Union Pacific Railroad v. Dept. of Rev. , 11 Or. Tax 165 ( 1989 )

These cases concern the true cash value of plaintiff's¹ railroad property located in Oregon for the tax years 1983 and 1984. The accepted method of arriving at the true cash value of an integrated unit of property such as plaintiff's railroad is to value the whole system and then allocate a portion of that value to Oregon. As will be seen, this is a complex and sophisticated process.

The magnitude and complexity of the property involved moves the analysis to a level of sophistication not previously experienced. The court finds itself pondering evidence that deals with industry-wide comparisons and evaluations, analysis of security markets and investment alternatives, finance theory and principles of statistics. Some of the theoretical disputes appear to be unresolvable by any human at this time. The court has avoided these disputes where possible, particularly with regard to principles of economics or finance, and has tried to move in the direction of distilling and simplifying, not expanding and complicating.

Background

Plaintiff is a subsidiary of Union Pacific Corporation (UPC). UPC also owns three other subsidiary corporations, all major businesses in their own right. Those subsidiaries are: Champlin Petroleum Company, which engages in exploration, production, refining, marketing and transporting of petroleum products; Rocky Mountain Energy Company, which engages in mining, primarily soda ash, low-sulfur coal and uranium; and Upland Industries Corporation, a real estate management and development company dealing mostly in industrial property for UPC. Upland Industries administers 1 million acres of land, 7 million acres of mineral rights and 53 industrial parks in 13 states.

Prior to 1983, Union Pacific Railroad had 9,300 miles of rail spread over 13 states and used some 1,600 locomotives to pull 57,000 cars. On December 22, 1982, Missouri Pacific Railroad and Western Pacific Railroad were merged into Union Pacific Railroad, virtually doubling the size of plaintiff's system. After the merger, plaintiff owned 22,000 miles of track spread over 21 states; using 3,200 locomotives to pull 108,000 freight cars. This combination of three railroads constitutes the third largest rail system in the United States, connecting most of the major cities between the Mississippi River and the West Coast. The merger brought significant benefits to the system as a whole, including improving service from Mississippi to the West Coast and reducing expenses by merging marketing, sales and repair facilities. The system is large enough that, in trying to establish its value, one has a sense of valuing a piece of the nation.

1. The property which is taxable by the statutes:

"[I]ncludes all property, real and personal, tangible and intangible, used or held by a company as owner, occupant, lessee, or otherwise, for or in use in the performance or maintenance of a business or service or in a sale of any commodity, as set forth in ORS 308.515, whether or not such activity is pursuant to any franchise, * * * but does not include items of intangible property that represent claims on other property including money at interest, bonds, notes, claims, demands and all other evidences of indebtedness, secured or unsecured, including notes, bonds, or certificates secured by mortgages, and all shares of stock in corporations, joint stock companies or associations." ORS 308.510(1).

2. In addition to the general description, the statute expressly subjects the properties of railroads to taxation. ORS308.517 provides that leased or rented property is to be assessed to the user. ORS 308.555 provides that a total railroad system can be valued as a unit but:

"[T]he department shall make deductions of the property of the company situated outside the state, and not connected directly with the business thereof, as may be just, to the end that the fair proportion of the property of the company in this state may be ascertained."

In valuing plaintiff's railroad system for ad valorem taxation, delineating the taxable unit from other property is not an easy task. This may be better understood in the context of the two approaches applied by the appraisers in this case, the income approach and the stock and debt approach. In the income approach, the appraisers define the unit as measured by income. This requires the appraisers to separate and exclude income not attributable to operation of the railroad system. Fortunately, plaintiff maintains complete accounting records for ICC regulation and income taxation.

In the stock and debt approach, the challenge is to determine what proportion of the total corporate value is attributable to the railroad. This requires determining the value of the plaintiff as a subsidiary and then excluding from that value all nonrail assets. The disputes between the parties as to the "taxable unit" will be resolved in the discussion of the two approaches.

Income Approach

The label "income approach" is an inadequate expression of what and how future benefits from property are measured in terms of present value. Two methods or theories were applied by the appraisers. The yield capitalization or discounted cash flow (DCF) was used by the appraisers for both parties. This method discounts all of the expected future net cash flow to its present value. In applying this method, it is necessary to distinguish net cash flow from other measures of income such as accounting earnings (as might be reported for income taxes). A second method, applied by defendant's appraisers only, is the direct capitalization (DCAP) method. This method divides next year's expected net accounting earnings by a capitalization rate based on the relationship between price and earnings, commonly known as the P/E ratio.

The dispute between the parties in the income approach extends not only to the appropriate data for each method but also to the methods themselves. Significant time was spent exploring the theories and financial assumptions underlying the different models, each side attempting to invalidate the other's model and application. At the heart of this contest is the problem of identifying expected increases in future cash flow which increases present value.

When determining the taxable unit, one must look to the particular sources of income. For example, plaintiff's appraiser uses net railway operating income (NROI) as a surrogate form of net cash flow. Although he agrees that certain tangible railroad operating property produces Account 510 (miscellaneous rent) income, he believes it is not part of the integrated whole. Consequently, he did not include the Account 510 income in the measure of net cash flow. Rather, he considered the Oregon property producing miscellaneous rent (Account 510) income as "nonunitary" property. He estimated a value for this property using an income approach and then added that indicated value to the allocated Oregon value derived from the whole system. The court does not believe this approach is advisable. It is inconsistent with the concept of valuing an integrated whole. In this case, there is no evidence that property producing the Account 510 income is not a part of the integrated whole. Plaintiff's appraiser admitted Account 510 income is from "tangible railroad operating property," although asserting that it was not part of the integrated unit.

Defendant, in its brief, radically departs from its position at trial and contends that its unit is not complete. Defendant maintains that its appraisers failed to include income from certain intangible assets because they did not understand Oregon law and did not have adequate information to separate the unit from nonunit income in other accounts. Defendant asks the court to modify the unit and require plaintiff to provide information necessary to make an accurate determination. This request comes too late and defendant's proposal is too inconclusive.

3. The court acknowledges the difficulty of defining the unit by sources of income. The mass and complex nature of the income renders that process forever imprecise. However, the unit concept is accurately described by defendant's appraisers and they appear to have reasonably assumed that integrated assets which function together may be valued as a going concern. Defendant's appraisers did as good a job as could be expected in this situation. For example, they included net working capital but excluded money at interest, all consistent with ORS308.510. To the extent that there is "indivisible income" from other tangible assets, defendant's appraisers considered the amount of that income "very insignificant."

Growth Issue

In addition to the sources of income to be included in cash flow, the parties disagreed as to the amounts of expected income. Plaintiff's appraiser used a five-year average of NROI as his expected net cash flow. It should be noted that he adjusted NROI by deducting taxes, deferred income taxes and income from lease of road and equipment and by adding rent for leased property. He viewed this approach as "conservative" and thought it probably "overstates economic earnings." Defendant's appraisers estimated net cash flow by taking the estimated cash flow from plaintiff's strategic plans and then using a computer simulation to achieve a statistically probable range of cash flow. These positions raise a number of issues, the most significant of which is growth.

Plaintiff's appraiser anticipated no future growth which would affect the present value of the property. In his view, the shift from industrial to service industries, the development of interstate highways for trucks, and other changes caused railroads to lose market share and financial ground. Although he recognized that the Staggers Act deregulated the railroads to a limited extent, he did not feel that it produced the benefits anticipated. In analyzing the railroad industry, with its high labor costs, capital intensive needs and limited options, he sees it as a mature industry with no significant prospects for expansion. More importantly, he considers the railroad to earn an inadequate return on investment. Defendant's appraisers agree with many of plaintiff's observations with regard to the fact that it is a mature industry, capital intensive and subject to competition by trucks and airlines. However, defendant's appraisers see the Staggers Act as a positive benefit which gives the railroad the benefit of productivity gains.

Based on all of the evidence, the court finds that the market expected growth in railroads. The Staggers Act did have a positive effect on railroads and their values. In addition, the merger of the three railroads, with numerous benefits including gains in productivity, should translate into higher income and therefore higher value. Plaintiff's heavy reliance upon historical facts and trends cannot be extended into the 1980s in view of the Staggers Act. The whole point of the Staggers Act was to stop the downward spiral of the railroads.

The court finds that plaintiff has failed to demonstrate that the railroads are unique or different from the rest of the United States' economy. The court finds that they are closely related to GNP and are expected to grow accordingly. The court is not persuaded that railroads are inflation negative and consistently fall further behind. If the railroads were the dinosaurs portrayed by plaintiff, they would have ceased to exist a long time ago. Unlike many industries, railroads have not been subsidized by the government. Rather, they have been regulated, to their detriment. Knowing that the railroads were freed of that regulation to a great extent after 1980, the market expected substantial growth.

Not only the market, but the Interstate Commerce Commission projected positive growth rates. Plaintiff's own projections expected not just expansion growth but real growth.

"While cash flow from operations is expected to exhibit steady growth, this growth will be accomplished without high levels of capital expenditures."

Plaintiff's projections which assumed favorable economic conditions expected growth rates of 12.7 percent. Its projections which expected unfavorable economic conditions, such as high inflation, high interest rates and large federal budget deficits, showed growth rates of 11.9 percent. Likewise, the stock prices of UPC were rising and not only because of the other subsidiaries. Mr. Eric Olsen, one of plaintiff's witnesses, testified that the Shareholder Value Study showed the warranted market value was below the stock price. This only proves that the market expectations were different from the assumptions used in the Shareholder Value Study. Where the issue is market value, the market is always right.

Deferred Income Taxes

In considering the adjustments to net cash flow, the parties dispute the treatment of deferred income taxes. Accelerated depreciation provisions in the income tax laws result in increased depreciation deductions early in the useful life of property. While this results in less income tax initially, it can result in higher taxes near the end of the useful life of the property. The ICC and other government regulatory bodies require regulated utilities, such as plaintiff, to "normalize" their income in order to get a true picture of the income over the long run. Thus, the accounting statements make a provision for deferred taxes which are deducted from current income in order to arrive at a true picture of the cash flow.² For example, the 1982 Form R-1 reports to the ICC show combined federal taxes of negative $44,549,000 (tax credits).³ Likewise, the combined deferred income taxes of all three railroad companies of $107,169,000 is the amount to be deducted from net cash flow in addition to the normal federal taxes.⁴

Plaintiff eliminates deferred income taxes (DIT) from net cash flow because plaintiff views the DIT as being payable in the future. On the other hand, defendant includes DIT in net cash flow because it believes DIT are like an interest-free loan from the government that never has to be paid back. While both parties seem to have assumed a direct relationship between DIT and capital expenditures, it is not that clear to the court. For example, plaintiff's figures show capital expenditures more than doubling from 1982 to 1988 and yet the amount of annual DIT does not increase; rather, it goes down. Defendant used these diminishing annual DIT amounts in its cash flow projections, yet a diminishing amount tends to support plaintiff's position. On the other hand, the total accumulated DIT shown on page 000373, Exhibit Y-3, does grow by the annual amounts and more than doubles from 1982 through 1988. This ever-increasing total tends to support defendant's position.

The court finds that both parties failed to properly account for DIT. The provision for deferred taxes, as found in the Form R-1s, is further explained on Schedule 450, entitled "Analysis of Taxes." That schedule states that under Part B are "the particulars which most often cause a differential between taxable income and pretax accounting income." Examining Exhibit I-3, at 62 and 65, the Form R-1 for plaintiff for 1982, discloses that not all of the deferred taxes are attributable to accelerated depreciation. In fact, a number of items such as contested property taxes are not long-term deferral items. Likewise, the Form R-1s for Missouri Pacific show greater amounts due to special depreciation allowed by the Economic Recovery Tax Act of 1980 than due to accelerated depreciation. Exhibit K-3, at 62, shows that, of $100,565,000 of DIT, only $25,777,000 can be attributed to accelerated depreciation. Likewise, on Exhibit N-3, at 61, of $97,507,000 of DIT, only $18,167,000 is accelerated depreciation. It would appear that specific adjustments underlying the provision for deferred income taxes must be analyzed before adjusting the cash flow.

The Economic Recovery Tax Act (ERTA), Public Law 97-34 enacted August 13, 1981, § 203(c), contained a transitional provision which allowed the railroads to write off their "frozen base" over five years, using accelerated depreciation. It would be error for defendant to add back the full amount of that special ERTA depreciation since it will not last indefinitely into the future. Likewise, it would be error for plaintiff to subtract all ERTA deductions since they do not entirely represent accelerated depreciation. To the contrary, in many ways the ERTA deductions were catching up with straight-line depreciation.

The court has additional questions concerning the continuance or permanency of deferred income taxes attributable to accelerated depreciation. Plaintiff assumes that depreciation equals capital expenditures and, therefore, there is no benefit from DIT. However, plaintiff acknowledges that, due to inflation, capital expenditures will generally exceed depreciation. Plaintiff appears to ignore that difference on the theory that it is caused by property being acquired in the future and, therefore, is not now subject to tax. Defendant, on the other hand, assumes depreciation is less than capital expenditures and so adds back the DIT.

4. It is appropriate to recast income to accurately reflect cash flow. Burlington Northern v. Dept. of Rev., 291 Or. 729,635 P.2d 347 (1980). Unlike other utilities, plaintiff's rates are not regulated and the appraisers are not using the cost approach. See, for example, Pacific Power Light Co. v. Dept.of Rev., 7 OTR 203 (1977), modified 286 Or. 529, 596 P.2d 912 (1979). In this case, it appears to the court that it would be reasonable to adjust cash flow to include the deferred income taxes which appear to be permanently sheltered by the ever-increasing asset base. This would exclude deductions attributable to ERTA. See Burlington Northern R.R. v. Bain,648 F. Supp. 91 (S D Iowa 1986). To reiterate, what we are looking for here is expected cash flow. The court finds that capital expenditures are expected to exceed depreciation, which means that the DIT will be permanently deferred. As long as the parties are using perpetuity models to value an existing pool of assets, then deductions from net cash flow such as DIT should have the same long-term implications.

Financial Models

Much of the testimony and evidence in this case dealt with the debate over the appropriate financial model to use. It seems to the court that either model, as applied in this case, appears to offer as many complexities as the famous E = MC². The long form of the discounted cash flow formula is:

I₁ I₂ I_n V_o = ---------- + ---------- + . . . + ---------- (1 + i)¹ (1 + i)² (1 + i)ⁿ

This formula can be shortened in various ways, depending upon the assumptions made. In this instance, plaintiff's appraiser assumed (1) a constant level of cash flow, (2) that plaintiff does not earn a return on its investments greater than its cost of capital, (3) that inflation does not benefit plaintiff and (4) that the only growth plaintiff can achieve is by purchase. Based on these assumptions, plaintiff's appraiser condensed the formula to:

NROI V_o = ---- k

This is referred to by the parties as an "expansion" model because the only growth anticipated by it would be growth that would cost at least as much as the value it adds. Any such growth would have no effect on the present value of the company. In fact, plaintiff actually sees the railroad as shrinking in value because it earns less than its cost of capital.

The court finds plaintiff's assumptions too limiting. Aside from the difficulty of delineating growth due to new investments from growth due to replacement investments, the court is persuaded that there can be gains or real growth from replacements due to changes in technology. The railroad can and does earn a return greater than the cost of capital on some specific investments. Plaintiff contends that the test is the aggregate new investment, not specific investments, and that plaintiff does not actually earn its projected returns. Plaintiff assumes it will never earn the cost of capital in the aggregate. The court does not believe that plaintiff invests and reinvests billions of dollars to operate a railroad, which enjoys perhaps the best reputation in the United States, on the basis of self-deluding projections and analysis and earns less than its cost of capital. If this were true, one would expect to find an extremely poor market image with low security prices. Also, if that were the case, one would also have to question how plaintiff has continued to exist for the past 30 years. As will be seen below, the contrary is true. Plaintiff enjoys an excellent market reputation.

In defending its model, plaintiff maintains that it is the same as Copeland and Weston's "most complicated and realistic" model, only "boiled down." The Copeland and Weston model may assume depreciation equals replacement costs, but it does not assume away all potential for growth as plaintiff does.⁵

Plaintiff also maintains that it loses ground during periods of high inflation. This appears to be true for most businesses. The question is, after the recession, what were the market expectations? The market appeared to be optimistic long-term. Plaintiff is a large asset-based concern which at least is perceived (if not in fact) as an inflation play stock.

The model used by plaintiff is also characterized as a "static situation" model. The court finds that plaintiff's true economic circumstances are not static and the model is not appropriate. It is apparent that the model's form is like that of a direct capitalization (DCAP) model, which uses a single year's income to indicate a value. However, the difference between the underlying assumptions of the two models are significant. The DCAP model determines the relationship between earnings and price in the marketplace and uses that relationship to convert next year's expected earnings into an indication of value. That model captures all of the known and unknown expectations of the market with regard to growth. By contrast, plaintiff's model requires the appraiser to identify specific expectations of the market with regard to risk, rate of return of capital, appreciation or depreciation and inflation. To the extent that the appraiser misperceives any of these expectations, the model will not give a true indication of value. More importantly, if the assumptions on which the model rests, such as no real growth, are not valid, the model will undervalue the property.

The court's concern with defendant's discounted cash flow (DCF) approach is not with its model, but with its estimates and application. Defendant used a "PRISM" computer modeling program. In the briefest of terms, the computer simulates the possible combinations of revenues, expenses and discount rates hundreds of times to develop a most likely range of value. The random sampling of the computer simulations is supposed to indicate the most probable results of an uncertain future. However, plaintiff's statistical expert pointed out that the PRISM simulation can produce illogical results. That is, an increase in the revenue factor alone will result in alower indicated value. Defendant's rebuttal explanation leaves much to be desired.

"That could be the problem. I tend to think that it's also a part of the problem that when you change the bottom scale, so if I change revenues going from, let's say, thirty two hundred million to forty-nine hundred million. I just raise them a hundred, what happens is that somehow the computer samples from a different point in the distribution than it should and that causes a problem."

In addition to the witness' uncertainty as to what the problem was, the court is unpersuaded that the computer simulates reality. Defendant's best explanation is that the computer program is accurate if the fuel costs and base operating costs are changed in the same percentage as the revenues are changed. However, this is not what occurs in reality. Fuel prices obviously do not rise or fall in the same percentages as other operating expenses. If the validity of the computer simulation relies upon a fixed relationship, it has none. Computer-assisted analysis undoubtedly looms large in the future of appraising, but the appraiser must be prepared to explain each step of the analysis.

Cost of Capital

In applying the discounted cash flow method, the appraisers must estimate the market yield or discount rate. Although overall they are concerned with a weighted average cost of capital, the most difficult element of the discount rate to estimate is the equity yield rate. That is, the appraiser must estimate the rate which the equity investor would use to discount future cash flow. Several methods can be used. Plaintiff's appraiser rejected using earnings-price ratios from the market because they gave an indicated cost of equity capital less than the cost of debt. Plaintiff's appraiser then looked to other sources and concluded that capital costs had returned to the highs experienced in 1980. Based upon his analysis of general capital markets, he concluded that the appropriate equity rate was 18 percent for 1983 and 17 percent for 1984. Defendant contends that plaintiff's approach is wholly unacceptable and unpersuasive. Recognizing that one can find support for almost any position in the financial market, plaintiff's approach is not particularly persuasive.

Defendant's appraisers used two accepted models or methods for determining the equity cost of capital: (1) the Gordongrowth model, which focuses on dividends to be received, including expected growth in dividends in relation to the price paid for the stock, and (2) the capital asset pricing model (CAPM), which attempts to determine the risk-free rate of investment return and then add an amount for risk associated with that particular stock. Defendant's appraisers did not give the details of how they reached their conclusions of an equity cost of capital of 15.5 percent for 1983 and 15 percent for 1984. While it may not be necessary to give details until questions are raised, the court is at a disadvantage to evaluate the appraiser's work. For example, plaintiff points out that in defendant's dividend growth model, defendant used dividend yields and growth rates for both years very close to those adopted by the ICC yet came to results that were quite different from the ICC. Although the court is also without any details on the ICC calculations, it would appear that defendant's conclusions may be low. The same can be said for defendant's CAPM approach. Defendant argues for use of the geometric mean over the arithmetic mean. It is not clear which base or points defendant's appraiser used. It does appear to the court that, in view of plaintiff's excellent reputation, a lower risk would be justified than might be applied to other companies.

As mentioned above, the yield rate or capitalization rate to be applied must be a weighted cost reflecting both equity and debt elements. In order to arrive at that number, the appraisers must determine the proportion of debt and equity to be financed. Plaintiff used a target capital structure of roughly two-thirds equity and one-third debt based upon plaintiff's view of the industry. Defendant used the actual capital structure or debt-equity ratio based upon the market values. In using this approach, defendant appears to have used roughly the same debt-equity ratio as plaintiff except for the year 1983. In that year, defendant shows an equity of 58 percent and debt of 42 percent for the DCAP method. The disparity between the parties in their capital structure is not particularly significant. Most of the difference in their final estimate of the discount rate is due to their cost of equity capital.

5. In the court's view, defendant's weighted average cost of capital for the discounted cash flow method is the most accurate. The cost of capital is primarily a function of risk. Plaintiff is a high quality, strong railroad with a good reputation. This should be reflected in its cost of capital. Plaintiff rejects "blind reliance on market value relationships" and quotes its appraiser to the effect that the stronger railroads have higher capitalization rates. The court believes plaintiff errs in this regard. It may be error to blindly rely upon the market but it is greater error to ignore the market.

Direct Capitalization Approach

Defendant's appraisers also used the direct capitalization approach. By regression analysis,⁶ they determined that the best correlation of indicators was between operating income and aggregate capacity. Based on the last five year's average aggregate capacity, defendant's appraisers then projected the net income for one year into the future. The net (accounting) income used consists of NROI, after taxes but before interest expense, including Account 510 miscellaneous rent income, plus an addition to reflect the value of non-capitalized leased equipment.

Two primary points of concern with this method are the estimate of next year's earnings and the capitalization rate used to divide into those earnings. Again, here the dispute is not with defendant's model but with the application. The logic of the approach is that the market-established relationship between a company's earnings and its stock price reveals the investors' expectations with regard to equity investments. The theory is that the appraiser can establish a direct capitalization rate from all of the comparable companies and then use that rate to give an indication of market value of the subject property.

Plaintiff contends that the ten railroads used by defendant to establish its price-earnings (P/E) ratios were not comparables. Some were very small "pure" railroads while some were large, diversified operating companies. The court finds that the small railroads, namely Kansas City Southern, Florida East Coast and Denver Rio Grande, are not comparable. On the other hand, the larger railroads such as the CSX Norfolk Southern, Burlington Northern, Southern Pacific and Santa Fe are comparable. Plaintiff argues that the commodity mix and diversified nature of these latter railroads render them noncomparable. The court disagrees. The commodity mix does not appear to be significant. For example, coal does not appear to make a railroad less valuable since the Denver Rio Grande has a higher P/E ratio than Kansas City Southern. Plaintiff contends that the nonrailroad subsidiaries of diversified corporations tend to bring the P/E ratio up, thus, overvaluing the railroads. The figures in defendant's stock and debt approach tend to support this argument. However, the P/E ratios of primary or pure railroads are not always lower than the diversified companies. The court finds that sufficient comparability can be established to permit use of the DCAP approach.

Plaintiff attacks defendant's DCAP approach on several other grounds. Plaintiff maintains that defendant mismatches the capitalization rate derived from historical P/E ratios to expected earnings in the future. If this is true, defendant would be doubling up on the investors' expectations. Defendant explained that it obtained the capitalization rate from P/E ratios determined by Value Line, a security reporting service, which uses six-month actual and six-month estimate of earnings. Defendant considered these to be an equivalent of estimated earnings. The court is unable to tell whether they are appropriately matched or not. The Value Line report of February 19, 1982, shows the P/E ratio of 8.4 used by defendant but it also shows a trailing P/E of 9.4 and a median of 11.

The answer to this dilemma is not to throw out the P/E ratio but to be conservative in the estimate of earnings. Accordingly, plaintiff contends that defendant should use historic earnings, which would reduce defendant's indications of value by $564 million in 1983 and $741 million in 1984. There appears to be some inaccuracy here on defendant's part to the extent that it used estimated earnings to obtain a P/E ratio for the subject but its P/E ratio was based on actual earnings for the other subsidiaries. Also, as plaintiff points out, defendant did use historical earnings in its stock and debt approach. The court finds that defendant's indications of value by the DCAP approach should be adjusted down to reflect a more conservative estimate of earnings, although not necessarily in the amounts computed by plaintiff.

It is necessary to distinguish the railroad from the oil and gas industry. Plaintiff claims that the high P/E ratios of UPC stock is due to its oil and gas business. However, the court finds that this is not true. The market also anticipated growth in the railroads alone. Exhibit Q-6 shows that due to the Staggers Act the railroad stock prices were up over other industrials. Exhibit B-7 and C-7 show different patterns for oil and gas than for the railroads. Exhibit D-7 shows that the oil composite is weak compared to the railroads. Moreover, plaintiff's Champlin subsidiary is an integrated oil composite, which tends to have a lower P/E ratio because refining is not as profitable as production and exploration.

Plaintiff attacks defendant's direct capitalization rate on the grounds that it is not logical or realistic. Plaintiff argues that the inverse of a price-earnings (P/E) ratio, the earnings-price (E/P) ratio, should be equivalent to the equity cost of capital. Since theoretically the equity cost of capital can never be less than the risk-free rate of interest, defendant's direct capitalization rates are, in plaintiff's view, obviously wrong. The court rejects plaintiff's underlying assumption.

If a company with low earnings issues common stock to raise equity capital and the stock sells for a high price, it would indicate that the market has high expectations for the future of that company. The expected rate of return or capitalization rate derived from the relationship between the price paid for the stock and the next year's expected earnings would be "low," perhaps even lower than the risk-free rate of interest. But stock prices are not set on the basis of the current risk-free rate or on past earnings. Rather, stock prices reflect the investor's optimistic expectations for the future. If the market expects substantially increased earnings and they are later realized, the investor's rate of return will be much higher than indicated by the initial E/P ratio.

Plaintiff's calculated E/P ratio and identical "rate of return" erroneously assume that the actual earnings will be the same as the expected. The factual basis for plaintiff's attack on defendant's DCAP approach is based on plaintiff's estimating the cost of equity capital from general market indicators. "The E/P ratio for the railroad industry falls out of that determination." The court views this as backing into a determination, resulting in a conclusion which is not worthy of much weight. Plaintiff acknowledges that when the E/P ratio is less than the equity cost of capital, it indicates that investors expect true growth or beneficial inflation. There is no relationship between the P/E ratio (or its inverse E/P ratio) and the equity cost of capital unless there areno growth opportunities anticipated. (Brealey Myers,Principles of Corporate Finance 58 (2d ed 1984).) What it comes down to is that plaintiff assumes the railroad is without any growth opportunities. The court finds to the contrary.

As noted, plaintiff's appraiser also developed some E/P ratios. For "pure" railroads, or primarily railroad companies, his E/P ratio for 1983 of 18 percent would translate into a P/E ratio of 5.55. Likewise, plaintiff's E/P ratio of 17 percent for 1984 would translate into a P/E ratio of 5.88. These P/E ratios appear to be low when compared with the P/E ratios of the railroads which plaintiff shows as "primarily" railroads in Exhibit 71, at 144. As indicated on Exhibit 139, the Value Line reported P/E ratios for those railroads were: CSX 12.1; Kansas City Southern 15; Norfolk Southern 10.9; Rio Grande 18.2; and Soo Line 12.8.

Plaintiff concedes that the E/P ratios, for primary railroads fell faster than for diversified companies, but contends that this was because of their low earnings. However, plaintiff also acknowledges that the railroad industry rebounded in 1983. Exhibit X-2 indicates that the primary railroads had higher initial costs of equity capital in the early 1980s but fell below the diversified group in 1983. The point of this comparison is that if the railroads had such poor earnings with no real prospects of growth, investors would have dropped their bid prices and the stock prices would have fallen to a point where the anticipated earnings would constitute a market return on the investment. Thus, if a market return on the investment is 18 percent, as plaintiff maintains, and the stock dividend is only $5 per share, the price of the stock would have to drop to $28 per share to constitute a market return. The facts are that stock prices rose, which indicates that the market expected growth. Thus, the present value is greater than the past earnings would otherwise indicate. The direct capitalization approach provides a valuable reference point to the market.

6. There are additional questions with defendant's application of the DCAP method. For example, in evaluating the oil and gas industry, defendant combined exploration and production with refining, which could produce a lower P/E ratio because refining might drag down the price. Plaintiff argues that the combination would result in a higher P/E ratio because the absence of earnings in the refining would result in an increased P/E ratio. Plaintiff's contention assumes that the market will price stock based on some asset value unrelated to its earnings experience.⁷ It seems just as logical that the lack of refining earnings would also lower the overall price, so that the overall P/E ratio reflects the overall relationship between price and earnings. This is not to say, however, that defendant's indicators of value are to be taken at face value. The apparent lack of in-depth analysis by defendant's appraisers before the trial raises concerns about the reliability of their conclusions. Defendant's efforts to bolster its case by Dr. Pettit's constant growth valuation model is of little help. The court finds that his model, as applied, was not a good indicator of the theoretical P/E ratio.

Plaintiff's rejection of the direct capitalization model as meaningless is not acceptable. Plaintiff seems to maintain that the direct capitalization approach can be discarded because it is unrelated to other investment opportunities. Plaintiff's appraiser reasoned that the decline in the E/P ratio (an increase in the P/E ratio) may mean only a reduction in the cost of capital. Yet, it is clear that an increase in the P/E ratio or a decline in the E/P ratio may also indicate an increase in value.

Plaintiff also sees diversified companies as having less risk than the pure railroads. Yet the evidence indicated that the pure railroads have a beta, a measure of volatility, of one or less while the diversified railroads have betas in excess of one.

Dr. Pettit's failure to explain the market P/E ratio by using beta, growth and payout merely points up the vagaries and multifaceted nature of the valuation perspective. The court does not dispute plaintiff's view that there was an unprecedented demand for capital in the 1970s and that the "competition for funds will be intense" in the future. However, if that is so and railroads are such a poor risk, then the prices of railroad securities should drop in the market to reflect their risk and poor potential. Instead, prices rose considerably, indicating an optimistic market.

Plaintiff's appraiser pointed out a mathematical problem with defendant's use of the P/E ratios. He pointed out that the average of the P/E ratios will vary from the average of the E/P ratios. This results in an upward bias. For example, in 1983 defendant's E/P ratio was 12.5 percent whereas it should be 13.872 percent. Likewise, for 1984 defendant used 8 percent when it should be 8.561 percent. No one was able to explain the mathematical reasons for this upward bias; it can be avoided by converting the P/E ratios to E/P ratios before they are averaged.

To summarize the income approaches, plaintiff used a five-year average NROI capitalized by a discount rate. The court finds plaintiff's work in that regard not persuasive in light of the evidence of the growth potential. More specifically, plaintiff's capitalization rate was too high when applied in the very restrictive expansion model with its unrealistic assumptions. Likewise, the court finds defendant's discounted cash flow approach unreliable. Not only did defendant use the PRISM simulation, which the court finds unreliable, it used it only for the equity portion of the value. Defendant used the known value of the debt and added that to the value indicated for the equity portion in order to get an indication of value for the whole property. While there is logic to this approach, it is not a pure income approach but rather constitutes a variation or mixture of the stock and debt and income approach. It would seem preferable to obtain an indication of value wholly by the income approach and an indication of value wholly by the stock and debt approach and then correlate those two indications of value rather than mix the methods and obtain a diluted result.

Defendant's direct capitalization approach provides some indication of value based on the market relationship between earnings and price. The primary weakness or danger in this regard is the mismatching of P/E ratios with the expected earnings. If defendant had used historic earnings instead of estimated future earnings, this would have lowered the indication of value. Defendant did use historic earnings in its stock and debt approach. Defendant vigorously maintains that it appropriately matched earnings in its direct capitalization method. The solution is not to throw the method out but to consider both measures when correlating the various approaches.

Stock and Debt Approach

The theory behind the stock and debt approach is that the market value of a company's securities represents the market's opinion of the value of the company's underlying assets. While appraisers recognize that there are many limitations and unanswered questions with this theory, it is an accepted method to indicate value. In order to apply the theory, the appraiser must ascertain the total market value of all publicly traded common stock, preferred stock and bonds. In this case, the debt securities are the least problematic.

7. As defendant points out, the parties are close in their debt values. For 1984 defendant calculated $1.449 billion and plaintiff calculated $1.448 billion. In 1983, defendant calculated $1.634 billion and plaintiff calculated $1.466 billion. The court finds that defendant made a number of small errors in its debt valuation. Plaintiff's witness pointed out, for example, that defendant double counted Chicago Heights Terminal Railroad Company debt of approximately $20 million, miscellaneous accounts payable of $11,700,000 and may have included nonrail of Western Pacific. If so, defendant double counted or overestimated debt value by $41 million. However, since there is only $1 million difference between plaintiff and defendant's figures for 1984, the court will accept plaintiff's figure for that year. In 1983, the difference between the parties is much greater, but most of that difference is attributable to the defendant's use of year-end values. The court believes that the values closest to the assessment date should be used and, therefore, the court accepts defendant's position for 1983. However, assuming that plaintiff's criticisms with regard to double counting are correct, the court will adjust the cost of the 1983 debt down to $1.6 billion.

The primary problem in valuing the stock of plaintiff is that it is not publicly traded. To overcome this problem, the appraisers must find the total value of all UPC stock and then allocate a portion of that equity value between plaintiff and the other subsidiaries. The court finds problems with both parties' determinations of equity value for the stock and debt approach. Plaintiff's appraiser divided the railroad net income by an equity capitalization rate in order to obtain an indication of equity value. That value was approximately 28 percent of the total known UPC equity (stock) value. He then applied a rounded percentage (30 percent) to find an indicator of value for the railroad (equity only) of $1.3 billion. Plaintiff admits that if he stopped there, there would be no real difference from that determination of value and his determination of value by the discounted cash flow method. Plaintiff's appraiser went on however, to "verify" the indication of value by using P/E ratios to find the value of nonrail segments. He then compared the total of the nonrail segments to the UPC value remaining after deducting the railroad $1.3 billion. Plaintiff claims that this method demonstrates that it has not overvalued UPC's nonrail subsidiaries.

Plaintiff's work is not persuasive. In essence, plaintiff's appraiser has backed into the basic number using the identical calculations he used to find the discounted cash flow indication of value. This dilutes the value of the stock and debt approach. Further, plaintiff's use of the P/E ratios raises many questions. Plaintiff values separately the two parts of Champlin's operations. Valuing the parts separately is not consistent with reality where the business appears to be an integrated operating entity. In establishing a value for Champlin for 1984, plaintiff maintains that low earnings for Champlin resulted in a higher P/E ratio. Accordingly, plaintiff's appraiser increased his estimate of the P/E ratio for Champlin's exploration and production operation and added two-thirds of the book value of the refinery. The problem with plaintiff's approach is that this is based more on reasoning than on market evidence. There is no evidence that the market reacts that way. In the court's view, plaintiff has ended up using a higher P/E ratio than the market would find acceptable. Plaintiff presented no evidence indicating that the market data supports a nonrail P/E ratio three times greater than the railroad's P/E ratio.

Plaintiff's testing of the values by a "fallout" P/E ratio is not persuasive. As defendant points out, if the railroad equals 30 percent of the total UPC equity, that would be equal to $2 billion plus. When added to $1.4 billion to $1.6 billion of debt, the total value of the railroad should be in the range of $3.5 to $4 billion. Plaintiff's opinion of value is closer to $2 billion, which would mean that plaintiff must be attributing less than $1 billion to the debt.

Defendant's approach to valuing plaintiff's stock is also subject to some basic criticisms. Defendant used the direct capitalization approach to find equity value, which is the same as its direct capitalization in the income approach. While it used other multiples, the court is not persuaded that they give a separate value. In fact, defendant's Case II, which provides slight modifications, makes it worse. That case provides for total parts which exceed the known total by 25-30 percent in 1983. Other errors in use of the income-influence and asset-influence methods, which separate out the nonrail assets, undermine the reliability of the conclusions.

One clear theoretical issue in the stock and debt approach was the treatment of preferred stock by the parties. The parent corporation, UPC, issued a significant amount of preferred stock as part of the purchase price of the Missouri Pacific Railroad. Defendant treats that preferred stock solely as representative of railroad assets. Plaintiff maintains that defendant's position is in error because all of the assets of UPC support the preferred stock, not just the railroad.

The court finds that it is appropriate in this case to attribute all of the preferred stock to the railroad. Except for the presence of the other subsidiaries, all of UPC's stock would be attributed to plaintiff. Where a block of securities can be clearly attributed to one company, there is no need to allocate it among all the subsidiary companies. In the stock and debt approach, it is not a question of where the income comes from to pay the preferred stock dividends. The attempt is to measure the value of the assets by segregating out the value of the securities. Since the preferred stock was issued solely to obtain railroad assets⁸ there is no need to try and allocate the preferred stock to other than the railroad. In using this approach, it is necessary however to reduce the income which defendant considered in determining the value of the equity portion of the stock and debt approach. Defendant concedes this point.

Overall Evaluation

If this were a contest between experts, plaintiff would win. The work of defendant's experts was pockmarked by numerous errors and weakened by lack of analysis. This may be attributable in part to the fact that two experts were attempting to cooperate to produce one appraisal. Such an endeavor requires great coordination of effort and thought. In contrast, plaintiff's single appraiser was perhaps the most persuasive and thorough expert to ever appear before this court in such a complicated field. He showed himself to be a master teacher and extremely well prepared. It is difficult to find any faults within the parameters of his presentation.

However, this is not a contest of experts alone but a process to determine the value of property. It is in this larger perspective that the court finds itself unable to accept plaintiff's position. Plaintiff's total value for the railroad system as of January 1, 1983, is $2,351,249,000. Yet, in the merger which took place nine days before that assessment date UPC paid approximately $1 billion in order to obtain the assets of Missouri Pacific Railroad Company, which were encumbered by in excess of $1 billion of debt. This would indicate that the underlying assets exceeded $2 billion in value. Even after deducting the nonrail assets from that total, when the Union Pacific Railroad (pre-merger) is added, the value for the total merged system should exceed $3 billion.

Although the parties did not do a cost approach, management's investments and reinvestment are some indication of management's judgment with regard to value.

"For UPRR, the economic value of the fixed asset base, restated to a 1983 cost level, maintained at about $8 billion between 1973 and 1980 and then declined slightly in 1981 and 1982 ($7.8 billion and $7.5 billion, respectively). On the same basis, MPRR's fixed asset base increased slightly between 1978 ($5.5 billion) and 1982 ($5.8 billion).

"Given the relative stability of UPRR's and MPRR's net asset bases over the study periods, incremental capital investments for roadway and equipment assets served to maintain the asset base by replacing, in total, capital consumption. For example, between 1973 and 1982, UPRR made incremental capital investments totaling $4.3 billion (stated in 1983 dollars), and MPRR made capital investments totaling $1.9 billion from 1978 to 1982. However, during the same periods, the value of the net asset bases (in 1983 dollars) changed very little, as indicated above."

It is difficult to believe that a system which had $6.2 billion of capital investments pumped into it over a 10-year period and a total replacement cost of twice that amount would have a true cash value of only $2.3 billion.

Plaintiff sees no real growth in the future despite the admitted benefits of the Staggers Act. In the court's view, the Staggers Act constituted a major departure from past history. The market expectations of growth were reflected in higher stock prices. Plaintiff's management was also optimistic, particularly in light of the merger and the benefits anticipated therefrom. The company's strategic plan, from which defendant derived its estimates of future cash flow, were optimistic and reflect anticipated growth. That optimism appears to have been based on hard reality, recognizing the pressure on prices due to the deregulation and competition.

"Despite the expected rate competition, the Railroad Group's ratio of operating expenses to revenue is projected to show steady improvement over the plan period * * *."

Finally, plaintiff attributes too much value to the other subsidiaries of the parent, UPC. The parent itself has indicated that: "Since 1978, however, the Corporation has experienced a 7% real annual compound growth rate in revenues, after eliminating the effects of inflation, despite declines in the current year." It appears from the same report that transportation (only a small part of which is nonrailroad) has consistently produced approximately 40 percent of the net income of the parent corporation from the years 1978 through 1982. During the same period of time, the percentage of total capital expenditures attributable to transportation diminished from 56 percent in 1978 to 12.3 percent in 1982 while capital expenditures for oil and gas increased from 42 percent in 1978 to 85.5 percent in 1982. Knowing that all industries go through economic cycles, it seems unreasonable to the court to view the railroad as worth substantially less than its share of the parent's corporate income. If the railroad consistently produces approximately a large percentage of the parent's income, it seems reasonable to require plaintiff to demonstrate why it isn't worth approximately the same proportion of the parent's total value.

Based upon the above, the court finds that plaintiff has failed to carry its burden of proof. However, the court has also found errors, with plaintiff's assistance, in defendant's estimate of values. This requires adjusting defendant's estimated market value of the total Union Pacific Railroad system. The court finds that the true cash value for the railroad system as of January 1, 1983, is $3,100,000,000 and, as of January 1, 1984, $3,600,000,000.

A proportion of the total system value must be allocated to Oregon for taxation. Although the differences between the allocation factors used by the parties are not great, there are differences. After reviewing defendant's calculation of its allocation factors and plaintiff's calculation of allocation factors, the court finds that defendant's allocation factor should be used. Therefore, the court finds that the total system value and the taxable value for Oregon for the years in question are as follows:

TOTAL SYSTEM ALLOCATION ALLOCATED YEAR VALUE FACTOR TO OREGON 1983 $3,100,000,000 4.06% $125,860,000 1984 $3,600,000,000 4.76% $171,360,000

Judgment will be entered consistent with this opinion.

Costs to neither party.

¹ To reflect the integrated nature of the property and for ease of reading, the separate corporate status of the other plaintiffs, subsidiaries of Union Pacific Railroad, are ignored.

² "Normalized" income could also be accomplished by deducting only straight-line depreciation rather than accelerated depreciation.

³ In looking at plaintiff's figures on page 105 of Exhibit 71, it appears that plaintiff did not include the figures for Western Pacific with those of Union Pacific and Missouri Pacific.

⁴ Again, plaintiff's figure of $110,078,000 on page 105 of Exhibit 71 appears to omit Western Pacific's information.

⁵ It might be noted that Copeland and Weston also add deferred income taxes back to measure net cash flows.

⁶ The court has found it unnecessary to attempt to resolve the dispute between the parties as to correct statistical theories and methods. Even so, it seems important to recognize the great caution with which statistics are to be used. As one writer has so aptly voiced the concern: "The organization's figures merely prove that if you torture statistics long enough, they will confess to anything." Bandow, Federal Appliance Standards, Wall St J, Feb 19, 1987, at 26.

⁷ This is exactly the opposite position plaintiff takes with regard to the railroads. There plaintiff maintains low earnings result in low values.

⁸ The court recognizes that adjustments need to be made for the nonrail assets owned by the railroad such as the Mississippi River Transmission Company.