Semi-Daily Journal
Brad DeLong's Thoughts of the Moment on Economics, and on Other Topics as Well
DeLong's Home Page
November 11, 2002
How to Account for Options
The Economist provides a nice little summary of the chief points at issue in deciding how companies should account for stock options they grant:
Economist.com: ...The IASB does not, though, specify exactly which method of fair-value option pricing should be used. And here academics dispute vigorously. Mr Merton helped to devise the Black-Scholes model for pricing options, named after Myron Scholes (who shared Mr Merton's Nobel prize) and Fisher Black (who would have done, had he lived a little longer). The main alternative is what is called the binomial pricing model, developed by John Cox, Stephen Ross and Mark Rubinstein. Both models were devised to price simple options traded on an exchange. So, as the IASB acknowledges, each needs to be modified to reflect several peculiarities of employee share options—such as longer lifespan or term (a typical executive option lasts ten years, against a few months to two years for an exchange-traded option), restrictions on when options may be exercised, and even the fact that they may not be sold. How these adjustments are made—and the IASB allows wide discretion—can make a big difference to the size of the expense.
The binomial method is much more complex than Black-Scholes, but if done properly, the price it produces is likely to be accurate more often. Of course, the very complexity may make it easier for the price to be manipulated by a company wanting to massage its profits. On the other hand, plenty of scope for creativity exists with the Black-Scholes method, even given the few simple assumptions that are used. These assumptions, in essence, are expected volatility (how much a share price is likely to fluctuate), term, the expected dividend yield, the risk-free interest rate, and the exercise price.
Some critics reckon that these vagaries of valuation undermine the entire case for expensing. Few economists agree, however. Whatever uncertainties there are in fair-value option pricing, goes the argument, they are smaller than the uncertainty in the value of other items already routinely expensed, such as depreciation and pension-fund gains or losses.
The biggest philosophical dispute among economists concerns when options should be expensed. The IASB wants it done once and for all from the date they are awarded to employees (in other words, the grant date). Mr Scholes, the Nobel laureate, agrees. But others, such as Mr Rubinstein, one of the creators of binomial pricing, do not. He argues for full expensing at the time options are exercised, ie, when the holder trades in the options for underlying shares. Under this approach, options would still initially be expensed on the grant date; but in subsequent public filings this estimate would be adjusted to take into account changes in their value. Upon exercise, the company would take a final extraordinary gain or loss to match up with the option's actual value when exercised. Mr Rubinstein argues that, under this method, there would be less incentive to manipulate option valuations, since any divergence from an option's final true value would, in the end, result in an extraordinary charge.
The issue boils down to this: is the granting of an option a once-only expense for the company, the equivalent of paying the employee in cash? Or is it a contingent liability, the potential cost of which to shareholders changes with the market price of the company's shares, and with the true cost becoming clear only when the option is either exercised or it expires? A once-only expense or a contingent liability: these are matters over which reasonable people can agree to differ. What is important is that share options are to be expensed at all.
Economics focus
So many options
Nov 7th 2002
From The Economist print edition
Have accounting regulators chosen the best way of expensing share options?
SEVERAL features of the wild bull market of the 1990s have since been branded as evil. Perhaps none more so than the billions of dollars of share options awarded to bosses and other employees. Once praised for their incentivising power, share options are now blamed for encouraging bosses to do all manner of bad things to prop up their company's share price and so keep their lucrative options packages in the money, as the jargon has it. In this, it is now generally agreed, executives have been abetted by accounting standards that did not require the cost of awarding options to be treated as compensation and lopped off a company's reported profits.
This week, to right this wrong, the International Accounting Standards Board (IASB) unveiled proposals for expensing options, ie, deducting their cost from a company's profits. In doing so, however, not only was the IASB declaring war on some powerful political opponents of expensing. It also took sides in a lively economic dispute.
Among economists, the debate is largely about how to value options and when to expense them, not about whether they should be expensed at all. (Only a few “renegades” disagree with expensing, declares Robert Merton, who won a Nobel prize in economics for his work on options pricing.) Economists mostly agree that options should be expensed using a so-called fair-value method, one that broadly reflects what the options would cost to buy in the market, were they available.
In this, the economists agree with the IASB, which has chosen fair-value accounting of options, ruling out several other methods. These include, for instance, intrinsic value: the difference between the market price of the underlying shares that the option confers the right to buy and the exercise price at which the underlying shares may be bought. If, when an option is issued, the exercise price equals the market price (and it often does), the intrinsic value is zero—which is nonsensical. Also ruled out is the minimum-value method, which is what somebody would willingly pay for an option if they knew that the firm's share price would be fixed for the life of the option. This ignores a big part of an option's value, namely the ability to cash in should the share price rise.
The IASB does not, though, specify exactly which method of fair-value option pricing should be used. And here academics dispute vigorously. Mr Merton helped to devise the Black-Scholes model for pricing options, named after Myron Scholes (who shared Mr Merton's Nobel prize) and Fisher Black (who would have done, had he lived a little longer). The main alternative is what is called the binomial pricing model, developed by John Cox, Stephen Ross and Mark Rubinstein.
Both models were devised to price simple options traded on an exchange. So, as the IASB acknowledges, each needs to be modified to reflect several peculiarities of employee share options—such as longer lifespan or term (a typical executive option lasts ten years, against a few months to two years for an exchange-traded option), restrictions on when options may be exercised, and even the fact that they may not be sold. How these adjustments are made—and the IASB allows wide discretion—can make a big difference to the size of the expense.
 | 
The binomial method is much more complex than Black-Scholes, but if done properly, the price it produces is likely to be accurate more often
|
The binomial method is much more complex than Black-Scholes, but if done properly, the price it produces is likely to be accurate more often. Of course, the very complexity may make it easier for the price to be manipulated by a company wanting to massage its profits. On the other hand, plenty of scope for creativity exists with the Black-Scholes method, even given the few simple assumptions that are used. These assumptions, in essence, are expected volatility (how much a share price is likely to fluctuate), term, the expected dividend yield, the risk-free interest rate, and the exercise price.
Some critics reckon that these vagaries of valuation undermine the entire case for expensing. Few economists agree, however. Whatever uncertainties there are in fair-value option pricing, goes the argument, they are smaller than the uncertainty in the value of other items already routinely expensed, such as depreciation and pension-fund gains or losses.
The biggest philosophical dispute among economists concerns when options should be expensed. The IASB wants it done once and for all from the date they are awarded to employees (in other words, the grant date). Mr Scholes, the Nobel laureate, agrees. But others, such as Mr Rubinstein, one of the creators of binomial pricing, do not. He argues for full expensing at the time options are exercised, ie, when the holder trades in the options for underlying shares. Under this approach, options would still initially be expensed on the grant date; but in subsequent public filings this estimate would be adjusted to take into account changes in their value. Upon exercise, the company would take a final extraordinary gain or loss to match up with the option's actual value when exercised. Mr Rubinstein argues that, under this method, there would be less incentive to manipulate option valuations, since any divergence from an option's final true value would, in the end, result in an extraordinary charge.
The issue boils down to this: is the granting of an option a once-only expense for the company, the equivalent of paying the employee in cash? Or is it a contingent liability, the potential cost of which to shareholders changes with the market price of the company's shares, and with the true cost becoming clear only when the option is either exercised or it expires? A once-only expense or a contingent liability: these are matters over which reasonable people can agree to differ. What is important is that share options are to be expensed at all.
Posted by DeLong at November 11, 2002 09:21 AM
| Trackback
A very nice summation of the controversy, but at end I don't think it is an issue on which reasonable disinterested people can disagree in principle.
The current standard practice (disclosure, off book) says that a corp with 1,000,000 shares and zero options outstanding is booked the same as a clone (same assets, same liabilities, same business prospects, etc.) with 1,000,000 shares plus 1,000,000 options (or 1 option, or 100,000,000 options) on shares outstanding, regardless of the strike price of said options. That ain't right.
The most popularly new-standard proposals (expensing at time of grant or vesting, on formulaic prospective basis) says that a clones with the same 1,000,000 shares and 1,000,000 options are equivalent value propositions, regardless of strike price. That ain't right neither.
Given equivalent corporate profiles (including number of options outstanding), rational valuation depends on the strike price of those option. Equivalently, it depends how the market price varies in relation to a fixed strike price.
A true contingent liability exists. Depending on uncertain future events (related to market price), the corp is obligated to shed cash
(or shareholder opportunity-cost equivalents) to maintain a comparable share-value profile. This true C.L. changes continually as the underlying share price varies, and can reasonably be "marked to market" periodically ... as would be done with any on-book contract derivative on an unrelated underlier (one of our clone corp's, for instance) that happened to exist in the corporate investment portfolio.
Underappreciated in all this is that most of the divergent fine points of option valuation pertain to valuation by the grantee (employee), not the grantor (corporation ... and in effect, all other shareholders and prospective shareholders). As such it is noise, not signal, with respect to the current discussion.
""but at end I don't think it is an issue on which reasonable disinterested people can disagree in principle.""
What nonsense. Most of the approaches for expensing will vastly overstate the cost of granting options. (see Economist, etc.) I will ignore the more technical point of the actual valuation methodology, as opposed to the timing of the expense recognition, but there are many sound reasons why expensing can be misleading.
Financial statements are supposed to capture the costs and expenses of the company. Simply granting or allowing options to vest is an expense of the company's existing shareholders (their shares have a contingent dilution). The options become an expense of the company (rather than the shareholders) only to the extent that the company uses cash to purchase shares. One of the more unremarked scandals of the past 10 years or so is that many companies engaged in share buybacks ostensibly to boost the share price but in reality as a vehicle to channel the companies cash to (insider) options holders. This would be a good avenue for more accounting guidance/disclosure, but the expensing issue has been so demagogued that reasonable discussion does not take place.
As for the assertion that reasonable people "cannot disagree", this is the rhetoric of the totalitarian left. This is further twisted by adding "disinterested". Is someone like John Biggs or Warren Buffet disinterested? If so, their views should count for nothing. There are actually fund managers (buy side investors) who are against expensing because of its distorting effect on financial statements. Many people, as shareholders, are perfectly capable of reading corporate financials and telling if they are being diluted or if management is engaging in preferential behavior. This should not be an ideological issue but apparently has become one.
By first-post resort to invocation of "the rhetoric of the totalitarian left", I believe a bright line is drawn between Josef and "reasonable people".
May I conclude the exchange of views with Josef by simply remarking "Hitler! Hitler! Hitler!"?
Now if any more reasonable discussants care to take up similar points, I'd be delighted. For instance, is the satisfaction of options using shares purchased explicitly for that purpose (with consequent departure of cash) different in essence from satisfaction of options via distribution of secondary issues or treasury shares (with consequent opportunity cost of cash that would have otherwise entered corporate coffers when the same shares went directly to market)?
FWIW, I'm perfectly willing to distinguish between "things reasonable people can disagree on" and "things reasonable people can disagree on after they've given them sufficient thought". The "accounting for options" problem may include "things" in both categories, but the "agree-to-disagree's" are things for footnotes, not things material in the current controversy. (We've already decided to expense options, we're simply haggling over the modalities.)
Or perhaps I've missed something. Try me. (Anybody but Josef, that is.) After all, I'm a reasonable person.
http://www.nytimes.com/2002/11/10/business/yourmoney/10OPTI.html
Joseph R. Blasi and Douglas L. Kruse, professors of human resource management at Rutgers, examined stock option grants and shareholder returns at the 1,500 largest American companies from 1992 to 2001. They found that companies dispensing significantly larger-than-average option grants to their top five executives produced decidedly lower total returns to shareholders over the period than those dispensing far fewer options.
As for the notion that options are primarily a rank-and-file perquisite — and that abandoning them would hurt lower-level employees — the study instead confirmed what many investors have suspected: in recent years, most options have gone to top executives. And that has been true at hundreds of companies....
The objection that one doesn't know how to evaluate option grants is extremely flimsy. Investment banks have long-dated options (> 10 years) on their books, and they mark them. Which is to say, not they necessarily mark them correctly, but if you think evaluating long-dated options is intrinsically impossible, you should be running to the hills already, because you should be worried about the safety of the banking system.
One tool the banks use is a privately-run company called Totem. Every month on all the major indices, usually up to five and sometimes up to ten years, investment banks send their prices to Totem, who makes averages and sends them back. The banks then use these prices to mark or to help mark their books.
Totem (honnestly, I don't work with them) would be more than happy to help evaluate company stock options. Companies with stock options would need to pay Totem a monthly fee, and Totem would use this money to pay (say) three investment banks to price the value of the option. Totem would then furnish the average to the company.
OK no one likes my brilliant ideas...
Right
Option grants can indeed be valued. As Warren Buffett always notes - "Berkshire Hathaway will gladly make an offer for the grants that supposedly can not be valued."
Option grants that are not valued by a company make it almost impossible for an investor to properly price a company. Try figuring out what Cisco Systems is really worth, but do not fail to account for the options granted.
Options should be valued by the issuing company, and pension accounting should be made transparent so that investors can easily note the effect of assumed pension returns on earnings.
Capitalism requires transparency for all unless it is merely capitalism for the wealthiest.
The real cost of options, IMO, is the dilution they will cause when exercised. I believe the two best alternatives are to revalue options each year (note that this smooths earnings) or calculate eps assuming all options are exercised (whether or not they are in the money).
Ignoring an expected expense does not give a true picture of a company's earnings. Neither does booking an expense when an option is issued but not revising that expense if an option expires unexercised.
The real cost of options, IMO, is the dilution they will cause when exercised. I believe the two best alternatives are to revalue options each year (note that this smooths earnings) or calculate eps assuming all options are exercised (whether or not they are in the money).
Ignoring an expected expense does not give a true picture of a company's earnings. Neither does booking an expense when an option is issued but not revising that expense if an option expires unexercised.
Calling for expensing of options is all well and good. But everyone should realize that it's no silver bullet. I'm guessing that a lot of the pressure to expense options of late is a reaction to the bear market. Perhaps this is obvious, but it's worth reminding everyone that while expensing options can help investors get a clearer view of company's finances, it won't stop crooks from pulling an Enron.
International Accounting Standards will likely require the expensing of options shortly, however such a change will not cover American companies.
If workers and shareholders are to be treated fairly, I do think we must expense options.
One other thought: the relation between GAAP earnings (even assuming no fraud) and economically meaningful earnings is tenuous, at best. There is too much subjectivity and judgment in GAAP for the numbers to be as useful as we'd like.
Standard and Poors has begun to post "core earnings" for companies which may be a fine help in more properly valuing companies. The point is for as much transparency on earnings as possible. There will always be some earnings data that is difficult to interpret, but we must be increasingly clear for the sake of workers and shareholders.
In my experience, there are always Wall Street and consulting firms that offer to price options and their track record is dubious.
Equity options are more complicated than most other derivatives and, in reality, there are relatively far fewer equity options in banks than other derivatives. Look at the OCC website (occ.treas.gov) for derivatives data and it is pretty clear that derivatives with a maturity of over five years are a tiny fraction of outstandings, so no, banks are not sitting on a lot of long-dates options! The notional amount of equity derivatives with maturity in excess 5 years is $20B out of total outstandings of $40TRILLION. Plus, interest rate and FX derivatives can be priced off implied forward rates and other forward prices, while the equity market (options) is short-term and illiquid.
The point of my first post was not necessarily to say that options cannot be priced (methods do exist, though slapping on inaccurate expense numbers will not clear up financials much), but more fundamentally that the COMPANY expense is when the company has to use its resources to make good on options. Prior to that point, it is entirely a liability to shareholders.
Finally, this rhetoric about capital markets being for the "rich" and that people must see this only one way are quite bizarre. I am as much for disclosure and transparency as anyone else, I just don't want to see ignorant, mob-like behavior impose ever more distorted accounting.
To quote a quite reasonable post above: "I'm guessing that a lot of the pressure to expense options of late is a reaction to the bear market. Perhaps this is obvious"
The problematic issues of valuing 10-year options are mitigated if we mark them to market/model periodically ... but maximized if we value them at grant time only.
The valuation problems of uncertain vesting, and of future dividend policy, and of uncertain future volatility, and of non-constant risk free rates of return are likewise mitigated. (In this connection, I'd book them initially at time of vesting, not at time of grant.)
Lack of an efficient market in 10-year restricted options is an inessential problem. So long as there is an efficient market in the underlying stock (giving best-estimate information on price and volatility of the underlier), we can crank economically reasonable options valuations -- again, from a grantor's perspective.
There is a genuine valuation problem in the case of a firm with no history and no market ... and this is where the favorable incentive effect of share options is most in play. Here, knowing neither the volatility nor the clearing price of the underlier, we would be justified in adopting ultra-conservative valuations (but re-marking these valuations at IPO time).
Richard correctly observes periodic marking would tend to smooth earnings. I believe it would also have a favorable demulcent effect.
In my experience, there are always Wall Street and consulting firms that offer to price options and their track record is dubious.
Consulting firms don't know diddly squat. If by "Wall Street" firms you mean the investment banks which actually trade the stuff, you did not give them proper incentive to price the stuff correctly.
Equity options are more complicated than most other derivatives
Plain vanilla equity options are not complicated. Exotic interest-rate derivatives are as complicated as they come.
and, in reality, there are relatively far fewer equity options in banks than other derivatives. Look at the OCC website (occ.treas.gov) for derivatives data and it is pretty clear that derivatives with a maturity of over five years are a tiny fraction of outstandings, so no, banks are not sitting on a lot of long-dates options! The notional amount of equity derivatives with maturity in excess 5 years is $20B out of total outstandings of $40TRILLION.
(1) I presume you know the difference between notional and premium.
(2) This data excludes bookings in foreign entities or foreign banks booking in foreign entities (or some combination thereof). I don't want to hurt your American pride, but long-dated options are pretty big in Europe, because retail investors have shown an appetite for capital-guarantee products (you get the performance of an index with initial capital outlay guaranteed).
(3) If you suppose that the top 4 American banks have the lion's share of this, there's $4-$5 billion notional over five years held by each bank, i.e. premium of anywhere between (say) $1.5 - $4 billion. That's still big enough where the numbers had better be reasonably coherent.
Plus, interest rate and FX derivatives can be priced off implied forward rates and other forward prices, while the equity market (options) is short-term and illiquid.
(1) Dax equity options (no divs and no repo) can be priced off the forward.
(2) The equity market is not short-term and illiquid.
""while the equity market (options) is short-term and illiquid."" My original post
Options on equities trade in a short-term and illiquid market. Index equity options are irrelevant to the subject at hand, which is pricing company specific restricted stock options.
Plain vanilla equity options are not hard to price: restricted options with maturities of ten years where the company has only been listed a couple years and where the the company's float is perhaps only 5% of it's total share base (e.g. your typical high-tech start up) are extremely difficult to price.
I know the difference between notional and premium. While the big banks in US do hold the lion's share of derivatives, they are traders (not end users)and try to hold a matched book (the notionals are mostly offsetting) I tossed out notionals as an illustrative example. The premium figures you tossed out are preposterous for a matched book.
Finally, never assume that I-Banks's strength lie in their pricing accuracy. If you have a good enough sales force you can make money in many circumstances.
Thought experiments that may shed light ...
Experiment #1: Suppose CrapCo, Inc. grants options to its valued employees, underwriting these options in the usual way (intending to satisfy them with treasury shares, secondary issues or open-market purchases at time of exercise).
Suppose CrudCo, Inc. (CrapCo's mirror image across the street) grants similar options to its valued employees, but arranges to satisfy them by purchasing equivalent options from an arms-length third party, MegaFinCo, and conveying these options to grantees.
Who is thereby relatively advantaged or disadvantaged, how, when and under what contingencies ... and how ought these similarities and diffences be accounted in the books of both firms, now and later?
Experiment #2: Suppose CrapCo -- for whatever reason -- vests its employees with options in CrudCo, and vice versa. How ought the effect of these transactions be valued and/or accounted, relative to issuance of similar options on one's own stock?
What many seem to be missing is that options are awarded in very large degree to the top executives of companies, in very small proportion to other workers. Also, options tend to distort management practices adversely for shareholders. Also, options that are not valued make proper valuation of a company most difficult giving an edge only to professional investors who can afford the time and have the expertise to get proper valuations.
Excessive use of options and hidden options costs have made the American stock market a significant problem for middle class investors.
We need stock and bond markets that are investor friendly for middle class sharholders. Other than through index funds or a bond fund at Vanguard or TIAA-CREF, I find investing badly pitted against middle class needs. This is a serious problem.
There were huge losses of retirement income during these last 34 months. These losses will not be made up for many months to come. Why then is there not more concern about why the loses have been so deep and long lasting? Much of the problem stemmed from American companies ignoring the best interests of workers and sharholders. To allow this point to be lost is to insure the continuation of an overly dangerous investment and work environment. Accounting reform and transparency are needed not to limit American companies, but to make them far more efficient for us all.
Options accounting: I actually support booking the options as a liability and expensing them as they are exercised. To engage in projection of their value will have a distorting effect on Net Income, removing it ever farther from the actual cash a company produces (or loses) and farther into the netherworld of estimates. And, yes, I know, the estimates would now be located on the Balance Sheet, but that's where they belong. The changes in value could be offset in the Unrealized Gains & Losses in Shareholder's Equity. In sum, I'd say that options represent neither a traditional cost nor a traditional expense and they shouldn't be treated like the others.
The poster from 12:44pm is overstating the impact of options on corporate earnings. The bubble -- an investment mania -- caused the vast losses for investors with options playing only an ancillary role. Middle-class investors have been told to invest through an index or managed mutual fund for decades now. You really have to leave it to the pros. Meanwhile, the conflict about corporate financial reporting shouldn't be pictured as the "mighty" vs. the "powerless." Institutional investors represent a powerful, sophisticated interest group. There are some problems with corporate governance that are causing sr execs at public corporations to have more power than they really should have but the system needs to be approached carefully. Overregulation poses even more risks than letting the status quo continue.
Thought experiment #3: CrudCo announces that it will no longer pay vendors, lendors or employees in cash or cash equivalents. Creditors will be compensated entirely in out-of-the-money options on shares, now and forever.
CrudCo further announces "yes, we have no expenses ..... we have no expenses toda-a-ay!", and further asserts "CrudCo has no unrealized losses on the balance sheet either".
Ron: Good question. So, as it stands now, CrudCo hasn't actually spent any of the investor's capital. Has it incurred any losses? Well, not until the options are actually vested and in-the-money. If they were to be exercised, then the expense would be run through the P&L. I think your example proves why options shouldn't be treated like other expenses. Interested in your response.
Thought Experiment #3 (extended) -- Capital? What capital? CrudCo's books are spotless. All fixed assets -- the crud plant, the crud-mills and dreck-separators on the plant floor, down to the last crud-packer's shovel -- were acquired the same way as all the other supplies and labor. They're carried on book at their accounting cost basis (zero), less depreciation (zero). Crud-in-process inventory is likewise carried at cost (zero), and as outgoing loads of crud change hands, the resulting cost-of-goods-sold is zero. Receivables are the only non-cash entry, and as these are collected, the cash just keeps piling up.
Any distribution of cash as dividends would unfairly impair the fair market valuation of CrudCo's outstanding options ... though (as an original CrudCo shareholder) I must confess it's tempting. In fact, it's awfully tempting to take the whole thing private, leaving option-holders in the lurch. Hmmmmm ...
Sorry, RonK, could you spell out the point you're making in the last post?
Thought Experiment #4 -- Mirror-image enterprises CrudCo and CrapCo adopt similar strategies, with one difference.
CrudCo, as above, pays for all necessities with out-of-the-money options on CrudCo shares.
CrapCo pays for all necessities with in-the-money CrapCo shares. (For convenience, let's set the strike price at $0.)
How does total market capitalization of CrapCo compare to CrudCo's market cap?
Thought Experiment #5 -- Mirror-image firms CrudCo and CrapCo maintain defined-benefit pension plans. Current annual financial statements for both firms show a nominal 10% return on pension assets, however -- due to a burst crap market bubble -- actual returns were negative 10% on pension assets. Consequently there is an unrecognized (in the accounting sense) $1B deficiency in each fund.
CrudCo proposes to remedy its pension deficiency by contributing an additional $1B in cash to its pension fund.
CrapCo proposes to remedy its pension deficiency by donating $1B in options on CrapCo shares (specifically the same special share options -- strike price $0 -- used in Experiment #4).
How are the comparative total market capitalizations of the two firms affected?
