Audit procedures on the use of fair value of share-based compensation.

In addition, the PCAOB Q & A addresses an inquiry as to how an auditor evaluates the qualifications of a valuation specialist. The PCAOB Staff refers to AU sec. 336.08(a) and (b), which states that auditors should examine and consider the valuation specialist's certification, licenses, and other documented competencies [1, 5]. More specifically, the PCAOB Staff believe the auditors should ascertain the external valuation specialist's knowledge, experience, and understanding of the valuation concepts central to the determination of fair value estimates. In addition, the auditors should evaluate the specialist's knowledge of the relevant laws, regulations, and FASB standards, as well as guidance from the Emerging Issues Task Force (EITF), American Institute of CPAs (AICPA), and the Securities Exchange Commission (SEC).

Testing an Entity's Share-based Payment Database and Activity

Once an understanding of the public company's process for developing estimates of the fair value of employee share options has been documented, the auditor should perform tests of the share-based payment database and recorded activity. A schedule presenting all share-based award activity should be obtained, including:

* The number of options outstanding at the beginning and end of the period;

* Options that are exercisable at the end of the period; and

* The number of options granted, exercised, forfeited, cancelled, or expired during the period.

In addition, exercise prices, including non-vested shares, should be included on the schedule. The schedule should contain options granted, along with the fair value on the grant date and the resulting compensation expense, which is reconciled to the general ledger. The auditor should examine if any patterns or conditions exist indicating any previously identified risks regarding share-based payment. In particular, items the auditor should examine include:

* The existence of a high percentage of grants awarded in a period;

* Arrangements where share-based compensation is a major portion of executive compensation arrangements;

* High variation in grant dates;

* Patterns of significant increases in stock prices subsequent to the grant date; and

* High levels of stock-price volatility.

After a preliminary examination of the current period's schedule, the auditors should use the prior-year working papers to determine that the outstanding stock options at the beginning of the period (on the schedule) for the current period, agree with the outstanding stock options at the end of the previous period. In addition, a sample of stock option awards should be selected from the beginning of the period (each sample should represent one individual issued at a single grant date), with balances based on the total unamortized fair value of those options.

To determine if the options selected remain outstanding, the auditors should obtain appropriate evidence corroborating that the employee is still employed with the entity, and that the option has not been cancelled or forfeited. The auditors should also determine if the terms of each option have been modified during the period, and whether or not each option is exercisable, and they should trace the option to ensure it agrees with the schedule of options exercisable at the end of the period. The PCAOB Staff noted that auditors should test the completeness of the various share-based payment award activities by making inquiries of responsible persons outside of the accounting function, including the corporate secretary and members of the board's compensation committee, and they should read the board's minutes.

Comparison of Two Option Pricing Models

The PCAOB staff writes that in testing the estimated value of employee share options, auditors should [5]:

* Evaluate the consistency of the process;

* Evaluate the reasonableness of the company's fair value model, and assumptions employed in the model, including expected term and expected volatility; and

* Verify both the accuracy and completeness of data underlying the fair value measurements.

Once the awards granted are corroborated for the current period in the schedule discussed above, the fair value of each award must be tested. Auditors are expected to evaluate the reasonableness of the option-pricing model selected by the company for calculating the fair value of employee options. The FASB states in SFAS 123 (R), Paragraph A13, "a lattice model (e.g., a binomial model) and a closed-form model (e.g., the Black-Scholes formula) are among the valuation techniques that meet the criteria required by this Statement for estimating the fair value of employee share options and similar instruments" [2]. The FASB, however, does not offer any preference as to which valuation technique a company may adopt. Initially the Board recommended the lattice model, but removed the preference from the exposure draft after receiving public comment [3]. The PCAOB staff further states that auditors should evaluate if the valuation model selected [5]:

* Is applied in a manner consistent with SFAS 123(R)'s fair value measurement objective;

* Is based on established principles of financial economic theory; and

* Reflects all of the substantive characteristics of the share options granted to employees.

In order to understand the substantive differences in each of the main option pricing models, both the Black-Scholes-Merton and Lattice option pricing models will be briefly discussed, respectively.

A. The Black-Scholes-Merton Option Pricing Model

The Black-Scholes-Merton (BSMOP) model, a closed-form option pricing model, measures the relationship between "call option value and ... factors that determine the premium of an option's market value over its expiration value" [6]. SFAS 123(R), Paragraph 18, specifies six assumptions used to calculate the fair value of share based payment [2]:

1. Expected term of the option;

2. Expected volatility of the price of the underlying share for the expected term of the option;

3. Exercise price of the option;

4. Current price of the underlying share;

5. The risk-free interest rate(s) for the expected term of the option; and

6. Expected dividends of the underlying share for the expected term of the option.

These six variables represent auditable components and are used in both the BSMOP and the lattice models. The BSMOP model assumes that stock option exercises occur only at maturity, with other variables (using weighted-average estimates) such as expected dividends, expected volatility, and risk-free interest rates remaining constant over the option term [5].

The BSMOP model is simpler to apply than the lattice-based models, but it has limitations in that it prevents the ability to consider "varying assumptions" over the option term; further, since most of the inputs used in the BSMOP model remain constant, none of the input data can describe any "unique" features of the employee stock option plans [4]. The BSMOP model was originally developed for valuing exchange-traded options; therefore, it does not take into consideration any differences between traditional exchange-traded stock options and those options granted to employees, possibly causing the value of stock options to be overstated. Baril et al. argue that BSMOP model's use is more appropriate for companies that grant relatively few stock options [4]. Unlike the static assumptions used in BSMOP model, the lattice-based models accommodate multiple dynamic assumptions.

B. The Lattice-Based Model

Because employee stock options cannot be transferred and are subject to vesting requirements and strict forfeiture clauses, more often such share options are exercised prior to their maturity. The lattice-based model takes into consideration many unique assumptions that reflect conditions under which employee options are typically granted, such as early exercise of the option. The most common lattice model is the binomial model (although trinomial and multinomial formulas also exist), where the period of time from the grant date to the maturity date is divided into small increments representing intervals. The lattice model estimates how changes in prices over the term of the option would affect the employee's exercise behavior during each interval. While BSMOP is simpler and more commonly used, the lattice model (binomial) requires considerable technical expertise.

As indicated, the six assumptions discussed earlier are also used in the lattice model, with one exception: the expected term of the option is an output of the lattice model, not an input. Specifically, the lattice model calculates the effects of changes in volatility factors (risk-free interest rate, dividend rate, and estimates of expected early exercise) over the option term, requiring more data analysis in building its assumptions. The PCAOB staff stated that the lattice model "might more fully reflect the substantive characteristics of a particular employee share option" [5].

The PCAOB Staff noted that auditors should be aware of circumstances in which the BSMOP model would not be appropriate, and it offers one specific example: "an exercise condition that is satisfied when the share prices exceeds a specified value for a specified period of days ... [Black-Scholes] is not designed to take into account that type of market condition" [5].