Strategic Management Journal Strat. Mgmt. J., 29: 701–721 (2008) Published online in Wiley InterScience (www.interscience.wiley.com) DOI: 10.1002/smj.681 Received 20 October 2006; Final revision received 2 January 2008

REAL OPTIONS AND REAL VALUE: THE ROLE OF EMPLOYEE INCENTIVES TO MAKE SPECIFIC KNOWLEDGE INVESTMENTS HELI WANG1 * and SONYA SEONGYEON LIM2 1 Department of Management of Organizations, School of Business and Management, Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong 2 Department of Finance, Kellstadt Graduate School of Business, DePaul University, Chicago, Illinois, U.S.A.

Real options reasoning emphasizes the strategic value of making flexible investments in a turbulent environment. Employees’ investments in specific human capital are often critical to the success of a real option project, but the very flexibility that allows a firm to change course in response to new information also affects employees’ incentives to make such specific human capital investments. We develop a model of real option investment that explicitly incorporates the role of employee incentives. The model suggests that the effect of investing in a real option project on employee incentives may be positive, further increasing the value of the project, or negative, sometimes more than offsetting the benefit of flexibility and resulting in reduced project value. Therefore, firms and managers should take into consideration the role of employee incentives when applying real options logic to investment decision making. Copyright  2008 John Wiley & Sons, Ltd.

INTRODUCTION Earlier strategy research, especially work in the tradition of the resource-based approach, has emphasized the importance of firm-specific resource commitment in gaining and sustaining competitive advantage (Amit and Schoemaker, 1993; Barney, 1991; Teece, 1982). However, with the increasing dynamism in a firm’s operational environments, the firm’s ability to effectively adapt to environmental changes has become one of the most important sources of competitive advantage. Real options theory emphasizes the value of flexibility in the face of environmental uncertainty, so it has, not surprisingly, aroused Keywords: real options; employee incentives; specific knowledge investments *Correspondence to: Heli Wang, Department of Management of Organizations, School of Business and Management, Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong. E-mail: [email protected]

Copyright  2008 John Wiley & Sons, Ltd.

considerable excitement among scholars of strategic management (e.g., Kogut, 1991; Kogut and Kulatilaka, 2001; Luehrman, 1998; McGrath, 1997; Trigeorgis, 1996). A real option investment provides a firm with the flexibility to abandon a project or make further investments as new information arrives. This generally gives a real option investment greater value under uncertainty than an inflexible fullscale investment. However, many projects that have been considered by strategy researchers as appropriate for applying real options logic, such as research and development (R&D) (McGrath, 1997), joint ventures (Kogut, 1991), or the development of firm capabilities or core competences in general (Kogut and Kulatilaka, 2001), cannot be carried out without the participation of the firm’s key employees.1 In such cases, although electing 1 The key arguments made in this paper may be extended to include other stakeholders of the firm, such as suppliers or customers, who play an important role in carrying out a firm’s

702

H. Wang and S. S. Lim

to abandon the project reduces the firm’s downside risk, it may increase the risk borne by the employees, as their human capital investments specific to the project may become less valuable when the project is abandoned. Employees’ concerns for such risks are most significant if they have been required to invest in project-specific (or more generally, firm-specific) human capital: skills tied to the particular project with very limited alternative use in other business settings (Becker, 1975). In this case, any increased likelihood of abandoning a real option investment raises a concern among the employees about losing the value of their specific human capital investments. Employees’ specific human capital investments in a real option project can take many forms. For example, in an R&D project, they may involve knowledge or skills about a unique technology, experience with a unique procedure, or familiarity with specific systems and processes. Such tacit knowledge can be specific to the R&D project because the employees can’t fully transfer them to alternative uses if the project is abandoned, or if the employees switch companies. Consider Intel’s effort to develop the first semiconductor dynamic random access memory chip (DRAM), the 1 kilobit DRAM 1103, in 1970. The project resembled a real option investment in that there was a high level of uncertainty associated with it, but at the same time it had the potential to give the firm a first foothold in a new technology area, since at that time no other companies had developed such technology. But despite the project’s attractiveness, Intel’s engineers were concerned about the potential failure of the project and the associated devaluation of their knowledge and skills specific to DRAM technology. According to Gordon Moore, the CEO of Intel at that time, ‘There was a lot of resistance to semiconductor technology on the part of the core memory engineers. The engineers didn’t embrace the 1103 until they realized that it . . . wouldn’t make their skills irrelevant.’ (Cogan, 1989: 2–3) Specific human capital investments by employees may also take the form of relationship assets or employees’ knowledge about group culture and intrafirm or interfirm informal norms operating in a particular project setting (Dyer and Singh, 1998). projects. But to keep the arguments focused, we use employees as representative of all the relevant stakeholder groups. Copyright  2008 John Wiley & Sons, Ltd.

For example, in a joint venture project that heavily relies on collaborative work teams formed by employees from different parent companies, each employee’s relationship assets are likely to be project-specific. To the extent such relational assets are tied to the interfirm cooperative network, they would be valued less in any other setting were the joint venture project discontinued. So employees’ concerns about devaluation of their specific human capital investments can affect their incentives to make investments necessary for the success of the project.2 Our model implies that the effect of real options on employee incentives to invest in specific human capital can be either positive or negative. The effect is positive when the employee’s incentive to increase the likelihood of project continuation outweighs her concern for a possible loss of her investment, and it is negative when the opposite happens. When the effect on employee incentives is positive, there is clearly a further increase in real option value. On the other hand, when a real option project decreases employee incentives, the project value is negatively affected, sometimes more than offsetting the value of increased flexibility. To formalize these arguments, we develop a model to explicitly examine the role of employee incentives and how the incentives vary in a real option project as compared to a full-scale project. The model suggests that some features of a real option that make it more valuable in a strategic setting are also often inextricably tied to the employee incentive problem. Thus firms often face a trade-off between obtaining flexibility and promoting employee incentives.

BACKGROUND A real option investment provides a firm with flexibility in an uncertain environment and enables it to alter a course of action in the light of new information (Dixit and Pindyck, 1994). It is a right without an obligation to invest further depending on how uncertainty resolves. A simplified version of a real option investment can be described as a two-stage process (Figure 1). At t = 0, a small, initial, irreversible investment is made as a commitment that 2 Note that in real organizational settings firm-specific human capital is not always required to carry out a project and moreover, it does not always affect project outcome positively. This discussion, however, will focus on situations where specific human capital has a positive effect on project outcomes.

Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options Firm observes signal s ∈ {sH, sL} about the project’s cash flow t=1

t=0

703

Project’s cash flow x ∈ {xH, xL} is realized t=2 xH

• Favorable signal

sH observed • Firm invests in real

• Firm makes

option project I0

additional investment I1

• Employee selects a

xL

• Unfavorable

level of human capital investment i ∈ {iH, iL}

signal sL observed • Firm abandons

Loss of initial investment

option

Stage 1

Figure 1.

An illustration of a simple two-stage real option investment

gives a firm the right to participate in the project (i.e., the firm purchases the option). At t = 1, the firm receives some information regarding the profitability of the project and chooses either to make an additional investment or to abandon the project (i.e., the firm exercises the option). If the firm decides to continue the investment, the project’s cash flow is then realized at t = 2. While real options have been considered a promising approach for making strategic investments (e.g., Bowman and Hurry, 1993; Kogut, 1991; Kogut and Kulatilaka, 2001; McGrath, 1997, 1999) and real options logic has been applied to a variety of strategic issues in organizations (Kogut, 1991; Chi, 2000; Kumar, 2005; Reuer and Leiblein, 2000; McGrath, 1999; McGrath and Nerkar, 2004; Mitchell and Hamilton, 1988; Kogut and Kulatilaka, 2001), prior studies have largely overlooked the effect of a real option investment on the incentives of employees who are involved in the project. This is not too important when the initial investment involves only physical or financial assets and human factors are not crucial in determining the outcome. For example, a real estate company may purchase a piece of land as a form of real option but only invest further in its development should the real estate retail market turn more favorable. In this case, whether the real restate retail market becomes favorable or not is independent of the employees in the real estate company. After an initial investment is made, the company Copyright  2008 John Wiley & Sons, Ltd.

Stage 2

may not do anything with regard to the land. It waits until uncertainty resolves, and then decides whether or not to invest further. However, in many other cases, such as investments in interfirm collaboration or in developing new technology, key employees are often required to invest in human capital specialized to the project in order to carry on to the second stage. Then, the eventual success of the project is not solely determined by how uncertainties resolve themselves. It is also influenced by employees’ willingness to make investments in human capital specific to the project. Before the second stage of investment, the firm’s key employees may be required to work closely on the project and make specialized human capital investments that will increase the likelihood of the project being favorably evaluated at the second stage. In this case, the role of employee incentives becomes an important issue that is worth special attention when applying real options logic in decision making. Employee incentives to invest in specific human capital have been a subject of interest for organizational economics researchers for decades. A line of enquiry that sheds light on this area is the literature on the property rights theory of the firm, best exemplified by the seminal work of Grossman and Hart (1986) and Hart and Moore (1990). Their fundamental argument is that when contracts are incomplete, the incentive for a transaction partner to specialize may be reduced as there may Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

704

H. Wang and S. S. Lim

be a threat of a holdup by the other partner in terms of ex post rent/profit appropriation. Grossman and Hart (1986) and Hart and Moore (1990) propose ownership of a physical asset as a solution to incentive problems, as asset ownership provides the partner that makes specific investments more secure rent appropriation. More recent work in this area by Rajan and Zingales (1998, 2001) argues that without the ownership of physical assets, an employee may still have the incentive to invest in specific human capital, if doing so makes the employee indispensable to the firm, and thus gives her more power to appropriate rents. In the strategy literature, the works by Castanias and Helfat (1991, 2001) can be considered important and rare contributions to incentive issues in a firm. Castanias and Helfat’s focus, however, is placed on top managers of a firm. Specifically, they argue that managerial resources do not automatically generate superior economic value if effort and motivation are lacking or misdirected. They further analyze the ability of managers of a firm to appropriate rents from their superior human capital, which in turn affects their incentives to generate rents in the first place. None of these studies, however, have attempted to apply the issue of employee incentives in making specific investments to the context of real options investments. Part of the reason for such an omission might lie in the primary focus of these previous studies on the threat of holdups acting as deterrents to employee investment. They have, on the other hand, overlooked the role of other risks, such as strategic change or project termination, in affecting employee incentives. This study aims to examine the issues of employee incentives in the context of real options investments, because the risk of project termination, or other changes in strategy, is likely to be more prevalent in such a context. Specifically, when an employee makes human capital investments that are specific to a real option project of a firm, the value of these specific investments in the particular business setting will depend on whether the firm continues or abandons the project while exercising the option. Since it is difficult for the employee to apply the specific human capital investment to a different business setting without incurring significant loss, the risk of abandoning a project, which is more prevalent in a project invested as a real option, imposes significant costs on the employee (Cornell and Shapiro, 1987; Miller, 1998; Titman, 1984). Therefore in Copyright  2008 John Wiley & Sons, Ltd.

this case, the employee may underinvest in specific human capital. A few studies have pointed out the potential negative effects of real options on employee incentives. For example, Busby and Pitts (1997) surveyed the use of real options in firms and found that some managers voiced the concern that real option investments may signal to employees a lack of organizational support, and thus that employees involved in such projects may reduce their commitment to the firm. Consistent with this view, Barnett (2003) has argued that by investing in a project as a real option, the firm is declaring that it will wait and see how events unfold before fully committing essential resources. If the employees do not believe that a firm has yet placed its full faith in a project, they will in turn be less willing to whole-heartedly devote themselves. While these prior studies have taken important steps toward recognizing the incentive effect of investing in projects as real options, they have mainly relied on intuition and anecdotal evidence for their arguments. In addition, they have discussed only negative effects of real options on employee incentives but overlooked the possibility of positive effects. Our study contributes to the real options literature by being the first to develop a formal model that shows that real option investments can have both positive and negative effects on employee incentives and on project value. Further, we identify specific conditions under which one effect dominates the other, which should add much greater clarity to the role of employee incentives in evaluating the costs and benefits of real options.

THE MODEL The main argument of this study is that a real option investment brings a firm the benefit of flexibility but also alters employees’ incentives to invest in specific human capital, which in turn affects the value of the real option project. In this section, we develop a model to formalize this argument. The basic setup We model a real option project as a simple twostage investment. There are two periods (stages) and three dates (t = 0,1,2) in the model (Figure 1). Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options

705

Project’s cash flow x ∈ {xH, xL} is realized t=0

t=1

• Make investment

t=2

xH

I0 + I1 • Employee selects the

level of human capital investment i ∈ {iH, iL} xL

Figure 2. An illustration of a full-scale investment up front

At date 0, the firm starts the project by making an initial investment I0 . Although I0 is essential for the start-up of the real option project, this investment alone is insufficient to carry the project through to the second stage (t = 1). Also at date 0, a representative employee invests in human capital specialized to the project. The employee chooses to invest in a level of specific human capital, i, for the implementation of the project. For simplicity, assume that i takes a binary value, as is commonly done in similar models. The employee can choose either to make a high level of specific human capital investment (i = i H ), or to make a low level of specific investment (i = i L ). At date 1, the firm observes a noisy signal s ∈ {s H , s L } about the future cash flow from the project, and decides whether to continue the project, which requires an additional investment I1 , or to abandon it. If the project is continued, it generates cash flow x ∈ {x H , x L }, x H > x L > 0 at date 2; if the project is abandoned, the cash flow is zero. In contrast, the firm will invest a lump sum amount I0 + I1 at date 0 if it decides to make a full-scale investment up front.3 No decision regarding project abandonment is made at date 1, and the project generates the final cash flow of x ∈ {x H , x L } at date 2. An illustration of the full-scale investment is shown in Figure 2.

3 This suggests that there is no additional cost of investing in the real option itself. Alternatively, I0 can be considered as the cost of the option. In that case, the amount of full-scale investment at date 0 will be I1 , instead of I0 + I1 . Introducing the cost of the real option reduces the value of a real option project by a constant amount, but does not alter the key insights from the model.

Copyright  2008 John Wiley & Sons, Ltd.

Employee incentives to make specific human capital investments Because an employee’s investments in specific human capital are often unobservable, or unmeasurable even if they are observable, it is unlikely that firms can design a compensation contract that directly compensates the employee for making an appropriate level of specific human capital investment. A common solution to such a problem is to use an incentive system that rewards the employee based on performance: performancecontingent compensation (Clinch, 1991; Levinthal, 1988). Under such a compensation scheme, the employee is rewarded for making specific human capital investments by capturing a portion of the performance gains resulting from her specific human capital investments. The stronger the link between pay and performance, the more effective a pay-for-performance compensation scheme should be in inducing employees to make specialized human capital investments. However, an employee can also derive utility from rewards that go beyond monetary compensation. It is very often the case that the reward is implicit, in the form of promotion opportunities or increased employee bargaining power in the future. The employee may also be motivated intrinsically, by feelings of satisfaction and fulfillment on the successful completion of the project (Argyris, 1964; Calder and Staw 1975; Coff, 1997; Deci, 1975; Osterloh and Frey, 2000). Such intrinsic motivation may reflect the personalities of the employees themselves; it can also be the result of positive corporate social capital and effective management coordination and control (Bartlett and Ghoshal, 2002; Ghoshal and Bruch, 2003). Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

706

H. Wang and S. S. Lim

Regardless of the forms the rewards take, the utility the employee obtains from making specific human capital investments is likely to be positively related to the outcome of the project. In other words, the value of the specific human capital investment to the employee is higher when the project outcome is more favorable. If the project is abandoned, the value of the specific human capital may be significantly reduced if it is not wholly transferable to other settings. For now, assume the employee will obtain a payoff of zero if the project is abandoned. This corresponds to the extreme case in which the employee’s human capital is highly specific to the project. We will later relax this assumption and consider a more general case in which at least a proportion of the human capital can be transferred to other settings. Assume a utility function for the employee with the following form. U E = kx − i

(1)

where x is the project outcome and kx can be considered the utility the employee derives from the project. k describes how sensitive the employee’s utility is to the project outcome.4 This formulation captures the idea that the value of the specific human capital investment increases with the project outcome (x), and that the employee incurs a higher cost (i ∈ {i H , i L }, i H > i L ) when making a higher level of human capital investment specific to the project. It is worth noting that the relationships among real options investment, employee incentives, and project outcomes are not unidimensional. On the one hand, employee incentives (the decision to invest in either a high or low level of specific human capital) are assumed to affect the project’s outcome. On the other hand, the employees take into consideration the expected project outcome when they decide how much to invest in specific human capital. Through the dynamic interactions 4 If the employee derives utility from her specific human capital investment solely through monetary pay-for-performance compensation, it is possible that the firm could choose an optimal level of k using a contract to induce the desired level of employee effort. However, the implicit payoffs for employee specific human capital investment, such as feelings of satisfaction, promotion opportunities and bargaining power, are hard to contract for and may not involve direct costs to the firm. Since in reality such implicit factors often account for a substantial portion of the payoffs for employees’ investments, it is reasonable to assume that k is not contractible.

Copyright  2008 John Wiley & Sons, Ltd.

of employee incentives and the project outcome, whether a project is undertaken as a real option or as a full-scale project affects employee incentives. As employee incentives translate into overall project value, firms can better decide whether to invest in a project as a real option or as a full-scale investment by taking employee incentives into consideration. Our model incorporates the dynamics of these effects. Employee incentives and project outcomes We begin the analysis by examining the effect of employee incentives to invest in specific human capital on the outcome of the project, invested either full-scale or as a real option. The cash flow from the project is likely to be positively affected by the employee’s specific human capital investment in the project. In the context of our model, the likelihood of a positive project outcome, that is, the probability of the project cash flow being favorable, is likely to be higher when the employee makes a high (i H ) rather than a low (i L ) level of specific human capital investment. Thus, it is reasonable to assume that the probability of a positive project outcome conditional on employee specific human capital acquisition is such that: Pr(x H |i H ) = Pr(x L |i L ) = q > 0.5

(2)

We first examine the effect of employee incentives if the project is invested as a real option. For a real option project, part of the uncertainty regarding the project outcome is resolved at date 1. The firm receives a signal about the project cash flow, s ∈ {s H , s L }, which is informative but not conclusive about the outcome of the project. Pr(s H |x H ) = Pr(s L |x L ) = p > 0.5

(3)

The higher the p, the more informative is the signal regarding the project outcome. To make the real option investment nontrivial, assume that the parameter values are such that the firm decides to continue the project if it observes s H and to abandon the project if it observes s L .5 In other words, 5 If the firm continues the project regardless of the signal value, the project is equivalent to a full-scale investment project. If it is optimal for the firm to abandon the project regardless of the signal value, the firm would not make the initial investment at date 0.

Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options we consider the case where the expected cash flow from the project is greater (less) than the amount of additional investment required conditional on observing s H (s L ). From Equations 2 and 3, we derive the expected cash flow from the project conditional on the level of employee specific human capital investment as follows: E[x|i H ] = Pr(x H |i H )[Pr(s H |x H ) · x H + Pr(s L |x H ) · 0] + Pr(x L |i H ) × [Pr(s H |x L ) · x L + Pr(s L |x L ) · 0] = qpx H + (1 − q)(1 − p)x L E[x|i L ] = Pr(x H |i L )[Pr(s H |x H ) · x H

U (i H ) − U (i L )

× [Pr(s H |x L ) · x L + Pr(s L |x L ) · 0]

= k(E[x|i H ] − E[x|i L ]) − i H + i L

= (1 − q)px + q(1 − p)x

L

(4)

Compared to the case of a real option project described in Equation 4, the effect of employee incentives on project outcome is simpler if the project is invested full-scale at the start. In the case of a full-scale project, the signal about project outcome at date 1 (p) is irrelevant, since there is no option to abandon the project at this date. Therefore, from Equation 2, we derive the expected cash flow from the full-scale project conditional on the level of employee specific human capital investment as follows:

= qx H + (1 − q)x L E[x|i ] = Pr(x |i )x + Pr(x |i )x H

L

H

L

L

L

= (1 − q)x H + qx L

(5)

The effect of a real option investment on employee incentives Employee incentives to make specific human capital investments are a function of the expected project outcome, so a real option investment affects employee incentives through its impact on project outcomes. In other words, the flexibility associated with a real option investment affects employee incentives because it introduces the possibility that the project may be abandoned, the likelihood of Copyright  2008 John Wiley & Sons, Ltd.

(6)

Plugging Equation 4 into Equation 6, U (i H ) − U (i L ) = k(2q − 1)[px H − (1 − p)x L ] − i H + i L (7) Thus, the employee chooses to make a high level of specific human capital investment i H rather than a low level i L when: U (i H ) − U (i L ) = k(2q − 1) × [px H − (1 − p)x L ] − i H + i L > 0 k>

E[x|i H ] = Pr(x H |i H )x H + Pr(x L |i H )x L L

which is, to some extent, influenced by the level of the employee’s specific human capital investment. We evaluate the effect of a real option investment on employee incentives by comparing the employee incentives when the firm makes a twostage real option investment (Figure 1) with those when the firm makes a full-scale investment at the start (Figure 2). We first examine the case of a real option investment by deriving the condition under which the employee chooses to make a high level of specific human capital investment (i H ) rather than a low level (i L ). Given employee utility function U E = kx − i, the difference between the employee’s utility with a high level of specific human capital investment and that with a low level can be written as follows:

+ Pr(s L |x H ) · 0] + Pr(x L |i L ) H

707

iH − iL (2q − 1)[px H − (1 − p)x L ]

(8)

The smaller (greater) is the right- (left-) hand side, the larger is the parameter set that satisfies the above inequality, implying a greater likelihood of the employee’s investing in a high rather than a low level of human capital specific to the project. Inequality 8 suggests that the employee is more likely to make a high level of specific human capital investment when her utility is more closely tied to the project outcome (larger k), when the exercise decision about the real option at date 1 is based on a more informative signal (larger p), when the employee’s specific human capital investment is more likely to determine the project cash flow (larger q), and when the additional cost for making a high level of specific investment is smaller (smaller i H − i L ). These predictions are Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

708

H. Wang and S. S. Lim

quite intuitive: employees have greater incentives to invest in specific human capital when their utility is more sensitive to the project outcome (larger k). Furthermore, the more informative is the signal (larger p) at date 1 or the greater is the probability that a high level of specific human capital investment leads to a favorable cash flow (larger q), the greater impact that the employee’s action has on the outcome of the project, increasing the incentive for the employee to make a high level of specific human capital investment to prevent costly abandonment. Lastly, when investing in a high level of specific human capital is less costly (i H − i L smaller), the employee is more likely to do so. We summarize the effect of each factor while holding others constant in the following proposition. Proposition 1: The employee is more likely to make a high rather than a low level of specific human capital investment in a real option project when a. the employee’s utility is more sensitive to the project outcome, b. the option exercise decision at date 1 is based on a more informative signal, c. the employee’s specific human capital investment has a greater impact on the project outcome, and d. the additional cost of investing in a high level of specific human capital is smaller. In order to examine how a real option investment affects employee incentives compared to a full-scale investment, we need to compare the condition under which employees choose to make a high rather than a low level of human capital investment in a real option project with that if the project was invested full-scale up front. To do so, we next consider the employee’s human capital investment decision in the case of a full-scale project at date 0, i.e., when there is no option to abandon the project at date 1. Plugging Equation 5 into Equation 6 gives the difference in the employee’s utility as a result of high and low levels of specific human capital investment: U (i H ) − U (i L ) = k(2q − 1)(x H − x L ) − i H + i L Copyright  2008 John Wiley & Sons, Ltd.

(9)

Without the option to abandon the project, the employee has an incentive to invest more in specific human capital when: U (i H ) − U (i L ) > 0 ⇒

k>

iH − iL (2q − 1)(x H − x L )

(10)

Inequality 10 suggests that if the project is invested full-scale up front, the employee is more likely to make a bigger specific human capital investment when her utility is more closely tied to the project outcome (larger k), when her specific human capital has a greater impact on the project cash flow (larger q), and when the additional cost for making a higher level of specific investment is smaller (smaller i H − i L ). These conditions are similar to those for the case of a real option investment (Inequality 8) except that there is no p term in Inequality 10, since the signal that the firm receives at date 1 is irrelevant for a full-scale project. Further comparing inequalities 8 and 10, we obtain the conditions under which the likelihood of the employee investing in a high level of specific human capital is greater when the project is invested as a real option rather than a full-scale investment. In other words, we identify the conditions under which the right-hand side of Inequality 8 is smaller than the right-hand side of Inequality 10, as a smaller right-hand side value implies a higher probability that the employee will make a bigger investment in specific human capital. From Inequalities 8 and 10, investing in a real option rather than a full-scale investment leads to a higher likelihood of employees investing more in specific human capital when px H − (1 − p)x L > x H − x L ⇒

p xH > L 1−p x

(11)

and in contrast, it leads to a lower likelihood of employees investing when the opposite prevails. Inequality 11 suggests that a real option investment increases the likelihood of the employees making larger investments when p becomes larger and/or when the difference between x H and x L becomes smaller. To provide intuitions underlying the above results, we identify two opposing effects of a real Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options option investment on employee incentives. To the extent that specific human capital investments by the employees positively affect project outcomes, which in turn reduce the likelihood of the firm’s abandoning the project, real options can increase the employees’ incentives to make greater specific human capital investments. We term this positive employee incentive-inducing effect the ‘motivating effect’ of a real option investment. On the other hand, a real option can also decrease the incentive for employees to make specific human capital investments. In particular, to the extent the employee’s human capital is specific to the project, the employee may get zero or little utility from her specific human capital investment if the project is abandoned. Knowing such risks, the employee may be reluctant to make specific human capital investments. Since the source of this negative effect is the specificity of the employee’s human capital, we term it the ‘specificity effect’ of a real option investment. When the positive ‘motivating effect’ outweighs the negative ‘specificity effect,’ a real option investment increases the likelihood that the employee invests in a high level of specific human capital compared to a full-scale investment. In Inequality 11, a larger p means that the signal received at date 1 is more informative, or there is a lower level of persistent uncertainty about the project outcome at the time the option is exercised. An informative signal implies that a high level of project cash flow is more likely to lead to project continuation, allowing the employee to capture some of the rent generated from her human capital investment. In contrast, when p is relatively small, the firm must decide whether or not to continue the project based on a noisy signal. This decreases employee incentives, because there is more uncertainty as to whether the employee’s investment will lead to project continuation. For example, the firm may observe s L and abandon the project even though the true cash flow is x H . Therefore, a larger p reduces the employee’s concern about losing the value of her investment due to project abandonment based on the wrong signal. In sum, when p becomes larger, the positive ‘motivating effect’ is more likely to dominate the negative ‘specificity effect’ and a real option investment results in a greater likelihood of the employee making a higher level of employee specific human capital investment. Copyright  2008 John Wiley & Sons, Ltd.

709

Proposition 2a: When the real option exercise decision is based on a more informative signal about the project outcome, a real option project is more likely to increase employee incentives to make a high level of specific human capital investment compared to a full-scale project. Inequality 11 also suggests that compared to a full-scale project, a real option project decreases the likelihood that employees will make large specific human capital investments when the gap between the two different levels of projected cash flows of the project (x H and x L ) increases. The reason is that under a full-scale project, the only uncertainty in the relationship between the employee specific human capital investment and her expected utility is the probability (q) that a high (low) level of specific human capital acquisition leads to a high (low) level of cash flow. In a real option investment, however, the possibility that the firm may abandon the project at date 1 accounts for an additional source of uncertainty in the employee’s expected payoff (which is relevant to p). This additional uncertainty changes the incremental effect of specific human capital investment on project outcome and, thus, the expected payoff to the employee. To better illustrate these points, consider two extreme cases. The first is where the difference between x H and x L is very small. In this case, the additional expected employee payoff from investing in a high level of specific human capital compared to investing in a low level is minimal for a full-scale investment, because the additional payoff comes solely from the difference between x H and x L . But the amount of additional expected payoff is much larger in the case of a real option project, because the additional payoff is not only determined by the difference between x H and x L , but also by the difference in the payoffs in the cases of project continuation and abandonment. Since the employee’s expected payoff is zero at the termination of the project and the probability of project continuation is higher when she makes a bigger specific human capital investment, a real option project is more likely to induce a high level of employee specific investment compared to a fullscale project when the difference between x H and x L is small. In the other extreme case, suppose x H is very large relative to x L (x H  x L , or x L = 0). If a high level of cash flow is realized, this high upside Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

710

H. Wang and S. S. Lim

potential is always captured in a full-scale investment. But it is captured only when the firm continues the project in a real option investment, which happens with a probability p. Thus, a project invested as a real option is likely to result in lower incentives for employees to acquire specific human capital compared to a full-scale project as the difference between x H and x L becomes larger. The above discussion provides additional intuition underlying Inequality 11, and can be summarized in the following proposition. Proposition 2b: When there is a large variation in possible project outcomes, a real option project is more likely to reduce employee incentives to make a high level of specific human capital investment compared to a full-scale project. Propositions 2a and 2b suggest that employee incentives are more likely to be negatively affected in a real option project when the signal that the firm receives at the time of option exercise is not sufficiently informative and when there is a large gap between the project’s upside and downside outcomes, in both cases the uncertainty remains largely unresolved at the time of option exercise. It is worth noting that this is more likely to happen for investments in human capital-based assets or core competences, due to the inherent causal ambiguity and social complexity associated with such investments (Coff and Laverty, 2001; Adner and Levinthal, 2004). For example, Coff and Laverty (2001) argue that for human capital-based assets, the signals about the project outcome are likely to be imperfect, or even conflicting. In addition, human capital-based strategic assets or core competences are also likely to have a greater level of variation in terms of their future outcomes, which adds additional uncertainty about the project’s potential outcomes. Therefore, the concern about reduced employee incentives is more severe for investments in human capital assets or core competences, situations to which the real options reasoning is commonly applied (e.g., Kogut and Kulatilaka, 2001). Real option investment and project value The discussion so far demonstrates how a project invested as a real option alters employee incentives compared to one invested as a full-scale project. Of ultimate interest, however, is how this change Copyright  2008 John Wiley & Sons, Ltd.

in incentives translates into overall project value (and therefore firm value), so that firms can better decide whether to invest in a project as a real option or as a full-scale investment. Lemma 1: Under the assumption of the model, the value of a real option project is greater than that of an otherwise identical full-scale project. Proof: See Appendix. Lemma 1 indicates that in the context of this model, when everything else is equal (i.e., given the same level of employee incentives), a project invested as a real option has a greater value than if it is invested full-scale. This result is quite intuitive since a real option investment takes advantage of the flexibility based on the additional information at date 1. Thus in a case in which real option increases employee incentives, the incentive effect enhances the value of the real option project even further. However, the effect becomes less straightforward when investing in a real option reduces employee incentives compared to a full-scale investment. The net effect will be negative when the value loss due to lower employee incentives is greater than the benefit of increased flexibility, and the net effect will be positive when the opposite holds. Lemma 2: Under the assumption of the model, the value of a real option project with a low level of employee specific human capital investment is smaller than the value of the project if invested full-scale with a high level of employee human capital investment. Proof: See Appendix. Lemma 2 implies that the value loss due to reduced employee incentives outweighs the value gain due to increased flexibility in our model, resulting in a net negative effect of the real option on project value. The model specifies, therefore, that the conditions for a real option to lead to a higher/lower project value compared to a fullscale project are simplified to the conditions for the real option to induce high/low levels of employee investment. Based on Lemmas 1 and 2, Propositions 2a and 2b can be rewritten in terms of project values as in the following proposition. Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options Proposition 3: When the signal at the time of option exercise (date 1) is not very informative about the project outcome, or when there is a large variation in the possible project outcomes, a real option investment is more likely to result in a lower project value compared to a full-scale investment. Proposition 3 suggests that the negative effect of reduced employee incentives can be large enough that the net value of a real option project is lower than that of a full-scale project.

EXTENSIONS AND IMPLICATIONS The discussion so far has considered situations in which a firm invests in a simple two-stage real option where the initial investment (I0 ) provides the firm with the exclusive right to exercise the option in the future. Moreover, the project is assumed to be implemented by a single representative employee. However, in real investment decisions, especially in complex decision making regarding investment in human capital-based assets and core competences, it is unlikely that such simple assumptions always hold. It may be useful, therefore, to examine how variations in the conditions and assumptions of the model may affect its implications. In this section, we consider several possible extensions of the model and discuss their implications. Employee specific human capital with a positive alternative value We have assumed that an employee obtains a payoff of zero from her specific human capital investment if the project is abandoned at date 1. It describes an extreme case where the human capital acquired by the employee is highly specific to the existing project, and thus not even a small portion can be transferred to other settings. In reality, however, it is more often the case that employee human capital specialized to a current project can, at least partially, be applied to a different business setting in the event of the abandonment of the current project. The loss to the employee in the event of project abandonment is thus reduced to the extent that the human capital acquired in implementing the project can be applied to other projects. Copyright  2008 John Wiley & Sons, Ltd.

711

Some previous research sheds light on the possible mechanisms through which employee human capital may be transferred within the firm (Cohen and Levinthal, 1990; Dierickx and Cool, 1989; Markides and Williamson, 1994). For example, Dierickx and Cool (1989) argue that firm capabilities are often accumulated internally, building upon the existing human capital and asset stock, suggesting that the development of new firm human capital is often closely related to the human capital used to implement current projects. Cohen and Levinthal (1990) further argue that a firm’s ability to develop new capabilities is in part determined by the firm’s absorptive capacity, which is greatest when what is to be learned is related to what is already known. This again implies that a firm is likely to be more successful in acquiring new human capital if it is built upon the human capital acquired from previous projects. Similarly, in a diversification setting, Markides and Williamson (1994) argue that in creating strategic assets to support a new line of business, existing strategic human capital assets and those involved in the new business often enhance each other, creating an economy of scope. Therefore, a more general analysis requires relaxing the assumption of zero expected employee payoff in case of project abandonment. Suppose the payoff to the employee at the time of abandonment depends on the level of specific human capital acquired by the employee, that is, it is r H if the level of specific human capital acquired by the employee is high (i = i H ) and r L if the level is low (i = i L ).6 Taking these additional expected payoffs into consideration in Equation 4, the difference in the employee’s utility under high and low levels of specific human capital investment becomes E[x|i H ] = Pr(x H |i H )[Pr(s H |x H ) · x H + Pr(s L |x H ) · r H ] + Pr(x L |i H )[Pr(s H |x L ) · x L + Pr(s L |x L ) · r H ] = qpx H + (1 − q)(1 − p)x L + [q(1 − p) + (1 − q)p]r H 6 The results are very similar under an alternative assumption that the alternative payoff to the employee depends on the project outcome, or the projected cash flow (r = r H when x = x H and r = r L when x = x L ): When r H is sufficiently large compared to r L , the incentive for the employee to make a high level of specific human capital investment is greater in the case of a nonzero alternative payoff than that of a zero alternative payoff. (Detailed proof is available upon request.)

Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

712

H. Wang and S. S. Lim

L

E[x|i ] = Pr(x H |i L )[Pr(s H |x H ) · x H + Pr(s L |x H ) · r L ] + Pr(x L |i L )[Pr(s H |x L ) · x L + Pr(s L |x L ) · r L ] = (1 − q)px H + q(1 − p)x L + [(1 − q)(1 − p) + qp]r L ⇒ U (i H ) − U (i L ) = k(E[x|i H ] − E[x|i L ]) − i H + i L = k(2q − 1)[px H − (1 − p)x L )] + [q(1 − p) + (1 − q)p]r H − [(1 − q)(1 − p) + qp]r L − i H + i L

(12)

Comparing Equation 12 with Equation 7, allowing nonzero payoffs at the abandonment of the project increases the employee’s incentives to make a high level of firm-specific human capital investment compared to the case of zero-payoffs at abandonment if [q(1 − p) + (1 − q)p]r H > [(1 − q)(1 − p) + qp]r L .

Proposition 4: Compared to a real option project with zero employee alternative payoff at the time of project abandonment, a real option project with a positive alternative payoff results in a stronger incentive for employees to make a high level of specific human capital investment and leads to a higher project value when the alternative employee payoff from a high level of specific human capital investment is sufficiently large compared to that from a low level of specific human capital investment.

The presence of the risk of preemption (13)

In general, when the employee’s payoff at project abandonment from a high level of specific human capital is sufficiently large compared to the payoff from a low level of specific human capital, i.e., r H  r L , or when the payoff from a low level of specific human capital is minimal, i.e., r L ≈ 0, the left-hand terms are greater than the right-hand terms. In these cases, the presence of a positive payoff to the employee at the time of project abandonment increases employee incentives to invest in specific human capital. In the other extreme case where the alternative value of specific human capital does not depend on whether the employee has invested in a high or low level of specific human capital, that is, r H ≈ r L , the presence of a positive employee payoff is likely to decrease employee incentives, since q(1 − p) + (1 − q)p < (1 − q)(1 − p) + qp with p > 0 .5 and q > 0 .5 . In real firm settings, it is reasonable to expect that the alternative payoff is likely to be higher when the employee obtains a high level rather than a low level of specific human capital. Therefore, to the extent that an employee’s investment in specific human capital retains part of its value when Copyright  2008 John Wiley & Sons, Ltd.

the project is abandoned and the retained value increases with the amount of specific human capital investment, it reduces the employee’s concern about losing the value of its specific human capital investment. When this reduced concern leads to a greater likelihood of investing in a high level of specific human capital, the overall value of the real option project is further enhanced. Therefore, we have the following proposition.

Another factor relevant to investments in real option projects is the firm’s competitive environment. The implicit assumption underlying our model is that the initial investment gives the firm full exclusivity in exercising the option in the future. However, competitive forces may make it unlikely in reality that a firm can obtain a fully exclusive right to exercise the option. Although the right to exercise a real option may to some extent be safeguarded by patents, legal contracts, or enforceable property rights, these legal protection mechanisms are not always available or sufficiently effective (e.g., Levin et al., 1987). Therefore, because the option to make additional investment in the future can be ‘killed’ by another competitor’s first move, the value of an option to delay an investment can be eroded when there is the risk that competitors may preempt (e.g., Weeds, 2002). This circumstance can result in two possible outcomes depending on whether the firm is able to directly react to the competitor’s preemption. First, in order to effectively preempt potential competitors, a firm invested in a real option may be able to react by exercising the option early, before the information about the project outcome is known to outsiders (Folta and Miller, 2002; Kulatilaka and Perotti, 1998). An early option exercise, however, Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options is often associated with insufficient information that is revealed at the time of the exercise. In other words, when a firm exercises an option early, it needs to base its decision with regard to the project outcome on a noisier signal (lower p). When the employee is aware of the presence of competitor preemption, and thus a possibility of an early option exercise, she will expect a lower value of p, which in turn has a negative effect on her incentives to invest in specific human capital at date 0 as predicted by Proposition 2a. The second possibility is that the firm simply bears the risk of competitor preemption, perhaps because it cannot advance the timing of the option exercise. In this case, we can extend the model to examine the effect of preemption risk on employee incentives and project value by introducing a positive probability (m) that the option becomes obsolete at the time of exercise due to competitor preemption, leading to a cash flow of zero. This implies that the expected employee payoff from the specific investment is scaled by (1 − m), because the project cash flow is realized with a probability (1 − m) conditional on the project’s continuation. The condition under which the employee invests in a high level of specific human capital then becomes U (i H ) − U (i L ) = k(2q − 1)(1 − m)[px H − (1 − p)x L ] − iH + iL

(14)

Compared to Equation 7, the above inequality indicates that a positive probability of the option becoming obsolete (m > 0 ) results in a smaller right-hand side, which decreases the likelihood that the employee invests in a high level of specific human capital in a real option investment. As discussed in an earlier section, the lower employee incentives imply a smaller project value. Therefore, when there is a risk of competitor preemption, regardless of whether the firm reacts to it by deciding to exercise the option early or by bearing the risk of preemption directly, employee incentives to invest in human capital and project value are negatively affected. We thus have: Proposition 5: When there is intense competition and/or weak property rights protection of a firm’s real option investment, employees have Copyright  2008 John Wiley & Sons, Ltd.

713

weaker incentives to invest in a high level of specific human capital, and this leads to a lower project value.

Employees in a team setting The model so far has examined a simplified setting in which the project is assumed to have only one representative employee. Normally, however, projects involve many employees. In a team setting, an individual employee’s willingness to make specific human capital investments is often influenced by the extent to which the project’s success depends on effective coordination among the team members and by how each employee evaluates the incentives of the other members of the team to make specific investments. We now extend the model to incorporate such group dynamics. Assume a team of two employees, A and B, works on the project.7 The outcome of the project will be influenced by their joint efforts and by both employees’ specific human capital investments. As in the original single-employee model, each employee can choose to make either a high or a low level of specific human capital investment (employee A: iA = iAH or iAL ; employee B: iB = iBH or iBL ). The employees’ utility functions and the signal that the firm observes at date 1 are UAE = kA x − iA ,

UBE = kB x − iB

(15)

Pr(s |x ) = Pr(s |x ) = p > 0.5

(16)

H

H

L

L

The probability of the project having a positive outcome (x H ) depends on the combination of the two employees’ specific human capital investments: Pr(x H |iAH , iBH ) = qH H ; Pr(x H |iAH , iBL ) = Pr(x H |iAL , iBH ) = qH L ; Pr(x H |iAL , iBL ) = qLL qH H ≥ qH L ≥ qLL

(17)

These probabilities are able to capture a wide range of group dynamics. For example, when cooperation among team members is extremely 7 The two-employee setting captures the interaction among employees without unnecessary complications. The insights can easily be extended to larger teams.

Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

714

H. Wang and S. S. Lim

important for project success (i.e., the investments of the two employees are highly complementary), it is likely that qH H > qH L ≈ qLL . In other words, the probability of a positive project outcome increases only if both employees make a high level of specific investment. On the other hand, when only one employee’s investment is needed to carry out the project (i.e., employee investments are highly substitutive), it is likely that qH H ≈ qH L > qLL . Now consider employee A’s decision (B’s decision is analogous). Assume that A assigns a probability of h ∈ [0, 1] to B’s investing in a high level of specific capital (and thus (1 − h) to B’s investing in a low level). In addition to the probability distributions (qH H ≥ qH L ≥ qLL ) as explained above, we further allow h to vary in order to capture additional team dynamics. Various factors may affect the probability h. For example, if employee B’s investment is largely in tacit knowledge, B’s incentive will not be easily measured or evaluated (Grant, 1996; Osterloh and Frey, 2000; Polanyi, 1966). On the other hand, employee A is likely to be better informed about B’s willingness to invest when there is effective communication and coordination in the team, or when there is a close personal relationship and trust between the two team members. In particular, h = 1 or h = 0 would indicate that A knows precisely how B will invest. h = 0.5 would indicate that A has no idea what B is going to do, so she assigns equal probabilities to high and low levels of investment. By allowing q (qH H , qH L , and qLL ) and h to vary, the model can be applied to teams in a broad range of settings. The probability distribution of project outcomes given the level of specific investment of employee A becomes Pr(x H |iAH ) = qH H · h + qH L · (1 − h); Pr(x H |iAL ) = qH L · h + qLL · (1 − h)

(18)

Plugging these probabilities into Equation 4 to obtain the expected project outcome given the specific investment level, we have the following condition under which employee A chooses to make a high level of specific human capital investment in a real option project. U (iAH ) − U (iAL ) = kA [qH H h + qH L (1 − 2h) − qLL (1 − h)] × [px H − (1 − p)x L ] − iAH + iAL > 0 Copyright  2008 John Wiley & Sons, Ltd.

−−→ kA >

iAH − iAL (19) [(qH L − qLL ) + h(qH H + qLL −2qH L )][px H − (1 − p)x L ]

Inequality 19 shows that several predictions of Proposition 1 (a, b, d) still hold in a team setting: an employee in a team setting is more likely to make a high level of specific human capital investment in a real option project when the employee’s utility is more sensitive to the project outcome (a larger k), the option exercise decision at date 1 is based on a more informative signal (a larger p), and the additional cost of investing in a high level of specific human capital is smaller (a smaller (i H − i L )). Predictions about the impact of employee specific investment on the project outcome are more complicated. In the single employee setting, Inequality 8 indicates that an employee’s incentive to make a high level of specific investment increases with q, the probability of a positive project outcome conditional on the high level of investment. But other factors need to be considered in the two-person case. From Inequality 19, employee A is more likely to invest in a high level of specific human capital when (qH L − qLL ) + h(qH H + qLL − 2qH L ) is greater, which can be rearranged as (qH L − qLL ) + h[(qH H − qH L ) − (qH L − qLL )] = h(qH H − qH L ) + (1 − h)(qH L − qLL ) (20) Here, (qH L − qLL ) can be interpreted as the incremental benefit (i.e., the increase in the probability of a good outcome) of employee A’s making a high level of investment when employee B makes a low level of investment; and (qH H − qH L ) is the incremental benefit when employee B makes a high level of specific investment. Since h is between 0 and 1, the right-hand side of Equation 20 indicates that employee A has a greater incentive to invest in a high level of specific human capital when the incremental benefit of a high level of investment, (qH L − qLL ) or (qH H − qH L ), increases. Intuitively, the employee has a greater incentive to invest in a high level of specific human capital if doing so has a positive effect on the project outcome. The left-hand side of Equation 20 implies that the effect of h on employee incentives depends on the relative size of these two incremental benefits. If (qH H − qH L ) > (qH L − qLL ), Equation 20 increases with h, suggesting that employee A has Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options a greater incentive when she assigns a greater probability of employee B investing in a high level of specific human capital. On the other hand, Equation 20 decreases with h when (qH H − qH L ) < (qH L − qLL ), implying that the incentive is lower when employee A assigns a greater probability of employee B investing in a high level of specific human capital. We can interpret the two scenarios described above in terms of the complementarity and substitutability of the two employees’ specific investments in a team setting. (qH H − qH L ) > (qH L − qLL ) means that the incremental benefit of one employee investing in a high level of specific human capital is greater when the other team member also invests in a high level, implying that the two employees’ investments are complementary. Therefore, the incentive for an employee to invest in a high level of specific human capital is greater when there is a higher probability that others will also invest in a high level. In contrast, (qH H − qH L ) < (qH L − qLL ) means that the incremental benefit of an additional investment is smaller, implying that the two employees’ investments are substitutive. In this case, the investment incentive of each employee decreases when there is a higher probability that the other will invest in a high level of specific human capital. The following proposition summarizes these effects. Proposition 6: In a team setting, each employee is more likely to make a high rather than a low level of specific human capital investment in a real option project when a. the employee’s utility is more sensitive to the project’s outcome, b. the option exercise decision at date 1 is based on a more informative signal, c. the additional cost of investing in a high level of specific human capital is smaller, d. the employee’s specific human capital investment has a greater incremental impact on the project outcome, holding the other employee’s investment constant, and e. the employee assigns a higher (lower) probability to the other employee investing in a high level of specific human capital when the investments of the two employees are complementary (substitutive). Copyright  2008 John Wiley & Sons, Ltd.

715

Predictions a–d are very similar to those of Proposition 1. Prediction e has interesting implications for teamwork and employee motivation. There are different and sometimes even opposing arguments with regard to whether teamwork increases productivity. For example, the social facilitation argument developed in social psychology suggests that as a result of an individual’s motivation to maintain a positive self-image in the presence of others, she may work harder in a team compared to working alone (Guerin, 1986; Harkins, 1987; Wicklund, 1975). On the other hand, one stream of the economics literature emphasizes the free-rider effect: when an individual thinks that other team members have taken sufficient action to achieve the team’s objective and thus his/her own actions become less necessary, she is likely to free-ride and thus be less motivated to contribute to the team production (Holmstrom, 1982; Kerr, 1983). Our model provides a unified framework that helps integrate these two opposing effects. The free-riding effect is likely to dominate when the employee’s skills are substitutive: the employee has less incentive to make a high level of specific human capital investment when other employees are likely to make high levels of investments. On the other hand, when the skills of team members are complementary, the employee has more incentive to make a high level of investment when she believes that other employees are also investing unreservedly. The resulting increase in employee productivity in a team setting is consistent with that predicted by the social facilitation argument, although the underlying mechanisms that lead to the conclusion might be different. We now move on to examine, in a team setting, the effect on employee incentives of investing in real options relative to investing in fullscale projects. If the project is invested full-scale, employee A will choose to make a high level of specific investment if U (iAH ) − U (iAL ) = kA [qH H h + qH L (1 − 2h) − qLL (1 − h)] × (x H − x L ) − iAH + iAL > 0

(21)

The comparison of Inequalities 19 and 21 shows that an employee is more likely to invest a high level of specific human capital in a real option Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

716

H. Wang and S. S. Lim

project rather than in a full-scale project when px H − (1 − p)x L > x H − x L ⇒

p xH > L 1−p x

This is exactly the same condition as Inequality 11, which was derived for a single employee. Therefore, the key results shown in Propositions 2a and 2b should still hold in the extended model. Although introducing team dynamics changes the mapping between employee specific human capital investment and project outcomes, it does not influence the precision of the signal at date 1 and the variability of future project outcomes. Since these two factors are unaffected by group dynamics yet determine the incremental effect on employee incentives of a project invested as a real option, it is not surprising that Propositions 2a and 2b continue to hold in a team setting. In summary, although team dynamics does affect employee investment incentives, it influences them in the same way as in a full-scale or a real option project. As a result, the incremental effect on employee incentive of a real option project over a full-scale one is not affected by whether the employee works alone or in a team setting.

DISCUSSION AND CONCLUSIONS Real options logic is often considered a superior approach to investment decision making, since it may sometimes justify investments that would be rejected under the traditional net present value approach. However, real option investments may be less justifiable when their impacts on employee incentives are taken into consideration. Many strategic decisions about such matters as R&D, the formation of multinational networks, joint ventures, or the development of a firm’s human capital-based core capabilities or competences are often considered crucial to firm profits and competitive advantage (e.g., Barney, 1991; Peteraf, 1993; Kogut and Zander, 1992). The effective implementation of such strategies, however, can rarely succeed without the active involvement of the firm’s key personnel. So if strategic investments are to be viewed as real options, employee incentives to invest in specific human capital should be an important consideration in evaluating such real option investments. Copyright  2008 John Wiley & Sons, Ltd.

This article seeks to develop a parsimonious model that demonstrates how employee incentives are affected by firms’ investments in real options, and how the incentives in turn influence the overall value of such real option investments. Several implications can be drawn from the results of the model. First, the very flexibility that allows a firm to benefit from changing its investment plans in response to new information also affects employee incentives to invest in specific human capital necessary to implement the real option project. When the incentive effect is negative, the benefit from the increased flexibility may not be sufficient to offset the negative effect on employee incentives. Thus, employee incentive considerations can rebalance the benefits and costs of real options, and failing to consider such incentive effects could result in suboptimal strategic decisions for firms. Second, employee incentives are most negatively affected when great uncertainty remains at the time when the option must be exercised, for example, when the signal that the firm receives at the time of exercise is not sufficiently informative or when there is a large gap between the project’s upside and downside outcomes. Interestingly, these conditions are most likely to apply to investments in human capital-based strategic assets or core competence, investments to which the real options reasoning is commonly applied (e.g., Kogut and Kulatilaka, 2001). This suggests that strategy researchers should use special caution when promoting the real options approach to strategic decision making. This study emphasizes the importance of incorporating employee incentives into strategic analysis. Research in strategic management, especially studies based on the resource-based and knowledge-based views of the firm (Barney, 1991; Peteraf, 1993; Kogut and Zander, 1992, 1996), explicitly or implicitly places human resources among the most important resources available to any firm, and central to any debate about how firms achieve competitive advantage (Coff, 1999; Snell, Shadur and Wright, 2001). Therefore, employee incentives to invest in firm-specific human capital should be an important area of study for strategic management scholars. By applying the issue of employee incentives to the context of real options investments and drawing some important strategic implications, this study reemphasizes the importance of incorporating employee incentive considerations into strategic decision making. It is Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options important to note that real options investment is not the only strategic setting where employees’ concerns about investing in firm-specific human capital may be applied. Related arguments may also be applied to some other setting as long as employees anticipate that their investments in specific human capital may be potentially at risk. For example, in a setting where a firm frequently changes its strategic directions and adjusts its business portfolios, the employees’ concern of losing their specific human capital may also be relevant. It is hoped that this study will constitute an important step toward a better understanding of the effect of employee incentives to invest in specific human capital on firm strategy and performance. Of course, the model developed here has some inherent limitations that require further research to advance its arguments. This parsimonious model has allowed us to examine some key factors with greater clarity, but it builds on assumptions that may limit its applicability in certain situations. In addition to the possible variations of the model that have already been discussed, there is still much room for future research to further refine the model and draw potentially richer implications. For example, models that incorporate more complicated real options investments, such as compounded options with more than two stages (e.g., Kim and Kogut, 1996; Miller and Folta, 2002) or multiple interrelated real options projects invested as bundles (e.g., McGrath, 1999; Vassolo, Anand, and Folta, 2004), can be explored. Future research might also consider the situation in which the project becomes increasingly embedded in the firm during the implementation process (Coff and Laverty, 2001). In this case, the firm’s decision to continue or abandon the project also evolves over time, since the abandonment cost would vary with the project implementation process. Future research may also consider the possibility of incorporating the resolution of endogenous uncertainty. In the context of this article, some of the uncertainty may be resolved endogenously either through investment by the firm or through employee learning. We expect that incorporating the resolution of endogenous uncertainty is likely to increase the value of real options because learning by the firm or employees reduces the uncertainty of the project. A more thorough analysis of the resolution of endogenous uncertainty, which we think may be made possible in a multiple-stage investment model, can be a Copyright  2008 John Wiley & Sons, Ltd.

717

promising future research direction. The model may also be extended to allow the firm to use compensation contracts that maximize the value of real options. Due to the influence of intrinsic motivation and other nonmonetary compensation on employee incentives, it would be impossible for the firm to perfectly align employee incentives through compensation contracts. Nevertheless, the firm may be able to design contracts (albeit still incomplete) that maximize the value of real options, taking into consideration the ambiguity of the link between compensation contracts and employee incentives. In addition, although the model as it stands allows for the possibility of implicit motivation and managerial coordination and control, these are not explicitly modeled in great detail. Future research may help further our understanding in this area by examining how implicit motivation and managerial control may, either independently or through interaction with employee incentives, affect real options investments. Lastly, the arguments made here may be extended from the incentives of employees (or other stakeholders) to those of the decision makers, since firm executives or project managers may also be required to make specific investments in company projects. The incentives of decision makers, however, may differ from those of other employees, since they are also involved in making decisions regarding the timing of the option exercise and whether to terminate or continue the project. To the extent that managers’ incentives may be misaligned with shareholders’ interests in making these decisions, the classic agency problem is also relevant to the analysis of real option investments. These considerations may be combined with the insights developed here about employees’ incentives to make specific investments to yield a more comprehensive understanding of real options.

ACKNOWLEDGEMENTS The authors would like to thank two anonymous reviewers for their constructive comments, and Jinyu He and Jiatao Li for their comments on an earlier version of the paper. We especially would like to express our gratitude to the following three people for their instrumental guidance and invaluable support: Jay Barney, Kent Miller, and Ed Zajac. Thanks also to KOZHIKODE Rajiv Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

718

H. Wang and S. S. Lim

Krishnan, who provided us with excellent research assistance.

REFERENCES Adner R, Levinthal DA. 2004. What is not a real option: considering boundaries for the application of real options to business strategy. Academy of Management Review 29: 74–85. Amit R, Schoemaker PJH. 1993. Strategic assets and organizational rents. Strategic Management Journal 14(1): 33–46. Argyris C. 1964. Integrating the individual and the organization. John Wiley & Sons: New York. Barnett ML. 2003. Falling off the fence? A realistic appraisal of a real options approach to corporate strategy. Journal of Management Inquiry 12: 185–196. Barney JB. 1991. Firm resources and sustained competitive advantage. Journal of Management 17: 99–120. Bartlett CA, Ghoshal S. 2002. Building competitive advantage through people. Sloan Management Review 43(2): 34–41. Becker GS. 1975. Human capital: a theoretical and empirical analysis with special reference to education. Columbia University Press, National Bureau of Economic Research: New York. Bowman E, Hurry D. 1993. Strategy through the options lens: an integrated view of resource investments and the incremental-choice process. Academy of Management Review 18: 760–782. Busby JS, Pitts CG. 1997. Real options in practice: an exploratory survey of how finance officers deal with flexibility in capital appraisal. Management Accounting Research 8: 169–186. Calder BJ, Staw BM. 1975. Self-perception of intrinsic and extrinsic motivation. Journal of Personality and Social Psychology 31: 599–605. Castanias RP, Helfat CE. 1991. Managerial resources and rents. Journal of Management 17(1): 155–171. Castanias RP, Helfat CE. 2001. The managerial rents model: theory and empirical analysis. Journal of Management 27(6): 661–678. Chi T. 2000. Option to acquire or divest a joint venture. Strategic Management Journal 21(6): 665–687. Clinch G. 1991. Employee compensation and firms’ research and development activity. Journal of Accounting Research 29: 59–78. Coff RW. 1997. Human assets and management dilemmas: coping with hazards on the road to resource-based theory. Academy of Management Review 22: 374–402. Coff RW. 1999. When competitive advantage doesn’t lead to performance: the resource-based view and stakeholder bargaining power. Organization Science 10: 119–213. Coff RW, Laverty K. 2001. Strategy process dilemmas in exercise decisions for options on core competencies. Paper presented at the Academy of Management Conference, Washington D.C. Copyright  2008 John Wiley & Sons, Ltd.

Cogan GW. 1989. Intel Corporation (A): The DRAM decision. (S-BP-256) Stanford University Press: Stanford, CA. Cohen WM, Levinthal DA. 1990. Absorptive capacity: a new perspective on learning and innovation. Administrative Science Quarterly 35: 128–152. Cornell B, Shapiro AC. 1987. Corporate stakeholders and corporate finance. Financial Management 16(1): 5–14. Deci EL. 1975. Intrinsic Motivation. Plenum Press: New York. Dierickx I, Cool K. 1989. Asset stock accumulation and the sustainability of competitive advantage. Management Science 35: 1504–1511. Dixit AK, Pindyck RS. 1994. Investment Under Uncertainty. Princeton University Press: Princeton, NJ. Dyer J, Singh H. 1998. The relational view: cooperative strategy and sources of interorganizational competitive strategy. Academy of Management Review 23: 660–679. Folta TB, Miller KD. 2002. Real options in equity partnerships. Strategic Management Journal 23(1): 77–88. Ghoshal S, Bruch H. 2003. Going beyond motivation to the power of volition. Sloan Management Review 44(3): 51–57. Grant RM. 1996. Toward a knowledge-based theory of the firm. Strategic Management Journal Winter Special Issue 17: 109–122. Grossman S, Hart OD. 1986. The costs and benefits of ownership. Journal of Political Economy 97(4): 691–719. Guerin B. 1986. Mere presence effects in humans: a review. Journal of Experimental and Social Psychology 22: 38–77. Harkins SG. 1987. Social loafing and social facilitation. Journal of Experimental Social Psychology 23: 1–18. Hart OD, Moore J. 1990. Property rights and the nature of the firm. Journal of Political Economy 98(6): 1119–58. Holmstrom B. 1982. Moral hazard in teams. Bell Journal of Economics 13: 324–340. Kerr NL. 1983. Motivation losses in small groups: a social dilemma analysis. Personality and Social Psychology 45: 819–828. Kim D, Kogut B. 1996. Technological platforms and diversification. Organization Science 7: 283–301. Kogut B. 1991. Joint ventures and the option to expand and acquire. Management Science 37: 19–33. Kogut B, Kulatilaka N. 2001. Capabilities as real options. Organization Science 12: 744–758. Kogut B, Zander U. 1992. Knowledge of the firm, combinative capabilities, and the replication of technology. Organization Science 3: 383–397. Kogut B, Zander U. 1996. What firms do? Coordination, identity, and learning. Organization Science 7: 502–518. Kulatilaka N, Perotti EC. 1998. Strategic growth options. Management Science 44: 1021–1031. Kumar MVS. 2005. The value from acquiring and divesting a joint venture: a real option approach. Strategic Management Journal 26(4): 321–331. Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options Levin RC, Klevorick AK, Nelson RR, Winter SG. 1987. Appropriating the returns from industrial research and development. Brookings Papers on Economic Activity. Brookings Institutution: Washington, D.C. Levinthal D. 1988. A survey of agency models of organizations. Journal of Economic Behavior & Organization 9: 153–185. Luehrman TA. 1998. Strategy as a portfolio of real options. In Managing in the New Economy, Magretta J (ed). Harvard Business School Press: Boston, MA; 91–111. Markides CC, Williamson PJ. 1994. Related diversification, core competencies and corporate performance. Strategic Management Journal Summer Special Issue 15: 149–165. McGrath RG. 1997. A real options logic for initiating technology positioning investments. Academy of Management. Review 22(4): 974–996. McGrath RG. 1999. Falling forward: real options reasoning and entrepreneurial failure. Academy of Management Review 24(1): 13–30. McGrath RG, Nerkar A. 2004. Real options reasoning and a new look at the R&D investment strategies of pharmaceutical firms. Strategic Management Journal 25(1): 1–21. Miller KD. 1998. Economic exposure and integrated risk management. Strategic Management Journal 19(5): 497–514. Miller KD, Folta TB. 2002. Option value and entry timing. Strategic Management Journal 23(7): 655–665. Mitchell GR, Hamilton WF. 1988. Managing R&D as a strategic option. Research and Technology Management 31: 15–22. Osterloh M, Frey BS. 2000. Motivation, knowledge transfer, and organizational forms. Organization Science 11: 538–550. Peteraf MA. 1993. The cornerstone of competitive advantage: a resource-based theory. Strategic Management Journal 14(3): 179–191.

Copyright  2008 John Wiley & Sons, Ltd.

719

Polanyi M. 1966. The Tacit Dimension. Routledge and Kegan Paul: London, UK. Rajan RG, Zingales L. 1998. Power in a theory of the firm. The Quarterly Journal of Economics 113: 387–432. Rajan RG, Zingales L. 2001. The firm as a dedicated hierarchy: a theory of the origin and the growth of firms. The Quarterly Journal of Economics 116: 805–851. Reuer JJ, Leiblein MJ. 2000. Downside risk implications of multinationality and international joint ventures. Academy of Management Journal 43: 203–14. Snell SA, Shadur MA, Wright PM. 2001. Human resources strategy: the era of our ways. In Handbook of Strategic Management, Hitt M, Freeman RE, Harrison JS (eds). Blackwell Publishing: Oxford, UK; 627–649. Teece DJ. 1982. Towards an economic theory of the multiproduct firm. Journal of Economic Behavior & Organization 3(1): 39–63. Titman S. 1984. The effect of capital structure on a firm’s liquidation decision. Journal of Financial Economics 13(1): 137–151. Trigeorgis L. 1996. Real Options: Managerial Flexibility and Strategy in Resource Allocation. MIT Press: Cambridge, MA. Vassolo RS, Anand J, Folta TB. 2004. Non-additivity in portfolios of exploration activities: a real optionsbased analysis of equity alliances in biotechnology. Strategic Management Journal 25(11): 1045–1061. Weeds H. 2002. Strategic delay in a real option model of R&D competition. Review of Economic Studies 69: 729–747. Wicklund RA. 1975. Objective self-awareness. In Advances in Experimental Social Psychology, Berkowitz L (ed). Academic Press: Orlando, FL; 233–275.

Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

720

H. Wang and S. S. Lim

APPENDIX

= −(1 − p)(1 − q)(x H − I1 )

PROOF OF LEMMA 1:

− pq(x L − I1 ) > 0

The assumption that the expected cash flow from the project is greater (less) than the additional investment conditional on s H (s L ) can be written as follows. Pr(x H |s H , i H )x H + Pr(x L |s H , i H )x L > I1 Pr(x H |s L , i H )x H + Pr(x L |s L , i H )x L < I1 Pr(x H |s H , i L )x H + Pr(x L |s H , i L )x L > I1 Pr(x H |s L , i L )x H + Pr(x L |s L , i L )x L < I1 (A1) These inequalities can be rewritten as follows: pq(x H − I1 ) + (1 − p)(1 − q)(x L − I1 ) > 0

PROOF OF LEMMA 2: We show in this lemma that the negative effect on project value due to the reduction in employee incentives dominates the positive effect of flexibility from making a real option investment. In other words, the firm value with a high level of employee specific human capital under a fullscale investment is higher than that with a low level of employee specific human capital under a real option investment, that is, E Options [V |i L ] < E F ull [V |i H ]. Since p > 0.5 and q > 0.5, (A3) implies (A5), and (A4) implies (A2). Combining (A3) and (A4) gives the following condition that can replace the set of inequalities (A2)-(A5).

(A2)

−

(1 − p)q(x − I1 ) H

+ p(1 − q)(x L − I1 ) < 0

(A3)

q(1 − p) L (x − I1 ) < (x H − I1 ) p(1 − q) <−

p(1 − q)(x − I1 ) H

+ (1 − p)q(x L − I1 ) > 0

(A4)

(1 − p)(1 − q)(x H − I1 ) + pq(x L − I1 ) < 0

(A5)

We show below that the conditions (A3) and (A5) imply that the expected value of a project is higher under a real option investment than a fullscale investment, holding the level of employee specific human capital constant.

− q(x H − I1 ) − (1 − q)(x L − I1 ) = −(1 − p)q(x H − I1 ) − p(1 − q) × (x L − I1 ) > 0 E

[V |i ] − E

F ull

(A6) [V |i ] L

= (1 − q)p(x − I1 ) + q(1 − p)(x − I1 ) H

− (1 − q)(x H − I1 ) − q(x L − I1 ) Copyright  2008 John Wiley & Sons, Ltd.

(A8)

A necessary condition for a nonempty parameter set that satisfies the above condition is that (x H − I1 ) > 0

&

(x L − I1 ) < 0

p(1 − q) q(1 − p) < p(1 − q) q(1 − p) ⇒

q 2 (1 − p)2 < p2 (1 − q)2

⇒

q
(A9)

E Options [V |i L ] − E F ull [V |i H ] − q(x H − I1 ) − (1 − q)(x L − I1 )

= qp(x H − I1 ) + (1 − q)(1 − p)(x L − I1 )

L

p(1 − q) L (x − I1 ) q(1 − p)

= p(1 − q)(x H − I1 ) + q(1 − p)(x L − I1 )

E Options [V |i H ] − E F ull [V |i H ]

Options

(A7)

L

= (p − q − pq)(x H − I1 ) + (q − pq + q − 1)
p(1 − q) × (x L − I1 ) < −(p − q − pq) (1 − p)q  L +(q − pq + q − 1) (x − I1 ) =

(2q − 1)[p(p − q) + q(1 − p)](x L − I1 ) (1 − p)q <0

(A10) Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj

Employee Specific Investments and Real Options The inequality in (A10) follows from the substitution of (x H − I1 ) with the second inequality in (A8), and the last line follows from (A9) that q < p and (x L − I1 ) < 0. Therefore, the expected value of the project if invested as a real option but

Copyright  2008 John Wiley & Sons, Ltd.

721

with a low level of employee specific human capital is smaller than the expected value of the project if invested full-scale up front but with a high level of employee specific human capital.

Strat. Mgmt. J., 29: 701–721 (2008) DOI: 10.1002/smj