The Accounting Onion

Stock Compensation, Tax Law, Financial Reporting and Facebook's IPO

Last August, I wrote a post in which I expressed my dismay that Carl Levin (D-Michigan) and Sherrod Brown (D-Ohio) are trying to make tax law based on the premise that FASB rules for stock options are caviar -- when the stark reality is that they are chopped liver (or worse). The senators are oblivious to the fact that the evils they decry lie in the financial reporting rules (hewn from decades of political wrangling, grandstanding and backroom dealing), and not the income tax rules; and worse, their proposal to reduce the deductions available to corporations while keeping the same taxable income for the individual who would exercise options has no precedent in tax policy, logic or fairness.

This current post is in response to Senator Levin's recent speech on the Senate floor in which he pressed the same dubious points. But, the really interesting part are the circumstances he relates regarding Facebook's upcoming initial public offering of its shares:

"According to its filings, when Facebook goes public, Mr. [Mark] Zuckerberg plans to exercise options to purchase 120 million shares of stock for 6 cents a share. Mr. Zuckerberg's shares, obviously, are going to be worth a great deal more than 6 cents, a total of about $7 million; they will apparently be worth more than 600 times as much, something in the neighborhood of $5 billion.

Here's where the tax loophole comes in. Under current law, Facebook can – perfectly legally – tell investors, the public, and regulators that the stock options he received cost the company a mere 6 cents a share – that's the expense shown on the company's books. [This is wrong – see Footnote.] But the company can also – perfectly legally – later file a tax return claiming that those same options cost the company something close to what the shares actually sell for later on – perhaps $40 a share. And the company can take a tax deduction for that far large [sic] amount. So the books show a highly profitable company – profitable, in part, because of the relatively small expense the company shows on its books for the stock options it grants to its employees. But when it comes time to pay taxes, to pay Uncle Sam, the loophole in the tax code allows the company to take a tax deduction for a far larger expense than they show on their books. …

Now, the end result is that a profitable U.S. corporation – a success story – could end up paying no taxes at all for years, even decades." [italics supplied]

To Levin, the Facebook IPO is a dramatic illustration of an inequitable "loophole" in the tax law. As Levin and Sherrod would have it, Facebook's tax deduction for using stock options to compensate executives – as opposed to any other form of compensation – would be essentially zero (that's probably a little dramatic on my part, but the point is the same); yet, Zuckerberg's tax liability when he exercises the options could be somewhere in the area of $3 billion.

When the Levin and Sherrod approach to taxation is compared to other forms of compensation to executives, rank and file employees, vendors and lenders, it makes no sense whatsoever. If the senators want to make stock options more costly to corporations than other forms of compensation, then so be it – but they should just come out and say so, and explain why it's a good idea.

But, most of what I have just written reprises my earlier post (and if you want even more background on the depressing politicis of the accounting question, see Jon Weil's article, here.) I'm covering this topic anew because Levin's Facebook IPO story raises some other interesting questions with implications for financial reporting.

My first question is this: If U.S. GAAP for stock compensation expense were based on the value realized upon exercise, just like the tax rules, how would that have affected Facebook's financial statements?

The strong implication of Levin's narrative is that the extra amount of expenses would have wiped out every dollar of Facebook's reported net income that it had ever 'earned.' On top of that, there could be other outstanding options held by Zuckerberg and other employees extending way down the organization, which are going to have the same effect on future reported net income.

Which brings me to my second question: For all practical purposes, could Zuckerberg be taking Facebook public at this point in time with no history of profitability, perhaps negative shareholders' equity, and perhaps no hopes for earnings for some for years to come?

My inclination is to say "no," but I can only speculate. So, I all answer that question with another question: Since the FASB's rules understate the economic cost of options granted to employees, did the FASB provide a perverse incentive to Facebook to grant more options (or at terms overly favorable) to employees than it should have?

If one accepts the maxim that "what gets mismeasured gets mismanaged," then the answer to that question should be "YES." The stock option problem lies not with the tax rules that eventually recognize the full cost of the options to shareholders, but with the financial reporting rules that allowed Facebook to grant options without recording their full cost.

Surely, the senators can see that different accounting rules would have wrought different compensation policies from Facebook. Senator Levin would not have had a story to tell and Mark Zuckerberg would be a few billion dollars poorer.

But, methinks that Facebook would still be Facebook.

* * * * * *

Footnote: It is my distinct pleasure to provide Senator Levin (and his staff) with yet another accounting lesson. The amount of expense to be reported by Facebook is not, as the senator claims, the exercise price of the options (i.e., 6 cents per share). Assuming that the options were issued without any intrinsic value, then their "grant-date present value" (i.e., the amount upon which periodic option expense is measured) could end up being more or less than 6 cents per share. Although this technical correction to Senator Levin's story it doesn't fundamentally change the message, it once again reveals a lack of understanding that makes one question if Senator Levin has an adequate grasp on the issues.

 

 

A Much Needed Accounting Lesson for Two Senators

Nothing good comes from bad accounting, and many of the bad things that happen are easy to predict: excess bonuses to management, negative NPV investments, underfunding of pension plans, etcetera, etcetera, etcetera. Yet, the instance I'm writing about in this post has a special flavor.

Savvy managers may know that the accounting rules they are exploiting fall far short of good, but others, in stark contrast to their status in society and evident sophistication in other matters, take the solemn pronouncements of the FASB for gospel. Recently, two U.S. senators have apparently forgotten or are oblivious to the fact that the making of financial reporting standards is more like the production of sausage than the engraving of the Ten Commandments by Moses on Mt. Sinai. Consequently, they have introduced a bill to model the tax deduction for stock-based compensation after the low-grade pork in a tube known to the accounting cognoscenti as Codification Topic 718 (formerly SFAS 123R).

The senators in question are Carl Levin (D-Michigan) and Sherrod Brown (D-Ohio). A couple of weeks ago, they introduced a bill to eliminate the so-called 'excess tax benefit' from issuing stock options to employees. The two senators note that the amount of expense reported for financial reporting purposes (for example, in SEC filings) is quite often less than the amount of the tax deduction taken per the tax return, and they seem to think that the so-called 'excess tax deduction' amounts to a government giveaway to corporations. If financial reporting measures stock-based compensation 'right' (and it surely does not), then by their logic it's wrong to receive a deduction that's greater than the expense reported for financial reporting purposes.

Bless their little hearts, but Levin and Brown (and surely other senators) are oblivious to the reality that the accounting evils lie in the financial reporting rules, and not the income tax rules. It is the financial reporting rules that are fraught with inconsistencies, for they were designed to systematically understate the cost of compensating employees.

The Economic Measure of Stock Compensation Cost

I'll be the first to admit that, when it comes to taxes, I'm sort of a babe in the woods myself. But, the complexity that is misleading the senators does not lie in the tax rules. The tax rules in this respect are simple, and a review of the fundamentals ought to be sufficient to set the stage for my accounting lesson to them.

There are two kinds of deductible expenses for tax purposes. The first type is the cost of productive resources that are consumed; examples include depreciation of productive assets, rent of productive assets, costs of inventory acquired and subsequently sold, and salaries of employees. The second type, financial cost, was at one time more controversial; and the most straightforward example is interest cost of debt.

Many decades ago, interest costs were seen to be fundamentally different than the costs of using productive resources. They were merely distributions to stakeholders, i.e., more analogous to dividends than costs of production. Eventually, though, in response to political pressure from railroads and other large corporations, taxing authorities adopted a shareholder perspective to tax deductions. In regards to interest costs, they should be deductible since they reduce the maximum amount that could be distributed to shareholders. In other words, from the perspective of the shareholder, expenditures on productive resources and the interest are both costs that reduce the amount of funds available to be distributed to shareholders.

Moreover, the tax law straightforwardly provides for the amount of interest that may ultimately be deducted as the difference between the total amount borrowed and the amount ultimately paid back to the lender. There is nothing magical about this, and it is consistent with the measurement of other tax deductible costs, including the stock-based compensation rules that are being challenged by Senators Levin and Brown.

So, one problem with their bill is that if you believe that the tax deduction for stock-based compensation should be consistent with interest costs, then you should agree that the ultimate cost of stock compensation should be the intrinsic value realized by the employee when the stock options held are recognized. And, in principle, the answer should not change whether the options are net settled, physically settled, or in the form of stock appreciation rights.

The Free Accounting Lesson

The foregoing is so mundane, I feel almost ashamed to have written it. But, I needed to do that to tee up the really interesting part: somehow, these two senators (and probably a few more) are either being grossly disingenuous or are laboring under the mistaken belief that U.S. GAAP on stock-based compensation (Topic 718) must be 'right,' since it was issued by the FASB after a decade of "due process." In point of fact, it is an embarrassing political compromise that was allowed to survive because its most outstanding feature is that it generally reports only a small portion of the full cost to shareholders of such programs (as I explained in the previous section).

Surely, the senators are aware that it was only relatively recently that U.S.GAAP began to require any recognition of compensation expense for stock options. The controversy was intense, and the best the FASB could do against stiff political resistance when they promulgated FAS 123R was to require that stock compensation expense be limited to the value of the options on the date they were granted – as opposed to either the date on which they were earned ('vested') – or most accurately and consistent with the measurement of other expenses, the date the options were exercised.

For example, assume that an employee is granted an "at-the-money" option to purchase one share of her employer's common stock for $8 per share, and she eventually exercises the option when the share price is $20 per share.

The option was granted on January 1, 20x1, at which time it was valued at $2 per share. It vests (i.e., becomes exercisable) four years later, and expires three years after that. The fact that a company's earnings reported to its shareholders is reduced by only $2 (and spread out over the four-year vesting period) while the amount of wealth transferred from shareholder to employee is $12 (i.e., six times greater than the $2 recognized for financial reporting) is one of the most glaring inconsistencies in gazillions of pages that comprise U.S. GAAP (or IFRS, for that matter).

* * * * * *

Bless these innocent children, who happen to be U.S. Senators, and keep them forever. Will they ever come to realize that the so-called 'excess tax deduction' (a financial reporting concept, not a taxation concept), which they have decried with indignation and sanctimony, is a blatant misnomer? A more accurate moniker would be 'insufficient financial reporting deduction', for the failure of GAAP to record the full amount of wealth ultimately transferred from current shareholders to employees.

Dear senators, GAAP is not gospel, and much of the time it is not even good. And, don't just take it from me. As one retired FASB member, David Mosso, now ruefully admits, GAAP (or IFRS – take your pick) is a "garbage truck."

SEC's “Simplified Method” for Estimating Stock Option Cost is Ripe for Manipulation

---

Would you like your staff to learn more about
SEC reporting? 
Click here for a sample in-house training agenda!

---

Call me cynical, but I am wary of accommodations that permit companies to freely choose between their best estimates and something less.  Even significant restrictions on the ability to elect a simplified approach may not be effective in preventing manipulation. Witness, for example, the  abuse of the short-cut method for demonstrating hedge effectiveness under FAS 133 by Fannie Mae et. al.  Accommodation may result in relief that is elected in good faith by many, but for some, it's just another opportunity to manage financial reporting.  The latest example is SEC Staff Accounting Bulletin 110, and the related earlier SAB 107.

Background

FAS 123R requires a company to estimate the grant-date fair value of its options, which in turn necessitates estimation of the expected exercise date of the options.  For companies that did not have access to adequate historical data about employee exercise behavior in order to make a realiable estimate of when options will be exercised, the SEC issued SAB 107 in March 2005, permitting use of a simplified approach until December 31, 2007.  The simplified approach was further restricted to options that meet criteria collectively referred to as “plain vanilla,” and essentially allowed companies to assume that the exercise date of an option would be halfway between its vesting date and the expiration date.

SAB 107 made the the simplified approach available only through December 31, 2007 – at which time the SEC staff expected companies to have sufficient historical data with which to more precisely estimate the expected exercise date.  As the Dec. 31, 2007 deadline in SAB 107 approached, the SEC staff was persuaded the detailed information about exercise behavior was still not readily available for many companies. Accordingly, SAB 110 was issued on December 12, 2007 to allow the continued use of the simplified method—provided a company concludes that its own historical share option exercise experience does not provide a reasonable basis for estimating expected term.

The Earnings Management Tactic

By the way, I am not a big fan of FAS 123R, because grant-date valuation often understates the actual costs of stock options.  It would make more sense to value the options on the date they are effectively 'issued' to employees, which is the vesting date.  But given that grant date valuation is the law of the land, I suppose that it is theoretically correct to base the valuation on the expected exercise date as opposed to the expiration date.   

Having said that, valuation based on assumed exercise at the expiration date -- as opposed to estimating an earlier exercise date -- would yield higher measures of compensation cost; and, especially given the concerns that the SEC has expressed in SABs 107 and 110, result in greater consistency of measurement.  It would also remove one very significant element that management could manipulate -- which is a big reason why companies lobbied for using the expected exercise date in the first place.  The simplified approach temporarily permitted by SAB 107 and extended by SAB 110 sweetens the deal.

Here's how the earnings management game works. Assume that (1) management of a company desires to minimize reported compensation cost, and (2) the expected exercise date of options is on the later side of the halfway point between vesting and expiration.  In this case, one would expect management to try their best to convince their auditors that they are unable to reliably estimate the exercise date -- that's because the simplified approach would produce lower compensation cost.  But, if the expected exercise date were on the earlier side of the midpoint, we would see just the opposite kind of self-interested rationalization: management trying to convince their auditor that their SWAG is practically as good as carved in stone.

Finally, why is the SEC making more work for itself?  Unlike SAB 107, there is no specific expiration date of the simplified approach in SAB 110.  The SAB 110 release states that once relevant detailed external information about exercise behavior becomes widely available for companies to make more refined estimates of expected term, the staff will no longer accept use of the simplified method.  In the interim, I suppose the SEC will be selectively asking companies to justify why they continue to use the simplified approach.  That's not a good use of anyone's time. 

If a company cannot make a reliable estimate of the exercise date, the SEC should require them to assume exercise at expiration.  That should get their attention -- and protect investors by cutting the legs out from under an obvious earnings management opportunity.

---

Would you like your staff to learn more about
SEC reporting? 
Click here for a sample in-house training agenda!

---

Are Stock Options an Efficient Form of Executive Compensation? - No!

Have you noticed how a large number of companies curtailed or eliminated their stock option grant programs when FAS 123R became effective?  You frequently hear proponents of stock option programs spout the line that stock options, like nothing else, provides incentives for management to create value for shareholders.  Well (can you hear Ronald Reagan?), that's a myth, and I'll prove it with a simple example:

The CEO of a company has one asset at the beginning of the year, $1,000,000 cash, and his only job is to decide how to invest the money.  The CEO's only compensation is in the form of an option to acquire 20% of the company at the end of the year.  The option is granted at the beginning of the year, and has an exercise price of $200,000.  (Note: since 20% of the company is worth $200,000 at the beginning of the year, we see that the option is granted at the money.)

The CEO has found only two projects to choose from, as follows:

• Project A will pay $1,050,000 for certain at the end of the year.  Therefore, the executive's option to purchase 20% of the company would be worth $10,000 (= 20% x $1,050,000 - $200,000).
• Project B has a 10% chance of paying $2,000,000 at the end of the year, and a 90% chance of paying $500,000.  Therefore, its year-end value is $650,000, and the executive's option would be worth nothing 90% of the time, but $400,000 (=20% x $2,000,000 - $200,000) 10% of the time.

Which project do you think the CEO will choose to invest in?  For sure, a lot of CEOs would be tempted to choose the value destroying Project B.  It also stands to reason, that the higher a CEO's cash salary (zero in this case, but rarely observed in reality), the more attractive a CEO would find Project B.

The point of the example is to demonstrate that the risk characteristics of options do not align with the risks that accompany holding shares of stock.  That's why so many companies have replaced options with shares, post the effectiveness of FAS 123R.  If the CEO's salary in my example was, say, shares of stock for 5% of the company, we could safely say that the CEO would always choose Project A--the choice in line with shareholders' interests.

OK.  I've now demonstrated that stock options as a form of executive compensation are inefficient.  Imagine the huge transfers of wealth from shareholders to managers in the supposed name of shareholder wealth creation, when just the opposite was happening. 

The analysis I have provided you is not groundbreaking, so it must be asked what has driven the propensity to award options?  For sure, the fantasy accounting under APB 25, which allowed companies to portray options as if they were costless to shareholders, was a huge factor.  Anyone (the APB, FASB, SEC and Congress--to name the principle classes of culprits) who had any involvement in establishing or perpetuating such a ridiculous accounting rule should feel ashamed for their complicity in a vast scheme to deceive the investing public and impair the competitiveness of the U.S. economy.

Before I conclude this post, I need to issue a caveat.  I am not saying that stock options are inefficient under all circumstances.  Labor markets may dictate that stock options be a part of the pay package for many key employees.  I am only pointing out the inefficiences created by giving stock options to executives who allocate corporate resources.  For other key workers, it is often the case that stock options are an essential means of retaining their employment. 

The Four Sins of Stock Option Backdating

I have a weak spot in my heart for my former students--especially if they were also my tennis or cycling buddies.  Chris, of the cycling variety, recently sent me this email message:

...So, today I was reading some posts on the Greg Reyes trial/verdict (he was the CEO of Brocade ...  I must confess that I've heard all the fuss, but don't really understand it (and haven't followed it very closely) ... was wondering what you make of it all.

Chris, I'm not going to remark about the specifics of the Brocade case, and would rather give you a broad outline of the general issues underlying illegitimate stock option backdating.  The following simple example is representative of the types of transactions that have been questioned:

The board of directors of BD Company met on December 1, 20x1, on which date the market price of BD was $50 per share.  The board authorized the granting of one at-the-money option to BD's CEO, to vest over two years. The CEO instructed the vice president of human resources to backdate the option to November 1, on which date the market price of BD was $40 per share.

The stock option backdating scandal involved four kinds of ripoffs, which I'll describe more fully below:

1. Shareholders ripoffs--By selecting a date in the past on which the stock price is lower than the current stock price, options are deceptively granted in the money.  In my simple example, the effect is to transfer more value to executives than the board of directors intended.  In some cases, the board may actually participate in the backdating scheme.
    
2. Accounting fraud--Options that were effectively granted in the money (because the real grant date is not the backdated one--it's the date the board authorized the award) are accounted for as if they were granted at the money.  Under APB 25 (recently superseded, but in effect during the halcyon days of scandal), expense is measured at the intrinsic value of options at their grant date. (I'm oversimplifying here, but am being specific enough to make the point.) In my simple example, recording zero expense overstates net income by $10, which is the intrinsic value of the option. Often, the accounting fraud has not had a material impact on earnings, except that it was an essential to misstate earnings in order to cover-up the shareholder ripoff (See #1, above).
   


3. Tax fraud--In addition to backdating grant dates, there is strong evidence of executives backdating their exercise of options to a date on which the stock price was low. In so doing, they could reduce the amount of payroll taxes they owed, and reduce the amount of profits they realized on exercise from being taxed at ordinary rates.  If they could then manage to hold the stock for at least a year, they would eventually pay taxes on the rest of their profits at the lower capital gains rate. 


4. Inadequate SEC regulations created a loophole the size of Arthur Andersen--Under SEC rules (promulgated under Section 16a of the Securities and Exchange Act of 1934), executives have to report (i.e., publicly announce) their personal acquisitions of financial instruments issued by their company, and also derivatives underlying those financial instruments.  Until Congress forced them to tighten the loophole (one of the many good things accomplished by SOX), executives had as much as 40 days to report the transactions.  In effect, that gave executives a window of more than a month from which to cherry-pick the lowest share price for option backdating purposes.  A funny thing happened when the report due date was shortened to two days: backdating disappeared. 

That last point, above, doesn't get a lot of press, but it rankles me the most.  I had criticized the SEC for years about the 40-day window--not because I was aware of the backdating frauds, but because I didn't see why investors had to wait so long to receive important information about executives trading in their own shares.  Of all the wonderful things Arthur Levitt tried to accomplish as SEC chair, I don't know why this one wasn't on his radar screen.  After all, it was during his watch that the SEC first floated the idea of accelerating the due dates of annual reports.  (Dear Mr. Levitt, if you perchance should read this, I would consider it an honor to publish your response.) 

Based on the number of SEC investigations and the academic research that exposed this repulsive scandal, there must be hundreds of executives who betrayed the trust of their shareholders with the some criminal indifference that possessed Enron's top guys.  Also similar to Enron, there are likely just as many dupes and facilitating lawyers, accountants, board members, etc. as there were executives who profited directly.  For those Pollyannas amongst you who thought that Sarbanes-Oxley was an overreaction to the crimes committed by a few, I hope the stock option backdating scandal has transported you closer to my world.

So, Chris, you owe me one ... or maybe I owed you one.  It doesn't matter.