This paper establishes the payoff models of the European option for research and development (R&D) projects with two enterprises in a research joint venture (RJV). The models are used to assess the timing and payoffs of the R&D project investment under quantified uncertainties. After the option game, the two enterprises can make optimal investment decision for the R&D project investment in the RJV.
1. Introduction
The success of a research and development (R&D) project will bring huge profits for the high technology enterprises. R&D project includes software project and hardware project. Both of them have some uncertainties that may come from the limitation of the R&D capabilities, external market volatility, or complexity of the project. Because of the existence of these uncertainties, it is not easy for an enterprise to seize the opportunities in R&D project investment.
To cope with the uncertainties and pressure of market competition in a R&D project, the enterprises will establish a research joint venture (RJV) for the alliance of financing and technology development before the commercialization of product. The competition among enterprises has been transformed from a zero-sum game to a non-zero-sum game. As discussed in Dong et al. [1], one of the best business partners usually is their largest competitors. However, due to self-interest, any alliance could collapse. If the only purpose of any enterprise is to take advantage of the technology from its partners without share, it will inevitably cause concealment of the achievement and the failure of the joint venture.
Kamien et al. [2] and Amir et al. [3] divide collaboration in R&D projects into three categories: R&D cartel, RJV, and cartelized RJV. The result of Liu and Zhang [4] shows that the RJV is more applicable in R&D projects. According to Sheng [5], the enterprises in a RJV fully share their R&D activities amongst the partners and seek to maximize their own business profits. This alliance is only for technology sharing but not for the strategic timing to invest. Therefore, in a RJV, the enterprises also compete with each other in investment decision making.
The various uncertainties in a R&D project are considered by real option theory, which can be used as a kind of business analysis and management tool in a R&D project investment decision making. The enterprises can have the right to invest or abandon the R&D project. When the internal or external conditions change and become not conducive to invest, the enterprises may give up the investment opportunities, and their losses are only the irreversible investment.
Azevedo and Paxson [6] review the real option game models for the last two decades. McDonald and Siegel [7] assess European option with dividends in their model. Cassimon et al. [8] extend compound option model to allow for phase-specific volatility estimates, but it does not discuss uncertainties for this model. Ghosh and Troutt [9] develop a more sophisticated multistage compound option model, but the model is hard to apply. Angelou and Economides [10] integrate compound real option and game theory and adopt price competition analysis for the optimal business strategy. Lukas and Welling [11] illustrate the influence of uncertainty on the associated first-mover advantage in a U-shaped pattern.
Due to the financial attributes of real option, therefore, in the transfer, investment, acquisition, and distribution of benefits, it is necessary to quantify the European option value for a R&D project. The real option method has already been used to quantify the option value. However, the calculations of the value and payoffs are not accurate when the uncertainties and other competitors’ strategies exist. Based on the above considerations, this paper from the perspective of enterprises investment in a R&D project firstly considers the impact of a RJV with other competitors, then quantifies and evaluates a variety of uncertainties in the R&D project, and finally builds the models of the investment decision payoffs. After the option game, the enterprises can obtain optimal R&D project investment decisions by Nash equilibrium point in the RJV.
2. European Option Model for R&D Project Investment under Market Uncertainty
According to the definition in Margrabe [12], European option holders may exchange asset I for asset V at time T. In the R&D project investment, it can be explained that the holder of the commercial investment I exchanges it for the market value V at time T. Increasing market demand will result in a scale expansion of production, which will decrease the production cost. Market uncertainty is the exogenous risks associated with acceptance by the market, which relates to the compatibility of a new technology with the preference of customers [13]. The main source of this uncertainty is from the market value VT of R&D project at time T. Taking the learning effect into account, the cumulative experience will cause unit profit QT=ebδ with the market penetration rate b (the learning speed) and unit cost reduction rate δ. The trend of VT should include the part of the increased profit (dQT/dT)(1/ΔQT)=bδ. Assuming that VT follows the Brownian motion process, after the application of Ito lemma, we have(1)dVT=μ-12σ2VTdT+σVTdZT+dQT.
The stochastic variable dZT follows a standard Wiener process. The market growth rate can be measured by μ=r+εσ, where ε and σ represent market price of risk and market volatility and r is the risk-free interest rate. Then we have(2)dVT=r-12σ2VTdT+σVTεdT+dZT+dQT.
By the Girsanov theorem, the process dZT∗=εdT+dZT is a new Brownian motion under risk-neutral measure, so (2) can be written as(3)dVTVT=r-12σ2+bδdT+σdZT∗.
A stochastic process is a sequence of probability distributions that provides the transition likelihood of future values. It adjusts the probabilities of future outcomes, which can be incorporated in the effects of risk [14]. Hence, dVT shows that the value of R&D project is also a stochastic process with drift r+bδ. Taking the market uncertainty into account, the project value will become(4)VT=Ver-1/2σ2+bδT+σZT∗.
The stochastic variable ZT∗ in (4) follows a standard Wiener process in which ZT∗~N(0,T), so that φ≡(-(1/2)σ2T+σZT∗)~N(-(1/2)σ2T,σT). Therefore, eφ is lognormally distributed with E[eφ]=e(-(1/2)σ2T+(1/2)σ2T)=1 and E[VT]=Ve(r+bδ)T, which is the unconditional expected value of VT.
The payoff of the R&D project at time T is maxVT-I,0 and the European option s(V,I,T) can be denoted as the value of investment opportunity. In this paper, we show that market uncertainty in a R&D project can be simplified to the European option model introduced by Geske [15] with increased discount rate:(5)sV,I,T=EmaxVer-1/2σ2+bδT+σuT-I,0.
Thus, we can obtain the European option model of a R&D project under market uncertainty:(6)sV,I,T=VebδTN1d1-Ie-rTN1d2,where d1=(logK+(r+σ2/2+bδ)T)/σT; d2=d1-σT; K=V/I; σV and σI are the volatility of V and I; σ=σV2-2ρVIσVσI+σI2; ρVI is the correlation between V and I; and N1 is a normal distribution.
Because the investment decisions in a R&D project can be made in stages, and the stages are nested to each other [16], the R&D project evaluation can be made in a two-phase compound option model. As above, the market uncertainty with the market factors outside the enterprises may cause marginal changes in the asset value of compound option. The terminal market value of the R&D project at time T1 is(7)VT1=Ver-1/2σ2+bδT1+σZT1∗.
The compound option valuation can be boiled down to the following:(8)cs,I1,T1=EmaxsV,I,T-I1,0.
From the derivation in Carr [17] and Paxson [18], we can obtain the compound option model under market uncertainty, given by(9)cs,I1,T1=VebδTN2d1KK∗,T1,d1K,T;τ-Ie-rTN2d2KK∗,T1,d2K,T;τ-I1e-rT1N1d2KK∗,T1,where d1(K/K∗,T1)=(log(K/K∗)+(r+(σ2/2)+bδ)T1)/σT1; d2(K/K∗,T1)=d1(K/K∗,T1)-σT1; T1 is the expiration date of the compound option; T2=T-T1, τ=T1/T, T>T1; I1 is the investment at time T1; K∗ is the critical price ratio that solves the equation I1/I=K∗ebδTN1(d1(K∗,T2))-e-rT2N1(d2(K∗,T2)); and N2 is a bivariate normal distribution.
3. The Payoffs of the R&D Project Investment Decisions
Under dynamic competition, the enterprises can generally be divided into leader, follower, or competitors who make the investment decisions at the same time. Although the follower often follows the leader to make decisions, it does not mean that the leader is always the winner. Having more time to observe the market and make decisions, the follower can also occupy favorable position in the market. Due to the constraints of the market size, the follower, when entering the market, will influence the benefits of the leader. Therefore, when making investment decisions, the leader should take the decisions of the follower into full account. Hence, it is necessary for the enterprises to assess all possible decisions of the competitors in the R&D project investment.
In the process of a R&D project investment, except market uncertainty, the technology uncertainty could exist [19]. It refers to the chance to successfully impel the transition at each R&D stage from technological concerns. Assume that there are two enterprises: A and B. In the RJV, they have the same success rate p [19]. To obtain the same payoffs in the R&D project, the competitor with stronger R&D capabilities needs relatively less investment. Assume that the R&D project investment amount of the two enterprises is IA and IB, and enterprise A is more competitive in the R&D project than enterprise B; that is, IA<IB.
In this paper, we assume the R&D project investment of the leader is made at time T0, earlier than that of the follower. The market shares of the two enterprises are different: the share of leader is α (α>0.5) and that of the follower is 1-α. The implicit condition is that the two enterprises occupy the entire market and become duopoly. In reality, there may be more than two competitors, where we can still apply this model by treating the homogeneous pioneers as one leader and all the rest as one follower.
According to the definition of the European option, all the decision points must be fixed, and all the investment has to be made at time T0. The two enterprises have to face the situation that invest in R&D projects at time T0 or delay investment until time T1, which requires the payoffs of all possible investment decisions to be calculated and analyzed accurately. In the RJV, there could be four investment decisions for an enterprise: being the leader at time T0; being the follower at time T1; investing together with the other at time T0; or waiting until time T1 to invest together with the other.
3.1. The Leader’s Payoff in the RJV
In the RJV of the R&D project investment, it can be assumed that enterprise A as the leader enters the market first at time T0 and enterprise B as the follower decides to postpone investment. In this situation, the leader claims the market share α and takes the option s(αV,I,T). For the payoffs in making the commercialization investment I, Pennings and Sereno [20] state that s(αV,I,T) must be larger than IA for the R&D project to succeed. According to (6), combining with the probability of success p, the payoff of the leader A at time T0 is the following: (10)LA=psαV,I,T-IA=pαVebδTN1d1K,T-Ie-rTN1d2K,T-IA.
Symmetrically, if we assume that enterprise B is the leader in the R&D project making an investment, at time T0, the payoff of the leader B is the following:(11)LB=psαV,I,T-IB=pαVebδTN1d1K,T-Ie-rTN1d2K,T-IB.
3.2. The Follower’s Payoff in the RJV
It can be assumed that the follower has to invest at time T1, so it takes R&D time T+T1 in the R&D project. In the RJV, the two enterprises fully carry out technology share, and the follower can obtain the technology from the leader, which can reduce the amount of investment at time T1. Assuming that the investment is used at constant rate over time, the amount of investment of the follower B reduces to IB∗=((T-T1)/T)IB. The follower takes the market share 1-α. Because the first stage is waiting from time T0, and the second stage is investing at time T1, the strike price of the option is s((1-α)V,I,(T+T1)) [16], the mature time is T+T1, and the payoff of the follower enterprise B is GB=c(ps((1-α)V,I,(T+T1)),IB∗,T1). According to (9), the follower’s payoff at time T0 is(12)GB=p1-αVebδTN2d1KKB∗,T1,d1K,T+T1;τ-Ie-rT+T1N2d2KKB∗,T1,d2K,T+T1;τ-IB∗e-rT1N1d2KKB∗,T1,where KB∗ is the critical value of the solution from ps((1-α)V,I,T)=IB∗; we can rewrite it as(13)IB∗I=p1-αKB∗ebδTN1d1KB∗,T-e-rTN1d2KB∗,T.
Similarly, in the RJV, the payoff of follower A at time T0 is the following:(14)GA=p1-αVebδTN2d1KKA∗,T1,d1K,T+T1;τ-Ie-rT+T1N2d2KKA∗,T1,d2K,T+T1;τ-IA∗e-rT1N1d2KKA∗,T1,where KA∗ is the critical value of the solution from(15)IA∗=pI1-αKA∗ebδTN1d1KA∗,T-e-rTN1d2KA∗,T.
3.3. The Payoffs When Both Enterprises Invest Simultaneously in the RJV
If the two enterprises A and B decide to invest in the R&D project at the same time T0, it can be assumed that the two enterprises can capture the same market share α=0.5. Because they have the same amount of commercialization investment in the R&D project, the enterprises hold the same option s(0.5V,I,T). According to (6), the payoff of A in the RJV is the following:(16)TA=ps0.5V,I,T-IA=p12VebδTN1d1K,T-Ie-rTN1d2K,T-IA.
Similarly, the payoff of B is the following:(17)TB=ps0.5V,I,T-IB=p12VebδTN1d1K,T-Ie-rTN1d2K,T-IB.
3.4. The Payoffs When Both Enterprises Wait to Invest Simultaneously in the RJV
In the RJV, when both enterprises simultaneously postpone their investments to time T1 at time T0, the maturity time will be postponed to T+T1. The two enterprises have the same market share of α=0.5. Both enterprises hold the same option s(0.5V,I,(T+T1)) as strike price of two-stage compound option [16]. After investments IA and IB in the second stage at time T1, according to (9), the payoff of enterprise A at time T0 is given by(18)DA=cps0.5V,I,T+T1,IA,T1=p0.5VebδTN2d1KKA∗∗,T1,d1K,T+T1;τ-Ie-rT+T1N2d2KKA∗∗,T1,d2K,T+T1;τ-IAe-rT1N1d2KKA∗∗,T1,where KA∗∗ is the critical value of the solution from the following:(19)IAI=p0.5KA∗∗ebδTN1d1KA∗∗,T-e-rTN1d2KA∗∗,T.
Similarly, the payoff of enterprise B at time T0 is given by(20)DB=cps0.5V,I,T+T1,IB,T1=p0.5VebδTN2d1KKB∗∗,T1,d1K,T+T1;τ-Ie-rT+T1N2d2KKB∗∗,T1,d2K,T+T1;τ-IBe-rT1N1d2KKB∗∗,T1,where KB∗∗ is the critical value of the solution from the following:(21)IB=pI0.5KA∗∗ebδTN1d1KA∗∗,T-e-rTN1d2KA∗∗,T.
4. Numerical Example
In this example, we use the following parameter values to conduct the numerical study: the investment of enterprise A with stronger R&D capabilities is IA=8000 in the R&D project, and the other enterprise B is IB=10000. The success rate of the R&D project is p=0.7. The market value and commercialization investment volatilities are σV=0.70 and σI=0.33, respectively, and the correlation between V and I is ρVI=0.15. The expiration time of the R&D project is T=3 years, and the investment of the follower is at time T1=0.5 years. The market share of the leader is α=0.60. The risk-free interest rate is r=0.05. The market value is V=30000. The market penetration rate is b=1.02 and the unit cost reduction rate is δ=0.2.
Table 1 shows the payoffs of the two enterprises for different commercialization investment in the RJV and the Nash equilibrium for the optimal investment decision after the option game. For the enterprises in the RJV, if less commercialization investment of the R&D project is required, such as I=5000, the two enterprises will anticipate investment at time T0. If the amount of commercialization investment increases, such as in the range 10000≤I≤25000, all kinds of payoffs will reduce, and the optimal investment decisions will change, and enterprise A with stronger R&D capabilities will try to invest in the R&D project at first. When the value becomes larger, such as I=30000, the optimal decisions for the two enterprises are the postponement for the investment at time T0 and then simultaneous investment at time T1, resulting in (DA,DB). As the optimal investment decisions for the R&D project, the results are consistent with practices. When the amount of commercialization investment of the R&D project is large, for the decrease of the sunk costs, the two enterprises will observe the investment behavior from other competitors and market dynamics for more market information. If the amount of commercialization investment is reduced, the enterprises with weaker R&D capabilities will wait for a better investment opportunity. If the amount of commercialization investment is further reduced, all enterprises in the RJV will invest immediately to compete and obtain more profits in the R&D projects.
The payoffs of different commercialization investment for enterprises A and B.
I
LA
DA
TA
GA
LB
DB
TB
GB
Optimum
5000
9465
6971
6159
5033
7465
5588
4159
3975
(TA, TB)
10000
7428
5507
4167
3709
5428
4344
2167
2877
(LA, GB)
15000
5754
4418
2558
2780
3754
3437
558
2124
(LA, GB), (LB, GA)
20000
4359
3589
1238
2111
2359
2758
−761
1593
(LA, GB)
25000
3179
2945
138
1621
1179
2237
−1861
1209
(LA, GB)
30000
2168
2436
−790
1254
168
1832
−2790
926
(DA, DB)
With the commercialization investment I=20000, Table 2 shows the payoffs of the investment decisions under different investment timing in the RJV and the optimal investment decisions after the option game. Because the two enterprises in the RJV fully share technology with each other, the success rate of the two enterprises is always the same. A shorter T1 means the follower enterprises have to make investment decision more quickly. Table 2 illustrates that the payoffs of DA, DB, GA, and GB could increase with the increase in T1. This is the reason that waiting can increase the option value and the enterprises will incline to collect more information before investment. However, the R&D project investment cannot be delayed for too long, because delays will result in reduced future market shares and future payoffs. Hence, the enterprises can select the most acceptable investment decision as the optimal decision and determine the optimal investment timing.
The payoffs of different T1 for enterprises A and B.
T1
LA
DA
TA
GA
LB
DB
TB
GB
Optimum
0.1
6419
3259
2778
939
4419
1951
778
414
(TA, TB)
0.3
6419
4177
2778
2062
4419
3095
778
1430
(LA, GB)
0.5
6419
4866
2778
2899
4419
3883
778
2272
(LA, GB), (LB, GA)
0.7
6419
5430
2778
3602
4419
4512
778
3003
(LA, GB)
0.9
6419
5909
2778
4221
4419
5040
778
3658
(LA, GB)
1.1
6419
6327
2778
4778
4419
5497
778
4256
(LA, GB)
1.3
6419
6695
2778
5288
4419
5899
778
4809
(DA, DB)
1.5
6419
7022
2778
5762
4419
6255
778
5327
(DA, DB)
5. Conclusions
In this paper, we apply the real option theory in R&D project investment decision making by considering various uncertainties. By using our model, combined with the competitive actions in the RJV, the payoffs of the enterprises can be computed accurately. After quantifying the market uncertainty and technology uncertainty of the R&D project investment, the two enterprises can make optimal investment decisions through option game from the payoffs of four kinds of investment decisions: as the leader, as the follower, investing at the same time, or deferring investment at the same time. If the expiration time of the compound option increases, the payoffs of follower and delayed investment will increase. The enterprises can make the optimal investment decision of R&D project and find the optimal investment timing. With the decrease in the commercialization investment in the R&D project, all the payoffs of the investment decisions will increase and the optimal decisions will be changed. Our theoretical and numerical results are consistent with what have been observed in practice in R&D project investment.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.
Acknowledgments
This work is supported by the National Natural Science Foundation of China (70871030), Yanshan University Science Foundation for Youths (13SKB013), the Science and Technology Bureau Foundation of Qinhuangdao (201402B041), the Development of Social Science Research Foundation of Hebei Province (2014031332), and the National Defense Science and Industry Bureau Technology Foundation Research Project of China (JZ20150006).
DongG.LiY.ZhouY.Roles of learning capacities and commitments in establishment of co-opetition relationshipsKamienM.MullerE.ZangI.Research joint ventures and R&D cartelsAmirR.NannerupN.StepanovaA.EguiazarovaE.Monopoly versus R&D-integrated duopolyLiuX.ZhangZ.The cooperative R&D based on research joint venture and enterprise R&D decision with spilloverProceedings of the 20th International Conference on Management Science and Engineering (ICMSE '13)July 2013Harbin, China1861186510.1109/icmse.2013.65865192-s2.0-84884304229ShengY.Spillover effect benefit distribution and choice of technical alliance-from the angle of Shapley valueAzevedoA.PaxsonD.Developing real option game modelsMcDonaldR. L.SiegelD. R.Investiment and the valuation of firms when there is an option to shut downCassimonD.EngelenP. J.YordanovV.Compound real option valuation with phase-specific volatility: a multi-phase mobile payments case studyGhoshS.TrouttM. D.Complex compound option models—can practitioners truly operationalize them?AngelouG. N.EconomidesA. A.Investment flexibility and competition modeling for broadband businessLukasE.WellingA.On the investment–uncertainty relationship: a game theoretic real option approachMargrabeW.The value of an option to exchange one asset for anotherHisschemöllerM.BodeR.KerkhofM. V. D.What governs the transition to a sustainable hydrogen economy? Articulating the relationship between technologies and political institutionsRonnE. I.GeskeR.The valuation of compound optionsAndergassenR.SerenoL.Valuation of N-stage investments under Jump-Diffusion processesCarrP.The valuation of sequential exchange opportunitiesPaxsonD. A.Sequential American exchange property optionsCassimonD.BackerM. D.EngelenP. J.WouweM. V.YordanovV.Incorporating technical risk in compound real option models to value a pharmaceutical R&D licensing opportunityPenningsE.SerenoL.Evaluating pharmaceutical R&D under technical and economic uncertainty