customers. MDU's projection was based on extending the 1984 decline in residential gas use per customer to 1986.
The PSC's expert witness, Dr. Larry Dobesh, disagreed with MDU's projected decline in residential gas use per customer. In support of his contention that MDU's projections for residential gas sales were too low, he developed econometric models to forecast residential gas use under alternative scenarios of rates, employment, seasonal and trend factors, and other determinants. Dobesh's econometric models projected residental gas sales from 9,976,593 to 10,765,124 Mcf.[1] Dobesh testified that an increase in the real price of gas causes a decline in use and conversely that a decrease in the real price of gas causes an increase in use. Dobesh testified that he believed, therefore, that the decline in the real price of gas in 1986 would lead to increased, not decreased, use per residential customer.
Dobesh also developed an "ARIMA Seasonal Time Series Model" to confirm his belief that MDU's projection was based on a continuation of the 1984 decline in use per customer. Dobesh stated that the ARIMA model ignored the determinants of usage and drew a trend line based on past experience without taking variable factors such as price or economic activity into account. Dobesh's "ARIMA" model projected residential gas sales at 8,352,987 Mcf.
After offering his models, Dobesh recommended a "known level" or "historical" residential gas use figure of 8,802,843 Mcf. for the test year because it showed "Mcf usage for MDU's North Dakota residential customers if we multiplied the usage per customer per day for the last known 12 months by the company's forecast of customers for 1986." While expressing confidence in his econometric models, Dobesh recommended that they not be used as a basis for projecting residential gas sales because he preferred more testing and a comparison with actual results.
The PSC agreed with Dobesh's assessment that a decline in the real price of gas would cause an increase in gas use. The PSC also concluded that Dobesh's econometric models consistently pointed to increased residential use during the test year and relied on those models and Dobesh's "known level" or "historical" residential gas use projection to arrive at a projected residential gas use of 9,196,010 Mcf. That figure was calculated by giving three times more weight to Dobesh's "known level" or "historical" usage than to the average of his econometric models.[2] Because Dobesh's "ARIMA Seasonal Time Model" relied on historical use as opposed to the determinants of usage, the PSC did not use *312 that model in calculating the projected residential gas use.
MDU argues that the PSC's projected residential gas use was arbitrary because it did not include Dobesh's "ARIMA" model. MDU also contends that the PSC ignored the limited purpose for which the econometric models were prepared and asserts that there is no evidence in the record to support the PSC's decision because Dobesh did not recommend using his econometric models as a basis for projecting residential gas use but testified that the econometric models simply tested whether MDU's projections were too high or too low. MDU argues that although the projections from Dobesh's econometric models appear in the record, his testimony refutes their validity for calculating the projected residential gas use. We disagree.
If the subject matter of a question before an administrative agency is of a highly technical nature, the agency expertise in that area is entitled to appreciable deference, and we are reluctant to substitute our judgment for that of the administrative agency on such matters. Triangle Oilfield Services, Inc. v. Hagen, 373 N.W.2d 413 (N.D.1985); Johnson v. Elkin, 263 N.W.2d 123 (N.D.1978). Projecting residential gas use to set gas rates is highly technical and involves several complex interrelated variables. The PSC's expertise in weighing those variables is entitled to deference, and if there is evidence in the record to support the PSC decision we will not substitute our judgment for that of the qualified experts in the PSC.
The PSC concluded that the real price of gas was a crucial variable in projecting residential gas use and agreed with Dobesh that the increase in the real price of gas in 1984 caused the decline in use per customer in that year. Dobesh's recommendation of a residential gas use of 8,802,843 Mcf. for 1986 was based on holding residential gas use constant despite declining prices in 1985 and 1986. The PSC determined that that decline in use per customer should not be projected to 1986 because the real price of gas would decline during that year resulting in increased gas usage. The PSC also determined that Dobesh's historical use projection and MDU's projection did not give sufficient credence to the effect of declining gas prices on projected use. Those determinations involve technical matters within the expertise of the PSC.
Given those determinations, we believe there was evidence in the record permitting the PSC to rely on Dobesh's econometric models which recognized the effect of declining gas prices on use.[3] We believe it *313 was appropriate for the PSC, in its expertise, to use Dobesh's econometric models along with his "historical" projection to calculate the projected residential gas use in the manner in which it did. Furthermore, in view of the PSC's determination that Dobesh's "ARIMA Seasonal Time Series Model" was based on past experience and did not take determinants such as price, employment, or economic activity into account, we do not believe the PSC's failure to use that model in calculating projected residential gas use was unreasonable. We conclude that the PSC decision was supported by a preponderance of the evidence.[4]
MDU next contends that the PSC erred in failing to adjust MDU's administrative and general (A & G) expense allocation because of changes known at the time of the formal hearing. A & G expenses include all salaries and expenses connected with general administration and management. In order to determine the appropriate allocation of A & G expenses for MDU's North Dakota gas rates, it is necessary to allocate MDU's A & G expenses to the electric, gas, or affiliate operations and also to each state jurisdiction.
MDU was an electric and gas production, transmission and distribution corporation until the Federal Energy Regulatory Commission approved the formation of Williston Basin Interstate Pipeline Company (Williston Basin) as a subsidiary gas production, storage, and transmission company, effective January 1, 1985. The corporate name was then changed to MDU Resources Group, Inc. and the MDU name was retained to identify the distribution division of MDU Resources Group, Inc. Thereafter MDU and Williston Basin were subsidiaries of MDU Resources Group, Inc.
As a result of Williston Basin becoming a gas production division and MDU becoming a gas distribution division, there was no actual experience regarding the appropriate allocation of A & G expenses between MDU and Williston Basin for North Dakota gas rates. In its 1985 application, MDU allocated itself 48% of its 1984 A & G expenses for determining its forecasted A & G expenses for test years 1985 and 1986. Based on that allocation, MDU forecasted A & G expenses of $2,269,000 for 1985 and $2,373,000 for 1986. At the time of the formal hearing in late 1985, MDU's books reflected A & G expenses of $1,945,000 for the first 8 months of 1985. Based on the data from MDU's books, PSC chief accountant Roy Paetzke prepared an exhibit forecasting $756,000 in A & G expenses for the last four months of 1985 for a total A & G expense of $2,701,000 for 1985, or $432,000 more than MDU projected in its application.
*314 During the formal hearing, MDU witness Donald Ball testified that if MDU's A & G expenses for the first eight months of 1985 were annualized by dividing $1,945,000 by eight and multiplying the result by twelve, the 1985 A & G expenses allocable to MDU would be about $2,928,000 or about $649,000 higher than the estimate in MDU's application. He further testified that if the 1984 results were restated to reflect the known circumstances, MDU would have included A & G expenses of about $2,837,000 for 1985 and $2,979,000 for 1986. He requested that the known changes in MDU's actual A & G expenses be considered in the PSC's rate determination.
The PSC refused to consider MDU's request, stating:
"The company did not amend the increased revenue requested in this case to reflect what it contended were increased expenses. Consequently, there are no company exhibits or workpapers to support the alleged increase nor was there any scrutiny of the dollar figure by the staff. Under these circumstances, it would not be appropriate for the Commission to base rates on the alleged misallocation."
MDU contends that because its request did not seek an increase in proposed rates or requested revenue, the PSC erred in not considering the known and measurable changes in A & G expenses. MDU asserts that it merely wanted the PSC to consider the known and measurable changes for purposes of establishing reasonable rates within the parameters of the original application. MDU argues that the information to make a determination concerning its A & G expenses was available through PSC witness Paetzke and his working papers, through MDU's rate application, and through the testimony of MDU's witnesses.
The PSC responds that the characterization of the A & G expenses as "known and measurable changes" is erroneous and that MDU did not meet its burden of proof regarding the validity of those expenses as legitimate operating expenses. The PSC does not dispute that test year data should be adjusted for known and measurable changes if the changes are shown to be reliable and certain. However, the PSC asserts that merely because the A & G expenses were carried on MDU's books in this case, does not establish that they should legitimately be used to establish MDU's North Dakota gas rates. The PSC argues that the A & G expenses carried in MDU's books could have been higher than MDU originally projected because of a misallocation of expenses between MDU and Williston Basin or because of nonrecurring expenses during the initial year of the reorganization. The PSC asserts that the methodology used for carrying A & G expenses on MDU's books was not investigated or explained, and therefore MDU did not meet its burden of proof.
Paetzke testified that prior to the formal hearing, the PSC's accounting staff examined MDU's books, including the A & G expenses. As a result of that examination, Paetzke prepared an exhibit which included a schedule of MDU's A & G expenses for the first eight months of 1985 and the projected A & G expenses for the last four months of 1985. Paetzke testified that he obtained those numbers in conjunction with his examination of MDU's books. However, Paetzke's prepared testimony, which was offered with his exhibit, contains the following qualification:
"Our examination included a review of the Company working papers as well as the tracing back of exhibit figures to the Company records previously audited by Arthur Young & Company. I have prepared an exhibit based on the results of the examination.
* * * * *
"The exhibit was prepared by me and under my direction and is true and correct to the best of my knowledge. I would like to qualify the extent of that knowledge, however, by pointing out that only the actual figures could be checked back to the historical records during the course of our examination."
Paetzke's prepared testimony reflects that during the course of his examination he reviewed only the actual figures carried on MDU's books. Paetzke's testimony *315 does not demonstrate that the actual figures and historical records were examined within the context of whether those A & G expenses were properly allocable to MDU's North Dakota gas rates and the nature of those expenses. While those A & G expenses may have been items carried on MDU's books, that fact does not, by itself, establish that those expenses were properly allocable to MDU and should be used to establish MDU's North Dakota gas rates. For example, some of those expenses may be attributable to salaries for MDU employees who perform work that is not related to MDU's North Dakota gas distribution. Without a breakdown and underlying analysis of those expenses, we cannot say that they should be used to establish MDU's North Dakota gas rates. Paetzke's qualifying statement reflects that his review did not entail that breakdown and analysis, and MDU did not present any other evidence to establish that breakdown.
Pursuant to N.D.C.C. § 49-05-06, the public utility has the burden to establish that an increased rate or proposed change of rate is just and reasonable. Based on this record and under these circumstances, we agree with the PSC that MDU did not meet its burden to establish the reasonableness of the changes in the A & G expenses. We do not believe that Paetzke's data or the changes in the A & G expenses requested by MDU represent known and measurable changes sufficient to support MDU's request for consideration by the PSC in setting North Dakota gas rates.
MDU also contends that the PSC's order regarding its employee discount constitutes an improper interference with company management. MDU allows a one-third discount to its employees on their monthly gas bills as part of their total compensation package. MDU purchases its gas from Williston Basin for approximately 80 percent of the price that it sells the gas to its North Dakota customers. Thus, the one-third employee discount results in MDU employees receiving gas service for less than MDU's cost. The PSC allowed the employee discount but directed MDU to take necessary steps when it negotiates a new labor agreement in 1987 to implement a discount level which will ensure that all employees pay for the actual cost of the gas to MDU.
MDU asserts that its employee discount is reasonable pursuant to N.D.C.C. § 49-04-07, because it is compensation in lieu of additional wages and its employees have not abused the discount. The PSC responds that the discount is not reasonable because it is less than MDU's wholesale cost of gas.
N.D.C.C. § 49-04-07, provides:
"49-04-07. Unreasonable preferences or advantages prohibited. No public utility shall make or give any undue or unreasonable preference or advantage to any particular person, company, firm, corporation, or locality, or to any particular character of traffic or service in any respect whatsoever, nor subject any particular person, firm, corporation, company, or locality, or any particular character of traffic or service to any undue or unreasonable prejudice or disadvantage in any respect. No public utility corporation, directly or indirectly, by any special rate, rebate, drawback, or other device or method, shall charge, demand, collect, or receive from any person, firm, or corporation, a greater or less compensation for any service rendered or to be rendered than it charges, demands, collects, or receives from any other person, firm, or corporation for doing a like and contemporaneous service under the same or substantially similar circumstances and conditions. Nothing in this chapter shall prohibit a public utility from entering into any reasonable agreement with its customers, consumers, or employees or from providing for a sliding scale of charges, unless the same is prohibited by the terms of the franchise or permit under which such public utility is operated. No such agreement or sliding scale shall be lawful unless and until the same shall be filed with and approved by the commission." [Emphasis added.]
The plain language of that statute manifests an intent to prohibit public utilities *316 from charging discriminatory rates or giving an undue or unreasonable preference. However, the statute provides authority for a public utility to enter into a "reasonable agreement" concerning rates with its employees.
The PSC made the following findings:
"The company's testimony indicated that cash compensation in lieu of the discount would cost the ratepayers 1½ times as much as the discount. In addition, we note that the average annual consumption of gas per employee is almost identical to that of a non-employee residential customer. This record does not support a finding that the discount leads to wasteful consumption of a finite natural resource. Furthermore, it indicates that the discount may have been more economical for ratepayers than equivalent cash compensation. Based on this record, we are reluctant to encroach upon management's perogative to use an employee discount as part of an overall compensation package."
Nevertheless, the PSC determined that "to avoid an unjustly discriminatory rate, the company, when it negotiates a new labor agreement in 1987, shall take the necessary steps to implement a discount level which will ensure that all employees at least cover the cost of gas."
In this case we do not believe the PSC's decision requiring negotiation of a different employee discount level is supported by its findings of fact. The PSC specifically found that the "record does not support a finding that the discount leads to wasteful consumption of a finite natural resource"[5] and that the discount "may have been more economical for ratepayers than equivalent cash compensation." These factual findings are supported by a preponderance of the evidence, but we do not believe they support a conclusion that the employee discount constitutes an unjust and discriminatory rate or is not a "reasonable agreement" within the meaning of N.D.C.C. § 49-04-07. In reaching this conclusion, we recognize that the overall employment compensation package between MDU and its employees is a matter left largely to the deference and judgment of management through its collective bargaining arrangement.[6]See Central Maine Power Co. v. Public Utilities Comm'n, 405 A.2d 153 (Me.1979), cert. denied, 447 U.S. 911, 100 S. Ct. 2999, 64 L. Ed. 2d 862 (1980). Accordingly, we conclude that the PSC's decision requiring MDU to negotiate a new discount level is not supported by the PSC's findings of fact and we reverse that portion of the PSC order.
We affirm in part, reverse in part, and remand for entry of an order consistent with this opinion. Statutory costs on appeal are awarded to MDU.
GIERKE, VANDE WALLE, LEVINE and MESCHKE, JJ., concur.
NOTES
[1] Dobesh's econometric models projected the following residential gas use:
Model 1 10,765,124 Mcf. Model 1a 10,392,585 Mcf. Model 2 10,682,404 Mcf. Model 2a 10,327,323 Mcf. Model 3 10,115,045 Mcf. Model 4 9,976,593 Mcf.
Dobesh outlined the following determinants for each model:
Model 1: Multiple regression with log of normalized use per customer per day as dependent variable and current rates.
Model 2: Multiple regression with log of normalized use per customer per day as dependent variable and actual rates.
Model 1a: Same as Model 1 except with proposed rates.
Model 2a: Same as Model 2 except with proposed rates.
Model 3: Same as Model 1 except no growth in employment for the forecast period.
Model 4: Same as Model 2 except no growth in employment for the forecast period.
[2] In its brief to this court, MDU calculates the average of the econometric models as 10,376,512. That average excludes Dobesh's ARIMA Seasonal Time Model. MDU then used the following formula, which gives " `three times more weight to [Dobesh's] historical usage' " than to the average of the higher econometric models, to determine the PSC's projected residental gas usage:
(3 × 8,802,843) + 10,376,512 = 9,196,260
----------------------------------
4
MDU notes that the minor discrepancy between 9,196,260 and the PSC's figure 9,196,010 is not accounted for, but is not significant.
[3] The Commission's decision relevant to this issue provides:
"Dr. Dobesh's econometric demand model under alternative scenarios of gas rates and employment show test year usage should be about 20 percent higher than the company's forecast for 1986 of 8,570,147 Mcf. Dr. Dobesh also used an ARIMA Seasonal Time Series Model which simply draws a trend line based on past experience without taking critical factors such as price or economic activity into account. That model projected residential usage of 8,352,987 Mcf, very close to the company's forecast. Thus it appears that the company's forecast is largely based on a continuation of the fall in usage per customer which the company experienced in 1984. Such `ruler' methods of forecasting ignore the structural determinants of the variable in question. We also note that the company's weather normalization methodology was characterized by Dr. Dobesh as `years behind the times.' Specifically, the company's normalization methodology does not take into account the effect on residential usage of gas price, income within its service area or appliance saturation. In view of those shortcomings we find that we can place more confidence in Dr. Dobesh's econometric model and testimony regarding projected residential usage than in the figure utilized by the company.
"Although all of his alternative models other than the ARIMA Model projected residential usage in excess of the company's projection by 1,406,446 Mcf to 2,195,000 Mcf, Dr. Dobesh recommended a far more conservative approach for this case. He expressed confidence in his modeling results, but did not recommend using them as a basis for setting rates in the current case. Rather, he suggested a residential usage projection of 8,802,843 Mcf based on the forecasted 1986 customers times 1986 billing days times the company's actual use per customer per day for May of 1984 through April of 1985. Dr. Dobesh's primary reason for suggesting reliance on historical figures rather than his projections was his concern that his model will not be validated by actual data until 1986 has passed. We appreciate Dr. Dobesh's desire to validate his models with actual data prior to recommending that the Commission implement any particular model as a basis for setting rates. However, we cannot adopt his recommendation of holding residential usage constant despite declining gas prices in 1985 and 1986.
"To do so would ignore the substantial and unrefuted evidence of his model results which consistently point to increased residential usage during the test year. We find that a more reasonable compromise for this case would be to project residential usage at a point between the historical usage figure of 8,802,843 Mcf and the average usage figure forecast by the econometric model of 10,376,512 Mcf. In doing so, however, we will give considerably more weight to the figure based on historical usage than to the figure forecast by the econometric model. This approach allows us to take into account the evidence of increased residential gas usage in the test year but also recognizes that it would be premature to base rates primarily on the econometric model at this point."
[4] MDU also contends that the PSC's failure to base its decision on evidence presented at the formal hearing and its failure to provide MDU with an opportunity to cross-examine and present rebuttal evidence at a further hearing denied MDU its right to due process of law in violation of the Fourteenth Amendment of the United States Constitution and Article 1, § 12 of the North Dakota Constitution. Because we have concluded that there was evidence presented at the formal hearing to support the PSC's decision, we need not address that issue. In any event, a party raising a constitutional claim must bring up his heavy artillery or forego the attack entirely. Southern Valley Grain Dealers Ass'n v. Bd. of County Commissioners of Richland County, 257 N.W.2d 425 (N.D.1977).
[5] For example, N.D.C.C. § 38-08-01 relating to the Industrial Commission's authority provides in part that it is "hereby declared to be in the public interest to foster, to encourage, and to promote the development, production, and utilization of natural resources of oil and gas in the state in such a manner as will prevent waste." While we have found no specific concomitant authority for the PSC, there can be no doubt that the PSC may take into account the conservation and prevention of waste of that resource. Implicit in the use of natural gas in this state is the conservation of that finite resource.
[6] Because of our resolution of this issue, we do not consider the argument by amicus curiae that this part of the PSC order is preempted by the National Labor Relations Act, 29 U.S.C. § 157 et seq.