Expand
When valuing an employee option under the Black-Scholes model, companies should use the option's expected term rather than the contractual term. SAB Topic 14 reinforces the guidance in ASC 718 that the nonhedgeability and nontransferability of most employee stock options is not considered in fair value, except as it affects the expected term assumption. Additionally, pre-vesting forfeitures should not be factored into the determination of expected term because they are taken into account by the company recognizing compensation cost only for those awards for which employees render the requisite service. As described in SC 9.3.10, certain other factors may be considered when a company develops its expected term assumption.
Companies should consider the following factors in developing an expected term assumption for use in the Black-Scholes model or in developing the group of assumptions related to the expected exercise patterns in a lattice or Monte Carlo model:
  • Vesting period of an award
  • Contractual term of an award
  • Historical exercise and post-vesting cancellation experience with similar company-specific grants (i.e., historical average holding periods)
  • Stock price history
  • Expected volatility (which may be inversely correlated with the expected term)
  • Blackout periods that may trigger automatic early exercise or delay exercise
  • Plan provisions that require exercise or cancellation of options shortly after employees terminate
  • The extent to which currently available information indicates that the future is reasonably expected to be similar or different from the past

Because employees typically cannot exercise an option until it vests, the vesting date represents the earliest end of the range of possible exercise dates, whereas the contractual term represents the latest end of the possible range. An analysis of historical exercise and post-vesting cancellation behavior is generally used to estimate where within this range the exercise or post-vesting cancellation may occur. If the award has an acceleration feature (e.g., immediate vesting upon a change in control of an IPO of a certain size), the vesting date used to determine the expected term should incorporate the probability that an award's vesting will be accelerated. A company should use its relevant historical experience for similar options and employee groups. If a company's specific historical data is insufficient, ASC 718-10-55-32 and SAB Topic 14 allow the company to use other publicly available data, such as financial statements of similar companies or published academic research. For example, if a company has a history of option grants and exercises only during periods in which the company's stock price was rising sharply, the exercise behavior related to those options likely would not be a sufficient basis to develop the expected term assumption because it would be unreasonable to assume that the stock price will continue to rise in a similar manner. In a case like this, the company might instead rely on academic studies or disclosures from similar companies with similar grants to similar employee groups, or the company might elect to use the simplified method (as discussed in SC 9.3.1).
When a company uses published academic research or industry data to estimate employees' exercise behavior, it should consider how the awards and companies that sourced the data compare to its own awards, including the following attributes:
  • Vesting periods
  • Contractual terms
  • Blackout periods
  • Stock-price volatility
  • Demographics of employee populations (which may affect employees’ attitudes toward risk and patterns of exercise)
  • Any other company-specific attributes that can affect employee exercise behavior

It may be difficult in some cases to identify similar companies that grant similar types of awards to similar populations of employees, but the objective is to ensure that the most relevant data available is used to inform management's judgments.

9.3.1 Simplified method for estimating expected term

SAB Topic 14 provides a simplified method for estimating the expected term for “plain vanilla” options that significantly reduces the analysis required to estimate expected term. This simplified method is only acceptable if (1) a company does not have sufficient appropriate exercise data on which to base its own estimate or (2) exercise data relating to employees of comparable companies is not easily obtainable. SAB Topic 14 also stipulates that the simplified method should no longer be used once a company has sufficient exercise data in which to base its own estimate or more relevant general information (e.g., published academic or industry-sponsored research) becomes available on employee exercise patterns.
Question 6 in SAB Topic 14-D.2 provides the criteria necessary for application of the simplified method, as follows:
  • Stock options are granted “at the money”
  • Exercisability depends only on completing a service condition (i.e., continuing to work through the vesting date)
  • Employees who terminate their service prior to vesting forfeit their options
  • Employees who terminate their service after vesting have only a limited time (typically 30-90 days) to exercise their stock options
  • Stock options are nontransferable and nonhedgeable

If a company grants awards that do not meet the SAB Topic 14 criteria, the simplified method cannot be used, and historical exercise data is required to be the starting point to develop the expected term assumption. See SC 6.2.3 for guidance regarding the use of the simplified method by nonpublic companies in ASC 718-10-30-20A through ASC 718-10-30-20B. Scenarios where use of the simplified method under SAB Topic 14 may be appropriate include:
  • Insufficient historical experience for option grants overall
  • Substantial changes in the contractual terms or vesting periods of options granted
  • Changes in a company’s business or employee population, rendering historical experience irrelevant to expectations for current grants

In addition, SAB Topic 14 specifically states that the simplified method is not intended to be applied as a benchmark in evaluating the reasonableness of more refined estimates of expected term.
The simplified method uses the mid-point between the vesting period and the contractual term for each grant (or for each vesting-tranche for awards with graded vesting) as the expected term. For awards with graded vesting, the time from grant until the mid-points for each of the vesting tranches may be averaged to provide an overall expected term.
See Figure SC 9-1 for an illustration of how a company would apply the simplified method of estimating the expected term of an award with a four-year, graded vesting schedule (see additional illustration in footnote 77 of SAB Topic 14). If the SAB Topic 14 criteria are met, this method can be used regardless of the attribution method used to recognize compensation cost (see SC 2.8 for information on attribution methods).
Figure SC 9-1
Application of the simplified method of estimating expected term
The following is the calculation of the expected term by vesting tranche:
The following is the calculation of the expected term for all vesting tranches:

9.3.2 Evaluating historical exercise data for expected term

Because most public companies have historical data on their employees' exercises of stock options, that should be the starting point for developing the expected term assumption. When completing the analysis, a company should (1) track behavior on an employee-by-employee basis from the grant date through the settlement date (e.g., exercise or post-vesting cancellation) and (2) make adjustments for any changes in award terms during the historical period in relation to current awards (i.e., where the history may not be indicative of the future). To appropriately develop the expected term assumption for a new award, a company should analyze historical information on options:
  • whose recipients would be expected to exhibit similar exercise and post-vesting termination behavior,
  • with a similar contractual term to expiration,
  • with a similar vesting schedule, and
  • with other contractual provisions similar to the award being granted.

Additionally, a company should consider whether it has an anomalous stock price history that may indicate that its historical exercise patterns may not be predictive of future exercise patterns (for example, if options were under water during most of the available exercise period or there was a sharp increase in the company's stock price over a long period of time). Once this information is collected and analyzed, a company can estimate a historical average holding period (period from grant to exercise) for its employee options. See SC 9.3.5 and SC 9.3.6 for information about adjustments for anomalous historical periods.
If the demographics of the groups of employees receiving options have changed over time, a company may need to make adjustments to historical exercise data before arriving at an expected term assumption for the latest grant. The company could still leverage its historical data but should adjust it to reflect the new demographics (for example, by using the historical data for only those employees who exhibit similar characteristics to the current covered group, or by appropriately re-weighting the various considerations underlying the expected term assumption). Similarly, if certain events or policy shifts have affected exercise behavior in the past, a company may have to isolate and remove portions of its historical data in favor of recent or more relevant information. In addition, the behavior of employees affected by a prior merger or spin-off may be different from what the company can expect from its current employees.
When analyzing historical exercise information, consideration should also be given to whether exercises generally happen evenly throughout the year or if there are seasonal effects. If exercises happen evenly throughout the year, this assumption can be used to simplify the historical calculation. If exercises are not spread evenly, a more refined approach to calculating the term of each exercise may be appropriate.

9.3.3 Pre-vesting forfeitures vs. post-vesting cancellations

The expected term assumption is intended to reflect the settlement of all vested options, including voluntary exercise, forced exercise (i.e., upon employee termination), and expirations. The term "post-vesting cancellations" refers to all events that may lead to a vested option not being exercised. These events, which occur once employees vest, need to be considered when developing the expected term assumption. In contrast, because previously recognized compensation cost is reversed for awards that are forfeited prior to vesting, a company would not consider pre-vesting forfeitures in determining the expected term assumption.
The expected term assumption should also reflect the possibility that some vested options may never be exercised because they will expire under water while the holder is still an employee. In computing historical average holding periods, a company should count those expired vested options as though they were exercised at expiration, because it reflects the period the awards were held by the employee.
Companies should consider the distinction between pre-vesting forfeitures and post-vesting cancellations when developing their expected term assumption. Some software packages used to administer stock-based compensation plans do not correctly segregate pre- and post-vesting events, which may inadvertently skew a company's expected term analysis by either incorrectly increasing or decreasing its expected term assumption. In addition, segregation of voluntary and forced early exercises (upon termination of employment) is generally necessary for development of the expected term assumptions under a lattice model.

9.3.4 Adjustments for partial life-cycles

Companies should make adjustments for potential bias due to recently granted unexercised options to account for what is called the partial life-cycle effect. For example, if a company typically issues options with a contractual term of ten years, the only exercise data that would be available for a full award lifecycle would be for options issued ten or more years ago. If the company does not make some adjustment for these outstanding options and instead calculates the average holding period based on partial exercise and post-vesting cancellation data, the expected term assumption and resulting fair value will most likely be too low, because it will not include the impact of outstanding options that will be exercised, expired, or cancelled (post-vesting) at a later date.
Additionally, companies should consider whether to only include option grants that are fully vested or to also include partially vested awards. This decision will depend on whether emerging experience is different from prior exercise experience as well as the amount of total data available.
Several methods of adjusting exercise data for the partial life-cycle effect exist, such as those listed below:
  • Exercised at expiration. While some recordkeeping software assumes outstanding options will be held until the end of their contractual term, this generally overstates the expected term assumption because, as practice has proven, there is no reason to believe that all outstanding employee options will be held until expiration. Accordingly, other approaches to adjust for the partial life-cycle effects, such as those described below, are generally more appropriate.
  • Exercised uniformly over remaining term (between the later of vesting date and date of the analysis, and the contractual expiration date of each option). This method is an impartial approach for estimating expected term, but it may not be appropriate in all situations. For example, if there is clear evidence that non-uniform exercise patterns occur in the later years of options' life-cycles, the uniform exercise approach method for dealing with outstanding options should not be used.
  • Marginal exercise rates. This more sophisticated method involves estimating marginal exercise rates to complete the life-cycle for each grant. Using this approach, a company determines the weighted-average percentage of options for each grant year (e.g., 20X1) that were exercised in a given period post-grant (e.g., in 20X4, 20X5) in relation to all options for that grant year eligible to be exercised in each given period. These percentages can be averaged over the grant years and then used to model a distribution of expected exercises that reflects all available data in an unbiased manner. If a company has only partial data (e.g., it grants ten-year options but has only five years of history), the marginal rates for the final years could also be estimated using published data, if available. If no published data is available, it may be reasonable to combine estimated marginal exercise rates for earlier life-cycle years with a uniform exercise assumption for later years, spreading outstanding options evenly over life-cycle years after the last year for which marginal rates could reasonably be estimated.

9.3.5 Adjustments for insufficient historical data

The sample size of historical exercises should be large enough to generate a reliable expected term assumption. The appropriate sample size of historical exercises depends on the inherent variability within the data and the number of adjustments a company has to make to that data. An otherwise large amount of data may not be sufficient if options were either significantly in-the-money or out-of-the-money during much of the observation period, or a significant company-specific event (e.g., downsizing) occurred that significantly affected exercise patterns.
If management believes that the expected term assumption derived from historical company-specific data is a poor indicator of future exercise patterns, it could use appropriate subsets of that data, or use data from other sources to replace or supplement the company's data. Some compensation consulting firms compile databases of exercise information collected from a large sample of companies of various sizes in different industries to (1) supplement the datasets from the limited number of academic studies on this subject and (2) provide companies with a wider dataset from which to build more refined expected term assumptions.
Companies that conclude they have inadequate exercise history and no access to alternative sources may use the simplified method discussed in SAB Topic 14 (see SC 9.3.1) if certain criteria are met. For example, if a company has a significant history of option grants, but nearly all of those grants have been continuously or nearly continuously out-of-the-money, the available exercise windows may yield only negligible exercise data. Another example is when a new company has made significant grants, but most are still unvested. Basing an expected term on the limited exercise data available may not yield a reasonable assumption.

9.3.6 Adjustments for stock price movements

Companies should consider whether exercise patterns are affected by shifting risk-preferences among employees or other external conditions. The most important external condition is stock price movements; employees' exercise decisions are frequently affected by stock price patterns.
Option pricing models implicitly consider several potential stock price paths. Accordingly, a company should not base the expected term of new options on historical data that reflects a unidirectional stock price trend – i.e., only rising (bull market) or falling (bear market) stock-price history. A predominantly bull market sample would tend to result in estimates that understate the expected term, while a bear market sample would tend to overstate it.
Lattice models, by design, as described in SC 9.3.11.1, directly address this over/understatement problem. But when the Black-Scholes model is used, adjustments may be necessary to deal with a largely unidirectional historical stock-price pattern. The following are three generally appropriate ways to address this situation:
  • Use more historical information to dilute the effect of periods strongly influenced by unusual market movements
  • Use data from academic or compensation consultants' studies as a basis or to supplement the historical data
  • Use an approach similar to the SAB Topic 14 simplified method (with appropriate adjustments to reflect the facts and circumstances of the award or grantee population).

In general, it would not be appropriate for companies to selectively use small portions of relatively recent historical exercise data, while excluding other portions based on unusual stock price movements. That approach would imply a forecast of future stock price movements, while financial theory assumes that future price-changes are not foreseeable. Historical exercise data that is strongly influenced by unusual stock-price movements should either be considered entirely irrelevant to future expectations, or possibly used to support an estimate that might be blended with estimates based on other sources, depending on how unusual the historical stock-price path is.
Companies should carefully observe the effect of stock price changes on exercise patterns, especially for more recent data, as the effects of stock prices might interact with the partial life-cycle effect. For example, if a company had a consistently rising stock price until five years ago, at which time the stock price began to fall, its pattern of exercises will likely indicate that employees are tending to hold their options longer for more recent grants. Due to the partial life-cycle effect, however, the average time until exercise for grants made in the past five years may still be much shorter than for older grants. If the outstanding options from these recent grants are extrapolated over their remaining lives, or alternatively, if more sophisticated marginal exercise rate analyses are employed on the data, a pattern of a lengthening holding period may become apparent. Observing this effect highlights the need to combine appropriately adjusted data from recent grants into the overall estimate of future holding periods.
Sometimes employees' appetite for risk and their exercise patterns change despite consistent stock performance. In such cases, a company should consider basing its estimates of future exercise behavior on data that largely reflects recent exercise patterns.

9.3.7 Using historical exercise data to calculate expected term

Once a company analyzes and, if necessary, adjusts its historical exercise data, it can use this data to calculate the expected term. This entails obtaining a weighted average of the holding periods for all awards (i.e., the average interval between the grant and exercise or post-vesting cancellation dates) adjusted as appropriate. While companies can sometimes group options by the month of their grant and/or exercise date, using the exact number of days between the grant and the exercise dates yields a more accurate expected term assumption.

9.3.8 Stratifying the employee population for expected term

This section has so far focused on developing a single expected term assumption for all grants made to the entire employee population. However, different types of employees (e.g., management vs. non-management) or employees of different ages or geographic location may have different appetites for risk and thus different propensities to exercise early. Thus, using a different expected term assumption for different groups of employees will likely yield a more refined estimate of exercise behavior. Stratification may be by position, salary range, geography, age, or any other factor that could affect exercise behavior.
Because fair values produced by the Black-Scholes model are not a linear function of the expected term, stratification of the employee population by the expected term assumption generally has less impact on the fair value of an option with a longer average expected term than one with a shorter average expected term. Typically, the average fair value estimate derived using different expected terms for different groups of employees is marginally lower than if a single expected term is used for all employee groups. Even though the average per share fair value weighted for class size may only be marginally different after stratification, the ultimate cumulative expense may be impacted to a greater degree if different groups of employees have significantly different rates of forfeiture. Therefore, if there are sub-groups of employees with significantly different expected exercise behavior and forfeiture experience, whose options represent a significant percentage of total company options granted, development of a separate expected term assumption should be considered for each one, provided there is relevant data upon which to develop these stratified assumptions.

9.3.9 Stratifying by vesting tranche for the expected term

ASC 718 allows valuation of options with graded vesting using a single expected term assumption for the entire grant or separate expected term assumptions for each tranche of the award. The practice of stratifying assumptions by vesting tranches will create an incremental (albeit small) reduction in aggregate compensation cost because stratifying by vesting tranche can separate early-exercising options (generally a lower fair value) from later-exercising options (generally a higher fair value). When analyzing exercises by vesting tranche, one potential challenge is that the option exercises are often not specifically tracked, or linked, to a specific tranche. For example, if an employee has vested in awards from two separate tranches and then exercises a portion of those vested options, typically there are no detailed records of which tranche of options were actually exercised. Although GAAP is silent, we believe it would be appropriate in these circumstances for companies to assume that the first exercises were from the first tranche to vest and that subsequently exercised options were from any remaining options in the first tranche, followed by options in later tranches, in order of vesting.
Regardless of the manner in which the expected term assumption is determined – i.e., by stratifying the options by tranche or by using a single expected term ─ companies can still avail themselves of the accounting policy election of either straight-line or graded attribution of the aggregate compensation cost over the requisite service period for awards with graded vesting and service conditions only.

9.3.10 Other considerations for expected term assumptions

Companies may consider using different volatility assumptions for different intervals of the overall expected term of an award because volatility may be expected to change over the expected term. Volatility that is assumed to change over time may also affect exercise patterns. Generally, only the more sophisticated lattice or Monte Carlo models can incorporate these relationships. However, when valuing options it is possible to adjust historical exercise data to reflect the assumption that future volatility will differ from recent stock-price volatility (see SC 9.4).
The expected term for option valuation may be impacted by the expected dividend yield. Because employees receiving options generally do not receive dividends on the underlying stock until they exercise, larger dividends offer an additional incentive to exercise options early. Companies should therefore consider adjusting the expected term assumption for significant differences between historical and expected future dividend yields.
Although ASC 718 acknowledges that blackout periods may affect the expected term assumption; it is rare that contractual or SEC-required blackout periods directly affect early exercise behavior or have a significant effect on the measurement of options' fair values. Such periods tend to be fairly short (e.g., six months) and, if they recur, will have already been incorporated into the exercise history.
Occasionally, for potential tax advantages, options may be exercisable prior to vesting. The exercise price is returned to the employee and the award is forfeited if the employee is terminated prior to vesting. For accounting purposes, this type of exercise is not considered substantive. Therefore, any historical analysis of exercise activity should reflect such an exercise as occurring at the vesting date, for those options that vest, and should exclude from the analysis any options that are forfeited.

9.3.11 Comparing expected term assumptions under different models

Lattice models are generally believed to produce a more refined estimate of fair value than the Black-Scholes model because they have the capacity to incorporate assumptions that vary over time and over potential stock price scenarios. Moving from the Black-Scholes model to a lattice model requires developing more complex assumptions concerning early exercise behavior. Because of the intricacies involved, lattice models are covered only in summary form in the discussion that follows, using examples to illustrate some of the considerations involved.

9.3.11.1 Modelling exercise behavior in relation to stock price

Lattice models replace the single expected term assumption of the Black-Scholes model with a set of assumptions that describe employees' early exercise behavior. That set of assumptions can range from a number of simple assumptions (similar to the expected term assumption under the Black-Scholes model) to an array that correlates the rate at which employees are expected to exercise their options to varying levels of stock price appreciation, as well as other factors. One typical difference between the Black-Scholes model and a lattice model is the manner in which a typical termination provision is handled. Most employee options include a clause that accelerates the contractual expiration of a vested award to a date 60 to 90 days after termination of employment, regardless of the remaining contractual term. The post-vesting termination rate is reflected indirectly in the single expected term assumption in the Black-Scholes model. However, a series of rates that change over the contractual term is generally a separate set of assumptions in a lattice model.
One approach to implementing a lattice model involves estimating the probability distribution of early exercise over two variables: the time that has elapsed between the grant date and the exercise date, and the assumed level of stock-price appreciation at the time of exercise. As described in SC 8.5, this latter variable is called the suboptimal exercise factor and is usually expressed as a multiple of the exercise price. Suboptimal exercise factors may (1) be single values, (2) be values that change over the life of an option, or (3) take the form of probability distributions.
A simple set of assumptions in a lattice model incorporating stock price appreciation is comprised of a single suboptimal exercise factor and fixed rate of post-vesting cancellations, along with the vesting period and contractual term of the option. The option would be assumed to (1) be exercised immediately at any point after vesting when the suboptimal exercise factor is reached; (2) be exercised on expiration if in-the-money but the suboptimal exercise factor is not reached; and (3) expire worthless if out-of-the-money.
A more elaborate set of assumptions to be used in a lattice model could involve either multiple suboptimal exercise factors (and/or post-vesting cancellation assumptions) that change over time, or probability distributions.
Figure SC 9-2 presents an illustrative distribution of the probability of exercise for an award that cliff vests after one year of service.
Figure SC 9-2
Illustration of probability distribution of early exercise
Years after grant
Suboptimal exercise factors
0–1
1–2
2–3
3–4
4–5
5–6
6–7
7–8
8–9
9–10
10
> 3.0
0%
100%
100%
100%
100%
100%
100%
100%
100%
100%
100%
2.8–3.0
0%
99%
99%
100%
100%
100%
100%
100%
100%
100%
100%
2.6–2.8
0%
98%
98%
99%
99%
99%
100%
100%
100%
100%
100%
2.4–2.6
0%
95%
95%
96%
96%
98%
99%
99%
99%
99%
100%
2.2–2.4
0%
88%
88%
92%
92%
95%
96%
98%
98%
98%
100%
2.0–2.2
0%
79%
79%
84%
84%
88%
92%
95%
95%
95%
100%
1.8–2.0
0%
66%
66%
73%
73%
79%
84%
88%
88%
88%
100%
1.6–1.8
0%
50%
50%
58%
58%
66%
73%
79%
79%
79%
100%
1.4–1.6
0%
34%
34%
42%
42%
50%
58%
66%
66%
66%
100%
1.2–1.4
0%
21%
21%
27%
27%
34%
42%
50%
50%
50%
100%
1.0–1.2
0%
12%
12%
16%
16%
21%
27%
34%
34%
34%
100%
In Figure SC 9-2, the early exercise probabilities are cumulative and correlate with various stock-price appreciation rates. If the stock price is between 2.0 and 2.2 times the exercise price between two and three years after the grant date, the model assumes that 79% of the options will have been exercised. Between three and four years, assuming the stock price remains constant, the proportion assumed to have been exercised climbs to 84%.

9.3.11.2 Expected term for awards with graded vesting

Typically, a company that offers options with graded vesting features would construct a separate probability distribution for each vesting tranche because the vesting date—the first date when exercises can occur—will be different for each tranche. The vesting date is an important input in lattice models because these models consider the possibility that if the stock price has risen significantly above the exercise price by the vesting date, it is very likely that employees will exercise their options immediately upon vesting. By contrast, the estimate of fair value under the Black-Scholes model is only indirectly affected by the vesting period in that the vesting period affects the expected term assumption.
Developing a probability distribution like the one shown in Figure SC 9-2 begins with an analysis of historical exercise data. In addition to elapsed time since grant date, this process considers the effect of stock-price appreciation on expected exercise. Generally, the early exercise distribution used in a lattice model will reflect the hypothesis that exercise becomes increasingly likely as the underlying stock's price appreciates. If a company does not have historical data to support this assumption, it may have to use another modeling technique or data from outside sources.

9.3.11.3  Sources of bias and adjustments to historical data

A company using a lattice model should understand its data requirements and the potential sources of bias in estimating the probability distribution of early exercise. Both Black-Scholes and lattice models can use the methods described earlier to address biases arising from an incomplete exercise history. However, extended periods of consistent upward or downward stock-price movement, lack of relevant data, historical data that does not fairly reflect future expectations and other factors can affect lattice models in more complex ways due to multiple assumptions about early exercise behavior and the addition of stock-price appreciation levels and other variables. For example, distributions of actual exercises based on recent historical data dominated by periods of extreme stock-price depreciation or appreciation relative to the prices on the grant dates are likely to overstate or understate how long employees are likely to hold their options in the future. Adjustments to historical data should be made in such cases in order to support a lattice model that reasonably reflects future expectations.
Lattice models may require different adjustments than the Black-Scholes model. For example, a historical stock-price path that was dominated by rapid appreciation (and high levels of early exercise that often accompany this scenario) might require further analysis and adjustment of the historical expected term under the Black-Scholes model (as noted in SC 9.3.2), because such rapid stock price appreciation is not expected to recur. Under the lattice model, the same historical stock price path might result in suboptimal exercise factors that are too high because simply applying the historical data to the new grants assumes that the historical stock price path will continue. The assumptions developed for lattice models will therefore have to be based on careful analysis, including adjustment for potential biases and mitigation of the impact of data affected by unusual stock price history that is not reflective of future expectations. Since lattice models typically will require more assumptions than those used in the Black-Scholes model, more analysis will generally be required to properly develop assumptions for lattice models.

9.3.11.4  Post-vesting cancellations and suboptimal exercises

Unlike the Black-Scholes model, lattice models treat post-vesting cancellations and voluntary early exercise behavior as two separate assumptions. Because the options of terminated employees may often be exercised earlier and at lower levels of stock-price appreciation than the options of employees who remain and are typically cancelled without any payoff if they are underwater during the post-termination exercise period (generally, 60 to 90 days), lattice models can reflect this assumption in more detail than the Black-Scholes model. The post-vesting cancellation assumption should be based on the actual behavior of a similar group of employees. In developing the probabilities of voluntary early exercise for a lattice model (unlike the development of expected term for the Black-Scholes model), the post-vesting cancellations should be excluded, because they are considered separately. Thus, when using a lattice model, an analysis should be performed to separate a company's history of employee exercise behavior into two categories: voluntary (early) exercise and forced exercise that results from termination of employment.
A simpler, less refined form of lattice modeling assumes that early exercise occurs 100% of the time when the stock price first reaches a level represented by a single suboptimal exercise factor. This factor is normally estimated by analyzing probabilities of early exercise over various historical periods in relation to stock price appreciation at the time of exercise. It may be necessary to adjust the data for possible biases due to unusual stock price movements, and there is some inherent unreliability in using a single exercise factor.

9.3.11.5 Limitations of only company-specific exercise history

Many companies will not have sufficient exercise history or the ability to analyze company-specific historical data that is necessary to support the exercise distribution assumptions required for lattice models. A company that decides to use a lattice model may need to hire outside consultants to assist with software, develop assumptions, and determine any adjustments necessary to mitigate data biases and deficiencies.
Finally, lattice models may incorporate other predictors of early exercise. Other variables tied to stock price performance (besides time and stock price) that may be used in an exercise-prediction model include recent stock price performance (over various windows) or recent stock price volatility. The same considerations may be applied when developing early exercise assumptions to be used in a Monte Carlo simulation model.
Expand Expand
Resize
Tools
Rcl

Welcome to Viewpoint, the new platform that replaces Inform. Once you have viewed this piece of content, to ensure you can access the content most relevant to you, please confirm your territory.

signin option menu option suggested option contentmouse option displaycontent option contentpage option relatedlink option prevandafter option trending option searchicon option search option feedback option end slide