Employee Stock Options and How the IRS Cannot Win a Court Case

The SEC has rejected an approach to value employee stock options (ESO) proposed by Cisco Systems:

Securities and Exchange Commission Chief Accountant Donald Nicolaisen is expected to reject Cisco’s Systems … The option valuation question is pressing because large U.S.-listed companies, under a new accounting rule, in fiscal years begun after June 15 had to start treating stock options as a routine business expense, such as salaries or bonuses. The new “option expensing” rule, adopted by accounting standard setters and backed by the SEC, was fiercely opposed by Cisco and other Silicon Valley powerhouses because it was expected to raise their operating costs and cut into profits … Cisco put forward its alternative, market-based idea six weeks later. The SEC said then it would evaluate the proposal. Cisco had argued that valuation methods proposed by regulators put too high a value on stock options.

While the Black-Scholes approach may be a reasonable means for valuing publicly traded options, the textbook version would tend to overvalue ESOs. Whether the approach suggested by Cisco Systems is a reasonable means for estimating the ex ante cost of issuing ESOs is an interesting question.

In the tax arena, consider the following “cost sharing arrangement” for a U.S. based Silicon Valley company that has agreed with a foreign subsidiary in a low-tax jurisdiction to share all R&D related expenses on the premise that the subsidiary owns the rights to the intangible assets in the Eastern Hemisphere and the U.S. taxpayer owns the rights to the intangible assets in the Western Hemisphere. One would think the IRS could make the case that the ex ante cost of issuing ESOs to its Silicon Valley R&D employees is part of this cost base. In fact, the IRS in its recent pronouncements is making this statement noting that it is litigating this issue in a pending Tax Court case Xilinx Inc. v. Comr. It seems the IRS has tried to litigate this issue for almost a decade with no success. And the really bad news is that the IRS just lost the Xilinx litigation.

The reason for their loss on this issue is clear if one reads the Judge’s decision. The IRS tried this case on two very different theories. One relates to valuing the ESOs on its ex post value after the employee exercises the option. As the Court noted:

Unrelated parties would not share the spread because it is difficult to estimate, unpredictable, and potentially large in amount. Petitioners’ uncontradicted evidence established that certainty and control are of paramount importance to unrelated parties involved in cost-sharing arrangements. Yet, the size of the spread is affected by a variety of factors, many of which are not within the control of the contracting parties.

OK, they failed to win on a theory that is clearly absurd, but what about its alternative theory:

Respondent, who had the burden of proof with respect to the grant date theory, presented no evidence that unrelated parties would

share the fair market value of ESOs at the time of grant. In other words, the IRS goes into Court with a reasonable theory but fails to bother to present any evidence? For almost a decade, the IRS has reasonably suggested that these costs should be included if they can be properly valued. Yet, the IRS cannot provide any evidence to such an obvious point? Then why do we even bother with tax law?