MEMORANDUM DECISION RE PLAINTIFFS’ MOTION FOR PRELIMINARY INJUNCTION.

The plaintiffs, an unincorporated association, and Jay Johnson, an individual acting in his own behalf and on behalf of persons similarly situated, seek to enjoin, pending the resolution of this matter on its merits, certain rental increases that have been assessed the residents of Yosemite National Park occupying living quarters owned and furnished by the United States government. The plaintiffs’ motion involves complex issues of law, and the Court has been hampered by a paucity of evidence that has been introduced by the parties that is material to the central legal issues underlying this dispute.

I

Statutory Background

In 1966, as part of the codification of the general and permanent laws relating to the organization of the government of the United States and its civilian officers and employees, the Congress passed 5 U.S.C. § 5911 , which reads as follows:

Congress did not, at any time, file any committee reports or indicate a purpose other than state that Public Law 89-544, § 1 was a part of the general effort to bring together into a single code all of the permanent laws relating to the organization of the government of the United States and pertaining to its civilian officers and employees. Significantly, this section has not been construed by any appellate court of the United States in a reported decision.

By Executive Order No. 11609 of July 22, 1971, the authority of the President under subsection (f) to issue the regulations provided for therein was delegated to the Director of the Office of Management and Budget (hereinafter OMB).

II

The Regulatory Scheme And Its Prior Implementation in Yosemite

The OMB designated Circular No. A-45, as revised, as the appropriate document setting forth the policies governing charges for rental quarters and related facilities. In 1973, the Department of the Interior adopted regulations governing the establishment of quarters rental rates. Generally, these regulations flesh out declarations of policy as contained in OMB Circular No. A-45, revised.

Appendix 1, OMB Circular A-45, provides as follows:

OMB Circular A-45 also provides that when the government furnished quarters (GFQ) were located more than five miles from an established community, basic rental rates will be set by one of two alternate methods:

(1) Comparability with the nearby representative private community. Rental rates in and adjacent to the nearby representative private community may be used as a base for establishing comparable rentals.

(2) Regional basis. Basic rental rates may be set in comparison with the average of rental rates for comparable private housing in an economically homogeneous area in which the rental quarters are located. The area selected should be large enough to permit an adequate sampling of comparable quarters (it may contain several communities), but small enough to maintain economic homogeneity and the area must be permanently defined. Extremely high and extremely low private housing rents should be excluded in computing the averagé in each classification of housing.

The Court finds that Mariposa, California, qualifies as an established community as well as a nearby representative private community.

Little evidence was offered as to the historical administrative enforcements of these statutes although some insight is provided by the declaration of George W. Sandberg, Chief of the Division of Real Property Management:

Elsewhere, the administrative record discloses that GFQ’s had previously been subjected to rental adjustments based upon the “nearby representative community” alternative contained in OMB Circular No. 45-B. The details of these previous adjustment processes were not detailed, but there was a suggestion that the community of Mariposa figured largely in the establishment of comparable rental rates for the GFQ’s at Yosemite.

III

The Methodology Employed by the Secretary To Arrive at the Rental for the GFQ in Yosemite

A. The Regional Survey

In January of 1982, a memorandum from the Secretary’s office directed all units of the Department of the Interior to proceed to implement the Congressional directive with regard to rents. Previously in California, rental increases had been determined by using individual appraisal, relying upon the use of a very small number of comparable units.

A determination was made by the appropriate officials of the Department to use the regional survey alternative authorized by OMB Circular No. A-45 revised.

The administrative justification for the regional survey is as follows:

In preparation for this, each agency of the Department of the Interior was re quested to submit the names of communities they had previously used in establishing their “comparables” for prior rental adjustments. According to Mr. Sandberg, a list of over 200 proposed comparable communities was compiled. From this list, the following communities were selected for the survey.

The contractor was instructed to omit certain classifications of housing from his survey:

First, if any community was found not to have a representative sample of any one of the classes of homes designated to be surveyed, that class would be omitted from the survey of any particular community.

Second, certain types of housing, i.e., lookouts, bunkhouses, tents, etc., were omitted entirely, on the logical grounds that such units are not customarily used outside of federally owned property.

Third, the contractor was instructed to exclude housing where the renter furnished some service, e.g., maintenance, repair, custodial service, etc.

The contractor was mandated to obtain four (4) — no more, no less — samples of each of the eleven (11) classes of housing being surveyed. The contractor was further mandated to obtain no less than fifteen (15) samples in a single community.

The survey form to be used was prescribed. The contract contemplated that approximately 1100 housing units would be surveyed, but there was no direction as to how they would be distributed among the surveyed communities.

B. The Yosemite Inventory of GFQ’s

The administrative record indicates that this survey was to be initiated by a physical inventory of GFQ's conducted by each of the Department units. Accuracy in this phase of the operation, of course, was critical to the accuracy of the end product.

The plaintiffs introduced evidence from which the Court concludes that the administration at Yosemite National Park has not been following the mandates of the Government Furnished Quarters Guideline (NPS-36) with reference to inspection of the GFQ’s in Yosemite National Park. In pertinent part, NPS-36 requires the Park Superintendent to implement a formalized procedure for the annual inspection of quarters. Further, the Superintendent is required to maintain and continually update quarters records necessary to provide an accurate quarters inventory that is a basic factor needed for the comparability surveys and regional surveys. The evidence is uncontradicted that the last time a physical inspection of all of the GFQ’s in Yosemite occurred was in 1976. None was made in preparation for this round of rental increases.

In order to rectify this deficiency, in anticipation of rent adjustments, that inspection frequently was only one unit per class of house rather than individual quarters. Using their historical “base” information, the Yosemite GFQ inventory was completed from a variety of sources, such as repair and replacement records, the familiarity of the officer-in-eharge of the Housing Office for the National Park Service in Yosemite, Mr. Vander Karr, personal telephone calls placed to some occupants at the discretion of the housing office staff, and if obvious discrepancies appeared to concern Mr. Vander Karr, a personal visit was made to the GFQ in question, ¿/'time permitted. It was this procedure which was employed by the Yosemite Park personnel to complete the inventory GFQ’s in Yosemite — this inventory being an essential and integral of the rent adjustments for Yosemite GFQ’s.

Testimony through government witnesses also established that 30% of the GFQ’s in Yosemite do not meet minimal insulation requirements. Indeed, at the time of the hearing, an insulation contractor was in the process of installing insulation in these units. The survey contractor, of course, had been directed to survey those dwelling units that were built to Housing and Development (HUD) minimum standards whenever possible.

The initial GFQ inventories received from Yosemite and other areas contained numerous errors, and requests were made for correction. The Court was provided with no information as to the nature of these errors, or what verification procedures the Department of the Interior personnel employed to assure themselves that the errors had been corrected.

C. The Establishment of the Basic Rental

After the Regional survey data was collected, categorized and fed into the computers, it was subject to a statistical process described as multiple regression procedure. It was the belief of the appropriate officers of the Department of the Interior that the use of this technique “allows the analyst to establish a basic rental rate for a given type of housing in the survey area.” In addition, it was designed to reflect, assess and adjust for the contributory values, plus or minus, of housing features that the private rental market indicates are significant, such as the identification and special treatment of relevant social and economic factors, and the impact of major recreational uses or mining and mineral operations.

The end product was a basic rental schedule on each category of housing which was furnished to each agency to compute the basic rental rate on each GFQ in its domain.

D. The Local Adjustment Process

After the basic rent for each type of housing was developed from the survey data, this material was sent to each local agency, including Yosemite National Park, to set the basic rent for each GFQ. Thereafter, certain required administrative adjustments are applied to the individual basic rent. This process, in Yosemite, went forward in fits and starts.

Factually, the following events occurred in Yosemite National Park. On or about February 6, 1983, the members of the plaintiff association and other occupants of GFQ in Yosemite were notified of the increased rental rates. This was first notification to park employees that such a procedure was being undertaken and the impact it would have upon them individually. As a result of this notification, an open meeting with the Yosemite employees was held on February 10, 1983. There is little evidence as to what occurred at that meeting.

As a result of this first meeting Superintendent Binnewies addressed a letter to his superiors in the Department requesting a review of the rental increases as they affected Yosemite. Apparently, he was dissuaded from a suggestion contained in his memorandum that the statewide survey was not equivalent to an economic homogeneous area within which the rental units (GFQ’s) were located, and instead a second meeting was set up at which representatives from the Washington Office of the National Park Service and the Regional Office of the National Park Service and the Rocky Mountain Regional Office of the National Park Service were in attendance.

After an outcry arose upon the first notification of the rental increases, the management at Yosemite Park formed a committee to evaluate the amenities available to each of the GFQ’s in Yosemite. It should be noted that this committee was not appointed and did not function until after each occupant of GFQ’s in Yosemite had been initially apprised of their adjusted rental rate.

At the time of the rentsetting process, and continuing up until the time of. the hearing in October of 1983, several of the units — particularly the trailer units at El Portal — were admittedly in need of substantial maintenance. No rental adjustment was made for lack of insulation, and a flat 13% reduction was made with regard to the trailer units in need of maintenance, regardless of the condition of the unit. One trailer in the Wawona district was declared obsolete, and the occupant rental was reduced by 50%. The occupant was also administratively made eligible for relocation to other quarters.

Presumably all of the foregoing adjustments would have been reflected had an adequate physical inventory have been maintained for the Yosemite GFQ’s.

41 C.F.R. 114-52.3 lists other adjustments to basic rental rates which must be made by each agency when applicable.

They are as follows:

Unusual transportation costs (41 C.F.R. 114-52.302). This deduction was made using Mariposa as the nearest established community. There was no evidence introduced from which the Court could ascertain that this determination was not correct.

Adjustment for differences in amenities (41 C.F.R. 441-52.303). A committee was formed to consider these deductions or additions after the fact. No one from the committee was called. The committee report indicates only a limited number of GFQ’s considered in its report. The record is silent on whether the remaining units were considered.

Invasion of privacy (41 C.F.R. 114-52.-304). Again, the administrative record would indicate that there was no attempt made to investigate whether any of the GFQ in Yosemite qualified for this deduction. Letters of appeal contained in the administrative record indicate that several persons claimed this deduction, but at the time of the hearing, the appeal process was completely bogged down.

Excessive size or quality (41 C.F.R. 114— 52.305). The rental adjustment scheme provides for increasing the occupant’s rent in the event that the GFQ occupied is excessive as to size and quality for the occupant’s needs or that of his family. The inventory as maintained by the housing office in Yosemite Park did contain information from which this determination was made.

Inadequate size (41 C.F.R. 114-52.306). This section sets forth the criteria for determining whether or not the quarters occupied by the occupant or his family are inadequate for his needs.

The Park administration, however, did make an initial determination of the adjusted rental to be charged for each GFQ in Yosemite effective May 1, 1983, and each tenant was notified of the adjusted rental for which they would be liable.

The Park administration, being somewhat insecure as to the correctness of its procedures, particularly with regard to its own inventory of the Yosemite GFQ’s, sent out a memorandum to each affected occupant on May 19, 1983. That memorandum, in pertinent part, stated:

In addition to adjustments based on this notice, further in-house adjustments were going on. For instance, the entire trailer rental village at El Portal was reduced from a category of “fair” to a category of “poor” resulting in a rental reduction for the occupants of those particular GFQ’s. The uncertainty created by this failure to follow established procedures is illustrated by the net rental rate for Unit 7818 G3, El Portal. The revised net monthly rental rate for this unit effective February 10, 1980, was $166.00. The occupant was next notified that effective May 1, 1983, the net monthly rental charge would be $242.37. Finally, the occupant was then notified that the rent was reduced to $191.80 per month, that being 20% of her salary.

Finally the Court notes that the administrative process allows for one appeal by an occupant who is dissatisfied with the final adjusted rental fixed for his or her quarters. 41 C.F.R. § 52.602 , provides in pertinent part that:

(a) Occupants shall be notified of their rights to appeal rental rates, adjustments or other changes when they believe the quarters have been improperly classified, rates have not been established within these guidelines, or rates do not reflect reasonable value to the employee-tenant;

(b) The appeal must be in writing;

(c) The appeal should be filed within 10 days following receipt of notice of the change in rate;

(d) A further appeal is permissible to the Office of Hearings and Appeals if the appellant is dissatisfied with the initial determination by a bureau official.

The Yosemite GFQ occupants were notified of their right to appeal, and many have done so. Complaint was made during the hearing of the speed, or rather the lack thereof, exhibited by the National Park Service in dealing with these appeals.

IV

The Plaintiffs’ Attack

The plaintiffs’ attack upon the adjusted rental rates, as gauged by the pleadings, centers upon the alleged failure of the defendants to comply with the federal regulatory procedures in arriving at the final rental charges in Yosemite; that the procedures that were used varied substantially from those which the OMB and the Department itself designed to carry out the process, rendering the entire procedure arbitrary and capricious, exceeding the administrative authority granted to the defendants; and finally, that regulatorily mandated adjustments were not made prior to the imposition of the rental figure on each GFQ.

The government, of course, has generally denied the allegations, and relies heavily upon the generally recognized principle that judicial deference be given to agency action. In addition, the government has filed the affirmative defenses of failure to state a claim, failure to exhaust administrative remedies, and failure to join indispensable parties.

A. The Statistical Model and the Multiple Regression Procedures Used in the Determination of the Yosemite Rates.

In the Summary contained in the Technical Standards for Multiple Regression Analysis For Use In Government Housing Rent Determination of the Model prepared by Dr. Ralph Brown for use in this case, Dr. Brown outlined the contractor’s duties as follows:

1. Variables — The variables used as independent variables should be based on the private rental survey form shown in Figure 1.

2. Variable form — The variables utilized should be a combination of continuous, dummy, interaction, and nonlinear variables.

3. Stepwise regression — A stepwise regression analysis should be used to estimate the rent equation. Preferably an in-an-out stepwise regression be [sic] used.

4. Minimum Standard Error Criterion— The equation should be chosen so as to minimize the standard error of estimate for the equation. The equation which meets this criterion will also maximize the corrected coefficient of determination and include only those variables whose absolute value of their t values are equal to or greater than one.

5. Heteroscedasticity — The regression equations should be tested for the existence of heteroscedasticity.

6. Multicollinearity — The researcher should be familiar with the concept and consequences of multicollinearity.

7. Outliers — Outliers should be examined and reasons for their existence should be studied. If they exist due to data error or poor samples they should be eliminated from the model. If no reasonable explanation can be determined for their existence they should be eliminated ■ only if it is felt that their inclusion will bias the regression coefficients.

8. Stratification — Stratification of the data into more homogeneous groups is legitimate, but it should not be carried to the point of over-parameterization of the data.

9. Ten Percent Error — In addition to the minimum standard error criterion the regression should provide a degree of precision of a 10 percent or less mean absolute percent error.

Plaintiffs’ first attack is upon the procedure employed by the government in creating the statistical model used to predict the base rents to be applied to the GFQ’s in Yosemite. The original model, as outlined (which is known in statistical circles as a multiple regression analysis) was found to be acceptable by the plaintiff’s expert, Dr. Stanley Taylor. However, Dr. Taylor complained that the government employees charged with the execution of the model did not follow the model originally outlined thus rendering the results undependable. The procedure as actually employed by the government in establishing the model was deficient in the following particulars:

1. The sought-after dependent variable, which in this model ideally would be rent, was established by dividing rent by predicted rent. Dr. Taylor was unable to find any justification for this method in statistical literature, and totally rejected the justification advanced by the government to use this method. He would have taken rent and regressed it on other variables, since rent is what the process was designed to predict.

2. There was a lack of validation studies in the work performed by the government employees. Dr. Taylor observed that if you are going to use a statistical model for actual predicting purposes, it should be tested to see how well it performs as compared with actual data.

3. An unusually low percentage of the dependent variables was used in each of the regression runs.

4. There were none of the interaction terms contained in the model actually used. The model, as written by Professor Brown, stated that interaction terms had to be used liberally. Brown originally explained that the use of interaction terms makes the model and all of its predictions more precise. Dr. Taylor found no interaction terms when he inspected the model.

5. Of the 24 variables included in the final model for houses, for instances, 14 of them were dummy variables. These dummy variables, Professor Taylor concluded, were not helpful in predicting rents in Yosemite National Park.

6. No validation study was ever run.

Over the weekend recess, the government contacted the author of the model. Professor Brown then consulted with the person who actually supervised the “construction” of the model, as well as some colleagues in the same discipline, and based on this ex parte communication, testified, in effect, that the deviations from the model, although not previously approved by him, were satisfactory and did not affect the validity of the model. Professor Brown was never asked, nor did he testify, why he included the omitted steps from his directions when he originally constructed the model. Professor Brown admitted that the error in housing data was 13.5%, and the error on mobile home housing was in excess of 10%. In some instances, errors could have been as high as 25% or 30%, positive or negative.

It should be noted that in his formulation of the model, Professor Brown wrote: “The objective should be' to choose the equation which minimizes the standard error of the estimate and provides a degree of precision so as to have an ‘underlying mean absolute error of 10% or less.’ ” Dr. Brown was never asked nor did he volunteer any satisfactory answer as to why a statistical model that admittedly displayed a substantially greater average error was acceptable.

Dr. Brown did not testify as to what percentage of error was acceptable in this field, nor did he attempt to answer the specific criticisms of Dr. Taylor.

In conclusion, Professor Taylor stated that without having the entire set of computer tapes, he could not say with any degree of certainty what the result would be if the statistical regression analysis had been completed precisely as Professor Brown had outlined it in his technical standards.

Clearly, the Court is entitled to an explanation, and, if possible, a demonstration that the omissions from the model, as constructed by Dr. Brown, did not materially alter the accuracy of the survey. While the Court recognizes that statistics is not an exact science, it also is aware that statistical procedures have been developed to enable economic phenomenon to be predicted within' fairly precise ranges. Further, there should be some explanation as to why a model designed to produce a “mean absolute error of 10% or less” was still acceptable when it produced a much greater error in all categories of housing surveyed.

B. The Problem of the Economically Homogeneous Area

At the outset, it should be noted that neither counsel addressed this problem in their presentation of oral or documentary evidence. However, the issue, in the Court’s view, is basic to this entire controversy.

First of all, it should be remembered that OMB Circular No. A-45, revised, dictates that when a regional survey is undertaken, that the following considerations are mandated in establishing the region. 41 C.F.R. 115-52.105(h) defines an economically homogeneous survey area as follows:

Homogeneous is defined by Webster’s New Collegiate Dictionary and Webster’s Unabridged Dictionary as “of the same or similar kind or nature,” or alternately, “of uniform structure or composition throughout.”

We turn to a consideration of the administrative record, in an attempt to determine whether or not the mandates of the regulatory admonitions have been met.

As previously noted, the appraisers, in organizing their appraisal, included a brief description of each community surveyed.

George Sandberg, Chief of the Division of Real Property, and responsible for policy development in real estate matters for the Secretary of the Interior, presumably was the person in charge of the entire program of regional surveys conducted by the Department of the Interior to implement the rental adjustments on GFQ’s. As previously stated, the foregoing 40 communities were chosen from a larger list of 200-400 submitted by the various units of the Department of the Interior who would be impacted by the proposed rental increases. Neither in his affidavit nor declaration, or alternately in his oral testimony, does Mr. Sandberg address his concept — and hence the Secretary’s — of the meaning of the term “economically homogeneous area.” Instead, he dwells on the numbers of comparable units surveyed, as well as other matters not germane to the question. Mr. Sandberg’s conclusion that the more units surveyed results in a higher degree of accuracy of predictable average rentals as reflected in the market place may be correct; however, that is not what the authors of the regulations contemplated. They contemplated, rather, that the rentals by which GFQ rents were to be fixed, were to come from an area that is economically homogeneous with the location of the affected GFQ’s, i.e., that the market sur veyed should be “of same or similar kind or nature” as the affected GFQ’s.

A quick glance of the descriptive material appended to each community, by the contract appraisers, immediately illustrates the disparate economies of these communities, and explicitly or inferentially the impact such conditions have on housing prices within the communities surveyed.

The Court, of course, is not the first to note this problem. On February 16, 1983, Superintendent Binnewies wrote to the Regional Director of the Western Region, in pertinent part, as follows:

When questioned about this letter, Superintendent Binnewies indicated that he was later convinced that this view was in error after the March 22nd meeting in Yosemite attended by representatives of the Washington office as well as the Regional office. It is not the purpose of the Court now to speculate on the methods and means used to change his opinion; however, it is the conclusion of the Court that the original opinion expressed by Superintendent Binnewies in his letter of February 16, 1983, is absolutely correct, and should have been heeded by his superiors.

Y

Plaintiffs’ Standing to Maintain this Action

5 U.S.C. § 702 , provides in pertinent part as follows:

The statutory and regulatory scheme providing for the periodic adjustment of rentals in GFQ’s is clearly agency action. By the alleged failure of the Interi- or Department to follow the prescribed administrative steps in adjusting the rentals, the plaintiffs here are clearly persons who are aggrieved by agency action. The United States Supreme Court in Data Processing Service v. Camp, 397 U.S. 150 , 90 S.Ct. 827 , 25 L.Ed.2d 184 (1970), enunciated a two-prong test to determine plaintiff’s standing to challenge agency action:

1. Has the plaintiff alleged that the challenged action has caused him injury in fact, economic or otherwise; and

2. Whether the interest sought to be protected by the plaintiff was arguably within the zone of interest to be protected or regulated by the statute or constitutional guarantee in question.

Clearly, plaintiffs’ allegations and proof satisfy these tests.

Further, we note that the Department of the Interior and its Secretary are subject to the provisions of the Administrative Procedures Act. See Glacier Park Foundation v. Watt, 663 F.2d 882 (9th Cir.1982).

VI

Are the Plaintiffs Required to Exhaust Their Administrative Remedies?

As pointed out above, an individual tenant quartered in the GFQ’s may apply for an administrative review of the rental imposed upon his or her quarters. The review, however, is limited to a consideration of whether or not the agency has properly classified the quarters with regard to either its amenities or other characteristics that would affect the ultimate rental paid. The administrative review does not enable the tenant to complain as do the plaintiffs here. The basic rental adjustment upon which further individual refinements are based is properly fixed.

Further, the Court notes that the agency in the instant case completely ignored the suggestion that the occupants be kept informed as to all steps of the adjustment process, although not actually allowed to participate in the ultimate decision. While this does not constitute an administrative remedy in the usual sense of the word, had it been followed, some of the defendant’s difficulties may have been diminished or completely avoided.

Finally, it should be noted that nowhere in the administrative process has there been any provision for an administrative review of the agency action fixing the basic rent for the various classifications of housing here under consideration. Hence, the controversy satisfies the standard requirements that the agency action is final, and there is no other adequate judicial or administrative remedy available to the plaintiffs in their quest for relief.

DISCUSSION

Clearly, the plaintiff has not established the Secretary’s action to be violative of any constitutional rights, powers, privileges or immunities afforded to the plaintiffs. The procedure embarked upon was authorized by Congress and undertaken pursuant to lawful authority granted to the executive branch by Congress. The act per se was not in excess of any statutory authority nor was it arbitrary or capricious. Neither was it an abuse of discretion inasmuch as the acts are not per se vested within the discretion of the Secretary; the ambit of agency action is precisely proscribed by OMB Circular 45, as revised.

There is substantial evidence that the rates set on the GFQ’s, which are in issue in this case, were not set in accordance with the administrative procedures which the Secretary has mandated be followed in the rent setting process. This failure, of course, may not result “in the charges for facilities made available in connection with the occupancy of the quarters, shall be based on the reasonable value of the quarters and facilities to the employee ...” ( 5 U.S.C. § 5911 (c), supra.)

It appears to the Court that the issue facing the Court in this case is akin the problem encountered by the Ninth Circuit Court of Appeals in Yong v. Regional Manpower Admin., U.S. Dept. of Labor, 509 F.2d 243 (9th Cir.1975). There, Yong was appealing denial of his application for alien employment certification. It appeared that in arriving at this decision, the Regional Manpower Administrator had not followed the regulations promulgated under the applicable statute, namely, 8 U.S.C. § 1182 (a)(14). There, the court stated: “The reviewing court must decide whether the agency observed the ‘procedure required by law’ ( 5 U.S.C. § 706 (2)(D)). The statute itself does not define the quoted terminology.”,

Although we deal with a different kind of administrative agency action here, there is really a close relationship between the two factual situations. In this case, as noted above, the rental which plaintiffs are being charged was based on a survey of 40 cities scattered all over the State of California. This regional survey, by the Secretary’s own regulations, was to have been confined to an “economically homogeneous area in which the housing is located.” Had the Secretary or his designee considered the issue of economic homogeneousness and made a finding thereon, the Court would have been required to give deference to the Secretary’s determination. Rather, this administrative record reflects that the Secretary purposely chose not to select an economically homogeneous community, and substituted an unauthorized methodology. The only evidence is that these cities were chosen from a list of cities previously used for comparable rent determinations is not such a finding. Significantly, there is no evidence that these 40 cities were all used previously for the determination of rental increases for GFQ’s at Yosemite. Indeed, the entire thrust of the defendant’s evidence is that administrative convenience, rather than regulatory compliance, was the sole consideration of those conducting the rental survey.

Our finding, however, that the requisite procedures were not followed in this ease does not end our inquiry. The last phrase in 5 U.S.C. § 706 directs that “due account shall be taken of the rule of prejudicial error.” The Ninth Circuit stated in N.L.R.B. v. Seine and Line Fishermen’s Union of San Pedro, 374 F.2d 974, 981 (9th Cir.1967), as follows:

This record is replete with evidence of the amount of the increases both individually and collectively, as well as their economic impact upon the individuals subjected to them. Indeed, the government’s evidence indicates that the current rental increases will produce approximately $144,-000 during a twelve month period. The Court also notes that the rental revenue generated is used for the maintenance, operational and administrative costs of the quarters program for the park. It does not produce a profit for the park or the government. As the Court stated at the time it denied plaintiffs’ application for a temporary restraining order, it is not the purpose of the Court to determine how much any or all of the tenants of GFQ’s in Yosemite National Park must pay. Alternately, it is not the function of the Court in the framework of these proceedings to determine the adverse affect upon the maintenance and operation of the quarters program in Yosemite National Park in framing its decision. Further, it should be noted that neither the statutory mandate (reasonable value) or the administrative mechanisms are geared to the prediction of any particular level of income for the maintenance and operation of the GFQ’s at Yosemite.

Having determined that procedural defects occurred with regard to the rent setting procedure, the Court must properly look to these facts to determine whether or not the Secretary’s actions constitute prejudicial error. The Court determines that prejudicial error has been demonstrated.

Other factors could be pointed to to establish plaintiff’s entitlement to relief. While the plaintiffs’ expert witness could not testify as to the precise differences, if any, occasioned by the failure of the government to precisely follow the regression analysis of Professor Brown, Profes sor Brown was forced to admit that the method employed by the government resulted in a substantially greater average error than he originally designed into his program. The defendants did not introduce any evidence that the greater margin of error had de minimus effect upon the ultimate rental structure.

Further, as is pointed out in the Court’s discussion above, the failure of the Yosemite Park administration to maintain and adequately update their quarters inventory greatly delayed their ability to make certain administrative adjustments and has completely prevented them from making others. These factors, too, inure to- the cumulative affect of the prejudicial effects suffered by the plaintiffs as a result of the administrative conduct.

REMEDIES

The Court’s power to fashion relief under 5 U.S.C. § 706 is limited. The specific relief being sought here by the plaintiffs, of course, is a preliminary injunction to maintain the status quo ante pending a determination of the action on the merits. The considerations that the Court must weigh before granting such relief are covered in this Circuit by Los Angeles Memorial Coliseum Com’n v. Nat. Football, 634 F.2d 1197 (9th Cir.1980), wherein the Court stated:

The Court finds that the plaintiff has adequately met the first prong of this test, namely, a combination of probable success on the merits and the probability of irreparable injury. While monetary loss is usually not considered irreparable injury, this case presents a unique fact situation. The monies which are collected from occupants of GFQ’s in Yosemite National Park are quasi-trust funds in the sense that they are used to operate, administer and maintain the quarters program in the Park. No other funds are currently available for that purpose. Unless the Park is assured of a stable income, it will be difficult, if not impossible, for the administration of the Park to properly budget for the maintenance and service of the quarters. It is important, therefore, that the Secretary be apprised, as quickly as possible, what the level of these funds will be.

As to the individual tenants, the administrative record reveals there are instances of individual hardship that could not be adequately rectified by a lump sum refund later on.

Serious questions are raised by this case. Plaintiffs’ exhibits indicate that the Department of the Interior already contemplates a similar exercise in the Spring of 1986, using the same statewide survey boundaries in order to establish a rental adjustment effective November 1, 1986. It is important that the mistake of 1983 not be repeated three (3) years hence.

Clearly, the balance of hardships tips sharply in the plaintiffs’ favor in this regard. It is important that the' program which the Department intends to apply to California in the future be regularized and stabilized so that the tenants, on the one hand, may know precisely what rental they will be required to pay in order to allocate their limited resources to meet their daily expenses.

For, all of the foregoing reasons, the Court finds that the plaintiffs are entitled to a preliminary injunction. That finding does not end the Court’s concern in the matter, however. The Secretary in this case has several alternatives. Some of them are as follows:

1. To break out from the existing regional survey an area which, after appropriate consideration, the Secretary finds to be economically homogeneous to Yosemite National Park; then, using the available data make further “computer runs” in the manner originally mandated by Professor Brown.

2. Revert to one of the other alternate methods of increasing rentals mandated by OMB Circular A-45, revised, pursuant to the procedures prescribed therefor contained in Title 41 of the Code of Federal Regulations.

3. Request an early date for a hearing on the merits (which the Court is willing to grant), and attempt to justify the procedures used in the instant ratesetting process.

4. Seek certification for an interlocutory appeal.

Perhaps there are other alternatives that the Court has not envisioned, inasmuch as the Court naturally is not privy to the considerations and concerns of the Secretary in this regard.

Accordingly, the Court is concerned that the preliminary injunction be fashioned in such a manner that the Secretary be given appropriate discretion and flexibility in responding to the problems that this preliminary injunction may cause.

Counsel for the parties are directed to appear before the Court on Monday, April 2, 1984, at 2:00 p.m., at Sacramento, in a Courtroom to be announced, to discuss with the Court the appropriate formulation of the injunction.

. Much of plaintiffs’ efforts are concentrated on the economic impact on the individuals affected; the government failed to introduce any evidence to justify the administrative decisions which are central to this dispute. . Established community is defined as follows: Established communities. For the purposes of calculating a deduction under paragraph 6c(l) of this Circular, an established community is a population center offering the minimal community services listed below on a year-round basis, or, alternatively, on approximately the same seasonal basis as the occupancy of the Federal rental quarters under consideration. Conformity with this definition, without regard to population size or other criteria, is the sole basis for identification of an established community. Services Minimum Medical............. l physician, 1 dentist Educational ......... Public elementary and high school (unless transportation is provided without charge to a county or district school). Services Minimum Public library, school library available to the public, or scheduled mobile library. Shopping ........... Grocery, drugs, clothing, hardware, and general household needs. Religious............ Congregations of two faiths or denominations. Public transportation .. .Connection with at least one major town or city by common carrier. Minimal social, cultural or entertainment facilities . OMB Circular No. A-45 defines "nearby representative private community” as follows: “For purposes of establishing comparability, a nearby representative private community is the nearest community to the rental quarters offering a rental housing market, together with the minimal services set forth in the preceding paragraph. It must be a community which is not unreasonably affected by conditions of seasonal agriculture or tourism, population explosion, severe economic depression, or other such conditions which may have created an inequitable rent structure in that community not shared by the general region in which the rental quarters are located. ‘Representative’ is intended to recognize that comparability in housing includes such things as the economic environment of the housing. Thus, for example, housing in a small Government reservation should not be directly compared with housing located adjacent to undesirable areas in a large city. It is intended that comparable housing be selected as nearly as possible to establish a rental rate that truly represents the ‘reasonable value’ of the quarters to the occupant.’ . What the HUD minimum standards were, of course, was not developed in the testimony. However, the Court notes many of the dwellings inventoried in Yosemite as well as those reported in the region survey were of an age that suggested they were built years before HUD or its predecessor agencies first were concerned with minimum housing standards. . Superintendent Binnewies apparently was not informed of the progress of the private market survey or the schedule for the implementation of the rental increases until early January, 1983. It is obvious, however, that the Department did not comply with its own regulations which wisely encourage the involvement of the affected employees during the rentsetting process. See 41 C.F.R. § 114-52.601 . . The Code of Federal Regulations in this regard are even more adamant. The guidelines suggest that a tenants' committee be in place during all stages of the rent adjustment process. See 41 C.F.R. § 114-52.601 . . 41 C.F.R. 114-52.303, provides in pertinent part: "An adjustment in the basic rental rate shall be made to reflect differences in amenities (higher or lower) which may exist for rental quarters in relation to those of the private housing used for comparison. Adjustment for amenities is limited to consideration of those amenities listed in § 114-52.105(f).” 41 C.F.R. 114-52.105(f), Amenities, provides: Amenities include: (1) Paved streets. (2) Street lighting at least at intersections. (3) Sidewalks. (4) Lawns, trees, and landscaping. (5) General attractiveness of the neighborhood. (6) Community sanitation services. (7) Reliability and adequacy of water safe for household uses. (8) Reliability and adequacy of electrical service. (9) Reliability and adequacy of telephone service. (10) Reliability and adequacy of fuel for heating, hot water, and cooking. (11) Police protection. (12) Fire protection. (13) Unusual design features of the dwelling. (14) Absence of disturbing noises on [sic] offensive odors. (15) Standards of maintenance. Presumably, the requirement of an annual physical inspection of each GFQ, if performed at Yosemite, would have contained this information and made this belated survey unnecessary. The Committee contemplated - by the regulations was designed to serve other purposes. . Apparently after the magnitude of the rental increases in relation to the salaries of the occupants of the Yosemite GFQ’s was made apparent, an administrative decision was made to limit the rental increases to 20% of the occupant’s salary. A minimal number of occupants did receive rental decreases as a result of this administrative decision. . The Government would justify its procedures here by testimony that the same methods were used in other states. Such a conclusion, of course, is not determinative of the factual and legal issues raised in this case. . There was no evidence that the cash flow produced by the previous rental rates on the Yosemite GFQ’s were inadequate to properly fund the maintenance, operations and administrative costs. . The Court’s reading of the applicable regulations suggests that the excess rental payments will be returned in the form of credits on a new rental schedule.