The Problem of Options Expensing: Lessons from Mercury and Electronic Arts

 |  Includes: EA, MERQ
by: David Harper, CFA

Will the backdating scandals promote option expense clarity? I doubt it. This week gave two uninspiring examples.

David Jackson shows why Mercury’s restatement doesn’t hurt as much as the ensuing radioactivity. Like most tech companies, Mercury (MERQ) responded to Sarbanes-Oxley in 2002 by stopping its egregious practices in favor of merely bad practices.

After 2002, the economic magnitude of their accounting error is barely noticeable. What’s surprising is the economic insignificance of the recent (last few years') restatements. The worst restatement I see, in average price terms, is only 3%. That is, an average, booked exercise price of $40.29 was restated to an actual price of $41.56. The others tranches barely register (less than 1% for all of them). But then, the restated numbers don’t help you get a grip on real option costs, either.

Mercury’s real crime was that they were overly diluting shareholders, proper accounting or no. Before they retreated in 2004, they were clocking an equity run-rate (i.e., annual grants as a percentage of outstanding) of about 6% per year. Until the Board, sensing perception problems attached to a 30% overhang (and surely, no easy way to get more incentive shares), layered in a cash long-term incentive plan. That’s a new tactic, to mix in additional plans and commingle costs (e.g., after the option expensing was finalized, a raft of consultants rushed into to recommend stock appreciation rights, SAR because they capitalize on loopholes in the rules). In Mercury’s case, they some much as explained they were doing this to save shares. Of course, cash incentive plans tend to consume a different resource that I like to call cash.

If you are trying to get a true cost and perform comparisons (year-to-year and peer-to-peer), some of these restatements just don’t help. It’s almost funny that, in restating their option expense, Mercury actually discovered plus $15 million to help with their total restated bill. In the restated filing, they found a way to reduce the average cost of an option granted in 2004 by $4.12 dollars to $24.16 (the difference multiplied by 3.3 million shares is a paper gain of +$15 million). That’s not better hindsight, since the value is based on known information at the historical point-in-time. Rather, they retroactively tweaked the Black-Scholes inputs (e.g., lower expected life, lower volatility) and discovered a less conservative but valid set of valuation model inputs for almost identical circumstances.

Large public scandals provide fine cover for titling obscure valuation models in your favor, I guess. It is pretty safe practice to swing the Black-Scholes around because it never gives the right answer for employee stock options (ESOs)—don’t get me wrong, it’s rails for traded options and certain circumstances when its assumptions are good—and so fortunately cannot give a wrong answer. Meanwhile, their newly introduced variable accounting is a bona-fide mark-to-market rule.

Mercury’s 5.9 million options under variable treatment implies that each $1 stock price increase translates into $5.9 million of additional non-cash expense (albeit amortized over the vesting period, so figure about +$1.5 million per year for every +$1 in stock price gain). That’s technically an uncapped liability. Variable accounting is a defensible method to value options—after all, it is arguably in the spirit of fair value accounting—except it no longer applies to the majority of equity (share-settled) awards under option expensing (FAS 123R) so it’s a legacy oddball that further clouds comparisons.

Circularity is the culprit that makes stock options hard to price: the cost of an option depends on the stock price, which inconveniently impacts the cost of the option. Take Electronic Arts’ (ERTS) proposal to exchange underwater options for full value but restricted shares. Although they create a non-cash drag on EPS, the underwater options have no current economic (dilutive) impact. The EA spokesperson reassures us that the voluntarily exchange will reduce overhang, should be favorably received by investors and elsewhere it’s implied that this might positively impact EPS because it reduces the number of outstanding options. One analyst asked whether this is good or bad for shareholders. I ran the numbers: cost of current underwater options: zero. Cost of new restricted shares: something. Additional net cost: something.

Diluted EPS is not perfect but won’t be fooled by this exchange. The underwater options do not show up in diluted EPS as long as they are underwater. The new restricted shares (nonvested shares, as FASB calls them) do count under diluted EPS (FAS 128, treasury stock method) at their full value. So out of the gate, diluted EPS is reduced. The other canard here is too long in the tooth and companies need to stop selling it. That’s the notion that shareholders benefit necessarily by a reduction in the numerical overhang from, say purely for example’s sake, 10 million outstanding but underwater options to only 3 million restricted but fully-valued shares outstanding (based on EA’s stated 3:1 or 4:1 exchange ratio of options: shares). Historically, numerical overhang made a little bit of sense when everyone was granting the same exact instrument, namely, four-year vested FMV options.

But numerical overhang was never economically useful and is now becoming irrelevant for comparisons (company versus peers) as companies diversify away from plain vanilla options. I don’t need to know the exact exchange ratio to know that investors economically prefer the premium-priced (underwater) options to the full value shares (disregarding the motivational benefits, of course). We can take a page for the institutional shareholder services [ISS] playbook here. For a long time, they’ve used a measure called shareholder value transfer [SVT] which could be called economic overhang. That’s overhang but it’s based on the value of the overhang not the number of instruments in the overhang.

To summarize the EA exchange: They suggest that maybe it isn’t a real cost, but surely it is; the accounting will show them as cost, but surely it will be wrong, even if properly stated the first time.