Real options and technology management: Assessing technology migration options in wireless industry

Telematics and Informatics 26 (2009) 180–192

Contents lists available at ScienceDirect

Telematics and Informatics journal homepage: www.elsevier.com/locate/tele

Real options and technology management: Assessing technology migration options in wireless industry Hak J. Kim a,*, Martin B.H. Weiss b, Benoit Morel c a b c

University of Houston-Clear Lake, MIS, Houston, TX 77058, United States University of Pittsburgh, Department of Information Science and Telecommunications, United States Carnegie Mellon University, Department of Engineering and Public policy, United States

a r t i c l e

i n f o

Article history: Received 10 September 2007 Received in revised form 21 May 2008 Accepted 26 May 2008

Keywords: Real options Technology migration 3G Wireless network

a b s t r a c t The wireless industry is currently undergoing a major transition from second generation (2G) to third generation (3G) wireless technologies. The paper attempts to assess wireless technology migration options using the real options approach (ROA) to support the wireless network operators’ strategic decisions: to migrate or not, if so, which migration path to take. The preliminary result shows that the evolution of wireless network technologies between generations is desirable, but not within generations. Finally, from a strategic perspective, we should consider the possible challenges that may hinder migration. By identifying these challenges, we can be more watchful of transition pitfalls and can choose a better alternative. Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction The US wireless industry is currently undergoing a major transition from second generation (2G) to third generation (3G) wireless technologies, which offer high speed data services. These high quality data services are the foundation of multimedia and interactive information systems that service providers hope will be a significant contributor to future profits. To that end, equipment providers have been developing the 3G technologies to support these services. For their part, the US wireless network operators must decide how to migrate their network, i.e., how to best deliver high-quality multimedia services at minimum cost over a smooth evolutionary path. Since the complete replacement of the existing wireless network architecture is generally not practical and since there is an economic trade-off involved in the choice of different technologies, the migration of the existing networks is challenging to network service providers: that is, which migration path to take and what to do once there. The US wireless industry now has four nationwide wireless service providers: Cingular-AT&T Wireless, Verizon Wireless, Sprint PCS-Nextel, and T-Mobile. In addition, there are a number of large regional players, including Western Wireless Corp., US Cellular, Dobson Communications Corp., and Alltel. Each firm uses a different technology or combination of technologies for their current networks, as shown in Table 1. Verizon Wireless uses AMPS and CDMA; and Cingular-AT&T Wireless use AMPS, TDMA, and GSM. Sprint Wireless and T-Mobile use CDMA and GSM, respectively. The evolutionary paths to 3G from the principal 2G technologies, GSM and CDMA are quite distinct and there are many alternative wireless network technologies, such as TDMA, GSM, GPRS, EDGE, WCDMA, cdma2000, etc. These abundant choices of wireless technologies require service providers to examine the options for their network evolution as a strategic

* Corresponding author. E-mail addresses: [email protected] (H.J. Kim), [email protected] (M.B.H. Weiss), [email protected] (B. Morel). 0736-5853/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tele.2008.05.004

181

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192 Table 1 US wireless firms’ technologies

AMPS TDMA GSM CDMA

Verizon

Cingular-AT&T

O

O O O

O

Sprint

T-Mobile

Nextel iDEN (integrated Digital Enhanced Network)

O O

decision. The main approach of this research, the real options approach (ROA), introduces a new perspective on technology policy issues in networks, such as network architecture and technology choice, network service provisioning, and network regulation and policy. Based on ROA, wireless network operators may find it worthwhile to evaluate new technologies as strategic options. This study intends to raise core issues concerning the transition to 3G and to resolve these both qualitatively and quantitatively. This paper explores two typical network migration alternatives: the ‘global systems for mobile communications (GSM)based’ network migration path and the ‘code division multiple access (CDMA)-based’ network migration path, as strategic options for facilitating the migration into a next generation network architecture. One calls for substantial infrastructure replacement (architectural innovation), while the other calls for upgrades to existing equipment towards 3G (modular innovation). In this study we blend a comparative study and a ‘what if’ study (or contingency study) by examining two wireless network migration scenarios. This paper also attempts to assess these migration path options towards the third generation (3G) wireless data service network architecture, using the real options approach (ROA). Since one of the implicit aims of this study is to understand how the real options approach can be used as a model for technology choice, we simplify matters where possible. For example, taking into account all the problems of reaching relevant data on technological development, we assume that the only available data are on current market shares of competing technologies in generation. We hope to use more refined and enriched data in future research. 2. Real options theories: a brief overview The field of real options has been recognized the fundamental importance in academics (McDonald and Siegel, 1986; Carr, 1988; Pindyck, 1989; Dixit and Pindyck, 1994; Trigeorgis, 1996; Amram and Kulatilaka, 1998; Benaroch and Kauffman, 2000; Benaroch, 2002; Kim and Sanders, 2002) and in actual practice (mining, oil, medicine, IT, etc.), as a strategic tool to evaluate business projects investment. Network industry (Alleman and Noam, 1999; Economides, 1999; d’Halluin et al., 2002; Basili and Fontini, 2003; Kim, 2004; Pateli and Giaglis, 2007) cannot be exception, especially capital investment issues. The most desirable situation is high uncertainty with much cost. Network industry fits well this environment. The demands for various services in markets and the rapid development of network technologies are challenging to network operators to follow these environment. These demand them to examine the options for their network architecture design as a strategic decision problem. The theory of real options provides a rigorous framework to analyze the optimal exercise of options (Trigeorgis, 1996). Real options have emerged by the criticism of the traditional investment approaches, such as a discounted cash flow approach (e.g., NPV and IRR). The traditional NPV method is called a now-or-never decision rule (Dixit and Pindyck, 1994). However, this simple rule is not appropriate for most investment projects. In reality, the decision for investment can be contingent on what kind of future unfolds. For instance a manager can wait for a time until market situation is improved, or decide not to invest if the market situation is bad, because less uncertainty about the investment. The ROA provides a structure linking strategic planning and financial analysis tools to evaluate potential opportunities and uncertainty (Dixit and Pindyck, 1994). For example, when managers evaluate new projects, they may face several choices beyond simply accepting or rejecting the investment. Other choices include delaying decisions until the market conditions are more favorable, or deciding to start small and expanding later if the results are good. The ROA is appealing to the firms because of its distinctive ability to capture managers’ flexibility (Dixit and Pindyck, 1995) in adapting their future actions in response to evolving markets or technological conditions. 3. Technology options in wireless networks Over the past decade, wireless networks have moved rapidly from first-generation (1G) analog, voice-only communications, to second generation (2G) digital, voice and data communications, and further to third generation (3G) wireless networks. The 1G wireless networks (Rapport, 1996) are based on the analog technology which is usually represented as a series of sine waves to show the fluctuations of the voice itself. The main 1G technology standards are AMPS, TACS and NMT (Dahlman et al., 1998). The 2G wireless networks replaced analog networks (1G) with digital, and allowed data to join the wireless world. The 2G wireless networks (Rapport, 1996) are started by the advent of digital technology which is a way to transmit or store data with a string of 0’s and 1’s. Digital technology forced to the most fundamental change in telecommunications.

182

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

Using this digital technology, voice signals are digitized and then sent as bits of data over radio waves. There are several 2G wireless technologies, such as TDMA, GSM, cdmaOne and PDC (Dahlman et al., 1998). One stage before 3G wireless systems comes 2.5G which is a technology that allowed second generation users to get a taste of what 3G would eventually present. 2.5G systems, such as GPRS, EDGE and HSCSD (Rapport, 1996), can be seen as straightforward upgrades of second generation networks, since in most cases, the 2G infrastructures underwent simple software/hardware developments. Nowadays the wireless network architecture is moving to the 3G wireless technologies, which is to provide the high-rate voice and data service. The 3G system (Prasad, 1998) is demanded to provide multi-megabit Internet access with an ‘always on’ feature and data rates of up to 2.048 Mbps for multimedia services. The 3G wireless system is currently split into two groups (Carsello et al., 1997; Dahlman et al., 1998; Garg, 2001): the UMTS group (3GPP) and the cdma2000 group (3GPP2): The Third generation Partnership Project (3GPP) is collaboration between organizational partners (OPs) which study the W-CDMA/TD-SCDMA/EDGE standards and the Third Generation Partnership Project 2 (3GPP2) is collaboration between OPs which examine the cdma2000 standards. Fig. 1 shows the possible technology transition scenarios. A carrier can implement one of three 2G technologies: TDMA, GSM, and CDMA. TDMA and CDMA are more popular in the US, while GSM is prevalent in Europe. For more high-speed data services, 2.5G technologies, GPRS, EDGE, and cdma2000-1XRTT, have been developed. 2.5G is always on, provides simultaneous voice and data, and delivers more speed than today’s 2G circuit-switched data connections. 2.5G offers more bandwidth than 2G but less than 3G. Wireless network operators can implement 2.5G much more cheaply than 3G because the former uses existing 2G spectrum and does not require a new network infrastructure, although some system upgrades are necessary. So, 2.5G is often considered a stepping-stone to 3G. As the wireless industry moves toward 3G technologies, the current coexistence of three major technologies will most likely evolve into two competing technologies within the 3G market: WCDMA and cdma2000. There are several migration path scenarios from 2G to 3G for the wireless network operators, but currently the 3G world is split into two alternatives: the cdma2000 which is an evolution of IS-95 (‘CDMA-based network migration strategy’) and the WCDMA/TD-SCDMA/EDGE whose standards are all improvements of GSM, IS-136 and PDC (‘GSM-based network migration strategy’). Still there is not clear which alternative is better towards the 3G. 3.1. Scenario 1: GSM-based network migration path Fig. 2 shows the migration path scenario from GSM (2G) to GPRS (2.5G) and to WCDMA (3G). When GPRS service is provided in the GSM network, several components are added, like SGSN and GGSN (yellow shaded boxes). Further, a transition from GSM/GPRS to UMTS (3G), access network section (blue shaded boxes) is totally changed or added in the carriers’ networks. In addition it is not possible to operate in a GSM mode within the same 5 MHz band. Since UMTS does not reuse the GSM base station hardware, operators must install new hardware cabinets adjacent to existing systems. Table 2 briefly summarizes what components are upgraded or replaced in GSM-based networks. In case of provisioning GPRS service, software upgrades with little hardware replacement is needed. For UMTS, in contrast, most access network facilities are changed because the technology in GSM/GPRS (TDMA-based) is totally different from UMTS’s technology (CDMA-based). So, it means for a significant investment for 3G under the GSM-based network architecture. 3.2. Scenario 2: CDMA-based network migration path Fig. 3 shows the evolution path from a cdmaOne(2G) to a cdma2000-3X(3G) network architecture. Since cdma2000 is the evolution of IS95-based systems, it is the natural 3G evolution of CDMA technology, requiring only minor upgrades to the network and small capital investment. Because of this, the transition from cdmaOne to cdma2000-1X is relatively easy for operators and transparent for consumers. Wireless service operators can gradually migrate from ‘cdmaOne’ to cdma2000

Fig. 1. Technology options in wireless networks.

183

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

PSTN

GMSC MSC/VLR

USIM

MS

Node B

SIM

MS

HLR

SGSN

RNC BTS

TE

GGSN

IP Network TE

BSC EIR

Signaling Interface Signaling and Data Transfer Interface

CGF

Billing System

CGF: Charging Gateway Functionality EIR: Equipment Identity Register

Fig. 2. GSM-based network architecture.

Table 2 Upgrade/new components in GSM-based networks category Category

GSM to GSM/GPRS

Mobile station (MS)/SIM Base transceiver station (BTS) Base station controller (BSC) Mobile switching center (MSC)/visitor location register (VLR) Home location register (HLR) Serving GPRS support node (SGSN) Gateway GPRS Support Node (GGSN)

GSM/GPRS to UMTS

HW

SW

HW

SW

Upgrade Upgrade Upgrade Upgrade Upgrade New New

Upgrade No change PCU interface No change No change New New

New New New No change No change No change No change

New New New Upgrade No change Upgrade No change

Fig. 3. CDMA-based network architecture.

Table 3 Upgrade/new components in CDMA-based Networks Category

Mobile station (MS) Base transceiver station (BTS) Base station controller (BSC) Mobile switching center (MSC)/visitor location register (VLR) Home location register Home agent (HA)/FA AAA server Packet data switching node (PDSN)

cdmaOne to cdma2000 1

Cdma2000 1 to cdma200 3

HW

SW

HW

SW

New No change No change No change No change New New New

New Upgrade Upgrade Upgrade No change New New New

No No No No No No No No

Upgrade Upgrade Upgrade Upgrade No change No change No change No change

change change change change change change change change

at the cdma2000-1X (1.2288 Mcps) rate. As users migrate to the new standard, network operators can swap out cdma2000 1X radios and insert a cdma2000-3X radio to increase cell capacity. They also have the choice of using three cdma2000-1X radios or converting to a single cdma2000-3X radio. The cdma2000 reuses the same 9.6 kbps vocoder from cdmaOne. As seen in Table 3, the transition from cdmaOne to cdma2000 requires channel card and software upgrades to cdmaOne base stations (older base stations may require some hardware upgrades) and the introduction of new handsets.

184

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

cdma2000-1X, which can be implemented in existing spectrum allocations, delivers approximately twice the voice capacity of cdmaOne, and provides average data rates of 144 kbps. The cdma2000-3X standard is used to signify three times 1.25 MHz or approximately 3.75 MHz. The cdma2000-3X multi-carrier approach, or wideband cdmaOne, is an important part of the evolution of IS95-based standards. In short, cdma2000-3X with data rates of up to 2Mbps offers greater capacity than cdma20001X. So, unlike the case of UMTS, cdma2000 does not require as much investment to provide 3G services. 4. The model In this section, we attempt to develop a theoretical framework to assess strategic options in the migration to the 3G wireless network architecture from 2G. The model is equivalent to options to exchange one risky asset for another in financial market (Margrabe, 1978) and is applied to assess technology options for migrating into the next generation network architecture and technologies in wireless industry. That is, we show the transition value when moving from generation-to-generation (inter-generational transition) and within the same generations (intra-generational transition). Let the value of technology investment in the revolutionary technology (i.e., CDMA-based) compared with the evolutionary (i.e., GSM-based) be ‘H’. Also let P and B be the net value of two alternatives of network migration by the choice of strategy at time t.: One (P) is a revolutionary technology change with a larger risk and investment (‘aggressive’) and the other (B) is a stepping-stone technology change with a smaller risk and investment (‘conservative’). Assuming that the level of investment for improving network performance is directly related to their revenues, the key issue in the choice of strategic options is how to quantify a trade-off between the level of performance improvement and the value of premium in a risk neutral situation. Risk neutrality means comparing one portfolio where an investment is in stepping-stone architecture with a premium to the other portfolio where an investment is in the revolutionary architecture with potentially higher value. We treat the choice between the two scenarios as a comparison between two wireless network technology migration portfolios. Again, let P correspond to a high level of uncertainty (potentially high value) with a much larger investment cost, and B correspond to a lower level of uncertainty with a much smaller investment cost. Two scenarios are defined as:  Revolutionary portfolio ðW REV Þ ¼ mP P (i.e., CDMA-based architecture)  Evolutionary portfolio ðW EVO Þ ¼ mB B (i.e., GSM-based architecture) where mP and mB are amounts invested in each scenario. To compare the two ‘‘portfolios”, we introduce a quantity HðP; BÞ which is defined as:

mH H þ W EVO ¼ W H þ W EVOt ¼ W REV Using the derivative, it can be described as:

mH dHðP; BÞ ¼ vP dP  mB dB Then, by definition,

W HðP;BÞ ¼ W REV  W EVO Rewritten it as follows:

mH HðP; BÞ ¼ vP P  vB B Using the derivative, it can be described as:

mH dHðP; BÞ ¼ vP dP  mB dB By combining the above two formula, we also can rewrite as:

WH

dH dP dB ¼ W REV  W EVO H P B

ð1Þ

Now let us consider the time horizon s to deal with a continuous option, like European-type option which can be exercised at s. This option is simultaneously a call option on asset one with. Clearly HðP; B; sÞ depends also on the time horizon s. Remembering that: W H ¼ W REV  W EVO ,

W H þ W EVO ¼ W REV So, Eq. (1) can be rewritten as:

dH dP dB ¼ ðW H þ W EVO Þ  W EVO H  P B dH dP dP dB WH  ¼ W EVO  H P P B     W EVO dP dB dH dP  ¼  P B H P WH

WH

ð2Þ

185

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

HðB; P; sÞ depend on the two stochastic variables P and B (i.e., it is a derivative) and on the time horizon s. Using Ito’s lemcan be written as: ma, the instantaneous rate of change of that derivative dH H

dH ¼ bdt þ cdz þ gdq H

ð3Þ

where

b¼

c¼

( ( )) 2 1 oH oH oH 1 2 2 o2 H o2 H 2 2o H þ lB þ aP þ d B þ þ 2 q r dBP r P H ot oB oP 2 oBoP oB2 oP 2

ð3aÞ

rP oH

ð3bÞ

H oP dB oH g¼ H oB

ð3cÞ

We make the unavoidable assumption that P and B follow a geometric Brownian motion with drift (we will have to meditate the validity of that assumption):

dP ¼ a dt þ r dz P dB ¼ l dt þ d dq B

ð4aÞ ð4bÞ

The fact that high technology has less variability here could mean that: 0 < d < r. To allow the possibility of correlations between the stochasticities of B(t) and P(t), we assume that: hdz  dqi ¼ qdt, where 1 6 q 6 1. Eq. (2) corresponds in fact to three equations. The coefficients of dt, dq and dz must separately satisfy the equation. Using Eqs. (4a), (4b), (3), and (2) yields the three equations:

W Bt ðb  aÞ ðc  rÞ g ¼ ¼ ¼ W H ða  lÞ r d

ð5Þ

Together with Eqs. (3b), (3c), and (5) (more precisely: rc þ gd ¼ 1) leads to

H¼P

oH oH þB oP oB

ð6Þ

One key observation is that Eq. (6) can be satisfied by assuming (with x ¼ BP):

HðB; P; sÞ ¼ P  hðx; sÞ

ð7Þ g

Another key observation stems from Eq. (5) combined with Eqs. (6) and (3a). Namely: ðb  aÞ ¼ ðl  aÞ d ¼ l

að1  HP oH Þ combined with Eq. (3a), leads to: oP

( ) 2 1 2 2 o2 H o2 H oH 2 2o H þ þ 2qrdBP d B þr P ¼0 2 oBoP ot oB2 oP2

B oH  H oB

ð8Þ

This equation is a differential equation for the derivative HðB; P; sÞ. Using x ¼ BP and Eqs. (7) and (8) become:

V 2 x2 o2 hðx; tÞ ohðx; tÞ þ ¼0 ox2 ot 2

ð10Þ

V 2 ¼ r2  2qrd þ d2 represents the infinitesimal variance of x.1 Rs Let T ¼ t V 2 ðsÞds be the cumulative uncertainty up until the time horizon s. By definition of T, dT ¼ V 2 ðtÞdt, and Eq. (10) can be written:

x2 o2 hðx; TÞ ohðx; TÞ ¼ ox2 oT 2

ð10aÞ

Eq. (10a) is the Kolmogorov backward equation for the stochastic process:

dx ¼ df  ðhdfi ¼ 0 and hdf2 i ¼ dTÞ: x If one defines: y ¼ logðxÞ,

dx x

¼ df becomes: dy ¼  dT þ df. The backward Kolmogorov equation for y is2 2

1 o2 hðy; TÞ 1 ohðy; TÞ ohðy; TÞ ¼  2 oy2 2 oy oT

1 2

  2 From the definition of x and Ito’s lemma: dx x ¼ l  a  qrd þ r dt þ ddq  rdz. S. Karlin, R. Taylor: Second Course in Stochastic processes (Academic, New York, 1981), p. 220.

ð11Þ

186

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

If f ðyÞ ¼ hðy; T ¼ 0Þ, the solution of Eq. (11) is:

1 hðy; TÞ ¼ pffiffiffiffiffiffiffiffiffi 2pT

Z

þ1

f ðnÞe

1

ðynþT Þ2 2 2T

dn;



This can also be written as with :g ¼

1 hðx; TÞ ¼ pffiffiffiffi

p

Z

1

f ðlogðxÞ 

1

 :

ðnyþT2Þ pffiffiffiffi 2T

pffiffiffiffiffiffi T 2 þ g 2T Þeg dg 2

ð12aÞ

What should we use as boundary conditions f ðyÞ ¼ hðy; T ¼ 0Þ? If we interpret hðx; TÞ as the maximum premium that should be paid to invest in high cost technology instead of conservative technology, investing in high technology makes sense only if the premium actually paid (B-P or 1x) is less than the value of H(P,B). In terms of the variable x, this means that hðx; TÞ must be larger than 1x. This implies that the zero uncertainty limits hðx; TÞ ¼ Max½0; 1  x: Remembering that y ¼ log, this implies that f ðy 6 0Þ ¼ 0 and pffiffiffiffi 2T T2

f ðy > 0Þ ¼ ey  1 ¼ xeg

1

Substituting this form for f(z) in Eq. (12a) eventually yields:

x hðx; TÞ ¼ pffiffiffiffi

p

Z

þ1

1 2 eg dg  pffiffiffiffi logðxÞþT p  pffiffiffi 2 2T

Z

þ1 logðxÞT 2 p 2T



ffiffiffi

2

eg dg

ð13Þ

Which can also be written as (this is our ‘‘basic formula”):

hðx; TÞ ¼ xUðd1 ðx; TÞÞ  Uðd2 ðx; TÞÞ

ð14Þ

With:



1 T d1 ðx; TÞ ¼ pffiffiffiffiffiffi LogðxÞ þ 2 2T

1 T d2 ðx; TÞ ¼ pffiffiffiffiffiffi LogðxÞ  2 2T Z d 1 2 UðdÞ ¼ pffiffiffiffi e g d g

p

ð15aÞ ð15bÞ ð15cÞ

1

Notice that hðx ¼ 0; TÞ ¼ 0. The form of hðx; TÞ is very similar to Black–Scholes. It differs in at least two important ways: x ¼ BP is dimensionless and the interpretation of hðx; TÞ. Remembering that x ¼ BP and HðB; P; TÞ ¼ P hðx; TÞ; the expression of HðB; P; TÞ in terms of the value of the evolutionary technology P and the value of the higher cost technology B, can be deduced from Eq. (14):

      B B HðB; P; TÞ ¼ B  U d1 ;T  P  U d2 ;T P P

ð16Þ

T ¼ ðr2  2qrd þ d2 Þs is the cumulative uncertainty over the time horizon ‘‘s”. When r  r, T r2 s. When the variability is zero, Eq. (16) becomes:

HðB; P; 0Þ ¼ Max½0; P  B . In Eq. (16), HðB; P; TÞ is the extra value of using high technology in risk neutral condition. If the premium associated with high technology, is exactly equal to HðB; P; TÞ, the investor is in a ‘‘risk neutral” situation. One way to interpret Eq. (16) is to interpret HðB; P; TÞ as the value of the option of investing in the revolutionary technology instead of the evolutionary one and to treat W H ¼ W REV  W EVO as the value of the premium that should be paid to accomplish higher network performance, under the assumption of risk neutrality. HðP; BÞ quantifies the maximum premium that should be paid to reduce the uncertainty associated with the evolutionary approach to technology migration. In other words, as long as the actual value of the premium paid for the higher network performance is smaller than HðP; BÞ, it is more advantageous to go for the revolutionary technology. Since our real option model can assess the transition value from old technology to new technology, we attempt to test the value of technology transition, from generation to generation (inter-generation transition), and within the same generations (intra-generation transition). For example, the parameters for our model consist of H (option value), B (the value new technology), P (the value of old technology) and T (accumulated uncertainty). Market shares in wireless market are used as the value of B and P because we consider that the value of technology is proportional to market power (i.e., market share). Modern, complex technologies often display increasing returns to adoption in that the more it is adopted, the more experience is gained with it and the more it is improved (which is directly connected to the value of technology) (Arthur, 1989).

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

187

5. Scenarios and data Fig. 4 illustrates the overall design of this study to determine the best technology transition path. Two types of technology migration paths are identified: (1) Inter-Generational Technology Migration Path and (2) Intra-Generational Technology Migration Path. First, Inter-Generational Technology Migration Path deals with moving from one generation technology to another, for example, analog-to-TDMA, analog-to-GSM, and analog-to-CDMA. The other type, Intra-Generational Technology Migration Path, i.e., movement within the same generation technology, includes cases such as TDMA-to-GSM, TDMA-CDMA, and GSMto-CDMA. Based on this structure, a total of sixteen scenarios have been constructed. The scenarios demonstrate the possible transitions of wireless technology and how they might change the value of networks. These scenarios are based on assumptions that had to be made about the future. That the future follows these suggestions is extremely unlikely, but still the scenarios may provide a firm’s manager with some new views and foster the own creativity in thinking about the influence of new technology. Several assumptions are applied when we construct these scenarios as follows: First, it is impossible to backward technologically. That is, a firm always prefers new technologies instead of old technologies. Second, a firm can only use one technology when it decides to migrate. Third, there is no limitation to technological choice. At present, GSM is standardized in Europe, but we allow that any technology can be chosen, as is the case in the US. Based on these assumptions, the following alternative technology migration paths are developed based on the wireless technology options diagram (Fig. 1), addressed in the previous section.                

Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario Scenario

1: Analog => TDMA => WCDMA 2: Analog => TDMA => cdma2000 3: Analog => TDMA => GSM => WCDMA 4: Analog => TDMA => GSM => cdma2000 5: Analog => TDMA => GSM => CDMA => WCDMA 6: Analog => TDMA => GSM => CDMA => cdma2000 7: Analog => TDMA => CDMA => WCDMA 8: Analog => TDMA => CDMA => cdma2000 9: Analog => GSM => WCDMA 10: Analog => GSM => cdma2000 11: Analog => GSM => CDMA => WCDMA 12: Analog => GSM => CDMA => cdma2000 13: Analog => CDMA => WCDMA 14: Analog => CDMA => cdma2000 15: Analog => WCDMA 16: Analog => cdma2000

Fig. 4. Research design.

188

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

Total GSM CDMA TDMA

19

19

93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04

Analog

92

Subscribers(Million)

US Wireless Industry 180.00 160.00 140.00 120.00 100.00 80.00 60.00 40.00 20.00 0.00

Year Fig. 5. US wireless market size.

US Wireless Market Share Market Share

100% 80% Analog 60%

TDMA CDM A

40%

GSM 20%

19 8 19 5 8 19 6 8 19 7 8 19 8 8 19 9 9 19 0 9 19 1 9 19 2 9 19 3 9 19 4 9 19 5 9 19 6 9 19 7 9 19 8 9 20 9 0 20 0 0 20 1 0 20 2 0 20 3 04

0%

Year Fig. 6. US wireless market share.

Fig. 5 plots the number of subscribers in each wireless technology from 1992 to 2004in the US. Unlike GSM’s dominant position in world wireless market, CDMA has experienced high growth and dominates US wireless market. TDMA also covers high market share, but will eventually obsolete as providers upgrade to more advanced technologies, such as GSM, GPRS, EDGE, and WCDMA. Analog is completely phased out after 2004 in the US wireless market. Based on the number of subscribers in generation (Fig. 5), Fig. 6 shows market shares for the various technologies in the US wireless industry. It provides a better picture of the relative size of US wireless market. The chart clearly shows the dramatic growth in CDMA and TDMA, while analog fades away. 6. Results and analysis Before we get into specific scenario results, it is worth making a note on interpreting the graphs in the subsequent sections. They are based on market share data, which represent actual consumer behavior and are thus backward looking, rather than on expected market share, which are forward looking. Thus these graphs do not have strong predictive power, but, in line with the objectives of the paper, are intended to illustrate how real options can be applied. 6.1. Intra-generational technology transition (2G => 2G) The scenario (Fig. 7) displays the value curve when moving from TDMA to GSM network technology. This analysis shows that the transition is undesirable because the premium value is positive continuously and the option value is always negative. Since the net option value fluctuates in the level of negative over time, transition should be delayed or never. Since TDMA and GSM is similar technology and do not need to invest in this transition. However, in reality, operators prefers to transit from TDMA to GSM as a stepping stone evolution, like AT&T Wireless. Another 2G scenario (Fig. 8) is the transition from TDMA to CDMA network technology. The premium value decreases rapidly and then decreases continuously because of CDMA’s popularity in the market. NOV is positive starting in 2001, and increases continually. NOV is achieved a peak in 2003 and then decreases gradually. So, the transition from TDMA to CDMA is most desirable in 2003 and less desirable after that, although NOV is positive. Fig. 9 shows the movement from GSM to CDMA network technology. This transition is recommended because the premium value is initially negative and continues to steadily negative and option value is positive continually. However,

189

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

Value Curve (TDMA=>GSM, US) 0.3000 0.2000 0.1000

10

Net Value

20

09

07

08

20

20

20

06 20

04

03

05 20

20

02

20

20

20

-0.1000

Premium 01

Value

OV 0.0000

-0.2000 -0.3000

Year Fig. 7. The value curve of technology transition (TDMA to GSM).

Value Curve (TDMA=>CDMA, US) 0.2500 0.2000 0.1500

Value

0.1000

OV

0.0500

Premium 0.0000 10 20

20

08

06 20

04 20

02 20

00 20

98

Net Value

19

-0.0500 -0.1000 -0.1500 -0.2000

Year Fig. 8. The value curve of technology transition (TDMA to CDMA).

Value Curve (GSM=>CDMA, US) 0.6000 0.5000 0.4000

20 10

20 09

-0.2000

20 08

20 01

-0.1000

20 07

Net Value 20 06

0.0000 20 05

Premium

20 04

OV

0.1000

20 03

0.2000

20 02

Value

0.3000

-0.3000 -0.4000

Year Fig. 9. The value curve of technology transition (GSM to CDMA).

NOV decreases gradually after a peak of 2003. So, the transition to move CDMA from GSM is desirable. This result is completely different from world market. This difference is clear because GSM dominates the market (over 70%) in world, while CDMA is more popular than GSM in the US market.

190

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

Value Curve (GSM=>WCDMA, US) 0.2500 0.2000 0.1500

Value

0.1000 0.0500 0.0000 -0.0500

2004

2005

2006

2007

2008

2009

2010

OV Premium NOV

-0.1000 -0.1500 -0.2000

Year Fig. 10. The value curve of technology transition (GSM-WCDMA).

Value Curve (CDMA=>cdma2000, US) 0.6000 0.4000

Value

0.2000 OV Premium

0.0000 2004

2005

2006

2007

2008

2009

2010

NOV

-0.2000 -0.4000 -0.6000

Year Fig. 11. The value curve of technology transition (CDMA-cdma2000).

6.2. Inter-generational technology transition (2 => 3) Fig. 10 shows the transition from GSM to WCDMA (3G) network technology. The premium value decreases continuously, and finally is negative after 2008. The option value is steadily negative, but positive after 2009. NOV is initially negative, but highly increases and positive after 2009. So, the transition is desirable starting in 2009 or later. The next scenario (Fig. 11) displays the value curve when moving from CDMA to cdma2000 network technology. These results show that the transition is undesirable because the premium value is positive continuously until 2010 (saturation point) and the option value is always negative. Since the net option value increases in the level of negative over time, so transition should be delayed or never. 7. Concluding remarks 7.1. Summary In this paper, we investigated the historical evolution of wireless technologies from the first generation (1G) to the second generation (2G) and the third generation (3G) wireless network technologies. Based on the real options approach (ROA), we also developed a model to assess the transition (replacement) from old technology (premium value) to new technology (option value). Finally, we assessed the migration options of wireless network technologies in the US using our technology transition assessment model as a case study. The results of all technology transition scenarios in the US wireless industry are summarized in Fig. 12. First, moving from analog to any 2G technology is desirable; however, the best choice for analog carriers is to move to CDMA in 2004 because it results in the highest option value (0.6978) of the three possibilities. Second, it is not desirable for the TDMA carrier to move into GSM because all transition values are negative. But CDMA is desirable because of the positive option value of 0.1972 in 2003. Note that this changes when market data from the world is used instead of just US data. Third, concerning the transition from TDMA to 3G technologies, there is not much difference in transition option value between WCDMA (0.0372) and cdma2000 (0.0289) in 2010. In the case of GSM carriers, moving to 2G CDMA is recommended

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

191

Fig. 12. Technology migration path diagram.

because of the positive transition option value (0.4654) in 2003, but, in reality, this transition costs are excessive and the technologies are incompatible. This is a limitation of this study since only market data is available for technology assessment. As with TDMA, the transition from GSM to 3G has a similar positive option value for WCDMA (0.1928) and cdma2000 (0.1840) in 2010. However, the majority of the GSM carriers is from Europe and only considers WCDMA migration for technical and political reasons. CDMA carriers do not consider 3G until 2010 because of the continuing negative transition values, but the transition will occur some time after arriving at the saturation point of current 2G CDMA market. 7.2. Strategic implications and limitations The findings of the study imply that strategic technology choice is extremely important determinant of firm’s competitiveness. Exploring the dimensions of strategic decisions proved to be valuable, as the study found that it is important for a firm to have strategic flexibility is extremely high for improving a firm’s value. The study also found that strategic technology choice is important regardless of the level of environmental uncertainty faced by the firm. Since the next generation wireless network technologies and architectures are still a subject of debate with no substantial implementation results, there is much work to do. With the further research, the scope of study can be expanded. However, this paper did not discuss some important factors, such as wireless spectrum allocation issues, political issues, and business environment, which might affect on the decision of wireless carriers. We recognize that the paper is still the fledgling stage to use our model into decision-making in complex business environment. 7.3. Future research The possibilities for future research on topics related to strategic technology management using the real options approach are extensive. Of them, a few of the possible extensions of the ideas covered in this paper. First, the US market with a suite of different technologies in use offers an interesting laboratory to test the real options approach as a strategic decision tool. Based on this preliminary practice of our real options model, we would like to develop a theory for a firm’s behavior analysis to solve strategic issues in the company level: for example, why do not all of the firms in wireless network industry to migrate or upgrade for the 3G services at the same time? Or why did some companies choose WCDMA instead of cdma2000 (i.e., AT&T wireless and Cingular), or else? Second, real option research is still very much a growing area. Thus there is much more that needs to be done. Although the conceptual foundation for real options is well established, there is scope for further research extensions to some of the basic theories, especially relating to valuation techniques. Options involving real technology choices and strategies are generally much more complex than simple financial options in stock market. First, the uncertainty may be due to several variables instead of simply one variable such as the price in financial options. Further, it may not always be easy to measure the value of underlying assets because of its dynamics and never traded in the market. These complexities may not allow one to find exact valuation model. Third, the other future research to come from this study will be the application of our real option theory and techniques to a variety of other industry to solve technology management problems, such as high-tech industry and medical industry. Conceptually any technology choice decision where significant uncertainties are present can be considered our strategic technology transition model using the real options approach. Finally, we hope that this study will take the form of helping wireless network service providers for a strategic decision to upgrade or migrate for the next generation network technologies and architectures, by resolving the ambiguity of the nature of network evolution. Finally, since still the areas of the next generation wireless network technologies and architectures

192

H.J. Kim et al. / Telematics and Informatics 26 (2009) 180–192

remains in its debating stages of development with no substantial implementation results, there is much work to do. With the further research, the scope of study can be developed. References Alleman, J., Noam, E., 1999. Real Options: The New Investment Theory and its Implications for Telecommunications Economics. Kluwer Academic Publishers. Amram, M., Kulatilaka, N., 1998. Real Options: Managing Strategic Investment in an Uncertain World. Harvard Business School Press. Arthur, W.B., 1989. Competing technologies, increasing returns and lock-in by historical events. Economic Journal 99, 116–131. Basili, M., Fontini, F., 2003. The Option Value of the UK 3G Telecom licenses. The Journal of Policy, Regulation and Strategy for Telecommunications 5 (3), 48– 52. Benaroch, M., 2002. Managing investments in information technology based on real options theory. Journal of MIS 19 (2), 43–84. Benaroch, M., Kauffman, H., 2000. Justifying electronic banking network expansion using real options analysis. MIS Quarterly 24 (2), 197–225. Carsello, R. et al, 1997. IMT-2000 standards: radio aspects. IEEE Personal Communications, 30–40. Carr, P., 1988. The Valuation of sequential exchange opportunities. Journal of Finance 23 (5), 1235–1256. d’Halluin, Y., Forsyth, A., Vetzal, R., 2002. Managing capacity for telecommunications network under uncertainty. IEEE/ACM Transactions on Networking 10 (4), 579–588. Dixit, A., Pindyck, R., 1994. Investment Under Uncertainty. Princeton University Press. Dixit, A.K., Pindyck, R.S., 1995. The Options Approach to Capital Investment, Harvard Business Review. Dahlman, E. et al, 1998. UMTS/IMT-2000 based on wideband CDMA. IEEE Communication Magazine, 70–80. Economides, N., 1999. Real Options and the Cost of the Local Telecommunications Network. In: Alleman, James, Noam, Eli (Eds.), The New Investment Theory of real options and its Implications for the Cost Models in Telecommunications. Kluwer Academic Publishers. Garg, V., 2001. Wireless Network Evolution: 2G to 3G, Prentice Hall Communications Engineering and Emerging Technology Series. Kim, Hak-Ju, 2004. The Real Options to technology management: Strategic Options for the 3G Wireless Network Architecture and Technologies, Ph.D. Dissertation. Kim, Y., Sanders, L., 2002. Strategic actions in information technology investment based on real option theory. Decision Support Systems 33, 1–11. Margrabe, W., 1978. The value of an option to exchange one asset for another. Journal of Finance 33, 177–186. McDonald, R., Siegel, D., 1986. The value of waiting to invest. Quarterly Journal of Economics 101, 707–727. Pateli, A., Giaglis, G. 2007. The impact of value on governance decisions for IT-based alliances: evidence from a joint venture in the wireless networks industry. In: Proceedings on the 40th Annual Hawaii International Conference on System Sciences, HICSS 2007, pp. 175b–175b. Pindyck, R.S., 1989. Irreversible investment, capacity choice, and the value of the firm. American Economic Review 2, 969–985. Prasad, R., 1998. An overview of CDMA evolution toward wideband CDMA, IEEE Communication Surveys, Fourth Quarter 1998, Vol. 1, No. 1, pp.1–29. Rapport, T., 1996. Wireless Communications Principles and Practice. Prentice-Hall, Inc. Trigeorgis, L., 1996. Real Options: Managerial Flexibility and Strategy in Resource Allocation. The MIT Press, Cambridge, Mass.