OPINION

BRUINIERS, J. —

Under California's version of the Uniform Division of Income for Tax Purposes Act (UDITPA; Rev. & Tax. Code, § 25120 et seq.),1 the portion of a national or multinational company's income that is subject to taxation in this state is determined by a formula that compares the company's payroll, property and sales in this state to its total payroll, property and sales. When the standard formula does not fairly represent the extent of the company's business in California, however, UDITPA also provides for application of a reasonable alternate formula in order to achieve an equitable result. (§ 25137; Microsoft Corp. v. Franchise Tax Bd. (2006) 39 Cal.4th 750, 757 [47 Cal.Rptr.3d 216, 139 P.3d 1169] (Microsoft).)

General Mills, Inc., and its subsidiary corporations (hereafter, General Mills) is a unitary group of corporations operating both within and outside California. It is a consumer foods company with its principal place of business in Minneapolis, Minnesota. In opposing this tax refund action, the Franchise Tax Board seeks to apply an alternative formula to income resulting from trading by General Mills in agricultural commodity futures. General Mills engages in such trades as a hedging strategy to protect against price fluctuations in the basic materials it needs for its business, the manufacture and sale of consumer food products, as well as flour and grain. In a prior appeal (General Mills v. Franchise Tax Bd. (2009) 172 Cal.App.4th 1535, 1548 [92 Cal.Rptr.3d 208] (General Mills I), we held that the proceeds from this activity were properly included as "gross receipts" under section 25120, subdivision (e) in the standard UDITPA sales apportionment factor. Since the trading activity did not occur in California, the inclusion would result in a reduction in California tax liability.

Because the trial court had not reached the issue of whether the UDITPA apportionment formula, including the trading proceeds, "does not then `fairly represent' General Mills's business activity within California, thus warranting imposition of an alternative formula pursuant to section 25137," we remanded for the trial court to decide that issue. (General Mills I, supra, 172 Cal.App.4th at p. 1548.) On remand, the trial court took additional evidence, considered further argument, and ruled that including overall gross receipts from futures trading in the standard UDITPA formula did not fairly represent the extent of General Mills's business activity in California. It allowed the Franchise Tax Board to impose an alternate formula that included only the net gains generated by General Mills from futures sales. We affirm.

We conclude that General Mills's hedging activity — while integral to General Mills's main consumer food business is both qualitatively different from General Mills's other sales that are made for profit and substantially distorts the percentage of General Mills's income that is apportioned to California. The Franchise Tax Board's alternate formula, including only the net gains from General Mills's futures sales, is reasonable and may be imposed consistent with UDITPA.

I. BACKGROUND

As our Supreme Court has observed, "... UDITPA's application is not always clear." (Microsoft, supra, 39 Cal.4th at p. 755, fn. omitted.) In General Mills I, we considered "whether commodity futures sales that are made to hedge against price fluctuations should be included in the sales factor of the [UDITPA]." (General Mills I, supra, 172 Cal.App.4th at p. 1537.) We described the background of this litigation as follows:

A. Our Holding in General Mills I

In General Mills I, we explained that the "... UDITPA defines `"sales"' as `all gross receipts of the taxpayer not allocated [as nonbusiness income].' (§ 25120, subd. (e).) A regulation interpreting section 25120, subdivision (e) reads, `[T]he term "sales" means all gross receipts derived by the taxpayer from transactions and activity in the regular course of such trade or business.' (Cal. Code Regs., tit. 18, § 25134, subd. (a)(1).)." (General Mills I, supra, 172 Cal.App.4th at p. 1543, fn. omitted.) We concluded that "the full sales price (number of bushels times price per bushel) of all of General Mills's futures sales contracts should be counted as gross receipts in the UDITPA sales factor." (Id. at p. 1544.) We based this conclusion on the fact that "a futures sales contract is a legally binding obligation to deliver a specified amount of a specified commodity at a specified price in a specified month." (Ibid.) Even when the contract is offset rather than consummated with delivery and payment, the offsetting party receives consideration for the contract in the form of relief from the obligation to consummate the sale. (Id. at pp. 1545-1546.) We rejected the Franchise Tax Board's arguments that such offset transactions were illusory and had no real financial value. (Ibid.)

We further explained that our conclusion that the full value of futures sales were gross receipts within the meaning of the sales factor was consistent with the purpose of the UDITPA sales factor: "The UDITPA sales factor is designed to reflect the taxpayer's `income-producing activity,' which regulations define as `transactions and activity directly engaged in by the taxpayer in the regular course of its trade or business for the ultimate purpose of obtaining gains or profit.' (Cal. Code Regs., tit. 18, § 25136, subd. (b);[7] see Rev. & Tax. Code, §§ 25134-25136.) General Mills's futures sales satisfy this definition. Its employees engage in futures trading (instructing brokers to enter into specified contracts) on a daily basis to hedge its actual or planned purchases of raw commodities that it will either resell for profit or process into flour or consumer food products to be sold for profit. Hedging allows General Mills to stay in business and make a profit despite frequent and significant fluctuations in the prices of the raw commodities. That is, hedging futures sales are made `for the ultimate purpose of obtaining gains or profit,' even though General Mills does not seek to make a profit on the futures trades alone. As the United States Supreme Court has written, `it is difficult to imagine a program more closely geared to a company's manufacturing enterprise or more important to its successful operation' than hedging in raw commodities on the futures market. (Corn Products Co. v. Commissioner (1955) 350 U.S. 46, 50 [100 L.Ed. 29, 76 S.Ct. 20].)" (General Mills I, supra, 172 Cal.App.4th at p. 1547.)

In sum, we held that General Mills's futures sales were gross receipts within the meaning of the UDITPA sales factor. As noted ante, we remanded for the trial court to decide whether imposition of an alternative formula pursuant to section 25137 was warranted because application of the standard apportionment formula would not then "`fairly represent'" General Mills's business activity within California. (General Mills I, supra, 172 Cal.App.4th at p. 1548.) We wrote, "In Microsoft, the Supreme Court held that an alternative formula could be imposed under section 25137 if the challenged activity both qualitatively differs from the taxpayer's principal business and quantitatively distorts the formula by a substantial amount (in that case `cutting Microsoft's California income tax nearly in half'). (Microsoft, supra, 39 Cal.4th at pp. 757, 766; Limited Stores, Inc. v. Franchise Tax Bd. (2007) 152 Cal.App.4th 1491, 1497 [62 Cal.Rptr.3d 191].)" (General Mills I, at p. 1548.)

B. Trial Court Proceedings on Remand

On remand, the trial court took additional evidence, including further expert testimony, and required additional briefing by the parties. On November 1, 2010, it issued a 68-page statement of decision (Statement of Decision) that upheld the Franchise Tax Board's use of an alternate formula pursuant to section 25137. The court ruled that the Franchise Tax Board could impose a formula that included only net futures sales gains in the sales factor ("Net Receipts (Gains Only)"; hereafter, Net Gains Alternative). The Statement of Decision became final November 12, 2010, and the court entered judgment in January 2011.

II. DISCUSSION

(1) The standard UDITPA formula "provides a rough but constitutionally sufficient approximation of the income attributable to business activity in each state." (Microsoft, supra, 39 Cal.4th at p. 756.) Section 25137 is a "relief provision" that allows the taxpayer to seek, or the Franchise Tax Board to impose, "an alternate method of calculation to achieve an equitable result." (Microsoft, at p. 757.) Section 25137 provides, "If the allocation and apportionment provisions of this act do not fairly represent the extent of the taxpayer's business activity in this state, ... the Franchise Tax Board may require, in respect to all or any part of the taxpayer's business activity, if reasonable: [¶] (a) Separate accounting; [¶] (b) The exclusion of any one or more of the factors; [¶] (c) The inclusion of one or more additional factors which will fairly represent the taxpayer's business activity in this state; or [¶] (d) The employment of any other method to effectuate an equitable allocation and apportionment of the taxpayer's income."

(2) "As the party invoking section 25137, the [Franchise Tax] Board has the burden of proving by clear and convincing evidence that (1) the approximation provided by the standard formula is not a fair representation, and (2) its proposed alternative is reasonable. [Citations.]" (Microsoft, at p. 765.)

The limited case law in this area consists primarily of a set of administrative decisions and court opinions focusing on whether including the gross receipts of a multistate or multinational corporation's treasury department misrepresents the extent of the corporation's business activity in California such that section 25137 applies. Microsoft, for example, dealt with tax treatment of income generated by investment of excess operating cash in short-term marketable securities. (Microsoft, supra, 39 Cal.4th at p. 757.) With the exception of a single case involving the treasury department of a financial services firm (Appeal of Merrill, Lynch, Pierce, Fenner & Smith, Inc. (1989) 89 SBE 017 [Cal. Tax Rptr. (CCH) ¶ 401-740] (Merrill Lynch)), all of these cases — hereafter, the treasury cases — held that section 25137 applied. (See Microsoft, at p. 771; Limited Stores, Inc. v. Franchise Tax Bd., supra, 152 Cal.App.4th 1491 [62 Cal.Rptr.3d 191] (Limited) [treasury department of chain of retail clothing stores]; Appeals of Pacific Telephone & Telegraph (1978) 78 SBE 028 [Cal. Tax Rptr. (CCH) ¶ 205-858] (Pacific T&T) [treasury department of telephone company]; In re: Buffets Holdings, Inc. (Bankr. D.Del. 2011) 455 B.R. 94, 99-100 (Buffets) [treasury department of buffet restaurant chain; applying California law].)

This case does not fit squarely within the treasury case paradigm. Here, the challenged gross receipts come not from short-term investment or currency transactions only peripherally related to the company's principal business, but from a support activity integral to the company's main line of business that nevertheless has no direct profit-making purpose. We conclude that the circumstances before us present a different and equally valid paradigm for application of section 25137.

A. Fair Representation

(3) As the trial court correctly observed, section 25137 applies "where the particular function or activity is qualitatively different from the taxpayer's principal business and the quantitative distortion from inclusion of the receipts of that function or activity ... is substantial." (Italics added, citing Microsoft, supra, 39 Cal.4th at p. 766 & Limited, supra, 152 Cal.App.4th at p. 1498.)

General Mills argues that qualitative difference and quantitative distortion are independent and separate requirements such that, if this court determines the futures sales do not meet a certain threshold of qualitative difference, the inquiry ends and the standard formula must be used. The trial court found: "While the parties in their proposed Statements of Decision have discussed two separate tests as requirements, the qualitative and quantitative, these decisions don't discuss separate qualitative and quantitative tests but rather the discussion concerns both effects." We agree. As Microsoft states, "the statutory touchstone remains an inquiry into whether the formula `fairly represent[s]' a unitary business's activities in a given state ...." (Microsoft, supra, 39 Cal.4th at p. 770; see id. at p. 771 [concluding the distortion caused by including treasury gross receipts "is of both a type and size properly addressed through invocation of section 25137" (italics added)].) The ultimate goal is assessing whether the standard formula fairly represents the company's business activity in California.

1. Standard of Review

The standard of review for a section 25137 determination is not clearly established. In Microsoft, the court applied de novo review after determining that the case involved the application of a tax law to undisputed facts.8 (Microsoft, supra, 39 Cal.4th at p. 758.) In Limited, we reviewed a summary judgment ruling and thus applied de novo review. (Limited, supra, 152 Cal.App.4th at pp. 1495-1496.) Buffets was also a summary judgment case. (Buffets, supra, 455 B.R. at p. 97.)

General Mills argues de novo review should apply in this appeal as it did in Microsoft, supra, 39 Cal.4th 750. Although, as here, the decision under review in Microsoft followed a bench trial, the Supreme Court there found the section 25137 issue to turn on undisputed and stipulated facts. (39 Cal.4th at pp. 757-758, 770.) Here we review the trial court's factual findings and conclusions of law following a trial where conflicting evidence, and contested expert testimony, was considered by the court, and where the court made specific credibility determinations. We do not agree that Microsoft dictates the appropriate standard of review in this different procedural context.

We agree with the trial court that "[t]he determination of the section 25137 issue presents a mixed question of fact and law." (Crocker National Bank v. City and County of San Francisco (1989) 49 Cal.3d 881, 888 [264 Cal.Rptr. 139, 782 P.2d 278] (Crocker).) "Mixed questions of law and fact concern the application of the rule to the facts and the consequent determination whether the rule is satisfied." (Ibid.) Here, the trial court made factual findings about the nature of General Mills's futures sales as compared to its sales to customers for profit, the difference in the company's financial results from those two types of sales, and the impact on the standard formula when futures sales are included in the sales factor. Once those findings were made, the court then needed to apply the "fair representation" legal standard to determine whether section 25137 relief was warranted.9 That is, the court applied a rule to facts and determined whether the rule was satisfied.

Our standard of review on a mixed question of fact and law depends on the type of inquiry involved. "If the pertinent inquiry requires application of experience with human affairs, the question is predominantly factual and its determination is reviewed under the substantial-evidence test. If, by contrast, the inquiry requires a critical consideration, in a factual context, of legal principles and their underlying values, the question is predominantly legal and its determination is reviewed independently. [Citation.]" (Crocker, supra, 49 Cal.3d at p. 888; see Ghirardo v. Antonioli (1994) 8 Cal.4th 791, 800-801 [35 Cal.Rptr.2d 418, 883 P.2d 960].) Here, the inquiry is primarily a legal one. In determining whether the facts demonstrate sufficient qualitative difference and quantitative distortion to justify departure from the standard formula, we must consider the impact of our ruling on national uniformity in state taxation of national and multinational corporations, predictability in commerce for business taxpayers and in tax collection for state governments, and the risk of creating exploitable loopholes in the tax system. These are public policy questions that favor de novo appellate review.

In sum, we will review the trial court's findings of historical and quantitative facts for substantial evidence and the court's conclusions about qualitative difference, quantitative distortion, and fair representation de novo.

2. Factual Disputes

The only significant factual dispute raised on appeal is whether hedging was critical to or merely important to General Mills's profitability and viability as a company. In General Mills I, relying on undisputed evidence in the record of the first trial, we wrote, "If General Mills did not hedge the price of grain, it would encounter severe fluctuations in its costs of goods. In such instances, General Mills would have to choose between selling at a loss or not selling at all, particularly for products such as flour where the cost of grain is about 85 percent of the selling price. Although General Mills may not make any profit on its futures trades, and may in fact experience a net loss, it would not be able to achieve its current profit margins on its ultimate product (e.g., flour and cereal) sales without the price protection of hedging." (General Mills I, supra, 172 Cal.App.4th at pp. 1539-1540.) The trial court, however, found "it was not true that [General Mills's] hedging through the use of futures trading saved the company from going out of existence. [Citation.] [¶] In fact, [General Mills's] business was extremely strong and profitable and it was in no danger of going bankrupt. [Citations.] ... Grain costs are only about 15% of the total costs of [General Mills]. [Citation.] Thus, though futures trading may have been important in managing the risk due to the volatility in the price of grain, it affected a relatively small part of [General Mills's] highly stable and profitable business.... [¶] ... Many other activities (e.g. food safety, collection activities, etc.) may be helpful or even necessary to a business, but they are essentially supportive in nature, just as the treasury activities were held to be supportive in nature in the treasury function cases."

General Mills vigorously disputes the trial court's finding as inconsistent with General Mills I and unsupported by substantial evidence. We find it unnecessary to address General Mills's arguments because the finding is not dispositive of the appeal. Even assuming hedging is critical to General Mills's profitability and viability, as General Mills argues, that does not end the inquiry. In Microsoft, the Supreme Court observed that trading marketable securities is "a critical aspect of the operations of the treasury departments of many large corporations" (Microsoft, supra, 39 Cal.4th at p. 755), but nonetheless concluded that mixing the gross receipts of short-term investment trading with the company's gross receipts from its other business activities seriously distorted income attribution under UDITPA. (39 Cal.4th at p. 770.) We conclude that, on the evidence presented to the trial court, inclusion of gross futures sales in the standard apportionment formula results in an unfair representation of the extent of General Mills's business activity in California.

3. Qualitative Analysis

We agree with the trial court that General Mills's futures sales are qualitatively different from their sales of end products to customers for profit.

a. Trial Court Statement of Decision

The trial court acknowledged that General Mills's futures trading "is a risk management tool. That is because, as a registered hedger, the purpose of [General Mills's] futures trading is not to make a profit directly [citation], but to lock in the cost of the grain that [General Mills] uses or sells. [Citation.]" The court identified several additional qualitative differences between General Mills's hedging activity and its main lines of business, including (1) ownership of commodities at the time the activity takes place, (2) intention to sell physical commodities during the activity, (3) delivery or nondelivery of physical commodities during the activity, (4) number of employees engaged in the activity, (5) plant and equipment needed for the activity, (6) storage space needed for the activity, (7) transportation needed for the activity, (8) production costs for the activity, (9) scalability of the activity, (10) time required to complete the activity, (11) time between purchase and sale in the activity, (12) capital requirements for the activity, (13) payment at the initiation of a sale, (14) net payments by and to General Mills during the activity, (15) necessity of consumer demand for the activity, (16) contribution of the activity to the company's profit and loss, (17) rate at which transactions are cancelled in the activity, (18) profit motive behind the activity, and (19) significance to shareholders as reflected in General Mills's annual reports.10

Summarizing, the court ruled that General Mills's "futures trading activities serve a primarily supportive function as a form of price insurance or risk management tool to control the price of the grain it needs which affects the costs of goods sold. Unlike sales of finished consumer products and flour in [General Mills's] main line of business (the gross receipts of which are already included in the sales factor), futures sales do not bear the same costs of equipment, storage facilities, transportation, or requisite ownership of inventory. Moreover, the considerable difference between the time (months versus minutes), labor, and costs required to produce [General Mills's] flour and consumer food products versus to trade futures contracts further supports the Court's finding that [General Mills's] futures trading is qualitatively different from its main line of business."

b. Analysis

Substantial evidence supports the trial court's factual findings, and we agree with the court's conclusion that General Mills's futures sales are qualitatively different from its sales of consumer food products, flour and grain for profit. Hedging futures sales serve a risk management function and are not sales for profit. They rarely result in actual delivery of and payment for goods. They serve an important and even a critical supportive function to General Mills's ultimate sales of profit because they protect against the risk of price fluctuations in basic commodities that are needed to produce the end products. However, they play only a supportive function and would be economically meaningless if separated from ultimate sales of grain, flour and consumer food products for profit.

(4) General Mills argues this conclusion is inconsistent with our rulings in General Mills I. We disagree. In General Mills I we clearly acknowledged that hedging serves a risk management function for General Mills, does not result in reliable or significant profits and often results in losses, and rarely results in actual sales of physical commodities. (General Mills I, supra, 172 Cal.App.4th at pp. 1539-1540.) Although we also stated that hedging was critical to the success of General Mills's primary lines of business, this fact does not preclude a finding of qualitative difference. As we explained in Limited, the qualitative inquiry does not turn on whether the challenged activity is fundamental or integral to the taxpayer's primary business. (Limited, supra, 152 Cal.App.4th at p. 1499.) The Limited Stores, Inc., had argued that its treasury department was integral to its retail operations because the company needed to manage a cash reserve to meet consistent cyclical needs to purchase seasonal inventory. (Ibid.) Without questioning the accuracy of the assertion, we concluded it was not dispositive: "It is almost always true that a treasury department's revenue production will be utilized to support or enhance the company's primary business. The qualitative test adopted in Microsoft would be illusory if The Limited's interpretation of it were adopted." (Ibid., italics added; see Buffets, supra, 455 B.R. at p. 99 [rejecting argument that treasury function was not qualitatively different because maintaining and investing large reserve of working capital was critical to operations].) As the trial court correctly concluded, the same analysis applies here.

General Mills argues it makes no sense to differentiate between its futures sales and its sales of grain, flour and consumer food products for profit and compare separate financial measures of those activities because, as we acknowledged in General Mills I, supra, 172 Cal.App.4th at p. 1547, the futures sales are an integral and even a critical part of General Mills's core business. As noted ante, General Mills correctly identifies an important distinction between this case and the treasury cases. In the treasury cases, the companies made use of their large pools of otherwise idle working capital by investing them in short-term securities. While unquestionably important and intended to be profitable, the investment activity is only incidental to the principal corporate business purpose. Accordingly, the Supreme Court approvingly cited language in a state Board of Equalization decision that "operation of a large treasury department unrelated to a taxpayer's main business is a paradigmatic example of circumstances warranting invocation of section 25137." (Microsoft, supra, 39 Cal.4th at p. 766, italics added, citing In the Matter of the Appeal of Crisa Corp. (2002) 02 SBE 004 [Cal.Tax Rptr. (CCH) ¶ 403-295].)

General Mills's hedging activity is not directly analogous. Its principal focus is not to serve as a profit center for the company, and it is not unrelated to General Mills's main business; it is a risk management tool that directly supports General Mills's main line of business. It does not follow, however, that section 25137 is inapplicable. Rather, as we explain further post, this case presents a different but equally valid paradigm warranting invocation of section 25137: sales activity that is not conducted for its own profit, and that has a substantially distortive effect on the standard apportionment formula, resulting in an unfair representation of the company's business activity in California.

General Mills argues that an activity is qualitatively different only if it has no value to the unitary business beyond the income it directly generates. General Mills relies on Pacific T&T's statement that the treasury gross receipts are appropriately excluded from the sales factor in part because "[w]e have serious doubts ... whether the turnover of assets in those pools [of working capital] has any value to the unitary business beyond the income that it generates directly." (Pacific T&T, supra, [1978-1981 Transfer Binder] Cal. Tax Rptr. (CCH) ¶ 205-858, p. 14,907-43.) In the trial court, General Mills offered expert testimony that its hedging activities were not qualitatively different than its principal business activities because of "synergies" that added income beyond that generated by the futures trading in isolation. Aside from the fact that the trial court found the testimony unpersuasive, the question again is not simply whether the activity has some demonstrated value to the unitary business beyond generation of income, and Pacific T&T does not stand for the proposition that income activity must be included in the sales factor if it has any value to the business beyond the income it generates directly.

(5) Finally, General Mills also argues that its hedging activity is not the sort of unusual, atypical fact situation for which section 25137 was designed. It cites California Code of Regulations, title 18, section 25137, which provides in part, "[Revenue and Taxation Code] [s]ection 25137 may be invoked only in specific cases where unusual fact situations (which ordinarily will be unique and nonrecurring) produce incongruous results under the apportionment and allocation provisions contained in these regulations." The Supreme Court, however, specifically rejected a similar argument in Microsoft: "Systematic oversights and undersights are equally a matter of statutory concern. Nothing in the language of [California Code of Regulations, title 18,] section 25137 persuades us otherwise. While Revenue and Taxation Code section 25137 `ordinarily' applies to nonrecurring situations, it does not apply only to such situations; the statutory touchstone remains an inquiry into whether the formula `fairly represent[s]' a unitary business's activities in a given state, and when it does not, the relief provision may apply. [Citations.]" (Microsoft, supra, 39 Cal.4th at p. 770, fn. omitted.)

In sum, we agree with the trial court that General Mills's futures sales are qualitatively different from General Mills's sales of grain, flour and consumer food products for profit.

4. Quantitative Analysis

While we differ from the trial court on certain specific aspects of its analysis, we agree with the court's ultimate conclusion that including futures sales in the sales factor quantitatively distorts the standard apportionment formula such that it does not fairly represent the extent of General Mills's business in California. In particular, we agree that a key quantitative metric — profit margin — weighs heavily in favor of a finding of substantial distortion. Although the percentage impact on the standard formula here is less severe than in the treasury cases, we conclude it is sufficient in all the circumstances to warrant application of section 25137.

a. Trial Court Statement of Decision

The trial court ruled, in summary, "that the quantitative metrics in the instant case are comparable to those in Pacific [T&T], Microsoft, and Limited and there is sufficient distortion to warrant section 25137 relief. On average for the years in issue, including futures gross receipts in the sales factor denominator results in assigning close to 9% ... of [General Mills's] entire business activities to Minnesota. [Citations.] On average for the years in issue, [General Mills's] futures trading activity produced no income (i.e., generated a negative percentage [-0.1353%] of [General Mills's] business income) but generated close to 19% [18.871%] of [General Mills's] total gross receipts. [Citations.] ... On average during the years in issue, [General Mills's] non-futures trading profit margin exceeded the futures trading profit margin by 81 times. [Citation.] Thus, this court holds that the [Franchise Tax Board] has proven by clear and convincing evidence that the rote application of the standard apportionment formula will result in an unfair representation of the extent of [General Mills's] business activity in California."11 (Fn. omitted.)

b. Analysis

We consider each of the quantitative metrics in turn.

i. Amount of Business Activity Attributed to a Single State

In Pacific T&T, the state Board of Equalization rejected the contention that the corporation's New York-based treasury department's gross receipts should be included in the UDITPA sales denominator: "we are unable to accept, even for a moment, the notion that more than 11 percent of [Pacific Telephone & Telegraph's] entire unitary business activities should be attributed to any single state solely because it is the center of working capital investment activities that are clearly only an incidental part of one of America's largest, and most widespread, businesses." (Pacific T&T, supra, [1978-1981 Transfer Binder] Cal. Tax Rptr. (CCH) ¶ 205-858, p. 14,907-43.) The Supreme Court in Microsoft cited Pacific T&T with approval, commenting, "If one substitutes `Washington' for `New York' and `24 percent' for `11 percent,' these words are equally applicable" to the Microsoft case. (Microsoft, supra, 39 Cal.4th at pp. 765-766.) In Limited, we held that sufficient distortion would be shown by evidence that inclusion of treasury department gross receipts would result in 9.25 percent of the company's business activities being assigned to Ohio. (Limited, supra, 152 Cal.App.4th at p. 1501.) In Buffets, the court noted that including gross receipts "would treat an average of over 38.5 percent of the [taxpayers'] income as though it were apportioned to Minnesota." (Buffets, supra, 455 B.R. at p. 101.)12

The trial court found that the inclusion of General Mills's futures trading gross receipts "causes the standard apportionment formula to assign to Minnesota ... a high of about 15 percent of [General Mills's] entire business activities (1996 and 1997) and a low of 2.8 percent of [General Mills's] entire business activities (1992). On average, including futures gross receipts in the sales factor denominator results in assigning close to 9% [8.722%] of [General Mills's] entire business activities to Minnesota."

We agree with the trial court that this factor weighs in favor of a finding of substantial distortion.

ii. Percentage of Income vs. Percentage of Gross Receipts

In Microsoft, the Supreme Court also found persuasive Pacific T&T's comparison of the percentages of income and gross receipts that were generated by the company's treasury department. (Microsoft, supra, 39 Cal.4th at p. 765.) In Pacific T&T, the treasury department "produced less than 2 percent of the company's business income, but 36 percent of its gross receipts." (Microsoft, at p. 765, fn. omitted.) Microsoft observes, "By comparison, the distortional impact is even greater here; Microsoft's short-term investments produced less than 2 percent of the company's income, but 73 percent of its gross receipts." (Id. at p. 765, fn. 17.) In Limited, we noted that the treasury department generated less than 1 percent of the company's income in both 1993 and 1994 but 62 percent and 52 percent, respectively, of the company's gross receipts. (Limited, supra, 152 Cal.App.4th at p. 1500.) In Buffets, the treasury department generated 5.4 percent of the company's income and 77 percent of its gross receipts. (Buffets, supra, 455 B.R. at p. 101.)

The trial court found that General Mills's hedging activities produced at most 2 percent of the company's income (and in two of six years operated at a loss) while it generated between 8 and 30 percent of the company's gross receipts.13 We agree this metric weighs heavily in favor of a finding of substantial distortion, especially because the minimal profit realized in General Mills's hedging activity is essentially serendipitous, as the ideal outcome of hedging is zero profit or loss.

iii. Profit Margin

(6) In Microsoft, the Supreme Court paid particular attention to the difference in profit margin in the challenged activity as compared to the primary business of the company. "[T]he problem is one of scale: short-term securities investments involve margins (i.e., differences between cost and sale price) that may be several orders of magnitude different than those for other commodities. When a short-term marketable security is sold or redeemed, the margin will often be, in absolute terms, quite small (though of course the annualized returns may well be perfectly respectable). Microsoft's treasury activities provide a perfect illustration. Its 1991 redemptions totaled $5.7 billion, while its income from those investments totaled only $10.7 million — a less than 0.2 percent margin. In contrast, its nontreasury activities produced income of $659 million and gross receipts of $2.1 billion, for a margin of more than 31 percent, roughly 170 times greater. [¶] This situation, when one mixes apples — the receipts of low-margin sales — with oranges — those of much higher margin sales — presents a problem for the UDITPA. The UDITPA's sales factor contains an implicit assumption that a corporation's margins will not vary inordinately from state to state.... [¶] ... [¶] ... [The] formula in effect estimates the income attributable to a state by multiplying the average worldwide margin by the in-state receipts to approximate the in-state income. [¶] This approximation works well enough in the absence of huge variations in state-to-state margins.... However, ... [u]nder the UDITPA, the operations and gross receipts of a treasury department are properly attributed to the state where the department operates — here, Washington. (See § 25136.)

(7) The nature of these operations means that Microsoft's true margin for its Washington operations will be much, much lower than the worldwide average, and its margin for every other state will be much higher than the worldwide average. Thus, rotely applying the worldwide average margin (Total Income/Total Sales) to each state's gross receipts would result in severely underestimating the amount of income attributable to every state except the state hosting the treasury department, for which state the income would be correspondingly severely overestimated. In such circumstances, rote application of the standard formula does not fairly represent the extent of a taxpayer's activity in each state, except in the rare instance when corresponding imprecision in the payroll and property factors may happen to balance out this distortion." (Microsoft, supra, 39 Cal.4th at pp. 767-769, fns. omitted.)

As noted, the Supreme Court in Microsoft found that the profit margin for the company's business was roughly 170 times greater than its profit margin for treasury activities. (Microsoft, supra, 39 Cal.4th at p. 767.) In Limited, the profit margin was less than 0.1 percent on treasury activities and more than 46 percent on the rest of the company's business (roughly 460 times greater), which we found to be a substantial quantitative distortion. (Limited, supra, 152 Cal.App.4th at p. 1500.) But in Buffets, the bankruptcy court found that a profit margin of 0.08 percent for the treasury department and 4.25 percent for the company's main business (53 times greater) was "well within the range of quantitative differences which the California courts have found support the application of section 25137."14 (Buffets, supra, 455 B.R. at pp. 100-101.)

The trial court in this case found that, on average for the tax years in issue, General Mills's nonfutures profit margin exceeded the futures profit margin by 81 times. In making this finding, the trial court relied on the Franchise Tax Board's calculation from trial exhibit No. 117 (Schedule I). This exhibit indicates that the profit margin on General Mills's futures trading was 0.75 percent at best, with an average of a 0.04 percent loss, whereas the profit margin on the rest of the company's business was on average 7.17 percent. For tax years in which General Mills's futures trading showed a profit, the nonfutures profit margin exceeded the futures profit margin by 153 times in TYE (tax years ending) 1993, 9 times in TYE 1994, 36 times in TYE 1995, and 126 times in TYE 1997.

It could be argued that the ratio of 81 times is misleadingly high because the profit margin figures under comparison are relatively small. However, one could also argue the factor is misleadingly low because futures sales are as likely to result in losses or come out even as to generate a profit. Nevertheless, there is no question this factor weighs heavily in favor of a finding of substantial quantitative distortion.

iv. Percentage Change in the Standard Apportionment Formula

The trial court did not consider the degree to which inclusion of the challenged activity ultimately affected the result under the standard apportionment formula, an issue we highlighted in our remand language in General Mills I. (See General Mills I, supra, 172 Cal.App.4th at p. 1548.) In Microsoft, the Supreme Court twice noted that including treasury gross receipts in the standard formula reduced Microsoft's California income tax nearly in half. (Microsoft, supra, 39 Cal.4th at pp. 757, 771.) In Buffets, inclusion of the treasury receipts reduced the company's California tax liability by an average of 44 percent over six years. (Buffets, supra, 455 B.R. at p. 101.) In Limited, the gross receipts from the treasury department decreased the company's California tax liability by 21 percent and 26 percent in the two years in issue, which we found sufficient to support a finding of substantial distortion. (Limited, supra, 152 Cal.App.4th at p. 1495, 1500.)

As noted ante, inclusion of General Mills's hedging gross receipts changed the apportionment formulas from about 10.9 percent to 10.5 percent for TYE 1992, from 11.2 percent to 10.8 percent for TYE 1993, from 11 percent to 10.3 percent for TYE 1994, from 10.4 percent to 9.5 percent for TYE 1995, from 10.8 percent to 9.3 percent for TYE 1996, and from 10.2 percent to 8.9 percent for TYE 1997. The percentage reductions in the standard apportionment figure thus ranged from 3.6 percent (TYE 1993) to 13.9 percent (TYE 1996), or an average of 8.2 percent.

Clearly, the ultimate impact on the standard formula here is less severe than in the treasury cases. However, the case law does not indicate that this quantitative metric, or any one metric, alone is dispositive.

5. Summary of Fair Representation Analysis

In sum, while some of the quantitative distortions in this case may not be as great as those cited in the treasury cases, they are nevertheless substantial. In the area of profit margin, which the Supreme Court identified as critical in Microsoft, the distortion arguably is greater here. And the treasury sales were made for the purpose of profit. Hedging for General Mills is not intended to be a profit center at all, although it is intended to facilitate the business of the company. If all works perfectly in such transactions, the profit will be zero. Given the inherent vagaries of the markets, futures sales sometimes return a profit or a loss. Although in several of the years at issue, the sales returned a net profit, it is a misnomer to refer to the activity's "profit margin" as if it were a goal or a reliable outcome of the hedging activity. For this reason, hedging both qualitatively differs from the General Mills's sales of grain, flour and consumer food products for profit and substantially distorts the standard formula's sales factor, which is designed to reflect the market for General Mills's goods. (See Microsoft, supra, 39 Cal.4th at p. 765, citing Pacific T&T, supra, Cal. Tax Rptr. (CCH) ¶ 205-858.) As General Mills itself argues, the purpose of hedging is to achieve the profit margins in the company's primary business. Using hedging gross receipts to dilute that profit margin, therefore, does not fairly represent California's market for General Mills's goods.

General Mills argues this conclusion is inconsistent with our statement in General Mills I that "including futures sales in the sales factor is consistent with the purpose of UDITPA." (General Mills, supra, 172 Cal.App.4th at p. 1547, some capitalization omitted.) Specifically, we held that including futures sales was consistent with the purpose of the UDITPA sales factor, which is "designed to reflect the taxpayer's `income producing activity,' which regulations define as `transactions and activity directly engaged in by the taxpayer in the regular course of its trade or business for the ultimate purpose of obtaining gains or profit.' (Cal. Code Regs., tit. 18, § 25136, subd. (b); see Rev. & Tax. Code, §§ 25134-25136.)" (172 Cal.App.4th at p.1547.) In other words, our focus was narrow, considering only the purpose of the gross receipts standard in calculating the sales factor. Our statement did not preclude a finding that the standard formula, with the inclusion of gross receipts from hedge trading, does not fairly represent General Mills's activity in California, and we remanded to the trial court for the specific purpose of making such a determination.

Finally, we comment on the public policy issues implicated by our holding. The trial court specifically found that allowing gross receipts from futures sales to be included in the UDITPA sales factor could result in "nowhere income," because the court received evidence that General Mills does not include futures sales in its Minnesota tax returns. Our holding does not substantially affect either the need for uniformity in state taxation of national corporations or the need for predictability in taxation and state tax revenue collection because the Legislature has already determined how hedging gross receipts will be handled in tax years beginning in and after 2011: they will be excluded from the sales factor. Moreover, our decision to uphold the Franchise Tax Board's use of an alternate formula in this circumstance helps to deter "creative" tax accounting that substantially distorts the standard formula's representation of a large company's business activity in California. These are not minor matters. Even an average 8.2 percent distortion of income is not insubstantial. The consequence may be differences in tax liability running into the millions of dollars — the very reason these questions are so vigorously litigated.

(8) We affirm the trial court's ruling the Franchise Tax Board proved by clear and convincing evidence that the standard UDITPA formula did not fairly represent the extent of General Mills's business activity in California, thus authorizing the Franchise Tax Board to impose a reasonable alternate formula under section 25137.

B. Reasonable Alternative

The Franchise Tax Board was also required to prove by clear and convincing evidence that the alternative formula it imposed was reasonable. This too is a mixed question of fact and law we review de novo. "If the [Franchise Tax] Board's proposal is reasonable, we are not empowered to substitute our own formula. [Citations.]" (Microsoft, supra, 39 Cal.4th at p. 771.)

1. Background and Statement of Decision

Before this litigation commenced, General Mills filed amended tax returns that included hedging gross receipts in the sales factor of the standard formula. The Franchise Tax Board, however, excluded the full amount of those receipts from the sales factor, reasoning that they did not qualify as gross receipts under UDITPA. After General Mills sued for a tax refund, the Franchise Tax Board defended its interpretation of gross receipts and also argued, in the alternative, that an alternate formula should be imposed under section 25137 that excluded the full amount of hedging receipts (the "No Futures Gross Receipts" alternative; hereafter, No Receipts Alternative). (See General Mills I, supra, 172 Cal.App.4th at pp. 1541-1543.) On remand following our decision in General Mills I, the Franchise Tax Board again proposed the No Receipts Alternative. However, the trial court expressed concern that our decision precluded adoption of that alternate formula and encouraged the Franchise Tax Board to propose different alternate formulas. The Franchise Tax Board then proposed the Net Gains Alternative and identified it as its preferred alternate formula.

In its Statement of Decision, the trial court considered seven alternate apportionment formulas presented by the Franchise Tax Board. It ruled that both the No Receipts and the Net Gains Alternatives were reasonable. The court then concluded that the Franchise Tax Board's then preferred alternative, the Net Gains Alternative, could be adopted.

2. Which Alternative Formula or Formulas Were Properly Before the Court

(9) General Mills argues the trial court exceeded its authority in prompting the Franchise Tax Board to propose a new alternate formula on remand and in approving that new formula as reasonable. General Mills argues the Franchise Tax Board was bound by the parties' joint stipulation of facts in the first trial, which stated, "Under the [Franchise Tax Board's] alternative methodology, pursuant to ... section 25137, none of the gross receipts from General Mills' trading activities in the commodity futures market is includable in General Mills' sales factor for the years in issue." The only authority General Mills cites to support this argument is Times Mirror Co. v. Franchise Tax Board, which held that the parties and the trial court were bound by a stipulation of fact regarding the taxpayer's purpose in acquiring a certain company. (Times Mirror Co. v. Franchise Tax Bd. (1980) 102 Cal.App.3d 872, 875, 877 [162 Cal.Rptr. 630].) Here, the cited stipulation did not concern the underlying facts of the tax case (e.g., the nature of General Mills's hedging activity) that otherwise would have been determined by the trial court, but rather described a legal position taken by one of the parties. A more relevant legal doctrine, therefore, is judicial estoppel, which "`"`precludes a party from gaining an advantage by taking one position, and then seeking a second advantage by taking an incompatible position.'"'" (MW Erectors, Inc. v. Niederhauser Ornamental & Metal Works Co., Inc. (2005) 36 Cal.4th 412, 422 [30 Cal.Rptr.3d 755, 115 P.3d 41].) General Mills has not shown that the Franchise Tax Board gained any unfair advantage by initially seeking a No Receipts Alternative formula in this litigation and later stating a preference for a Net Gains Alternative.

(10) In any event, the stipulation at issue did not state that the Franchise Tax Board sought only a No Receipts Alternative. General Mills's argument is premised on the assumption that the Franchise Tax Board must identify one alternative formula, and that its case thereafter must rise or fall on the reasonableness of that alternative. We are not persuaded. The statute provides that, if the standard formula does not fairly represent the extent of the taxpayer's business in the state, the Franchise Tax Board "may require, in respect to all or any part of the taxpayer's business activity, if reasonable: [¶] ... [¶] ... [t]he employment of any ... method to effectuate an equitable allocation and apportionment of the taxpayer's income." (§ 25137, italics added.) Nothing in the statute precludes the Franchise Tax Board from proposing alternative methods to effectuate an equitable apportionment during tax litigation.

3. The Net Gains Alternative Is Reasonable.

We agree that the Net Gains Alternative is reasonable. As the trial court observed, allowing consideration of net trading gains still gives some representation to the futures trading activity in the sales factor and the formula treats General Mills consistently with the taxpayers in the treasury cases.

Because hedging is a form of risk management for General Mills's core business and is not a direct profit-seeking enterprise on its own, it would be reasonable to apportion hedging gross receipts to each state according to the portion of General Mills's sales for profit that take place in the state, rather than apportioning them all to the home state. As a mathematical matter, such an apportionment is equivalent to excluding the hedging gross receipts from the sales factor altogether. The Franchise Tax Board's decision is more generous to the taxpayer in including hedging net gains so that hedging activity is represented to some degree in the sales factor as well as in the payroll factor (the cost of hedging personnel) and property factor (the office facility costs associated with hedging). It is therefore also reasonable. (Cf. Microsoft, supra, 39 Cal.4th at p. 771 [noting that in some cases including only net treasury receipts in the sales factor "may go too far in the opposite direction and fail the test of reasonableness"].)

C. Constitutional Issue

(11) General Mills halfheartedly argues that excluding hedging gross receipts from the sales factor violates constitutional limits on state taxation of interstate companies. As the Franchise Tax Board correctly observes, a constitutional violation occurs only if the apportionment formula imposed by the Franchise Tax Board is "`"out of all ... proportions to the business transacted ... in that State."'" (Container Corp. v. Franchise Tax Bd. (1983) 463 U.S. 159, 170 [77 L.Ed.2d 545, 103 S.Ct. 2933]; see Microsoft, supra, 39 Cal.4th at p. 765, fn. 16.) General Mills made no real effort in the trial court to make such a showing, and makes none here.

III. DISPOSITION

The judgment is affirmed. General Mills shall bear the Franchise Tax Board's costs on appeal.

Simons, Acting P. J., and Needham, J., concurred.