Southern Railway Co. v. State Board of Equalization

MEMORANDUM OPINION HORACE T. WARD, District Judge. Introduction The above cases are consolidated actions brought for relief under the provisions of 49 U.S.C. § 11503 (“Section 306”) by seven interstate carriers by rail which own property in the state of Georgia and by eleven private rail carlines (hereinafter “Railroads” or “plaintiffs”). Defendants are the Georgia State Board of Equalization, its members, and the Georgia Commissioner of Revenue (hereinafter “defendants”). The defendants bear the responsibility under Georgia law for the valuation and assessment of rail transportation property for ad valorem tax purposes. The Railroads are seeking to permanently enjoin the defendants from assessing Railroads’ property for purposes of ad valo-rem taxation for the 1985 tax year pursuant to Notices of Assessment dated October 7, 1985, to the extent that such assessments are contrary to 49 U.S.C. § 11503. The Railroads also seek a declaration pursuant to 28 U.S.C. § 2201 that the defendants’ assessments of the plaintiffs’ property for the 1985 tax year, and the certification and distribution of such assessments, violate 49 U.S.C. § 11503. *1560 Jurisdiction is based upon 49 U.S.C. § 11503, 28 U.S.C. § 1337, and 28 U.S.C. § 1331. On application of the Railroads, the court granted preliminary injunctive relief, enjoining the defendants from assessing plaintiffs’ property for 1985 at a level higher than 28% of true market value. Railroads contend that defendants are assessing their transportation property for the 1985 tax year at a higher ratio of assessed value to true market value than the ratio of assessed value to true market value at which commercial and industrial property has been assessed in Georgia for the 1985 tax year. Specifically, they contend that defendants have assessed Railroads’ property at no less than 40% of true market value, even though other commercial and industrial property has been assessed for the same tax year at no more than 28% of true market value. The defendants contend that the Railroads must show that the assessment level of “all other commercial and industrial property” (i.e., non-railroad property) in each Georgia county is more than 5% less than the railroad’s 40% assessment level. They assert that Section 306 provides that the Railroads must satisfy the burden of proof provided by state law, and must prove the level of assessment of all other commercial and industrial property in each county of the state of Georgia. Defendants contend that the assessment jurisdiction should be on a county-by-county basis rather than a statewide basis as asserted by Railroads. The defendants further contend that the evidence they presented shows that the statute has not been violated in most Georgia counties, and that at a minimum the 28% assessment level urged by the plaintiffs is not an accurate measure of any relief to which they might be entitled. The matter was tried by the court, sitting without a jury, with the court taking the case under advisement pending the filing of post-trial briefs. FINDINGS OF FACT All nonexempt real and personal property located in the state of Georgia on January 1 of each year is subject to ad valorem taxation. O.C.G.A. §§ 48-5-3 & 48-5-10 (Michie 1982). All tangible property subject to taxation is required to be assessed for property tax purposes at 40% of its true market value. O.C.G.A. § 48-5-7 (Michie 1982 & Supp.1988). Public utility property, which by statutory definition includes railroad property, is assessed annually for property tax purposes by the Commissioner of Revenue. O.C.G.A. § 48-5-511 (Michie 1982). The assessment of a railroad’s property is apportioned among the counties in which the railroad owns property by the Commissioner under O.C.G.A. § 48-5-521 (Michie 1982 & Supp.1988). All other tangible commercial and industrial property is assessed for taxation by the tax receiver or tax commissioner in the county where the property is located. O.C.G.A. § 48-5-11 (Michie 1982). Railroad property, in practice, is reappraised annually by the Commissioner and is assessed each year at no less than 40% of true market value. The parties have stipulated that for the 1985 tax year Railroads’ property has been assessed at 40% of true market value. In the case of motor vehicles, the counties are required to determine the taxable value of motor vehicles on a uniform basis in accordance with schedules prepared annually by the Commissioner of Revenue. O.C.G.A. § 48-5-442 (Michie 1982). The tax is collected by the county at the time the motor vehicle owner purchases his or her license plate. O.C.G.A. § 48-5-473 (Mi-chie 1982 & Supp.1988). The parties called several expert witnesses who testified regarding certain theories and methodologies being advanced by the parties in support of their respective positions. Three of the principal expert witnesses are identified at the outset. The Railroads called Dr. Frederick A. Ekeblad as their statistical witness. Dr. Ekeblad has taught in the fields of economics and applied statistics. His experience included a professorship at Northwestern University and post of Dean of the College of Business Administration at the University *1561 of Bridgeport. The Railroads also called Dr. Dick Netzer, an economist and professor at New York University. He was called to testify regarding the level of assessment of commercial and industrial personal property in Georgia, based on a method that he developed. The defendants called Dr. Allison R. Manson, a long-time professor of statistics at North Carolina University, as their statistical witness. He testified about various issues raised in the case. Dr. Manson was initially retained by the defendants to make calculations designed to determine the overall assessment level of commercial and industrial property in all Georgia counties for 1985. The defendants contend that the evidence shows that the ratio of assessed value to true market value of commercial and industrial personal property (other than railroad transportation property) for 1985 is 40% in Bibb, Chatham, Cobb, DeKalb, Floyd, Hall, Richmond, and Whitfield counties. The evidence relating to those counties showed the methods used in assessing personal property. The individual assessors from these counties testified that in their opinion personal property in the respective counties was assessed at or near 40% in 1985. Their opinions are challenged by other evidence in the record. Some of these assessors also admitted that they failed to comply with certain assessment practices and procedures outlined by Steve Pruitt and Lyman Martin. Some did no trending and others did not conduct site audits. Lyman Martin, Tax Assessor of Hall County, which had one of the best systems, testified that while the ratio in his county might be close to the required level, his county had not achieved a 40% ratio as to personal property. Steve Pruitt, a defense witness, when asked about these eight counties, testified that they have the capabilities of assessing personal property at 40%. As to personal property ratios in the remaining counties in Georgia, defendants urged the court to assume that the personal property ratios were at the same level as real property ratios. The evidence in the case demonstrates that in a large number of counties in Georgia, assessment of personal property is of a poor quality when compared to the assessment of real property. The Railroads presented evidence regarding personal property assessment practices in 65 counties which supports this finding. Further, the evidence indicates that the assessment level of personal property in many Georgia counties is lower than that of real property. According to Dr. Ekeblad, this historically has been the case throughout the nation as a whole. Congress designated the sales assessment ratio study as an evidentiary device in section 306 litigation to demonstrate discrimination in taxation. The sales assessment ratio study compares the assessment on each sample of locally assessed real property to the price at which the property sells. The total of assessments of all samples of property is divided by the total of selling prices. The resulting percentage represents the average ratio of assessed value to true market value of commercial and industrial real property in the assessment jurisdiction. The Georgia Department of Audits conducts a sales assessment ratio study in each county in Georgia each year. 1 Railroads have been provided with the underlying data from the Department of Audits study, which compared sales occurring in calendar year 1984 to assessments on the sold properties as of January 1,1985. Railroads’ statistical witness, Dr. Frederick A. Ekeblad, has made an analysis of this data from the Department of Audits, using the ratio of aggregates as the average, and has determined that the average level of assessment of locally assessed commercial and industrial real property in the state of Georgia for the 1985 tax year was 27.8% The court has heard testimony from the director of the Department of Audits ratio *1562 study, Mr. Charles Ridley. The evidence showed that Mr. Ridley directs a full time staff of ten competent real estate appraisers who collect the data for the ratio study each year, and that the data was edited in accordance with standard ratio study editing principles. The Railroads retained Dr. Dick Netzer to prepare an estimate of the level of assessment of commercial and industrial personal property in Georgia for the 1985 tax year. Dr. Netzer used data compiled by the Department of Commerce and by that department’s Bureau of the Census as his source. Dr. Netzer estimated that the market value of all machinery, equipment, and inventories in Georgia as of January 1, 1985 was $40,487,000,000. The revenue department’s tax digest statistics establish that the assessed value of the same property for 1985 was $9,152,000,000. According to Dr. Netzer, when the total assessed value is divided by the total market value, the level of assessment of commercial and industrial personal property in Georgia for 1985 was 22.6%. Dr. Netzer admitted in his testimony that there was room for error in his calculations, stating that they were within about a 10% error rate. In order to arrive at a combined assessment ratio, it is necessary to include data regarding both real and personal property. In reaching a combined ratio, Dr. Ekeblad utilized the data developed by Dr. Netzer as to personal property and the Department of Audits’ data as to real property. At the trial, he testified that, in his opinion, the average level of assessment of locally assessed commercial and industrial real and personal property in Georgia as of January 1, 1985 was 25.2%. He did not include in this combined assessment commercial and industrial motor vehicles or the property of appealing or non-appealing utilities. Dr. Ekeblad later testified that if the property of non-appealing utilities was included the ratio went up to 25.5%, and if defendants’ data on motor vehicles is included the ratio went up to 25.6%. The Railroads concede that the property of non-appealing utilities should be included. The property of utility companies which have appealed the assessment ratios set by the revenue department for 1985 have been excluded by the Railroads from their ratio of aggregates calculations. It appears that this property represents 80% of the dollar value of the digest of centrally assessed property for non-railroad taxpayers. The Railroads contend that there is no way to know the actual figures until the appeals are decided. The defendants argue that this property should be included at a 40% ratio. The total assessed value of commercial and industrial motor vehicles was not included by Railroads in the comparison class of centrally assessed non-railroad property. The Railroads agree this property must be included, as a rule, but they contend that it is not required under the facts of this case. The Railroads do not seek relief with respect to their motor vehicles, which were assessed at the county level in the same method as were the motor vehicles of every other taxpayer. In support of their position, the defendants provided a study which indicates commercial and industrial vehicles in Georgia are assessed at 37.5% of true market value. The accuracy of the study is questioned by the Railroads. Dr. Manson testified that the best estimate of assessment level of commercial and industrial real property (other than rail transportation property) in each Georgia county for 1985 is the median. Dr. Manson thus used the median as the appropriate measure to estimate the market value of real property. His calculations were made on a county-by-county basis for the 159 counties in Georgia. In order to arrive at a combined ratio, Dr. Manson utilized ratios for real property, motor vehicles, utilities other than railroads and personal property, all on a county-by-county basis. The ratios for non-railroad utilities and personal property were determined to be 40%. These two ratios were based on figures supplied to Dr. Manson and not developed by him. The combination of the various ratios was performed by the use of ratio of aggregates. The resulting assessment ratios ranged from 20.7746% to 40%, with most ratios falling between 34% and 40%. *1563 The defendants did not use the sales assessment ratio study conducted by the Department of Audits for 1985. They, instead, used the study conducted by Assessment Analysis Associates, Inc. (“the AAA study”). This study was a sales assessment ratio study of real property assessment levels in each Georgia county for the 1985 tax year. It was commissioned by the revenue department. The AAA study compared sales prices on parcels of real property (for sales occurring between January 1, 1984 and December 31, 1984) to assessed values of such parcels as reflected on county assessment records for January 1, 1985. The sales data used in the AAA study were comprised of a random sampling of the sales data developed for the 1985 Department of Audits study and contained on the Audit Department’s computer tape. The data in the AAA study was analyzed by Dr. Manson and used by him as the basis for his calculations regarding the estimated market value of real property. The evidence shows that the Department of Audits’ assessment ratio study has a data base much larger than the revenue department ratio study, and the court finds it to be a reliable indicator of local real property assessment levels in Georgia for the 1985 tax year. The Railroads have offered evidence in an effort to establish that the average assessment level of commercial and industrial property in Georgia contains two upward biases. First, they sought to show that the sales assessment ratio for the data used by their expert produced an upwardly distorted result because of the widespread practice of “sales chasing” in Georgia. “Sales chasing” occurs when a local tax assessor changes assessments only on property which actually sells in a given year, while leaving the assessments on property which does not sell unchanged. The Railroads contend that because the sales assessment ratio study is based almost entirely on information concerning sold properties, the average produced by the study in a county which chases sales will be higher than the actual assessment level of all property (sold and unsold) would be. Secondly, the Railroads seek to show that although Georgia is a state with vast timber resources, standing timber is under-assessed in most counties and not assessed at all in some counties. They assert that these under-assessments of larger amounts of property have not been reflected in the evidence produced by the Railroads as to the assessment level of real property. The court has considered this evidence. While it appears to be relevant on the issues for which it was offered, it has not been a determining factor in any decisions reached in this case. CONCLUSIONS OF LAW As pointed out above, the Railroads filed this lawsuit contending that the defendants had attempted to assess the plaintiffs’ property for the 1985 tax year in violation of 49 U.S.C. § 11503. Section 11503 confers jurisdiction upon United States district courts to prevent states, or any authority acting for a state, from assessing rail transportation property at a higher ratio of assessed value to true market value than the ratio applicable to all other commercial and industrial property in the same assessment jurisdiction. This statute was enacted by Congress in 1976 to prohibit states from discriminating against rail transportation property in the assessing, levying, and collecting of property taxes. Railroad Revitalization and Regulatory Reform Act of 1976, Pub.L. No. 94-210, § 306, 90 Stat. 31, 54. Section 306 was first codified at 49 U.S.C. § 26c. The language of § 306 was changed and recodi-fied at 49 U.S.C. § 11503 as part of the Revised Interstate Commerce Act of 1978, 49 U.S.C. §§ 10101-11917 (West Supp. 1988). The statute will be referred to as § 306 rather than § 11503. The Railroads argue that the evidence clearly shows that the railroads have been victims of tax discrimination for the 1985 tax year, and it is only for the court to determine the extent of relief to be accorded plaintiffs. They further contend that the defendants’ own evidence supports their claim of tax discrimination. There is a definite dichotomy between the defen *1564 dants and the Railroads as to the approach the court should use in deciding various factual and legal issues raised in this case. These include the two basic issues as to whether the matter should be considered on a statewide or a county-by-county basis and whether the ratio of aggregates or median should be used as the measuring technique. Also, there is a major dispute as to whether the Railroads have carried their burden of proof regarding personal property, motor vehicles and appealing public utility property. The parties in their pretrial order have identified seven issues to be decided by the court in this case. The court will address each issue as follows. 1. Should the Court Use the State or the Individual Counties As the Assessment Jurisdiction? The Railroads contend that the average level of assessment of commercial and industrial property should be determined by aggregating all of the Department of Audits sales and assessments data into a single statewide average. The defendants contend that assessment levels of each individual county should be used to determine the amount of equalization relief to which Railroads are entitled. Georgia, like many other states, assesses railroad property for ad valorem taxation by a central agency, in this state by the revenue department. The statewide assessment is then allocated among the Georgia counties within which the railroads operate using traditional allocation formulas. The Railroads in Georgia pay property taxes based upon the revenue department assessment to both the counties and to the state. Whether the state or the counties should be considered to be the assessment jurisdiction is an issue which has been raised in Section 306 litigation in at least four other United States district courts. These courts concluded that where railroads are centrally assessed by the state, the assessment jurisdiction within the meaning of section 306 is the entire state, not the individual counties. State of Arizona v. Atchison, Topeka and Santa Fe Railroad Co., 656 F.2d 398 (9th Cir.1981); Burlington Northern Railroad Co. v. Department of Revenue, 604 F.Supp. 1575 (W.D.Wisc.1985); Atchison, Topeka and Santa Fe Railway Co. v. Lennen, 552 F.Supp. 1031 (D.Kan.1982), aff 'd in part and rev’d in part on other grounds, 732 F.2d 1495 (10th Cir.1984); Southern Pacific Transportation Company v. State of California, Nos. C-81-4848 W, etc., filed April 5, 1984. In Georgia, as the court has previously observed, railroad property is centrally assessed for property tax purposes. That fact standing alone is sufficient to support a finding that the state as a whole should be considered to be the assessment jurisdiction when the above cited cases are considered. Although the above cited cases are not binding, this court sees no reason why long established precedent in Section 306 litigation should not be followed in this case. Because railroads are centrally assessed in Georgia, this court concludes that the assessment jurisdiction in this case is the state and not the individual counties. There is another reason to treat the state as the assessment jurisdiction. Railroads under Section 306 are entitled to relief based upon a comparison of railroad property assessment levels to the average commercial and industrial property assessment level. In small, rural counties, very few commercial and industrial sales occur in a given year. As a result, it is virtually impossible to calculate a statistically reliable average for those counties. On the other hand, when the entire state is considered to be the assessment jurisdiction, a sufficient number of sales exists to calculate an average level of assessment with a high degree of confidence. The defendants have cited to the court two decisions in Section 306 litigation in North Carolina as precedent for the use of the county as the assessment jurisdiction. Clinchfield Railroad Company v. Lynch, 527 F.Supp. 784 (E.D.N.C.1981), aff'd., 700 F.2d 126 (4th Cir.1983) (“Clinchfield I”); Clinchfield Railroad Company v. Lynch, 605 F.Supp. 1005 (E.D.N.C.1985), aff'd., 784 F.2d 545 (4th Cir.1986) (“Clinchfield II”). *1565 They are correct in stating that individual counties were used as assessment jurisdictions in those two cases. The cases do not serve as useful precedent, however. It does not appear that the question of whether the state or the counties should be the assessment jurisdiction was ever raised in those cases. Also, it appears that neither the district court nor the court of appeals in either Clinchfield I or Clinchfield II even considered whether the state or the county should be treated as the assessment jurisdiction for the purpose of measuring relief from discrimination. 2. Does Section 306 Limit Relief From Discrimination by Permitting States and Counties to Assess Commercial and Industrial Property at a Level of Assessment Five Percentage Points Below the Railroad Assessment Level? Section 306(2)(c) contains the following language: No relief may be granted under this section unless the ratio of assessed value to true market value with respect to transportation property exceeds by at least 5 per centum the ratio of assessed value to true market value, with respect to all other commercial and industrial property in the same assessment jurisdiction. In Georgia all nonexempt, tangible real and personal property must be assessed at 40% of true market value. O.C.G.A. § 48-5-7 (Michie 1982 & Supp 1988). The defendants contend that Railroads are entitled to a reduction in their assessments only to the extent that the level of assessment of commercial and industrial property is below 35%. The defendants argue that Section 306(2)(c) should be interpreted to mean that discrimination is permitted within five percentage points of the railroad assessment level of 40%. The Railroads contend that: (a) Section 306(2)(c) simply provides for a jurisdictional threshold, just like the jurisdictional amount of $10,000.00 contained in the diversity statute, and that once the 5% threshold has been satisfied, the court is empowered to prevent all discrimination against the Railroads; and (b) the 5% threshold is calculated by measuring 5% of 40% (which yields two percentage points, not five), thereby establishing 38% as the jurisdictional threshold. In other words, Railroads contend that if the proof shows that commercial and industrial property is assessed no higher than 38% of true market value, this court has jurisdiction to grant Railroads full relief from discriminatory taxes. The decision of a United States District Court in Louisiana provides some guidance on this issue. In Louisville and Nashville Railroad Company v. Louisiana Tax Commission, 498 F.Supp. 418 (M.D.La. 1980), the state by statute taxed railroads at 25% of value and commercial and industrial property at 15% of value. The state argued that railroad assessments should only be reduced to 20% of value, since Section 306 permitted five percentage points of discrimination. The Court held as follows: Defendants overlook the provisions of subsection (b) of the Act which unequivocally provide that a state may not assess rail transportation property at a value that has a higher ratio to the true market value than the ratio used in assessing other commercial and industrial property. The 5% differential was included simply because the Congress was not concerned with insignificant variations that might develop in state assessment practices and therefore authorized injunctive relief only where the differential actually exists. That provision, however, does not restrict in any way the reach of the relief. The statute prohibits all discrimination. The 5% variance is simply a threshold authorization for injunctive relief. 498 F.Supp. at 423. This court’s interpretation of the five percent threshold calculation appears consistent with the decisions of United States district courts in Section 306 and related litigation in Ogilvie v. State Board of Equalization, 492 F.Supp. 446, affid, 657 F.2d 204 (8th Cir.), cert. denied, 454 U.S. 1086, 102 S.Ct. 644, 70 L.Ed.2d 621 (1981). In Ogilvie, supra, the Court found that a *1566 16.8% assessment level for railroads exceeded a 12.2% assessment level for commercial and industrial property by 38%, not by just 4.6%. In Arkansas-Best Freight System, Inc. v. Cochran, 546 F.Supp. 904 (M.D.Tenn.1981), the Court found that a 55% assessment level for motor carriers exceeded a 40% assessment level for commercial and industrial property by 88%, not by just 15%. This court concludes that the 5% differential should be calculated as Railroads contend. The jurisdictional threshold in this case is 38%, not 35%, because the parties have stipulated that Railroads are assessed at 40%, which exceeds an assessment at 38% by at least 5%. 3. Should Commercial and Industrial Motor Vehicles Be Included in the Comparison Class. If So, Who Has the Burden of Proving the Level of Assessment of Motor Vehicles? The defendants have produced a study which indicates that commercial and industrial motor vehicles in Georgia are assessed at 37.5% of true market value. They contend that the total assessed value of commercial and industrial motor vehicles should be added to the numerator of the ratio of aggregates calculation and the total estimated market value of such vehicle should be added to the denominator. To do so would increase the statewide ratio of aggregates because motor vehicles appear to be assessed at a higher level of assessment than any other category of locally assessed property. The Railroads contend that motor vehicle assessed values and estimated market values should be ignored altogether, because Railroads seek no equalization relief with respect to their own motor vehicles. In Georgia the motor vehicles of centrally assessed taxpayers are assessed identically to motor vehicles of all other taxpayers. A railroad pays the tax when it purchases license plates. The railroad motor vehicle assessment is based upon the same manual used for all other taxpayers. When the revenue department values a railroad’s property each year for tax purposes, the revenue department deducts from its valuation the value of the railroad’s motor vehicles, because those motor vehicles have already been valued for taxation when license plates were purchased. Another reason advanced by the Railroads as to why commercial and industrial motor vehicles should not be included in the comparison class is that defendants have failed to prove either the actual assessed value or the estimated market value of commercial and industrial motor vehicles in the state. The evidence offered by defendants to support inclusion of motor vehicles is deficient. In the first place, defendants were unable to devise a method to identify the assessed value of all commercial and industrial motor vehicles. Mr. Steve Pruitt, a defense witness, testified that while separate classifications exist for automobiles, trucks and motorcycles, there is no way to tell which automobiles, trucks, or motorcycles were commercial and industrial vehicles. Mr. Pruitt decided not to use the truck classification as the best data base to determine the total assessed value of commercial and industrial motor vehicles. Instead, he utilized a hybrid classification consisting of certain vehicles selected according to their tag prefixes. This grouping did not include all commercial and industrial motor vehicles. Thus, defendants used as the total assessed value the amount of assessments of a portion of commercial and industrial motor vehicles. In the second place, no ratio study was performed to determine the assessment level of a sample of the hybrid class chosen. The information regarding motor vehicles developed by Mr. Pruitt was supplied to Dr. Manson, and utilizing these figures Dr. Manson developed his motor vehicle ratio. Also, there is a burden of proof problem involved at this juncture. The Fourth Circuit Court of Appeals held in section 306 litigation that once railroads have shown the existence of discrimination by a real property sales assessment ratio study, the state has the burden of proving a different and higher level of assessment for personal property. Clinchfield R. Co. v. Lynch, 700 F.2d 126, 131, 133-34 (4th Cir.1983). The court agrees with the position of the Rail *1567 roads that motor vehicles not be included under the facts of this case. 4. Should Assessed Values and Market Values of Centrally Assessed Property for Taxpayers Whose Proposed Assessments Are Under Appeal Be Excluded from the Ratio of Aggregates Calculation? It has long been decided that federal courts in Section 306 litigation should include centrally assessed, non-railroad utility property in the commercial and industrial property class for the purpose of comparing the level of assessment of railroad property with the average level of assessment of commercial and industrial property. Ogilvie v. State Board of Equalization, 657 F.2d 204 (8th Cir.), cert. denied, 454 U.S. 1086, 102 S.Ct. 644, 70 L.Ed.2d 621 (1981); Clinchfield Railroad Company v. Lynch, 700 F.2d 126 (4th Cir.1983). The Railroads contend, however, that in this case it is not possible to include the assessed values for most of the utility digest, because utilities representing approximately 80% of the dollar value of the digest have appealed the amount of their assessments. Their assessed values simply cannot be determined until the appeals have been concluded, the Railroads argue. The defendants, on the other hand, argue that the property of appealing utilities should be aggregated into the ratio of aggregates at the amount of the proposed assessments, and these proposed assessments should be presumed to be at 40% of true market value as required by state law. The court disagrees. If the large utilities in Georgia are successful in their equalization claims under state law, the resulting ratio of aggregates would be too high because the final assessed value of utility property would be substantially reduced. The court determines that the property of these utilities whose assessments are under appeal should at this time be excluded from the statewide ratio of aggregates calculation in this case. The court will retain jurisdiction of the case until the appeals are concluded in order to determine whether adjustments should be made. 5. Have the Eight Georgia Counties Designated by the Revenue Department Assessed Commercial and Industrial Personal Property for the 1985 Tax Year at b0% of True Market Value? The defendants contend that 8 of the 159 counties in Georgia assess commercial and industrial personal property at the statutory level of 40%, and, therefore, at a higher level of assessment than that of real property. The defendants ask the court to accept an increased ratio for these eight counties over the real property ratio. Because the court has concluded that the assessment level to be used for equalization purposes is the average level of assessment of commercial and industrial property statewide, this issue might be moot. Even if these eight counties are assessing personal property at 40%, which the court doubts, no adjustment is required where a state average is used instead of county averages, because this would require the court to assume that personal property ratios in the other counties were equal to the ratios of real property. This is an assumption the court is not prepared to make, as the evidence indicates that in many counties the level of assessment of personal property is below that of real property. There is, however, a related question which should be addressed. The defendants contend that the railroads are required by Section 306 to prove the average level of assessment of all commercial and industrial property, not just real property which is the subject of a sales assessment ratio study. The State relies upon the following language contained in Section 306(2)(d): [t]he burden of proof with respect to the determination of assessed value and true market value shall be declared by the applicable State law ... The Court of Appeals for the Fourth Circuit has reached a contrary result in Clinchfield I. On the burden of proof issue as to personal property, the Court held as follows: The answer to our question requires a consideration of where the burden of proof fell once discrimination, even *1568 though shown only as to real estate, had been established. We are satisfied that once North Carolina was shown to have practiced discrimination with regard to real property under its statute which applies a single undifferentiated assessment to both real property and personal property, the state assumed the burden of establishing facts sufficient both to warrant a different conclusion with respect to personal property and to enable the district court to fashion a decree that would not frustrate efforts to alleviate the discrimination already proven as to real property. Clinchfteld I, 700 F.2d at 131. The Court of Appeals based its holding upon North Carolina law, but the court considered the same question in the context of federal law. The court strongly indicated that as a matter of federal policy, district courts should not require railroads to prove the level of assessment of personal property where discrimination as to real property has been established by use of a sales assessment ratio study. The court observed as follows: While it does not affect in any way our decision, an observation is merited that the result would be unfortunate were we to agree with the state’s contention that the burden to produce the relevant figures as to personal property lies with the taxpayer. Bearing in mind the remedial purposes of the Railroad Revitalization and Regulatory Reform Act of 1976, it is sensible that the burden of proof has been allocated to the state. It also is sensible that the state’s legislators be afforded an incentive to effect revisions necessary to eradicate the inherently discriminatory practices evidently imbedded in the present version of their state’s law. Clinchfteld I, 700 F.2d at 133-134. The court will not require Railroads to prove the level of assessment of commercial and industrial personal property in order to obtain relief under Section 306 if sales assessment ratio study evidence proves discrimination as to real property. If the defendants seek to establish a higher level of assessment for personal property than that established by a sales assessment ratio study for real property, they have the burden of proof. On the other hand, if railroads seek to establish a level of assessment for personal property which is lower than that for real property, as in this case, railroads have the burden of proof, which Railroads have acknowledged at the trial of this case. The court further observes that the only evidence of the level of assessment of locally assessed commercial and industrial personal property on a statewide basis is the evidence offered by Railroads through Dr. Netzer. The defendants raised several objections to Dr. Netzer’s study and testimony. They contend that it is not probative of the level of personal property assessments and is not persuasive evidence for the court to consider. Defendants argue that Dr. Netzer’s basic assumption regarding business and industry is not well-founded and that the Railroads have offered no scholarly support for his study. The court has carefully considered defendant’s objections, along with the Railroads’ response regarding the testimony of Dr. Netzer. The court is not convinced that it has no probative value and will not reject it. As is often the case with expert testimony, Dr. Netzer’s methodology is subject to various limitations, but that goes to its overall probative weight. This court has made a determination similar to one made in Clinchfteld II, supra, 784 F.2d 545, 554, n. 11, that Dr. Netzer’s study is “basically sound” and provides some evidence of the level of personal property assessments. Further, it is supported by other evidence in the record indicating that commercial and industrial personal property in Georgia is assessed at a lower level than commercial and industrial real property. 6. and 7. What Statistical Measure or Average Should Be Used to Determine the Market Value of Commercial and Industrial Real Property1 The expert statistical witnesses who testified for both sides agree that the *1569 ratio of aggregates must be used as the average to determine the level of assessment of commercial and industrial property when different types of property are combined to calculate an overall assessment level. The expert witnesses also agree that in order to derive the ratio of aggregates for commercial and industrial property, it is necessary to estimate the fair market value of both commercial and industrial real property and commercial and industrial personal property. These witnesses differed as to how the fair market values of real property should be estimated. There are two types of averages which statisticians agree can be used to determine the level of assessment of real property from sales assessment ratio study data. These two averages are the median and the ratio of aggregates. To calculate the median, the statistician converts each sample to a percentage of assessed value to selling price. For example, if a parcel of property in the statistical sample is assessed at $10,-000.00 and sells for $40,000.00, the ratio applicable to that parcel is 25% ($10,000 divided by $40,000 X 100). The individual ratios contained in the sample are arrayed in descending order to find the median, which is the ratio in the middle. When the median is used as the average, 50% of the ratios in the sample fall above the median and 50% of the ratios in the sample fall beneath the median. The ratio of aggregates, as the name applies, is an aggregate calculation. All assessments in the edited data base are summed. All selling prices are summed. The sum of the assessments is divided by the sum of the selling prices. The result is then converted to a percentage (assessed values divided by selling prices x 100). The defendants’ expert witness, Dr. Manson, testified that the median should be used to estimate the market value of real property because it is the middle of the ratio and is not influenced by extremes to the same extent that the ratio of aggregates is influenced by extremes. The Railroads’ expert, Dr. Ekeblad, disagreed. Dr. Ekeblad testified that the ratio of aggregates, which is dollar weighted, is the only average which should be used to estimate market value of real property under the facts of this case. One problem with Dr. Manson’s position on this point is that it ignores the reality of assessment trends in Georgia. Dr. Ekeb-lad testified in rebuttal testimony that in Georgia high valued commercial and industrial property tends to be assessed at lower levels than low valued commercial and industrial property. Accordingly, the use of the median to estimate market value ignores such a tendency and invariably results in underestimating the market value of commercial and industrial property, where the assessment ratio for the most highly valued commercial and industrial properties is lower than the median. 2 It appears that most experts agree with plaintiffs’ witness on this point. Dr. Ekeb-lad testified that he did not know of anyone qualified as an expert in the field of sales assessment ratio studies other than Dr. Manson who used the median to estimate the value of real property. Dr. Manson was asked about the use of the ratio of aggregates to estimate fair market valúe of property. Q. (Mr. Gibson): Well, my question is, don’t most people in the field, if asked for any purpose to estimate the total market value of the tax digest or the tax roll, don’t most people in the field use a ratio of aggregates for that purpose? A. (Dr. Manson): I would say that’s probably a correct statement. There is judicial precedent for the use of the ratio of aggregates in a case like the instant case. This was decided by the *1570 Court of Appeals for the Tenth Circuit in a case arising under Section 306, as follows: A median has no significance if the items on the continuum are not alike in some relevant way. When, as here, categories of property that are subject to different methods of appraisal must be factored together to produce the assessment ratio for “all other commercial and industrial property,” each category should be factored in proportion to its share of the total true market value of all such property. Atchison, Topeka and Santa Fe Railway Company v. Lennen, 732 F.2d 1495, 1504-1505 (10th Cir.1984). General Overview As indicated above, in this case the parties have disagreed on all major factual and legal issues. In the course of their presentation, the respective sides enlisted expert witnesses to support and defend their positions, such as assessment jurisdiction (county versus statewide) and measuring ratios (median versus ratios of aggregate). When there is such a sharp dichotomy of theories and positions being advanced, the court is confronted with the prospect of coming down on one side or the other. This is especially troubling when such heavy reliance is placed upon the testimony of expert witnesses. These witnesses were chosen by the parties for reasons best known to them. While the court has the authority to call expert witnesses, it is generally not done in this court’s experience except on the motion of a party. 3 The court has carefully considered the challenges and responses made by the lawyers on the qualifications and opinions of the chief experts. In general, the court found all to be qualified and scholarly in their various endeavors and was impressed by their testimony. As between the main statistical expert witnesses, Dr. Ekeblad for the Railroads and Dr. Manson for the defendants, while giving full regard to the testimony of each, the court has accepted the testimony of Dr. Ekeblad as being more persuasive and convincing on the major issues raised in this case. One reason why the court has found Dr. Ekeblad’s opinions more persuasive is that they appear to be substantially supported by other evidence in the record. As to Dr. Netzer’s study and testimony relating to the difficult issue of personal property assessment, the court has found it to be basically sound and the best evidence on the estimate of the statewide personal property assessment ratio in this case. In effect, what a federal district court judge is called upon to do in a case under section 306 is to determine the assessment level for purposes of taxation regarding the commercial and industrial property of railroads. The court realizes that this is very serious and complicated undertaking, and it has not been taken lightly. In the course of reaching a decision, it has not been possible for the court, to arrive at precise assessment level percentages in this case. Upon a consideration of the evidence, the court has simply concluded that the greater weight of the evidence supports the Railroads’ contention that non-railroad commercial and industrial property in Georgia for 1985 was assessed at a level not higher than 28% of true market value, while similar property of the *1571 Railroads was assessed at 40%. The Railroads have not sought relief on an assessment ratio lower than 28%. It might have been possible for the court to have arrived at a higher percentage than 28% if justified by the evidence. The main thrust of defendants' evidence was premised on a county-by-county assessment jurisdiction with the use of the median as a measuring technique, neither of which was adopted by the court. In the form of rebuttal testimony through Dr. Manson, defendants did present alternative calculations on a statewide basis using both the median and the ratio of aggregates. Defendants’ expert testified that both calculations, whether by median or ratio of aggregates, were unweighted. The statewide calculations using the ratio of aggregates testified to by Dr. Manson are contained in Defendants’ Exhibit 27. These calculations for all non-railroad commercial and industrial property (excluding appealing utilities) produced assessment ratios of 33.43% and 33.85%, depending on the configuration of real property used (either R-2, R-4, R-5, R-6, or R-4 and R-5). The court has considered these calculations, but was not persuaded that they should be substituted for the Railroads’ calculations or used to raise the final assessment ratio. One reason is that these calculations required the court to assume that personal property assessment levels equal the assessment ratio of real property in 150 counties. This position is flawed as it has been previously determined in this memorandum opinion that in a substantial number of counties in Georgia personal property is assessed at a lower level than real property. CONCLUSION In view of the foregoing, the court concludes that the evidence establishes that the level of assessment of nonrailroad commercial and industrial property in Georgia for the 1985 tax year does not exceed 28%. Accordingly, a separate order will be entered permanently enjoining the defendants from assessing railroad commercial and industrial property for the 1985 tax year in excess of 28% of true market value. This order will contain a proviso reserving jurisdiction in the event the decisions regarding appealing utilities require adjustments in the 28% assessment ratio. SO ORDERED. Footnotes [1] The state auditor’s sales assessment ratio study is conducted by the Georgia Department of Audits each year for the purpose of being used in determining how to allocate state funds to the counties in Georgia for school purposes. Thus, its origin and purpose has no direct relationship to the instant litigation between the Railroads and the state defendants. [2] Another reason asserted by the Railroads as to why the median should be rejected is that the use of any average other than the ratio of aggregates in this case is mathematically incorrect. They assert that the ratio of aggregates is the algebraic expression of the ratio of assessed value to true market value. They assert that since the ratio of aggregates is determined by the amount of the market value of property, it necessarily follows that to rearrange the algebraic equation to solve for market value, the ratio of aggregates is the average which must be used as a devisor. The court simply notes this argument, but neither accepts or reject it. [3] See McCormick on Evidence 3rd Ed., West Publishing Co., (1984). The following passage at p. 43 deals with expert testimony: There are two chief points of weakness in the use of experts. The first is the choice of experts by the party, who will naturally be interested in finding not the best scientist, but the "best witness.” The remedy for the first weakness is not far to seek. It lies simply in using the trial judge’s common law poser to call experts. Cases are recorded as early as the 14th century — before witnesses were heard by juries — of the summoning of experts by the judges to aid them in determining scientific issues. The existence of the judge’s power to call witnesses generally and expert witnesses particularly seems well recognized in this country. The principle is implemented in the carefully drafted Model Expert Testimony Act approved in 1937 by the Commissioners on Uniform State Laws, and embodied in abbreviated form in the former and present Uniform Rules of Evidence and Federal Rule of Evidence 706_