expensing options
Ramesh Ponnuru offers a relatively fair presentation of the options expensing debate, but as a supporter of expensing, I'd like to respond to his arguments against.

Corporations already disclose all the information that the reformers want, albeit not as prominently as they want. Moreover, the impact of the options on price-per-share is revealed the moment the options are exercised — which is to say, the moment a cost is actually incurred. Opponents also say that there is no reliable method of estimating what options will cost years before they are exercised.

All of this, Ramesh indicates, is true; and, in its way, it is. On the other hand, corporations do release earnings statements, and we might as well make them mean something.

The calculation of a bottom-line "profit" is not necessary, perhaps; investors can read all the raw data themselves and do with them what they like. So long as there is to be a bottom line reported, however, it should as accurately as possible reflect the value generated by the company during the reporting period. The cost is not actually incurred when the option is excercised (though that may be the closest it gets to a cash-flow drain, if the company buys back shares to prevent dilution); the option is issued in exchange for labor over a certain period of time, and should be marked against earnings over the same period. That the number assigned to that cost cannot be derived with precision, and that

companies will have to make choices, not only about the model to be used but the various inputs that the model requires

creating a risk of legal liability if they use the wrong one applies no more to this than it does to estimates of pension obligations or depreciation; a company should be immune to prosecution in any case for any assumptions that are not simply untenable. An example of a value for the typical issued option that is untenable is 0; a Black-Scholes model (which understates option values) using 10% annual volatility — a low value for any stock — will still produce a better estimate than what is being used by many companies now.

Update: My brother writes to note that accountants refer to this as "matching"; revenues should be temporally matched with the expenses that generate them. I kind of knew this, but the term "matching" is simply too prosaic for my tastes; if they had called it something in Latin I would have used that term.
- posted by dWj @ 3:45 PM