Washington Mutual, Inc. v. United States, 2014 Del. LEXIS 115 (March 14, 2014)

In a case arising out of the savings and loan debacle, the successor-in-interest to a “healthy” thrift that agreed to take over a number of failing ones sued the government for a tax refund based on tax deductions and losses for certain intangible assets, specifically branching and RAP rights. The case hinged on whether the plaintiff could establish a cost basis for those rights. The plaintiff turned to a leading valuation expert who used an income approach to determine the fair market value of a specific group of branching rights. The government’s expert attacked the valuation model’s key assumptions and also tried to undercut its validity with a market analysis.

Another outgrowth of the S&L crisis. This case, said the court in its introduction, was yet another chapter in the “continuing saga that has its genesis in the savings and loan crisis of the late 1970’s to early 1980’s.” In late 1981, a “healthy” Southern California thrift agreed to a supervisory merger with three failing thrifts: Two of them were in Missouri and one in Florida. Under the transaction terms, the acquirer assumed the liabilities of the failing institutions in return for a set of incentives from the government, including cash contributions, indemnities related to covered assets, and other payments that went through a special reserve account. Moreover, it received branching rights in Missouri and Florida. And it received the RAP right, meaning it was permitted to count the excess of the failing thrifts’ liabilities over their assets as “supervisory goodwill” and apply it to its minimum regulatory capital requirement. Finally, it was able to structure the transaction as a tax-free “G” reorganization such that it could carry over the tax basis of the acquired assets and realize substantial built-in losses by selling the loans of the acquired thrifts.

In 1998, Washington Mutual (WaMu) acquired the healthy thrift and subsequently filed an amended income tax return claiming refunds for 1990, 1992, and 1993 related to the thrift’s branching and RAP rights. After the Internal Revenue Service denied the requests, WaMu sued in federal court (W.D. Wash.) There, the government argued that the claim failed as a matter of law because WaMu was unable to show that the thrift had a tax basis in the rights. The court agreed, which triggered the plaintiff’s appeal with the 9th Circuit. The Court of Appeals “return[ed] to the very basics of tax law” and began by defining the term “basis,” which was “a taxpayer’s capital stake in an asset for tax purposes.” Here, the branching and RAP rights were part of the consideration the thrift received in exchange for taking over the three failing thrifts. The cost to the acquiring thrift “was the excess of the three failing thrifts’ liabilities over the value of their assets.” The 9th Circuit found the thrift’s cost basis in the branching and RAP rights was “equal to some part of the total amount of that excess liability.” (emphasis in original)

After the Court of Appeals found the plaintiff, in theory, had a cost basis, it sent the case back to the district court where the plaintiff had to show with reasonable certainty what the cost basis was before qualifying for any refund. To determine how much the thrift paid for the incentive package, the plaintiff had to ascertain: (1) the total value of the three failing thrifts’ liabilities; and (2) the total value of their assets. As the district court put it: “While one may expect that this calculation would be straightforward (at least relatively speaking), the parties dispute each variable in the calculation.”

The plaintiff claimed that the purchase price did not represent the fair market value (FMV) of the incentive package the thrift received as part of the supervisory merger. This claim, the district court emphasized, meant the plaintiff conceded that the value of the incentive package was more than its purchase price. It explained that, under case law, where one lump sum buys a conglomeration of assets, the plaintiff had to determine the cost of each asset “by apportioning the purchase price among the assets according to each asset’s relative fair market value at the time of the acquisition.” If the plaintiff was unable to establish the FMV of one of the assets in the package, it would be impossible to calculate the package’s total FMV and by extension to determine the pro rate share of each asset, the court went on to say.

At trial the focus was on valuing the Missouri branching right. The plaintiff presented expert testimony as to its fair market value to a hypothetical buyer; the government offered an expert to critique the valuation and present a counteranalysis.

Five-step analysis to determine excess earnings. The plaintiff’s valuation expert used an income approach and specifically “a discounted cash flow model known as the ‘excess earnings’ approach.” He determined the excess earnings or cash flows a hypothetical buyer would expect from the branching right beginning in late 1981, excluding charges for the use of contributory assets, and discounted to present value by performing a five-step analysis. The steps include:

Step 1. Determine the projected net operating income for the Missouri branching right by projecting the Missouri deposit market growth rate and calculating the hypothetical buyer’s expected share of that market. The expert projected Missouri’s new deposit market would increase 47% in 1982, 39% in 1983, 32.3% in 1984, 26.5% in 1985, and 21.85% in 1986. He suggested the increase was due to inflation, which meant there was no actual growth in real-dollar terms. He described the hypothetical buyer as a “well-capitalized thrift” that would take control of part of the market by advertising and offering new products and a sense of “safety.”

Because thrifts at the relevant time were not allowed to operate branches outside their home market, the expert drew on the acquiring thrift’s actual expansion into Northern California in the 1970s to predict what the hypothetical buyer could expect to accomplish. With its headquarters in Southern California, the thrift originally was limited to open branches within 100 miles of there. The expert believed Northern California was a new market in the way Missouri would be a new market for the hypothetical buyer. He concluded that the buyer would capture a market share equal to the institution’s capture of 7% of the Northern California market after 10 years.

He also assumed that the buyer would use all the new deposits to originate new adjustable rate loan mortgages (ARMS). He assumed interest rates would remain flat but did not believe this would prevent the hypothetical buyer from generating new loans because ARMS would be attractive to consumers. He acknowledged that the rate could “impact” the “deposit market” and “tone down the amount that was available for new loans.”

To project the “spread” the buyer would make on the new mortgage loans, he calculated the difference between the interest charged on the mortgage loans and the cost to the thrift for obtaining and processing the deposits. He assumed the buyer would obtain deposits from Missouri (it’s less expensive than getting them from California) and make ARMS in California to achieve “a little bit more than the contract spread.” From the cost of funds in San Francisco and Des Moines, Iowa, in 1981, he assumed a spread of between 2.25% and 2.75%. He adopted a midrange of 2.5% and arrived at net revenues of nearly $2.8 million.

Step 2. Determine the projected net income of the right. He determined that the tax rate applicable to the incremental income would be 36% and subtracted the corresponding amount from the operating income.

Step 3. Determine the cash flow attributable to the branching right. To isolate the cash flow, he separated income that was attributable to assets from the acquired thrifts. For this purpose, he subtracted capital charges from the projected income.

Step 4. Determine the value of the branching right prior to adjustments. Acknowledging that developing projections was a “risky business,” the plaintiff’s expert used a 22% discount rate to compute the present value of the right. He then calculated the residual period cash flow and converted it to present value, concluding it was $37 million before adjustments. He considered this figure conservative.

Step 5. Determine the FMV of the branching right. Recognizing that a hypothetical buyer had to make two “upfront investments” to capitalize on the Missouri branching right, the expert calculated $1.2 million in transaction costs to take over existing branches and acquire new branches in the state. The costs went toward advertising, signage changes, and training the workforce. There also was “assemblage value” related to the buyer’s ability to buy the acquired thrifts’ existing branches and not to have to find branch locations and assemble and build the infrastructure necessary to operate the business. This value was $5 million, he concluded. After subtracting the transaction costs and assemblage value, he arrived at an FMV for the branching right of $28.8 million.

Government attacks four ‘value drivers.’ The government contended the valuation model was unreliable. Broadly speaking, it failed to account for the gloomy economic situation that existed in late 1981. In particular, there were four “value drivers,” or key assumptions, underlying the expert’s cash flow projections that did not square with reality and together inflated the value of the right. Moreover, the government expert’s market analysis brought into relief the model’s lack of reliability.

The court considered the arguments and counterarguments in turn.

1. Unrealistic deposit market growth rate. The government’s expert claimed that the opposing expert’s projected high rate of deposit growth did not align with the dire situation in which the thrift industry found itself in 1981. As a result of high interest rates and regulations as to the rates thrifts could pay depositors, deposits actually were diverted from the thrift industry—a development known as “disintermediation.” The acquiring thrift in fact lost about $720 million in deposits in 1981. The situation only improved due to Congressional action in late 1982.

The plaintiff argued that a product called “All Savers Certificate” justified the expert’s “robust” deposit projections. Also, its expert’s model actually indicated nominal deposit growth when adjusted for inflation. Finally, the hypothetical buyer would be able to attract depositors from other thrifts because, unlike them, it suggested “safety.”

The court found the growth rate problematic. It credited the testimony of the government’s expert about disintermediation and noted that the allure of the “All Savers Certificate” was nonexistent. The product, it said, “was quickly labeled a dud.” Further, the plaintiff’s inflation adjustment argument was only true for the model’s statewide deposit projections. Multiplying the projections by the hypothetical buyer’s projected share of the deposit market showed the model “more than compensates for inflation.” And it was likely that depositors willing to move funds would have moved them to banks that not only were just as “safe,” but also were able to offer a product with a higher rate of return—something the hypothetical buyer was not able to do until late 1982.

2. Unrealistic 7% market share. The government pointed out that the rivaling expert’s growth rate analysis was based on the acquiring thrift’s expansion into Northern California, which took place before disintermediation, in the 1970s. Then the thrift industry operated in a much more favorable environment.

Also, in terms of the thrift’s Northern California expansion, much of the growth was the result of its buying existing institutions. In contrast, the plaintiff expert’s model relied on an “organic expansion,” assuming the hypothetical buyer would grab market share solely by opening new branches.

The plaintiff tried to downplay this conceptual difference by claiming it did not matter in terms of outcome whether the buyer opened new branches or bought up existing branches. Either way, the buyer would have had the same market share at the 10-year mark.

The government also introduced a letter from the acquiring thrift’s then CFO to the Securities and Exchange Commission (SEC) written three days before the transaction and projecting the thrift would have $638 million in new Missouri deposit funds by the end of 1986—whereas the plaintiff’s expert forecast the hypothetical buyer would have $1 billion by then.

The court agreed that the projected 7% rate was unrealistic. The Northern California expansion was “a wholly unreliable indicator” of the hypothetical buyer’s ability to capture a share of the Missouri deposit market in 1981, it said. The plaintiff’s expert also “did not properly account for the expense” of the rapid growth he projected. Finally, the court pointed out that “lowering [the model’s] market share rate in Missouri to a level that aligns with [the CFO’s] projections for 1986 reduces the model’s computation of the Missouri branching right fair market value by almost fifty percent.”

3. Unrealistic quantity of new loans. The government claimed the record contradicted one of the most critical assumptions underpinning the model: that the hypothetical buyer would be able to use the new deposits to originate ARMS, particularly in California where, according to the plaintiff’s expert, demand for loans remained high. To do so was important because a high influx of new deposits without the ability to generate new loans would have a negative effect on the thrift’s earnings as it would have to shoulder the cost of the interest on the deposits without receiving mortgage payments to offset it.

As the government told it, in 1981 the thrift industry was unable to originate mortgage loans to any measurable degree. The thrift’s own statistics indicated that it originated a mere $342.6 million in new mortgages in 1981, compared to $2.3 billion in 1976. This steep decline, said the government, was a direct result of “sky-high” interest rates and the inability of homebuyers to finance their purchases. Also, the thrift’s CFO testified that the thrift industry was loan-driven: If demand for loans was weak, a high influx of deposits would bring down earnings.

The plaintiff justified the assumption by noting that thrifts “historically had been able to use all of their deposits to make mortgages.”

The court found the plaintiff’s argument unconvincing. Relying on the institution’s prior ability to originate loans ignored the reality at the time of the transaction. Also, given the industry’s loan-driven approach, it was not surprising that in the past the thrift managed to originate loans from all of its deposits. According to the court, “this simply meant [it] was good at judging its need for deposits.” The court emphasized that the plaintiff expert’s model was deposit-driven: The hypothetical buyer would try to capture as many deposits as possible and then try to originate mortgages from those funds. Based on this approach, the hypothetical buyer could well generate more deposit funds than it could turn into loans, the court said.

4. Unrealistic 2.5% net interest spread. The government claimed that, since the plaintiff expert’s model was based on new loans with new deposits, the hypothetical buyer would have to pay a going rate for a new market rate CD instead of the lower costs indicated in the 11th District Cost of Funds Index, which informed the expert’s calculation.

The plaintiff countered that the buyer would have been able to bring in new deposits simply by “marketing and [its] secure reputation” in line with the thrift’s earlier experience in Northern California.

The court dismissed the argument, repeating that the Northern California experience was not a reliable indicator of what a buyer could accomplish in the 1980s.

Government expert rebuts with market analysis. Besides critiquing the plaintiff expert’s model, the government’s expert performed a market analysis of some 200 in-state thrift mergers that took place around the date of the transaction. The impetus was the plaintiff’s claim that the acquiring bank paid a premium because it wanted the branching rights. By the government expert’s reasoning, if that were the case, then the in-state mergers should not show a similar premium because they did not include branching rights. But since these mergers did in fact indicate that the acquiring thrifts paid a similar premium, it was for some other intangible asset. Therefore, the opposing expert erred when he attributed a significant percentage of the premium for the transaction to the Missouri branching right, the government’s expert concluded.

The plaintiff disparaged the market analysis as “worthless.” It included thrifts that were “qualitatively different” from the failing thrifts that the acquiring institution took over. The latter involved assistance from the Federal Savings and Loan Insurance Corp. (FSLIC), whereas the market analysis also considered unassisted mergers.

However, the plaintiff’s expert examined 55 in-state mergers that were part of the market analysis to determine “why the acquiring thrifts in those cases might have been willing to pay a [premium].” He concluded they paid for “synergy,” which would be “some trade name value or market positioning.” The government’s expert dismissed that conclusion, noting that his own study involved small and large thrifts paying similar premiums. Synergy would result in far less gain for the larger thrifts than the small ones, but they still paid a premium on par with the one in the transaction at issue.

The court found the plaintiff expert’s explanation unsatisfactory and said the market analysis “significantly undercuts the trustworthiness” of his model.

Because the plaintiff had failed to establish the FMV of the Missouri branching right, the court found itself “unable to ascertain the cost basis for either the branching rights or the RAP right, or, for that matter, to the incentive package.” Consequently, it dismissed the plaintiff’s refund claim.