ASPIRATIONS AND REAL OPTIONS: A BE HAVIORAL THEORY OF STRATEGIC DECISION MAKING By RICHARD JOHN GENTRY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006
Copyright 2006 by Richard John Gentry
This work is dedicated to my family.
iv ACKNOWLEDGMENTS Anytime someone embarks on a quest of this magnitude, it is the people around him that bear the disproportionate share of the difficulty. I have been working toward this PhD for 6 years. In that time, I have been fortunate enough to muster the support of some truly wonderful people. To begin wit h, this research would not have begun without the generous support of the Ka uffman Foundation and the Public Utility Research Center. Their support has helped me to fund this research, and th eir generosity resulted in a very detailed and complex paper to which I am proud to put my name. Secondly, this work owes a great deal to the patience of my committee, in particular Wei Shen. Dr. Shen took me on as his student upon his arri val at the university either out of charity or simple need, but hi s strong support and never ending tolerance for my antics has helped me develop a fascination in research and a str ong critical eye. I can only hope that I will serve as e ffectively as a mentor in the future to one who was as desperately in need as I was. Henry Tosi has been with me through the entire process, and he has never failed to offer his honest as sessment of my progre ss, a rare and difficult thing to find in academics. Srikanth Paruchur i has generously offered his expertise to my work when it was probably beyond reasonable to ask him for any more assistance. Heather Elms is also owed a great deal of th anks for soliciting my entry into the program, even after I told her I was not interested. Sanford Berg and Mark Jamison have been enthusiastic supporters of me and my career since before I began this program. Without their help and cheer, I would never have fi nished. In fact, much of the computer
v programming skills I have used in this project were learne d while working for Dr. Berg and Dr. Jamison in the early years of my program. However, I take great pride in knowing that I will continue all of th ese relationships well into the future. In addition to academic help, I have enjoye d the benefit of work ing with wonderful people at the university. From Mitzi Calver t who joyfully handles my whimsical and purposively nave approach to university bur eaucracy to the seemingly limitless joy that the people at the Public Utility Research Ce nter express for my work, I can truly say I have been very lucky in my friends. I shall miss being able to bumble into their office for some laughs when I am stressed. I know th at I make them laugh, but I hope that I can express to them effectively how wonderful it has been to have their support. I shall miss Cynthia Stehouwer, Edith Myrick, and Araceli Castaneda a great deal. I have enjoyed the collegial help of Nathan Podsakoff, w ho helped me deal with the drudgery that sometimes defines our work. Giorgia Dâ€™Allura and Irene de Pater have been wonderful friends and colleagues to me, and their frie ndship has helped me realize the scope and magnitude of the profession I am about to en ter. Suzanne Taylor, J. B. Loane, Rhys Williams, Steven Leonard, Val Watson, and Ch ris Melley have been wonderful and supportive friends who never fail ed to provide me a refuge when I had forgotten what incredible people act like. Matthew Mats en, Matthew Wilson, and Troy Quast have provided indispensable help in my da ta collection and programming efforts. Finally, however, I want to thank my pare nts and my brother. I do not know how I could have done this without their help. I ha ve not a few faults, one of which is letting go of my temper and angst upon those I care about th e most. In particular, I have been hard on my mother. Her dedicati on and strength in handling my personality through this
vi process has been inspiring. It is only through the example of her tenacity that I knew people can achieve things such as this. Th e process of graduate school has been hard, almost debilitating at times. The hours and rest rictions that I have placed on myself have cost me tremendously. At times, I have been unfair to myself, but my goal has always been clear. I was never going to allow this process to get the best of me, and I was always true to what I set out to do. I entered this profession in spirit sitting in a campground with my parents outside of Ann Arbor, Michigan in 2000. I was unhappy with my job at Ford, and I did not want to work in a career like that. I wanted a pr ofession like my fatherâ€™s. I did not envy the money: only the time. I envied the time b ecause I knew how important it was to me and my brother growing up. Without a professi on which allowed me some freedom, I might never have the possibility to be as good a fa ther to my children as he is to me: a happy man, a strong role model to his two sons, and powerful enough in will and character to ignore the irrelevant. While he is only huma n, it cannot be said that he does not inspire others try to be more. Finally, I want to thank my br other. Nothing is more satisfying to me than going to Nicholasâ€™s house. Being in a rarified world of academics can be very constraining and suffocating. It is my brother who knows me best and my brother who always knows what to say. He may not always know that he is doing it, but my br other has helped me keep my head through this process. Ou r exchanges are often short and seemingly meaningless, but I truly wish to be more lik e my brother. In the people who have truly mattered in this life, I find my brother and may it always be so.
vii TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES.............................................................................................................x LIST OF OBJECTS...........................................................................................................xi ABSTRACT......................................................................................................................xii 1 INTRODUCTION........................................................................................................1 Real Options.................................................................................................................5 Aspirations....................................................................................................................7 2 ASPIRATIONS..........................................................................................................13 Aspirations and Information.......................................................................................15 Aspiration Adaptation.................................................................................................22 Innovation...................................................................................................................24 Organizational Search.................................................................................................29 3 REAL OPTIONS........................................................................................................34 Real Options...............................................................................................................37 Shadow Options..........................................................................................................40 Performance equal to aspiration..........................................................................41 Performance below aspiration.............................................................................42 Performance above aspirations............................................................................44 Risk and Uncertainty..................................................................................................46 Thresholds...................................................................................................................49 Survival Bias...............................................................................................................53 4 HYPOTHESES AND MODEL..................................................................................58 Within Group Aspirations...........................................................................................59 Imitation and Option Behavior...................................................................................62 Trait-based Imitation...........................................................................................62
viii Scope of Purchase................................................................................................64 Option execution.................................................................................................64 Sample........................................................................................................................65 Dependent Variables...................................................................................................67 Independent Variables................................................................................................68 State Characteristics............................................................................................70 Option Characteristics.........................................................................................72 Option uncertainty........................................................................................72 Option value.................................................................................................73 Option similarity..........................................................................................73 Market Characteristics.........................................................................................75 Data Considerations....................................................................................................75 Model..........................................................................................................................79 Hypotheses 1 and 2..............................................................................................80 Hypothesis 3........................................................................................................80 Hypothesis 4........................................................................................................81 Hypothesis 5........................................................................................................82 Hypothesis 6........................................................................................................82 5 RESULTS AND DISCUSSION.................................................................................89 A LOGIT ESTIMATION OF MARKET ENTRY.......................................................118 B ALTERNATIVE ATTAINMENT DISCREPANCY DEFINITIONS.....................121 LIST OF REFERENCES.................................................................................................127 BIOGRAPHICAL SKETCH...........................................................................................135
ix LIST OF TABLES Table page 4-1 Option Uncertainty Measures...................................................................................85 4-2 Option Value Measures............................................................................................85 4-3 Option Similarity Measures.....................................................................................86 4-4 Market Characteristic Measures...............................................................................87 5-1 Option data summary statistics..............................................................................102 5-2 Market data summary statistics..............................................................................105 5-3 Option by year data summary statistics..................................................................107 5-4 Market by year data summary statistics.................................................................108 5-5 Maximum likelihood estimates of the likelihood of option purchase....................109 5-6 Poisson model of option acquisition......................................................................111 5-7 Maximum likelihood estimates of th e likelihood of option purchase based on dissimilarity............................................................................................................112 5-8 Poisson estimation of the tendency to enter markets.............................................114 5-9 Maximum likelihood estimates of the likelihood of market entry.........................115 5-10 Summary of results.................................................................................................117 A-1 Logit estimates of the likelihood of market entry..................................................119 B-1 Maximum likelihood estimates of the like lihood of market entr y with attainment discrepancy defined just in terms of market and option differences......................123 B-2 Maximum likelihood estimates of the like lihood of market entr y with attainment discrepancy defined just in terms of market differences........................................125
x LIST OF FIGURES Figure page 1-1 The proposed theoretical model...............................................................................12 2-1 Expanded model of attainment discrepancy.............................................................33 3-1 Model of corporate decision process........................................................................56 3-2 Option behavior relative to attainment discrepancy.................................................56 3-3 Prospect theory power curve....................................................................................57 4-1 Observed theoretical model......................................................................................84 4-2 Delay behavior following option purchase..............................................................84 5-1 Market competitor count by origin.........................................................................100 5-2 Entry and exit graphs by origin..............................................................................100 5-3 Relationship between attainment discrepancy and markets entered......................101
xi LIST OF OBJECTS Object page 1 The Visual Basic program code used to generate the dataset..................................79
xii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ASPIRATION AND REAL OPTIONS: A BE HAVIORAL THEORY OF STRATEGIC DECISION MAKING By Richard John Gentry May 2006 Chair: Wei Shen Major Department: Management Recent debates regarding theories of stra tegic decision making have attempted to limit the application of real options theory to only those decisions where uncertainty is exogenous and time horizons are fixed. Many pr ojects in a corporate setting do not fit this exclusive definition, so real options theory must either be restricted to a particular context or expanded to accommodate these proble ms. Rather than restrict the theory, this paper tries to expand real options theory to incorporate managerial behavior and thus attempts to resolve the problems with real options theory. To do so, this dissertation incorporates aspiration theory into real options theory to develop a behaviorally based perspective of option identification, devel opment, and execution. For an empirical context, this research used a subfield of the teleco mmunications industry from 1996 through 2004. The study separated competitors in this market into two subgroups to highlight behavioral differences between groups. To highlight the importance and dynamics of entrepreneurial market entry, this dissertation divided the industry into two
xiii origin-based groups. This study found interest ing behaviors based on the histories of the firms considered. Rather than support within group differen ces, this study found aspirations have a consistent influence on option purchase and market entry. While the effects of aspiration differences appear homogenous across groups, some results suggest that firms imitate within these groups. In addition, the firms pos ition relative to other firms in the industry influences its tendenc y to purchase options and enter markets.
1 CHAPTER 1 INTRODUCTION Since March and Simon (1958), organizati on researchers have sought concise, formal techniques to describe managerial decision making which explicitly incorporate the behavioral tendencies of the managers i nvolved. One of the most difficult challenges in developing a theory that describes decision making is incorporating managerial perceptions of uncertainty. Environmental uncertainty unsettles the mental models managers use to frame the environment and complicates theorizing. To date, two approaches have emerged which hold pr omise for a more complete theory, institutional/learning theory and real options theory. Institutional and learning theories are based on the obs ervation that environmental uncertainty leads firms to imitate other firms (DiMaggio & Powell, 1983; Greve, 1998a; Haveman, 1993; Henisz & Delios, 2001). When in an uncertain and changing environment, firms adjust for risk increases by following strategies which have already been attempted by competitor firms. Ultimately then, firms compensate for uncertainty by following similar strategies and gauging th eir success relative to their competitors. These theories have found that external factors moderate th e tendency for firms to borrow strategies from one another, one of which is performance relative to an internally defined aspiration level. Firms tend to change stra tegies as performance fluctuates around the aspiration (Greve, 2003a; Massini, Lewin, & Gr eve, 2005), a finding consistent with The Behavioral Theory of the Firm (Cyert & March, 1963). However, while learning and institutional theories have grown in theore tical importance and f ound empirical support,
2 these theories do not offer a prescriptive assessment of the decision making process nor do they provide a basis for extending the th eoretical domain to incorporate between-firm differences. For instance, while research shows that search behavior increases as performance falls below aspiration level, there are few studies to suggest that performance influences how the firm invests the results of search. Furthermore, while aspirations are defined both internally and ex ternally, there have been no suggestions as yet for what managers can do to influence th e process of their formation. Learning from performance feedback is an important theore tical domain, but it l acks formalization and the ability to describe incremental investments. Another recent development in the managerial decision making literature, real opt ions theory, might provide a way to bridge these gaps. Real options theory explicitly incorpor ates an investmentâ€™s uncertainty. The application of real options theory to corporate deci sion making provides scholars a concise way to devolve decisions into discrete packets. The theory provides a simple conceptual way to divide decisions into steps, each of which is an option or choice which can be ignored, deferred, or executed. The th eory has its roots in the finance literature and provides the ability to frame decisions as discrete choices in much the same way as a financial option provide the holder with the abil ity to take discrete action. Real options theory has proven versatile enough for research ers to frame the minutest decisions or the grandest corporate scheme as a management choice similar to a financial option (Kogut, 1991; Kogut & Kulatilaka, 2001). Real opti ons theory incorporates the risk and uncertainty associated with the decision, a prope rty that is lost in traditional net present value analysis, and it has encouraged crossdisciplinary research in the literatures on
3 economics (Dixit, 1992; Teisberg, 1993), fina nce (Trigeorgis, 1993) and management (McGrath, 1997). However, while the theorizing of real options has grown to incorporate many different corporate strategies such as preempting market competitors (Miller & Folta, 2002) or research and development (McGrat h, Ferrier, & Mendelow, 2004), the theory is still disconnected from the sociology of firm decision making and search behavior. There has been no explicit work that looks at the bi ases of management or other organizational problems in the use of real options both as a pr actical and as a theoretical tool. It is not clear if the process of option purchase changes as the performance of the firm changes. In addition, real options theory suggests that as uncertainty in the environment increases, the options of the firm should increase in value, but the context under which managers evaluate options is not in corporated in the theory. In short, this theory is underdeveloped, but it is underdeveloped in some very particular ways. Most approaches to the theory focus on the economic choices involved in executing or striking a r eal option while generally i gnoring the more social and behavioral components of firm-level decisi on making. As such, the theory cannot be used to model corporate decision making, in its current state it can only be used to describe very precisely defined decisions, su ch as acquisitions or research projects. While real options theory has incredible pot ential, it has not yet been developed enough to create a theoretical link between the ba sis of firm strategy, resources, and the application of those resources to decision making, real options. Until the theory is refined to incorporate a link from real options to a firm â€™s underlying resources, real options theory will continue to offer little in the way of practi cal help to managers and lay
4 exposed for theoretical criticism (Adner & Levinthal, 2004) because it simplifies reality in some troubling ways. This paper creates this li nk by incorporating aspirations theory, drawn from the behavioral theory of the firm, to explain unde r what circumstances firms extract strategic options from their resource stocks and how environmental uncertainty changes this choice. The theory presented here holds that as the firm fails to live up to its aspirations, managers look for more ways to employ the firm â€™s internal resources. A firmâ€™s relative performance drives its efforts to find and de velop strategic options , a behavior known as search. As search increases, the likelihood of a firm discovering a new strategic option increases. These options then create a pool of potential strategic actions which are the basis for considering strategic alterna tives in the decision making process. Aspirations then further influence the st riking or execution of strategic options by influencing managerial risk tolerance and performance expectations. These thresholds represent a) the managerâ€™s preference for one option over another and b) the managerâ€™s preference for taking any strategic action. This paper will expand real options theory by incorporating a behavioral theory that offers a link between firm resources and manage mentâ€™s need to grasp for the brass ring of success. The theory presented here will explicitly incorporate environmental uncertainty and managerial perception of uncertainty to generate a more robust explanation of not only how behavior changes around the aspirati on level but how uncertainty changes the behavior of firms with rega rds to their strategic option. This chapter will begin with a discussion of real options and the current state of research in one of the most cross-disciplin ary theories since Prospect Theory used
5 psychology to revolutionize th e economics and finance literature (Kannaman and Tversky, 1979). Following an examination of re al options, the causes of aspiration levels and their influence on real option creation will be examined in greater detail. Finally, this chapter will summarize, by inco rporating other research, a process governing how firms take options created through search and move them to execution. Real Options Bowman and Hurry (1993) first conceptu alized corporate strategy through an option-based perspective. In their model, an option confers preferential access to an opportunity for investment choi ce. In other words, an op tion provides to its holder the right and ready access to strate gic assets and choices. Options can be held to be exercised at a later date or disregarded. Options, like decisions, can be as mundane to management as the ability to change maintena nce schedules or as strategic as entering a completely new market. All decisions within a firm can be characterized as an option and the firm itself can be characterized as a collection of strategi c options of varying value. The first step in modeling strategy using real options theory is identifying how options are created. Until management seeks to find new options, strategic options exist as shadow options within the firm. A sh adow option is any possible combination of resources within the firm. Importantly, thes e combinations do not necessarily have a positive expected payoff. Not all options are profitable or even sensible. This horde of possible resource combinations creates a ne ar endless portfolio of possible firm behaviors. However, only firms with valuable resources will have valuable options; the value of the firm derives from the value of th is unique combination of options. However, determining the constituent parts of this valuation is a serious problem for researchers,
6 but this limitation can be addressed by exam ining the constituent parts of an option, real or otherwise. As with any theory borrowed from another field, financial economics in this case, the theory must be adapted to be appli cable to management theory. Although real options explicitly accounts for the uncertain ty and risk implied in any decision, several characteristics of financial option theory do not easily translate to real options. First, financial options have a fixed date at wh ich the option is no longer valuable, its expiration date. Real opti ons do not have a fixed expi ration date, managers must decipher information from the external market and internal corporate circumstances when the value of the option is about to expire. Because the expiration date of an option, and thus its value, depends on the perceptions of management, the use of real options as a management tool can lead to escalation of commitment (Adner & Levinthal, 2004). The second major problem with applying real options to managerial decision making was suggested by Adner & Levinthal (2004). Uncertainty is one of the key components of pricing an option. However, unlike environmental uncertainty, how much uncertainty an option contains is determined by the perceptions of management. Because the uncertainty of the option is determined be fore the option is purchased, theories that seek to explain option purchase need to suggest a way that ma nagerial perceptions influence decision making. The third major problem, connected with the first two, is estimating the price of the real option. Although managers use a logic that is consis tent with real options by approaching decisions in incremental and discrete steps (McGrath & Nerkar, 2004), managers tend to systematically undervalue the real options with in their firm. The
7 intuitive nature of option valuation leads manage rs to incorrectly frame their costs. This incorrect valuation leads managers to assume that any particular real option will cost less than it actually will, payout more than it actual ly will, and is less ris ky than it actually is (Miller & Shapira, 2004). Thus, managers tend to display a self-serving or optimistic bias in option valuation. In short, the problems with applyi ng real options theory to the practice of management rests on the behavioral assumptions of the managers involved. Under-pricing and escalation of commitment are both behavioral characteristics of managers, and both are largely ignored in the real options literature. A more complete theory of real options will incorporate thes e managerial biases, and this paper suggests one mechanism for doing so, aspirations theory. Aspirations Aspirations were first de veloped by March and Simon (1958) and later Cyert and March (1963). Within their behavioral th eory of the firm, they noted that firm management tends to express a preference fo r a particular performance level and this performance level seems to be persistently greater than zero. In short, management desires projects that are not just positive net present va lue, but projects that are significantly greater than zero ne t present value. The level they wish to obtain is a sociological comfort level of pr ofits referred to as the firmâ€™s aspiration level. Aspiration levels are the borderline be tween perceived success and failure and denote the starting point of doubt and conflict in decision making (Greve, 1998b). The difference between realized performance and the aspiration leve l is attainment discrepancy (Lant, 1992). Managers set their aspirations, or success re quirements, from a pr oject through aspiration formation.
8 Aspirations are determined by three primary factors. The first, suggested by Cyert, March, and other researchers (Glynn, Lant, & Mezias, 1991; Greve, 2002; Wilsted & Hand, 1974), is past performance levels. Mana gers would prefer to earn as much on any new project as they earned on past projects . Thus, matching or exceeding historical performance becomes a goal within the organization. The second determinate of managerial pr eferred performance, as suggested by Greve (1998b), are the performance levels of co mparative firms. If the profit structure of an industry is changing or consumer tastes are shifting radically, the firmâ€™s historical aspiration level may no longer be appropriate. In these situa tions, managers tend to focus their attention on the performance of competito rs. If competitors are earning returns or achieving market share consistent with the fo cal firm, the firm will not experience as much attainment discrepancy. Not only do mana gers seek to earn returns at historical levels, but they also desire returns consistent with competitor returns. Figure 1-1 suggests a third source of rele vant performance objectives. In the absence of publicly available performance, firm aspirations will be driven by comparing market positions to competitors. In s hort, firms will seek to reduce attainment discrepancy by mimicking the investments of competitors. In the absence of performance, firms use other visible success cr iteria to determine an aspiration (Greve, 2002). When managers cannot discern the outcom es of competitorsâ€™ strategies, they will try to mimic these strategies to maintain competitive parity. This will increase the likelihood managers will perform as well if not better than th eir nearest competitors once information does become available.
9 The managementâ€™s cognitive models determine which firms are imitated and which are ignored. These groups are developed around firms that are visible to mangers (Reger & Huff, 1993) or dominate in an industry (Haunschild & Miner, 1997). This paper will suggest that in this particular context, firms will tend to display a tendency towards homogeneity and this similarity is due to the cognitive models developed by managers. The uncertainty in the environment forces mana gers to look more at the traits of their competitors than the prestige of market competitors. Greve (1996; 1998b) suggested organization change and geogra phic proximity encourages firms to imitate one another under new market conditions. He argues that learning from other firms is easier when those firms are proximal competitors. However, Greveâ€™s (1998b) supported hypotheses do not really speak to similarity between fi rms within the industry, only their level of contact and the characteristics of the market. If managers are concerned with maintaining performance parity with similar firms, thei r behaviors should reflect a bias towards similarity seeking with firms of similar size , origin, and market presence. This paper suggests that managers will not be as concerne d with firms in the same markets as they will with firms in a similar resource position. These similar resource positions should derive from the firmâ€™s origin, whether it is entrepreneurial or a preexisting established firm. This suggestion is an extension to th e behavioral models which have heretofore focused on how firms adjust to their position wi thin an entire industry instead of within groups. This paper will suggest that organizations imitate th e behavior of similar firms and will suggest that similarity between firm s in terms of origin will encourage imitation in the absence of other performance information.
10 This argument echoes Haveman (1993) who observed mimetic behavior among firms. Her sample looked at established savi ngs and loan corporati ons in California. While isolating one kind of entry decision in one kind of firm, she found that firms in one industry followed the best performing firms into new market segments. This paper suggests that only when the environment of fers no relevant goal information do firms engage in explicit mimetic investment behavior such as the ones modeled in this study. Thus, this paperâ€™s theory borrows more from the managerial/cognitive perspective offered by Greve (Greve, 2000; Greve & Tayl or, 2000) than the in stitutional pressure perspective developed by Haveman (1992; 1993). This paper argues that firms are influenced by more than the institutional pressures that encourage homogeneity (Greve, 2003b). As historical levels become more difficult to achieve and less relevant to the industry context, managers seek other sources of performance information ultimately relying solely on the behavior of competitor firms. Managers have limited ability to conceptuali ze the entire competitive dynamics within an industry and will engage in more and more si mplifying behavior as the industry becomes more difficult to interpret. By combining re al options theory with managerial decision making, this paper provides a systematic m odel of how the firmâ€™s prior performance influences not only the creation of new options but also how those options are implemented. It suggests how uncertainty ch anges this process, and is a first step towards integrating research on managerial cognition which focuses on process learning, the resource based view which emphasizes th e importance of idiosyncratic historical paths and their role in performance heterogeneity, and aspiration theory which describes the firmâ€™s motivation and tendencies.
11 The dissertation, to follow, is the first model in the literature to suggest the importance of aspirations in real option s earch, option pooling and option execution. The studyâ€™s model is depicted in Figure 1.1. The first step in developing this model is incorporating the importance of information, a nd lack there of, into aspirations theory. This is the objective of Chapter 2. Aspira tions motivate the firmâ€™s tendency to look for new uses of the firmâ€™s resources. When th e market does not provide enough information to support aspiration formation, firms will define their aspirations in terms of other firms, a tendency that limits search and creates homogeneity among competitors. The third chapter will suggest why aspira tions are the missing link in re al options research and how the explicit incorporation of The Behavioral Theory of the Firm (Cyert & March, 1963) can inform the application of real options th eory and create a more unified theory of managerial decisions resulting from resource stocks. The theory is tested in the telecommunications industry by compari ng two groups of new market entrants, entrepreneurial firms and established entran ts, to derive aspira tion differences which motivate market behavior differences.
12 Figure 1-1. The Proposed Theoretical Model. Attainment discrepancy leads to changes in search and the tendency to execute strategic options.
13 CHAPTER 2 ASPIRATIONS Aspirations theory was developed by Cyer t and March in 1963. This theory has survived several shifts in research attenti on and underscores a very basic principle of organization behavior: failure and success al ways depend on the context in which they are considered. When maki ng evaluative judgments about success and failure, people have specific reference point s they use to decide whether an outcome was a success or failure. How people feel about success and where they set the explicit cutoff point for what determines failure depends on the inform ation they consider important when they are making the determination. A basic principle now, but this theory followed close behind March and Simonâ€™s (1958) work suggesting that individuals are not perfectly rational in the economics sense of the word, where decision makers with full information and zero transaction costs must fairly incorporate all info rmation and choose dispassiona tely between alternatives.1 Instead, Marchâ€™s work asserted that peopl e are not completely rational; they are boundedly rational. Individuals only search within a particular region, one that is familiar to them, and they stop search when they find a solution to their problems which is just good enough. People generally do not arrive at the optimal solution to any particular problem; instead they balance off the possibility of finding something better by continuing to search. In the end, people economize and arrive at solutions which are 1 The framing of economics literature as a system of specifiable variables contrasts with March and Simon who emphasize the impossibility of specifying the system even for a decision making with the most complete information in the game and massive quantities of time to make the decision.
14 locally optimum and most likely globally subo ptimum. Cyert and March took this local maximizing principle and applied it to manage ment decisions to investigate what the principle would mean for the market behavi or of firms. They concluded that the application of a satisficing or locally optimiz ing principle to managers within the firm may speak to why firm performance seems to be remarkably consistent from one quarter to the next (Greve, 1998b). The firmâ€™s managers determine the level at which the firm is performing acceptably, the point where the firm should be performing. Managers generate this acceptable value for performance and then mark everything against this value. This acceptable value, also called an aspiration, is the yardstick against which the firmâ€™s efforts are measured. Of course, the partic ular aspiration for any given period changes depending on the context. If the firm ha s been performing particularly well, the aspiration moves up and similarly it slides down if the firm has been underperforming. Managers grow expectant of further success just as they can grow complacent about failure. In recent years, the scholarly interest in aspiratio ns has increased (e.g. Baum & Lant, 2003; Chen, 2003; Greve, 2003a). Scholars have identified what particular pieces of information managers use to set aspirations, what causes as pirations to adjust and at what speed they adjust. However, although aspirations theory has re cently enjoyed considerable attention in the organizational learning literature, stud ies have not yet examined what managers do in the absence of relevant performance information for use in setting goals. Without performance information, how do managers set their aspiration level? Do they change it at all? Studies of industries following a di scontinuous change have suggested that firms
15 tend to cope with this change by copying the behavioral patterns of other firms, most notably large firms, but the motivations for this process have not been developed. The importance of where managers get their information is particularly relevant now that environmental uncertainty in most industries seems to be increasing (D'Aveni & Gunther, 1994). This chapter will explore the aspiration theo ry literature and construct a theory of aspirations under uncertainty in the absence of relevant performance information. The absence of performance information drives firms to look at other for performance information. This external comparison is based on several factors, one of which is similar history. It will begin with a discus sion of information and aspiration formation. Following this development, the chapter will discuss the process through which aspirations change and how innovation influences this process. Finally, the chapter will discuss how aspirations influe nce organizational search rou tines and its implication for the overall theory. Aspirations and Information While the idea of an aspiration is motivati onally very similar to a goal, it is not shaped in the same way as th e goals often discussed in motiv ation theory (Locke, Saari, Shaw, & Latham, 1981). While a goal is indivi dually created or a ssigned, aspirations represent the compromise be tween political sub-parties within the organization. Aspirations are socially constructed through tacit agreement among the dominant collation. It is a process th at is largely formalized by organizational processes and primarily, although not entirely, outside mana gementâ€™s control. Aspirations are influential within an organization and shap e behavior, they create motivation for actors within the organization and shap e decision processes. However, they are not the same as
16 individual goals. Other studies use goal and aspiration interchangeably (Lant & Montgomery, 1987; Lant, 1992) as will I. Howeve r, it is important to remember that aspirations are created at the organization level not at the individual level (Cyert & March, 1963). Aspirations are the organizationâ€™s general consensus for an acceptable performance level. Firms tend to focus on particular pieces of information from the general environment. For instance, when the chairman of American Express was replaced with James Robinson, earnings had begun to deteri orate. Morale at the company suffered greatly when the firm failed to report in come growth for the first time in 29 years (Grossman, 1987). This was an arbitrary goal, but one formalized by social processes within the organization and referred to internally as â€œThe Record.â€ McDonalds Restaurants, who experienced a similar de cline in 2002 and recognized its first ever quarterly loss, also suffered great shifts in morale even though the survival of the company was not in jeopardy (Doonar, 2004). The concept of how managers approach an aspiration is not very complex; organizations have a performance objective. Because managerial decision making is characterized by bounded rationality, the managers within an organization use a heuristic to find a goal quickly that is just sufficient to express the multiple objectives of the firm, a principle known as satisificing. They need a quick way to understand relative performance. The use of an aspiration level is just the application of heuristics to the evaluation of performance. The organizati onâ€™s goals are primarily derived from two different sources which together form th e socially constructed aspiration level.2 2 Greve specifies direct learning as a third option in his 2003 book. I do not discuss direct learning here because it is not particularly relevant in a context where performance information is not available. In
17 The first source of relevant performance information is the organizations past performance (Cyert & March, 1963) . Organizations that have been able to perform at a high level for a long time will normally wish to maintain that performance level and this performance level will become their objective. For orga nizations who are solely concerned with maintaining their historical as piration level, such as American Express, the performance of other firms is not an im portant factor. Provided that the firm can consistently perform at a level equal to its as piration level, it will not trouble itself with how well or how badly other or ganizations are performing. Us ing prior performance as a basis for aspiration formation is one of the ba ses for organizational in ertia. Organizations will not change their operati ng procedures if they are not motivated to change by poor performance relative to their aspiration level. It is when the orga nization is unable to maintain its prior performance that it be gins to look at the performance of other organizations (Greve, 2003b). The second source of performance informati on is the performance of other firms in the industry. Research s uggests that organizations fo rm comparison groups from competing firms to evaluate their perf ormance. Greve (1998b) has shown that organizations display a lower tendency to unde rtake risky behaviors as their performance approaches that of geographically-local firms. The performance of similarly sized firms also motivates aspirations as does the presen ce of a parent corporation mandating certain performance standards (Greve, 1998b). In s hort, instead of looking at all possible competitive organizations in the market, fi rms seem to pay more attention to the strategies of organizations much like them selves. Organizations frequently form a addition, this omission maintains clarity and avoids delving into an aspiration source that is not well understood.
18 reference group of important performance-releva nt firms from the organizations in their social group. Firms monitor reference organiza tions and managers are able to recite their characteristics and resu lts (Reger & Huff, 1993). Research has confirmed the tendency of firms to continue to change strategies until their performance was consistent with firm s within their social group (Bromiley, 1991; Greve, 1998b). When the organization is invest igating the performance of other firms, it will be less likely to undertake risky change in an effort to improve performance if its current performance is about average within its group. Organizations display similar tendencies regarding their histor ical performance, and the tend ency for firms to fixate on a social or historical aspiration or a mixture of the two, a lthough moderated by the environment, appears to be idiosyncratic. Firms balance between the two aspiration points depending on their absolute difference from each. If firms are out-performing the industry, they may have a tendency to look inte rnally while if they are performing at the same level historically but underperforming the industry, they may tend to focus on the industry (Greve, 2003). Generally, these aspirations are stated in te rms of profits, but it has been shown that organizations seem to set aspiration levels based on other perform ance-related criteria (Greve & Taylor, 2000). Setting aspiration le vels using the two mechanisms outlined above assumes that organizations are able to define performance in a given context. Not all markets offer verifiable performance info rmation that can be used as a basis for comparison. The early dot-com period and co mpetition among private start-ups are but two examples of when performance informati on is not available and aspirations are based on other sources of information. Particul arly when entering new markets or when
19 innovations transform the market, the firmâ€™s performance will need to be discussed in more qualitative terms. For example, the use of page views was a popular organizational performance measure during the Internet bubble of the late 1990s (Graham, Cannice, & Sayre, 2002). Even though this performance measure was not related to profits (Graham et al., 2002), organizations and analysts alike used this m easure as a viable, simple and universal way to compare organizations within the industry. In this context then, page-views became the performance measure, and they formed a foundation for comparison between firms. Thus, there is a form of the second as piration mechanism which functions when organizations are unable to define performan ce or define a strategic path towards good performance. In situations of high uncerta inty, organizations and managers may not be able to define good performance nor lay out a cl ear path to success. In such situations, the externally defined goal will be competitive parity. Competitive parity in this context refers to a firmâ€™s ability to respond to a competitive market action with a reaction of equal strength and effectiveness. As an as piration, competitive parity means that firms seek to be as strong as their rivals. In orde r to ensure competitive parity, the firm will try to keep pace with the average organization in the industry using other variables of strategic importance. This implies that the fi rm will use a series of intermediate variables to describe success. In the absence of performance information, the firm will set aspirations according to the competitive mean s rather than the competitive ends. When the outcomes of behavior are not clear, firm s will tend to focuses on the means to that outcome.
20 Greve (2000) suggests this process by showi ng that radio stations often use market share as a comparison goal when revenues or pr ofits are not publicly available. Haveman (1993) and Hausanchild (Haunschild, 1993; Ha unschild & Miner, 1997) have also shown that organizations in situations of high uncertainty are like ly to mimic the behaviors of others in order to gather legitimacy from ex ternal stakeholders rather than pursue growth opportunities that might lead to comparable performance (Carroll, 1993). Aspirations theory suggests that this kind of mimicry is driven by the desire to maintain competitive parity more than it is driven by the tendency of firms to gather and maintain institutional legitimacy (Greve, 2003b). Studies have found a tendency to copy behaviors from firms in the same industry (Guler, Guillen, & MacPherson, 2002) and also from firms who share director appointees (Westphal, Seid el, & Stewart, 2001). However, in this situation, where investors and the public are ju st as incapable as managers at defining a good strategy or good performance, following a similar strategy can be motivated by the need to maintain aspiration consistency just as it can be motivated by the need to gather legitimacy (Greve, 1998a). In uncertain conditions, the literature s uggests that organizations will tend to copy the behaviors of other firms (Winter, 2000). Where there is weak information, the ability of the firm to skillful select adjustments to the environment is restricted, encouraging more vicarious learning a nd heuristical thinking. Th e absence of performance information propels firms to compare themselv es to others using heuristic thinking, of which mimetic investment is just one outco me. Looking at the tendency for mimicry as the outcome of an aspiration pr ocess can help explain why some firms seem to engage in differentiation in addition to mimicry (Greve , 1998a; Korn & Baum, 1999). Firms might
21 mimic other firms to meet their aspiration requirement while expanding into different markets to secure that position. The most important factor then, is information. The availability of public information constrains the information that managers use to develop their mental models of potential competitiv e moves. In addition, bounded rationality and heuristics will govern this organization proce ss as it tends to govern every other process within the organization. Aspirations shoul d conform to those measures which are available and relevant to the manager. A comparison metric could be a good proxy for performance, such as market share for radi o firms or page-views for web companies. Often times, however, this information will be quite distal from financial firm performance. Thus, organizations in a highly uncertain environment will more likely set their aspirations according to the most important information source freely available to them. Information used in setting an aspiration mu st be public and the organization must be able to find information about all relevant fi rms while monitoring that information easily. Literature on competitive dynamics has stressed the importance of available information in monitoring competitive positions (Smith, Grimm, Gannon, & Chen, 1991), and I suggest it here as a key way for organizations to develop relevant goals. As performance information in the environment becomes incr easingly rare, firms will latch onto the most public information available to ascertain th eir performance. If the environment is changing or the industry is new, the firmâ€™s ow n historical performan ce may be irrelevant, and the firm will search for new informati on. The firm will use the most public and visible information available to measure its relative performance.
22 Proposition 2-1 In the absence of relevant perf ormance information, organizations will set their aspiration level based on th e most public and available information. This proposition, stated formally here, has been observed by Greve (2003b) who observed that organizations sometimes colla borate to form performance measures where none exist. He notes the J.D. Powers automo tive quality rankings as an example of firms focusing on performance criteria that are form ed by an outside pa rty and accepted by the industry as a whole. I am supplementing th at by suggesting that industry norms will create such a measure where no informati on otherwise exists. Figure 2-1 shows an expanded model of aspiration formation. Aspiration Adaptation In order to set an aspiration level, the or ganization must be ab le to change their aspirations. When the organization is not meeting its aspiration level, it is undergoing attainment discrepancy (Lant, 1992). Not being able to live up to the socially constructed aspiration level places considerable stress on individuals within the organization. Stress drives individuals to adapt a nd find a way to remove the stressors. The prior section discussed how organizations not performing up to their aspiration le vel construct a new one from outside information. However, this may not lower the attainment discrepancy within the organization and may in fact cause the or ganization more stress. Organizations cannot exist in a state of stress, individuals leave the firm and morale suffers as the panic associated with ba d performance begins to deteriorate the organization (Huff, Huff, & Thomas, 1992). In this situation, firms normally engage in a sense-making process where they lower their aspirations (Greve, 2002). Inst ead of increasing actual performance or finding an external rationalization for pe rformance through a social comparison, the
23 collective processes within the firm lowe r the aspiration level towards the current performance level in order to raise satisfaction with poor perf ormance. A similar process takes place when organizations are performing consistently higher than their aspiration level, but the process of dow nward adjustment is the most salient to the following discussion. Aspiration adjustments take time. If the firm is incapable of performing at its aspiration level, managers are still often unwilling to change performance targets. Particularly in business organi zations, individuals have a grea t deal to lose by revising aspirations levels. Stock options might not ap preciate, political capital might be lost and career advancement slowed by not being able to increase organizational performance. For instance, managers in the American Express example preferred consistent performance over time and were unwilling to accept sudden changes in performance or aspirations. In addition, organizational routin es bring the attention of the organization back to prior periods through budget cycles a nd other organizational rules and learning (Cyert & March, 1963; Greve, 2002). Although they are difficu lt to change, laboratory and empirical studies have found that aspi rations move down as performance falls consistently below aspirations and aspira tions move upward after prolonged success (Lant & Montgomery, 1987; Lant, 1992). S o, aspiration adjustme nt seems to be symmetric across situations of high or low performance. The speed of this adjustment also has an important influence on the organizationâ€™s performance. Greve (2002) suggests that or ganizations that are able to keep their aspirations from moving too suddenly up or dow n are more likely to perform higher than organizations who allow rapid swings in thei r aspirations. Managers have some control
24 over the organizationâ€™s attention, and they can use this attention to slow or speed up the adjustment process, a process generally gove rned by inertia (Gresov, Haveman, & Oliva, 1993). In sum, aspiration adaptation is a slow pr ocess. Organizations are structured to prevent aspirations from changing on a regular basis, but they do shift given enough time. Large differences in performance relative to aspiration inspire fast movements while small differences do not encourage such rapid shifts. One example of an event that will encourage rapid aspiration adjustment is disc ontinuous innovation in the industry. Such a change creates large performance changes and will more likely inspire organizations to change their aspirations. Innovation Often ignored in the discussion of aspi ration adjustment is the importance of innovations in the external e nvironment on the changes in as pirations within the firm (Greve, 2003a). Because innovations rapidly ch ange the industryâ€™s competitive structure, they are a source of attainment discrepancy. Rapid changes in pe rformance levels can create high levels of attainme nt discrepancy, but the result of this attainment discrepancy is asymmetric. If the organization is pe rforming above its aspi ration level and an innovation in the market creates a situation where the firm is performing well above its competitors, the firm will adjust its aspiration up and tend to seek that level of performance in the future (Greve, 2003b). However, if the innovation causes the firm to perform below its aspiration level, the firm will experience significant attainment discrepancy. The firm will need to take immediate competitive action to reduce its at tainment discrepancy. Empirically, in
25 response to revolutionary innovations la unched by other firms in an industry, organizations have been observed to take immediate action (Greve & Taylor, 2000) to protect their competitive position. However, the recognition of these revolutionary changes is dependent on the organizationâ€™s fo cus. Major innovations and organizational responses to innovations tend to occur in gr oups, where a new change is quickly adopted by other organizations within the indust ry (Romanelli & Tushman, 1994; Tushman & Anderson, 1986). The speed with which organizations re spond to innovations and how they respond has been of interest to researchers for a long time (Chen & Hambrick, 1995). The speed with which organizations undertake this response is dependent on how much the organization is learning from the market, it s tendency for environmental scanning. In uncertain periods, this tendency will be higher, but in stable periods firms will generally focus on their own performance rather than the market behaviors and performance of other firms (Greve, 2003b). The tendency of the firm to watch and learn from its environment will moderate its ability to implement strategic actions following a radical shift in the environment (Smith et al., 1991). The previous section spoke about the importance of the social reference group in aspiration formation. Similarly, there are diffe rences between firms in their tendency to form aspirations by looking at the external environment. Some firms are better at absorbing information from the environment than others. Researchers have documented the myopia of managers who tend to look only internally for market information (Levinthal & March, 1993), and this tendency will limit the ability of firms to adjust their aspirations to incorporate external market information. This external information
26 included the performance and technologies of smaller firms who will tend to be the firms that launch innovations which can destr oy the managerâ€™s business (Tushman & Anderson, 1986). Studies have shown that firm s who look at more competitors tend to be the firms who are less likely to be surpri sed by new innovation (Gar g, Walters, & Priem, 2003). Aspirations generate the motivation and te nsion for creative destruction within the firm. Aspirations, which underlie the behavi ors of the firm, must shift before the behaviors of the organizations will cha nge. Once, the recognition of the new environment has been made and new information incorporated, aspirations will adjust in a manner consistent with how much new inform ation must now be included. If the firm has been focused internally or has otherwise missed information from the outside market, it is more likely to make a rapid adjustment to its aspiration level once its cognitive perspectives shifts enough to allow any ki nd of recognition (Gavetti & Levinthal, 2000; Gresov et al., 1993). If the firm has been monitoring its environment, sudden innovations will not cause a rapid aspiration adjustment or high levels of attainment discrepancy because the firmâ€™s aspirations will already in corporate more external information. Thus, sudden shifts in the sources of the organizati onâ€™s aspiration level will lead to high levels of attainment discrepancy a nd large subsequent changes to aspirations. These shifts might come from new competitors entering a market forcing the firm to incorporate them in their aspiration definition or new innovati ons in the market which destroy a firmâ€™s profit In short, while aspirations tend to slowly migrate towards actual performance as time passes, sudden shifts in an industry can l ead to rapid changes in aspirations. This
27 effect is moderated by firm specific factor s, such as inertia, but innovations which threaten a firm can be dealt with in aspira tion theory. In order to accommodate highly inertial firms who are forced out of the ma rket, this model restricts the effects of innovation to only those firms who survive a discontinuous change. Proposition 2-2 For firms who survive discontinuous change, sudden shifts in organization information and discontinuous mark et innovations will lead to high levels of attainment discrepancy and rapid aspiration adjustment. Proposition two suggests, in effect, that as pirations are more than simply period-byperiod comparisons, but they instead encompa ss a long-term perspective. Although there are firm-specific differences in the mix mana gers chose between in ternal and external focus, firms are not entirely myopic. The average firm will monito r both the short-term efficiency of competitors as well as the i nnovations produced in the market. Firms who are not monitoring the environment are likely to be surprised and require a massive readjustment while firms who expended the resources to monitor the environment will require less of an adjustment after an uncerta inty inducing event, such as a technological innovation (Abrahamson, 1991). As market uncertainty increases, the firmâ€™s aspiration will tend to be defined in terms of increasing competitive parity rather than simply increasing performance. The firm will seek to change its strategy to counter that of other, more innovative firms within the market. Thus, as the level of strategi c uncertainty increases from a competencedestroying innovation or other change in market dynamics, the more likely the firm will define its aspiration in terms of other firms rather using pe rformance targets. Once, the organization recognizes the environmental sh ift and assigns it a high importance, the
28 uncertainty associated with the new innovati on will force managers to further employ a heuristic regarding its importanc e. Managers need to use some kind of mental model to cope with these changes, the most likely of which is simply copying the behavior of the other firms following the innovation (Gavetti & Levinthal, 2000). In summary, proposition 1 sugge sted that as uncertainty increases in an industry, firms will be more likely to use alternative performance measures that are employed as proxies for performance or seen as primar y drivers of performance. Proposition 2 augmented this argument by suggesting that larg e shifts in aspirations often result from massive shifts in the industry and the more my opic the firm has been, the larger this shift is likely to be. Large shifts in performance relative to aspiration levels confuses management and leads to stress within the firm , reducing this stress requires the firm to change its aspiration levels. The followi ng proposition suggests that as uncertainty increases and firms try to close the gap between other firms, they begin to define their aspirations in terms of the market positions of other firms. Managers grasp for a heuristic to explain their performance. Because simply reducing the attainment discrepancy is not a clear goal, managers use the behavior of other firms as a goal. The high uncertainty and strain within organizations resulting from high levels of attainment discrepancy leads firms to define aspirations in terms of di fferent outcomes. Thus, consistent with proposition one and two: Proposition 2-3 The higher a firmâ€™s attainment discrepancy, the more likely the firm is to redefine its aspiration level in terms of market pos ition and technological offerings.
29 The speed of the change is very import ant here. If an organization sets its aspirations using the performance of other firms, a social aspiration instead of a historical one, it will have been updating its aspirations vis--vis other firms the entire life of a new innovation. Thus, the more a firm engages in environmental scanning, the less rapid an adjustment will occur to its aspiration level. Aspirations levels would have been adjusting to the innovation and the uncertain ty it creates. If a new innovation is a surprise to a firm, its aspiration levels will now be wildly out of configuration with the environment. Firms will now undertake a pr ocess to reduce attainment discrepancy through new strategies. Large attainment discrepancy leve ls force firms to undertake action to close this gap. New strategies only occur after a period of organizational search. During the search process, the organization develops new ideas through external scanning and internal analysis. Organizational Search This section will discuss the process of organizational search, a well established organization behavior. Search is the pro cess of scanning both in ternal and external environments through which organizations r ecognize their opportuni ties. The section will, however, stop short of discussing th e implementation of innovation generated during search. Simply searching for an i nnovation does not determine the organizationâ€™s tendency to actually implement that innovati on (Greve, 2003a). However, search does have some interesting characteristic s of importance to this exploration. First, search is motivated by attainment discrepancy. Managersâ€™ search activity can be either directed towards a problem or undirected and wandering. Sometimes, organizations undertake search ju st as a dalliance to try out something new or experiment
30 with a laboratory innovation. Th is kind of search process is called slack search because it generally funded with the excess resources (slack) generated through prior good performance. Slack search is not intended to solve any problems, merely to expand the organization. When it comes to money, or ganizations cannot have enough. However, the process they go through for finding new sources of revenue and profits depends on the context of their search. Slack search is not a stressful time for the organization and the results are not critical to the firmâ€™s surv ival or continued ability to meet aspirations (Cyert & March, 1963). Slack search has been found in empirical research which supports a positive trend between organizati onal slack and research and development activity (Chen, 2003; Greve, 2003a). An importa nt point that will be referenced in the next chapter, slack search will generally not lead to innovation implementation. Although organizations might make new discove ries through the proce ss of search, they will not change the organizations strategy and abandon the strategy that generated the slack in the first place (Greve, 1996). The other kind of search occurs when organizations are not meeting aspiration levels, attainment discrepancy is positive. In this search process, the organizations are under more stress and it is often referred to as problemistic search because organizations are trying to solve a problem with their pe rformance level (Dimaggio & Powell, 1983). Managers need to solve a problem and they embark on a survey of possible solutions to bring their performance level back in line with aspirations. Di rected search, or problemistic search, is a response to poor perfor mance. The extent of managersâ€™ effort in this search process is driven by how far belo w aspirations the company is performing. As
31 performance declines, invest ments in problemistic search increase (Cyert & March, 1963). The second major component of search is that it is directed. Search is an organizational process used to rectify attainment discrepa ncy. As aspirations become more defined towards the market behaviors rather than the market performance of comparable firms, search will be more direct ed towards the processes of competitors. If attainment discrepancy is high, managers will en gage in an extensive search process. If attainment discrepancy is low, managers will not employ as extensive a search and will often find solutions that are incremental or ve ry similar to current operating requirements (Chang, 1996). In short, the motivation and direction of search is fuel ed by the firmâ€™s attainment discrepancy. If attainment discrepancy is positive (the firm is under-performing its aspirations), search will be more extensive a nd more energetic as di screpancy increases. If discrepancy is small, organizations will search locally for a sufficient solution to reduce discrepancy (Greve, 2003b). If discrepa ncy is large, organizations will search extensively for a way to make the discrepa ncy smaller. The na ture of the strategy depends on how the discrepancy is defined orga nizationally. If the di screpancy is defined by performance, the organization has a much br oader search set than if discrepancy is defined in terms of market position or t echnologies. How the organization defines discrepancy is a function of how much information the firm has access to. Negative discrepancy (out-performing aspira tions) can also spark search. Here the organization is exploring applications of its slack resources to new environments or technologies. As performance increases mana gers and employees have more free time to
32 spend on projects that may not be immediately applicable to the business or that offer small efficiency gains to the current business. This chapter suggested the sources and pr ocesses that result from aspirations. Aspirations are formed using available inform ation about historical performance and the performance of socially comparable firms. However, the envir onment often changes, forcing firms to focus on comparable firm s for information regarding appropriate performance. Frequently, this compar ison process does not reveal performance information, instead offering only informa tion regarding the market positioning of competitors. In the absence of performance information, firms shift their aspirations so that they reflect a desire to maintain a co mparable market position with their competitor firms. The size of this adjustment depends on the amount of environmental monitoring the firm has been engaged in prior to the marketâ€™s shift. Differences between the aspiration level and cu rrent performance spark search behavior. Now that this paper has focused on the theory of aspiration development under stable and chaotic situations, the paper will focus on how firms deal with this attainment discrepancy under stable and chaotic situations . The environment in which the firm is operating frames the firmâ€™s aspiration level and ultimately shapes the results of the search process. The result of that search, st rategic options, is the focus of chapter 3.
33 Figure 2-1. Expanded Model of Attainment Discrepancy. Attainment discrepancy is influenced by three factors: part perf ormance, comparison firm performance, and the expected performance of comparison firms.
34 CHAPTER 3 REAL OPTIONS The prior sections developed aspiration theo ry and suggested that aspirations form the basis for many competitive behaviors. The unease which permeates the organization when performance is below aspirations motiv ates the organizationâ€™s decision makers to invest and get out of the poor performing situat ion. However, in order to do this, it is important to characterize the decision making process. Decisions to undertake a competitive action have two stages. The first stage is a decision to investigate or learn about a potential action while th e second is the actual decisi on to enter and execute the behavior. In short, in or der to apply aspirations to decision making, a comprehensive theory must account for the sequential nature of decision making while also theorizing about aspirationsâ€™ influence on the decision outcomes. As a first step in this process, this paper will propose a static model of r eal option decision making under attainment discrepancy. One recent theory which incorporates the sequential nature of the decision making process is real options theory. Real options theory characterizes corporate decisions as financial instruments similar in nature to financial options, which can be divided, analyzed and implemented incrementally (K ogut, 1991). Most major corporate decisions have an impact on a companyâ€™s cash flow a nd most incorporate the firmâ€™s need to prepare for future strategic contingencies, a ch aracteristic well suited to explanation by a theory which divides decisions into discrete un its. The theory of real options has proven very effective in describing se veral kinds of corporate deci sions including joint venture
35 buyout, entrepreneurial failure, and technol ogy investment (e.g. Kogut, 1991; McGrath, 1999; Miller & Arikan, 2004). While options logic has been expanded to explain the day-to-day business decisions managers undertake (Bowman & Moskowitz, 2001), its conceptual domain does not yet incorporate th e social entity of an organization (Kogut & Kulatilaka, 2001). What is missing from the analysis of real options is a descript ion of the behavior which underlies their use. Work by Miller and Chen (2004) has discussed individual differences in the perspectives managers take on options and has suggested that managers tend to value options too optimistically at time s. At other times, managers show poor judgment in ascribing the risk to options that they objectively contain. However, although there seems to be work on individua l difference models in option valuation, none of this work has been applied to the larg er firm context or what it means for firm behavior. The first portion of this chapte r will describe the pr ocess through which options are recognized and valued. The model suggested by this section is described in Figure 3-1. Management influences this process through recognition and threshold setting. The control levers held by manage ment are surrounded by a dashed line in the figure. In later sections, this model will form the basis for considering the importance of aspirations in option execution. Briefly, options arise initially from the resources within a firm (Bowman & Hurry, 1993). Management can combine resources in unique ways, and this combination is the foundation of competitive advantage (Wernerfe lt, 1984). It is through the ability of management and the motivation of the firm that options within the resource pool are recognized and codified. Options create th e basis for a firmâ€™s strategy going forward;
36 firms do not have strategic options unless they have a recognized pool of options to draw from (Bowman & Hurry, 1993). However, ma ny of these options are not valuable. Some of them are too risky; others explo it opportunities that ar e outside the business domain of the firm. For each firm, options will be assessed against a firm-specific threshold which determines each firmâ€™s propensity to execute the option. Managers have two important roles in this model. The first is the recognition of shadow options. Shadow options or unrecogni zed options have not received a great deal of attention since they were first described by Bowman and Hurry (1993), but they will form the basis for conceptualization here. Managers must recognize options from the resource pool and describe their risk and retu rn characteristics. The second decision step, the threshold, has only recently entered the options discussion in management, but it forms an important step in the organizationa l process (Folta & O'Brien, 2005). In this step, managers must decide if the optionâ€™s value is greater than the firmâ€™s minimal threshold. Thresholds are a firm-specific value where firms compare their investment needs against what an individual option offers . If an optionâ€™s value is greater than the threshold, it will be implemented or struck. The model will stop at the decision to st rike a particular option and leave the performance implications of this process to other papers. Risk, the likelihood of losing money, figures heavily in this section. A closer examinati on of the differences between idealistic risk (as operationalized by the capita l asset pricing model) and perceived risk (as discussed in the managerial behavior l iterature) will help to shed light on the problems confronting real options resear ch and expose where good progress towards clarity can be made.
37 Real Options Research into the valuation and behavior of financial options has been ongoing for over a century. Surveys point out that the early work in this field derives from models of Brownian motion in fluid diffusion devel oped by Albert Einstein and Kiyoshi It (Chance & Peterson, 1997). Fundamentally an op tion is the right to ei ther buy an asset (a call option) or sell one (a put option) at a fixed price at some future date. Options thus have a one-sided risk profile. Once the option is purchased, its value can increase infinitely (at least conceptually) but it will never decline below zero. Theoretically, the price of the option will incorporate the risk of changes in the underl ying assetâ€™s value. Thus, the risk of the underlying asset and the va riance of its market value is a key part of the option valuation model. Using fungible asse ts as a basis, financial economists Fisher Black and Myron Scholes were able to deve lop clear models of how the value of an option is based on five characte ristics: The price at which th e asset can be purchased in the future, the interest rate at the time, the time until the option expires, the current market value of the option, and the variance in the underlying assetâ€™s market value. The genius of the Black-Sc holes (1973) option prici ng model is how it devolves characteristics of an asset into a uni-dimensi onal value, price. Provided that the option can be characterized by these five traits, rese arch can suggest the assetsâ€™ optimal price. Not only is this an incredibly powerful tool for use in the derivation of financial pricing models, but it has clear applic ation for managerial decisions . An option is nothing more than a contract with certain characteristics . Many management decisions have similar characteristics, uncertainty, upfront costs, time pressure, and potential payoffs. It was not long before scholars discovered that the analysis of real options was an insightful way to analyze corporate decision ma king (Bowman & Hurry, 1993). A real option is simply an
38 investment that gives the decision maker th e right but not the ob ligation to make a decision, the option can be postponed until the e xpiration date of the option or it can be executed immediately. The upfront cost can be characterized as either an acquisition (Vassolo, Anand, & Folta, 2004), investment in developing a new project (McGrath, 1997), or a series of expected cash-flows from which the project should not vary (Bowman & Moskowitz, 2001). Scholars have found several examples of r eal options-like behavi or in managerial decision-making. McGrath and Nerkar (2004) have used the development of research patents as an example. Pate nts require an upfront cost to develop, but once developed they have no downside risk. The informa tion contained in a patent, once approved, does not have a downside risk while it does ha ve a potential payoff and an expiration date.3 Similarly, Kogut (1991) looked at the tendency for firms to acquire control of a joint venture when the environment changed. Cons istent with a real option perspective, changes in environmental uncertainty altered the tendency for firms to buy-out their joint venture partners. Joint ventures, then, can be conceptualized as both firms buying an option to acquire a market position while spli tting the risk with another party. While these two operationalizations of real options have been the most popular, several scholars have used real options logic to examine the tendency for firms to enter new markets. The 3 Ignored here is the possibility of litigation from a fraudulently developed patent, a simplifying assumption. Incorporating into the patentâ€™s value a separate kind of option addresses this problem. The structure of litigation can by incorporated by viewing the option development as the selling of an open put option where other firms have the right to put a legal action to the firm at any given time. In short, the firm opens up an unrestricted downside risk if it lacks prop er controls on fraud within its organization. Legal maneuvering by the firm and a history of legal action will place a lower bound on this option and might be associated with higher levels of dishonest behavior. Most risk structures, once fra med appropriately, can be conceptualized as options.
39 market entry decision, while a unique manage ment decision context, is not outside the realm of real options tho ught (Miller & Folta, 2002). Market entry is a complex behavior whic h has many different intensity levels and motivations. Strictly speaking, if entry is the option, real opti ons logic does not incorporate the ability of the firm to mainta in several different sized investments in the market. Thus, the market entry will be both small and constant (consistent with purchasing the right but not th e obligation to undertake subs tantial investment in the future) or it will be large and aggressive (cons istent with striking the option to enter). However, closer examination of the market entry decision offers insight into the complexity that a real options le ns can explain in firm behavior. Folta and Oâ€™Brien (2004) broke down the market entry decision into what are essentially two options that compete for manage rial attention and res ources. The first is the option to grow its investment. Although a firm might have th e ability to enter a market, it is not required to do so. By ma king a small investment in a market, it is purchasing the ability but not the obligation to potentially grow in that market. Purchasing this option has costs of investme nt and an uncertain cash flow, but any loss from the operation will be limited to the ini tial investment. The second option suggested in Folta and Oâ€™Brienâ€™s work and developed formally by McDonald and Siegel (1986) is the option to defer an investment. At any gi ven moment, a firm can put off investment in a new market and pay for that option with the cost of monitori ng the market and the opportunity costs of not making an investment. In this cas e, the option will most likely remain unobserved to outsiders who might exam ine a large-scale secondary database, but there are actions within the firm that constitute the purchase of an option.
40 In short, as the complexity of decision s modeled through real options increases, the importance of real options in the developmen t of management theory will increase. The underlying logic of real opti ons has a demonstrated prac ticality in three different managerial decisions. Charact erizing decisions as options rather than simple one-off choices, offers researchers the ability to inco rporate tradeoffs, time, and risk into static observations of management choices. The ne xt section will discuss where options come from and how management chooses one from another. It will be followed by a discussion of the process manage rs follow to strike the option. Shadow Options Every firm has a collection of resources. When combined resources form the basis for competitive advantage (Barney, 1991). E ach firm has a unique pool of resources which ensures that its performance will differ from other firms in the market. Resources form the foundation of performance hetero geneity (Hoopes, Madsen, & Walker, 2003). The resources within a firm develop over time , and are highly history (path) dependent (Dierickx & Cool, 1989). Each firmâ€™s unique experience determines its resources. The unique experiences form the basis for a resour ce pool within the firm. These can be such things as special processes within th e firm (Teece, Pisano, & Shuen, 1997), unique capital assets (Barney, 1986), human assets (Hitt, Bierman, Shimizu, & Kochhar, 2001), special market locations (Lippman & Rume lt, 2003), or unique synergies (Kalnins & Chung, 2004). Resources then form the basis for possibl e action by the firm. Not all of these potential actions are discer nable by management; management may not always know the optimal combination of resources nor may it know all the combina tions. The options literature characterizes these unrecognized opt ions as shadow options. A shadow option
41 becomes a real option when it is identified by a person within the organization (Bowman & Hurry, 1993). Recognition resu lts from the efforts of management to develop options regarding a particular investment and ch aracterize the payoffs and risks in that investment. In other words, the recognition proc ess, like the search process, is a directed one where management directs resources toward s the codification of a particular options bundle. As with the search process discu ssed in the prior chapter, recognition is a motivated exercise; management must be searching for options to find them. The recognition step also involves the firm pur chasing the options. Many times, the costs associated with purchasing the option are nothing more than the salary of the analyst who developed the plan, such as an in-house IT project. Other times, the firm must outlay cash for the option, such as an operating license or permit. It is the framing of the option during the recognition process that shapes the optionâ€™s value and the firmâ€™s tendency to purchase the option. Shadow options, once recognized, are characterized by uncertainty, upfront costs, tim e pressures, and potential payoffs. The conditions under which the real option is being examined will determine what kinds of options the firm is looking for and what kinds it acquires. These will now be examined in detail. Performance equal to Aspiration As a base case, consider the firmâ€™s pur chase of strategic real options when performance is equal to aspiration levels. There are two reasons firms will not generate new strategic options in this case. First, when the organization is performing at its aspiration level, it will not be as interest ed in expanding its market position (Greve, 2003b). Search will not be a prominent compon ent of the firmâ€™s strategy, and its overall orientation will not encourag e or reward search within the organization. In the
42 knowledge literature, this is th e exploitation phase of organi zational behavior as opposed to exploration (March, 1991; Park, Chen, & Gallagher, 2002). It will not undertake activity to find new projects and it will not se arch out profitable opportunities in other markets. In short, it will become static in its market approach. The satisfaction with performance leads to a lower tendency to purchase options. While the firm itself has not lost the ability to generate ideas, in this state the firm does not have the motivation to undertake search. Satisficing managers have met their goal. Proposition 3-1 When organizations are performing equal to their aspiration level, they will show a lower tendency to acquire strategic real options. Secondly, because firms are performing at their aspiration level (have zero attainment discrepancy), they will not be generating a lot of organizational slack. Although profit levels depend on the aspiration level, the firm is likely to be profitable. With zero attainment discrepancy, the gains from performing at as pirations are divided amongst the various constituencies within the organization. There would be little funding left to generate new options and search. Because managers in this situation will not be as motivated to explore and tend to behave risk neutrally, we wi ll not see managers in these cases purchasing exploratory strategic options. Instead, the contentment that results from achieving their aspiration level will lead to acquiring similar options to the ones already in the firmâ€™s portfolio. Proposition 3-2 When organizations are performing equal to their aspiration level, acquired options will be very similar to options already in the firmâ€™s portfolio. Performance below Aspiration Cyert and March (1963) first framed the syst ematic analysis of firm behavior under poor performance. Organizations are social enti ties that enjoy perfor ming at a particular
43 level. When performance does not equal that level, the organization will take action to bring performance back up to their aspiration level rather than adjust their aspiration level. In the short term, managers s eek to increase performance by creating new opportunities through the use of resources. As discussed in Chapter 2, poor performa nce creates attainment discrepancy and inspires the generation of strategic options. While firms engage in aspiration adjustment, this effect is longer term th an the immediate effect of mo tivating search behavior. The search for new options will be motivated to try and close the gap between aspiration and performance. As the gap increases, the manage rs will need to engage in a more extensive search of the surrounding environment in order to close the gap. The extensive search of the environment will lead to strategic options that are highly unique to the options already in the firmâ€™s possession. The uncertainty in the payoff from the op tion will be fundamentally influenced by the ability of managers within the organization to predict the cash flows from the project. The relatedness of the option to current assets will, in part, determine the uncertainty in the project. Especially if the underlying asset needs to appreciate in value extensively for the option to be valuable, uncertainty in opti ons increases the price. Paradoxically, more uncertain projects become more valuable to the firm as search becomes broader. Proposition 3-3 When attainment discrepancy is high and positive, managers will tend to acquire options that are less rela ted to options currently held by the firm. Certainly, this relationship is moderated by managerial characteri stics as different authors have suggested that poor performance and managerial behavior can lead to a downward spiral of overreaction to the ma rket (Miller & Chen, 2004; Simon, Houghton,
44 & Savelli, 2003). Firms in a downward spir al are purchasing every option the search process generates and ultimately spreading themse lves too thin to generate a capability in any. Within the conception of Shapira (March & Shapira, 1992; Shapira, 1995), it is possible for different managers to focus di fferentially on the sometimes conflicting goals of decreasing attainment discrepancy while si multaneously securing the firmâ€™s survival. These conflicting goals will be analyzed in a later section. Performance above Aspirations The outcome of search activities and st rategic option acquisition will be similar when the organization is in high or low perf ormance. Although the outcomes will appear the same, the motivations are different. Cons istent with Figure 3-3, performance above the aspiration level leads managers to try to secure and maintain those gains by becoming increasingly risk-adverse (Kahneman & Tver sky, 1979). However, when organizational performance is far beyond aspiration level, the firm will try out new things. Good performance leads to slack search (Cyert & March, 1963). The firm will be generating a lot of slack from its superior performance, and the firm will tend to use (absorb) this slack rather than immediately return it to shareholders. Research has shown, in the patent litera ture, that firms w ho are performing far above their aspiration level engage in a grea t deal of search and generate new patents (Chen, 2003; Greve, 2003a). This finding s uggests that firms performing above their aspiration will search for and purchase new st rategic options. Aspirations are a social construct that represent the social consensus of where performance should be. They function as a goal for the organization, but once the aspiration is achieved the different political constituencies within the firm try to divide the excess resources amongst themselves. This division is a functi on of the dominant collation which develops
45 aspirations. This division of spoils will lead to different groups within the organization generating options to use the resources to thei r own best interest. Slack search is the deployment of excess resources to different groups within the firm who then use those resources to expand their own business. This growth will represent consistent growth for the different groups within the firm, but it is more likely that the options when considered as a whole will be unrelated to one another and do not represent a strategic vision or patterned search for the firm. Greve suggests that slack search tends to be located in areas more distal from the firmâ€™s current operations (Greve, 2003b). This model suggests that it is a result of the nature of aspirations encouraging political groups within the firm to grow their base. While some firms ha ve found success by institutionalizing search and idea generation by engineering staffs, it is also the case that these firms could just as easily spend their excess resources on other proj ects outside of engineering staffs. As a result, the division of slack resources is ultimately a political process, although not necessarily a destructive one. Proposition 3-4 Firms performing in excess of th eir aspiration level will acquire more strategic options that are less re lated to the options already possessed. The above propositions constructed search and options generation as a process governed by the firmâ€™s attainment discrepanc y. As performance declines, managers will search in broader areas to find new options that will enable the firm to close the gap with its goals. As performance increases to appr oach aspiration levels , the search process declines and managers do not search as br oadly. New options, when they are generated at all, are very similar to those already owned by the firm. As performance increases beyond the aspiration level, various constituenci es in the firm employ the slack generated
46 by the excess performance to expand their own options portfolio and potentially their own business. This leads to a growth in option accumulation and a decline in the relatedness of the individual options. These propositions are presented graphically in Figure 3-2. The next section looks at the risk tendenci es of managers and their influence on the final leg of the model, managementâ€™s tende ncy to implement stra tegic options. Risk preferences imply considerably different things for the execution of options than for the purchase of options. Risk and Uncertainty Risk and the behaviors of individuals confr onting risk have been a central issue in organizational research for many years. Schol ars as early as Mill in 1848 and Knight in 1927 (O'Brien, Folta, & Johnson, 2003) have studie d the risky behavior in organizations (O'Brien et al., 2003). Since then, scholars ha ve been interested in how managers and organizations respond to choices with differe nt variance. While academic research has generally used the term risk in referen ce to uncertainty about outcomes, managers generally use the word risk only to discuss potential losses (Sitkin & Pablo, 1992). Firms are loss-averse and generally prefer projects th at have a limited potenti al to lose the firm money (Miller & Leiblein, 1996). When gain and loss are within range of one another (the firm is not taking a bet in order to make a potentially massive gain) the loss-potential takes presence in decision making over the gain-p otential. This is a consistent definition of risk to that used in th e Prospect Theory model which explicitly incorporates the tendency for people to respond negatively to lo ss despite the potential for gains. Figure 3-3 displays the power curve relationship be tween attainment discrepancy and the value of each unit of wealth. It is also consistent with experi mental evidence that suggests
47 utility functions and behaviors change when gamble choices are affine-transformed (Payne, Laughhunn, & Crum, 1980). Risk is a delicate word in management re search, and authors (R uefli, 1990) have criticized the tendency for management res earchers to employ a mean-variance measure of risk as it unfairly assumes a constant le vel in firm returns. Because variance and means are arithmetically related in the risk-retu rn realm, management researchers tend to treat risk inappropriately. In this theory, I hold risk and risk preferences as firm-level concepts. They describe the managersâ€™ te ndency to ignore loss and seek gain. Losses and gains have separate prope rties to managers depending on where the managers are along the graph in Figure 3-3. As organizations reach high levels of gain or loss, the value they place on each incremental dollar changes. Managers beco me gain-seeking as their performance drops below their aspiration while managers become loss-adverse as their performance rises above their aspiration. Thus, managers may feel the same about losing $9 billion as losing $8 billion. Subsequently, managers ar e not as concerned about losing one more dollar when they are at the bottom, but they will feel a great deal of satisfaction from gaining one dollar. In shor t, the loss functions are asymmetric. Similarly, if the managers are above their aspirations levels, thei r reaction to gain is only slightly smaller (or flatter) than their reaction to loss. Wh en managers are above their aspiration level, they will not be as concerned about gains a nd losses. Individuals only display symmetric loss functions at or around the aspiration leve l, where they feel equally passionate about gaining or losing one dollar. In the language of economics, they are indifferent between
48 the two. This asymmetric preference for gains or losses depends on the past, and I characterize that as risk. To separate terms and avoid falling into re search pitfalls, I will separate risk, a firm-level concept that depends on the manage rs, from uncertainty, a characteristic of the options which are analyzed by management. Risk preference will lead to different valuations regarding uncertainty. When mana gers are risk-seeking, the uncertainty in an option will become its most attractive characte ristic. As managers move deeper into the domain of losses, they will tend to acquire op tions with a bigger potential pay-off as they seek to improve the firmâ€™s performance. Ma nagers begin to think that they can control the loss potential of project s, a feeling exacerbated by poor performance (March & Shapira, 1987). The uncertainty in the option becomes e ndogenous with the managerâ€™s perceptions of control. March and Shapir a (1987) found that managers t ypically separate a projectâ€™s risk from a project return. While most finance theories, in addition to options theory, tend to treat risk as related to return, mana gers do not follow that logic. To them, the uncertainty in an option is something th at can be controlled through effective management while the return is something that can be maximized. March and Shapira (1987) cite an example of a manager who accep ted the potential returns from a project as given but returned the estimates of the projectâ€™ s uncertainty to his st aff for reevaluation. Because uncertainty and pr eference are chosen endogenous ly, we can use aspiration theory to further examine real options. The uncertainty in an option is first characterized during the search process. Earlier, search was characterized as the or ganization surveying the competitive landscape
49 to generate viable alternativ e actions. Search is at its highest intensity when the organization is experiencing attainment discre pancy. When organizations are either over or under performing their aspirations level, they will look for options to grow their market. The end result of this search pro cess will be the acquisiti on of strategic real options. Only the absolute magnitude of the attainment discrepancy is important for the search process. Risk perceptions change within the organizati on as the firm moves further and further from its aspiration level. The tendency for i ndividuals to exhibit nonlinear risk preference was first char acterized by Kahneman and Tversky (1979). The asymmetric loss functions hypothe sized in the Khanneman & Tverksyâ€™s Prospect Theory suggests that risk toleran ces will shift depending on the distance from the aspiration level. As attainment discrepa ncy increases, managers should become gainseeking as their performance suffers and lo ss-avoiding as their performance increases relative to their aspirations. Particularly on high return projects, managersâ€™ perceptions of their own abilities to mitigate risk changes, and they begin to act as if they have greater control of a projectâ€™s risk as they become mo re gain-seeking. This shift will influence the thresholds managers employ when deciding whether to execute a strategic option. Thresholds Once managers have acquired strategic options during search, these options will need to be evaluated. Not all options will be viable, and many will be unprofitable. Search is a stochastic process where random moments of inspirati on lead managers to purchase options and create opportunities (Stuart & Podolny, 1996). It is up to management to decide which options to strike and which options to defer. The decision between either executing an option of deferring it is determined by the firmâ€™s investment threshold. At different points in time, firm s will demand a higher return than other firms
50 for the same project, and this will infl uence the managementâ€™s tendency to strike particular options. A real op tion is a right, but not an obl igation for strategic action. Management must evaluate its options before executing one. This section suggests that firms have a threshold which determines the likelihood of a firm striking an option. These threshol ds are determined by the firmâ€™s attainment discrepancy, or performance rela tive to its aspirations. The concept of a threshold places a barrier between search and execution. The th reshold is where managers decide whether an option meets their risk profile and if the op tion moves the firm closer to its goals. The premise is suggested in the economics literat ure which holds that the market entry is often determined by a marketâ€™s competitiv e conditions at a given moment in time (Bresnahan & Reiss, 1991). This finding sugge sts that potential entrants to the market are waiting until the time is right to enter. In the options language, these firms are characterized as deferring investment until the option value has changed sufficiently for them to enter. Similarly, in the mana gement literature, research has found that entrepreneurs have different levels at whic h they will quit an industry. Entrepreneurs with high levels of human capital will leave an industry, in the face of poor performance, much quicker than will entrepreneurs who do not have a lot of human capital (Gimeno, Folta, Cooper, & Woo, 1997). This implies that the different option por tfolios held by entrepreneurs shape their observed behaviors. As their performance expectations from continuing in their entrepreneurial venture decreased, those entrepreneurs who could find employment outside (those with more op tions) exited the market and sh ut down their firms. Those entrepreneurs, who did not posses an option to leave, stayed while performance continued
51 to decrease. However, even those entreprene urs who had the option to do something else did not act until the industryâ€™s performance bega n to suffer. Working papers suggest that these differences are evident at the firm leve l as well. These papers suggest that firms differ in their tendency to enter a market and these thresholds vary from one period to the next within firms. Thresholds are firm-sp ecific traits which mark the indifference point between executing an option and deferring the option. These thresholds vary from period to period and are dependent on th e context, one important por tion of which is attainment discrepancy. High absolute levels of attainment discre pancy leads to high levels of search and option generation, but this does not mean that fi rms always strike the options they acquire during search. Instead, Figure 3-1 suggests an entirely different mechanism, also driven by aspirations, influences the tendency to execu te a strategic option. The level at which firms execute options is a threshold, a le vel which is influenced by attainment discrepancy. The model presented in Figur e 3-1 contains a mechanism for explaining why the tendency for firms to execute strate gic options shifts with aspiration levels. When firms are performing outside their aspi ration level, search will have generated many different options. Now, while the fi rm may use these options and the options themselves have value, research suggests th at firms will not actually execute new options when they are performing above aspiration le vels (Greve, 2003a). As the graph of Prospect Theory in Figure 3-3 suggests, when managers are performing consistently above aspiration levels, they will value a dollar lost more than a dollar gained. Management becomes loss-adverse when perf ormance is high. They should be unwilling to execute more options and risk losing money (Bromiley, 1991; Miller & Bromiley,
52 1990). They will value the initi al investment too highly and pay too much attention to the potential loses depicted in the projectâ€™s uncertainty. Similarly, firms performing below their as piration level will be attracted by the possibility of high risk/return projects. Because managers tend to separate the uncertainty of a project from its return, managers will feel more in control of the uncertainty inherent to a project as perf ormance decreases (March & Shapira, 1987). Managers will then tend to underweight the un certainty in a project and execute options which they might not otherwise. This is the final component of the downward spiral decision process outlined in learning rese arch (Levinthal & March, 1993; McDonald & Westphal, 2003). Here, firms have acquire d many different options through a broad search process but at the same time they misvalue those options and execute more of them in an effort to increase performance. Proposition 3-5 Option execution and attainment discrepancy will be negatively related such that the likelihood of executing an option decreases as performance relative to aspiration increases. Proposition 3-6 Option execution and attainment discrepancy will be negatively related such that the likelihood of executing an option increases as performance relative to aspiration decreases. As important as risk tendencies to opt ion execution is the go al the organization wishes to achieve with stra tegic options. Chapter 2 di scussed how aspirations are sometimes defined in the context of particular market positions or technologies. How the firm defines its goals determines which op tions it strikes as much as how much uncertainty it is willing to tolerate. Partic ularly when the market experiences a dramatic
53 change, aspirations are likely to be framed in more qualitative terms, such as producing a particular product or mimicking a market position. To close these gaps, firms will undertake similar actions. As a result, a lthough the firm may have explored other options, the best way to close the attainment discrepancy gap is to execute options similar to those of other firms. Proposition 3-7 As attainment discrepancy increases in absolute terms, the options a firm strikes will become more and more similar to the options of other firms. Survival Bias In addition to the changes in goals suggest ed above, there is a second form of poor performance conceptualized in the risk liter ature. While, this chapter has focused on aspirations as the organizationâ€™s primary goal in prosperous times, firms are also very concerned with survival as a going concern. As performance begins to decline too close to bankruptcy, managers begin to focus on the likelihood of the firm going bankrupt rather than the likelihood of the firm incr easing its performance back to its aspiration level. Shapira (March & Shapira, 19 87, 1992; Shapira, 1995) suggests that the performance can decline only so much before managers begin to focus more on the likelihood of survival rather than on attainme nt discrepancy. Where managers focus on survival as a goal rather than on aspirations , they will behave much more conservatively as they will work to preserve the firm (Mu llins & Forlani, 2005) rather than attain its aspiration level. The potential for failure freezes their ability to conceptualize new options and increasingly forces them into he uristic thinking. This is the property behind the threat-rigidity hypotheses (S taw, Sandelands, & Dutton, 1981). For firms performing so far below their aspi ration level that they begin to focus on their survival barrier, options are much more difficult to execute. Now, managers are
54 concerned that every dollar the firm spends is driving it closer and closer to bankruptcy. The simple cost of options and the fact that managers will be re sponding to the increasing threat with less imagination implies that th eir risk tolerance will once again decrease. Now firms will not execute as many options as they would have if their risk tolerance was strictly increasing in the performan ce domain. Instead, because risk tolerance decreases as performance declines beyond a certain point, options execution will show a similar decrease as performance declines. Proposition 3-8 As performance decreases to a level close to bankruptcy, the likelihood of a firm executing a strategic opti on unrelated to its current portfolio will decrease. Bankruptcy, or the avoiding of bankruptcy, could be termed as an aspiration unto itself and as such is a special case of the th eory above. The firm is faced with working against two conflicting goals, either of which requiring a di fferent mind set. Trying to meet the primary aspiration of profit suggest s risk-taking while th e survival aspiration suggests risk-avoidance because the firm, if operating, is always operating above its survival aspiration. This model conceptuali zes the two as distinct rather than place survival as another simple aspiration because the survival-aspiration presents a different payoff structure than a typical aspiration. Attainment disc repancy on the survival goal leads to bankruptcy and exit not search and risk -taking. If the survival goal is not met, bankruptcy will result and the firm will be unobserved. Because survival is a unique form of an aspiration, although its behaviors ar e consistent with the model, I have held this case outside the formal model becau se survival has an asymmetric payoff.
55 In summary, this chapter developed a model of real options generation and execution. Aspirations create th e motivation for the firm and different behaviors result as this motivation changes. As attainment disc repancy increases, firms search for ways to grow their firm. As performance increases, managers become increasingly loss-adverse and will no longer spend the f unds to execute a real opti on, instead deferring more options to the future. If performance relativ e to aspirations decreases, firms will begin to execute more options and behave in a more gain-seeking way. The next chapter will place this discussion into an empirical context. Because aspirations in an uncertain environment re volve around matching market positions rather than profits, we can isolate this mechan ism by looking at a new market environment where turbulence restricted managementâ€™s abil ity to look at performance and encouraged it to focus on other sources of comparison. The next chapter will compare two different social groups to examine their aspiration behavi or and the market positions that resulted.
56 Figure 3-1. Model of Corporate Decision Pro cess. Resources are transformed into an option pool through recognition. The deci sion to execute an option is based on a firm specific threshold. 0 20 40 60 80 100 120 -10010 A ttainment Discrepanc y Option Purchased 0 20 40 60 80 100 120Option Similarity Number of Options Purchased Option Similarity Figure 3-2. Option Behavior Relative to A ttainment Discrepancy. As attainment discrepancy increases, firms purchase increasingly dissimilar options, but they execute increasingly similar options.
57 -2 -1 0 1 2 -15-10-5051015 Attainment DiscrepancyValue A spiration Point Figure 3-3. Prospect Theory Power Curve. As attainment discrepancy moves beyond the aspiration point, the relative value of pr ofits and losses changes. To the left of the aspiration point, the relative value of a profit is higher than that of a gain. To the right, the relative value of a loss is higher than that of profit. So, firms are risk seeking to the le ft and risk averse to the right.
58 CHAPTER 4 HYPOTHESES AND MODEL This chapter will set the preceding chapters in an empirical context. To do this, the modelâ€™s scope must be limited. An industry w ith high levels of ma rket uncertainty and low amounts of competitor information was used. So this study limits, by the choice of a context, the importance of past performa nce and the performance of other firms on aspiration formation. This study examined market entry and option purchase decisions in the competitive local exchange carrier industry (CLEC, pr onounced â€œC-lekâ€). This extraordinarily competitive environment was legalized only in 1997 with the passage of a The Telecommunications Reform Act in 1996. Some market entrants were entrepreneurial while others were active competitors in othe r industries. Because the industry was so new, there was a lot of market entry and a considerable number of firms. The FCC documented over 200 firms in this industry by 2000, when it was only three years old (Rangos & Lynch, 2001). Because of the high uncertainty in this industry and the predominance of private firms who do not report performance info rmation publicly, the empirical context provides a uni que opportunity to examine the tendency for firms to set aspirations based on the expected performance of competitor firms, the competitive parity goal. This paper makes a contribution to existi ng aspiration literature which must be isolated from existing theoretical influences , thus this paper has used the specialized context of a new industry. The theory must be tested in an environment where firms are
59 either robbed of relevant performance informa tion or that information is easy to identify and isolate. In the CLEC i ndustry, the availability of fina ncial and market performance information was severely restricted to such an extent that such hist orical performance and comparative firm performance measures were not available. Figure 4-1 presents the observed component of the theoretical model de veloped above. In this industry, there is no consistent context from which to judge pa st performance because it was so new. In addition, there was no public performance inform ation to judge other firms. As a result, organizations could only use competitive parity or relative market positions as their goal. In addition, search, although it is a key process variable in the model, cannot be directly observed in this c ontext. The empirical contex t only provides information on attainment discrepancy and the outcome of the search process, option purchase. Similarly, thresholds are not obs ervable in this context, so this study investigated whether attainment discrepancy moderates the relationship between option purchase and execution. The sections to follow will develop the importance of firm origin as a source for aspiration formation. This will be done to isolate two behavioral groupings in the hope to generating a more robust test of the theory. Following this development, this chapter will propose specific hypotheses relating to option purchase and execution based on attainment discrepancy. Within Group Aspirations The prior chapters developed a model whic h suggests that systematic aspiration differences between firms motivate the acquisi tion of real options and the execution of those options. As uncertainty in the market increases, firms define their aspirations in terms of competitor market positions. Mana gers learn about their competitorâ€™s market positions by watching their competitors in the market (Greve, 1998a) and set their
60 aspirations through this learning process. As uncertainty increases , the magnitude of a firmâ€™s attainment discrepancy depends on the firms that managers use to define their competitors. Different firms have different peer groups, and the level of their discrepancy depends on which firms they are tr ying to compete with. As this attainment discrepancy increases, risk prof iles change and search increases. One byproduct of this behavior and a result of this model is mimetic investment. Because success is defined in terms of comparable market positions, managers will seek to close the gap with competitors by a dopting similar market positions to their competitors. If the firm does not current closely imitate the market position of its competitors, mangers behave increasingly risk-seeking as this gap increases. In the presence of uncertainty, firms borrow best practices from one another because no individual firm has a good idea of what will actually be the best strategy in a given market context. In the end, firms te nd to look very similar to one another as bandwagon effects encourage firms into sim ilar strategies. Scholars have examined firmsâ€™ tendencies to make these kinds of investments (Mitchell, 1989), the types of imitative investment firms undertake (Haunsch ild & Miner, 1997) and from whom the firm is most likely to borrow investment strate gies (Greve, 1998a). As uncertainty in the market increases, underperforming firms begin to attempt imitation of competitor firms and will do increasingly aggressive things to get th ere. In this sense, market imitation is a byproduct of an aspiration process. These studies build on the differences betw een firms in terms of prestige or size, but they often ignore the inhe rent differences that firms have when they enter a new industry. Entrepreneurs form a distinguishab le subgroup within any market context.
61 Despite the acknowledged importance of or igin in studies into population ecology (Carroll, Bigelow, Seidel, & Tsai, 1996; Klepper & Simons, 2000), the importance of origin and prior experience has not been tr anslated into the isomorphism research domain. Learning and mimicry in the isomor phism field has largely been based on the construction of social comparison groups thr ough the use of observa ble and verifiable divisions. However, these groupings ignor e the importance of social identity and cognitive groupings that managers form usi ng classifications which are not as easily discernable ex ante (Reger & Huff, 1993). This study expands this notion by sugges ting one particular cognitive grouping, origin of new market en trants (Carroll et al., 199 6). Entrants can be classified as either entrepreneurial firms (de novo entrants) or es tablished firms (de alio entrants). The following section suggests that entrepreneuria l firms are more likely to learn practices and set aspiration levels from each other. Wh ile market behavior of established firms in an uncertain industry should be highly related to one another as they seek to close the performance gap, entrepreneurs should not di splay this tendency to borrow behavior from other firms, as their mental models will be more heterogeneous than the mental models of managers in established firms. The first se ction and series of hypot heses will suggest one relevant cognitive grouping used by management, origin. This section will also explore the impact of these groupings. As managers begin to define aspirations in terms of competitors, the options purchased by firms will become increasingly similar within the cognitive grouping. The second set will suggest th at as attainment discrepancy increases, managers will execute more options. Attain ment discrepancy and aspirations will be defined in terms of comparative firms, and th e end result will be imitation among firms of
62 similar origin moderated by the difference be tween current market position and average market position within the group. Imitation and Option Behavior Firms imitate best practices from one a nother based on the tr aits of the firm undertaking the action. The impor tance of similarity is magnified by the uncertainty in the market. There are three major compone nts of market behavior that encourage imitation (Haunschild & Miner, 1997). The firs t group is the similar ity between the focal firm and other firms in the market. Studies have suggested that firms learn from other firms of similar size, simila r geographic location, and sim ilar market competence (Greve, 1995, 1996). The second predictor of imitation is the frequency of that action through the population, and the last is the tendency for that action to result in positive outcomes (Haunschild & Miner, 1997). Th is section will describe the tendency of firms to engage in the first type, trait-based imitation. In the second type, frequency-based imitation, firms imitate actions they see more frequent ly. The third type, outcome-based imitation, where firms imitate based on outcomes obser ved from a competitorâ€™s action is unobservable in this environment. Trait-based Imitation To imitate other firms, firms first gather information by observing the behaviors of other firms. By watching these firms, managers at the focal firm get a better understanding of what is in th e best interest of their firm. The more firms the manager watches, the better his or her understanding of their market environment becomes. However, managers are boundedly rational and cannot monitor every relevant firm (Cyert & March, 1963). They have to set boundari es and often end up focusing on firms who are very similar to th eir own firm. Studies have shown that firms use proximity and size
63 to identify these relevant se ts (Greve, 1995, 2000). This pa per suggests the addition of firm origin as one characteristic that f acilitates these relevant group distinctions. Research, both recent and classic, has established the importance of reference groups to the development of trait-based im itation. Greve (1995) found that firms were likely to abandon a strategy if that strategy wa s also abandoned by firms in a focal firmâ€™s reference group. Other studies found that th e presence of a large firm in a profitable market encourages other large firms to enter into that market (Haveman, 1993), suggesting that firms with similar resource pool s imitate one another. The importance of origin as an imitation driver has been suggested by St uart and Sorenson (2003) who found that new venture creation in a geographi c location was highly related with the prior success of other new ventures in that locati on. There has been no formal hypothesis that firms of like origin will imita te each otherâ€™s behavior although there has been work to suggest they share similar survival likelihoods (Carro ll et al., 1996). The model above suggests that firms in the same reference group will seek to maintain a similar competitive position to fi rms in the group. Their aspirations will be defined in terms of competitor market posit ions and the outcomes of that grouping will be similarity in their option purchases. The tendency for firms to group one another into competitor groups will lead them to purchase options which are similar to the options purchased by other firms within the group. Hypothesis 1 In uncertain environments, the op tions purchased by entrepreneurial firms will be more related to other entrepreneurial firms than to established firms. Hypothesis 2 In uncertain environments, the options purchased by established firms will be more related to other establis hed firms than to entrepreneurial firms.
64 Scope of Purchase The model developed in Chapter 3 also s uggests that firms performing at different levels relative to their aspiration level will behave differently in purchasing option. Specifically, as their attainment discrepancy increases, their search process will be much wider and their tendency to pur chase unrelated options high er. Thus as attainment discrepancy increases, either positively or negatively, firms should purchase options that are increasingly different from the options already in their portfolio. This suggests Hypothesis three and four: Hypothesis 3 As absolute levels of attainment discrepancy increase, firms will tend to purchase more options. Hypothesis 4 As absolute levels of attainment discrepancy increase, firms will tend to purchase options which are increasingly unr elated to their current option portfolio. Option Execution In addition to learning and imitation mo tivated through attain ment discrepancy, Chapter 3 also suggested that not all of these imitative investments will be followed through to actual market entry. Firms undertak e an investment to learn about the market and learn about the competition, but the aspi rations of the firm and the attainment discrepancy will predict under what conditions the firm will actually strike the option and enter the market. The model developed in Chapter 3 suggests that as discrepancy decreases (performance increases above aspirations) the firm will not strike options because it is loss-adverse. Similarly, as the attainment discrepancy increases and performance falls, the firm will be more gain-seeking as it tries to keep up with competitors by investing heavily.
65 Hypothesis 5 As attainment discrepancy increases (decreases), more (less) options will be struck. As attainment discrepancy increases, the fi rm has closed its search process down to a very small segment of possible alternatives . The firm will purchase options that are similar to other firms and it will seek to cl ose the attainment discrepancy gap with those firms. This will lead it to execute options th at are similar to other firms. Its tendency to define its strategy and success in terms of other firms will lead to homogeneity. So, although the options the firm purchased will be quite different from those of competitors as a result of its broadened search patterns, the options that it actually strikes will be very similar to those of competitors because success is defined in terms of similarity. Hypothesis 6 As attainment discrepancy incr eases (decreases), executed options will be more (less) like the options executed by comparative firms. The reader will note that these hypotheses do not incorporate Proposition 3-8. This proposition is not testable in the current data set because the ambiguity in the environment means that public information regarding proxi mity to bankruptcy is not available. Secondly, recent empirical work by Miller and Chen (2004) has not displayed this threatrigidity tendency. As such, this particular component of the mode l will not be tested here. While no single piece of research can inva lidate a theory such as the threat-rigidity hypothesis, there is sufficient r eason to leave further investigation of th is topic to other papers. Sample To test these predictions, this study used a competitive subfield of the telecommunications industry, the competitive local exchange carrier industry (CLEC). This very competitive and dynamic industry began primarily following the
66 Telecommunications Act of 1996 to offer local telephone service in competition with the bell operating companies established after th e break-up of AT&T. Industry data for this study comes from industry reports published annually by New Paradigm Research Group (NPRG) and the Local Exchange Routing Guide (LERG). The New Paradigm Research Group publishe d yearly consulting reports on this industry for each year and listed a comprehe nsive database of firms who considered themselves members of the facilities-based C LEC industry. There were at least two other strategies opened up by the communications act , and this study limits its examination to one sub-industry, the facilities-based CLECs. For the purposes of this sample, firms w ho enter a market unde r a strictly retail arrangement and do not build a switch will not be counted as entering a market. This was done to isolate firms competing in the industr y using the same techniques, which seek the same customer base, and operate under the same regulatory arrangement. Retail-only firms sought different customers and were subj ect to a different leasing arrangement with the incumbent carriers. As a result, this study held them aside. The extensive consulting reports cover th e industry from 1995 through 2005 for the 50 United States. These reports detail firm performance information such as revenue as well as the industry presence of each firm, measured by such variables as the number of networks a firm has and how many miles of fi ber optic lines a firm has in place. The firms in these consulting reports form the basi s of the sample. After excluding firms for which no market information could be obtai ned, the final sample was 156 firms over 9 years.
67 Dependent Variables This study has two dependent variables, th e purchase of a strategic option and entry into a market. The purchase of a strategic option in a market will be measured by the acquisition of an operating company number (OCN) from Telecordia, the administrator of the Local Exchange Routing Guide (LER G). Entry into the LERG database means that the firm appears in the North Ameri can telecommunications network. The LERG is the database used by phone companies to rout e traffic across the ne twork, and a presence on the network is required in order to operate. The OCN is the first step in that listing process. The purchase of an OCN will se rve as a proxy for certification by the state regulatory authority. It is a viable option because it is 1) visible 2) has no incremental revenue associated with owning an OCN 3) is not necessarily transferable and 4) is not expensive beyond the process of state cer tification. Once a CLEC has attained certification to operate within a state, the CLEC sends a copy of its certifica tion letter to the LERG administrator along with $200. The LERG administrator then returns an OCN number to the firm. The firm then uses this OCN to notify the LERG administrator if it places any telecommunications equipment within the state. Market entry and market behavior were also assessed using the Local Exchange Routing Guide, or LERG tables, which li sts the switching loca tions of every phone switch in the country in addition to a table of certified operators. When firms entered new markets, they installed a switch in the local telephone network. These switches varied in their capacity and complexity of service offerings. Because the tables are updated nightly, the information is a public and visible way for competitors to communicate competitive behaviors. In addi tion, because the information is used to route telephone calls and is m onitored by the FCC, it is hones t and accurate information.
68 In this study, market entry was defined as the installation of a new telecommunications switch (voice or data) in a metropolitan se rvice area (MSA). Th e metropolitan service areas employed in this study are those defined in the 2000 census. Independent Variables Origin in this study is a dichotomous variab le. In this study, a firm can either be entrepreneurial or established. Origin wa s derived from the New Paradigm Research data. These reports detail firms in the CLEC industry who competed by building new facilities. In addition to help ing to isolate a relevant sample , the short description of each firm identified which ones were entrepreneurial. The classification of the firms into subgroups was confirmed by internet and database searches on each firm. Entrepreneurial firms are defined as any fi rm who entered the industry de novo with no funding from an established firm and was stil l operated by the firmâ€™s founder. A firm that received funding from any source aff iliated with a telecomm unications, electric utility, cable television comp any or other kind of going c oncern involved in consumer services or products was not classified as en trepreneurial. Information on the supply of credit from input suppliers was not available co nsistently in our sear ch, so we did not use this as a criterion for classi fication. However, credit from an input supplier probably did not have a large influence on our sample as most firms tended to get some kind of credit from their equipment suppliers. Funding wa s an important separa tor of the sample because of the prevalence of corporate ve nture capital in this industry, a technique sometimes used by firms to establish a toe hold investment in a new industry (Wadhwa & Kotha, in press). As such an entreprene urial firm who received corporate venturing funding may actually be better desc ribed as a strategic asset of an established firm rather than as an entrepreneurial firm. Because th e use of definite requirements on what firms
69 are entrepreneurial and those that are not, the sample incl udes firms who were at first classified as entrepreneurial and we re later classified as established. The attainment discrepancy measures seek to capture the difference between the focal firmâ€™s market position and those of its competitors. This was done by measuring a firmâ€™s differences along several key market va riables versus its competitors. Attainment discrepancy was analyzed along four market dimensions. The relevant competitor group was defined within two different groups. Th e reference groups were classified as the whole industry and firms within the industry w ith the same origin. The first measure of attainment discrepancy is the difference betw een the number of options held by the focal firm and those held by the average firm in the reference group. Th e second measure is the difference between the numbe r of markets which the focal firm operates in and the markets operated in by the average firm in th e reference group. The third measure is the difference between the population reachab le by the focal firm and the population reachable by firms in the reference group. The fourth measure is the difference between the number of business establishments reachabl e by the focal firm and the average firm in the reference group. Thus attainment discrepa ncy is the difference of the firmâ€™s position from the market position of a relevant comp arison group. Market position is measured in four ways, and these four measures were aggr egated into a single measure using principal components analysis. When aggregating these values, the facets were first verified to load on a single factor. Because only one ei genvalue was greater than one, the facets were rotated using an oblique rotation. The resulting factoring weighting was applied to each set of four variables in each observation and the resulting aggregated variable was used as the attainment discrepancy variable.
70 These are only potential indicators of this construct. This study employed four measures of attainment discrepancy, all of which are suggested by the existing work in aspiration theory. By using four dimensi ons and aggregating them using a factoring procedure, this study has tried to ensure that multidimensionality of expected market performance is captured and that the measur e employed is as complete as possible. Using competitive characteristics, such as option and market holdings as well as market characteristics such as the potential populat ion and businesses which a firm can connect seeks to balance out the size differences acros s markets with the size differences across firms. A firm might be in twenty markets, but all of those markets together might not be as large as New York City. Using a weigh ting scheme, as this study did, allows for a balance between competing measures. While th is is a measure which exists outside the firm and may not capture the actual feelings of managers within th e firm, it is a measure which uses a new technique while still bei ng consistent with the existing literature. State Characteristics This study employed measures of population derived from the decennial census. In addition, the study incorporated the number of business establishments and total employment figures within each state from the county level business pattern data provided by the Census Bureau. Because desc riptive statistics for these variables are important in the calculation of option and market characteristics, their use will be described in the tables discussed below. This study did not employ measures of geographical size or land area of the states because these figures do not change over the study period and are not us eful in estimating the statistical models employed. Finally, at a state level, it is important to control for the regulatory characteristics of each state. Interconnection rates were used instead of a qualitative measure describing
71 the political leanings of th e stateâ€™s regulatory commission because the interconnection rate is a closer measure to the actual re gulatory policies towards entry which affect business conduct in this state. Interconnec tion rates are overseen by the individual state public service commissions. These rates are of fundamental importance when describing a firmâ€™s business prospects in a state because they list how much money a firm is allowed to charge for connecting telephone traffic. Each state has the right to decide its own rate. This data is provided by each state individually and some years are missing, particularly in the early stages of the industryâ€™s devel opment. In these cases, the study employed averages. When the state did not list an interc onnection rate in one year but listed a rate in a prior year and a subsequent year, the ol der rate was carried forward until there is a new rate to replace it. In this way, the st udy assumes that a lack of new information means that the rate has not changed. This is a better assumption than using an interpolated value because state regulators are inclined to take no action until there are court challenges or industry pr essures. Thus rates are more likely to remain unchanged rather than the drifting that would be implie d if price were modele d using interpolation. When the states did not list an interconnecti on rate and never liste d one (the observation was missing), the national average rate for th at year was used. Two states, Alaska and Hawaii, did not list interconnection rates. As a result, the yearly national average was used for the rates in these states. This will effectively remove the influence of interconnection rates the decisions modeled for these two states. The rates used in this study were for interconnected traffic in dense areas as opposed to an adjusted average for the entire state because the relevant market entry
72 event on which this study is trying to descri be option purchase is entry into the dense markets defined by the MSA. Option Characteristics Because this study uses a type of option that has not been dealt with in the literature before, one that depends on the product ma rket more than the stock market, it is necessary to generate a consid erable number of va riables to classify the uncertainty and value associated with each market. This op tion has some very usef ul characteristics. First, the use of state cer tification as an option implie s a time horizon. Most states require that a company who seeks certificat ion must open operations within one to two years. Secondly, the option describes the state, so the certification by an individual state means that the firms might be purchasing a bundl e of options to enter the different MSAs within the state. Because the option is at the state level, uncertainty and value of the option, its two defining characteristics, need to be described at the st ate level. Table 4-1 through 4-4 offer the measures generated for each variable group. In addition, the righthand column details the source from which the data was collected. Option uncertainty In this study, market uncertainty is the cha nge in market characteristics over time. These measures are meant to capture the cha nges in market conditions from one period to the next and are described at the state level. These measures are included to control for the static and dynamic uncertainty in the market option. The column on the left side of Table 4-1 is the data item wh ile the right-hand side lists th e source for the data. These items are combined in the analysis using the same principal component procedure discussed above into a single measure of option uncertainty and the growth in that uncertainty overtime.
73 Option value An optionâ€™s value is the potential revenues from the market. This can be assessed using physical characteristics of the market and the growth in those characteristics. Option value also included measures of the num ber of initial public o fferings in a state as well as how much capital was issued to that st ate last year. This is included because research shows that firms enter markets that have just experienced a growth in capital issuance (Stuart and Sorenson, 2003). Tabl e 4-2 lists the measures employed in measuring an optionâ€™s value. The values wi ll be included as control variables in the model. Option similarity Similarity is a measure of how differen t the option is from the average of a particular reference point. The reference point is the average the firm uses to determine how different this option is from either the op tions that it holds, firms of similar origin hold, or the industry as a whole holds. For in stance, competitor similarity to the focal firm is the difference between the number of competitors in this market and the average number of competitors in the other markets th e firm operates in. Distance similarity to the origin group is the average distance of this option from the option held by other firms with the same origin as the focal firm. The focal marketâ€™s similarity with the options already held by the firm was determined along five dimensions. This was done to be as inclusive as possible about the dimensions along which simila rity could be assessed. Competitor similarity is a measure of th e size differences between the focal firm and the firms in the market. Population and business environment similarity measures differences in the population count and busine ss establishment count between the markets
74 the focal operates in and the potential new market. Distance similarity is a geodesic distance from the focal firmâ€™s markets to th e potential new market. Regulation similarity is the difference between the interconnecti on rate of the new market and the firmâ€™s existing markets. These measures, once calculated were aggr egated using the same factor analysis procedure employed to aggregate the attainme nt discrepancy measure. All aggregated variables had nothing less than a 0.7 correlation between the items and the analysis did not reveal two factors. A two factor soluti on is one where two or more eigenvalues are greater than 1. Table 4-3 lists the name of the aggregated measure, the individual data item, and the source of the data item. Similarity is a weighted score that in corporates average difference, squared difference, and minimum difference. This wa s done because of the lack of theoretical clarity on the issue. The use of differences is meant to assess the extent to which a firm would need to stretch itself e ither geographically or operationa lly to meet the needs of a market. By this definition, the minimum diffe rence would seem to be the best indicator of the difficulty of entering a market. A ma rket close to a potential market should be more similar to the new market and thus offer more information than a market quite distant. However, the average is the most common measure used in the literature to assess the extent to which the new market is different from the center of a firmâ€™s operations. For example, consider a firm w ith five markets on the east coast who enters two markets on the west coast of the United Stat es. If the markets in the west are only a few miles apart, a minimum distance measure would be considerably different than an
75 average. There is no theoretical resolution to this, so this study in cludes both and weights each according to its convergen ce to a central construct. Hypotheses one and two discussed the similarity between options owned by different firms. While the within-firm si milarity is a differe nce between individual markets, the between-firm similarity is the difference between the firm and the entire industry. Thus, an option which is geographica lly quite distant from the firm, and this distance is above the average distance for the new options in th e entire industry, the option is dissimilar. As with the above measures, dissimilarity is assessed on three dimensions for each group and these groups we re aggregated using factor analysis. Dissimilarity was measured within the industr y as a whole (overall) and within origin groups (origin). Market Characteristics The data to describe the market is very si milar in its nature to the information about the option. Once the option to en ter a market is purchased, the data needs to describe a) what firms are in the market and b) the c onditions which model entry into the market. Table 4-4 presents those values. These measures will not need to be aggreg ated. They can be entered in a model to look at the different influences of each on the likelihood of entry having controlled for the likelihood of purchasing the option. Data Considerations Values from the LERG were extracted ba sed on the criteria se t out in published studies using the LERG tables (Brown & Zimmer man, 2004). This was to ensure that the study measured entry using the proper equipment as some items in the LERG are not switches but line aggregators (multiplexers). To select data points from the LERG, the
76 NPRG data was matched to the table listi ng all the operating comp any numbers in the LERG database, the LERG 1 table. B ecause the LERG database did not assign individual companies with an overarc hing number until after 2003, every operating company number (OCN) in the LERG for 1996 through 2005 was manually match with a list of companies in the NPRG data. This matching was manually co rrected and checked. The matching was not straightforward. Of the 321 firms in the initial database, many companies were acquired (96 firms) and some changed their names (46 firms). Furthermore, the OCN numbers and names did not always match a company the NPRG data had listed as being in existence. For instance, if a firm was acquired, it was often still listed in the LERG as bei ng part of the independent company, as if it had never been acquired. Many times, acquiring firms did not go to the effort to reregister their certifications with the LERG database. In these cases, firms were matched, in order of importance, on the old company name, on the name of any holding company, and finally on the name of the person listed in the databa se as the contact representative for that number. This served as the measure of owni ng an option on the market. As a measure of when a firm was in the market, LERG 6 and LER G 7 (other tables in the database) were used to determine if the firms was offeri ng phone numbers and was operating a switch in a particular market respec tively. If the firm was both offering phone numbers and operating a switch in the market, the firm was counted as operating in the market. There were several duplicates and err oneous data items in the LERG. Many duplicates were simply values which had been replaced by an updated entry. In these cases, the older observation was dropped and the updated observation was retained. True double entries were grouped a nd one observation was deleted at random. Observations
77 where the switch was marked as blank or the operating company was marked as â€œTRA 14â€ were also dropped. The LERG 7, the table which provides the information about the physical location of the switch, presented a different problem . Here, although the street address would seem correct, the zip code assigned to the swit ch would change over time. As a result, it was impossible to determine in what MSA the switch was located by using the data in the LERG. To get around this, the vertical and horizontal (V&H) coor dinates provided for each switch were used for the switchâ€™s loca tion. The V&H coordinate s are the results of a flat projection of the Eart h and are used by phone companies to determine distance as an alternative to using longitude and latit ude. Using the V&H projections, each switch was mapped to the current zip code division s and those divisions were the basis for assignment to an MSA. The matching was done by finding the zip code for which the distance between the switch and the cen ter of the zip code was minimized. In addition, the extensive use of Decennial Census data may be a problem. To achieve information for intervening years between 1990 and 2000, values were interpolated at the zip code leve l and then aggregated to the st ate or market level. For the years 2000 through 2005, the 2000 value was multiplied by the growth rate projected for each state by the census bureau. The aggrega tion of these measures will be discussed later. While the use of interpolated data may cause concern, it is used here for two reasons. First, the census is the publicly ava ilable information which is most likely to be used by smaller and medium sized companies when determining the size of a market. 4 A TRA1 designation indicates that the OCN is curre ntly operated by Telecordia, the company in charge of the LERG. Telecordia is not a competitor in the industry; it is a quasi-government entity that only manages the database.
78 Firms only need rough estimates of population gr owth, and the census data provides that. Secondly, by using interpolated census data th is study can measure market information at a much more detailed level than that provi ded on a yearly basis by the census bureau. County level business pattern data, the LERG database, a nd the SDC database were collected on a yearly basis and no interpolation was necessary. The initial sample consisted of 164 firms after controlling for acquisitions, name changes, and actual operation of a switch. Many firms listed as facilities-based CLECs in the initial database never actually began ope ration, and they were dropped. Of the remaining firms for whom the initial databa se listed as operating, eight could not be located in the LERG tables, so they were dropped. Of the remaining firms about 35% each year were entrepreneurial firms. The final sample consisted of 156 firms over the 9 years from 1997 through 2005. The year 1996 wa s a base year used only for variable calculation. There were 266 potential MSA markets and 50 potential state option purchases. This study used four datasets fr om these variables. Hypotheses which looked at individual markets used e ither a state by firm by year or market by firm by year dataset. There were 41,664 observations in the state by firm by year dataset and 269,192 observations in the MSA by firm by year datase t. Hypotheses which looked at firm-level questions employed summation of these two da tasets where either option totals were created or market totals we re created within firm by year groupings. There were 893 observations in the pooled state by firm by y ear dataset and 884 observations in the MSA by firm by year dataset. The innumerable calculations and summary statistics as well as the database programming required presented quite a problem in the development of this dataset.
79 Because prior extensive computer programming efforts, such as this, have provided the computer code and to offer interested partie s the ability to double check the accuracy of this programming effort, the Microsoft Access 2003 Visual Basic code used to construct the dataset is offered in Object 1. Object 1. The Visual Basic program code us ed to generate the dataset (266 KB, Visual Basic Code.txt). Model This study has developed a model of two stage investment decisions where the decision to purchase and option and the d ecision to enter a market are intimately connected and not conceptually divorced fr om one another. However, the models presented below will provide a sequence of models that take these decisions apart. While it is possible to model these decisions join tly using a censored bivariate probit model (Boyes, Hoffman, & Low, 1989), this model is difficult to apply in this case because MSAs are not always nested within the states . Thus, there are a few MSAs which require multiple options to enter. For instance, en tering every facet of the New York market might require options on New Jersey, New York, and Connecticut. This study did not employ this model because the end result will not be as conceptually clear as would a series of one-stage models. This section will describe the models which will be used to models each stage in the decision and what is required by these models to support the hypotheses. In addition, when helpful, a ma thematical representation of the model is provided. All independent variables are lagged one year such that the entry decision in any year is based on variables in the preceding year.
80 Hypotheses 1 and 2 In addition to the importance of the link between option purchase and option execution, this study has suggested that the op tions purchased by firms will be connected to the options purchased by firms in the same reference group. To that end, Hypotheses 1 and 2 suggest that entrepreneurial firms were more likely to purchase options similar to other entrepreneurial firms and established fi rms were more likely to purchase options similar to other established firms. Thes e hypotheses were tested using a parametric hazard model. The use of proportionality al lows for the possibility that the option purchase behavior of firms does not follow a typical distribution, but the proportional hazard assumption was shown to be violated by a test employing Schonfeld residuals in other models. Because proportionality did no t hold for all models, a parametric model was used for all models instead. An expone ntial distribution was used to model the failure event for study, the purchase of an op tion. Significant and ne gative coefficients on the interaction between opti on similarity when defined w ithin groups and origin will support these hypotheses. Similarity with othe r firmsâ€™ options was measured as geodesic distance, business environment difference s, population differences, competitor differences, and regulation differences. These will be measured as differences from the focal firmâ€™s position. This similarity scale in creases as the option is less similar from the firmâ€™s other options, so it is better termed a dissimilarity measure. Hypothesis 3 The third hypothesis suggests that the number of options purchased will increase as attainment discrepancy increases. The depe ndent variable in this model is a count variable that cannot go below zero. Because it is necessary to correct for the datasetâ€™s panel structure, the results will employ a pa nel-corrected fixed-effect Poisson model.
81 Positive and significant coefficients indicate a greater tendency to purchase options. Hypotheses 3 will be supported by significant co efficients on the squared attainment discrepancy variables. i 1, 1 t 4 1 t 3 1 t 2 1 t 1 t in _same_orig iscrepancy tainment_d squared_at iscrepancy tainment_d squared_at igin cy_same_or _discrepan attainment cy _discrepan attainment options Number_of_ (1) Hypothesis 4 The fourth hypothesis suggests that as atta inment discrepancy increases, firms will be more likely to purchase dissimilar options . Firms purchase increasingly unrelated options in the hope of returning to a stable performance level. This was modeled using the same parametric hazard model employed to test Hypotheses 1 and 2. Just as the model in Hypotheses 1 and 2, the dependent vari able in this equati on is the purchase of an option. The relevant coefficient is the in teraction of attainment discrepancy and the option similarity measures. If the coefficien ts are positive and signi ficant, the likelihood of purchasing an option increases as the si milarity of the optio n decreases and the magnitude of attainment discrepancy increases. In short, a positive coefficient suggests that the firm is less likely to buy similar options. In this model, attainment discrepancy will be a squared value rather th an a linear value. The squared value is used instead of an absolute term as a matter of preference and to aid in coefficient interpretation. i 1, 1 t 1 t 6 1 t 1 t 3 1 t 2 1 t 1 t igin cy_same_or _discrepan attainment * igin cy_same_or _discrepan attainment cy _discrepan attainment * cy _discrepan attainment alue expected_v y uncertaint purchase 1 t 7 1 t 5 1 t 4ilarity option_sim ilarity option_sim ilarity option_sim (2)
82 Hypothesis 5 Hypothesis 5 hypothesizes the conditions unde r which an option is struck and when firms are more likely, in the aggregate to st rike market options. It suggests that the option strike behavior is a lin ear relationship of attainment discrepancy. As with the prior models, this was modeled as a fixed-eff ect Poisson model. To control for the twostage nature of the game, the models will onl y be estimated for firms who have purchased an option. This will reduce the sample size bu t is a more appropriate method. Here, with the inclusion of several market characteristics as control, attainment discrepancy should increase the tendency to execute an option. Further, attainment discrepancy when defined within a group with the same origin should have an additive effect on the likelihood of market entry. Equatio n 3 depicts the empirical model. i 1, 1 t 4 1 t 3 1 t 2 1 t 1 t in _same_orig iscrepancy tainment_d squared_at iscrepancy tainment_d squared_at igin cy_same_or _discrepan attainment cy _discrepan attainment tered Markets_en (3) Hypothesis 6 The final hypothesis required two analys es. The first, which employed a logit model, looked only at options which were executed immediately following purchase or one year after purchase. As discussed in Chapter 3, the decision to execute an option immediately must be isolated from the decisi on to defer an option. Because the model in this study is static, a first attempt to model entry behavior should l ook at entry decisions that were tightly connected with the option purchase beha vior. Looking at immediate entry decision lowers the observed behavi or being influenced by the shifting of aspirations after prolonged periods of di screpancy. Figure 4-2 shows the number of options executed at each year of ownership. The tendency to execute an option decreases
83 as the period of ownership increases, but th ere are substantial numbers of options which are executed only after the firm has held onto them for a while. The second part of Hypothesis 6 will be modeled just as Hypothesis 4, using a parametric hazard model while using option strike as the failure event rather than option purchase. In this hypothesis, as attainment discrepancy increases, the similarity of the markets entered should decline. Thus, if the interactions of attainment discrepancy and the similarity measures are above positiv e and significant, this hypothesis will be supported. i 1, 1 t 1 t 6 1 t 1 t 3 1 t 2 1 t 1 t igin cy_same_or _discrepan attainment * igin cy_same_or _discrepan attainment cy _discrepan attainment * cy _discrepan attainment alue expected_v y uncertaint ike option_str 1 t 7 1 t 5 1 t 4ilarity market_sim ilarity market_sim ilarity market_sim (4) Chapter 5 will discuss the results of these models and the implications for the literature. Table 4-5 summarizes the datasets and models used to test each of the hypotheses.
84 Figure 4-1. Observed Theoretical Model. Items outlined in a dashed line are not observable in the empirical context. Items circled with a solid line are modeled. 0 200 400 600 800 012345678 Years DelayedNumber of Markets Figure 4-2. Delay Behavior Following Option Purchase. Years delayed is the period between when the option was purchased and when that option was finally executed. The Y-axis displays a simp le count of how many markets were entered after that period of delay. Th e graph displays is a decreasing but important of deferring options to later periods.
85 Table 4-1. Option Uncertainty Measures Aggregated Data Point Measure Source Market Variances Variance (5 years) in employment County level business pattern data from the Census Bureau Variance (5 years) in population Decennial National Census Variance (5 years) in business establishments County level business pattern data from the Census Bureau Variance (5 years) in payro lls County level business pattern data from the Census Bureau Market variance growth Growth in employment variance in year prior County level business pattern data from the Census Bureau Growth in population variance in year prior Decennial National Census Growth in business establishment variance in year prior County level business pattern data from the Census Bureau Growth in payroll variance in year prior County level business pattern data from the Census Bureau Table 4-2. Option Value Measures Aggregated Data Point Measure Source Market Size Number of firms who own the option LERG Database Population Decennial National Census Number of business establishments County level business pattern data from the Census Bureau Payroll in the market County level business pattern data from the Census Bureau Market Growth Growth of the market (population) in the year prior Decennial National Census Growth in the payroll totals for each market County level business pattern data from the Census Bureau Growth in the number of business establishments in the year prior County level business pattern data from the Census Bureau IPOs in Market IPOs in the state SDC Database IPO Captial IPO capital delivered to state SDC Database
86 Table 4-3. Option Similarity Measures Aggregated Data Point Data Item Source of data Competitor Similarity The arithmetic mean difference between the number of competitors in the focal market and the firmâ€™s other markets. LERG database The arithmetic mean squared difference between the focal marketâ€™s number of competitors and the firmâ€™s other markets LERG database The minimum difference between the number of competitors in the focal market and the firmâ€™s other markets LERG database Population Similarity The arithmetic mean difference between the focal marketâ€™s population and the firmâ€™s other markets. County level business pattern data from the Census Bureau The arithmetic mean squared difference between the focal marketâ€™s population and the firmâ€™s other markets. County level business pattern data from the Census Bureau The minimum difference between the focal marketâ€™s population and the firmâ€™s other markets. County level business pattern data from the Census Bureau Business Environment Similarity The arithmetic mean difference between in the focal marketâ€™s count of business establishments and the firmâ€™s other markets. County level business pattern data from the Census Bureau The arithmetic mean squared difference between the focal marketâ€™s count of business establishments and the firmâ€™s other markets. County level business pattern data from the Census Bureau The minimum difference between the count of business establishments in the focal market and the firmâ€™s other markets. County level business pattern data from the Census Bureau Distance Similarity The arithmetic mean geodesic distance from the focal market and the firmâ€™s other markets Latitude and longitude listed in the Decennial National Census The arithmetic mean squared geodesic distance from the focal market and the firmâ€™s other markets Latitude and longitude listed in the Decennial National Census
87 Table 4-3. Continued Aggregated Data Point Data Item Source of data Distance Similarity The Minimum geodesic distance from the focal market to the firmâ€™s other markets Latitude and longitude listed in the Decennial National Census Regulation Similarity The arithmetic mean difference between the focal marketâ€™s interconnection rates and the firmâ€™s other markets. Individual state public service commissions The arithmetic mean squared difference between the focal marketâ€™s interconnection rates and the firmâ€™s other markets. Individual state public service commissions Minimum difference between the focal marketâ€™s interconnection rates and those of the firmâ€™s other markets. Individual state public service commissions Table 4-4. Market Characteristic Measures Variable Measure Source Prior year entry (overall) Market entry in year prior LERG database Prior year entry (origin) Mark et entry in the year prior by firms of the same origin LERG database, NPRG database Count of competitors (overall) Number of firms in the market (overall) LERG database Count of competitors (origin) Number of firms in the market with the same origin LERG database Growth of competitor count (overall) Growth in the number of firms (overall) LERG database Growth of competitor count (origin) Growth in the number of firms with the same origin LERG database
88 Table 4-5. Model and Data set for Hypothesis Testing Hypothesis Model Dataset 1 Correlations and parametric hazard model Option by year by firm 2 Correlations and parametric hazard model Option by year by firm 3 Fixed-effect Poisson model for count data Year by firm 4 Parametric hazard model Option by year by firm 5 Fixed-effect Poisson model for count data Year by firm 6 Parametric hazard model Market by year by firm
89 CHAPTER 5 RESULTS AND DISCUSSION This paper produced some very complex and involved hypotheses and measures. The suggestion is essentially that firms bo rrow best practice from one another and are more likely to do that when they are underp erforming. The definition of practice and performance are complex. Attached to this section ar e nine tables which outline th e results of the analysis. Table 5-1 presents the summary correlations a nd statistics for the data used in analyzing the option behavior. Table 5-2 presents the same information for the information developed at a market level. Table 5-3 pres ents the option purchase data when pooled at the firm by year level and Table 5-4 does th e same for the market entry data. These correlations are not adjusted for the panel nature of the data. The first step in this analysis was to d eal with the extremely high intercorrelations between overall attainment discrepancy a nd attainment discrepancy calculated across groups displayed in Table 5-1 and 5-2 as well as the high correlation between the similarity measures. It is possible that the high correlations ar e the result of erroneous results in one of the groups for one of the years. Figure 51 and 5-2 present a summa ry of option purchase and market entry behavior within groups over this studyâ€™s period. Figure 5-1 presents a simple count of th e number of market competitors in each year for each group. Within the studyâ€™s pe riod, established firms always represented
90 about 65% of the market competitors in the i ndustry. Entrepreneurial firms, while always a smaller group, were a sizable and consistent group of competitors within this industry. Figure 5-2 presents four charts detailing market behaviors within each group. Entrepreneurial firms and established firms are fairly consistent with one another in terms of their option purchase behavi or, their market entries, a nd market exits. Established firms, in this industry, tended to be the fi rms who entered new markets, markets with no prior competitors. This is a unique result to this industry which suggests that the established firms were the firms who invested in new markets while entrepreneurial firms tended to follow behind. This trend revers es itself late in th e study period, with entrepreneurial firms entering twice as many ne w markets. This might be the result of the high levels of venturing financing avai lable in 2001 or the re sult of a regulatory change. The exact behavior deserves atten tion in a later analysis, but for this study entrepreneurial firms and established firms will be treated as having only one aspiration, the industry average. The charts do not suppor t measuring attainment discrepancy within groups, and the correlations suggest that meas uring attainment discrepancy within groups is essentially the same as measuring it across the industry. Because of this consistent high correlati on across groups, the results presented in Tables 5-5 through 5-9 offer re sults which only include the at tainment discrepancy results calculated at the overall indus try. In addition, the high co rrelation between population and business environment similarity meant that these two variables could not be included in the same model; as a result population m easures were not used in the statistical analysis.
91 Because the attainment discrepancy and si milarity measures within groups are so highly correlated with the measures at the i ndustry level, and because summary analysis does not indicate a difference within origin groups, the analysis results do not support Hypothesis 1 and 2. There is no de tectable difference within groups. Table 5-5 presents the results which examine the tendency for firms to purchase options which are similar to those owned by the average firm in the industry. These models were conducted using a hazard with robust standard errors which are corrected for potential clustering within ea ch firm. In this model, th e relevant hazard spell was a market-firm pair. Thus, there are multiple spel ls for each firm within the dataset. The clustering option in Stata was used to adjust for the relationship between these spells. The tables display exponentiated coefficients not hazard ratios. Negative coefficients in this model suggest a lower likelihood of entry. Model 1 presents the control model and M odel 2 presents the model after including the attainment discrepancy variable. This model suggests that firms are less likely to purchase options when their atta inment discrepancy is positive. In other words, firms do not seem to purchase more options when they are performing below the industry average. Instead, they are less likely to purchase options if they are performing below the average. They also tend to purchase more options wh en they are performi ng above the industry average. Model 3 incorporat es the dissimilarity measures , and its coefficient suggests that firms are less likely to purchase options on markets which are geographically distant and operate under a different regulatory scheme. In an effort to identify any relevant di fference in option purchase behavior, Model 4 through 7 present the interaction of the dissimilarity measures with firm origin. Model 4
92 suggests that entrepreneurial firms are less likely than established firms to purchase options which are geographically distant. Mo del 5 suggests that these two types of firms show similar option purchase tendencies with respect to the business environment. Model 7 suggests that entrepreneurial firms tend to seek markets with a large in competitor size. Entrepreneurial firms tend to purchase options on markets populated by different sized competitors. This model also suggests that established firms purchase options on markets with similar sized competito rs. Model 7 suggests that entrepreneurial firms show a slightly higher te ndency to avoid markets with different regulation schemes than current markets. This analysis suggests that attainment discrepancy has an influence on option purchase. However, it is not entirely consis tent with the hypothesized results. While the hypotheses suggested that similarity and orig in have an important influence on option purchase, the results in Table 5-5 have suggested that simila rity in terms of geographic proximity and competitor characteristics influence option purchase in a direction consistent with theory. Regulation and busine ss environment similarity did not behave in a manner consistent with the theorized model. Similarity differences had no different influences within origin groups. Attainme nt discrepancy had an influence on option purchase but one that consistently decreased option purchase the further behind a firm became. In this model, the lower a firm â€™s performance the lower its likelihood of purchase options became. To further examine option purchase, Table 5-6 presents the results of regressing attainment discrepanc y on the total number of options purchased. The model to test Hypothesis 3 is shown in Table 5-6. This is a Poisson regression; a positive coefficient indicates a tendency to purchase more options. Model 1 presents
93 the control model. The variab les used for control in this analysis are measured at the firm-year level as opposed to the firm-marketyear level in the prior analysis. Model 2 suggests that increasing attainment discrepanc y reduces the number of options purchased. If attainment discrepancy is squared, the number of options purchased also declines. In short, Table 5-6 suggests contrary support for Hypothesis 3. Firms seem to purchase fewer options overall as their performance falls below the industry average and the tendency to purchase options also declines in an inverted-U pa ttern as attainment discrepancy increases in absolute terms. Th ese results are the same in their implication as the results presented in Table 5-5 which suggested outperforming the industry lead to higher option purchase. This analysis, conduc ted a higher level, suggests the same. In short, these results suggest an overall relationship between attainment discrepancy and option purchase, but the overall trend is not in support of Hypothesis 3. Hypothesis 4 suggested that different levels of attainment discrepancy wi ll lead to the purchase of different kinds of options. Hypothesis 4 suggested that as the absolu te value of focal firmâ€™s attainment discrepancy increases, it is more likely to purchase options which are increasingly dissimilar to the options it already owns. Ta ble 5-7 presents the results of interacting squared attainment discrepancy with dissimilarity. The squared values in this model are mean-centered prior to being squared to elimin ate collinearity with th e main effect. The results in Table 5-7 do not support this hypothesi s. While there are main effects for the dissimilarity measures, the interaction of a ttainment discrepancy and these variables is not different than zero. Firm option purchas e behavior does not seem to change with changes in attainment discrepancy. The square d value of attainment discrepancy in these
94 models had a negative influence on the like lihood of option purchase as does the main effect. In summary, while Table 5-5 and 5-6 su ggested that attainment discrepancy might have a differential impact on option purcha se depending on the option, the results in Table 5-7 was not able to identify if these effects indeed exist. Table 5-8 shows the investig ation of Hypothesis 5. This hypothesis suggested that the tendency to enter a market was the result of a positive influence of attainment discrepancy. If the firm is underperformi ng the industry, Hypothesis 5 suggested that the firm would be more likely to enter the market. Similarly, if the firm is outperforming the market, Hypothesis 5 suggested that it would be less likely to enter a new market. The results in Table 5-8 support this hypothesis. Attainment discrepancy had a consistent influence on the likelihood of market entry. As attainment discrepancy increased, market entry became more likely. Model 3 examines this result further and fi nds that high levels of squared attainment discrepancy have a small negative effect on market entry. The strongest result in this model, however, is the positive linear result. Firms who are in a weaker market position (high attainment discrepancy) are more likely to enter markets than firms who are in a better market pos ition. Figure 5-3 charts this relationship between the number of markets entered and a ttainment discrepancy over the range of the data in the sample. Hypothesis 6 suggested that attainment discrepancy would increase the likelihood of executing options which are similar to thos e held by other firms. Table 5-9 presents the examination of this hypothesis using a hazard model. This hypothesis was not supported. Table 5-9 suggests that attainment discrepancy does not hold as a relationship with market entry on any given market through the entire sa mple period. Although
95 market dissimilarity discourages entry, the interaction of the dissimilarity measures with attainment discrepancy did not influence the lik elihood of entry in thei r empirical model. Firms might be less likely to enter markets in the early stages of th eir ownership if they are far behind their competitors. While th e overall number of ma rkets entered can be predicted by attainment discre pancy, the individual market s entered cannot be predicted consistently. The similarity of the potential market with the markets of other firms in the industry did not influence the firmâ€™s tendency to enter. As a result, firms did not display an increased tendency to ente r the markets mimetically in this industry based on their attainment discrepancy. This analysis was also conducted on a limited sample, using only the first two years of data af ter option purchase. These resu lts are presented in Appendix A. Discussion and Conclusion The intent of this study was to show that option behaviors are influenced by a ttainment discrepancy. The results are summarized in Table 5-10. While the study suggests some intere sting potential relationships and finds significance at a level greater than that expe cted based on chance alone, it does not fully support the model. While there are several pot ential causes for this, the results do present some interesting findings which can be generalized out to the options and entrepreneurship literature. This section will first discuss some of the contextual reason that the results were not more robust and then move on to some of the im plications for the findings that were significant. This paper has approached a complex and dynamic industry. Any time such a complex industry is modeled empirically, there is the possibility that its complexity will be lost. This study employed over 70 variab les and reduced these through statistical
96 techniques to a tractable 25 variables. While no solution is ever complete, this was a first best attempt to model the industry. As an initial problem, the analysis is extremely focused on the product market. The options as well as the execution event are only measured on a product market level. Measurement was conducted at a higher leve l because of the theoretical importance of measuring items on a level which is visibl e to all firms in the industry, but there may have been different items which could have been employed. For instance, the option acquisition event, here measured as the cer tification by state regulatory agencies, would have been better measured by looking at an investigation event conducted prior to the state certification. For instan ce, firms truly investigate ma rket possibilities by asking consumers what they want or hiring local attorn eys to assist in the market entry process. These are both options on the market, and th ese events, which are prior to OCN purchase, might have better supported the theory. Figure 4-2 suggests that firms did indeed us e state certific ation as an option. While nearly 700 entry events occurred within a ye ar of state certifi cation, there were many events which occurred after se veral years, a finding consis tent with option behavior. While these 700 events occurred in 171 markets, there were an additional 92 markets for which entry was not observed immediately follo wing certification. While all but three of the potential markets in this study did under go entry at some point in the study period, this study might be expanded to model the delay that firms used after attaining certification. Secondly, the attainment discrepancy vari able, which was inconsistent in its support for the paperâ€™s theoretical propositions, could have been measured differently.
97 This paper has operationalized attainment discre pancy as a combination of four variables. A firmâ€™s shortcomings on either one of th ese contributes to increasing attainment discrepancy. However, measuring the constr uct this way has some problems. First, by treating attainment discrepancy as multidimensional, this paper weighted all of the factors uniformly across the industry. Firms who are below the industry average in terms of the number of options owned are in cluded with firms who are be low the industry average in terms of the number of markets they have entered. Because attainment discrepancy is inherently determined within the firm, the we ights for these two factors may differ across firms. This study measured attainment discre pancy in this way to better handle the two stage nature of the game and to avoid us ing an industry average as an independent variable. If this study had us ed an industry average as a in dependent variable, any result might have been explained as regression to the mean. Alterna tive definitions of attainment discrepancy and the empirical resu lts for those definitions are developed in Appendix B. The second problem with attainment disc repancy measurement, suggested in the results, is also one of th e studyâ€™s contributions. Attain ment discrepancy measurement might not only differ between firms but also between decisions. Prior studies have motivated the market entry decision using as piration levels (Greve, 1998b). Others have shown that aspirations can st retch across two decisions such as those modeled here (Greve, 2003a), but none have looked at the use of aspiration levels in highly dynamic environments. How firms use their reference gr oups to decide on their market actions is the intent of this study. However, differe nt groups might have different impacts on decision making. Attainment discrepancy in this study, depending on how it was defined,
98 had different influences on the firmâ€™s d ecision process. Table 5-9 and Appendix A suggest that firms do enter markets which are populated by firms of similar origin; they do tend to follow one another. However, it is not clear how they define their aspirations to justify this outcome. This suggests a potential contribution to the real options and entrepreneurship literature. If firms purchase options to l earn about an environmen t and strengthen their market position, they do so in a manner that is as different from their competitors as possible. They explore markets with the inte nt of differentiating. Firms in the reference group influence the tendency to seek out opportuniti es to learn, but not in predictable ways. However, once these learning opportun ities are explored, firms seek to exploit options which will get them to the forefront of their industry. They enter markets which are popular with competitors a nd particularly with competitors with the same origin. They compete within groups with the ultimate goal of being the industry leader. They enter markets which are very similar to the average market for the industry as a whole and enter more markets when their market posit ion is behind that of competitors. While the lowliest market competitor mi ght keep tabs on its nearest ri val, it still yearns to be the big firm. For the real options literature, this suggest s that firms might be using real options logic to expand their potential to learn. Howe ver, once they have finished the search process, the most highly valued options are those options which are also held by competitors. In addition, this model did suggest that firms limit their tendency to enter new markets if they are performing below the industry average. Their tendency to
99 engage in search is limit by high levels of attainment discrepancy, but discrepancy increases their likelihood of market entry. For the entrepreneurship literature, th is has important suggestions. First, entrepreneurs as a group are an important subg roup of an industry. While there was not overwhelming support for the notion that entrep reneurial firms do more of one behavior than another, there was support that origin de fines groups within this industry. Table 5-9 suggests that firms show different tendencies to enter markets depending on what kind of firms were already in the market. The entrepreneurship literature, which tends to emphasize the importance of entrepreneurial firms creating new and better alternatives, might benefit from this study and others like it, which suggests that entr epreneurial firms try new things because it benefits their business, but ultimately entrepreneurial firms are trying to maintain competitive parity with a subgroup. If this subgroup stampedes down the wrong path and invests in a dying industry, th e cognitive groupings which these firms have adopted will be the cause of their failure. Ultimately, th is is largely what happe ned in this industry; the rush to invest led to a glut of capacity and customer exhaustion. Billions of capital dollars were lost and it was the startup firms who were leading the charge.
100 0 20 40 60 80 100 199719981999200020012002200320042005 YearNumber of Firms Established Entrepreneurial Figure 5-1. Market Compet itor Count by Origin. 0 50 100 150 200 25019 96 19 97 19 98 19 99 20 00 2001 20 02 20 03 20 04 20 05YearOptions Purchased Established Entrepreneurial A 0 50 100 150 20019 96 19 97 19 98 19 99 20 00 2001 20 02 20 03 20 04 20 05YearMarkets Entered Established Entrepreneurial B 0 5 10 15 20 25 30 35 19971998199920002001 YearNumber of Markets Established Entrepreneurial C 0 20 40 60 80 100 120 140 160 18019 96 1997 1 998 1999 2000 2001 200 2 2003 200 4 2005YearMarkets Exited Established Entrepreneurial D Figure 5-2. Entry and Exit Graphs by Origin. A) The number of options bought by origin B) The number of market entries by origin C) Entry into new markets by origin D) Market exit by origin
101 -1 -0.5 0 0.5 1 1.5 2 2.5 3 -50510 Markets EnteredAttainment Discrepancy Figure 5-3. Relationship Between Attainme nt Discrepancy and Markets Entered. Although there is a signif icant squared term in the relationship between attainment discrepancy and the number of market entries, this relationship has no visible effect over the range of the data in the sample.
102Table 5-1. Option data summary statistics Variable Number Variable Mean Std. Dev. 1 2 3 4 5 6 1 Distance dissimilarity (overall) 0.00 1.71 2 Population dissimilarity (overall) 0.00 1.00 0.00 3 Business environment dissimilarity (overall) 0.00 1.00 0.06* 0.00 4 Competitor dissimilarity (overall) 0.00 1.57 -0.03* -0.01 -0.02* 5 Regulation dissimilarity (overall) 0.00 1.63 -0.04* 0.12* 0.18* 0.10* 6 Distance dissimilarity (origin) 0.00 1.71 0.94* 0.01 0.07* -0.05* -0.04* 7 Population dissimilarity (origin) 0.00 1.67 0.00 1.00* 0.00 -0.01 0.12* 0.01 8 Business environment dissimilarity (origin) 0.00 1.64 0.05* 0.62* 0.79* -0.02* 0.22* 0.06* 9 Competitor dissimilarity (origin) 0.00 1.60 -0.09* -0.01 -0.02* 0.76* 0.05* -0.12* 10 Regulation dissimilarity (origin) 0.00 1.62 -0.04* 0.12* 0.18* 0.15* 1.00* -0.04* 11 Population dissimilarity (focal) 0.03 1.69 0.00 1.00* -0.01* 0.00 0.11* 0.01 12 Business environment dissimilarity (focal) 0.04 1.65 0.05* 0.61* 0.78* -0.01 0.21* 0.06* 13 Regulation dissimilarity (focal) 0.05 1.64 -0.04* 0.11* 0.16* 0.12* 0.98* -0.04* 14 Distance dissimilarity (focal) 0.06 1.69 0.50* 0.12* 0.19* -0.09* 0.10* 0.56* 15 Competitor dissimilarity (focal) 0.06 1.54 -0.02* -0.01 -0.02* 0.87* 0.05* -0.03* 16 Attainment discrepancy (origin) 0.00 1.92 0.00 0.04* 0.03* 0.16* 0.00 0.00 17 Attainment discrepancy (overall) 0.00 1.92 0.00 0.04* 0.02* 0.16* 0.00 0.00 20 Firm age 8.03 15.99 0.00 -0.03* -0.01 0.05* 0.02* 0.00 21 Total option purchased 1.79 4.26 0.00 -0.05* -0.01* -0.06* -0.04* 0.00 22 Market size 0.00 1.92 -0.10* 0.22* 0.23* -0.16* -0.08* -0.09* 23 IPOs in market 5.60 14.41 0.01 0.07* 0.25* -0.09* -0.06* 0.02* 24 IPO capital 915.41 2086.77 -0.05* 0.08* 0.21* -0.12* -0.05* -0.04* 25 Variance in market dimensions 0.01 1.86 0.13* 0.05* 0.38* -0.11* -0.06* 0.16* 26 Growth in market variances 0.00 1.22 0.06* 0.02* 0.08* -0.10* -0.08* 0.09* 27 Origin 0.37 0.48 0.00 0.03* 0.10* -0.04* -0.02* 0.02* * p < 0.05
103Table 5-1. Continued Variable Number 7 8 9 10 11 12 13 14 15 16 17 20 21 8 0.61* 9 0.00 -0.02* 10 0.12* 0.22* 0.05* 11 0.99* 0.61* 0.00 0.11* 12 0.61* 0.99* -0.01* 0.21* 0.61* 13 0.11* 0.20* 0.07* 0.98* 0.12* 0.21* 14 0.12* 0.22* -0.12* 0.10* 0.13* 0.23* 0.11* 15 -0.01 -0.02* 0.68* 0.05* 0.01* 0.01* 0.08* 0.01* 16 0.04* 0.05* 0.15* 0.00 0.04* 0.05* 0.00 -0.14* 0.11* 17 0.04* 0.04* 0.15* 0.00 0.04* 0.04* 0.00 -0.15* 0.11* 1.00* 20 -0.03* -0.03* 0.04* 0.01* -0.03* -0.02* 0.02* 0.04* 0.11* -0.20* -0.20* 21 -0.05* -0.04* -0.05* -0.04* -0.05* -0.04* -0.04* -0.03* -0.13* -0.07* -0.07* 0.01* 22 0.22* 0.32* -0.12* -0.08* 0.22* 0.32* -0.07* -0.04* -0.12* 0.00 0.00 0.00 0.00 23 0.07* 0.24* -0.08* -0.06* 0.06* 0.23* -0.07* -0.01* -0.15* 0.00 0.00 -0.02* 0.01* 24 0.08* 0.21* -0.09* -0.05* 0.08* 0.21* -0.05* -0.01 -0.08* 0.00 0.00 0.00 -0.01 25 0.05* 0.33* -0.09* -0.06* 0.05* 0.34* -0.06* 0.10* -0.06* 0.00 0.00 0.01 -0.01* 26 0.02* 0.07* -0.06* -0.08* 0.02* 0.08* -0.08* 0.12* 0.14* -0.01* -0.01* 0.08* -0.11* 27 0.01 0.10* -0.08* -0.01 0.03* 0.10* -0.02* 0.05* -0.03* 0.00 -0.03* -0.13* 0.02* * p < 0.05
104Table 5-1. Continued Variable Number 22 23 24 25 26 23 0.65* 24 0.64* 0.77* 25 0.72* 0.67* 0.64* 26 0.15* 0.08* 0.21* 0.27* 27 0.00 0.00 0.00 0.00 -0.01 * p < 0.05
105Table 5-2. Market data summary statistics Variable Number Variables Mean Std. Dev. 1 2 3 4 5 6 1 Attainment discrepancy (overall) -1.59 2.70 2 Attainment discrepancy (origin) -1.59 2.71 1.00* 3 Distance dissimilarity (origin) -0.09 1.04 -0.02* -0.02* 4 Population dissimilarity (origin) 0.15 1.49 -0.01 0.00 0.02* 5 Business environment dissimilarity (origin) 0.15 1.44 0.00 0.01 0.11* 0.68* 6 Competitor dissimilarity (origin) -0.01 1.76 0.08* 0.08* -0.07* 0.00 -0.01* 7 Distance dissimilarity (overall) -0.09 1.03 -0.02* -0.02* 0.99* 0.02* 0.10* -0.06* 8 Population dissimilarity (overall) 0.17 1.49 -0.01 0.00 0.02* 0.99* 0.67* 0.00 9 Business environment dissimilarity (overall) 0.16 1.44 0.00 0.01* 0.11* 0.67* 0.99* -0.01* 10 Competitor dissimilarity (overall) -0.06 1. 65 0.10* 0.10* -0.04* 0.00 -0.01 0.57* 11 Population dissimilarity (focal) 0.21 1.47 -0.04* -0.03* 0.02* 0.95* 0.64* 0.01* 12 Business environment dissimilarity (focal) 0.19 1.43 -0.02* -0.02* 0.10* 0.63* 0.97* -0.01* 13 Distance dissimilarity (focal) -0.22 1.09 -0.25* -0.25* 0.43* 0.26* 0.34* -0.05* 14 Competitor dissimilarity (focal) 0.13 1.61 0.02* 0.02* 0.00 0.00 0.00 0.47* 15 Number of firms in market (overall) 5.05 5.64 0.07* 0.07* 0.02* 0.06* 0.08* -0.08* 16 Number of firms in market (origin) 2.59 3.06 0.07* 0.07* 0.02* 0.06* 0.07* -0.05* 17 Growth in number of competitors (overall) -0.07 0.32 -0.02* -0.02* 0.03* 0.01* 0.02* -0.04* 18 Growth in number of competitors (origin) -0. 04 0.27 -0.02* -0.02* 0.03* 0.00 0.01* -0.13* 19 Entry in prior year (origin) 0.24 1.10 0.01* 0.01* -0.03* -0.01* -0.02* 0.12* 20 Entry in prior year (overall) -4.35 5.51 -0.06* -0.06* -0.04* -0.07* -0.09* 0.09* 21 Firm age 9.41 16.49 -0.26* -0.26* 0.03* 0.05* 0.02* -0.04* 22 Market growth -0.05 1.87 0.03* 0.03* 0.07* -0.02* 0.00 -0.03* 23 Market size -0.01 1.83 0.08* 0.08* 0.05* 0.12* 0.26* -0.10* 24 Market variance 0.00 1.59 0.04* 0.04* 0.09* 0.06* 0.20* -0.06* 25 Market variance growth 0.02 2.28 0.01* 0.01* 0.02* -0.01 0.00 -0.04* * p < 0.05
106Table 5-2. Continued Variable number 7 8 9 10 11 12 13 14 15 16 17 18 19 8 0.02* 9 0.10* 0.67* 10 -0.04* 0.00 -0.01 11 0.02* 0.97* 0.65* 0.01* 12 0.10* 0.64* 0.98* 0.00 0.67* 13 0.43* 0.26* 0.34* -0.03* 0.27* 0.34* 14 0.00 0.00 0.00 0.84* 0.09* 0.05* 0.04* 15 0.02* 0.06* 0.08* -0.20* 0.12* 0.12* 0.05* -0.04* 16 0.03* 0.05* 0.07* -0.18* 0.11* 0.11* 0.05* -0.03* 0.94* 17 0.04* 0.01* 0.02* -0.10* 0.02* 0.01* 0.05* 0.05* -0.12* -0.12* 18 0.03* 0.00 0.01* -0.07* 0.01* 0.01 0.04* 0.05* -0.12* -0.16* 0.62* 19 -0.03* 0.00 -0.01* 0.07* -0.01* -0.01 -0.04* -0.07* 0.24* 0.29* -0.51* -0.78* 20 -0.04* -0.07* -0.09* 0.22* -0.14* -0.13* -0.07* -0.01* -0.94* -0.88* -0.11* -0.09* 0.04* 21 0.03* 0.05* 0.02* -0.03* 0.05* 0.02* 0.06* 0.00 -0.08* -0.06* 0.05* 0.04* -0.04* 22 0.08* -0.02* 0.00 -0.03* -0.03* -0.01 0.04* -0.04* 0.03* 0.03* 0.00 -0.01 0.00 23 0.05* 0.12* 0.27* -0.14* 0.12* 0.26* 0.03* -0.12* 0.54* 0.51* -0.03* -0.03* 0.09* 24 0.09* 0.06* 0.20* -0.08* 0.06* 0.20* 0.05* -0.06* 0.34* 0.32* 0.00 -0.01 0.03* 25 0.02* -0.01 0.00 -0.04* 0.00 0.00 0.02* -0.01* 0.10* 0.09* 0.02* 0.02* -0.04* * p < 0.05 Table 5-2. Continued Variable number 20 21 22 23 24 21 0.06* 22 -0.03* 0.00 23 -0.51* -0.03* 0.01* 24 -0.34* -0.01* 0.01 0.66* 25 -0.12* 0.00 0.00 0.03* 0.64* * p < 0.05
107Table 5-3. Option by year data summary statistics Variable number Variables Mean Std. Dev. 1 2 3 4 5 6 1 Total options purchased 1.79 4.26 2 Firm age 7.86 16.05 0.01 3 Attainment discrepancy (overall) 0.00 1.88 0.06 0.19* 4 Attainment discrepancy (Origin) 0.00 1.88 0.06 0.20* 1.00* 5 Market size 0.00 1.63 -0.04 -0.02 -0.01 -0.01 6 Market variances 0.00 1.88 -0.04 0.00 0.07* 0.06 0.71* 7 Market variance growth 0.00 1.42 -0.16* 0.17* 0.13* 0.13* 0.20* 0.36* 8 Market growth 0.00 1.48 0.03 0.06 0.12* 0.12* 0.31* 0.27* 9 Absolute value of attainment discrepancy (overall) 1.37 1.29 -0.01 0.32* 0.67* 0.68* -0.08* 0.01 10 Absolute value of attainment discrepancy (origin) 1.37 1.29 -0.01 0.32* 0.67* 0.68* -0.08* 0.01 11 Squared attainment discrepancy (overall) 3.55 7.31 -0.02 0.35* 0.76* 0.77* -0.07* 0.01 12 Squared attainment discrepancy (origin) 3.55 7.41 -0.02 0.35* 0.76* 0.76* -0.06* 0.01 13 Origin 0.37 0.48 0.02 -0.12* 0.00 -0.02 0.12* 0.18* * p < 0.05 Table 5-3. Continued Variable Number 7 8 9 10 11 12 8 0.17* 9 0.24* 0.00 10 0.25* 0.01 0.99* 11 0.19* 0.02 0.93* 0.93* 12 0.19* 0.02 0.93* 0.93* 1.00* 13 0.03 0.09* -0.05 -0.05 -0.07* -0.07* * p < 0.05
108Table 5-4. Market by year data summary statistics Variable Number Variable Name Mean Std. Dev. 1 2 3 4 5 1 Markets entered 11.84 27.27 2 Firm age 7.00 15.58 0.38* 3 Attainment discrepancy (overall) 0.00 1.90 0.74* 0.22* 4 Attainment discrepancy (origin) 0.00 1.90 0.74* 0.22* 1.00* 5 Market size 0.00 1.63 0.02 0.02 0.12* 0.12* 6 Market variances 0.00 1.73 0.01 0.01 0.13* 0.12* 0.86* 7 Market variance growth 0.00 1.99 0.06 0.03 0.10* 0.09* 0.04 8 Market growth 0.00 1.81 0.10* 0.04 0.15* 0.15* 0.27* 9 Absolute value of attainment discrepancy (overall) 1.23 1.45 0.73* 0.34* 0.73* 0.73* 0.10* 10 Absolute value of attainment discrepancy (origin) 1.23 1.45 0.73* 0.34* 0.73* 0.74* 0.11* 11 Squared attainment discrepancy (overall) 3.62 9.56 0.83* 0.34* 0.82* 0.82* 0.05 12 Squared attainment discrepancy (origin) 3.62 9.66 0.83* 0.35* 0.81* 0.82* 0.05 13 Origin 0.39 0.49 -0.01 -0.13* -0.01 -0.02 0.13* * p < 0.05 Table 5-4. Continued Variable Number 6 7 8 9 10 11 12 7 0.27* 8 0.29* 0.03 9 0.08* 0.14* 0.12* 10 0.09* 0.15* 0.13* 1.00* 11 0.04 0.09* 0.09* 0.93* 0.93* 12 0.05 0.09* 0.09* 0.92* 0.93* 1.00* 13 0.15* 0.07* 0.07* -0.05 -0.04 -0.06 -0.06 * p < 0.05
109Table 5-5. Maximum likelihood estimate s of the likelihood of option purchase Variables Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Distance dissimilarity (focal) * origin -0.04 (0.09) Business environment dissimilarity (focal)* origin -0.01 (0.07) Competitor dissimilarity (focal)* origin 0.09 (0.08) Regulation dissimilarity (focal) * origin -0.01 (0.07) Distance dissimilarity (focal) -0.06 (0.04) -0.06 (0.04) -0.06 (0.04) -0.06 (0.04) -0.04 (0.05) Business environment dissimilarity (focal) 0.11** (0.03) 0.11** (0.03) 0.11** (0.03) 0.12** (0.04) 0.12** (0.03) Competitor dissimilarity (focal) -0.12** (0.04) -0.12** (0.04) -0.16** (0.06) -0.12** (0.04) -0.12** (0.04) Regulation dissimilarity (focal) 0.06** (0.02) 0.06 (0.04) 0.06** (0.02) 0.06** (0.02) 0.06** (0.02) Attainment discrepancy (overall) -0.10** (0.01) -0.10** (0.01) -0.10** (0.01) -0.10** (0.01) -0.10** (0.01) -0.10** (0.01) Firm age 0.01 (0.01) 0.00 (0.01) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) Total options purchased 0.09** (0.01) 0.09** (0.01) 0.10** (0.01) 0.10** (0.01) 0.10** (0.01) 0.10** (0.01) 0.10** (0.01) Market size 0.11** (0.01) 0.13** (0.02) 0.11** (0.02) 0.11** (0.02) 0.11** (0.02) 0.11** (0.02) 0.11** (0.02) Market growth -0.01 (0.05) -0.01 (0.05) -0.03 (0.04) -0.03 (0.04) -0.03 (0.04) -0.03 (0.04) -0.03 (0.04) IPOs in market -0.00 (0.00) -0.00 (0.00) -0.00 (0.00) -0.00 (0.00) -0.00 (0.00) -0.00 (0.00) -0.00 (0.00) IPO capital 0.00 (0.00) 0.00 (0.00) 0.00* (0.00) 0.00* (0.00) 0.00* (0.00) 0.00* (0.00) 0.00* (0.00) p<.10; * p<.05; ** p<.01. Robust standard errors are in parentheses. Z-tests are two-tailed, hypothesized effects are one-tai led. N=36,348.
110Table 5-5. Continued Variables Model 1 Model 2 Model 3 Model 5 Model 6 Model 7 Model 8 Market variances -0.00 (0.02) 0.01 (0.02) -0.00 (0.02) -0.00 (0.02) -0.00 (0.02) -0.00 (0.02) -0.00 (0.02) Market variance growth -0.28** (0.08) -0.32** (0.09) -0.29** (0.09) -0.29** (0.09) -0.29** (0.09) -0.29** (0.09) -0.29** (0.08) Origin 0.12 (0.23) 0.09 (0.20) 0.10 (0.19) 0.10 (0.18) 0.14 (0.19) 0.10 (0.19) 0.08 (0.17) Constant -4.13** (0.15) -4.11** (0.13) -4.15** (0.13) -4.15** (0.13) -4.17** (0.13) -4.15** (0.13) -4.15** (0.12) Log pseudolikelihood -4107.87 -3648.36 -3105.12 -3132.82 -3131.77 -3132.85 -3130.20 Model 2 195.43** 950.99** 1189.35** 1077.21** 1076.35** 1072.32** 1062.79** p<.10; * p<.05; ** p<.01. Robust standard errors are in parentheses. Z-tests ar e two-tailed, hypothesize d effects are one-ta iled. N=36,348.
111Table 5-6. Poisson model of option acquisition Variables Model 1 Model 2 Model 3 Squared attainment discrepancy (overall) -0.02* (0.01) Attainment discrepancy (overall) -0.39** (0.02) -0.30** (0.04) Firm age -0.06** (0.02) 0.13** (0.02) 0.13** (0.02) Market variances -0.07 (0.05) -0.00 (0.05) -0.00 (0.05) Market growth 0.09** (0.02) 0.16** (0.02) 0.16** (0.02) Market variance growth -0.58** (0.04) -0.60** (0.04) -0.61** (0.04) Market size 0.26** (0.05) 0.17** (0.05) 0.16** (0.05) Log pseudolikelihood -1470.67 -1294.87 -1291.81 Model 2 431.11**726.77**708.37** p<.10; * p<.05; ** p<.01; *** p<.001. R obust standard errors are in parenthe ses. Z-tests are two-tailed. N=893.
112Table 5-7. Maximum likelihood estimates of the lik elihood of option purchase based on dissimilarity Variable Model 1 Model 2 Mode l 3 Model 4 Model 5 Model 6 Squared attainment discrepancy (overall) * competitor dissimilarity (focal) 0.00 (0.00) Squared attainment discrepancy (overall) * distance dissimilarity (focal) 0.00 (0.00) Squared attainment discrepancy (overall) * regulation dissimilarity (focal) -0.00 (0.00) Squared attainment discrepancy (overall) * business environment dissimilarity (focal) -0.00** (0.00) Distance dissimilarity (focal) -0.08* (0.04) -0.08* (0.04) -0.08* (0.04) -0.09* (0.04) -0.08* (0.04) Business environment dissimilarity (focal) 0.12** (0.03) 0.15** (0.03) 0.12** (0.03) 0.12** (0.03) 0.12** (0.03) Competitor dissimilarity (focal) -0.04 (0.03) -0.03 (0.03) -0.04 (0.03) -0.04 (0.03) -0.05 (0.03) Regulation dissimilarity (focal) 0.04 (0.02) 0.04 (0.02) 0.05* (0.02) 0.04 (0.02) 0.04 (0.02) Squared attainment discrepancy (overall) -0.00** (0.00) -0.01** (0.00) -0.00** (0.00) -0.00** (0.00) -0.00** (0.00) Attainment discrepancy (overall) -0.16** (0.02) -0.17** (0.02) -0.17** (0.02) -0.17** (0.02) -0.17** (0.02) Firm age 0.01 (0.01) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) 0.00 (0.00) Total options purchased 0.09** (0.01) 0.10** (0.01) 0.10** (0.01) 0.10** (0.01) 0.10** (0.01) 0.10** (0.01) Market size 0.11** (0.01) 0.11** (0.02) 0.12** (0.02) 0.11** (0.02) 0.11** (0.02) 0.11** (0.02) Market growth -0.01 (0.05) -0.04 (0.04) -0.04 (0.04) -0.04 (0.04) -0.04 (0.04) -0.04 (0.04) IPOs in Market -0.00 (0.00) -0.00 (0.00) -0.01** (0.00) -0.00 (0.00) -0.00 (0.00) -0.00* (0.00) IPO Capital 0.00 (0.00) 0.00** (0.00) 0.00** (0.00) 0.00** (0.00) 0.00** (0.00) 0.00** (0.00) p<.10; * p<.05; ** p<.01; *** p<.001. Robust standard errors are in pa rentheses; z-tests ar e two-tailed. N = 39,228.
113Table 5-7. Continued Variable Model 1 Model 2 Mode l 3 Model 4 Model 5 Model 6 Market variances -0.00 (0.02) -0.00 (0.02) 0.01 (0.02) -0.00 (0.02) -0.00 (0.02) -0.00 (0.02) Market variance growth -0.28** (0.08) -0.26** (0.08) -0.23** (0.08) -0.26** (0.08) -0.26** (0.08) -0.25** (0.08) Origin 0.12 (0.23) 0.10 (0.18) 0.10 (0.18) 0.10 (0.18) 0.10 (0.18) 0.10 (0.18) Constant -4.13** (0.15) -4.02** (0.13) -3.98** (0.13) -4.02** (0.13) -4.02** (0.13) -4.01** (0.13) Log pseudolikelihood -4138.75 -3778.73 -3761.92 -3778.23 -3778.31 -3777.58 Model 2 202.71** 844.45** 849.87** 874.51** 858.59** 892.58** p<.10; * p<.05; ** p<.01; *** p<.001. Robust standard errors are in parentheses. Z-tests are two-tailed; hypothesized effects are onetailed. N = 39,228.
114 Table 5-8. Poisson estimation of the tendency to enter markets Variable Model 1 Model 2 Model 3 Squared attainment discrepancy (overall) -0.01** (0.00) Attainment discrepancy (overall) 0.13** (0.01) 0.21** (0.02) Firm age 0.20** (0.00) 0.14** (0.01) 0.15** (0.01) Market variances 0.06* (0.03) -0.01 (0.03) -0.03 (0.03) Market growth 0.02** (0.01) 0.02* (0.01) 0.01 (0.01) Market variance growth -0.06** (0.01) -0.04** (0.01) -0.04** (0.01) Market size 0.03 (0.02) 0.12** (0.02) 0.12** (0.02) Log pseudolikelihood -2304.63 -2130.06 -2118.99 Model 2 2068.50** 2144.40** 2127.80** p<.10; * p<.05; ** p<.01. Robust sta ndard errors are in parentheses. Z-tests are two-tailed. N=884.
115 Table 5-9. Maximum likelihood estimate s of the likelihood of market entry Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Competitor dissimilarity (overall)* attainment discrepancy (overall) -0.00 (0.02) Business environment dissimilarity (overall) * attainment discrepancy (overall) -0.03 (0.03) Distance dissimilarity (overall) * attainment discrepancy (overall) -0.02 (0.01) Attainment discrepancy (overall) -0.02 (0.07) -0.02 (0.07) -0.02 (0.08) -0.02 (0.08) Distance dissimilarity (overall) 0.05 (0.05) -0.06 (0.10) 0.05 (0.05) 0.05 (0.05) Business environment dissimilarity (overall) -0.19** (0.07) -0.19** (0.07) -0.21** (0.07) -0.19** (0.07) Competitor dissimilarity (overall) -0.13** (0.04) -0.13** (0.04) -0.13** (0.04) -0.13* (0.05) Count of competitors (overall) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) Count of competitors (origin) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) Growth in competitors (overall) -1.03** (0.33) -1.10** (0.42) -1.09** (0.42) -1.10** (0.42) -1.10** (0.42) Growth in competitors (origin) -1.27** (0.40) -1.54** (0.55) -1.54** (0.55) -1.53** (0.54) -1.54** (0.55) Prior year entry (overall) 0.17* (0.08) 0.16* (0.07) 0.16* (0.07) 0.16* (0.07) 0.16* (0.07) Prior year entry (origin) 0.08 (0.04) 0.11* (0.04) 0.11* (0.04) 0.11* (0.04) 0.11* (0.04) Firm age 0.01** (0.00) -0.02 (0.03) -0.02 (0.02) -0.02 (0.03) -0.02 (0.03) Market growth 0.01 (0.01) -0.02 (0.05) -0.02 (0.04) -0.02 (0.05) -0.02 (0.05) p<.10; * p<.05; ** p<.01. Robust standard errors are in parentheses; z-tests are two-tailed, hypothesized effects are one-ta iled. N=44,087.
116Table 5-9. Continued Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Market size 0.13** (0.02) 0.17** (0.03) 0.17** (0.04) 0.17** (0.04) 0.17** (0.04) Market variances -0.11** (0.04) -0.11** (0.03) -0.10** (0.04) -0.11** (0.03) -0.11** (0.04) Market variance growth -2.99 (1.76) -2.65 (1.93) -2.49 (1.89) -2.60 (1.90) -2.65 (1.93) Origin 0.09 (0.30) 0.24 (0.35) 0.24 (0.34) 0.24 (0.34) 0.24 (0.34) Constant -5.96** (0.29) -5.82** (0.39) -5.82** (0.38) -5.83** (0.40) -5.81** (0.39) Log pseudolikelihood -2245.98 -1539.98 -1537.45 -1538.18 -1539.96 Model 2 528.28** 712.01** 807.05** 719.10** 809.08** p<.10; * p<.05; ** p<.01. Robust standard errors are in parentheses; z-tests are two-tailed, hypothesized effects are one-ta iled. N=44,087.
117Table 5-10. Summary of results Hypothesis Statement Table/Model Result 1 Options purchased by entrepreneurial fi rms will be related to those purchased by other entrepreneurial firms Table 5-1 and 5-2 No support 2 Options purchased by established firms w ill be related to those purchased by other established firms Table 5-1 and 5-2 No support 3 As attainment discrepancy increases, option purchases will increase Table 5-6 Contrary support 4 As attainment discrepancy increases, firms will purchase options which are unrelated to current option holdings Table 5-7 No support 5 As attainment discrepancy increases, firm s will execute more options Table 5-8 Support 6 As attainment discrepancy increases, firms will execute options similar to those of comparison firms Table 5-9 No support
118 APPENDIX A LOGIT ESTIMATION OF MARKET ENTRY This appendix displays the results of estimating market entry by using a logit analysis. The logit analysis is conducted on observations in the same year or the year after the acquisition of an option on the market (t0 or t+1). The results presented below are largely consistent with those presented in Table 5-9 with an important caveat. In contrast to Table 5-9, attainment discrepa ncy has a significant effect. Attainment discrepancy (underperforming the market) reta rds entry. Firms are less likely to enter markets when they are behind their competitors. The fact that this is contrary to Tabl e 5-9 while looking at a different time period suggests that attainment discrepancy may ha ve different impacts on the firms depending on when it is observed. Here, attainment di screpancy may be an important factor is striking the option later on, but it does not encourage firms to immediately enter. Instead, firms take a wait-and-see approach to markets when they are underperforming. This analysis has been held out as an appe ndix because it confuses the results in the prior sections, but these results are not without implication.
119Appendix Table A-1. Logit estimates of the likelihood of market entry Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Competitor dissimilarity (overall)* attainment discrepancy (overall) 0.01 (0.02) Business environment dissimilarity (overall) * attainment discrepancy (overall) -0.02 (0.05) Distance dissimilarity (overall) * attainment discrepancy (overall) -0.01 (0.02) Attainment discrepancy (overall) -0.19** (0.06) -0.19** (0.06) -0.19** (0.06) -0.19** (0.06) -0.19** (0.06) Distance dissimilarity (overall) 0.04 (0.04) -0.00 (0.08) 0.04 (0.04) 0.04 (0.04) 0.04 (0.04) Business environment dissimilarity (overall) -0.24* (0.10) -0.24* (0.10) -0.26** (0.09) -0.24* (0.10) -0.24* (0.10) Competitor dissimilarity (overall) -0.14** (0.05) -0.14** (0.05) -0.14** (0.05) -0.12 (0.07) -0.14** (0.05) Count of competitors (overall) 0.09** (0.03) 0.09** (0.03) 0.09** (0.03) 0.09** (0.03) 0.09** (0.03) Count of competitors (origin) 0.02 (0.06) 0.02 (0.06) 0.02 (0.06) 0.02 (0.06) 0.02 (0.06) Growth in competitors (overall) -1.42** (0.34) -1.42** (0.34) -1.42** (0.34) -1.42** (0.34) -1.42** (0.34) Growth in competitors (origin) -1.26** (0.46) -1.26** (0.45) -1.25** (0.45) -1.26** (0.46) -1.26** (0.46) Prior year entry (overall) 0.29** (0.08) 0.29** (0.08) 0.29** (0.08) 0.29** (0.08) 0.29** (0.08) Prior year entry (origin) -0.00 (0.05) -0.00 (0.05) -0.00 (0.05) -0.00 (0.05) -0.00 (0.05) Firm age 0.02** (0.00) 0.02** (0.00) 0.02** (0.00) 0.02** (0.00) 0.02** (0.00) Market growth 0.13* (0.05) 0.13* (0.05) 0.13** (0.05) 0.13* (0.05) 0.13* (0.05) p<.10; * p<.05; ** p<.01. Robust standard errors ar e in parentheses; z-test s are two-tailed. N=15,260.
120Appendix Table A-1. Continued Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Market size -0.01 (0.01) -0.01 (0.01) -0.01 (0.01) -0.01 (0.01) -0.01 (0.01) Market variances -0.05 (0.04) -0.05 (0.04) -0.06 (0.05) -0.06 (0.04) -0.05 (0.04) Market variance growth -0.30 (0.62) -0.26 (0.62) -0.30 (0.62) -0.29 (0.62) -0.30 (0.62) Origin 0.18 (0.37) 0.18 (0.37) 0.18 (0.37) 0.18 (0.37) 0.18 (0.37) Constant -5.01** (0.59) -5.02** (0.58) -5.01** (0.58) -5.02** (0.59) -5.01** (0.59) Log pseudolikelihood -2825.25 -2824.57 -2824.85 -2825.09 -2825.25 Model 2 862.21** 896.07** 872.08** 864.25** 862.21** p<.10; * p<.05; ** p<.01. Robust standard errors ar e in parentheses; z-test s are two-tailed. N=15,260.
121 APPENDIX B ALTERNATIVE ATTAINMENT DISCREPANCY DEFINITIONS This appendix contains two tables whic h summarize an investigation into the definition of attainment discrepancy. Becau se attainment discrepancy was defined as a four factor model in Chapter 4 and 5, there is the potential that alternative definitions would alter the results presented for the hypot heses. In particular, it is likely that defining attainment discrepancy using both characteristics of markets (population and business environment) as well as characteristic s of relative market position (options and markets) has polluted the results. The charac teristics of a firmâ€™s market portfolio are quite distinct from the summary characteris tics of the firm itself. To look at the possibility that this definition has biased th e results of this study, appendix table B-1 and B-2 present the results for hypothesis 6 using two alternative measures of attainment discrepancy. The first, which was used in Appendix Tabl e B-1, looked at the changes in results from defining the hazard entry models used to explore hypothesis 6 by defining attainment discrepancy only in terms of ma rket and option differences with the industry as a whole. Hypothesis 6, which dealt w ith market entry, was not supported in the models presented in the full di ssertation. The similarity of a particular market to the marketâ€™s held by industry competitors did not encourage entry. In addition, attainment discrepancy did not interact with these variables, sugges ting that similarity does not encourage entry even under different levels of attainment discrepancy.
122 The tables presented do not change this result. Attainment discrepancy did not influence the market entry decision. In addi tion, defining attainment discrepancy only in terms of market count differences had no eff ect on the interpretation of these results. These results are shown in Appendix Table B-2.
123Appendix Table B-1. Maximum likelihood estimates of the likelihood of market entry with attainment discrepancy defined just in terms of market and option differences Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Competitor dissimilarity (overall)* attainment discrepancy (overall) -0.00 (0.02) Business environment dissimilarity (overall) * attainment discrepancy (overall) -0.04 (0.04) Distance dissimilarity (overall) * attainment discrepancy (overall) -0.03 (0.02) Attainment discrepancy (overall) 0.02 (0.12) 0.01 (0.11) 0.02 (0.12) -0.02 (0.08) Distance dissimilarity (overall) 0.05 (0.05) -0.03 (0.09) 0.05 (0.05) 0.05 (0.05) Business environment dissimilarity (overall) -0.18** (0.07) -0.18** (0.07) -0.20** (0.07) -0.19** (0.07) Competitor dissimilarity (overall) -0.14** (0.04) -0.14** (0.04) -0.14** (0.04) -0.13* (0.05) Count of competitors (overall) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) Count of competitors (origin) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) Growth in competitors (overall) -1.03** (0.33) -1.11** (0.42) -1.10** (0.42) -1.11** (0.42) -1.10** (0.42) Growth in competitors (origin) -1.27** (0.40) -1.55** (0.55) -1.55** (0.55) -1.54** (0.54) -1.54** (0.55) Prior year entry (overall) 0.17* (0.08) 0.16* (0.07) 0.16* (0.07) 0.16* (0.07) 0.16* (0.07) Prior year entry (origin) 0.08 (0.04) 0.11* (0.04) 0.11* (0.04) 0.11* (0.04) 0.11* (0.04) Firm age 0.01** (0.00) -0.02 (0.02) -0.02 (0.02) -0.02 (0.02) -0.02 (0.03) Market growth 0.01 (0.01) -0.02 (0.05) -0.02 (0.05) -0.03 (0.05) -0.02 (0.05) p<.10; * p<.05; ** p<.01. Robust standard errors ar e in parentheses; z-test s are two-tailed. N=44,087.
124Appendix Table B-1. Continued Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Market size 0.13** (0.02) 0.17** (0.03) 0.17** (0.03) 0.17** (0.04) 0.17** (0.04) Market variances -0.11** (0.04) -0.10** (0.03) -0.10** (0.03) -0.10** (0.03) -0.11** (0.04) Market variance growth -2.99 (1.76) -2.66 (1.94) -2.55 (1.91) -2.63 (1.92) -2.65 (1.93) Constant -5.96** (0.29) -5.79** (0.37) -5.79** (0.37) -5.80** (0.38) -5.81** (0.39) Log pseudolikelihood -2245.98 -1540.29 -1538.7 -1539.16 -1540.22 Model 2 528.28** 675.41** 744.42** 674.14** 741.32** p<.10; * p<.05; ** p<.01. Robust standard errors ar e in parentheses; z-test s are two-tailed. N=44,087.
125Appendix Table B-2. Maximum likelihood estimates of the likelihood of market entry with attainment discrepancy defined just in terms of market differences Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Competitor dissimilarity (overall)* attainment discrepancy (overall) -0.00 (0.02) Business environment dissimilarity (overall) * attainment discrepancy (overall) -0.04 (0.04) Distance dissimilarity (overall) * attainment discrepancy (overall) -0.03 (0.02) Attainment discrepancy (overall) 0.02 (0.12) 0.01 (0.11) 0.02 (0.12) -0.02 (0.08) Distance dissimilarity (overall) 0.05 (0.05) -0.03 (0.09) 0.05 (0.05) 0.05 (0.05) Business environment dissimilarity (overall) -0.18** (0.07) -0.18** (0.07) -0.20** (0.07) -0.19** (0.07) Competitor dissimilarity (overall) -0.14** (0.04) -0.14** (0.04) -0.14** (0.04) -0.13* (0.05) Count of competitors (overall) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) 0.06 (0.04) Count of competitors (origin) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) 0.01 (0.08) Growth in competitors (overall) -1.03** (0.33) -1.11** (0.42) -1.10** (0.42) -1.11** (0.42) -1.10** (0.42) Growth in competitors (origin) -1.27** (0.40) -1.55** (0.55) -1.55** (0.55) -1.54** (0.54) -1.54** (0.55) Prior year entry (overall) 0.17* (0.08) 0.16* (0.07) 0.16* (0.07) 0.16* (0.07) 0.16* (0.07) Prior year entry (origin) 0.08 (0.04) 0.11* (0.04) 0.11* (0.04) 0.11* (0.04) 0.11* (0.04) Firm age 0.01** (0.00) -0.02 (0.02) -0.02 (0.02) -0.02 (0.02) -0.02 (0.03) Market growth 0.01 (0.01) -0.02 (0.05) -0.02 (0.05) -0.03 (0.05) -0.02 (0.05) p<.10; * p<.05; ** p<.01. Robust standard errors ar e in parentheses; z-test s are two-tailed. N=44,087.
126Appendix Table B-2. Continued Variable Names Model 1 Mode l 2 Model 3 Model 4 Model 5 Market size 0.13** (0.02) 0.17** (0.03) 0.17** (0.03) 0.17** (0.04) 0.17** (0.04) Market variances -0.11** (0.04) -0.10** (0.03) -0.10** (0.03) -0.10** (0.03) -0.11** (0.04) Market variance growth -2.99 (1.76) -2.66 (1.94) -2.55 (1.91) -2.63 (1.92) -2.65 (1.93) Constant -5.96** (0.29) -5.79** (0.37) -5.79** (0.37) -5.80** (0.38) -5.81** (0.39) Log pseudolikelihood -2245.98 -1540.29 -1538.7 -1539.16 -1540.22 Model 2 528.28** 675.41** 744.42** 674.14** 741.32** p<.10; * p<.05; ** p<.01. Robust standard errors ar e in parentheses; z-test s are two-tailed. N=44,087.
127 LIST OF REFERENCES Abrahamson, E. 1991. Managerial Fads and Fashions: The Diffusion and Rejection of Innovations. Academy of Management Review , 16(3): 586-612. Adner, R. & Levinthal, D. A. 2004. What is not a real option: Considering boundaries for the application of real op tions to business strategy. Academy of Management Review , 29(1): 74-85. Alexander, D. L. & Feinberg, R. M. 2004. Entry in local telecommunication markets. Review of Industrial Organization , 25(2): 107-127. Barney, J. 1991. Firm Resources an d Sustained Competitive Advantage. Journal of Management , 17(1): 99-120. Barney, J. B. 1986. Strategic Factor Markets Expectations, Luck, and Business Strategy. Management Science , 32(10): 1230-1241. Baum, J. A. C. & Lant, T. K. 2003. Hits a nd misses: Managers' (mis)categorization of competitors in the Manhattan hotel industry. Advances in Strategic Management , 20: 119-156. Black, F. & Scholes, M. 1973. Pricing of Options and Corporate Liabilities. Journal of Political Economy , 81(3): 637-654. Bowman, E. H. & Hurry, D. 1993. Strategy throu gh the Option Lens an Integrated View of Resource Investments and th e Incremental-Choice Process. Academy of Management Review , 18(4): 760-782. Bowman, E. H. & Moskowitz, G. T. 2001. Real options analysis and strategic decision making. Organization Science , 12(6): 772-777. Boyes, W. J., Hoffman, D. L., & Low, S. A. 1989. An Econometric-Analysis of the Bank Credit Scoring Problem. Journal of Econometrics , 40(1): 3-14. Bresnahan, T. F. & Reiss, P. C. 1991. Entry and Competition in Con centrated Markets. Journal of Political Economy , 99(5): 977-1009. Bromiley, P. 1991. Testing a Causal Model of Corporate Risk-Taking and Performance. Academy of Management Journal , 34(1): 37-59.
128 Brown, K. S. & Zimmerman, P. R. 2004. The effect of Section 271 on competitive entry into local telecommunications ma rkets: an init ial evaluation. Information Economics and Policy , 16(2): 215-233. Carroll, G. R. 1993. A Sociological View on Why Firms Differ. Strategic Management Journal , 14(4): 237-249. Carroll, G. R., Bigelow, L. S., Seidel, M. D. L., & Tsai, L. B. 1996. The fates of de novo and de alio producers in the Amer ican automobile industry 1885-1981. Strategic Management Journal , 17: 117-137. Chance, D. & Peterson, P. 2006. The Scientific Evolution of Finance. Retrieved from http://www.finance-andphysics.org/Library/Articles3/sciencea ndfinance/science.htm on February 28, 2006. Chang, S. J. 1996. An evolutionary perspe ctive on diversific ation and corporate restructuring: Entry, exit, and economic performance during 1981-89. Strategic Management Journal , 17(8): 587-611. Chen, M. J. & Hambrick, D. C. 1995. Speed, Stealth, and Selective Attack How Small Firms Differ from Large Firms in Competitive Behavior. Academy of Management Journal , 38(2): 453-482. Chen, W. R. 2003. Determinants of Technol ogical Search Behavior of the Firm. Academy of Management Proceedings : T1-T6. Cyert, R. M. & March, J. G. 1963. A behavioral theory of the firm . Englewood Cliffs, N.J.: Prentice-Hall. D'Aveni, R. A. & Gunther, R. E. 1994. Hypercompetition : managing the dynamics of strategic maneuvering . Maxwell Macmillan. New York, New York. Dierickx, I. & Cool, K. 1989. Asset Stoc k Accumulation and Sustainability of Competitive Advantage. Management Science , 35(12): 1504-1511. DiMaggio, P. J. & Powell, W. W. 1983. The Iron Cage Revisited Institutional Isomorphism and Collective Rationality in Organizational Fields. American Sociological Review , 48(2): 147-160. Dixit, A. 1992. Investment and Hysteresis. Journal of Economic Perspectives , 6(1): 107132. Doonar, J. 2004. McDonalds's: Life in the fast lane, Brand Strategy , Vol. 20.
129 Folta, T. & O'Brien, J. 2005. Firm-S pecific Performance Thresholds, University of Purdue Working Paper . West Layfette, IN. Folta, T. B. & O'Brien, J. P. 2004. Entry in the presence of dueling options. Strategic Management Journal , 25(2): 121-138. Garg, V. K., Walters, B. A., & Priem, R. L. 2003. Chief executive scanning emphases, environmental dynamism, and manufacturing firm performance. Strategic Management Journal , 24(8): 725-744. Gavetti, G. & Levinthal, D. 2000. Looking fo rward and looking backward: Cognitive and experiential search. Administrative Science Quarterly , 45(1): 113-137. Gimeno, J., Folta, T. B., Cooper, A. C., & Woo, C. Y. 1997. Survival of the fittest? Entrepreneurial human capital and the persistence of underperforming firms. Administrative Science Quarterly , 42(4): 750-783. Glynn, M. A., Lant, T. K., & Mezias, S. J. 1991. Incrementalism, Learning, and Ambiguity: An Experimental Study of Aspiration Level Adaptation. Academy of Management Proceedings : 384-389. Graham, C. M., Cannice, M. V., & Sayre, T. L. 2002. The Value-Relevance of Financial and Non-financial Information for Internet Companies. Thunderbird International Business Review, , 44(1): 47. Gresov, C., Haveman, H. A., & Oliva, T. A. 1993. Organizational Design, Inertia and the Dynamics of Competitive Response. Organization Science , 4(2): 181-208. Greve, H. R. 1995. Jumping Ship th e Diffusion of Strategy Abandonment. Administrative Science Quarterly , 40(3): 444-473. Greve, H. R. 1996. Patterns of competition: The diffusion of a mark et position in radio broadcasting. Administrative Science Quarterly , 41(1): 29-60. Greve, H. R. 1998a. Managerial cognition and the mimetic adoption of market positions: What you see is what you do. Strategic Management Journal , 19(10): 967-988. Greve, H. R. 1998b. Performance, aspira tions, and risky organizational change. Administrative Science Quarterly , 43(1): 58-86. Greve, H. R. 2000. Market niche entry deci sions: Competition, learning, and strategy in Tokyo banking, 1894-1936. Academy of Management Journal , 43(5): 816-836. Greve, H. R. 2002. Sticky aspirations: Organizational time perspective and competitiveness. Organization Science , 13(1): 1-17.
130 Greve, H. R. 2003a. A behavioral theory of R&D expenditures and innovations: Evidence from shipbuilding. Academy of Management Journal , 46(6): 685-702. Greve, H. R. 2003b. Organizational learning from perf ormance feedback : a behavioral perspective on innovation and change . Cambridge: Cambridge University Press. Greve, H. R. & Taylor, A. 2000. Innovations as catalysts for organizational change: Shifts in organizational cognition and search. Administrative Science Quarterly , 45(1): 54-80. Grossman, P. Z. 1987. American Express : the unofficial history of the people who built the great financial empire (1st ed.). New York: Crown Publishers. Guler, I., Guillen, M. F., & MacPherson, J. M. 2002. Global competition, institutions, and the diffusion of organizational practices : The international spread of ISO 9000 quality certificates. Administrative Science Quarterly , 47(2): 207-232. Haunschild, P. R. 1993. Interorganizational Imitation the Impact of Interlocks on Corporate Acquisition Activity. Administrative Science Quarterly , 38(4): 564592. Haunschild, P. R. & Miner, A. S. 1997. Mode s of interorganiza tional imitation: The effects of outcome salience and uncertainty. Administrative Science Quarterly , 42(3): 472-500. Haveman, H. A. 1992. Between a Rock and a Hard Place Organizational-Change and Performance under Conditions of Fundam ental Environmental Transformation. Administrative Science Quarterly , 37(1): 48-75. Haveman, H. A. 1993. Follow the Leader Mimetic Isomorphism and Entry into New Markets. Administrative Science Quarterly , 38(4): 593-627. Henisz, W. J. & Delios, A. 2001. Uncertaint y, imitation, and plant location: Japanese multinational corporations, 1990-1996. Administrative Science Quarterly , 46(3): 443-475. Hitt, M. A., Bierman, L., Shimizu, K., & Kochhar, R. 2001. Direct and moderating effects of human capital on strategy and performance in pr ofessional service firms: A resource-based perspective. Academy of Management Journal , 44(1): 13-28. Hoopes, D. G., Madsen, T. L., & Walker, G. 2003. Guest editors' introduction to the special issue: Why is there a resour ce-based view? Toward a theory of competitive heterogeneity. Strategic Management Journal , 24(10): 889-902.
131 Huff, J. O., Huff, A. S., & Thomas, H. 1992. Strategic Renewal and the Interaction of Cumulative Stress and Inertia. Strategic Management Journal , 13: 55-75. Kahneman, D. & Tversky, A. 1979. Prospect Th eory Analysis of Decision under Risk. Econometrica , 47(2): 263-291. Kalnins, A. & Chung, W. 2004. Resource-seeki ng agglomeration: A study of market entry in the lodging industry. Strategic Management Journal , 25(7): 689-699. Klepper, S. & Simons, K. L. 2000. Dominance by birthrig ht: Entry of prior radio producers and competitive ramifications in the US television receiver industry. Strategic Management Journal , 21(10-11): 997-1016. Kogut, B. 1991. Joint Ventures and the Option to Expand and Acquire. Management Science , 37(1): 19-33. Kogut, B. & Kulatilaka, N. 2001. Ca pabilities as real options. Organization Science , 12(6): 744-758. Korn, H. J. & Baum, J. A. C. 1999. Chance, imitative, and strategic antecedents to multimarket contact. Academy of Management Journal , 42(2): 171-193. Lant, T. K. & Montgomery, D. B. 1987. Learni ng from Strategic Success and Failure. Journal of Business Research , 15(6): 503-517. Lant, T. K. 1992. Aspiration Level Adap tation an Empirical Exploration. Management Science , 38(5): 623-644. Levinthal, D. A. & March, J. G. 1993. The Myopia of Learning. Strategic Management Journal , 14: 95-112. Lippman, S. A. & Rumelt, R. P. 2003. The pa yments perspective: Micro-foundations of resource analysis. Strategic Management Journal , 24(10): 903-927. Locke, E. A., Saari, L. M., Shaw, K. N., & Latham, G. P. 1981. Goal Setting and TaskPerformance 1969-1980. Psychological Bulletin , 90(1): 125-152. March, J. G. & Simon, H. A. 1958. Organizations . New York,: Wiley. March, J. G. & Shapira, Z. 1987. Managerial Perspectives on Risk and Risk-Taking. Management Science , 33(11): 1404-1418. March, J. G. 1991. Exploration and Expl oitation in Organizational Learning. Organization Science , 2(1): 71-87.
132 March, J. G. & Shapira, Z. 1992. Variable Risk Preferences and the Focus of Attention. Psychological Review , 99(1): 172-183. Massini, S., Lewin, A. Y., & Greve, H. R. 2005. Innovators and imitators: Organizational reference groups and adoption of organizational routines. Research Policy , 34(10): 1550-1569. McDonald, M. L. & Westphal, J. D. 2003. Ge tting by with the advice of their friends: CEOs' advice networks and firms' stra tegic responses to poor performance. Administrative Science Quarterly , 48(1): 1-32. McDonald, R. & Siegel, D. 1986. Th e Value of Waiting to Invest. Quarterly Journal of Economics , 101(4): 707-727. McGrath, R. G. 1997. A real options logic for initia ting technology positioning investments. Academy of Management Review , 22(4): 974-996. McGrath, R. G. 1999. Falling forward: Real options reasoning and entrepreneurial failure. Academy of Management Review , 24(1): 13-30. McGrath, R. G., Ferrier, W. J., & Mendelow, A. L. 2004. Real options as engines of choice and heterogeneity. Academy of Management Review , 29(1): 86-101. McGrath, R. G. & Nerkar, A. 2004. Real op tions reasoning and a new look at the R&D investment strategies of pharmaceutical firms. Strategic Management Journal , 25(1): 1-21. Miller, K. D. & Leiblein, M. J. 1996. Corporat e risk-return relations : Returns variability versus downside risk. Academy of Management Journal , 39(1): 91-122. Miller, K. D. & Folta, T. B. 2002. Option value and entry timing. Strategic Management Journal , 23(7): 655-665. Miller, K. D. & Arikan, A. T. 2004. Technology search investments: Evolutionary, option reasoning, and option pricing approaches. Strategic Management Journal , 25(5): 473-485. Miller, K. D. & Chen, W. R. 2004. Variable or ganizational risk prefer ences: Tests of the March-Shapira model. Academy of Management Journal , 47(1): 105-115. Miller, K. D. & Shapira, Z. 2004. An empirical te st of heuristics and biases affecting real option valuation. Strategic Management Journal , 25(3): 269-284. Mitchell, W. 1989. Whether and When Proba bility and Timing of Incumbents Entry into Emerging Industrial Subfields. Administrative Science Quarterly , 34(2): 208-230.
133 Mullins, J. W. & Forlani, D. 2005. Missing th e boat or sinking the boat: a study of new venture decision making. Journal of Business Venturing , 20(1): 47-69. O'Brien, J. P., Folta, T. B., & Johnson, D. R. 2003. A real options perspective on entrepreneurial entry in the face of uncertainty. Managerial and Decision Economics , 24(8): 515-533. Park, S. H., Chen, R., & Gallagher, S. 2002. Firm resources as moderators of the relationship between market growth and st rategic alliances in semiconductor startups. Academy of Management Journal , 45(3): 527-545. Payne, J. W., Laughhunn, D. J., & Crum, R. 1980. Translation of Gambles and Aspiration Level Effects in Risky Choice Behavior. Management Science , 26(10): 1039-1060. Rangos, K. & Lynch, K. 2001. Telecommunication Provider Locator . Federal Communications Commission, 2001. Reger, R. K. & Huff, A. S. 1993. Strate gic Groups a Cogni tive Perspective. Strategic Management Journal , 14(2): 103-123. Romanelli, E. & Tushman, M. L. 1994. Organi zational Transformation as Punctuated Equilibrium an Empirical-Test. Academy of Management Journal , 37(5): 11411166. Ruefli, T. W. 1990. Mean-Variance Approaches to Risk-Return Relationships in Strategy Paradox Lost. Management Science , 36(3): 368-380. Shapira, Z. 1995. Risk taking : a managerial perspective . New York, N.Y.: Russell Sage Foundation. Simon, M., Houghton, S. M., & Savelli, S. 2003. Out of the frying pan ...? Why small business managers introduce high-risk products. Journal of Business Venturing , 18(3): 419-440. Sitkin, S. B. & Pablo, A. L. 1992. Reconceptua lizing the Determinants of Risk Behavior. Academy of Management Review , 17(1): 9-38. Smith, K. G., Grimm, C. M., Gannon, M. J., & Chen, M. J. 1991. Organizational Information-Processing, Competitive Res ponses, and Performance in the UnitedStates Domestic Airline Industry. Academy of Management Journal , 34(1): 6085.
134 Staw, B. M., Sandelands, L. E., & Dutton, J. E. 1981. Threat-Rigidity Effects in Organizational-Behavior a Multilevel Analysis. Administrative Science Quarterly , 26(4): 501-524. Stuart, T. E. & Podolny, J. M. 1996. Local se arch and the evolution of technological capabilities. Strategic Management Journal , 17: 21-38. Stuart, T. E. & Sorenson, O. 2003. Liquidity events and the geogra phic distribution of entrepreneurial activity. Administrative Science Quarterly , 48(2): 175-201. Teece, D. J., Pisano, G., & Shuen, A. 1997. Dynamic capabilities and strategic management. Strategic Management Journal , 18(7): 509-533. Teisberg, E. O. 1993. Capital investment strategies under uncertain regulation. RAND Journal of Economics , 24(4): 591-604. Trigeorgis, L. 1993. Real Options and Inte ractions with Financial Flexibility. Financial Management , 22(3): 202-224. Tushman, M. L. & Anderson, P. 1986. Technolog ical Discontinuities and Organizational Environments. Administrative Science Quarterly , 31(3): 439-465. Vassolo, R. S., Anand, J., & Folta, T. B. 2004. Non-additivity in portfolios of exploration activities: A real op tions-based analysis of equ ity alliances in biotechnology. Strategic Management Journal , 25(11): 1045-1061. Wadhwa, A. & Kotha, S. In press. K nowledge Creation through External Venturing: Evidence from the Telecommunication Equipment Manufac turing Industry. Forthcoming at Academy of Management Journal. Wernerfelt, B. 1984. A Resource-Based View of the Firm. Strategic Management Journal , 5(2): 171-180. Westphal, J. D., Seidel, M.-D. L., & St ewart, K. J. 2001. Second-order imitation: Uncovering latent effects of board network ties. Administrative Science Quarterly , 46(4): 717-747. Wilsted, W. D. & Hand, H. H. 1974. Determinan ts of Aspiration Levels in a Simulated Goal Setting Environment of Firm. Academy of Management Journal , 17(1): 172-177. Winter, S. G. 2000. The satisficing pr inciple in capability learning. Strategic Management Journal , 21(10-11): 981-996.
135 BIOGRAPHICAL SKETCH Rich was born in Jacksonville, FL on August 12, 1977. He completed primary education in St. Augustine, FL. Rich was a competitive international shotgun shooter for many years and passed up an opportunity to sh oot professionally with the United States Army to begin his studies in industrial engineer ing at The University of Florida. After a short time studying international business at the Rijksuniversite it Groningen in the northern portion of Holland, a summer intern ship with the Ford Motor Company, and another with Asea Brown and Boveri, he comp leted his bachelorâ€™s degree in industrial engineering along with a Masterâ€™s degree in Business Administration in 2001. Wishing to do more meaningful work than was possibl e in industry, Rich decided to return the university to pursue his PhD. In the fall of 2001, he returned to the University of Florida for his graduate studies in strategic management at the Warrington College of Business. During his time as a PhD candidate, Richar d was awarded a Kauffman Dissertation Fellowship to pursue work in entrepreneursh ip and was recognized as one of five outstanding student researchers at the Strategic Management Societyâ€™s Annual Meetings in 2005. Through is strong connections with th e Public Utility Rese arch Center, Rich also has lectured in London and Nigeria on st rategic management. Richâ€™s first job after graduate school is as an a ssistant professor of strategy a nd entrepreneurship at West Virginia University.