WHAT IS A DERIVATIVE WORTH?

The Los Alamos Valuation Model

By Viva Hammer, Esq, Price Waterhouse LLP

Not everything produced by the government is burdensome and wasteful. Take, for example, electronic mail and space-food sticks. However, examples to the contrary are so numerous, I will not even begin to list them. We have yet to see into which category the new software, developed by ex-rocket science physicists to value "securities" under IRC section 475, fits.

In 1993, Congress introduced section 475 which requires "dealers in securities" to mark to market all securities not held for investment at year-end. A dealer is anyone who--

* regularly purchases securities from or sells securities to customers in the ordinary course of a trade or business, or

* regularly offers to enter into, assume, offset, assign, or otherwise terminate positions in securities with customers in the ordinary course of a trade or business.

A security includes stock, interests in publicly traded partnerships, bonds, notional principal contracts, or any interest in a derivative financial instrument used to hedge one of the above.

After determining whether a taxpayer is a dealer, the central question under IRC section 475 is "what is the year-end mark?" Since the mark-to-market concept is one that has been adopted from the securities industry, taxpayers assumed that the marks used by the industry, if fair and truthful, would be accepted by the IRS. The IRS felt, however, that it had to reinvent the wheel, or perhaps pressure was brought to bear to gainfully employ all the physicists dispossessed after the end of the Cold War. Whatever the impetus, the Department of Energy agreed that the IRS could utilize the scientists at the Los Alamos National Laboratory to develop software to value financial products. On April 3, 1997, taxpayers were given a demonstration of the model.

The software developed at Los Alamos is designed to generate a range of reasonable values for financial products. Any discrepancy between the numbers generated by the software and a taxpayer's numbers will not lead to an automatic adjustment, according to the IRS, but rather will provide the framework for follow-up discussions with the taxpayer. The software is designed to be able to take into account the information and modeling capacities that existed in the year of the audit, through the use of interest rate and exchange rate databases. The IRS chose to base its software on five seminal derivative models: Ho-Lee, Hull-White, Black-Derman-Toy, Black-Karasinski, and Heath-Jarrow-Morton, each incorporating one level higher complexity than the last. In addition, an agent will be able to pick an alternative model if that seems more appropriate for the taxpayer's circumstances.

Taxpayers who will be audited using this software have reacted to the IRS presentations with extreme nervousness. They argue that the figures which taxpayers use for regulatory, accounting, reporting, and compensation purposes should be sufficiently accurate to be accepted without quibble, since these figures are generated for parties that often represent diverse interests. In addition, taxpayers claim that the information and methodologies with which they constructed their original derivative valuations have long since been lost and there is no way that they can be recreated with accuracy. Taxpayers are also concerned that although the scientists developing the software are sophisticated, it is unlikely that the IRS auditors using the software will have the expertise or experience to understand the nuances in the models, and understand where they should be modified or abandoned entirely. Taxpayers are afraid of large numbers of adjustments leading to disputes that will be almost impossible to resolve.

Countering these arguments, the IRS has said that if taxpayers would be more accommodating in providing information and technical data, there would not be a need for the expected "battle of the models." The argument that taxpayers have long disposed of their valuation methodologies is hardly legitimate, because taxpayers have been on notice of the possibility of having to justify their numbers since the enactment of IRC section 475. The IRS also points out that although large institutions are more likely to have fair values for their derivatives because competing parties use information generated by their systems, that safeguard does not apply to small taxpayers that are not subject to CFTC or other regulatory scrutiny. Finally, the IRS reminds taxpayers that the more they are involved in the development of the software and its implementation, the less likely it is that they will suffer unnecessarily at the hands of auditors.

Whatever the outcome of the current dialogue, there is no question that some version of the software is here to stay. Taxpayers with significant interest in the outcome of the Los Alamos project are well advised to associate with the Securities Industry Association, bar associations, or other organizations which have organized responses to the development of the project.

And there is always a slim chance that the excellent minds at Los Alamos will come up with something so original in the field of finance that all our lives will be improved. *