Northwest Airlines, Inc. v. Dept. of Rev.

943 P.2d 175 (1997)

325 Or. 530

NORTHWEST AIRLINES, INC., United Airlines, Inc., Trans World Airlines, Inc., Horizon Air Industries, Inc., Continental Airlines, Inc., American Airlines, Inc., United Parcel Service Company, Federal Express Corporation, and Alaska Airlines, Inc., Appellants, v. DEPARTMENT OF REVENUE, State of Oregon, Respondent.

OTC 2892, 3137, 3276, 3482, and 3688; 3019, 3136, and 3481; 3274 and 3479; 3135 and 3476; 3140; 3013, 3483, and 3689; 3478 and 3685; 2902, 3001, 3275, and 3474; SC S42867.

Supreme Court of Oregon.

Argued and Submitted November 4, 1996.

Decided August 7, 1997.

*176 Stephen D. Goodwin, of Baker, Donelson, Bearman & Caldwell, Memphis, TN, argued the cause and filed the briefs for appellants. With him on the briefs were Gregory J. Miner and Bruce Cahn, of Bogle and Gates, P.L.L.C., Portland, and James W. McBride, of Baker, Donelson, Bearman & Caldwell, Washington, DC.

Marilyn J. Harbur, Assistant Attorney General, Salem, argued the cause for respondent. With her on the brief was Theodore R. Kulongoski, Attorney General, Salem.

Before CARSON, C.J., and GILLETTE, VAN HOOMISSEN, FADELEY, GRABER and DURHAM, JJ.

*177 CARSON, Chief Justice.

Plaintiffs are several interstate air carriers (the airlines) that own property, consisting of aircraft and other personal property, located within the State of Oregon. They challenge several tax assessments on their property completed by the Department of Revenue (the department) for various tax years. The issues presented involve the interpretation and application of a federal statute, 49 U.S.C. section 1513(d),[1] which prohibits discriminatory state taxation of air carrier transportation property. For the reasons that follow, we affirm the decision of the Tax Court, which rejected the airlines' claims.

I. BACKGROUND

The airlines' personal property located within the state is subject to ad valorem state taxation, ORS 307.030(1),[2] and is centrally, rather than locally, assessed each year by the department. The department assesses the airlines based upon a "unit" method, which requires determining a nationwide total value for each airline and then allocating part of that value to Oregon.[3] ORS 308.555. Utility property and certain industrial personal property [4] also are centrally assessed by the department. ORS 308.515(1). Utilities are assessed under the unit method, and industrial property is assessed in a manner similar to local assessments, as described in the following paragraph.

By contrast, all commercial personal property and most industrial personal property is locally assessed by county assessors. ORS 308.210(1). Each year, owners of such property must provide their county assessor's office with personal property returns, which list their personal property owned during that year and its original cost. The county then adjusts the original reported costs upward for inflation and downward for depreciation. The resulting figure is the "assessed value" for each piece of property, which each county uses to impose personal property tax upon property owners. Under that "self-reporting" system, local assessments are based upon information that is reported to county assessors by property owners.[5]

The department completed assessments for the airlines' property for varying tax years. The airlines challenged those assessments by filing complaints in the Tax Court,[6] claiming that the assessments violated 49 U.S.C. section 1513(d). That statute prohibits states from assessing air carrier transportation property at a higher ratio to true market value than the ratio of assessed value to true market value for other similar commercial and industrial property. The airlines each sought a set-aside of the department's assessments and a determination of the proper ratio of assessed value to true market value for other commercial and industrial property, as compared to their property.

At trial, the parties stipulated that: (1) because about 95 percent of airline property in Oregon is located in Multnomah County (the county), the parties would present evidence concerning only the assessment ratio of other commercial and industrial personal property located in that county; (2) the Tax Court's determination of the assessment ratio for the 1993-94 tax year would apply to all the tax years at issue; and (3) the department *178 assessed the airlines' property at 100 percent of its true market value. Delta Air Lines, Inc. v. Dept. of Rev. (I), 13 OTR 357, 361, 1995 WL 617652 (1995). Consequently, the central issue before the Tax Court was whether the ratio of assessed value to true market value of other commercial and industrial personal property located in the county during the 1993-94 tax year was less than 100 percent.

The airlines presented a study prepared by their economics expert, Dr. Bahl, which demonstrated that, during the 1993-94 tax year, the county had assessed other commercial and industrial property far below its true market value (the Bahl study). The Bahl study, discussed in greater detail later in this opinion, was based upon national data that reflected the estimated value of equipment nationwide, obtained from the United States Bureau of Economic Analysis (BEA). The study broke down the BEA data to state and county levels, using employment figures for different types of industries, in order to estimate the total market value of commercial and industrial property located in the county during the 1993-94 tax year. The airlines also presented testimony by a county auditor that supported their contention that a large amount of property was not reported by its owners and, therefore, was neither assessed nor taxed by the county.

The department presented testimony by employees of both the county and the department that described the assessment and auditing practices of the county and, in the case of utilities and high-value industrial accounts, the department.[7] The department also presented a personal property ratio study that it had conducted in response to the airlines' claim of discrimination. The department's study, also discussed in greater detail later in this opinion, was based upon a sampling of 123 county accounts, all of which were audited by the department as part of its study. To a small extent, the department's study accounted for some property that had not been reported to the county. After conducting a statistical analysis of the audit results, the department's study concluded that, during the 1993-94 tax year, the county assessed other commercial and industrial property at nearly 100 percent of its true market value.

The Tax Court concluded that "the airlines failed to show that their air carrier transportation property is assessed at a higher ratio to its [true] market value than the ratio of other commercial and industrial personal property" and, consequently, failed to prove a violation of 49 U.S.C. section 1513(d). Id. at 371. Accordingly, the Tax Court entered a judgment in favor of the department. Ibid.

The airlines appealed to this court. ORS 305.445. We review the Tax Court's decision de novo. ORS 19.125(3). The airlines have the burden of proving their claims by a preponderance of the evidence. ORS 305.427.

II. 49 U.S.C. SECTION 1513(d)

49 U.S.C. section 1513(d) provides, in part:

"(1) The following acts unreasonably burden and discriminate against interstate commerce and a State, subdivision of a State, or authority acting for a State or subdivision of a State may not do any of them:

"(A) assess air carrier transportation property at a value that has a higher ratio to the true market value of the air carrier transportation property than the ratio that the assessed value of other commercial and industrial property of the same type in the same assessment jurisdiction has to the true market value of the other commercial and industrial property;

" * * * *

"(2) In this subsection

"(A) `Assessment' means valuation for a property tax levied by a taxing district;

"* * * * *

"(C) `air carrier transportation property' means property * * * owned or used by an air carrier providing air transportation; *179 "(D) `commercial and industrial property' means property, other than transportation property and land used primarily for agricultural purposes or timber growing, devoted to a commercial or industrial use and subject to a property tax levy[.]" (Emphasis added.)

Subsection (d)(1)(A) prohibits discriminatory taxation of air carrier transportation property by requiring states to assess such property at a ratio to true market value that does not exceed the ratio of assessed value to true market value for "other commercial and industrial property." In other words, the "assessment ratio" for airline property must not be higher than the "assessment ratio" for other similar property. The process by which tax assessing authorities ensure that the former does not exceed the latter is referred to as "equalization."

As noted, the parties stipulated before the Tax Court that the department assessed the airlines' property at 100 percent of its true market value. The equalization issue in this case, therefore, involves the determination of the assessment ratio for the "other commercial and industrial property" located in the county for the 1993-94 tax year, which we refer to in this opinion as the "comparison class." [8]

Oregon law requires that non-exempt personal property be assessed at 100 percent of its true market value. ORS 308.232 and 308.250(1).[9] In accordance with that statutory mandate, the assessment ratio for the comparison class should be 100 percent, the equivalent of the airlines' assessment ratio. The airlines contend, however, that the assessment practices followed by both the county and the department resulted in the substantial undervaluation of property in the comparison class. The airlines further argue that, according to their evidence, the comparison class was assessed at only 75.5 percent of its true market value. The department responds that the assessment ratio for the comparison class was between 97.7 and 100.65 percent. We address the issues presented by those competing contentions below.

III. DETERMINING WHETHER "ESCAPED PROPERTY" IS INCLUDABLE IN THE COMPARISON CLASS

A. Interpretation of 49 U.S.C. section 1513(d)

The airlines first contend that the true market value of all "escaped property," that is, comparison class property that is not assessed because property owners fail to report it to tax assessing authorities, still should be included in the comparison class for the purpose of determining whether the department assessed the airlines' property in a discriminatory fashion. In so contending, the airlines focus upon the definition of other "commercial and industrial property" set forth in 49 U.S.C. section 1513(d)(2)(D):

"`commercial and industrial property' means property, other than transportation property and land used primarily for agricultural purposes or timber growing, devoted to a commercial or industrial use and subject to a property tax levy [.]" (Emphasis added.)

In the airlines' view, escaped property still is "subject to a property tax levy" (emphasis added), despite the fact that it escapes taxation *180 due to nonreporting by its owners. It follows, they argue, that the department should add the value of such property to the total true market value of "other" reported property, in order to determine the assessment ratio for the comparison class. Such an increase in the total true market value of the comparison class property would lower the assessment ratio for that class, supporting the airlines' claim of discrimination under section 1513(d).

The department responds that a violation of section 1513(d) requires a state action and that, because the airlines' reading of subsection (d)(2)(D) relies upon the failure of nongovernmental property owners to report their property, that argument must fail.[10] The Tax Court agreed that escaped Property was not includable in the comparison class. 13 OTR at 366-67.

The issue before us is one of statutory construction. That is, we must determine whether Congress intended, for the purpose of comparing assessment ratios under 49 U.S.C. section 1513(d), that the phrase "property * * * subject to a property tax levy" in subsection (d)(2)(D) should include nonexempt property that is located within the assessment jurisdiction, but that is not taxed because its value is not reported to tax assessing authorities. In interpreting that wording, we examine the statute's text and structure and, if necessary, its legislative history. See Department of Revenue of Ore. v. ACF Industries, Inc., 510 U.S. 332, 339-46, 114 S. Ct. 843, 847-51, 127 L. Ed. 2d 165 (1994) (examining the text, structure, and legislative history of a similar federal enactment); Burlington No. R. Co. v. Okla. Tax Comm'n, 481 U.S. 454, 461, 107 S. Ct. 1855, 1860, 95 L. Ed. 2d 404 (1987) (Court refused to examine legislative history when the text of a similar federal enactment "plainly declare[d] the congressional purpose"); see also Shaw v. PACC Health Plan, Inc., 322 Or. 392, 400, 908 P.2d 308 (1995) (court followed United States Supreme Court's methodology when interpreting a federal statute).

A textual analysis of the words "property * * * subject to a property tax levy" demonstrates the conflicting interpretations advanced by the parties. It is plausible, as the airlines contend, that all escaped property is includable in the comparison class because, although such property is not taxed, it still is "subject to" taxation in that it is not exempt from taxation. In reality, however, no tax will be levied upon escaped property because that property is not assessed. Under that latter reading of the statute, escaped property is not "property * * * subject to a property tax levy" and, therefore, should not be considered as part of the comparison class.

The United States Supreme Court has not analyzed the specific provision of section 1513(d) at issue in this case. We note, however, that section 1513(d) is modeled after similar provisions in two other federal enactments: Section 306 of the Railroad Revitalization and Regulatory Reform Act of 1976 (4-R Act), codified as 49 U.S.C. section 11503, and section 31 of the Motor Carriers Act of 1980 (MCA), codified as 49 U.S.C. section 14502.[11] Case law interpreting those provisions and their legislative history therefore are helpful to our construction of section 1513(d). See Western Air Lines v. Board of Equalization, 480 U.S. 123,130-31, 107 S. Ct. 1038, 1042-43, 94 L. Ed. 2d 112 (1987) (when interpreting another provision of section *181 1513(d), Court examined the legislative history of the 4-R Act); ABF Freight System, Inc. v. Tax Div. of Arkansas, 787 F.2d 292, 293 n. 1 (8th Cir.1986) (because the MCA "was patterned after and is virtually identical to" the 4-R Act, "cases construing the 4-R Act and its legislative history are relevant to an analysis of the MCA").

The United States Supreme Court has examined the text of the corresponding provision of the 4-R Act, although in a context different from that before us now. In ACF, the issue before the Court was whether exempt property was "property * * * subject to a property tax levy" under the 4-R Act. 49 U.S.C. § 11503(a)(4). In examining that text, the Court stated:

"The statute does not define th[e] phrase [`subject to a property tax levy'], which on its face could bear one of two interpretations: (1) taxed property; or (2) taxable property, a broader category consisting of the general mass of property within the State's jurisdiction and power to tax, including property that enjoys a current exemption." 510 U.S. at 341, 114 S. Ct. at 848-49.

The court concluded, after examining the context and structure of the 4-R Act, that "the words `property subject to a property tax levy' must mean `taxed property.'" Id. at 342, 114 S. Ct. at 849 (emphasis added). Consequently, because exempt property is, by definition, not taxed property, the Court further concluded that exempt property was not includable in the comparison class. Ibid.

At first glance, escaped property should fall within the latter category described in ACF, that is, it is not taxed property but, rather, is property within the state's jurisdiction and power to tax. Under that reading of ACF, because the Supreme Court held that only taxed property is includable in the comparison class, the airlines' argument fails.

Escaped property, however, is not the same as exempt property. In American Airlines v. County of San Mateo, 12 Cal. 4th 1110, 51 Cal. Rptr. 2d 251, 912 P.2d 1198 (1996), the California Supreme Court was faced with an issue similar to that before us here. In rejecting the argument that, under ACF, escaped property should be treated the same as exempt property, that court explained:

"Here, the Counties assert that `[l]ike exempt property, personal property that inadvertently escapes taxation (notwithstanding the good faith efforts of the assessor) is not "taxed property" and therefore should not be included as part of the comparison class.' This is not entirely correct. Escaped property, unlike exempt property, is by definition taxable; however, that tax has not been assessed or paid." 12 Cal. 4th at 1133, 51 Cal. Rptr. 2d at 266, 912 P.2d at 1214 (emphasis added).

We agree that escaped property should not necessarily be treated the same as exempt property under the reasoning set forth in ACF. Instead, escaped property logically could be viewed as property that is "subject to a property tax levy."

However, another provision of section 1513(d) supports the department's interpretation. Subsection (d)(2)(A) provides that, for the purpose of section 1513(d), "`[a]ssessment' means valuation for a property tax levied by a taxing district." (Emphasis added.) That definition suggests that an "assessment" is part of the procedure that results in the taxation of property. The root word "assess" is used in subsection (d)(1)(A), which prohibits assessing airline property at a higher ratio "than the ratio that the assessed value of other commercial and industrial property * * * has to the true market value of the other commercial and industrial property." (Emphasis added.) When viewed in conjunction with the definition of "assessment" in subsection (d)(2)(A), the words "assessed value" in subsection (d)(1)(A) refer to the value that will be used to impose taxes upon other commercial and industrial property. Under that reading, the phrase "other commercial and industrial property," that is, other such property "subject to a property tax levy," includes only nonexempt property that eventually will be taxed, i.e., property that has been reported to assessing authorities.

As part of our textual and structural analysis, we also note that, unlike section 1513(d)(2)(A), section 306 of the 4-R Act *182 originally defined "commercial and industrial property" as "all property * * * which is devoted to a commercial or industrial use and which is subject to a property tax levy." Pub.L. 94-210, § 306, 90 Stat. 55 (1976) (emphasis added). That section also prohibited states from assessing railroad property at a higher ratio to true market value "than the ratio which the assessed value of all other commercial and industrial property in the same assessment jurisdiction bears to the true market value of all such other commercial and industrial property." Pub.L. 94-210, § 306, 90 Stat. 54 (1976) (emphasis added). In light of that original wording, the omission of the word "all" in section 1513(d) suggests that Congress intended the definition of the comparison class in that provision to be read less broadly than the corresponding provision in the 4-R Act. As explained below, however, the context of the 4-R Act does not support that conclusion.

Section 306 of the 4-R Act was recodified in 1978, as part of a comprehensive revision to the Interstate Commerce Act. See 325 Or. at 539 n. 11, 943 P.2d at 180 n. 11 (discussing and citing that revision). At that time, Congress deleted the word "all" from section 306, as quoted in the preceding paragraph. Pub.L. 95-473, § 11503, 92 Stat. 1445-46 (1978). In passing the recodification, Congress specifically noted that it "ma[de] no substantive change in the law." H.Rep. No. 1395, 95th Cong., 2d Sess., reprinted in 1978 U.S.Code Cong. & Ad.News 3009, 3018. See also Burlington No. R. Co., 481 U.S. at 457 n. 1, 107 S. Ct. at 1858 n. 1 (the recodification "may not be construed as making a substantive change in the laws replaced" (internal quotation marks omitted)); Burlington Northern R. Co. v. James, 911 F.2d 1297, 1298-99 (8th Cir.1990) (citing the original version of section 306 because the recodification made no substantive changes). Consequently, the deletion of the word "all" from section 306 and the subsequent adoption of the revised wording in section 1513(d) cannot be significant in our interpretation of that provision. See State of Ariz. v. Atchison, T. & S.F.R. Co., 656 F.2d 398, 404 (9th Cir.1981) (rejecting a different argument that also relied upon the word "all" in the original 4-R Act).[12]

We conclude that, for the purpose of comparing assessment ratios, the text and structure of section 1513(d) do not clearly reveal whether escaped property is includable in the comparison class. We now turn to the legislative history to determine whether Congress spoke to that question.

Congress enacted section 1513(d) in 1982 as part of the Airport and Airway Improvement Act (AAIA),[13] which was "part of a series of congressional actions dedicated to improving the Nation's air transportation system." Western Air Lines, 480 U.S. at 124, 107 S. Ct. at 1039. The AAIA made certain changes to an earlier federal enactment, the Airport Development Acceleration Act of 1973. One change was to prohibit states from imposing discriminatory taxes upon airline property. Id. at 125, 107 S. Ct. at 1039-40. As noted, section 1513(d) was patterned after similar provisions relating to railroads and motor carriers, set forth in the 4-R Act and the MCA.

Most of the relevant legislative history was developed during the 15 years before Congress enacted the 4-R Act in 1976.[14]*183 The significant history begins in 1961, with a Senate-sponsored study conducted by the Special Study Group on Transportation Policies in the United States (the Doyle Report). The Doyle Report analyzed many aspects of national transportation policy, including the practice and effect of discriminatory state taxation of interstate carriers, and recommended curative legislation. S.Rep. No. 445, 87th Cong., 1st Sess. (1961), 451-91. Following publication of that report, legislators introduced several bills over the next six legislative sessions, containing provisions substantially similar to those eventually included in the 4-R Act, the MCA, and the AAIA, that would have amended the Interstate Commerce Act to prohibit discriminatory state taxation of interstate carriers.[15] As have other courts, we look to that collective legislative history to assist our statutory analysis. See Western Air Lines, 480 U.S. at 130-31, 107 S. Ct. at 1042-43 (examining that history when interpreting another provision of section 1513(d)); Atchison, T. & S.F.R. Co., 656 at 404 n. 6 (because the earlier bills "are nearly identical to section 306 [of the 4-R Act] as it was enacted," the legislative history of those bills is helpful in interpreting section 306).

We begin with a general overview of the purpose of section 1513(d) and its counterparts in the 4-R Act and the MCA. In Western Air Lines, the Supreme Court observed:

"The legislative history of the antidiscrimination provision in the 4-R Act demonstrates Congress' awareness that interstate carriers `are easy prey for State and local tax assessors' in that they are `nonvoting, often nonresident, targets for local taxation,' who cannot easily remove themselves from the locality. S.Rep. No. 91-630, p. 3 (1969). The Department of Transportation had observed that `State and local governments derive substantial revenues from taxes on property owned by common carriers.' Id., at 4. It is this temptation to excessively tax nonvoting, nonresident businesses in order to subsidize general welfare services for state residents that made federal legislation in this area necessary." 480 U.S. at 131, 107 S. Ct. at 1043.

Congress intended the legislation to prohibit discriminatory state assessment practices, as well as the imposition of higher tax rates upon interstate carriers' property. See S.Rep. No. 630, 91st Cong., 1st Sess. (1969), 21 (the purpose of the legislation is to require states to treat interstate carriers' property "in the same fashion as other such property"); S.Rep. No. 1483, 90th Cong., 2d Sess. (1968), 3 (discrimination can arise in the form of higher assessments and tax rates). Further, the legislation prohibits both de jure and de facto discrimination. See Louisville & Nashville R. Co. v. Dept. of Rev., Etc., 736 F.2d 1495, 1498 (11th Cir.1984) (so stating). De facto discrimination occurs "when a single rate of tax is applied to full true market value for [interstate transportation] property but to something less than true market value for non-[interstate transportation] property." Ibid. Moreover, discriminatory intent is not required for a violation to occur. Rather, the statutes prohibit any act that results in discriminatory assessment or taxation. See Burlington No. R. Co., 481 U.S. at 463-64, 107 S. Ct. at 1861 (4-R Act does not require a showing of discriminatory intent).

After reviewing the legislative history, we conclude that Congress did not specifically address the issue before us, that is, whether the comparison class should include the value of all escaped property. The legislative history *184 does, however, provide some assistance in making that determination.

First, it is clear that discriminatory assessment practices, a form of de facto discrimination, were of great concern to Congress when it drafted the legislation. See S.Rep. No. 1085, 92d Cong., 2d Sess. (1972), 4 (one cause of state discrimination "lies in long-established procedures for assessing property"); S.Rep. No. 445, 87th Cong., 1st Sess. (1961), 485 (discrimination can result from the lack of prescribed standards for assessing property). Local assessment practices particularly were scrutinized. See S.Rep. No. 1483, 90th Cong., 2d Sess. (1968), 3 (local assessment can permit considerable variation in the valuation of comparison class property).

The legislative history further reveals that the spirit of the legislation was to eradicate all forms of discriminatory assessment against interstate carriers. See 116 Cong. Rec. S2023 (January 30, 1970) (statement of Senator Clifford Hansen) (the legislation "will encourage fair and equitable taxation in those few instances where localities either overtly or covertly tax transportation property at a higher * * * assessed valuation than for other segments of the community"). Because the airlines contend here that they are subject to discriminatory assessment practices, it appears that Congress would have intended section 1513(d) to protect them from such practices.

Further, the legislative history provides some indication that Congress intended the comparison class to include all non-exempt property located within an assessment jurisdiction, possibly including all escaped property. For example, the Doyle report gave the following description of comparison class property that is "subject to" taxation:

"All tangible property * * * is subject to ad valorem property taxation in most States, three characteristics of which are:

"(a) The tax is legally applicable to all property not specifically exempt;

* * * * *

"(c) The tax rate of each government is uniform on all taxable property within its jurisdiction." S.Rep. No. 445, 87th Cong., 1st Sess. (1961), 451 (emphasis added).

That report further stated that "[t]he practical objective in equalization is to appraise all property in a jurisdiction at a constant level of opinion as to market value and keep all properties in their proper relationship one to the other." Id. at 457 (emphasis added). The report also used several similar phrases interchangeably to refer to property in the comparison class. See, e.g., id. at 448 ("other property * * * in the same jurisdiction"); id. at 460 ("property subject to general property taxation"); id. at 465 (property "subject to the same local tax rates"). Those varying descriptions could support the conclusion that the phrase "property * * * subject to a property tax levy" means all non-exempt property, whether or not that property is reported. At the least, that legislative history indicates that Congress assumed that all "other" property owners fully report their property to tax assessing authorities.[16]

The legislative history discussed thus far supports the airlines' contention that the value of all escaped property is includable in the comparison class. Other aspects of the legislative history, however, support the opposite conclusion.

First, we again note that Congress considered several causes of discriminatory state assessment and taxation, including a number *185 of specific problems with local assessment practices.[17] It also considered discrimination caused by state property classifications and inadequate equalization procedures. S.Rep. No. 1483, 90th Cong, 2d Sess. (1968), 3, 5. However, the deficiency alleged in this casethat a significant amount of "other" property escapes assessment and taxation while the airlines' property is more fully assessedis not mentioned in the legislative history as a problem leading to the need for legislation. Such a lack of discussion supports the conclusion that Congress did not intend to alter assessment systems that are based upon self-reporting. See ACF, 510 U.S. at 344, 114 S. Ct. at 850 (in concluding that exempt property was not part of the comparison class, the Court noted that the legislative history did not suggest that Congress was concerned with state exemption policies); Union Pacific R. Co. v. Public Utility Com'n, 899 F.2d 854, 857 (9th Cir. 1990) (in concluding that levies used only to finance railroad regulations were not a "tax" under the 4-R Act, the court found it significant that the legislative history neither mentioned that common state financing method nor suggested that it was a problem that the 4-R Act was intended to solve).

Moreover, the legislative history indicates that Congress did not intend the legislation to have a significant impact upon the states' manner of assessing and collecting taxes. For example, one report concluded:

"The committee wishes to emphasize that this bill would in no way alter the freedom of a State to tax its taxpayers so long as interstate carriers are accorded equal tax treatment with other taxpayers. In the majority of States that now grant equal justice to all taxpayers, State property tax assessments, collections or rates would in no way be affected by passage of this bill." S.Rep. No. 630, 91st Cong., 1st Sess. (1969), 15 (emphasis added).[18]

See also S. Rep. No. 1483, 90th Cong., 2d Sess. (1968), 14 (same). In light of that legislative history, it again appears that Congress did not intend to alter the practice of assessing property based upon a self-reporting system.[19]

*186 In short, the legislative history lends support to the contentions of both the airlines and the department. Although Congress may not have considered the issue of escaped property, it clearly intended to prohibit assessment practices that result in discrimination against interstate carriers.

Two judicial decisions assist us in applying section 1513(d) in this case, in a manner consistent with the legislative history. First, in Burlington No. R. Co., the Supreme Court observed in dictum that discriminatory practices prohibited by section 306 of the 4-R Act include "the systematic undervaluation of other commercial and industrial property." 481 U.S. at 463, 107 S. Ct. at 1861 (emphasis added). Governmental assessment practices that regularly allow a significant amount of comparison class property to escape assessment and taxation, be they intentional or unintentional, therefore may violate section 1513(d).

Second, in American Airlines, the California Supreme Court also examined the issue whether the undervaluation of "other" property, due to lack of reporting, constituted discrimination under section 1513(d). After reviewing the legislative history and the interpretation in ACF of the words "subject to a property tax levy," that court concluded:

"[D]iscrimination under former section 1513(d) is not demonstrated merely by the existence, through no fault of the assessor, of escaped commercial and industrial personal property. By its terms, former section 1513(d) applies to `a State, subdivision of a State, or authority acting for a State or subdivision of a State,' not fraudulent taxpayers. Hence, the action or inaction complained of under former section 1513(d) must be one ascribable to government officials.

"Moreover, any claim of de facto discrimination based on the unequal enforcement of otherwise neutral tax laws must consist of some practice or policy on the part of the assessor that has a discriminatory impact or effect. * * * As with legislation exempting specific personal property, so long as those enforcement decisions are reasonable, and are neither targeted to disadvantage nor have a disparate effect on a particular group such as the Airlines, we do not believe that Congress intended to condemn the resulting assessment ratios as acts that `unreasonably burden and discriminate against' interstate commerce within the meaning of former section 1513(d)." 12 Cal. 4th at 1133-34, 51 Cal. Rptr. 2d at 266-267, 912 P.2d at 1214 (emphasis added).

We agree with the California Supreme Court that the failure of property owners to report their property to tax assessing authorities, by itself, cannot result in a violation of section 1513(d). Rather, there must be a showing that state or county assessment practices result in the "systematic undervaluation," Burlington No., 481 U.S. at 463, 107 S. Ct. at 1861, of comparison class property. That conclusion is consistent with the text of section 1513(d) and the legislative history, which focuses upon discriminatory assessment and taxation of interstate carriers by assessing and taxing authorities.

At bottom, the question still remains whether, within that framework, the value of all escaped property should be included in the comparison class. In American Airlines, the California Supreme Court concluded:

"[T]he [plaintiffs] must demonstrate that any unassessed * * * personal property results from a practice or policy on the assessor's part that is either targeted to disadvantage or that has an unreasonably disparate effect on the [plaintiffs]. Under these circumstances, unassessed * * * property [then] is properly included in the comparison class * * *." 12 Cal. 4th at 1136, 51 Cal. Rptr. 2d at 268, 912 P.2d at 1215-16 (emphasis added).

We agree with that approach. Because section 1513(d) requires a state action, escaped property should not be included in the comparison class merely because some property owners fail to report their property. State and local assessors are, however, responsible for assessing all nonexempt property, which includes making an effort to ensure that all such property is accounted for. See ORS 308.210(1) (providing, in part, that local assessors "shall proceed each year to assess the value of all taxable property within the county, except property that by *187 law is to be otherwise assessed." (emphasis added)); ORS 308.515(1) (requiring the department to assess certain utility and industrial properties). If, but only if, state and local assessment practices, either intentionally or unintentionally, systematically undervalue such property, resulting in discriminatory treatment of the airlines, then the value of all escaped property located within the county must be included in the comparison class, for the purpose of determining the assessment ratio for that class.

B. Evidence Relating to Assessment and Valuation

We now must analyze the evidence to determine whether, during the 1993-94 tax year, commercial and industrial personal property located in the county was systematically undervalued by either the county or the department. As we shall explain, we conclude that the assessment practices followed by the county and the department, although problematic in some respects, did not rise to the level of systematic undervaluation.

1. Assessment Practices Followed by the County and the Department

We begin by describing the county's assessment practices during the 1993-94 tax year. All commercial and most industrial property owners must file personal property returns with the county each year, which list all the owners' assets and the original cost of those assets, among other things.[20] When it receives the returns, the county assigns a value to each reported asset, using valuation schedules developed by the department, with some additional adjustments by the county. Those schedules, which are based upon Issimarket studies conducted by the department for different types of assets, adjust the original cost of the assets for both depreciation and inflation. The resulting value for each assetits "assessed value"provides the basis for imposing a tax upon the property owner.

If the county did not receive a return for an existing account during the 1993-94 tax year, it carried the total assessed value forward from the earlier year's filing, together with any appropriate adjustments. If no earlier personal property return were available and the property owner continuously failed to file, then the county filed a "dummy" return for that owner, based upon an onsite viewing of the owner's property.

Accounts routinely were flagged if anything on the returns appeared suspicious, such as illogical property listings or substantial changes in value from previous years. Employees followed up on those flagged accounts through the course of the year, generally by contacting the taxpayer and, if necessary, by reviewing the books and records of the company, i.e., conducting a book audit, to verify the original costs and current values of the assets. Accounts with dummy returns were set aside for similar follow-up.

The county employs one full-time auditor who, during the 1993-94 tax year, conducted random audits of personal property accounts with total assessed values exceeding $100,000. Those audits always consisted of book audits and sometimes included on-site visits, during which the auditor would spot-check the assets. Other appraisal staff members also conducted audits to ensure compliance with reporting requirements.

The county followed certain practices during the 1993-94 tax year for discovering new accounts to add to the county's tax roll. One employee, who specifically was responsible for discovering new accounts, regularly traced returned mail and reviewed other sources of information, including the Portland business listings, newspapers, business journals, and Uniform Commercial Code filings. The appraisal staff also conducted field verifications, covering about one-third of the accounts each year, during which the staff visited every known business address to verify its current occupant. Those visits sometimes resulted in flagging an account for audit, if an appraiser thought that the business might own but did not report property.

*188 According to the record, a five-year adjustment period is in place for all the accounts on the county's tax roll. Consequently, the county may look for escaped property for up to five years after a certain tax year and make necessary adjustments to the assessed values for that year. Audits also frequently go back as much as five years.

Finally, the department conducts performance reviews of all county assessment offices in the state, to evaluate their assessment procedures, among other things, and to recommend improvements when necessary. In a 1993 review, the department concluded that the county could improve some of its assessment procedures. Those recommendations, however, did not relate to the discovery of new accounts or, for the most part, escaped property.[21] An expert witness also testified that the county's discovery and assessment procedures were consistent with standards developed by the International Association of Assessing Officers (IAAO), an organization with expertise in the area of property assessment.

The above-described evidence supports the tentative conclusion that the county sufficiently valued commercial and industrial property during the 1993-94 tax year. The following evidence, however, demonstrates that escaped property posed a problem for the county.

First, as noted above, the county employed only one full-time auditor to conduct random audits during the 1993-94 tax year. Because that auditor reviewed only those accounts with total assessed values exceeding $100,000, smaller accounts were not audited unless a specific problem was noted during the assessment process. The auditor testified that it was most cost effective to audit the larger accounts because it was more likely that they contained large errors, which, in turn, would generate more revenue for the county.

The department's ratio study, however, concluded that "it was in the smaller value ranges that the relatively overvalued and undervalued accounts were found."

Further, the auditor selected accounts based upon his judgment, rather than according to any objective standard or statistically random process. He also conducted on-site visits only if the property owner requested that he visit the business in order to review the books and records. Once on the premises, however, he would not specifically look for nonreported property or inspect assets to determine their value. Between 1989 and 1993, the auditor discovered an average of $14,208,112 worth of escaped property each year. He testified that, if the county had had more auditors, then more escaped property likely would have been discovered during the 1993-94 tax year.

With the exception of those random audits of larger accounts, the county generally would not audit an account unless a return contained irregularities. Consequently, there was little likelihood that most property owners would be subject to an audit. Moreover, the county often conducted only book audits of questionable accounts, rather than conducting book audits together with on-site inspections.

The lack of meaningful penalties also might have caused some property to escape assessment during the 1993-94 tax year. The statutory penalty for late filing of a personal property return was $1[22] for each $1,000 of assessed value, with a minimum penalty of $10 and a maximum of $250. ORS 308.295(2). The penalty for nonfiling was $10 per day. ORS 308.300. Although the county regularly imposed the late filing penalty, it never imposed the nonfiling penalty.[23] Further, no penalty was imposed for knowingly providing incorrect information on a *189 personal property return.[24] The lack of significant penalties, particularly when considered in light of the small likelihood of being audited, supports the airlines' contention that there was little incentive for property owners to report property information truthfully to the county.

As can be seen, the evidence demonstrates that, during the 1993-94 tax year, the county did not assess all commercial and industrial personal property. On balance, however, the evidence does not demonstrate that the county systematically undervalued commercial or industrial property. The county followed accepted and sound procedures for assigning assessed values to reported property, including the application of market-based valuation schedules. Further, although escaped property posed a problem, the county had several procedures in place to discover and assess such property. For example, the county routinely discovered new businesses and, therefore, new assets, through its discovery, verification, and auditing practices. The county also has five years to locate and assess escaped property and adjust accounts accordingly. Although those procedures offered a somewhat hodgepodge approach, taken together, they formed a complete, albeit imperfect, assessment system that regularly led to the discovery and assessment of escaped property during the 1993-94 tax year. In short, although it is clear that the county did not discover and assess all commercial and industrial property, we conclude that that failure is not indicative of a systematic assessment problem.

The airlines also contend that the department did not fully value industrial property from high-value accounts located in the county during the 1993-94 tax year. The department's assessment procedures were similar to the county's, in that the department relied upon information contained in industrial property reports filed by property owners and also applied its valuation schedules to determine an assessed value for each reported asset. The department also conducted on-site appraisals to verify property values. If a return were not filed, then the department carried forward the total assessed value from the prior year and assigned the account to an appraiser for investigation and completion of an on-site appraisal. The department used the same penalty structure as the county and relied upon the county for discovery of new accounts.

Because the department's procedures essentially mirror the county's, and because the airlines did not present any other evidence tending to show that escaped property posed a problem specifically for the department, we conclude that the department did not systematically undervalue industrial property from high-value industrial accounts during the 1993-94 tax year.

2. The Bahl Study

The airlines further contend that, according to the Bahl study, the assessment ratio for the comparison class was between only 50 and 75 percent. In the airlines' view, that is indicative of a systematic assessment problem. As we shall explain, we conclude that the Bahl study is not persuasive.

The Bahl study estimated the total market value of commercial and industrial personal property located in the county during the 1993-94 tax year, based upon national figures obtained from the BEA, which is a division of the United States Census Bureau. Through general surveys of manufacturers, the Census Bureau determines the total value of equipment sold to businesses nationwide during any given year. The BEA then uses that data to estimate the total value of capital stock in the United States, by examining the total value of all equipment, broken down in detail by industry and by type of asset. The BEA's purpose in compiling and analyzing that data is to calculate the gross national product (GNP)the measure of total output for the national economyand to calculate other national product and income figures.

The BEA data are compiled on a national level only. Consequently, the Bahl study relied upon employment information obtained from the Census Bureau, in order to break the data down to state and local levels. That *190 employment information set forth the number of workers employed by different industries in each state and county in the nation. The study then divided the total amount of equipment in each industry by the number of workers employed in each industry in both the state and the county, to calculate a "capital-labor ratio" for each industry in the state and the county. In theory, that ratio represented the average amount of equipment that supported each worker in a particular industry during the 1993-94 tax year.[25] Finally, the Bahl study calculated an estimated value of equipment for each industry in the state and county, by multiplying the capitallabor ratio for each industry by the number of workers in the state and county for that industry. The sum of those estimated values for the county made up a final estimated value of the personal property located in the county during the 1993-94 tax year.

After calculating a total estimated market value, the Bahl study compared that figure to the county's assessment figures and preliminarily concluded that the comparison class was assessed at only about 22 percent of its true market value. The study then made several adjustments to the estimated market value figure, to make the data more consistent with state and county assessment practices.[26] After making those adjustments, the study concluded that the assessment ratio for the comparison class was 75.5 percent.

The airlines are correct that, if the Bahl study is accurate, the resulting ratio of 75.5 percent indicates that property in the comparison class might have been systematically undervalued. After reviewing all the evidence, however, we conclude that the Bahl study is not an accurate measure of the true market value of that property.

For example, the BEA classifies property as either "structures" or "equipment," in contrast to classifying it as real or personal property. Oregon's definition of real property includes a significant amount of manufacturing, industrial, and commercial equipment that the BEA classifies as equipment. Although the Bahl study excluded a substantial amount of that type of property, the differences in classification illustrate a significant problem with using the BEA data to estimate the value of personal property located in both Oregon and the county.

The BEA also uses different "service lives" from the department and the county to calculate depreciation. For many classes of assets, the BEA's service lives are longer, resulting in a higher estimation of market value in the early years of ownership. In respect of service lives, unlike the department and the county, the BEA does not account for certain forms of noneconomic obsolescence and also does not reflect the actual retirement of assets.

Most significantly, the Bahl study is based upon national estimated figures that are gathered for the purpose of analyzing the GNP and calculating other national product and income figures. The data are drawn from a general, voluntary, unsworn survey of manufacturers nationwide, with no assurance that the information provided is correct or complete. Further, the study relies upon the assumption that the capital-labor ratio for workers in the county is the same as the national average for each industry. Although that assumption may be reliable enough in the economic forecasting arena, the evidence demonstrated that it is problematic in calculating estimated values of property.[27]

*191 In short, the airlines contend that the comparison class should include the value of all personal property located in the county. Their calculation of that value, however, is based upon economic assumptions and estimated national data that are collected for an entirely different purpose. Stated another way, the airlines are trying to prove the value of apples with evidence consisting only of oranges.[28]

3. Conclusion

To summarize, we conclude that the airlines have not demonstrated that either the county or the department systematically undervalued property in the comparison class during the 1993-94 tax year. Accordingly, the value of all escaped property does not need to be added to the true market value of the comparison class, for the purpose of determining the assessment ratio for the comparison class. We now turn to the remaining issues surrounding the calculation of that ratio.

IV. CALCULATING THE ASSESSMENT RATIO FOR THE COMPARISON CLASS A. The Department's Ratio Study

As discussed earlier, the department conducted a ratio study to determine the county's level of assessment of comparison class property during the 1993-94 tax year. To complete its study, the department audited 123 county accounts, out of a total of 23,707 accounts, and used the results of those audits to estimate the true market value for all the comparison class property and to calculate a final assessment ratio for that class. The department's study used a stratified sampling that organized the 123 accounts into five separate strata based upon their assessed values, with the first stratum containing the largest accounts and the fifth stratum containing the smallest.[29] In designing and conducting its study, the department consulted with an expert on ratio studies, Dr. Gloudemans.

The department's audits consisted of conducting book and on-site audits of several assets randomly selected from the 1993-94 personal property return for each sample account (sample one), as well as an audit of several additional assets examined during an on-site inspection (sample two). The sample one audit primarily served to verify that the county had valued reported property correctly. The sample two audit served to broaden the sample base and to identify any additional assets that might have been omitted from the owner's personal property tax return.

After completing the audits, the department applied its valuation factors to calculate an audited value, i.e., an estimated true market value, for each sample of selected assets. For some accounts, the auditors also noted any lump sum adjustments that, in their judgment, would ensure that the results of the audit accurately reflected the true market value of all the property in the account. Next, the department applied a formula to those numbers, which was developed by the department and approved by Gloudemans, in order to estimate the total true market value for each sample account. The department then calculated an assessment ratio for each account, by dividing the total assessed value *192 for each account by the estimated true market value for each account.

The department next used the assessment ratios for each sample account to determine an overall assessment ratio for each of the five strata. To do so, the department calculated three measures of central tendency,[30] that is, three methods for determining the most central ratio figure, for each stratum: the median, the mean, and the ratio of aggregates. The median, or midpoint, simply represented the middle ratio figure in each stratum. The mean, or average, was calculated by adding all the ratios in each stratum and then dividing that figure by the number of accounts included in each stratum. The ratio of aggregates, which computes an assessment ratio based upon aggregated total sums, was calculated by adding the assessed values for all the accounts in each stratum and then dividing that figure by the sum of the estimated true market values for those accounts, as determined by the department's formula, mentioned in the preceding paragraph. The department then calculated three measures of central tendency for the entire sample, which ignored the effects of the stratification, in an analogous manner.

Finally, the department used the results from each stratum to calculate a composite result for the entire study that weighted the results under each stratum by the corresponding estimated market value of that stratum. The department first calculated an estimated market value for each stratum, again based upon each of the three measures of central tendency, by adding the total assessed values for the accounts in each stratum and dividing that figure by the median, mean, and ratio of aggregates for that stratum. To determine three final, weighted assessment ratios for the entire study, the department then added the assessed values of all 123 accounts and divided that figure by the sum of the estimated true market values of the five strata, calculated based upon the three measures of central tendency, as described in the preceding sentence. The results of the study are illustrated as follows:

             Assessed                 Number of                     Ratio
             Value of       Total       Sample                        of
             Accounts     Number of    Accounts   Median    Mean   Aggreg's
Stratum     in Stratum     Accounts      Drawn     (%)       (%)      (%)
----------------------------------------------------------------------------
  1        $3,520,680 +       38         16       100.00    101.16   100.12
  2        $951,250 to       162         20       100.00    101.93   101.36
            3,520,679
  3        $273,150 to       568         20        97.70     96.30    96.00
            $951,249
  4        $66,960 to       2,141        20       100.00     99.75   100.40
            $273,149
  5        $3,000 to        20,798       47       100.00    104.48    91.32
            $66,959
----------------------------------------------------------------------------
Totals                      23,707      123       100.00    101.53   100.11
----------------------------------------------------------------------------
  Final
Weighted                                           99.53    100.65    97.69
 Results

According to the department's study, then, the assessment ratio for the comparison class during the 1993-94 tax year was between 97.69 and 100.65 percent. The department contends here that, because the median is the most appropriate method of central tendency to use in this case, the assessment *193 ratio for the comparison class was 99.53 percent.

At trial, the airlines challenged the department's study in several respects. For example, the department's audit reports in some instances were not complete and, therefore, did not always reliably reflect accurate values of the sample assets. The evidence also demonstrated that the auditors did not conduct a sample two audit for some accounts. Further, the auditors made the lump sum adjustments, some of which were substantial, based solely upon their judgment, as opposed to following any established procedure. Finally, the airlines also challenged the validity of the department's formula for calculating the estimated market value for each sample account and demonstrated that it had been applied differently in a few instances.

We agree that the airlines have demonstrated several problems with the department's ratio study. However, of all the evidence presented at trial, we conclude that the department's study provides the best estimate of the assessment ratio for the comparison class. As discussed earlier, the Bahl study was based upon national figures compiled for the purpose of estimating the GNP and other national product and income figures, and is unpersuasive. By contrast, the department's study, even with its flaws, is based upon actual audits of personal property that is part of the comparison class. In our view, although neither study demonstrates precisely the value of comparison class property and thus of the accompanying assessment ratio, the department's study provides the more reliable estimate of those figures.

The airlines make two additional challenges to the department's study that may affect the final calculation of the assessment ratio for the comparison class. We now turn to those issues.

B. Appropriate Measure of Central Tendency

As explained above, the department calculated three measures of central tendencythe median, the mean, and the ratio of aggregatesfor each of the five strata in its study and for the study as a whole. The airlines do not challenge the department's calculation of the three measures for each individual stratum. They contend, however, that only the ratio of aggregates should be used to calculate an estimated market value for each stratum, in order to merge the data from the five strata into a final result. See 325 Or. at 560, 943 P.2d at 192 (explaining that calculation). It follows, they argue, that the final assessment ratio based upon the ratio of aggregates is the only valid result of the department's study.

The department responds that the selection of the appropriate measure of central tendency is a matter of judgment, depending upon the nature of the sample used. The department further contends that the final assessment ratio based upon the median is the most representative result of its study. The Tax Court agreed. 13 OTR at 369-70.

The evidence presented at trial offers some support for the contentions of each party. Gloudemans testified that the ratio of aggregates can be advantageous as a measure of central tendency because it weights each ratio based upon the estimated value of each sample account and, therefore, will reflect any systematic biases, represented by a series of extreme ratios, in either larger or smaller accounts. However, the ratio of aggregates also will be affected adversely by isolated, extreme ratios that are not reflective of a systematic assessment bias. On the other hand, according to Gloudemans, the median is not affected by extreme high or low ratios, be they isolated or reflective of a systematic bias. The median, therefore, is the more appropriate measure of central tendency if the sample is relatively small and if it contains isolated, extreme ratios.

Gloudemans also testified that other jurisdictions most predominantly use the median in ratio studies, rather than the ratio of aggregates. He concluded that the median was the most appropriate measure to use in this case, for two reasons. First, the sample size was relatively small. Second, because the study was stratified into five strata based upon equal assessed values, and because he had determined that no systematic assessment *194 biases existed within any of those strata, the advantages offered by the ratio of aggregates did not apply to the study.

Bahl testified that the ratio of aggregates is the only appropriate measure to use when calculating the estimated market value for each stratum, in order to merge the data into a final result. Although Gloudemans maintained that the final assessment ratio based upon the median was the most appropriate result here, he agreed that, in many cases, it is better to use the ratio of aggregates to calculate estimated market value.

The parties also offered different passages from the same IAAO text, Property Appraisal and Assessment Administration, to prove the appropriate measure of central tendency. Chapter 20 of that text, entitled "Sales Analysis and Mass Appraisal Performance Evaluation," which was written in part by Gloudemans, explains the advantages and disadvantages of each measure. The airlines emphasize the following passage, which relates to the general use of the ratio of aggregates, from that chapter:

"The [ratio of aggregates] weights each ratio in proportion to its [audited value], whereas the mean and median give equal weight to each [audited value].

"* * * * *

"Because of this weighting feature, the [ratio of aggregates] is the appropriate measure of central tendency for estimating the total dollar value of a population of parcels." International Association of Assessing Officers, Joseph K. Eckert, Ph. D., ed., Property Appraisal and Assessment Administration, 529-30 (1990) (emphasis added).

The department, however, points to an appendix to chapter 20 that describes the process of combining different groups of property from a stratified sample and states, in part:

"[T]he [ratio of aggregates] gives equal weight to each dollar and is appropriate if the objective is determination of full market value. If an equalization action is being contemplated, however, then the median could be used, because it is less influenced by outlier [i.e., isolated and extreme] ratios and thus tends to be the most stable measure of central tendency." Id. at 611, Appendix 20-5 (emphasis added).

Finally, the parties cite case law that they claim supports their competing contentions. We do not find that case law to be directly on point. For example, two circuits have concluded that the ratio of aggregates must be used to calculate a final assessment ratio in sales ratio studies conducted in 4-R Act litigation. Those cases, however, involved merging the values of different types of property assessed in different manners, such as real with personal property or centrally assessed with locally assessed property, a task for which the ratio of aggregates is better suited than the median. See Atchison, Topeka and Santa Fe Ry. Co. v. Lennen, 732 F.2d 1495, 1505 (10th Cir.1984) ("When, as here, categories of property that are subject to different methods of appraisal must be factored together to produce the assessment ratio for [the comparison class], each category should be factored in proportion to its share of the total true market value of all such property," by using the ratio of aggregates); ACF Industries, Inc. v. State of Ariz., 714 F.2d 93, 95 (9th Cir.1983) ("[T]he median approach would ignore legitimate classifications which are fundamental to Arizona's tax system."). In Southern Ry. Co. v. State Bd. of Equalization, 712 F. Supp. 1557, 1568-70 (N.D.Ga.1988), the court approved using the ratio of aggregates to estimate the value of real estate in the state's ratio study. However, the court emphasized that the ratio of aggregates was most appropriate in that case because the median would ignore the state's assessment trends. Id. at 1569.

The Fourth Circuit has approved the use of the median in determining final assessment ratios in two cases. However, the cases did not involve stratified samples and, consequently, are not helpful to our determination here. See CSX Transp., Inc. v. Board of Public Works, 95 F.3d 318, 324 (4th Cir. 1996) (because the properties at issue were of the same type and were assessed under the same assessment method, and due to practical variations in the assessment process and fluctuations in market values, the median was the most appropriate measure of central tendency); Clinchfield R. Co. v. Lynch, 700 *195 F.2d 126, 130 n. 5 (4th Cir.1983) (the standard practice in sales-assessment ratio studies is to use the median).

After considering all the evidence, we conclude that the department may use either the ratio of aggregates or the median to estimate the true market value of the property in each stratum, in order to merge the data and calculate a final assessment ratio. The most persuasive evidence relating to this issue is contained in the IAAO text, quoted above, which states that, in studies conducted for equalization purposes that use stratified samples, either the ratio of aggregates or the median may be used to estimate market value. In short, the evidence does not demonstrate that the ratio of aggregates, rather than the median, must be used exclusively. Consequently, we reject the airlines' contention to the contrary.

C. Exclusion of Account Number 83

The final issue in this case involves the department's decision to exclude one account, number 83, from its final calculations. The Tax Court concluded that that account properly was excluded. 13 OTR at 370. As we shall explain, we conclude that it was permissible for the department to exclude the account.

Account number 83 initially was included as a sample account in the fifth stratum of the department's ratio study. The department's audit of that account demonstrated that its total assessed value was far below its estimated market value. The department later learned the following history behind the account.

Before the 1993-94 tax year, the property owner for account number 83 had failed to file a personal property return. Consequently, a county appraiser assigned an assessed value of $2,950 to the account, based upon a viewing of the assets observable in the front area of the owner's business. The county coded the account, under its letter coding system, as having a non-final assessed value, and the owner was fined a $10 late-filing penalty. Because the owner again refused to file a personal property return in 1993, the county rolled the earlier assessed value forward, together with an additional adjustment, and assigned an assessed value of $3,050 to the account for the 1993-94 tax year. The county also changed the account status code to a "final," "penaltyno filing received" account.

After the account status code had been changed to "final," an auditor visited the site to audit account number 83 for the department's study. At that time, the department did not know the meaning of the county's letter status codes, including the codes assigned at various times to account number 83. The auditor concluded that the audited value for the account was $363,950, which was almost 12,000 percent higher than the county's assessed value. The resulting assessment ratio for account number 83 was 0.838, or less than one percent.

The department's initial ratio study results, which included the assessed and audited values for account number 83, demonstrated that, under the ratio of aggregates, the county assessed the comparison class property at only about 87 percent of its true market value. The initial median result, however, was not significantly affected. After consulting with Gloudemans, the department decided to draw 24 additional samples for the fifth stratum and, after completing those audits, failed to discover any accounts as grossly undervalued as number 83.[31] At about the same time, the department also contacted the county and learned the history of that account. After consulting again with Gloudemans, the department concluded that it was appropriate to exclude account number 83 from its study. The department then recalculated the results without that account and reached the conclusions set forth earlier in this opinion.

The airlines contend that the department improperly excluded account number 83 from its ratio study. They point to Bahl's testimony that the account was a valid data point and that the department violated accepted *196 statistical principles when it omitted a valid data point after drawing its sample. The airlines further argue that the department provided conflicting explanations as to why the account was excluded. That conflict suggests, in the airlines' view, that the department's exclusion of the account was suspect.[32] They also accuse the department of excluding the account for the sole reason of obtaining a favorable result under the ratio of aggregates.

The department responds that it properly excluded account number 83 from its study. The department emphasizes that the account was so highly undervalued, when compared to any other accounts in the study, that it must be "deemed * * * an aberration which was not at all representative of the other accounts within the [fifth] stratum." Consequently, including the account would skew the results of the study unfairly.[33] The department also contends that the account was analogous to accounts under appeal, which were excluded from the study at the outset due to their lack of finality. Finally, the department points to Gloudemans' testimony that, under standards developed by the IAAO, it is permissible in most cases to exclude extreme data points, such as account number 83, from a ratio study.[34]

We conclude that it was permissible for the department to exclude account number 83 from its study, for two reasons. First, Bahl and Gloudemans offered conflicting testimony, with no published documentation to support their respective contentions, that it either was or was not permissible to exclude the account. Faced with credible conflicting testimony from two qualified experts, we cannot conclude that the airlines have proved that excluding the account violated sound statistical principles. Second, we already have concluded that it is permissible to accept the department's final results based upon the median, rather than the ratio of aggregates, which was not affected by the extreme ratio of account number 83.

D. Final Ratio

After reviewing the department's ratio study and the trial record, and having considering all the airlines' contentions challenging the validity of the study, we agree with the department that its study showed that, for the 1993-94 tax year, the county assessed comparison class property at 99.53 percent, or nearly 100 percent, of its true market value.

V. CONCLUSION

We conclude that the airlines' personal property was assessed at the same ratio to true market value as the ratio of assessed to true market value for other commercial and industrial property located in the county during the 1993-94 tax year. We therefore hold that the department did not assess the airlines' property in a discriminatory fashion and, consequently, did not violate 49 U.S.C. section 1513(d).

The judgment of the Tax Court is affirmed.

NOTES

[1] That section was recodified without substantive change in 1994 as 49 U.S.C. section 40116(d). Pub.L. 103-272, 108 Stat. 745, 1111 (1994). We refer to the original codification, section 1513(d), throughout this opinion. The relevant parts of that statute are set forth below. See 325 Or. at 535-36, 943 P.2d at 178-79.

[2] ORS 307.030(1) provides:

"All * * * tangible personal property situated within this state, except as otherwise provided by law, shall be subject to assessment and taxation in equal and ratable proportion."

[3] The department's methods for valuing the airlines' property are not at issue in this case.

[4] According to the record, the department centrally assesses industrial personal property included in manufacturing accounts that have more than $1 million in value of real property.

[5] Oregon's self-reporting system is discussed in greater detail later in this opinion. See 325 Or. at 550-51, 943 P.2d at 186-87.

[6] Each airline challenged the department's assessments for different tax years. Collectively, those challenges covered the tax years between 1989 and 1994-95.

[7] For ease of reference, throughout this opinion we refer to the industrial property centrally assessed by the department as industrial property from "high-value" accounts. See 325 Or. at 532 n. 4, 943 P.2d at 177 n. 4 (describing the industrial property assessed by the department).

[8] The Tax Court determined that, in this case, the words "other commercial and industrial property of the same type," 49 U.S.C. § 1513(d)(1)(A), required comparing the airlines' personal property to "other commercial and industrial tangible personal property." Delta Air Lines, Inc. v. Dept. of Rev. (I), 13 OTR 357, 361, 1995 WL 617652 (1995). We agree with that conclusion and, noting that neither party disputes it, adopt it for the purpose of our analysis here.

[9] ORS 308.232 provides that "[a]ll real or personal property within each county shall be valued and assessed at 100 percent of its real market value." ORS 308.250(1) provides that "[a]ll personal property not exempt from ad valorem taxation shall be valued at 100 percent of its real market value, as of July 1, at 1:00 a.m. and shall be assessed at 100 percent of its real market value." Before the adoption of Ballot Measure 5 in 1990, Or. Const., Art. XI, § 11b, ORS 308.250(1) (1989) required property to be valued as of January 1, at 1:00 a.m.

Although those statutes use the term "real market value," we follow the wording of section 1513(d), which uses the term "true market value," noting no substantive difference for the purpose of this opinion.

[10] We note that, as mentioned earlier and discussed later in this opinion, the department's ratio study did account for the existence of some escaped property. However, the study made only relatively minor adjustments for nonreported property belonging to known taxpayers and did not account for any property belonging to unknown taxpayers. Consequently, the results under the department's study did not reflect the value of all escaped property that may have existed during the 1993-94 tax year.

[11] Section 306 of the 4-R Act was written as an amendment to the Interstate Commerce Act. Pub.L. 94-210, § 306, 90 Stat. 54 (1976). In 1978, Congress reorganized several provisions of the Interstate Commerce Act and related laws as subtitle IV of title 49, United States Code, entitled "Transportation." Pub.L. 95-473, 92 Stat. 1337 (1978). The MCA was enacted as part of that subtitle. Pub.L. 96-296, § 31, 94 Stat. 823 (1980).

The relevant provisions of the 4-R Act and a more detailed history of all the anti-discrimination legislation are set forth below. Unless otherwise noted, throughout this opinion we refer to the original section of the 4-R Act, section 306, rather than the later codification.

[12] The 4-R Act of 1976 also provided that section 306 was not to become effective until three years after the date of enactment. Pub.L. 94-210, § 306, 90 Stat. 54 (1976). Consequently, the original wording of section 306 never became effective. Rather, only the wording of the 1978 recodification was enacted. See State of Ariz. v. Atchison, T. & S.F.R. Co., 656 F.2d 398, 404 (9th Cir.1981) (discussing the recodification).

Different parts of the legislative history also use the adjective "all" to describe property in the comparison class. The significance of that legislative history is discussed in the text below. See 325 Or. at 545-46 and n. 16, 943 P.2d at 184 and n. 16.

[13] The AAIA was part of the Tax Equity and Fiscal Responsibility Act of 1982 (TEFRA), which was a comprehensive bill that enacted several changes to the tax laws. Pub.L. 97-248, 96 Stat. 324 (1982); S.Conf.Rep. No. 530, 97th Cong., 2d Sess. (1982), 1, 8.

[14] There is little legislative history relating to the motor and air carrier provisions. The legislative history for section 1513(d), set forth in the committee reports, provides:

"This section provides that States may not tax at a level which unreasonably burdens or discriminates against interstate commerce. The provision makes current law[,] which prohibits the assessment, levying, or collecting of taxes on motor carrier property in a manner different from that of other commercial and industrial property, applicable to air carriers."

H.Conf.Rep. No. 760, 97th Cong., 2d Sess., reprinted in 1982 U.S.Code Cong. & Ad.News 1190, 1484. See also S.Rep. No. 494, 97th Cong., 2d Sess., reprinted in 1982 U.S.Code Cong. & Ad.News 1156, 1188 (setting forth that same description).

[15] The following bills were introduced during that time frame, only three of which were reported out of committee: HR 736, 88th Cong., 1st Sess.; HR 4972, 89th Cong., 1st Sess.; S 2988, 89th Cong., 2d Sess.; HR 1480, 90th Cong., 1st Sess.; S 927, 90th Cong., 1st Sess.; S 2289, 91st Cong., 1st Sess.; S 3945, 92d Cong., 2d Sess.; and S 1891, 93d Cong., 1st Sess. None of those bills was enacted into law, and only S 2289 passed one chamber of Congress.

[16] Other parts of the legislative history similarly indicate that Congress intended the phrase "property * * * subject to a property tax levy" to be read broadly. See, e.g., H.Rep. No. 1069, 96th Cong., 2d Sess., reprinted in 1980 U.S.Code Cong. & Ad.News 2283, 2327 (section 31 of the MCA "prohibits the assessment, levying, or collecting of taxes on motor carrier property in a manner different from that of other commercial and industrial property" (emphasis added)); S.Rep. No. 1085, 92d Cong., 2d Sess. (1972), 6 (the legislation "would require the assessment of transportation property to be at the average assessment ratio of all property in the jurisdiction" (emphasis added)); Hearings on S 927 Before the Subcomm. on Surface Transportation of the S.Comm. on Commerce, 90th Cong., 1st Sess., 23 (August 7, 1967) (prepared statement of James N. Ogden, in behalf of the Association of American Railroads) ("True equalization requires that each parcel of property be assessed at a figure which is the same proportion of its full value as the assessment of every other parcel of property in the same taxing district is of its full value.") (emphasis added).

[17] The legislative history identifies several problems with local assessments. For example, the market value of locally assessed real property generally increases while the assessed value remains relatively constant until the property is sold. As a result, the assessed value does not reflect the true market value of the property until sale. S.Rep. No. 445, 87th Cong., 1st Sess. (1961), 458. In addition, local assessors often are under political pressure to keep assessed values relatively low. See Atchison, Topeka & Santa Fe Ry. v. State of Ariz., 559 F. Supp. 1237, 1244 (D.Ariz.1983) (the legislative history discusses the tendency of local assessment offices to bow to political pressures to keep locally assessed values down, despite state laws requiring assessments at market value); Hearings on S 927 Before the Subcomm. on Surface Transportation of the S.Comm. on Commerce, 90th Cong., 1st Sess., 68 (August 7, 1967) (testimony of Rolf A. Weil, President, Roosevelt University) (local assessors try to keep assessments down in order to keep taxes low and to help local school districts qualify for financial assistance). Similarly, central assessing agencies might be under political pressure to maintain a higher revenue flow, in the face of increasing governmental costs and tax-related needs. Id. at 57 (testimony of Broley E. Travis, Consulting Valuation Engineer).

[18] That wording was reiterated by Senator Vance Hartke, a co-sponsor of the legislation, during floor debate in the Senate. 116 Cong. Rec. S2024 (January 30, 1970).

[19] The following testimony in behalf of an earlier bill also supports that conclusion:

"Let me emphasize that [the legislation] does not suggest or require a State to change its assessment standards, assessment practices, or the assessments themselves. It merely provides a single standard against which all affected assessments must be measured in order to determine their relationship to each other. It is not a standard for determining value; it is a standard to which values that have already been determined must be compared. This standard is `true market value' (also the generally accepted standard for assessment purposes) and the requirement is that carrier property be assessed at the same proportion of such value as the proportion at which all other property subject to the same tax rates is assessed." Hearings on S 927 Before the Subcomm. on Surface Transportation of the S.Comm. on Commerce, 90th Cong., 1st Sess., 29 (August 7, 1967) (prepared statement of James N. Ogden, in behalf of the Association of American Railroads) (emphasis added).

See also Clinchfield R. Co. v. Lynch, 784 F.2d 545, 552 (4th Cir.1986) ("Certainly, the 4-R Act does not encroach upon, the State's right to tax its citizens as it sees fit, as long as that tax does not discriminate against railroads.").

[20] For new industrial accounts, the county conducts an initial site inspection to determine the amount of real and personal property owned by the property owner. Personal property returns filed thereafter update the information acquired by the county at that initial inspection.

[21] The department's review did conclude that the county took no action when basic taxable supplies were not reported and, accordingly, recommended the adoption of procedures concerning the assessment of such supplies.

[22] The department's witnesses testified that the late filing penalty was $10 for each $1,000 of assessed value. The amount set by statute, however, was $1 for each $1,000. ORS 308.295(2) (1993).

[23] A county valuation manager testified that the county never imposed a nonfiling penalty because it was difficult to prove intent to evade taxation on the part of the property owner. See ORS 308.300 (requiring intent).

[24] The county valuation manager further testified that "I believe the requirement is for an individual to file on time and there's no requirement for the * * * return to be correct."

[25] The capital-labor ratio assumes that the average worker in a particular industry in the state or county, regardless of the wage rate, must have the same amount of equipment available to perform his or her job as does the average worker in that industry nationwide.

[26] In making those adjustments, the Bahl study excluded a significant amount of data representing property that, in Oregon, would be or might be centrally assessed or categorized as real property. The study also deducted $3,000 from the estimated true market value for each commercial and industrial property owner in the county, because the county does not assess accounts that have less than $3,000 total value in personal property. Bahl made all the above-described adjustments because he intended the resulting assessment ratio under his study to be as conservative an estimate as possible.

[27] For example, a state economist testified that productivity per employee in Oregon is less than the national average, which might suggest that the capital-labor ratio for the state also is lower than the national average.

[28] Other courts have reached the same conclusion relating to studies conducted by Bahl and other economists, using BEA-based data. See Duluth, Missabe & Iron Range Ry. Co. v. State of Wis., 100 F.3d 69, 75 (7th Cir.1996) (in concluding that the lower court permissibly found that the BEA data did not provide a reliable basis for determining the value of taxable property in a 4-R Act case, the court stated that "data may be reliable for one purpose but not another"); Southern Ry. Co. v. Stair, 801 F. Supp. 37, 47-49 (W.D.Tenn.1992) (rejecting a similar study in a 4-R Act case); CSX Transp. v. Tennessee State Bd. of Equalization, 801 F. Supp. 28, 33-35 (M.D.Tenn.1992) (same).

[29] The department initially audited 100 accounts, 16 accounts in the first stratum, 20 in the second, third, and fourth strata, and 24 in the fifth stratum. After calculating results based upon that sampling, the department decided to sample an additional 24 accounts in the fifth stratum, because the ratios in that stratum varied significantly. One of the accounts in that stratum eventually was excluded from the final calculations, resulting in a final sample size of 123 accounts.

[30] A measure of central tendency "describe[s] the typical level of appraisal by a single number or statistic." International Association of Assessing Officers, Joseph K. Eckert, Ph.D., ed. Property Appraisal and Assessment Administration, 527 (1990).

[31] The department's witnesses also testified that the additional samples were drawn because the ratios in the fifth stratum varied considerably.

[32] As partial support for that contention, the airlines presented evidence that the department's final calculations included another account that had a background factually similar to account number 83.

[33] Gloudemans testified that, based upon its audited value, account number 83 should have been included in the third stratum, rather than the fifth stratum. He explained that, if the account was left in the fifth stratum, under the ratio of aggregates, it inappropriately would have received 20 percent of the total weight of the fifth stratum and over 10 percent of the weight of the study as a whole.

[34] As noted, the airlines contend that the department's conflicting explanations suggest that the decision to exclude account number 83 was suspect. After reviewing the record, we conclude that the department's explanations all are reasonable and do not demonstrate improper conduct on the department's part.