Valuing Employee Stock Option (ESO) With Sequential Execution Using Lattice Method Erwinna Chendra 1,2 , Dai Tian-Shyr 3 , Kuntjoro A. Sidarto 4 1,4 Industrial and Financial Mathematics Group, Institute of Technology Bandung, Indonesia [email protected], [email protected]2 Department of Mathematics, Parahyangan Catholic University, Indonesia [email protected]3 Institute of Finance, National Chiao Tung University, Taiwan [email protected]Abstract Employee Stock Options (ESOs) are frequently used by a company to encourage its employees without the burdens of immediate cash payout. How to evaluate ESOs accurately from employees’ point of view is critical for designing effective salary systems. Empirical studies suggest that ESOs are sequentially exercised in parts by employees. Our paper analyzes this sequential execution phenomenon by considering the impacts of employee’s income tax and share dilution due to exercises of ESOs. Specifically, we develop a novel forest model that analyzes an employee’s execution strategy to maximize the present value of the lump sum of his future disposable incomes, which is defined as his salary plus ESOs payoff minus the progressive income tax. The forest is composed of several lattices and each lattice simulates one possible dilution scenario. Each node of the lattice contains a table for analyzing the optimal exercising strategy at that node. The transition among the nodes in the forest models the sequential executions of an employee and the corresponding dilutions. Numerical experiments are given to analyze and to verify the robustness of the lattice. 1. Introduction Employee stock option (ESO) is a call option on the issuer’s stock granted by an issuing company to its employees. Employees can exercise ESOs in part and may lose their unexercised options once they resign. Thus ESOs are useful and popular financial instruments for a firm to retain key employees and motivate them to improve the company’s performance. Unlike usual options that can be traded by typical institutional or individual investors, employees cannot sell or transfer their ESOs to other investors. To maintain employees’ long-term incentives, an ESO usually has a long maturity ranged from 5 to 15 years. It is not exercisable during the first few years of the option’s life (called the vesting period). In case an employee leaves the company during the vesting period, then his ESOs is forfeited (i.e. ESOs become worthless). Otherwise, the holder can repeatedly exercise his ESOs in part to optimize his/her benefits. Analyzing the benefits for granting an employee ESOs is critical for designing effective salary systems.
24
Embed
Valuing Employee Stock Option (ESO) With Sequential ... Employee Stock Option (ESO) With Sequential Execution Using Lattice Method Erwinna Chendra1,2, Dai Tian-Shyr3, Kuntjoro A. Sidarto4
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Valuing Employee Stock Option (ESO) With Sequential Execution
Using Lattice Method
Erwinna Chendra1,2
, Dai Tian-Shyr3, Kuntjoro A. Sidarto
4
1,4
Industrial and Financial Mathematics Group, Institute of Technology Bandung, Indonesia
The “after-tax” payoff, or the marginal increment of disposable income, for exercising ESOs to
gain $10,000 “pre-tax” income should be $8,245, which is the difference between $34,206.25
(i.e., the disposable income for $40,000 taxable income) and $25,961.25. Obviously, exercising
too many ESOs at the same time would incur higher marginal tax rate; on the other hand,
postponing executions of ESOs would result in loss of time value. The optimized ESO execution
strategy can be quantitatively analyzed as the one to maximize the lump sum of expected present
value of future disposable incomes, which is also the after-tax present value of the holder’s ESO
and future salaries evaluated under the risk-neutral valuation method. The value of ESOs can be
evaluated as the lump sum expected present values of their after-tax payoffs implied by the
aforementioned optimal execution strategy.
The optimal execution strategy and hence the value of ESOs for an arbitrary node (𝑖, 𝑗) in the
state transition lattice can be determined backwardly from the last time step n to time step 0 by
applying the following recursive formulas. Let 𝑀𝐷𝑥𝑖,𝑗
denote the maximum present value of
disposable incomes received from time step i to maturity given that 𝑥 𝑚⁄ % of ESOs are already
exercised prior to time step 𝑖. At the maturity (i.e. 𝑖 = 𝑛), an ESO holder will choose an optimal
accumulated execution ratio 𝑦′ that maximize 𝑀𝐷𝑥𝑛,𝑗
as follows
𝑀𝐷𝑥𝑛,𝑗
= 𝐷𝑥,𝑦′𝑛,𝑗
= max𝑥≤𝑦≤𝑚
[𝐷𝑥,𝑦𝑛,𝑗
] = max𝑥≤𝑦≤𝑚
[(𝑊 + 𝐸𝑥,𝑦𝑛,𝑗
) − 𝑇𝑎𝑥𝑥,𝑦𝑛,𝑗
]. (5)
The after-tax value for ESO at node (𝑛, 𝑗) implied by the optimal strategy 𝑦′ is
VESO𝑥𝑛,𝑗
= ESO𝑥,𝑦′n,𝑗
= 𝐷𝑥,𝑦′𝑛,𝑗
− (𝑊 − 𝑇𝑎𝑥𝑊). (6)
For a lattice node prior to maturity, its optimal execution strategy and corresponding values of
ESOs and disposable incomes can be calculated by the backward induction procedure. Note that
an ESO holder would pick his optimal exercise amount 𝑦 − 𝑥 at node (𝑖, 𝑗) to maximize the
present value of disposable incomes received from time step 𝑖 to maturity described as follows
𝑀𝐷𝑥𝑖,𝑗
= max𝑥≤𝑦≤𝑚[𝐷𝑥,𝑦𝑖,𝑗
+ EP𝑉𝑦𝑖,𝑗
], (7)
where EP𝑉𝑦𝑖,𝑗
denotes the present value at node (𝑖, 𝑗) of future disposable incomes received from
time step 𝑖 + 1 to ESO’s maturity given that the accumulated execution ratio is (𝑦/𝑚)%. Based
on the outgoing binomial structure of the CRR lattice illustrated in Figure 1, EP𝑉𝑦𝑖,𝑗
can be
evaluated by the backward induction procedure described as follows
EP𝑉𝑦𝑖,𝑗
= 𝑒−𝑟∆𝑡[𝑝 ∙ 𝑀𝐷𝑦𝑖+1,𝑗+1
+ (1 − 𝑝) ∙ 𝑀𝐷𝑦𝑖+1,𝑗
]. (8)
Let the optimal accumulated execution ratio that maximize the right hand side of Eq. (7) be 𝑦′. The after-tax ESO value under the optimal exercise strategy can be evaluated as
VESO𝑥𝑖,𝑗
= ESO𝑥,𝑦′𝑖,𝑗
+ 𝑒−𝑟∆𝑡[𝑝 ∙ VESO𝑦′𝑖+1,𝑗+1
+ (1 − 𝑝) ∙ VESO𝑦′𝑖+1,𝑗
]. (9)
By repeatedly applying Equations (7), (8), and (9) alternatively from time step 𝑛 − 1 back to
time step 0, we can obtain the optimal execution strategy, the corresponding payoff for
exercising ESOs, and the disposable income at each node. The fair value of ESO evaluated by
the state transition lattice is the after-tax ESO value at the root node; that is, VESO00,0
.
The above model can be extended to deal with the vesting period and the turnover rate. Note that
a holder cannot exercise his/her ESOs during the vesting period and complex backward
induction for determining optimal sequential execution is not required. Take the vesting period
that ends at time step 1 in Figure 2 for example. An ESO holder cannot exercise any ESOs at
time step 0 and only one state is required at the root node. The dimension of previous
accumulated execution ratio of ESOs for the nodes in time step 1 is not required since no ESOs
can be exercised prior to time step 1. The structure of state transition lattice (see Figure 1) is
applied at the end of the vesting period as illustrated in the latter part of the lattice in Figure 2.
The turnover rate can be incorporated by inserting “leaving the company” branch to each node of
the lattice. The branching probability 𝛿Δ𝑡 denotes the probability for the employee to leave
within a time step Δ𝑡, where 𝛿 is the annual turnover rate. Note that, an employee can exercise
his/her ESOs after the vesting period before he/she leaves, but he/she cannot exercise the
remaining unexercised ESOs after he/she leaves the company.
Figure 2 Dealing with the vesting period and the turnover rate
4. Dilution or Expanses due to ESOs Execution
The evaluation of ESOs should consider execution costs or dilution effects due to contingent
executions, especially when the outstanding amounts are large. Specifically, the execution of an
ESO can be either settled in cash or in equity. Cash-settled ESOs is recognized as liability
because the execution of ESOs would require the issuer to pay cash to holder. On the other hand,
the equity transferred to holders that exercise equity-selected ESOs can be obtaining by either
purchasing from the stock market or by issuing new shares. Whether ESOs are treated as
liabilities or equities may raise the so-called “double accounting” problem mentioned in Bodie et
al. (2003). That paper mentions a New York Times column4 written by the venture capitalist
4 It was published in April 5, 2002.
John Doerr and FedEx CEO Frederick Smith that reads “the impact of options would be counted
twice in the earnings per share: first as a potential dilution of the earnings, by increasing the
shares outstanding, and second as a charge against reported earnings. The result would be
inaccurate and misleading earnings per share”. To fairly evaluate the impacts of ESOs executions
on the prevailing stock price and hence the payoffs of executions without causing the
aforementioned double accounting problem, two evaluation methods that consider solely
execution expenses and dilution effects are discussed as follows. Let us consider a simple case
that the firm grants 𝛽 shares to ESOs holder at maturity due to executions. If these shares are
purchased from the market, the after-settlement stock price 𝑆′ should be
𝑆′ =𝑉−𝐷−𝛽𝑆+𝛽𝐾
𝛼 , (10)
where 𝑉 is the firm value, 𝐷 is the amount of due debt, 𝑆 denotes the price for the firm to
purchase the stocks, 𝐾 is the ESO strike price, and 𝛼 is the number of outstanding shares. To
solve Equation (10) with two unknowns 𝑆 and 𝑆′, we add one more constraint 𝑆 = 𝑆′ due to the
arbitrage-free argument described as follows. Assume that 𝑆′ > 𝑆 then someone can borrow 𝛽𝑆
dollars to buy stock and at ESO maturity he gets 𝛽(𝑆′ − 𝑆) dollars. If 𝑆′ < 𝑆 then he can sell 𝛽
shares to get 𝛽𝑆 dollars and at ESO maturity he gets 𝛽(𝑆 − 𝑆′) dollars. So 𝑆′ should be equal to
𝑆 because if they are not equal then there is an arbitrage opportunity. If the company issues new
shares, then at debt and ESO maturity, the stock price becomes
𝑆′ =𝑉−𝐷+𝛽𝐾
𝛼+𝛽 (11)
For cash-settled ESO, at debt and ESO maturity we have new equity value as
𝑉 − 𝐷 − 𝛽(𝑆 − 𝐾)
and if we divide it by the number of outstanding shares, we also have equation (10). So we can
say that equity-settled ESO by purchasing old shares is economically equivalent to cash-settled
ESO. But this is not true if the company issues new shares. From equation (10), we can see that
𝛽(𝑆 − 𝐾) is the cost and there is no dilution in the denominator of equation (10). As well as the
equation (11), there is no cost in the numerator and the dilution effect occurs from the
denominator of equation (11). So we can argue there is not “double accounting” problem with
equity-settled ESO or cash-settled ESO from model point of view.
Next, we consider dilution effect of ESO in our sequential execution model. We will use the
forest model to price ESO with sequential execution and dilution effect. The forest model is
composed of several lattices and each lattice simulates one possible dilution scenario. The
transition among the nodes in the forest models the sequential executions of an employee and the
corresponding dilutions. We modify the bino-trinomial (BTT) method by Dai and Lyuu (2010)
that uses trinomial structure to connect the diluted stock price to others CRR binomial lattices
that represent possible capital structures.
The following discussions only focus on equity-settled ESO by issuance new shares. Again we
assume that all ESOs can be divided into 4 indivisible units (𝑚 = 4). Then we have 5(= 𝑚 + 1)
layers of CRR binomial lattice, which each lattice represents one possible capital structure.
Assume the initial stock price for the 1st CRR lattice (when 0% of ESO being executed) is 𝑆0
1,
then we can build the x-th (𝑥 = 2, … , 𝑚 + 1) CRR lattice with
𝑆0𝑥 =
𝑆01𝜔+𝐾
𝑥−1
𝑚𝑏
𝜔+𝑥−1
𝑚𝑏
(12)
as the initial stock price, where K is strike price, 𝜔 is the number of outstanding shares before the
exercise of ESO and 𝑏 is the total number of ESO. Then the stock price at node (𝑖, 𝑗) in the x-th
layer CRR lattice is:
𝑆𝑖,𝑗𝑥 = 𝑆0
𝑥𝑢𝑗𝑑𝑖−𝑗 (13)
with 𝑢, 𝑑, and p as before. At the maturity date (time step n), the pre-tax for exercising 𝑦 − 𝑥
units of ESOs payoff is
𝐸𝑥,𝑦𝑛,𝑗
=𝑦−𝑥
𝑚𝑏 (
𝑆𝑛,𝑗𝑥 [𝜔+
𝑥−1
𝑚𝑏]+𝐾
𝑦−𝑥
𝑚𝑏
𝜔+𝑦−1
𝑚𝑏
− 𝐾) (14)
0 ≤ 𝑗 ≤ 𝑛, and 0 ≤ 𝑥 ≤ 𝑦 ≤ 𝑚. An ESO holder will choose an optimal execution ratio to
maximize his/her disposable income 𝐷𝑥,𝑦𝑛,𝑗
; that is, maximized disposable income 𝑀𝐷𝑥𝑛,𝑗
can be
evaluated using Eq. (5) and the after-tax ESO value under the optimal exercise strategy is given
by Eq. (6).
Define the stock price at node (𝑖, 𝑗) of the 1st CRR lattice as 𝑆𝑖,𝑗
1 . Assume if 25% of ESO is
executed then the stock price is diluted to 𝑆𝑖,𝑗∗ . Note that 𝑆𝑖,𝑗
∗ may not be in the 2nd
CRR lattice.
Thus we need to connect the node 𝑆𝑖,𝑗∗ with next period (𝑖 + 1) stock prices at the 2
nd CRR
lattice. We can use trinomial structure to connect 𝑆𝑖,𝑗∗ to a unique choice of three proper
following nodes, say nodes 𝑆𝑖+1,𝑘−12 , 𝑆𝑖+1,𝑘
2 and 𝑆𝑖+1,𝑘+12 , in the 2
nd CRR lattice as in Dai and
Lyuu (2010) (see Figure 3).
Define the mean function 𝜇 and the variance function Var as (see Dai an Lyuu (2010))
𝜇(𝑧) ≡ (𝑟 −𝜎2
2) 𝑧
Var(𝑧) ≡ 𝜎2𝑧
and the V-log-price of stock price V’ as ln(𝑉′ 𝑉⁄ ). By the lognormality of the stock price, the
mean and the variance of the 𝑆𝑖,𝑗∗ -log-price of 𝑆𝑖+1,𝑘−1
2 , 𝑆𝑖+1,𝑘2 and 𝑆𝑖+1,𝑘+1
2 equal to 𝜇(∆𝑡) and
Var(∆𝑡), respectively. Assume the 𝑆𝑖,𝑗∗ -log-price of 𝑆𝑖+1,𝑘
2 as �̂�, which is the closest one to 𝜇(∆𝑡)
among the 𝑆𝑖,𝑗∗ -log-price of all nodes at period (𝑖 + 1) of the 2
nd CRR lattice. Define also 𝛽, 𝛼,
and 𝛾 as the difference between �̂� and 𝜇(∆𝑡), �̂� + 2𝜎√∆𝑡 and 𝜇(∆𝑡), and �̂� − 2𝜎√∆𝑡 and 𝜇(∆𝑡),
respectively. Then the branching probabilities of node 𝑆𝑖,𝑗∗ (i.e. 𝑝𝑢, 𝑝𝑚, and 𝑝𝑑) can be derived by
matching the first two moments of the logarithmic stock price process and the sum of the
branching probabilities equal to one. In Appendix A of Dai and Lyuu (2010), we can see that the
above probabilities are valid probabilities.
Figure 3 Three layers of CRR capital structure
For 𝑖 = 𝑛 − 1, 𝑛 − 2, … ,0 and 𝑗 = 0,1, … , 𝑖, the pre-tax for exercising 𝑦 − 𝑥 units of ESOs
payoff is
𝐸𝑥,𝑦𝑖,𝑗
=𝑦 − 𝑥
𝑚𝑏 (
𝑆𝑖,𝑗𝑥 [𝜔 +
𝑥 − 1𝑚 𝑏] + 𝐾
𝑦 − 𝑥𝑚 𝑏
𝜔 +𝑦 − 1
𝑚𝑏
− 𝐾)
Again an ESOs holder will choose a strategy that optimizes his disposable income using Eq. (7).
Note that if 𝑥 = 𝑦, we do not need trinomial structure. But if 𝑥 ≠ 𝑦 then we need trinomial
structure to connect the diluted stock price at period 𝑖 with stock price at period 𝑖 + 1 in the x-th
layer CRR lattice. So if 𝑥 = 𝑦 then EPV𝑦𝑖,𝑗
in Eq. (7) is equal to Eq. (8), but if 𝑥 ≠ 𝑦 then
EPV𝑦𝑖,𝑗
= 𝑒−𝑟∆𝑡 [𝑝𝑢 ∙ 𝑀𝐷𝑦𝑖+1,𝑘+1 + 𝑝𝑚 ∙ 𝑀𝐷𝑦
𝑖+1,𝑘 + 𝑝𝑑 ∙ 𝑀𝐷𝑦𝑖+1,𝑘−1]
Let the optimal execution ratio that maximize 𝑀𝐷𝑥𝑖,𝑗
in Eq. (7) be 𝑦′. Then for 𝑥 = 𝑦 the after-
tax ESO value under the optimal exercise strategy can be evaluated using Eq. (9), but for 𝑥 ≠ 𝑦
the after-tax ESO value under the optimal exercise strategy is
VESO𝑥𝑖,𝑗
= ESO𝑥,𝑦′𝑖,𝑗
+ 𝑒−𝑟∆𝑡[𝑝𝑢 ∙ VESO𝑦′𝑖+1,𝑘+1 + 𝑝𝑚 ∙ VESO𝑦′
𝑖+1,𝑘 + 𝑝𝑑 ∙ VESO𝑦′𝑖+1,𝑘−1]
5. Numerical Results
This section evaluates the ESO value with sequential execution. Before that, we give a simple
example with two time steps to illustrate the sequential execution phenomenon. Assume that
𝑆 = 100, 𝐾 = 80, 𝑇 = 2, 𝑟 = 0.1, 𝜎 = 0.4, 𝑊 = $50,000, 𝑏 = 10000, and 𝑚 = 4. First we
assume that there is no dilution effect on our model. So we can build the two period of CRR
binomial lattices as follow:
Figure 4 Two Period CRR Binomial Lattice
Note that because we assume 𝑚 = 4, then each node of CRR binomial lattice contains a table
with five rows and five columns as in Figure 1. First, we calculate the ESO benefit before tax for
node D, E, and F using equation (2) and then choose the optimal disposable income using
equation (3) for each node. Figure 5 shows the disposable income value for node D, E, and F at
maturity date.
Figure 5 Disposable incomes without dilution (at maturity)
For node D and E, the optimal disposable income is on the fifth column, which means the
optimal strategy is executing all remaining ESOs. But for node F, the optimal strategy is not
doing execution at all. So the ESO value at these nodes is the ESO benefit corresponding to the
optimal disposable income. Then we continue backward induction to node B and C. Again, we
calculate the ESO benefit before tax and the optimal disposable income for each node using
equation (4) and equation (5), respectively. Figure 6 shows the disposable income value for node
B and C.
Figure 6 Disposable incomes without dilution (at period 1)
Again, for node C, the optimal strategy is not doing execution at all. But for node B, we can see
the sequential execution phenomenon. For example, if we have executed 25% of ESOs at the
past, then the optimal strategy is executing 25% more. But if we have executed 75% of ESO,
then the optimal strategy is not doing any execution at period 1. Similar procedure for node A
and we can get the ESO value as the ESO benefit corresponding to the optimal disposable
income. Note that there are negative disposable incomes for node F and node C in Figure 5 and
Figure 6. These because of the ESO benefit before tax at node F and C are negative. It means
that if we execute ESO at node F and C, then we will lose money, because the stock price at
these nodes is lower than the exercise price. If we get negative disposable income then the tax
will be zero.
Next, we also give a simple example to illustrate the dilution effect. We use the same data as
before and assume that the number of outstanding shares before the exercise of ESO, 𝜔, is 250.
Because we assume 𝑚 = 4, then we have five layers of CRR binomial lattices. The initial stock
prices for the 2nd
to 5th
layers can be calculated using equation (8) and others node can be
calculated using equation (8). However, extra nodes are required in low layer lattices for
trinomial lattice connection. Because they are also CRR binomial lattices, then for the l-th CRR
lattice, we can add m extra nodes above stock price 𝑆𝑖,𝑖𝑙 with 𝑆𝑖,𝑖
𝑙 𝑢2, 𝑆𝑖,𝑖𝑙 𝑢4, … , 𝑆𝑖,𝑖
𝑙 𝑢2𝑚 and m extra
nodes below stock price 𝑆𝑖,0𝑙 with 𝑆𝑖,0
𝑙 𝑑2, 𝑆𝑖,0𝑙 𝑑4, … , 𝑆𝑖,0
𝑙 𝑑2𝑚. Figure 7 shows the five layers of
CRR binomial lattices. Node X and Y are the extra nodes at period 1.
In Figure 7, we can see that if we have 10000 numbers of ESOs and assume that 100% of ESOs
is executed then the original stock price $100 (node A1) becomes $80.4878 (node A5). But if
we only have 1000 numbers of ESOs then the original stock price becomes $84. Thus the
number of ESOs that companies give to employees will change the company capital structure.
Figure 7 The 5 layers of CRR two period binomial lattices
Figure 8 Disposable income with dilution (at maturity)
At time T (period 2), we calculate the pre-tax for exercising ESOs for each node using equation
(14). Note that we use node D1 if the accumulated exercised units of ESOs one-step prior to the
present time step is 0% (𝑥 = 1), we use node D2 if the accumulated exercised units of ESOs
one-step prior to the present time step is 25% (𝑥 = 2), and so on. Then we choose the optimal
strategy that maximizes our disposable income. Figure 8 shows the disposable income value at
the maturity. We use node D1 for the first row of the top table, node D2 for the second row of
the top table, node D3 for the third row of the top table, and so on. Similarly, we use node E1 for
the first row of the middle table and node F1 for the first row of the down table, and so on.
At period 1, for example, we consider about the first row of the table (if the accumulated
exercised units of ESOs one-step prior to the present time step is 0%). Assume that the
accumulated exercised units of ESOs up to present time step is 25%, then we use trinomial
branch to connect the diluted stock price from node B1 to three unique nodes at the 2nd
CRR
lattice (node X2, D2, E2, or node D2, E2, F2). To choose the three unique nodes, first we
calculated B1-log price of all nodes and the mean-variance at period 1. Then we calculated the
“distance” between those values and the mean, and we choose the closest node as the middle
node. For that example, we get B1-log price of X2 = 0.9993, B1-log price of D2 = 0.1993, B1-
log price of E2 = -0.6007, B1-log price of F2 = -1.4007, mean = 0.02, and Var = 0.16. We can
choose node D2 as the middle node. So we connect B1 to node X2, D2, and E2 and we have
𝛽 = 0.1793, 𝛼 = 0.9793, and 𝛾 = −0.6207. The branching probabilities can be derived by
matching the first two moments of the logarithmic stock price process and the sum of the
branching probabilities equal to one. We get 𝑝𝑢 = 0.0381, 𝑝𝑚 = 0.6998, and 𝑝𝑑 = 0.2622.
Thus, at node B1, if the optimal strategy is executing 25% with past accumulation of execution is
0% (at the 1st-row and the 2
nd-column), then we calculate the pre-tax for exercising ESOs using
the trinomial structure that connect diluted stock price at node B1 with node X2, D2, and E2. In
case if the accumulated exercised units of ESOs up to present time step is 50% then we use
trinomial branch to connect the diluted stock price from node B1 to three unique nodes at the 3rd
CRR lattice (node X3, D3, and E3 or node D3, E3, and F3).
For other example, we consider about the second row of the table (the accumulated exercised
units of ESOs one-step prior to the present time step is 25%), If we assume that the accumulated
exercised units of ESOs up to present time step is also 25% then we do not need any trinomial
branch because we do not have any diluted stock price. So we just use CRR binomial
probabilities to connect the node B2 to node D2 and E2, but if the accumulated exercised units of
ESOs up to present time step is 75% then we use trinomial branch to connect the diluted stock
price from node B2 to three unique nodes at the 4th
CRR lattice (node X4, D4, E4 or node D4,
E4, and F4).
The algorithm is implemented using MATLAB on computer with processor Core™i5, RAM 4
GB, and Windows7 32-bit OS. Given the data: 𝑆 = 100, 𝐾 = 100, 𝑇 = 5, 𝑟 = 0.1, 𝜎 = 0.3,
𝑊 = $50,000, 𝑚 = 4, 𝜔 = 250, and 𝑏 = 10000. The ESO price without dilution effect is
given by Figure 9 and the ESO price with dilution effect is given by Figure 10 for different time
steps. We can see that the ESO price with dilution effect ($11.8258) is cheaper than the ESO
price without dilution effect ($46.016). If we use the same data but 𝑏 = 1000, then the ESO
price with dilution effect is $14.9944 and the ESO price without dilution effect is $46.0159. We
can see that the number of ESOs makes significant impact on the ESO price with dilution effect.
.
Figure 9 ESO Price without Dilution vs Time Step
Figure 10 ESO Price with Dilution vs Time Step
We also analyze sensitivities of ESO’s price without dilution respect to the model parameters.
Figure 11, Figure 12, Figure 13, Figure 14, and Figure 15 show how the ESO price behave as the
interest rate, the volatility, the maturity date, the amount of ESO distribution, and the total
number of ESO changes. Again, the calculations are based on the same data used by Figure 9
except for the parameter that is varied in each panel and 𝑛 = 5000.
Figure 11 ESO Price vs Interest Rate
Figure 12 ESO Price vs Volatility
Figure 13 ESO Price vs Maturity Date
Figure 14 ESO Price vs Amount of ESO Distribution
Figure 15 ESO Price vs Total Number of ESO (with m=4)
Just as standard options, the increasing interest rate, volatility, and maturity date gives the
increasing ESO price, as shown in Figure 11, Figure 12, and Figure 13. Furthermore, the
increasing amount of ESO distribution gives the decreasing ESO price, as shown in Figure 14.
But, generally, the increasing total number of ESO also gives the increasing ESO price, see
Figure 15. There is something interesting from Figure 15, we can see that the ESO price
descends if the total number of ESO is in between 1000 to 5000. Figure 16 shows the ESO price
behaves as the total number of ESO changes between 1000 and 5000. We can see that there is a
minimum price before the ESO price increases continuously.
Figure 16 ESO Price vs Total Number of ESO with m=4 (Zoom In)
But this phenomenon does not occur if we take 𝑚 = 1 (i.e. we can only execute 0% or 100% of
ESO), the increasing total number of ESO gives the increasing ESO price continuously (see
Figure 17). We can conclude that the sequential execution will be effectively useful if the total
number of ESO is large enough.
Figure 17 ESO Price vs Total Number of ESO (with m=1)
6. Conclusion
References
[1] Aboody, D. 1996. Market valuation of employee stock options. Journal of Accounting and
Economics 22, 357-391.
[2] American Accounting Association (AAA) Financial Accounting Standards Committee
(FASC). 2004. Evaluation of the IASB's proposed accounting and disclosure requirements
for share-based payment. Accounting Horizons vol. 18 no. 1, pp. 65-76.
[3] Ammann, M. and R. Seiz. 2004. Valuing employee stock options: does the model matter?
Financial Analysts Journal, vol. 60, no. 5, pp. 21-37.
[4] Bajaj, M., S.C. Mazumdar, R. Surana, and S. Unni. 2006. A matrix-based lattice model to
value employee stock options. The Journal of Derivatives, vol. 14, no. 1, pp. 9-26.
[5] Balsam, S. 1994. Extending the method of accounting for stock appreciation rights to