WHAT IS NOT A REAL OPTION: CONSIDERING BOUNDARIES FOR THE APPLICATION OF REAL OPTIONS TO BUSINESS STRATEGY

Page created by Paul Morris
 
CONTINUE READING
姝 Academy of Management Review
2004, Vol. 29, No. 1, 74–85.

                      WHAT IS NOT A REAL OPTION: CONSIDERING
                      BOUNDARIES FOR THE APPLICATION OF REAL
                          OPTIONS TO BUSINESS STRATEGY
                                                         RON ADNER
                                                           INSEAD

                                             DANIEL A. LEVINTHAL
                                  The Wharton School, University of Pennsylvania

                      We argue that the greater the extent to which choice sets evolve as a consequence of
                      firms’ exploration activities, the less structured the firms’ abandonment decisions
                      become and, in turn, the less distinguishable a real option is from more generic
                      notions of path dependence—a sequential stream of investment in and of itself does
                      not constitute a real option. While organizational adaptations can extend the appli-
                      cability of real options, they impose tradeoffs that may lead to the underutilization of
                      discoveries made in the course of exploration.

   The concept of real options has created con-                      behavioral theory of the firm (Cyert & March,
siderable excitement in the management litera-                       1963) or evolutionary economics (Dosi, 1982; Nel-
ture in recent years (Kogut & Kulatilaka, 2001;                      son & Winter, 1982; Penrose, 1959) who have long
McGrath, 1997). The appeal of real options is                        been sensitive to path dependence. What is dis-
quite natural. Firms face uncertain futures, and                     tinctive about the use of real options as both an
the investment opportunities they face are, to an                    analytic tool (Dixit & Pindyck, 1994; Trigeorgis,
important degree, a function of their prior in-                      1997) and a strategic heuristic (Luehrman, 1998;
vestment commitments. Thus, at a surface level,                      McGrath, 1997) is that the real options logic of-
the real options framework appears to precisely                      fers the prospect of assigning actual values (car-
fit firms’ strategic challenges by linking current                   dinal in the former case or ordinal in the latter
actions to uncertain futures. In addition, relative                  case) to stage-setting investments.
to other justifications for exploration, such as                        However, while the logic can always be used
rationales for slack search (March & Simon,                          to generate values and orderings, the validity of
1958), real options provide a rhetoric that com-                     these results depends on some key assumptions.
fortably fits the standard language of corporate                     The technical violations of strict real options
finance. Less obvious than the benefits offered                      assumptions and the methods for their accom-
by the real options framework but equally im-                        modation are, by now, well known (e.g., Cope-
portant are its inherent limitations in guiding                      land & Antikarnov, 2001). More subtle, however,
organizations’ decision making under uncertainty.                    and more limiting to the validity of real options
   The underlying logic of the real options frame-                   as a tool in organizational decision making are
work is based on the realization that future in-                     violations that stem from organizational pro-
vestment opportunities are contingent on prior                       cesses in the face of different modes of uncer-
investment commitments. Thus, in contrast to                         tainty resolution.
net present value analysis, real options analysis                       The real options framework is intended to ex-
accounts for the sequential nature of choice pro-                    ploit the flexibility inherent in sequential in-
cesses. Such an observation is not alien to re-                      vestments. We argue that this flexibility stems
searchers working within the tradition of the                        from the possibility of abandoning investment
                                                                     initiatives, rather than from the simple substitu-
                                                                     tion of a stream of smaller payments for a larger
   We thank Bert Cannella, Javier Gimeno, Bruce Kogut,               lump sum payment. This implies that the effec-
Christoph Loch, Anand Swaminathan, Peter Zemsky, partic-
ipants at the 2001 Strategy Research Forum and the Jones
                                                                     tive management of real options requires a high
Center Workshop at Wharton, and our three reviewers for              degree of rigidity in the specification of the
their valuable comments on prior drafts of this manuscript.          agenda of the initiatives and the criteria for
                                                                74
2004                                        Adner and Levinthal                                        75

their success. We consider the implications, pos-      approaches and more generic path-dependent
itive and negative, of such constraints and com-       processes, and we suggest how they might be
pare them to more generic path-dependent ap-           distinguished empirically.
proaches to managing investment under
uncertainty.
                                                                       REAL OPTIONS
   In the next section we briefly lay out the for-
mal structure of real options. We then examine            Real options investments are characterized by
some of the critical assumptions that underlie         sequential, irreversible investments made un-
the application of options to firms’ investment        der conditions of uncertainty (Dixit & Pindyck,
decisions. When uncertainty resolution emerges         1994). The framework suggests that purchasing
as an outcome of firm action, the sharp temporal       a real option on a strategically important oppor-
demarcation made in the options literature be-         tunity allows firms to postpone commitment un-
tween “Stage 1” and “Stage 2” investments is           til a substantial portion of the uncertainty about
called into question. In such settings, beyond         the opportunity has been resolved. After making
learning about a specific initiative, the flexibil-    an initial investment, management is then to
ity associated with later investment decisions         turn its attention to other matters and wait for a
often stems from the possibility of discovering a      signal as to whether or not it is appropriate to
wide variety of related opportunities, even in the     harvest or cultivate the initial investment.
face of unfavorable initial outcomes. We show             Consider the events that transpire prior to the
that the greater the extent to which initiatives       exercise of financial options, on which the real
are open ended, the more problematic the appli-        options model is based (Figure 1). First, an in-
cation of the real options framework is. Flexibil-     vestor purchases an option (Stage 1). Then, dur-
ity in search can undermine the flexibility asso-      ing the course of the holding period, the value of
ciated with abandonment. Abandonment is                the option changes in response to external
essential for limiting downside risk, a key virtue     events. Throughout the holding period, the fi-
attributed to real option investments. We con-         nancial markets provide a clear signal as to the
sider how organizational processes, such as the        current value of the option. Finally, events tran-
allocation of decision rights that limit the range     spire so that the investor chooses to exercise the
of possible action and the specification of well-      option, or, alternatively, the expiration date
defined temporal boundaries, can extend the            specified in the option contract is reached and
applicability of real options reasoning but force      the option expires (Stage 2). Investments with
strategic tradeoffs of their own. We conclude by       this structure are optionlike in that Stage 2 in-
considering differences between real options           vestments are not a necessary consequence of

                                               FIGURE 1
                                     The Structure of Real Options
76                                   Academy of Management Review                                 January

having made an initial Stage 1 investment but,                        FIGURE 2a
rather, can be conditioned on the realization of       Boundaries of Applicability for Net Present
interim information.                                           Value and Real Options
   Note two critical features of this process: (1)
the value of the option (and the underlying as-
set) is exogenous to the investor’s activity—the
investor cannot take steps to make the intrinsic
characteristics of the asset more attractive; (2)
the market signal of option value is readily ob-
servable and is independent of the investor’s
behavior. If these properties carry over to an
investment of a nonfinancial sort (such as in
plant and equipment, or in technology licenses),
then the logic of options can directly carry over.
The greater the extent to which these properties
are violated, the more problematic the applica-                       FIGURE 2b
tion of an options framework is.                      Boundaries of Applicability for Real Options
   The “wait and see” setting of financial options        and Path-Dependent Opportunities
represents the extreme case of investment envi-
ronments, but one for which the methodology is
ideally suited. Because much attention in the
management literature is focused on the ways
in which the firm can affect outcomes and vari-
ances (e.g., McGrath, 1997), it is important to
examine what happens to the applicability of
options logic as we move away from a world of
wait and see to a world of “act and see,” in
which uncertainty resolution is endogenous to
firm activity.

       BOUNDARIES OF REAL OPTIONS
                                                     the application of real options becomes more
   The boundaries of the logic of real options are   challenging analytically and, focal to this pa-
often considered in relation to the breakdown of     per, organizationally.
net present value analysis: to the degree that          When target markets and technical agendas
investment choices have the property of high         are flexible (see Figure 2b), the discrete invest-
uncertainty and irreversibility, a real options      ment logic of real options is eroded, and activities
valuation provides a better characterization of      may be characterized more appropriately as more
the investment’s true value than does a net          generic path-dependent processes that fall under
present value calculation because of the latter’s    such labels as probe and learn (Lynn, Morone, &
inability to account for the value of delaying       Paulson, 1996), incremental search (March & Si-
commitments (see Figure 2a).                         mon, 1958; Nelson & Winter, 1982), or innovation
   Less examined but no less important are the       journeys (Van de Ven, Pollery, Garud, & Venkat-
boundaries along which real options logic is         araman, 1999). Alternatively, if the scope of the
itself strained. As we move from a world of real     option investment is fixed a priori—that is, if the
options on tradable assets, in which the firm has    opportunities on which one is taking an option can
no hand in resolving uncertainty and the set of      be clearly specified at the inception of the op-
possible actions in response to this uncertainty     tion—then the decision to abandon an initiative
resolution can be specified at the time of the       can be clearly articulated and the flexibility asso-
initial investment, to real options on strategic     ciated with an option investment can be readily
opportunities whose outcomes are intimately          maintained.
linked to firm action, the clean demarcations           In the context of financial options, one can
between investment stages begin to blur and          clearly state a priori when a given option will be
2004                                                  Adner and Levinthal                                          77

“in the money” and worth exercising. However,                     lenges to abandoning options that can deter
in the case of real options on strategic opportu-                 firms from exercising the very flexibility that
nities in which new possibilities are identified                  made the real options approach attractive in the
as a consequence of a firm’s actions, such a                      first place.
prior specification may not be possible or even                      The degree to which an initiative is circum-
desirable. Experiments, even unsuccessful ones,                   scribed is itself dependent on organizational
not only provide information about intended in-                   choices concerning project scope. If an initiative
vestment paths but also provide information                       is highly circumscribed in terms of what appro-
about other possibilities—possibilities that may                  priate target markets are and what temporal
not even have been envisioned at the time of the                  and technical boundaries must be respected,
initial investments. Outcomes that are negative                   then the degree of flexibility in the potential
vis-à-vis the initial proposed initiative may still              directions of the initiative is limited, but at a
suggest or engender other possible actions. For                   price. Imposing rigid criteria for abandonment
instance, the failure of a technology develop-                    may result in the underutilization of discoveries
ment effort to reach a given technical hurdle in                  made in the context of initiatives that are fail-
a specified time horizon does not preclude addi-                  ures with respect to their initial agenda but that
tional efforts or different approaches; similarly,                introduce promising possibilities not previously
if a new product fails to win acceptance in a                     imagined.
given target market, it may still be successful in                   This highlights the need for proponents of real
other possible target markets.1                                   options logic to consider a more nuanced organ-
   At a basic logical level, the boundless set of                 izational perspective that incorporates the dif-
possibilities associated with a strategic invest-                 ferent views that exist within an organization.
ment initiative presents a problem analogous to                   The firm cannot be regarded as a unitary actor.
that posed by Popper (1959) in the context of                     Managers charged with pursuing an opportu-
hypothesis testing. Popper’s proposition is that a                nity and executives charged with evaluating a
hypothesis is never proved; rather, it is not dis-                portfolio of opportunities will differ in their be-
proved in any given test. Conversely, in a given                  liefs as to when an option ceases to be attrac-
attempt to realize an opportunity, we can only                    tive. Such differences are a function of their in-
prove that an initiative can be successful in a                   centives and opportunity structures, and they
particular setting; we cannot use negative out-                   serve to compound the psychological biases re-
comes to preclude the possibility that it will be                 garding sunk costs and escalating commitment
successful in other target markets or with fur-                   that can act to deter strategic redirection.
ther technical refinements.                                          These considerations are largely overlooked
   This “impossibility of proving failure” is an                  in the theoretical (e.g., Dixit & Pindyck, 1994;
inherent feature of firm initiatives under uncer-                 Trigeorgis, 1997) and managerial (e.g., Amram &
tainty and poses a fundamental challenge to the                   Kulatilaka, 1998; Luehrman, 1998) work on real
applicability of the real options framework. Fur-                 options in strategy, in which researchers tend to
ther, because actors at different levels of an                    treat the firm as a monolithic actor. They are
organization have different perspectives on the                   also largely overlooked in empirical work in
attractiveness of a given opportunity, they will                  which scholars tend to look at investment deci-
disagree as to the proper framing of termination                  sions that are not integrated within the firm’s
decisions. Thus, the open-ended nature of the                     organizational activities (e.g., investments in ac-
search for success raises organizational chal-                    quisitions and joint ventures). The importance of
                                                                  organizational factors in determining the appli-
                                                                  cability of options logic increases as real op-
   1
     Knight’s (1921) distinction between risk and uncertainty,
                                                                  tions theory is extended from the evaluation of
which lies at the heart of notions of entrepreneurship (Kirz-     investments in physical assets, for which the
ner, 1997), is closely related to the distinction we are making   resolution of uncertainty is exogenous to firm
here. Real options are well suited to incorporate risk, an        action and the scope of possible firm response to
uncertain realization from a well-specified probability dis-
                                                                  this uncertainty is relatively constrained (e.g.,
tribution. In contrast, the inherent unknowability that char-
acterizes Knightian uncertainty poses a significant chal-         making use of a plant in a foreign country to
lenge to characterizing means and variances of key option         potentially take advantage of swings in relative
analysis parameters.                                              wages), to the evaluation of investments in stra-
78                                   Academy of Management Review                                  January

tegic opportunities, for which the resolution of     to the asymmetry of proving and disproving a
uncertainty is largely endogenous to firm action,    hypothesis in empirical research. Popper (1959)
and the scope for possible modifications in the      argues that it is logically impossible to prove
initial initiative is vast.                          that a hypothesis is true; we can only show that
                                                     it is not false, up to a specified probability level.
                                                     In contrast, a managerial initiative can only
       CHALLENGE OF ABANDONMENT                      be shown to be successful in that some market
                                                     or technical criteria are met; an initiative,
Ambiguity of Stages in Real Options                  such as the development of a new technology,
   An investment’s flexibility is revealed in the    cannot conclusively be shown to be incapable
project abandonment decision. Options are flex-      of succeeding.
ible not because they substitute a stream of            Consider what we learn when a new technol-
smaller payment for a larger lump sum payment        ogy is “tried.” We may learn that, in a given time
but because the payment stream can be aban-          period and with a given level of investment,
doned in light of negative outcomes. The less        researchers were not able to meet certain tech-
well-defined the abandonment decision, the           nical milestones. Similarly, we may learn that a
less valid the perception of flexibility and the     particular market context was not receptive to
less appropriate the nominal application of real     the technology in its current form. However,
options logic. In the absence of proofs of failure   these results are not impossibility theorems.
and strict, structured abandonment deadlines,        They do not demonstrate that, with more time
firms face a difficult organizational challenge to   and more resources, these technical milestones
exploit the flexibility offered by the sequential    could not be met. Nor do such results demon-
nature of optionlike investments. As noted ear-      strate that other market settings might not re-
lier, there is an important asymmetry between        spond positively to the technology in its current
positive and negative signals that is analogous      form or that enhancements to the technology or

                                             FIGURE 3
                    Malleability of Technical and Market Focus to Firm Activity
2004                                                  Adner and Levinthal                                                       79

a reduction in cost will not lead to success in the               efforts to enhance the value of the initial initia-
current market setting.                                           tive, time to expiration becomes an endogenous
  Thus, the significance of any specific market                   choice. From a valuation perspective, longer op-
response is constrained in market space to a                      tion durations lead to higher option values and
specific segment, and constrained in time to a                    increase the risk of overvaluation. From an or-
specific development state. Figure 3 maps the                     ganizational perspective, flexibility in duration
potential malleability of a project to interim                    leads to negotiations over termination criteria,
feedback, in terms of the opportunity to affect                   which increase the risk that firm investments
the technical agenda of the project and the mar-                  will be governed by influence processes and
kets in which the project will be evaluated, after                idiosyncratic justification, rather than by a co-
an option has been purchased. Again, what is                      herent portfolio strategy.
critical is that these new directions need not                       A fundamental difficulty raised by broad flex-
have been identified at the time of the inception                 ibility in response to interim signals is that
of the initial option investment. They were not                   while real options valuation techniques are
part of the justification for the initiation of the               well tailored to offering go/no-go guidance on
option, but they became critical for the continu-                 project initiations, they are relatively equivo-
ation of the investment effort. The greater the                   cal regarding what to do after an initial “go”
potential for ex post discovery of possible new                   recommendation.
directions for the original initiative, the greater                  Clearly, the real options valuation framework
the difficulty in deciding when to abandon the                    can be reapplied to the evaluation of a given
opportunity.                                                      initiative at a future time, when some uncer-
  The problem of abandonment is compounded                        tainty is resolved (e.g., compound real options).
by the fact that most options on strategic oppor-                 Structured reevaluation, however, becomes in-
tunities do not have an explicit, exogenous ex-                   creasingly difficult as uncertainty gets resolved
piration date. How long the firm can keep the                     in increasingly incremental steps that are a
option open, whether in persisting in develop-                    function of firm activity.3 Incremental discovery
ment efforts in nominally promising pharma-                       leads to difficulty in drawing precise distinc-
ceuticals or in market creation efforts for emerg-                tions among a series of independent options, a
ing technologies, is often difficult to specify ex                single compound option, and more generic path-
ante.2 When there is no explicit time limit on                    dependent search activities. The division of time
                                                                  into discrete decision points is typically artifi-
                                                                  cial. Does the pharmaceutical firm that started a
   2
     The fact that a real option may lack an explicit expira-     research project on Monday make a conscious
tion date need not imply that the option persists indefinitely.   choice on Tuesday to continue with the project?
In particular, competitive forces may result in the effective
                                                                  Ultimately, the structure of the discovery activity
expiration of an option. However, even in cases where the
competitive environment drives option expiration (e.g.,
McGrath, 1997; Trigeorgis, 1997), the timing of expiration
cannot be specified ex ante. In addition, when the opportu-       are impossible to assess with accuracy for many strategi-
nity can be exploited in multiple markets, a rival’s entry into   cally important opportunities. Furthermore, for analytic trac-
one market does not necessarily lead to the expiration of the     tability, these models necessarily ignore many important
option in other markets, so even in these settings ex post        factors, such as the complexity and duration of competitive
endogeneity is present. Strategic interactions also present       responses, identified in recent work on competitive dynam-
another dimension of endogeneity to the problem of evalu-         ics (e.g., Ferrier, Smith, & Grimm, 1999).
                                                                     3
ating options that leads to an additional set of challenges.           Trigeorgis (1997), for example, notes that a rich set of
As opposed to “standard” real options settings, where ex-         compound options poses analytic and computational chal-
pected payoffs are exogenously determined, “competitive”          lenges to the actual evaluation of real options. Consider that
real options analysis requires an accounting for the ways in      of the six factors that determine option value (value of un-
which the firm’s actions affect rivals’ responses. The benefits   derlying risky asset, exercise price, time to expiration, vol-
of delay and exploration that are derived from sequential         atility of the value of the underlying asset, foregone cash
investments must be weighted against the potential benefits       flows, discount rate), the first four are especially difficult to
offered by commitment and preemption (Ghemawat, 1991).            specify with confidence for the case of strategic opportu-
   Although still in infancy, theoretical attempts to marry       nities. See Bowman and Moskowitz (2001) for a reexami-
real options with game theory (e.g., Grenadier, 2001; Kulati-     nation of the effectiveness of the application of option
laka & Perotti, 1998) suggest that analytic results can be        analysis at Merck, and Lander and Pinches (1998) for a
obtained. These results, however, are very sensitive to un-       discussion of the modeling challenges to making real
derlying assumptions about parameter values — values that         options analysis “practical.”
80                                    Academy of Management Review                                         January

                                             FIGURE 4
                  ‘‘Option Traps’’ That Hinder the Abandonment of Opportunities

and the scope of the authority of the project team    positive in the sense of increasing the likelihood
define these distinctions and thereby draw the        of success, clearly hampers their ability to aban-
boundary of applicability for options logic.          don initiatives when the learning outcomes are
                                                      negative (Garud & Van de Ven, 1992). For this
                                                      reason, the way in which the selection and re-
Organizational Factors
                                                      source allocation mechanisms are manifested
   Because holding options open entails both or-      throughout the organizational hierarchy is fun-
ganizational and financial maintenance costs          damental to the challenge of exploiting the flex-
(Garud & Nayyar, 1994), managing abandon-             ibility inherent in real options.
ment has direct implications for firms’ success          These organizational drivers are compounded
in managing real options. The “option traps”          by psychological deterrents to abandonment.
presented in Figure 4 highlight the forces that       Given the difficulty firms have in incorporating
undermine timely project abandonment in dif-          the logic of sunk costs (Russo & Schoemaker,
ferent uncertainty resolution regimes. The chal-      1989), their tendency toward escalating commit-
lenge of project abandonment is complicated by        ments (Staw, 1981) and overconfidence (Camerer
the nature of organizational resource allocation      & Lovallo, 1999), the political impetus not to
(Bower, 1970) and the different perspectives and      show failure (McGrath, 1999; Sitkin, 1992), and
incentives of stakeholders at different levels in     the natural desire to succeed, the challenge of
the organization.                                     abandonment— giving up on an opportunity
   This contrast is most succinctly characterized     that has a chance for success—is a large one.4
as the difference between “holding the option”        This, essentially, is the dark side of managing
and “being the option.” For executives at high        projects with product champions (Maidique,
levels in the organization within whose purview       1980) and skunk works (Kanter, 1988), where the
lie a number of distinct initiatives, an individual   systems and support mechanisms put in place
initiative may have an optionlike quality in that     to create an impetus for starting innovations
the abandonment of the particular project may         act directly against their objective reassess-
not entail significant consequences. However,         ment and termination. Indeed, the challenges
the managers focused on that particular project       associated with abandoning projects can be
may see greater potential in its pursuit, both
because they are not aware of the larger set of
alternative investments available to the firm
                                                        4
and because of the career consequences associ-            Investment in joint ventures creates an organizational
                                                      boundary that should make it easier to monitor, separate,
ated with its abandonment. These consider-
                                                      and abandon an option investment (e.g., Kogut, 1991). Even
ations, in turn, act to increase these managers’      in this ideal setting, however, the evidence of firms’ effec-
dedication to achieving success with respect to       tiveness in managing exit and minimizing downside risk is
a particular initiative. This dedication, while       mixed (Reuer & Leiblein, 2000).
2004                                               Adner and Levinthal                                           81

greater than those associated with initiating                  exploiting the lessons learned from unsuc-
them (Brunsson, 1982; Garud & Van de Ven,                      cessful or only partially successful initiatives
1992).5                                                        (e.g., Sitkin, 1992).
   Real options logic makes a fundamental con-                    Our aim is not to dispute the importance of
tribution to the structuring of risk. When de-                 continued search activity. Indeed, we note that it
ployed in real organizations, however, this logic              is at the heart of well-admired processes such
must be complemented with appropriate con-                     as skunk works, probe and learn, and so forth.
trols (Block & MacMillan, 1993; McGrath & Mac-                 Rather, our aim is to highlight the difficulty such
Millan, 2000). These controls need to reflect the              processes raise for conceptualizing and manag-
evaluation traps created by flexible sets of pos-              ing real options. Specifically, the more freedom
sible responses to interim signals and the way                 afforded to an option manager (i.e., the more
in which initiatives may be usefully redirected.
                                                               active his or her role is in resolving uncertainty),
Given the vast set of possible paths that an
                                                               the more likely the option is to take on a life of
investment effort may take, the “flexibility” of
                                                               its own, independent of the requirement speci-
abandonment that is central to the evaluation of
                                                               fied by the logic governing the firm’s portfolio of
real options calls for an off-setting rigidity in
specifying the set of allowable courses of action              options. When the selection and evaluation cri-
at the time of the initial investment.                         teria differ across levels of the organization,
   The degree of flexibility in response to interim            which is increasingly the case as managers
signals is likely to vary with the unit of analysis            have more control and are more vested in their
being considered and where the boundaries of                   projects, freedom of action at the level of the
the decision-making entity lie. At the level of the            project is not consonant with the consistent ap-
firm, it is difficult to presume that the firm can             proach to portfolio management at the level of
commit itself to a narrow range of response to                 the firm. Making investment decisions under the
information that may emerge for an initial “op-                assumption of clear-cut abandonment points,
tion” investment. Indeed, such a commitment                    when organizational conditions are such that
may not be sequentially rational (Selten, 1975).               abandonment is unlikely to occur or unlikely to
However, at lower levels of the organization,                  be timely, is likely to render an initial real op-
subunits may be constrained in the range of                    tions valuation to be quite misleading. Since
initiatives that they pursue (Galunic & Eisen-                 real options framework is a theory of allocation
hardt, 1996), whether limited in the range of tar-             and control, this raises a question as to the mer-
get markets that can be addressed or the set of                its of its application under such conditions.
technical approaches that can be explored. It is                  In sum, firms invest in the active develop-
at these lower levels that real options may re-                ment of technology, and not in passive bets as
side. Disciplined project management can foster                to winners and losers. Given the impossibility
organizational environments that facilitate real               of proving failure and the absence of formal
options strategies. Projects can be artificially               expiration dates for an option investment, a
constrained through milestones, strict dead-
                                                               firm’s internal selection regime dictates its
lines, and regimented market focus to over-
                                                               ability to manage options. A well-managed
come the challenge of the impossibility of
                                                               real options strategy must necessarily guard
proving that all possible avenues of opportu-
                                                               against the natural momentum that builds up
nity are unpromising and, in turn, that the
abandonment of the initiative is appropriate.                  around hopeful activity.
Such practices, however, run the risk of under-                   Another way of characterizing our argument
                                                               about the importance of abandonment is to rec-
                                                               ognize that, from the perspective of the option
  5
    In our own conversations with managers in high-
                                                               initiative, the key is to avoid false negatives
technology firms known for the depth and breadth of their      (type II errors), in which valuable opportunities
innovative capacity, they consistently spoke of the chal-      are foregone. However, from the perspective of
lenge of shutting down projects. At one firm, “death row”      managing the option portfolio, the key is not to
was used to describe a set of projects that the organization
had formally recognized as in need of termination but
                                                               forget about avoiding false positives (type I er-
that, for various reasons, kept getting stays of execution     rors), which tie up resources that are better used
and surviving.                                                 elsewhere.
82                                   Academy of Management Review                                      January

     ALTERNATIVE SEARCH PROCESSES                       A natural question to ask is when might such
                                                     less specified discovery processes be preferred
   While the notion of real options as a way to
                                                     to the more explicit decision calculus of real
frame decision makers’ choices under uncer-
                                                     options? One complexity in considering this
tainty is relatively new to the management lit-
                                                     contrast is what one is to take as a character-
erature, the problem of organizations confront-
                                                     ization of a real options process. In our view, as
ing uncharted worlds is well established. A
                                                     a behavioral matter, the use of real options may
long-standing strand of the literature has high-
                                                     not vary so greatly from the behavior that stems
lighted the role and importance of search pro-       from slack search. Indeed, one of the merits of
cesses. Search may be local and problem              the real options framework is that it provides a
driven, but it also may comprise more opportu-       procedurally rational justification for processes
nistic, and less local, exploration. This latter     of slack search. In that sense, real options are a
notion of slack search (March & Simon, 1958) has     powerful tool for the “technology of foolishness”
been picked up and interpreted in terms of ideas     (March, 1988).6
of autonomous innovation (Burgelman, 1983) and          Alternatively, we can make the reference
skunk works (Kanter, 1988)—ideas that have not       point decision processes that more closely ad-
only been offered to capture existing behavior       here to the precepts of real options decision
but also held out as normative suggestions for       making. One virtue of the real options approach
how organizations may cope with uncertain            is that since the initial stage-setting invest-
futures.                                             ments are more explicit and, in turn, more visi-
   These images of search efforts, relatively        ble to higher-level actors within the organiza-
loosely controlled but modest in financial com-      tion, the organization will have a better sense of
mitments, lack the procedural rationality of real    its portfolio of initiatives. An organization may
options decision making. No explicit decision is     feel that it wants some initial exploration in a
made as to what constitutes the appropriate          variety of technologies or markets and would
amount of slack search at a given time; rather,      have a better sense of its overall exposure under
the organization is viewed as having, or is en-      the real options approach than under a system
couraged to develop, a heuristic or norm regard-     of slack search
ing the appropriate level of resources to be al-        The tradeoff to this benefit is, as we have
located to such efforts.                             argued above, that to make exploration initia-
   Another clear distinction between this broad      tives conform to the structure of a real option,
category of path-dependent principles and the        the boundaries of these initiatives should be
real options framework is that specific initia-      tightly specified ex ante. If an initiative to ex-
tives are not endorsed or necessarily examined       plore the value of a particular technology or
by higher-level actors within the organization.      particular market application wanders in re-
Slack is introduced into the resource allocation     sponse to feedback from initial efforts or shifts
system such that modest initiatives can occur        in markets and technologies, the discipline re-
without direct corporate oversight. Of course, if    quired by real options is lost. However, such
such initiatives identify opportunities that ap-     discovery processes are often heralded as valu-
pear to be worth pursuing and worth commer-          able means by which innovation occurs (Adner
cial development, it is likely that greater re-      & Levinthal, 2002; Day, 1990; Lynn et al., 1996).
source commitments will be necessary, and, as        While these more generic path-dependent pro-
a result, higher-level organizational actors will    cesses of exploration do not address the aban-
need to be convinced of the initiative’s merits.     donment challenges we have highlighted, un-
This commitment of more substantial resources        like real option approaches, they do not make
and the organizational approval of such com-         implicit assumptions about termination that are
mitments is akin to the “striking” of a option.      critical to their validity.
Again, though, it is important to note that the
initial initiative was never chosen by the or-
ganization; rather, the organization chose to al-      6
                                                         By “technology of foolishness,” March (1988) refers to
locate sufficient slack such that such initiatives   mechanisms that legitimate the allocation of resources for
might emerge.                                        search and discovery processes.
2004                                        Adner and Levinthal                                             83

  Clearly, to the extent that the potential paths          1. stricter action mandates for business units
a stage-setting investment may take are inher-                and project teams;
                                                           2. formalized milestones and go/no-go proce-
ently limited, the real options framework is more             dures;
readily applied. In addition, in such settings the         3. incentive systems, organizational cultures,
potential downside of its application in terms of             and allocation mechanisms that are more
foregone opportunities that are discovered ex                 tolerant of failure; and
post is sharply diminished. Such a perspective             4. review procedures that are more sensitive
                                                              to the presence of different incentives at
suggests that real options may be better suited               different levels of the organization.
to well-specified investments, such as overseas
production facilities and innovation licenses,            More broadly, in terms of empirical research
than to less structured opportunities, such as the     on the use of real options in organizations, we
development of new product technologies in the         would hope to see demonstrations of patterns of
face of a wide set of possible technical solutions     systematic and structured decision making that
and market applications.                               demonstrate divergence from both net present
                                                       value analysis and unstructured path depen-
                                                       dence, as well as data that consider project ter-
                 CONCLUSIONS
                                                       mination, not just the initiation and pacing of
   In some sense, our argument simply places           investment.
the notion of real options in the broader context         Firms face complex and highly uncertain in-
of sequential decision making. Actions are             vestment environments. The real options frame-
taken, beliefs are revised, and subsequent             work usefully highlights the links between cur-
choices are made. One is then left with the em-        rent actions and the set of future possibilities.
pirical question of how to distinguish real op-        However, to the extent that exit criteria are not
tions from the broader class of sequential deci-       well posed in a world of action and open-ended
sion-making processes. Demonstrating that              search, this framing can be overly seductive.
firms’ investment paths involve a sequential           While real options logic may justify investments
process of scaling up is, in itself, insufficient.     that would be rejected under the calculus of net
There are a wide variety of processes that would       present value, these “justified” investments may
generate such a pattern, such as adjust-               well destroy value when implicit assumptions
ment costs and time compression diseconomies           about abandonment flexibility are wrong. In set-
(Dierickx & Cool, 1989), and incremental learn-        tings where the range of responses to the reso-
ing (Cyert & March, 1963), that are not directly       lution of technical and market uncertainty is
linked to ideas of real options.                       largely unconstrained, the utility of applying op-
   As a result, we must move beyond measures           tions logic is unclear. For such strategic invest-
of the performance characteristics of firms’           ments, in contrast to financial options, exit cri-
project portfolios, whether average project            teria are not self-evident. Rather, in such
losses at real options firms are smaller (because      circumstances it may be more useful to identify
they have limited downside risk) or average            the possible sequence of experiments that will
project gains relative to losses are larger at real    test the most promising market and technical
options firms (because losses are limited but          paths available to the firm.
upside risk is unbounded). Rather, we should              To be clear, our purpose is not to question the
examine the way in which the portfolio is ad-          internal logic of real options. We do feel, how-
justed over time and, in particular, firms’ ap-        ever, that grouping all path-dependent activity
proaches to abandonment.                               under the real options label overextends real
   Thus, in terms of observed behavior, we would       options logic as tool, framework, and even as
expect real options firms, relative to their tradi-    metaphor, and undermines its effective applica-
tional counterparts, to                                tion. We believe that understanding a theory’s
                                                       boundaries serves to make it more powerful and
   1. abandon projects earlier and                     more precise. We argue that the cause of options
   2. have higher project abandonment rates.
                                                       thinking is best served in considering the
  Similarly, in terms of organizational pro-           boundaries of the domain of applicability of this
cesses, we would expect real options firms to          logic for business strategy, and in defining its
have                                                   place within the broader set of tools that are
84                                             Academy of Management Review                                             January

available to address decision making under un-                        Continual structuring by intertermporal technology
certainty. In this regard, we suggest that the                        transfer. Strategic Management Journal, 15: 365–385.
answer to the question “What is (and is not) a                    Garud, R., & Van de Ven, A. H. 1992. An empirical evaluation
real option?” has as much to do with the or-                         of the internal corporate venturing process. Strategic
                                                                     Management Journal, 13: 93–109.
ganization as it does with the opportunity.
                                                                  Ghemawat, P. 1991. Commitment. New York: Free Press.
                                                                  Grenadier, S. 2001. Game choices: The intersection of real
                      REFERENCES                                     options and game theory. London: Risk Books.
                                                                  Kanter, R. M. 1988. When a thousand flowers bloom: Struc-
Adner, R., & Levinthal, D. 2002. The emergence of emerging
                                                                     tural, collective, and social conditions for innovation in
   technologies. California Management Review, 45(1):
                                                                     organization. Research in Organizational Behavior, 10:
   50 – 66.
                                                                     169 –211.
Amram, M., & Kulatilaka, N. 1998. Real options: Managing
                                                                  Kirzner, I. M. 1997. Entrepreneurial discovery and the com-
   strategic investment in an uncertain world. Boston: Har-
                                                                      petitive market process: An Austrian approach. Journal
   vard Business School Press.
                                                                      of Economic Literature, 35: 60 – 85.
Block, Z., & MacMillan, I. 1993. Corporate venturing: Creating
                                                                  Knight, F. H. 1921. Risk, uncertainty and profit. Boston:
    new businesses within the firm. Boston: Harvard Busi-
                                                                      Houghton Mifflin.
    ness School Press.
                                                                  Kogut, B. 1991. Joint ventures and the option to expand and
Bower, J. L. 1970. Managing the resource allocation process.
                                                                     acquire. Management Science, 37: 19 –33.
   Boston: Harvard Business School Press.
                                                                  Kogut, B., & Kulatilaka, N. 2001. Capabilities as real options.
Bowman, E. H., & Moskowitz, G. T. 2001. Real options analy-
                                                                     Organization Science, 12: 744 –758.
   sis and strategic decision making. Organization Sci-
   ence, 12: 772–777.                                             Kulatilaka, N., & Perotti, A. 1998. Strategic growth options.
                                                                      Management Science, 44: 1021–1031.
Brunsson, N. 1982. The irrationality of action and action ra-
    tionality: Decisions, ideologies, and organizational ac-      Lander, D. M., & Pinches, G. E. 1998. Challenges to the prac-
    tions. Journal of Management Studies, 19: 29 –34.                tical implementation of modeling and valuing real op-
                                                                     tions. Quarterly Review of Economics and Finance, 38:
Burgelman, R. A. 1983. A process model of internal corporate
                                                                     537–567.
    venturing in the diversified major firm. Administrative
    Science Quarterly, 28: 223–244.                               Luehrman, T. A. 1998. Strategy as a portfolio of real options.
                                                                     Harvard Business Review, 76(5): 89 –99.
Camerer, C., & Lovallo, D. 1999. Overconfidence and excess
   entry: An experimental approach. American Economic             Lynn, G. S., Morone, J. G., & Paulson, A. S. 1996. Marketing
   Review, 89: 306 –319.                                             and discontinuous innovation: The probe and learn pro-
                                                                     cess. California Management Review, 38(3): 8 –37.
Copeland, T., & Antikarnov, V. 2001. Real options: A practi-
   tioner’s guide. New York: Texere.                              Maidique, M. 1980. Entrepreneurs, champions, and technolog-
                                                                     ical innovation. Sloan Management Review, 21(2): 59–76.
Cyert, R. M., & March, J. G. 1963. A behavioral theory of the
   firm. Englewoods Cliffs, NJ: Prentice-Hall.                    March, J. G. 1988. Decisions and organizations. Oxford:
                                                                     Blackwell.
Day, G. S. 1990. Market driven strategy: Processes for creating
   value. New York: Free Press.                                   March, J. G., & Simon, H. A. 1958. Organizations. New York:
                                                                     Wiley.
Dierickx, I., & Cool, K. 1989. Asset stock accumulation and the
    sustainability of competitive advanatage. Management          McGrath, R. G. 1997. A real options logic for initiating tech-
    Science, 35: 1504 –1511.                                         nology positioning investments. Academy of Manage-
                                                                     ment Review, 22: 974 –996.
Dixit, A. K., & Pindyck, R. S. 1994. Investment under uncer-
    tainty. Princeton, NJ: Princeton University Press.            McGrath, R. G. 1999. Falling forward: Real options reasoning
                                                                     and entrepreneurial failure. Academy of Management
Dosi, G. 1982. Technological paradigms and technological
                                                                     Review, 24: 13–30.
    trajectories: A suggested interpretation of the determi-
    nants and directions of technical change. Research Pol-       McGrath, R. G., & MacMillan, I. C. 2000. The entrepreneurial
    icy, 11: 147–162.                                                mindset. Cambridge, MA: Harvard Business School
                                                                     Press.
Ferrier, W. J., Smith, K. G., & Grimm, C. 1999. The role of
    competitive action in market share erosion and industry       Miller, K. D., & Reuer, J. J. 1998. Firm strategy and economic
    dethronement: A study of industry leaders and challeng-           exposure to foreign exchange rate movements. Journal
    ers. Academy of Management Journal, 42: 372–388.                  of International Business Studies, 29: 493–514.
Galunic, D. C., & Eisenhardt, K. M. 1996. The evolution of        Nelson, R., & Winter, S. 1982. An evolutionary theory of eco-
   intracorporate domains: Divisional charter losses in               nomic change. Cambridge, MA: Harvard University
   high-technology, multidivisional corporations. Organi-             Press.
   zation Science, 7: 255–282.                                    Penrose, E. 1959. The theory of the growth of the firm. New
Garud, R., & Nayyar, P. R. 1994. Transformative capacity:             York: Oxford University Press.
2004                                                   Adner and Levinthal                                                        85

Popper, K. R. 1959. The logic of scientific discovery. New York:   Sitkin, S. B 1992. Learning through failure: The strategy of small
   Harper Torchbook.                                                    losses. Research in Organizational Behavior, 14: 231–266.
Reuer, J. J., & Leiblein, M. J. 2000. Downside risk implications   Staw, B. M. 1981. The escalation of commitment to a chosen
   of multinationality and international joint ventures.               course of action. Academy of Management Review, 6:
   Academy of Management Journal, 43: 203–214.                         577–587.
Russo, E., & Schoemaker, P. J. 1989. Decision traps: Ten bar-      Trigeorgis, L. 1997. Real options: Managerial flexibility and
   riers to brilliant decision-making and how to overcome              strategy in resource allocation. Cambridge, MA: MIT
   them. New York: Doubleday.                                          Press.
Selten, R. 1975. Reexamination of the perfectness concept for      Van de Ven, A. H., Polley, D. E., Garud, R., & Venkataraman,
    equilibrium points in extension games. International              S. 1999. The innovation journey. New York: Oxford Uni-
    Journal of Game Theory, 4: 25–55.                                 versity Press.

                     Ron Adner is an associate professor of strategy and management at INSEAD. He
                     received his Ph.D. from The Wharton School, University of Pennsylvania. His research
                     examines the impact of the demand environment on technology evolution and com-
                     petition, as well as the organizational challenges that confront firms competing in
                     heterogeneous demand environments.

                     Daniel A. Levinthal is the Julian Aresty Professor of Management and Economics at
                     The Wharton School, University of Pennsylvania. He received his Ph.D. from the
                     Graduate School of Business, Stanford University. His research interests focus on
                     issues of organizational adaptation and industry evolution, particularly in the context
                     of technological change.
You can also read